Abstract

Background. Heat shock protein 90 (HSP90) appears to have a pivotal function in ischemic preconditioning. Pioglitazone preconditioning (PioC) attenuates myocardial ischemia/reperfusion (I/R) injuries.

Objectives. The current study aims to investigate the role of HSP90, complement C3 and C5a, and nuclear factor kappa-B (NF-κB) in PioC-induced cardioprotection.

Materials and methods. A total of 80 rats were randomly categorized into 4 groups, as follows: sham, I/R, PioC, and PioC+HSP90 inhibitor geldanamycin (PioC+GA). The sham group rats had a thoracotomy, in which the ligature was passed by the heart with no ligation (150 min). The other 3 groups were exposed to ischemia (30 min) followed by reperfusion (2 h). In the PioC group, pioglitazone (3 mg/kg) was administered intravenously 24 h before ischemia. In the PioC+GA group, after being pretreated with pioglitazone, GA was administered (intraperitoneally, 1 mg/kg) 30 min before ischemia. Myocardial infarct sizes (ISs), apoptosis rates, creatine kinase-MB (CK-MB), lactate dehydrogenase (LDH), and cardiac troponin I (cTnI) serum levels were determined. The HSP90, C3, NF-κB, C5a, B-cell lymphoma-2 (Bcl-2), and Bax expression levels, as well as interleukin (IL)-1β, IL-6, intercellular cell adhesion molecule-1 (ICAM-1), and tumor necrosis factor alpha (TNF-α) mRNA levels were measured.

Results. The myocardial ISs, serum CK-MB, cTnI and LDH levels, apoptosis rates, IL-1β, IL-6, TNF-α, ICAM-1 release, as well as Bax, C5a, C3, and NF-κB protein expression were considerably lower in the PioC group than in the I/R group (p < 0.05). The Bcl-2 and HSP90 expression was higher in the PioC group than in the I/R group (p < 0.05). Geldanamycin inhibited the effects of PioC. These data strongly demonstrate that the PioC-induced is dependent upon HSP90 activity.

Conclusions. The HSP90 is indispensable for PioC-mediated cardioprotection. The HSP90 attenuates I/R-induced ISs, apoptosis of cardiomyocytes and myocardial inflammation through C3, C5a and NF-κB activation inhibition.

Key words: NF-κB, C3, HSP90, C5a, pioglitazone preconditioning

Background

Myocardial ischemia/reperfusion (I/R) injury is a serious acute myocardial infarction (AMI) complication after revascularization. Research has focused on methods and strategies to reduce myocardial I/R injury.¹ The cardioprotective impact of peroxisome proliferator-activated receptor gamma (PPAR-γ) has gained increasing interest in this area.^{2, 3} Pioglitazone belongs to the thiazolidinedione class and is a second-line drug in the treatment of type 2 diabetes acting by activating PPAR-γ to increase insulin sensitivity and improve glucose and lipid metabolism.^{4, 5} Pioglitazone exerts beneficial cardiovascular outcomes influencing cardiac function, metabolism, cardiac remodeling, etc.^{6, 7} In a previous clinical trial, I/R injury was reduced in diabetic patients after AMI who were pretreated with pioglitazone.⁸ Previous observations have shown that pioglitazone-induced cardioprotection may activate extracellular signal-regulated kinase (ERK1/2) signaling pathways using cyclooxygenase-2 (COX-2) as the downstream target, depend on endothelial nitric oxide synthase (eNOS) and inducible NOS (iNOS), activate Akt, and increase the expression of cytosolic phospholipase A2 (cPLA2).^{5, 9, 10} Nevertheless, the cardioprotective mechanism underlying pioglitazone in I/R injuries has not been fully elucidated.

The complement system, a key player in innate immunity, causes inflammation by recruiting immune cells and inducing the release of inflammatory factors during myocardial injury. Activation of the complement cascade in I/R injury causes myocardial tissue damage.¹¹ Emerging evidence indicates that complement components C3 and C5a promote the release of inflammatory factors by inducing nuclear factor kappa-B (NF-κB) signaling, which further aggravates myocardial I/R injury.^{12, 13} According to the literature, ischemic preconditioning impedes the upregulation of complement C3 and C5a induced by I/R injury in vivo.¹⁴ There is a paucity of information regarding pioglitazone preconditioning (PioC)-induced cardioprotection in terms of its inhibitory effects on C3, C5a and NF-κB.

Heat shock protein 90 (HSP90) is a highly conserved molecular chaperone and a vital effector molecule for cardioprotection during ischemic preconditioning.¹⁵ The inhibition of HSP90 aggravates complement-mediated cell lysis.¹⁶ The HSP90 can bind to molecular chaperones, which are resistant to complement-induced cell death.¹⁷ In our recent study, we demonstrated a correlation between HSP90, C5a, C3, and NF-κB signaling in ischemic postconditioning.^{18, 19} However, whether HSP90 contributes to the protective effects of PioC in myocardial I/R injuries by suppressing C3, C5a and NF-κB remains unclear. Therefore, in the present study, we investigated the roles of HSP90, C3, C5a, and NF-κB in PioC-induced cardioprotection.

An I/R injury rat model was constructed to study the anti-inflammatory impact of PioC. We assessed interleukin (IL)-1β, IL-6, intercellular cell adhesion molecule 1 (ICAM-1), and tumor necrosis factor alpha (TNF-α) messenger RNA (mRNA) levels. Western blot analysis was conducted to determine HSP90, C3, C5a, B-cell lymphoma-2 (Bcl-2), NF-κB, and Bax proteins levels in the sham, I/R, PioC, and PioC+geldanamycin (GA) groups.

Objectives

The current study aims to investigate the role of HSP90, complement C3 and C5a, and NF-κB in PioC-induced cardioprotection.

Materials and methods

Animals

Guangxi Medical University Experimental Animal Center (certificate No. SYXK (Gui) 2020-0001; Nanning, China) provided 80 adult male Sprague Dawley rats (8 weeks old) weighing between 250–280 g and kept in a 12:12-hour light/dark cycle, 50 ±15% humidity and 25 ±2°C temperature conditions with ad libitum access to food and water. The National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals was followed. The Guangxi Medical University Animal Protection and Use Committee approved this study (approval No. 201909028).

Experimental model

Anesthesia was provided using intraperitoneal sodium pentobarbital injection (50 mg/kg). A small animal ventilator (ALC-V8; Alcott Biotech Co., Shanghai, China) was used to ventilate the animals. The heart was exposed by opening the chest through the left 5^th intercostal area. Suture ligation of the left anterior descending (LAD) coronary artery was used for 30 min to create myocardial ischemia. The suture was removed, and reperfusion was performed for 2 h. An ST-segment elevation on electrocardiogram (ECG) and myocardium whitening in the LAD blood supply area indicated the successful construction of the model. The rats were euthanized immediately after the reperfusion. Blood samples were drawn from the left ventricular anterior wall close to the ventricular myocardium apex for further research.

Experimental groups

The rats were randomly divided into 4 groups (n = 20 per group): 1) sham group, where a ligature was passed around the LAD with no ligation (150 min); 2) I/R group, where the rats had 30 min of ischemia and then 120 min of reperfusion; 3) PioC group, where pioglitazone (3 mg/kg²⁰) was administered intravenously 24 h before ischemia, followed by 30 min of ischemia and then 120 min of reperfusion; 4) PioC+GA group, where pioglitazone (3 mg/kg²⁰) was administered intravenously 24 h before ischemia, and 30 min before ischemia, an intraperitoneal injection of GA (1 mg/kg²¹) was administered, followed by 30 min of ischemia and 120 min of reperfusion.

Myocardial infarct size

Each group was tested separately to calculate myocardial infarct size (IS) (n = 5). After reperfusion, we tightened the LAD and stained the inferior vena cava with 2% Evans blue dye. The IS was determined after the stained heart was frozen, sectioned into 2-millimeter pieces, and treated for 15 min with 1% 2,3,5-triphenyltetrazolium chloride (TTC; Sigma-Aldrich, St. Louis, USA) at 37°C. Using digital imaging software, the IS as a percentage of the left ventricle (LV) was calculated quantitatively.

Lactate dehydrogenase, creatine kinase-MB and cardiac troponin I plasma levels

The serum was obtained by collecting 5-milliliter blood samples and centrifuging them at 3000 × g. The cardiac troponin I (cTnI), creatine kinase-MB (CK-MB) and lactate dehydrogenase (LDH) serum concentrations were determined utilizing respective enzyme-linked immunosorbent assay (ELISA) kits (cTnI: CSB-E08594r, CK-MB: CSB-E14403r, LDH: CSB-E11324r; all acquired from CUSABIO, Wuhan, China) according to the manufacturer’s instructions.

TUNEL staining

Cardiomyocyte apoptosis was assessed using the terminal deoxynucleotidyl transferase-mediated dUTP-X nick end labeling (TUNEL) detection kit. Stained samples were visualized under a light microscope (model CKX41SF; Olympus Corp., Tokyo, Japan). A minimum of 5 randomly selected fields containing apoptotic cells were scored and recorded. Apoptosis was indicated by brown-stained nuclei (TUNEL-positive cells) and its index was computed using this formula: the number of TUNEL-positive cells / total number of myocytes × 100%.

mRNA expression quantification

Trizol reagent (Invitrogen, Carlsbad, USA) was used to extract 1 µg of total RNA from the myocardial tissue in the ischemic areas, and it was purified. The PrimeScript™ RT reagent Kit (Takara, Kusatsu, Japan) synthesized complement DNA (cDNA) and quantitative polymerase chain reaction (qPCR) was performed using the SYBR standard qPCR mix (Takara) on an ABI Prism 7500 System (Thermo Fisher Scientific, Waltham, USA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-actin served as the housekeeping genes. Statistics for the β-actin are presented in the Supplementary material (https://doi.org/10.5281/zenodo.7766218). The 2^−ΔΔCT method was utilized for analyzing relative mRNA expression. The following primers were used:

IL-1β gene:

forward: 5’-TAGCAGCTTTCGACAGTGAGG-3’,

reverse: 5’-CCACAGCCACAATGAGTGAC-3’;

IL-6 gene:

forward: 5’-CTGGTCTTCTGGAGTTCCGTT-3’,

reverse: 5’-GCATTGGAAGTTGGGGTAGGA-3’;

TNF-α gene:

forward: 5’-ATGGGCTCCCTCTCATCAGT-3’,

reverse: 5’-GCTTGGTGGTTTGCTACGAC-3’;

and ICAM-1 gene:

forward: 5’-TGTCGGTGCTCAGGTATCCATCC-3’,

reverse: 5’-TTCGCAAGAGGAAGAGCAGTTCAC-3’.

Western blot

After reperfusion, myocardial tissue from the ischemic areas was lysed with radioimmunoprecipitation assay (RIPA) lysis buffer (Solarbio, Beijing, China), treated with an ultrasonic tissue homogenizer (Servicebio, Wuhan, China), and then centrifuged for 15 min at 4°C and 12,000 × g. The enhanced bicinchoninic acid (BCA) protein assay kit (Beyotime Biotechnology, Shanghai, China) was used to determine protein concentrations. Equal protein amounts were subjected to gel electrophoresis before being moved to the polyvinylidene difluoride (PVDF) membranes (Merck Millipore, Bedford, USA), which were blocked at room temperature using 5% bovine serum albumin for 1 h before being treated with the primary antibodies against HSP90 (1:5000; Proteintech, Chicago, USA), Bcl-2 (1:500, Proteintech), Bax (1:1000, Cell Signaling Technology, Danvers, USA), β-actin (1:5000; Proteintech), C5a (1:5000; Invitrogen), NF-κB p65 (1:1000; Invitrogen), and C3 (1:1000; Abcam, Cambridge, UK) for 1 full day at 4°C, and then washed by Tween 20 and treated with the horseradish peroxidase (HRP)-labeled goat anti-rabbit immunoglobulin (1:12000) for 1 h. For the detection of protein bands, the FluorChemFC3 imaging system (ProteinSimple, Santa Clara, USA) and ImageJ v. 1.8.0 software (National Institutes of Health, Bethesda, USA) were used.

Statistical analyses

The IBM SPSS v. 23.0 software (IBM Corp., Armonk, USA) was utilized to analyze the data, which were presented as mean ± standard deviation (M ±SD) for normally distributed variables. The Shapiro–Wilk test was used for the normal distribution of continuous variables. A one-way variance analysis (ANOVA) was utilized to compare several groups. The Tukey’s honestly significant difference (HSD) test was performed as a post hoc test. A value of p < 0.05 was considered statistically significant.

Results

Two animals died due to ventricular fibrillation and machine malfunction – 1 in the I/R group and 1 in the PioC+GA group. The final results correspond to the data obtained on 78 rats.

PioC upregulated HSP90

To explore the potential link between HSP90 and PioC, we identified the HSP90 expression in myocardial tissue for each group. The PioC group had a higher HSP90 protein level than the I/R group (84.60 ±3.31% compared to 69.13 ±5.58%, p < 0.001, ANOVA, F = 57.2; Figure 1 and Table 1, Table 2). Notably, treatment with the HSP90 inhibitor, GA, counteracted the effects of PioC.

PioC alleviated I/R-induced myocardial IS via HSP90

The PioC group had lower IS/LV rate than the I/R group (21.12 ±3.12% compared to 43.00 ±3.40%, p < 0.001, ANOVA, F = 65.6; Figure 2 and Table 3, Table 4). Geldanamycin reversed the PioC protective effects (21.12 ±3.12% compared to 42.68 ±3.83%, p < 0.001, ANOVA, F = 65.6). Neither the I/R group nor the PioC+GA group had a statistically significant difference in IS/LV (p = 0.988, ANOVA, F = 65.6).

PioC alleviated I/R-induced cardiomyocytes apoptosis via HSP90

The cardiomyocyte apoptotic index was examined to determine whether HSP90 has a role in PioC anti-apoptosis. The PioC group had lower apoptosis rates than the I/R group (23.71 ±2.08% compared to 45.98 ±1.70%, p < 0.001, ANOVA, F = 866.0; Figure 3A and Table 5, Table 6). The PioC group had higher Bcl-2 and lower Bax expression levels than the I/R group (66.69 ±3.83% compared to 22.46 ±1.71%, p < 0.001, ANOVA, F = 332.9; 54.09 ±5.26% compared to 71.97 ±3.75%, p < 0.001, ANOVA, F = 141.0; Figure 3B and Table 5, Table 6), respectively. Geldanamycin inhibited the PioC anti-apoptotic effect on cardiomyocytes. These data suggest that HSP90 could effectively alleviate cardiomyocyte apoptosis and play a critical role in the anti-apoptotic effect of PioC.

PioC alleviated myocardial injury
after I/R via HSP90

In comparison with the I/R group, CK-MB (1271.2 ±89.66 U/L compared to 828.6 ±53.13 U/L), LDH (1608.6 ±101.34 U/L compared to 824.0 ±22.36 U/L) and cTnI (325.44 ±34.67 ng/mL compared to 138.79 ±7.40 ng/mL) levels were considerably lower after I/R in the PioC group (p < 0.001, ANOVA, F = 168.9; p < 0.001, ANOVA, F = 104.0; p < 0.001, ANOVA, F = 160.8, respectively (Table 7, Table 8). Geldanamycin reversed the beneficial effect of PioC in mitigating myocardial injury.

PioC alleviated I/R-induced activation of complement and NF-κB via HSP90

To explore whether PioC modulated the complement system and NF-κB signaling pathways through HSP90, the complement components C3 and C5a and NF-κB expression were assessed in the presence of GA. The complement components C3 and C5a and NF-κB expression were assessed. The C3, C5a and NF-κB protein levels in the PioC group were lower than in the I/R group (65.03 ±3.32% compared to 78.62 ±3.04%, p = 0.001, ANOVA, F = 105.2; 57.31 ±4.71% compared to 75.35 ±5.49%, p < 0.001, ANOVA, F = 190.8; and 47.98 ±9.34% compared to 69.35 ±2.75%, p < 0.001, ANOVA, F = 72.4, respectively, Figure 4 and Table 9, Table 10). The mRNA levels of TNF-α, IL-1β, IL-6, and ICAM-1 in the PioC group were significantly lower than in the I/R group (p < 0.001, ANOVA, F = 108.1; p < 0.001, ANOVA, F = 38.1; p < 0.001, ANOVA, F = 163.9; p = 0.017, ANOVA, F = 13.5, respectively; Figure 4B and Table 9, Table 10). Geldanamycin negated the anti-inflammatory effects for PioC. These data demonstrate that PioC protects the heart from I/R-induced inflammatory responses by suppressing the activation of the complement system and NF-κB. Furthermore, the effects of PioC were closely linked to HSP90 activity.

Discussion

The major finding of this study was that PioC significantly decreased the I/R-induced activation of complement components C5a and C3 and NF-κB through HSP90. The PioC significantly alleviated I/R-induced myocardial injury, increased Bcl-2 expression and increased HSP90 expression. In rat hearts treated with GA, the cardioprotection was prevented as demonstrated by an increase in IS and apoptosis, CK-MB, cTnI, and LDH serum levels, C3, C5a, Bax, and NF-κB levels, and the inflammatory mediator expression of IL-6, TNF-α, ICAM-1, and IL-1β. These data highlight the pivotal function of HSP90 in inhibiting C3, C5a and NF-κB in PioC.

Pioglitazone, a PPAR-γ agonist, is widely used to manage type 2 diabetes. Based on previously published evidence, it can be stated that PPAR-γ has the potential as a new therapeutic target in I/R injury.^{22, 23} Clinical studies have confirmed the benefits of pioglitazone in alleviating reperfusion injury in diabetic patients with AMI.⁸ Previous studies showed that PioC prevents I/R injury by increasing the expression of HSP32 (also called heme oxygenase 1) and HSP72.^{24, 25} Thus far, no studies have investigated whether PioC exerts a regulatory effect on HSP90, so herein, we constructed an I/R rat model, and found that PioC significantly upregulated the expression of HSP90 to prevent I/R injury, and the treatment with the HSP90 inhibitor GA reversed this effect.

The C3 is a vital component in the activation of the complement system, as demonstrated in reperfused myocardium, where C3 levels are markedly increased. Activation of C3 induces the production of C3a and C5a, which act as inflammatory mediators promoting the release of inflammatory cytokines (TNF-α, IL-1β and IL-6), which leads to an inflammatory response, and eventually the cleavage of fragment C5b. The C5b binds with the other components (C6, C7, C8, and C9) to form C5b-9, also known as the membrane attack complex, which causes the subsequent cardiomyocyte damage and myocardium destruction.^{11, 26, 27} Ischemic preconditioning and postconditioning inhibit C3 and C5a.^{14, 18} The PPAR-γ activation inhibits the inflammatory response by suppressing the production of inflammatory factors.^{3, 28} To date, it was unclear whether PioC had potent inhibitory effects on C3, C5a, TNF-α, IL-1β, IL-6, and ICAM-1. Our data revealed that the PioC group had lower C3, C5a, IL-1β, TNF-α, IL-6, and ICAM-1 levels than the I/R group. These data highlight the potential role of PioC in the suppression of C3, IL-1β, C5a, TNF-α, IL-6, and ICAM-1, anti-inflammatory induction, and cardioprotective effects.

The NF-κB is the key player in the inflammatory response after myocardial I/R injury.²⁹ It has previously been shown that NF-κB regulates complement component C3 to induce complement system activation and contribute to the complement-mediated immune–inflammatory response.³⁰ Other studies have revealed that NF-κB promotes the expression of pro-inflammatory cytokines in response to complement activation, which aggravates myocardial damage,^{31, 32} and that pioglitazone inhibits inflammation via suppression of NF-κB signaling.^{33, 34} In addition, pioglitazone was shown to combat myocardial I/R injury by inhibiting the overexpression of NF-κB activation.²⁰ Nevertheless, a correlation between complement components and NF-κB signaling in PioC-mediated cardioprotection remained unclear. Our results revealed that PioC lowers the protein expression levels of NF-κB, C3 and C5a. These data show that the PioC inhibitory effects on NF-κB signaling could be related to the regulation of the complement system.

The HSP90 is upregulated under stress and protects the heart from I/R injury during pharmacological conditioning and ischemic postconditioning.^{21, 35} Relevant investigations revealed a direct link between HSP90 and complement-mediated cell death events.^{16, 17} The C3 and C5a can be suppressed by preconditioning.¹⁴ Previously, we found that ischemic postconditioning inhibited C3, C5a and NF-κB through HSP90.^{18, 19} These findings supported the view that PioC inhibits C3, C5a and NF-κB by upregulating HSP90. In the current study, the PioC group had higher HSP90 levels and lower C5a, C3 and NF-κB expression levels than the I/R group. Furthermore, GA reversed the PioC-mediated C3, C5a and NF-κB downregulation.

Cardiomyocyte apoptosis induced by I/R contributes to myocardial I/R injury dependent on the ratio of anti-apoptotic (Bcl-2) to proapoptotic (Bax) proteins.³⁶ Previous evidence demonstrated that pioglitazone reduced the levels of proapoptotic Bax and enhanced those of anti-apoptotic Bcl-2 proteins to attenuate myocardial apoptosis during I/R.²⁰ We reported that the Bax and Bcl-2 expression in postconditioning depended on HSP90.²¹ In mast cells, the inhibition of HSP90 by GA inhibited the interaction of HSP90 with Bcl-2 and resulted in apoptosis.³⁷ In the present study, pioglitazone combined with GA reversed PioC-induced upregulation of Bcl-2, suggesting that the PioC anti-apoptotic effect is linked with HSP90.

These data support the evidence that HSP90 is a central player in the C3, C5a, NF-κB, and Bax inhibition during PioC. To the best of our knowledge, our observation is a pioneer in reporting the relationship between HSP90, NF-κB, C5a, C3, Bax, and Bcl-2 in PioC, which indicates that cardioprotection induced by PioC is closely related to HSP90 and the complement system. However, the detailed mechanism of the direct effect of PioC on HSP90 and the complement system needs to be clarified.

Limitations

There are some limitations to this investigation. First, we only applied the HSP90-specific modulator GA. The application of over- or underexpression of HSP90 using viral vectors to explore its role in PioC could be carried out in future studies. Second, our current work only showed the effect of HSP90 on complement activation in PioC; the application of complement inhibitors could further explore its effects on HSP90 and downstream signal transduction mechanisms of complement. Third, further research is needed to explore the detailed mechanism of the direct effect of PioC on HSP90 and the complement system. Fourth, because the sample size was very small, the Shapiro–Wilk test had low test power; thus, larger sample sizes should be applied in future studies.

Conclusions

The HSP90 plays a pivotal role in PioC-mediated cardioprotection, likely through the inhibition of C3, C5a and NF-κB activation, leading to a decrease in I/R-induced IS, myocardial inflammation and cardiomyocyte apoptosis. The current findings highlight the potential of HSP90 in treating I/R injury.

Supplementary data

The supplementary materials are available at https://doi.org/10.5281/zenodo.7871084. The package contains the following files:

Supplementary Tables 1–4. Statistical analysis results of Figure 1, Figure 2, Figure 3, Figure 4.

Supplementary Table 5. Data points of the study.

Supplementary Tables 6–12. Statistical analysis results of Supplementary Fig. 1.

Supplementary Fig. 1. The mRNA expression of inflammatory factors with β-actin as a housekeeping gene.