Male 8-week-old Sprague Dawley (SD) rats (weighing 200–220 g) were utilized in the present study.

The AMI model was generated via left anterior descending coronary artery (LAD) ligation in SD rats. Briefly, rats were anesthetized with 1% pentobarbital sodium. After left thoracotomy, the heart was exteriorized, and the LAD was ligated approximately 2 mm below the left atrium with a 6–0 silk suture. AMI was confirmed by elevation of the ST segment on an electrocardiogram and bulging of the relevant segment of the left ventricle (LV). In the sham group, the suture was removed without tying, and no infarction was generated. After establishment of the AMI model, rats were divided into six groups (n = 14–16 per group) by a random number table: sham, model, diltiazem (8.1 mg/kg), allicin (14 mg/kg), allicin (7 mg/kg), and allicin (14 mg/kg) + PAG (32 mg/kg) groups (Cui et al., 2020). All groups received intraperitoneal injection once a day for 7 days.

Cardiac function was assessed using a Vevo 3100 echocardiography system (Visual sonics Inc, Toronto, Canada). Rats were anesthetized with 1.5–2% isoflurane via continuous inhalation and warmed on a heated pad (37°C). Ultrasound transmission gel was applied to the chest, and echocardiography (M-mode and B-mode imaging) was performed. The LV internal diameter and thickness of the anterior wall at end-diastole (LVID d, LVAW d) and end-systole (LVID s, LVAW s), as well as LV fractional shortening (FS), ejection fraction (EF), and stroke volume (SV) were measured in each rat in a blinded manner. All values were averaged using three to five cardiac cycles per rat. Rat hearts were harvested after 1% pentobarbital sodium overdose via intraperitoneal injection and sliced into five sections of 1-mm thickness across the left ventricular long axis under the ligature. To identify the infarction area, heart slices were incubated with nitro-blue tetrazolium chloride (Sigma-Aldrich) for 3 min at 22 ± 2°C. Infarction areas were measured using Image-Pro Plus software (Version 6.0; Media Cybernetics, Silver Springs, MD, United States) and presented as a percentage of infarct area to ventricular area or total area.

Before the rats were sacrificed, blood samples were collected from the abdominal aorta. Serum was incubated at 22 ± 2°C for 30 min and centrifuged at 975.87 ×g for 10 min. The supernatant was collected for determination of serum cardiac troponin I (cTnI), lactate dehydrogenase (LDH), and CSE levels. Levels of serum cTnI, LDH, and CSE were separately quantified with commercially available cTnI (Medical Discovery Leader, Beijing, China, 159632), LDH (Medical Discovery Leader, Beijing, China, 164752), and CSE ELISA kits (Bluegene, Shanghai, China, E02C0834); antibody and chromogenic agent were added according to the manufacturer’s instructions. Absorbance was measured at 450 nm, detected by a microplate tester. Levels of LDH, cTnI, and CSE were calculated according to the standard curve (Chen et al., 2019; Li J. et al., 2020; Wu et al., 2020).

The border zone of myocardial infarction tissues were fixed with 4% (v/v) paraformaldehyde and incubated with dimethylbenzene for 30 min before serum blocking for 60 min. Specimens were incubated with CSE antibody (Proteintech, Wuhan, China, 12217-1-AP) for 24 h at 4°C prior to incubation with goat anti-rabbit IgG (H + L) fluorescein isothiocyanate-conjugated polyclonal antibody (20200321, Bai Aotong Experimental Materials Center, Luoyang, China) in the dark at 37°C for 60 min. After washing specimens in phosphate buffer solution, nuclei were stained with 4′, 6-diamidino-2-phenylindole (Sigma-Aldrich). Images were obtained using an upright fluorescence microscope (DM-LFS, Leica, MH, Germany) under ×400 magnification.

Levels of H₂S in serum and the border zone of myocardial infarction tissue were measured using methylene blue spectrophotometry at 665 nm according to manufacturer’s instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Hearts were harvested, weighed, washed in phosphate buffer solution, fixed in 4% paraformaldehyde overnight, and embedded in paraffin. Each paraffin-embedded heart was cut into 4-µm thick sections through the infarction area and stained with hematoxylin and eosin (H&E) for morphological observation. Specimens were stained with Masson’s trichrome stain to evaluate collagen volume. Sections were imaged using a stereomicroscope (Olympus SZ61, Tokyo, Japan).

The coronary arterial rings (diameter: 100–300 µm) of SD rats were then isolated and placed in a cold Krebs buffer [composition (mM): NaCl, 119; KCl, 4.6; CaCl₂, 1.5; NaH₂PO₄, 1.2; MgCl₂, 1.2; NaHCO₃, 15; Glucose, 5.6; pH 7.4]. RCARs were prepared and performed as previously described (Cui et al., 2020).

RCARs were contracted with KCl (6 × 10^–2 M) until a plateau of contraction was reached. The rings were then divided equally into two groups (8 rings each): PAG or control. RCARs were incubated with PAG (10^–2 M) or the same volume of saline, respectively, for 5 min. Both groups were treated with allicin (10⁻⁵–10^–4.2 M) and cumulative concentration-response curves were obtained.

KCl mediates the opening of voltage-dependent Ca²⁺ channels (VDCCs), whereas U46619, 5-HT, and ET-1 mediate the opening of receptor-dependent Ca²⁺ channels (RDCCs). In this study, RCARs were divided into three groups (8 rings each): allicin, allicin + PAG, and control groups. Rings were incubated with allicin (10^–4.8 M), allicin (10^–4.8 M) + PAG (10^–2 M), or the equivalent volume of saline, respectively, for 5 min. Cumulative concentration-response curves for KCl (10^–1.54–10^–1.42 M), U46619 (10⁻⁸–10^–5 M), 5-HT (10⁻⁸–10^–4 M), and ET-1 (10⁻⁹–10^–6 M) were obtained.

To investigate the contribution of Ca²⁺-sensitive potassium channels (K_Ca), voltage-dependent potassium channels (Kv), inwardly rectifying potassium channels (K_ir), and ATP-sensitive potassium channels (K_ATP) to allicin-induced vasodilation, the corresponding inhibitors, TEA (10^–3 M), 4-AP (10^–3 M), BaCl₂ (10^–5 M) and Glib (10^–5 M) were applied. RCARs were divided into two groups (8 rings each): control and inhibitor groups. The rings were incubated with the four respective inhibitors for each channel or the same volume of saline for 5 min. Cumulative concentration-response curves for allicin (10⁻⁵–10^–4.2 M) were obtained.

If the maximum vasodilatory effect in the inhibitor group was lower than that in the control group, subsequent experiments were conducted, as follows. The rings were divided into three groups (8 rings each): inhibitor, PAG, and PAG + inhibitor groups. Cumulative concentration-response curves for allicin (10⁻⁵–10^–4.2 M) were obtained as described above.

The grouping and intervention for RCARs were performed as described above (See “Effect of Allicin on Ca²⁺ Channel-induced Contraction” section). Contraction-response curves for caffeine (3 × 10^–2 M) were obtained.

Isolated cardiomyocytes were loaded with 2 μM Fura-2 AM (Sigma-Aldrich) in the dark for 30 min at 22 ± 2°C. Cells were washed, resuspended twice in Tyrode’s solution (concentration in mM: 137.0 NaCl, 1.2 NaH₂PO₄, 5.0 KCl, 1.2 MgCl₂, 10.0 HEPES, 10.0 glucose, and 1.2 CaCl₂ [pH 7.4]), and placed in a cell chamber. Myocytes were stimulated to contract at a pacing frequency of 1 Hz with 4 ms of electrical stimulation. Myocytes were exposed to 340 or 380 nm excitation wavelengths, and the emitted fluorescent signal was detected at 510 nm. Sarcomere length and fluorescence intensity (a proxy of Ca²⁺ concentration) were synchronously recorded with a cell contraction-ion detection system (IonOptix, Westwood, MA, United States). Contractility parameters including amplitude, peak time, systolic half-time of decay (T₅₀), diastolic T₅₀, and myofilament sensitivity were measured. Ca²⁺ transient parameters including amplitude, maximum ascending and descending velocity, and Ca²⁺ decline time constant were also recorded.

SR Ca²⁺ content is associated with caffeine-sensitive Ca²⁺ release (Santulli et al., 2017). Short puffs of 10 mM caffeine were applied to completely empty the SR, following a train of 1-Hz field stimulation to achieve steady-state SR Ca²⁺ loading in ventricular myocytes. SR Ca²⁺ content was assessed by measuring the amplitude of caffeine-elicited Ca²⁺ transients (△F/F₀).

Rapid and continuous application of 10 mM caffeine was employed to induce SR Ca²⁺ release and assess the contribution of the Na⁺/Ca²⁺ exchanger (NCX) and slow transport systems (mitochondrial Ca²⁺ uniporter and sarcolemmal Ca²⁺-ATPase). The contribution of the slow transport system to Ca²⁺ removal is only 1% and is often overlooked (Puglisi et al., 2014). With continuous caffeine superfusion, a decrease in fluorescence (F340/380) indicates Ca²⁺ removal, which is predominantly attributable to the NCX. Ca²⁺ removal was predominantly achieved by sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) uptake and NCX Ca²⁺ efflux with superfusion of Tyrode’s solution. Based on these factors, the time of SERCA-mediated Ca²⁺ removal was calculated (Tau).

Ca²⁺ leakage levels were assessed by perfusing myocytes with 1 mM tetracaine and reperfusing in 10 mM Na-, Ca-free Tyrode’s solution (containing Li⁺ and EGTA instead of Na⁺ and Ca²⁺). Levels of SR Ca²⁺ leakage were calculated based on the difference in cytosolic Ca²⁺ concentration before and after tetracaine perfusion.

Ventricular size and function were measured to assess the effect of allicin treatment. There was no statistical difference in LVID d among the groups. LVID s was significantly greater in the model group than in the sham group (Figure 1A), whereas the values of LVAW s, LVAW d, EF, FS, and SV were significantly lower in the model group than in the sham group. On the contrary, LVID s was significantly lower, and LVAW s, LVAW d, EF, FS, and SV were significantly higher, in the diltiazem 8.1 mg/kg and allicin 14 mg/kg groups than in the model group. LVID s was significantly higher, and LVAW s, LVAW d, EF, FS, and SV were significantly lower in the allicin 14 mg/kg + PAG group than in the allicin 14 mg/kg group (Figures 1B‒F). Hence, as shown in Figure 1G, we found that allicin significantly improved the cardiac function of AMI rats, and PAG partially weakened this effect.

The percentages of myocardial infarction area to ventricular area (ITV) and total heart area (ITT) were significantly lower in the sham group (0.42 ± 0.26% and 0.23 ± 0.16%, respectively) than in the model group (30.89 ± 8.26% and 17.49 ± 4.75%, respectively). ITV and ITT were significantly reduced by treatment with 14 mg/kg of allicin (7.67 ± 5.15% and 4.33 ± 2.84%, respectively) and 8.1 mg/kg of diltiazem (8.27 ± 7.65% and 4.67 ± 4.57%, respectively). However, in the allicin 14 mg/kg + PAG group, ITV and ITT were significantly higher (17.57 ± 8.17% and 9.97 ± 4.63%, respectively) than in the allicin 14 mg/kg group, but significantly lower than in the model group (Figures 2A,B). Hence, allicin significantly decreased the myocardial infarction area of AMI rats, and PAG partially weakened the effect of allicin (Figure 2E).

cTnI and LDH levels were significantly higher in the model group than in the sham group (cTnI: 120.53 ± 13.93 ng/L vs. 59.39 ± 9.81 ng/L; LDH: 291.81 ± 15.91 µg/L vs. 146.62 ± 16.20 µg/L), indicating that the model successfully mimicked AMI conditions. Levels of cTnI and LDH were significantly lower in the diltiazem 8.1 mg/kg group (82.12 ± 5.09 ng/L and 215.06 ± 16.16 µg/L, respectively), allicin 14 mg/kg group (82.82 ± 8.30 ng/L and 220.77 ± 19.85 µg/L, respectively), allicin 7 mg/kg group (94.04 ± 3.84 ng/L and 231.55 ± 24.56 µg/L, respectively), and allicin 14 mg/kg + PAG group (94.14 ± 6.28 ng/L and 252.20 ± 20.53 µg/L, respectively) than in the model group. cTnI and LDH levels were significantly higher in the allicin 14 mg/kg + PAG group than in the allicin 14 mg/kg group. These data suggest that treatment with diltiazem or allicin significantly alleviated the changes in cardiac function induced by AMI, and PAG partially reversed this effect (Figures 2C,D).

H&E and Masson staining revealed that in the sham group, the structure of cardiomyocytes remained intact, the transverse and fiber striations of the myocardium were clear, and cells were arranged regularly. No degeneration, necrosis, hemorrhage, inflammatory cell infiltration, or collagen deposition was observed in the sham group. In the model group, the arrangement of myocardial fibers was disordered, swollen, and disjointed; the septum of the fiber bundles was widened; transverse lines of cells were absent, striations were disordered, normal cell structure was disrupted; and extensive cardiomyocyte necrosis, interstitial vascular hyperplasia, hyperemia, edema, inflammatory cell infiltration, and collagen deposition were observed. These features are typical of myocardial infarction. In the allicin 14 mg/kg and diltiazem 8.1 mg/kg groups, the arrangement of cardiomyocytes was slightly disordered, myocardial fibers were neatly arranged, the interstitium was broadened, the septum of the fiber bundles was not significantly widened, no cell necrosis was observed, and minimal cardiomyocyte edema and collagen deposition were noted. Hence, the degree of pathological damage observed in the allicin 7 mg/kg and allicin 14 mg/kg + PAG treatment groups was intermediate compared with that noted in the model and allicin 14 mg/kg groups (Figures 2F,G).

The levels of the H₂S in serum and the border zone of the myocardial infarction tissues were significantly lower in the model group (36.23 ± 8.96 nM/ml and 342.18 ± 48.77 nM/g, respectively) than in the sham group (109.39 ± 11.44 nM/ml and 682.93 ± 56.83 nM/g, respectively) (Figures 3A,B). On the other hand, the levels of H₂S in the serum and myocardial tissue were significantly higher in the diltiazem 8.1 mg/kg (86.21 ± 7.03 nM/ml and 543.91 ± 49.15 nM/g, respectively), allicin 14 mg/kg (85.73 ± 9.06 nM/ml and 563.91 ± 34.28 nM/g, respectively), allicin 7 mg/kg (68.35 ± 7.18 nM/ml and 469.82 ± 48.25 nM/g, respectively), and allicin 14 mg/kg + PAG groups (62.00 ± 8.48 nM/ml and 490.53 ± 47.75 nM/g, respectively) than in the model group, suggesting that these treatments alleviated myocardial infarction symptoms. High levels of allicin were more effective than lower levels, as H₂S serum and tissue levels were significantly lower in the allicin 7 mg/kg and allicin 14 mg/kg + PAG groups than in the allicin 14 mg/kg group. Finally, the levels of H₂S in the serum and myocardial tissue were significantly higher in the allicin 14 mg/kg group than in the model group.

Immunofluorescence analysis revealed that the average optical density (AOD) of CSE in myocardial tissue was significantly lower in the model group (0.039 ± 0.014) than in the sham group (0.141 ± 0.028). CSE showed a significantly higher level in the diltiazem 8.1 mg/kg (0.091 ± 0.009), allicin 14 mg/kg (0.101 ± 0.008), allicin 7 mg/kg (0.077 ± 0.010), and allicin 14 mg/kg + PAG groups (0.073 ± 0.003) than in the model group. On the other hand, CSE levels tended to be lower in the allicin 7 mg/kg and allicin 14 mg/kg + PAG groups than in the allicin 14 mg/kg group, but this trend did not reach statistical significance (Figure 3C). Similarly, serum CSE levels were significantly lower in the model group than in the sham group (CSE: 1.95 ± 0.52 µg/L vs. 4.07 ± 0.88 µg/L). Serum CSE levels were significantly higher in all treatment groups than in the model group. CSE levels were significantly lower in the allicin 14 mg/kg + PAG group than in the allicin 14 mg/kg group (Figure 3D). Figure 3E suggests that in rats with AMI, allicin treatment improved CSE levels partly via inducing H₂S production.

Allicin-induced dilation in rat coronary arteries precontracted with KCl in a dose-dependent manner. The maximum vasodilation reached 85.11 ± 2.11% of the pre-contraction amplitude. After the application of PAG, the maximum relaxation induced by allicin was 49.37 ± 6.94%, which was significantly lower than that in the control group (Figures 4A,B). The inhibitory rate of PAG on the vasodilation effect of allicin was 42.4%. There was no significant difference in pEC₅₀ between the two groups.

KCl induced contractions in rat coronary arteries in a dose-dependent manner. No significant changes in E_max and pEC₅₀ of KCl-induced concentration-contraction curves were observed after administration of allicin or allicin + PAG (Figures 4C,D).

Following administration of allicin, the E_max values of the concentration-contraction curves induced by 5-HT, U46619, and ET-1 were 101.86 ± 2.16%, 102.19 ± 15.24%, and 120.77 ± 13.98%, respectively. These values were significantly lower than those of the control group (158.73 ± 12.63%, 146.56 ± 18.23%, and 173.51 ± 14.09%, respectively). Following administration of allicin + PAG, the E_max values of the concentration-contraction curves induced by 5-HT, U46619, and ET-1 were 133.91 ± 8.65%, 124.09 ± 6.64%, and 144.02 ± 5.22%. These values were significantly lower than those of the control group but were significantly higher than those of the allicin group (Figure 5).

The R_max of allicin in the TEA, 4-AP, and BaCl₂ groups was not significantly different compared with that of the control group (82.90 ± 6.22%), whereas the R_max of allicin was significantly lower in the Glib group (46.98 ± 3.86%) than in the control group. No significant differences were observed in the R_max of allicin among the PAG, PAG + Glib, and Glib groups. Altogether, these data suggest that the response to allicin is mediated by receptor-dependent, rather than voltage-dependent, Ca²⁺ channels (Figures 6A‒D).

Caffeine (3 × 10^–2 M) induced a transient and rapid contraction of rat coronary arteries in Ca²⁺-free solution. The contraction amplitude was 40.70 ± 10.09% of the contraction induced by KCl (6 × 10^–2 M). After administration of allicin, the contraction amplitude induced by caffeine was 28.45 ± 5.42%, which was significantly lower than that in the control group. Caffeine-induced contraction amplitude was significantly higher in the allicin + PAG group (40.28 ± 6.05%) than in the allicin group (Figure 6E).

Figure 7 shows that the contraction amplitude (3.85 ± 1.49%) was significantly lower, whereas peak time (0.17 ± 0.04 s), systolic T₅₀ (0.056 ± 0.017 s), and diastolic T₅₀ (0.128 ± 0.044 s) were significantly longer in the model group than in the sham group (8.47 ± 1.80%, 0.13 ± 0.02 s, 0.043 ± 0.009 s, and 0.069 ± 0.014 s, respectively). Contraction amplitude in the allicin 14 mg/kg (6.76 ± 2.43%) and allicin 14 mg/kg + PAG groups (4.65 ± 1.53%) was higher, whereas peak time, systolic T₅₀, and diastolic T₅₀ in the allicin 14 mg/kg group (0.13 ± 0.02 s, 0.041 ± 0.010 s, 0.081 ± 0.033 s) was significantly lower than that in the model group. The addition of PAG significantly reduced the contraction amplitude in the allicin 14 mg/kg + PAG group (4.65 ± 1.53%) compared with the allicin 14 mg/kg group. Similarly, peak time, systolic T₅₀, and diastolic T₅₀ were significantly longer in the allicin 14 mg/kg + PAG group (0.16 ± 0.04 s, 0.062 ± 0.016 s, and 0.122 ± 0.046 s, respectively) than in the allicin 14 mg/kg group, indicating that PAG significantly alleviated the effects of allicin.

AMI significantly reduced the amplitude, maximum ascending velocity, and maximum descending velocity of Ca²⁺ transients in the model group (0.13 ± 0.03 RU [ratio unit], 8.02 ± 3.20 RU/s, and 0.52 ± 0.28 RU/s, respectively) compared with the sham group (0.18 ± 0.05 RU, 11.36 ± 3.27 RU/s, and 0.80 ± 0.28 RU/s, respectively). The Ca²⁺ decline time constant in the model group (0.35 ± 0.08 s) was significantly longer than that in the sham group (0.23 ± 0.07 s) as well. Ca²⁺ transient amplitude in the allicin 14 mg/kg group (0.15 ± 0.03 RU) and maximum ascending and descending velocities of Ca²⁺ transients in the allicin 14 mg/kg (10.26 ± 3.42 RU/s and 0.79 ± 0.34 RU/s, respectively) and allicin 14 mg/kg + PAG groups (9.39 ± 3.26 RU/s and 0.68 ± 0.38 RU/s, respectively) were significantly higher than in the model group. The Ca²⁺ decline time constants in the allicin 14 mg/kg (0.25 ± 0.10 s) and allicin 14 mg/kg + PAG groups (0.28 ± 0.08 s) were significantly shorter than in the model group. Compared with the allicin 14 mg/kg group, the allicin 14 mg/kg + PAG group exhibited significantly lower Ca²⁺ transient amplitude and maximum ascending and descending velocities of Ca²⁺ transients, as well as significantly longer Ca²⁺ decline time constant (Figure 7).

Diastolic curves with good linear relationships were obtained from the nonlinear fitting curve between sarcomere length and Ca²⁺ transient amplitude. The slope of this curve was calculated to reflect myofilament sensitivity. Myofilament sensitivity of cardiomyocytes in the model group (2.46 ± 0.77) was significantly lower than in the sham group (3.86 ± 1.76). By contrast, myofilament sensitivity in the allicin 14 mg/kg and allicin 14 mg/kg + PAG groups was significantly higher than that in the model group. Myofilament sensitivity was significantly lower in the allicin 14 mg/kg + PAG group than in the allicin 14 mg/kg group (Figure 7I).

Tau_NCX and Tau_SERCA were significantly higher in the model group (6986.2 ± 5238.6 ms and 349.5 ± 63.5 ms, respectively) than in the sham group (2732.3 ± 1275.1 ms and 200.7 ± 93.6 ms, respectively). Compared with the model group, the Tau_NCX and Tau_SERCA of the allicin 14 mg/kg group (3202.9 ± 1473.1 ms and 246.4 ± 63.1 ms, respectively) and allicin 14 mg/kg + PAG group (3584.4 ± 768.0 ms and 246.4 ± 29.1 ms, respectively) were both significantly decreased. No significant differences were observed in Tau_NCX and Tau_SERCA between the allicin 14 mg/kg and allicin 14 mg/kg + PAG groups (Figures 8A,B).

The SR Ca²⁺ content of myocardial cells was threefold lower in the model group (0.53 ± 0.16 RU) than in the sham group (1.51 ± 0.37 RU). On the other hand, the allicin 14 mg/kg (1.43 ± 0.18 RU) and allicin 14 mg/kg + PAG (1.17 ± 0.13 RU) treatments induced significantly higher SR Ca²⁺ content when compared with the model group. SR Ca²⁺ content was significantly lower in the allicin 14 mg/kg + PAG group than in the allicin 14 mg/kg group.

Ca²⁺ leakage was significantly higher in the model group (0.097 ± 0.033 RU) than in the sham group (0.052 ± 0.040 RU), whereas allicin 14 mg/kg (0.056 ± 0.05 RU) and allicin 14 mg/kg + PAG (0.073 ± 0.044 RU) administration significantly decreased leakage compared with the model group. Ca²⁺ leakage was significantly higher in the allicin 14 mg/kg + PAG group than in the allicin 14 mg/kg group (Figures 8C,D).