The cardioprotective effect persisting during recovery from cold acclimation is mediated by the β2-adrenoceptor pathway and Akt activation.

The infarct size-limiting effect elicited by cold acclimation (CA) is accompanied by increased mitochondrial resistance and unaltered β1-adrenergic receptor (AR) signaling persisting for two weeks at room temperature. As the mechanism of CA-elicited cardioprotection is not fully understood we examined the role of the salvage β2-AR/Gi/Akt pathway. Male Wistar rats were exposed to CA (8 °C, 5 weeks), while the recovery group (CAR) was kept at 24 °C for additional 2 weeks. We show that the total number of myocardial β-ARs in the left ventricular myocardium did not change after CA but decreased after CAR. We confirmed the infarct size-limiting effect in both CA and CAR groups. Acute administration of β2-AR inhibitor ICI-118551 abolished the protective effect in the CAR group but had no effect in the control and CA groups. The inhibitory Giα1/2 and Giα3 proteins increased in the membrane fraction of the CAR group, and the p-AktSer473/Akt ratio also increased. Expression, phosphorylation, and mitochondrial location of the Akt target glycogen synthase kinase (GSK3β), were affected neither by CA nor by CAR. However, GSK3β translocated from the Z-disc to the H-zone after CA, and acquired its original location after CAR. Our data indicate that the cardioprotection observed after CAR is mediated by the β2-AR/Gi pathway and Akt activation. Further studies are needed to unravel downstream targets of the central regulators of the CA process and the downstream targets of the Akt protein after CAR.


INTRODUCTION
Despite intensive research, acute myocardial infarction remains a leading cause of death and disability. New approaches are therefore needed to reduce myocardial infarct size and its severity. Recently, it was reported that cold acclimation (CA) possesses an infarct size-limiting effect (1,2). We have documented that exposure of rats to 5 wk of gradual mild cold acclimation (CA), followed by 2 wk of recovery resulted in reduction of myocardial infarct size without apparent side effects such as hypertension or hypertrophy. We suggested that b-adrenergic signaling plays a significant role in the development of cardioprotection triggered by CA, but its mechanism is not known (1).
There are three subtypes of b-adrenoceptors (b-ARs) in the ventricular myocardium, b 1 -ARs, b 2 -ARs, and b 3 -ARs (3). The predominant b 1 -ARs, coupled to G s a are required for hormone-stimulated cAMP generation by adenylyl cyclase, which stimulates the activity of PKA. Acute activation of PKA results in phosphorylation of target proteins and promotes positive inotropic and lusitropic effects; its persistent activation is, however, detrimental (4). Our recent study showed that cold-induced cardioprotection implies an unaltered b 1 -AR/adenylyl cyclase/cAMP pathway, including stable levels of phosphorylated PKA and G s . We confirmed the finding by acute administration of b 1 -AR antagonist (metoprolol), which did not affect the CA-elicited cardioprotection. Simultaneously, increased b 2 -AR translocation to the membrane fraction was observed after CA, and it remained upregulated for 2 wk of recovery at 24 C following CA (1).
It is known that b 2 -ARs, unlike b 1 -ARs, signal via both G s and G i proteins (5). Thus, b 2 -ARs can activate not only G s / cAMP-dependent PKA signaling pathways but also G i protein-dependent pathways. This includes, for example, phosphoinositide-3-kinase/protein kinase B (PI3K/Akt) (6) and cytosolic phospholipase A 2 that dampen the b 2 -AR/cAMP/ PKA signaling in the heart in a pertussis toxin-sensitive manner (7,8). Activation of b 2 -ARs is implicated in the cardioprotective effect of preconditioning (9), and both downstream kinases, PKA and Akt, were found to play an important role in this process (10). Phosphorylation of Ser 473 of Akt is believed to be a crucial step in its activation, since it stabilizes the kinase domain in its active conformation (11). The b 2 -AR/G i /PI3K/Akt signaling axis that promotes cell survival (12) has been repeatedly reported to be cardioprotective (13)(14)(15). The Akt-elicited antiapoptotic effect is mediated mainly via mitochondria by inhibition of the mitochondrial permeability transition pore (mPTP) opening, presenting a key terminal effector of cardioprotection and cell death (16,17).
In the present study, we examined the role of b 2 -AR/G i signaling and the possible involvement of Akt kinase in the CA-elicited infarct size-limiting effect. The results of our study indicate that the b 2 -AR pathway plays a protective role in cardioprotection during the recovery phase after CA.

Cold Acclimation and Ischemia-Reperfusion Injury
The cold acclimation and ischemia-reperfusion protocol was performed as previously described (1). Male Wistar rats (7-wk old, 200 g body weight, Velaz, Prague, Czech Republic) were housed in pairs in well-bedded cages to minimize environmental and social stress. All experiments were performed in the "winter-spring" season (from November to April). The animals were handled in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication, 8th ed., Revised 2011). The experimental protocol was approved by the Animal Care and Use Committee of the Faculty of Science, Charles University, Prague, Czech Republic.
Rats divided into two experimental groups were acclimated to 8 ± 1 C for 8 h/day during the 1st wk and then for 24 h/day during the following 4 wk, either without (CA, n = 24) or with 2 wk of recovery at 24 ± 1 C (CAR, n = 24). The control group (n = 24) was kept at 24 ± 1 C during the whole experiment. At the end of acclimation, the animals were anesthetized (thiopental, 60 mg/kg) at the respective acclimation temperature to prevent acute thermoregulatory response. Twelve animals from each group were treated after anesthesia by the b 2 -AR selective inhibitor ICI-118551 (Sigma-Aldrich; dissolved in saline), which was administered intraperitoneally at 1 mg/kg of body weight 20 min before coronary occlusion. The same volume of saline was administered in the respective control subgroups. Anesthetized animals were intubated, connected to a rodent ventilator (Ugo Basile, Italy), and ventilated at 60-70 strokes/min (tidal volume of 1.2 mL per 100 g of body weight). Systemic blood pressure was monitored by cannulation of the carotid artery and a single-lead electrocardiogram was recorded using PowerLab and LabChart Pro software (ADInstruments). The rats were subjected to left thoracotomy, followed by 10 min stabilization, left coronary artery occlusion for 20 min, and subsequent 3-h reperfusion. Then the hearts were excised, and the area at risk was delineated by perfusion with 5% potassium permanganate. Frozen hearts were cut to 1mm thick slices and infarct size was visualized by staining with 1% 2,3,5-triphenyltetrazolium chloride (Sigma-Aldrich) as described previously (18). Infarct size and area at risk were quantified by the Graphic Cell Analyzer software (19).

Homogenate and Crude Membrane Fraction Preparation
Hearts from individual groups of animals (n = 8) were collected under the same conditions. Briefly, hearts were excised from anesthetized rats (thiopental, 60 mg/kg) and quickly washed in ice-cold saline. Left and right ventricles and the septum were separated. The left ventricle (LV) free wall was snap-frozen in liquid nitrogen, weighed, and stored in liquid nitrogen until use as described previously (20). Each frozen LV sample was placed in five volumes of ice-cold TMES buffer (20 mM Tris-HCl, 3 mM MgCl 2 , 1 mM EDTA, 250 mM sucrose; pH 7.4) containing protease and phosphatase inhibitors (cOMPLETE and PhosSTOP, Sigma-Aldrich), cut into small pieces, and homogenized on ice using an Ultra-Turrax (IKA, 24,000 rpm, 15 s), and a glass homogenizer with a motor-driven Teflon pestle (1,200 rpm, 2 min). Aliquots of homogenate from each sample were stored in liquid nitrogen for further analyses. The homogenate was centrifuged (2,100 rpm, 10 min, 4 C, Hettich Universal 320 R, Hettich, Germany). The supernatant was collected, and the pellet was homogenized in the same volume of TMES and centrifuged again. Both supernatants were mixed and centrifuged (23,500 rpm, 30 min, 4 C, Beckman Optima L.90K, rotor Ti 50.2 Beckman). The pellet (crude membrane fraction) was homogenized in TMES buffer without sucrose (13), and stored the aliquots at À80 C until use. Protein concentration was assessed using the Bradford protein assay (Sigma-Aldrich).

b -Adrenoreceptor Binding Assay
The total number of myocardial b-ARs was determined by the radioligand binding assay with the b-AR antagonist [ 3 H] CGP-12177 as previously described (21). Briefly, samples of crude membrane fraction (100 μg protein) were incubated in the medium (total volume of 0.5 mL) containing 50 mM Tris-HCl, 10 mM MgCl 2 , and 1 mM ascorbic acid at pH 7.4 with the b-AR antagonist [ 3 H]CGP 12177 (ARC) in descending concentrations from 5 nM to 0.15 nM at 37 C for 1 h. The reaction was terminated by adding 3 mL of ice-cold washing buffer (50 mM Tris-HCl and 10 mM MgCl 2 ; pH 7.4) and subsequent filtration through GF/C filter pre-soaked for 1 h with 0.3% polyethyleneimine. The filter was washed twice with 3 mL of ice-cold washing buffer. After addition of 4 mL scintillation cocktail EcoLite (MP Biomedicals), radioactivity retained on the filter was assessed by liquid scintillation counting for 5 min. Nonspecific binding (background signal) was defined as the signal that was not displaced by 10 μM L-propranolol. It represented $30% of the totally bound radioligand. The proportion of b 2 -ARs (% of total b-ARs) in the crude membrane fraction was determined using a competitive binding assay with the b 2 -AR selective antagonist ICI-118551. Crude membrane samples were incubated with 1 nM [ 3 H] CGP 12177 at increasing concentrations of the selective b 2 -AR antagonist ICI-118551 (10 À10 À10 À4 M). Binding characteristics (B max and K d ) and the percentage of b 2 -ARs in the membrane fraction were calculated and statistically analyzed using GraphPad Prism 8 software.

Western Blot Analysis
Individual samples (n = 8) of crude membrane fractions from each group (20 mg protein/lane) were resolved by SDS-PAGE using 12% polyacrylamide gel at constant voltage (200 V) and Mini-Protean Tetra Cell (Bio-Rad), and subsequently blotted onto nitrocellulose membranes (0.2-mm pore size; Bio-Rad) at constant voltage of 100 V for 90 min using the Wet Blot Module (Bio-Rad) as previously described (22). After blocking with 5% nonfat milk in Tris-buffered saline (20 mM Tris-HCl, 0.5 M NaCl, and 0.05% Tween 20) for 1 h, the membranes were incubated overnight at 4 C with polyclonal antibodies against G i a 1/2 , G i a 3 [RCS antibody (23)], and Akt (sc-8312, Santa Cruz Biotechnology) or phosphorylated (p-)Akt (Ser 473 ) (no. 9271S, Cell Signaling), GSK-3b (sc-9166), and mouse monoclonal antip-GSK-3b (Ser 9 ) (sc-373800). The membranes were then washed and incubated with HRP-conjugated anti-rabbit antibody (A9169, Sigma-Aldrich) or HRP-conjugated anti-mouse antibody (no. 31432, Invitrogen). Protein bands were visualized by enhanced chemiluminescence (ECL) substrate SuperSignal West Dura Extended Duration Substrate (ThermoFisher Scientific) using the LAS-4000 imaging system (Fujifilm). Protein band intensity was quantified densitometrically using Quantity One Software (Bio-Rad). At least four samples from each group were always run on the same gel, quantified on the same membrane, and normalized to the total protein content per lane determined by Ponceau S staining, b-actin and GAPDH were used as a loading control in crude membrane fractions and homogenates, respectively. The accuracy and reproducibility of the chemiluminescence signal was validated by loading samples in ascending concentrations of 10 À 40 mg protein/lane, and each determination was performed at least three times. All figures show representative images of individual Western blots.
COLD ACCLIMATION PERSISTING CARDIOPROTECTION BY b 2 -AR/Gi/Akt phalloidin (A22287, Life Technologies). Sections were mounted in ProLong Gold Antifade Reagent with DAPI (Invitrogen). Images taken from at least five randomly selected positions of each section were scanned sequentially using a wide-field inverted fluorescence microscope (IX2-UCB, Olympus). Each position was optically sectioned at 0.5-mm steps resulting in $8-12 Z-stack layers, depending on the specimen thickness. Images were processed in Nikon Microscope Imaging Software (NIS-Elements, Japan). Colocalization analyses were performed as previously described (1) using ImageJ software (26).

Statistical Analysis
Statistical analysis was performed using the GraphPad Prism 8 software (GraphPad). The distribution of data was analyzed by Shapiro-Wilk and Komogorov-Smirnov normality tests. For parametric data, one-way ANOVA with Dunnett's multiple comparison test was used to identify significant differences between mean values of individual groups. Two-way ANOVA with Dunnett's multiple comparison test (effect of low temperature and/or ischemia-reperfusion) and Sidak's multiple comparison test (effect of ICI-118551) was used for

RESULTS
Proper acclimation to low temperature was confirmed by significantly increased weight of brown adipose tissue (BAT) after CA (by 90% compared with controls; Table 1). The increase in BAT weight persisted even after cessation of CAR (by 62%).
Body weight of the CA group animals was slightly but significantly lower (8%) compared with controls, and body weight of CAR group (2-wk older animals) increased by 12%. Heart weight and the ratio of heart/body weight did not differ between the groups (Table 1), which excludes potential heart hypertrophy. Consistent with this, baseline values of mean arterial blood pressure and heart rate were affected neither by CA nor by CAR. Acute administration of a specific b 2 -AR inhibitor had no effect on hemodynamic parameters ( Table 2).
To evaluate the role of b-ARs in cardioprotection elicited by CA, we assessed the number of total b-ARs. Analysis of saturation binding curves (Fig. 1A) indicated that the total number of b-ARs did not change as a result of CA but decreased by $16% after CAR (Fig. 1B). In parallel, the binding affinity of [ 3 H]CGP 12177 to b-ARs somewhat increased after CAR, which was reflected by a 29% drop in the value of K d (Fig. 1C). The ratio of b 1 -and b 2 -ARs was not affected by CA, but the b 2 -AR fraction increased from 20% to 25% after 2 wk recovery (Fig. 1, D and E).
We next evaluated an extent of myocardial infarction in the presence of b 2 -AR selective inhibitor ICI-118551 in order to confirm or exclude the role of b 2 -AR in CA-elicited cardioprotection. Infarct size represented 50% of the area at risk in control rats. CA reduced infarct size to 25% and the protection persisted in the CAR group (to 27%). Administration of b 2 -AR inhibitor abolished the protective effect in the CAR group, but it had no significant effect in the control and CA groups ( Fig. 2A). These results confirm a role of b 2 -ARs in the infarct-size lowering effect of CAR but not that of CA. The average area at risk normalized to the LV (AR/LV) was 47%-52%, and it did not differ among the groups (Fig. 2B).
Subsequent analysis of the b 2 -AR downstream pathway showed that, when compared with the control group, the protein level of G i a 1/2 and G i a 3 in the crude membrane fraction increased after CAR by 62% and 55%, respectively (Fig.  3, A and B). Interestingly, we found that after CAR the level of the Akt protein decreased by 17%, whereas phosphorylation of p-Akt (Ser 473 ) increased by 27%. Intriguingly, the p-Akt (Ser 473 )-to-Akt ratio increased by 50% in the CAR group (Fig. 4, A-C).
In the next set of experiments, GSK-3b, a target of Akt that regulates the mitochondrial proapoptotic pathway, was assessed. However, neither the protein nor its form phosphorylated on Ser 9 was affected by CA (Fig. 5, A-D). Fluorescent signal of p-GSK-3b (Ser 9 ) appeared mostly in transversal stripes in longitudinal sections of LVs, and we did not observe any differences in the colocalization of p-GSK-3b (Ser 9 ) with mitochondria among the groups (Fig. 5, E and F).
Analysis of total GSK-3b localization in sarcomeres revealed its translocation from the middle of I-band (Z-disk) to the H-zone after CA, which was reversed after CAR (Fig.  5G), and this was confirmed by a significant decrease in the colocalization of GSK-3b with phalloidin after CA but not  , and the p-Akt Ser473 -to-Akt ratio (C) were determined in the crude membrane fraction (n = 8, males) from control rats (Cont) and those acclimated for 5 wk to cold (CA) and subsequently recovered for 2 wk at 24 C (CAR). D: images of representative membranes for Western blot analysis of Akt, Akt Ser473 , and GAPDH as a loading control. Data were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. Values are means ± SD. * P < 0.05 and *** P < 0.0001 vs. Cont. COLD ACCLIMATION PERSISTING CARDIOPROTECTION BY b 2 -AR/Gi/Akt CAR (Fig. 5H). Application of another GSK-3b antibody (antimouse monoclonal sc-377213) confirmed this unexpected observation (data not shown). Analysis of p-GSK-3b (Ser 9 ) localization detected similar translocation pattern to the total GSK-3b (Fig. 5, I and J). The role of GSK-3b translocation in CA is unclear at present, and we are pursuing experiments to resolve this issue.

DISCUSSION
The present study builds on a recent observation that cardioprotection is induced by mild CA that persists for at least 2-wk period of normothermic recovery with a possible role of b 2 -ARs in the mechanism (1). Here, we show that the mechanism of the infarct size-limiting effect elicited by CA is not identical when analyzed immediately after CA and after CAR. A competitive receptor-binding assay demonstrated increased proportion of b 2 -AR within total b-AR in crude myocardial membrane fraction after CAR but not after CA. This effect was confirmed by acute administration of the specific b 2 -AR inhibitor ICI-118551, which abolished the infarct size-lowering effect persisting after CAR, but not that elicited by CA.
The observed increase in the b 2 -AR-to-b 1 -AR ratio after CAR is apparently a result of upregulation of b 2 -ARs and downregulation of b 1 -ARs. The increased b 2 -/b 1 -AR ratio under stress conditions may be involved in the protection of the heart against damage caused by overstimulation, since the b 2 -AR-subtype couples to both G s and G i proteins (27,28). As we demonstrated previously, the G s /adenylyl cyclase/ PKA pathway remained unaltered in both the CA and CAR groups (1). Hence, in the present study we focused on the G i / Akt pathway downstream of b 2 -AR. Expression of G i a proteins (Gia1/2 and Gia3) markedly increased after CAR, and this was reflected by activation of Akt; the p-Akt-to-Akt ratio increased considerably in this group. Activation of the b 2 -AR/G i /Akt pathway is related to intracellular signaling associated with antiapoptotic effects and improved cell survival (12,29), which have been repeatedly reported as cardioprotective (13,14,30). Activated Akt phosphorylates several regulatory proteins, including caspase-9, Bcl-2family proteins, and GSK-3b (31). Akt-mediated phosphorylation inactivates caspase-9, resulting in suppression of mitochondria-dependent apoptosis (32). Moreover, inactivation of the proapoptotic protein Bad by phosphorylation contributes to stabilization of outer mitochondrial membrane. Akt also inhibits a conformational change of the proapoptotic Bax protein and its translocation to mito-chondria (33). Hexokinase 2, phosphorylated on Thr 472 by Akt, increases its association with the outer mitochondrial membrane, which maintains mitochondrial membrane potential by preferential ADP supply to complex V and prevents association of the proapoptotic Bax protein with the voltage-dependent anion channel (25,34).
Akt activation can protect mitochondria and prevent release of proapoptotic proteins, i.e., cytochrome c or the apoptosis-inducing factor (35), which restrain increase in oxidative stress and lower the probability of mPTP opening (36). As we demonstrated previously, both CA and CAR lead to increased mitochondrial resistance to Ca 2 þ overload (1). This suggests that active Akt can contribute to the cardioprotective effect elicited by CAR via increasing the resistance of mPTP to Ca 2 þ -overload. Nevertheless, our data suggest that mitochondrial protection in CA and CAR groups involves different mechanisms, whose precise delineation is a subject of our further studies.
In this study, we also focused on GSK-3b, a key component of the PI3K/Akt pathway, which may contribute to mPTP opening when translocated to mitochondria (37). The prosurvival PI3K/Akt signaling negatively regulates GSK-3b and thus may participate in mitochondria-linked cardioprotection (38)(39)(40)(41). We have observed, however, that expression, phosphorylation, and location of GSK-3b in the mitochondrial compartment were affected neither by CA nor by CAR. This finding may exclude GSK-3b as an Akt target in the CA-elicited cardioprotective mechanism related to prevention of mPTP opening.
Importantly, a detailed inspection of the striated distribution pattern of GSK-3b in longitudinal sections of LVs revealed its translocation from the Z-disk to the H-zone compartment after CA and back to the original position during the recovery phase. To the best of our knowledge, this phenomenon has not been reported to date. There is strong evidence that GSK-3b-targeted proteins located mostly within the Z-disk of the sarcomere mediate the increase in myofilament Ca 2 þ sensitivity in the failing heart when the kinase is rephosphorylated during cardiac resynchronization therapy (42,43). This phosphorylation is accompanied by restoration of contractility. Interestingly, Akt was excluded as an upstream kinase in this process (42). In line with this, our data suggest that the CA-elicited translocation of GSK-3b to the H-band occurs in an Akt-independent manner (as Akt is activated only after CAR). To the best of our knowledge, there is no data suggesting that GSK-3b phosphorylates its substrate proteins within the central part of the H-band, i.e., the M-line. However, a possible candidate is myosin binding Figure 5. Effect of cold acclimation (CA) on myocardial glycogen synthase kinase-3b (GSK-3b) expression, phosphorylation (p-GSK-3b Ser9 ), and location. Relative protein levels of total GSK-3b (A), p-GSK-3b Ser9 (B), and the p-GSK-3b Ser9 -to-total-GSK3b ratio (C) were determined in homogenates (n = 8, males) prepared from control rats (Cont) and from rats acclimated for 5 wk to cold (CA) and subsequently recovered for 2 wk at 24 C (CAR). D: images of representative Western blots of total GSK-3b, p-GSK-3b Ser9 , and GAPDH as a loading control. E, F: representative micrographs of longitudinal cryosections of left ventricles stained for p-GSK-3b Ser9 (green) and OXPHOS (red; E) and quantitative analyses of their colocalization expressed by Mander's M1 correlation coefficient (n = 5, males; F). G, H: representative micrographs of longitudinal cryosections of left ventricles stained for total GSK-3b (red) and with phalloidin (green) (G, top) and quantitative analyses of their colocalization expressed by Mander's M2 correlation coefficient (n = 5; H). Intensity profile along the test line (yellow) show the translocation of total GSK3b from the I-band region to the H-zone after CA and its normalization after CAR (G, bottom). I, J: p-GSK-3b Ser9 translocation within the sarcomere (n = 5). Scale bar = 10 mm. Data presented in graphs were analyzed by one-way ANOVA with Dunnett's multiple comparisons test. Values are means ± SD. * P < 0.05 and ** P < 0.01 vs. Cont. COLD ACCLIMATION PERSISTING CARDIOPROTECTION BY b 2 -AR/Gi/Akt protein C (MyBP-C), a key regulator of cardiac contractility that is located in the C-zone of sarcomere, which is positioned laterally within the H-band (44). MyBP-C was shown to be located adjacent to the actin-positive I-band, which is similar as we document for GSK-3b colocalized with the phalloidin-positive staining after CA in the present study. Importantly, GSK-3b-mediated phosphorylation of MyBP-C on Ser 133 increased the contractility of permeabilized human cardiomyocytes (45). Generally, the MyBP-C phosphorylation maintains thick filament spacing and structure, which was approved as cardioprotective (46). In summary, changes of GSK-3b location elicited by CA may affect calcium sensitivity and contractility of sarcomeres, whereby contributing to the cardioprotective effect. These plausible hypotheses require experimental verification, which is a subject of our future studies.
Our results raise the following question: What is the reason for the difference between CA and CAR in establishing cardioprotection? CA animals feature high level of insulation that prevents heat loss and enlarged BAT that increases heat production, both contributing to improved thermal homeostasis (47). Specific signaling pathways elicited by CA that are activated at the cellular level are still not fully understood (1). At the early stages of CAR, the shift back to room temperature presents a sudden temperature rise by 16 C that leads to a transient overheating episode of the acclimated animals. This event is reminiscent of the well-known "heat stress" phenomenon associated with heat acclimation-mediated cytoprotective memory. Under these conditions, additional cardioprotective signaling pathways are activated (48). Importantly, increased Akt phosphorylation has been reported as a fundamental player in the heat acclimation-elicited protection (49)(50)(51).
We conclude that the mechanism of cardioprotection observed after CAR is mediated via the b 2 -AR-G i pathway and Akt activation. This could be related to the additional transient heat episode occurring when the cold acclimatedsubjects return to the room temperature. Further studies are needed to unravel downstream targets of the central regulators of the CA process and the downstream targets of the Akt protein after CAR.