Increased Beta2-Adrenoceptors in Doxorubicin-Induced Cardiomyopathy in Rat

Background The toxicity of doxorubicin, leading to an irreversible heart failure, limits its use as chemotherapeutic agent. The beneficial effects of early administration of β-blocker were reported in patients with heart failure due to doxorubicin, suggesting an important role of β-adrenoceptors (β-ARs). This study aimed to identify a putative target (β-AR and/or its effectors) at the early phase of a chronic doxorubicin-induced cardiomyopathy (Dox-CM) in a rat model. Methodology Dox-CM was induced by six doxorubicin injections (cumulative dose: 15 mg.kg−1) and validated by echocardiography and left ventricle (LV) catheterization. The β-AR protein expressions in LV were evaluated by western-blot at days 35 (d35) and 70 (d70) after the first doxorubicin injection. Ex vivo cardiac contractility (dP/dtmax, dP/dtmin) was evaluated on isolated heart in response to specific β-AR stimulations at d35. Results At d35, Dox-CM hearts were characterized by mild LV systolic and diastolic dysfunctions, which were exacerbated at d70. In Dox-CM hearts, β3-AR expression was only decreased at d70 (-37±8%). At d35, β1-AR expression was decreased by 68±6%, but ex vivo β1-AR function was preserved due to, at least in part, an increased adenylyl cyclase response assessed by forskolin. β2-AR expression was increased both at d35 (+58±22%) and d70 (+174±35%), with an increase of ex vivo β2-AR response at d35. Inhibition of Gi protein with pertussis toxin did not affect β2-AR response in Dox-CM hearts, suggesting a decoupling of β2-AR to Gi protein. Conclusion This study highlights the β1/β2-AR imbalance in early Dox-CM and reveals the important role that β2-AR/Gi coupling could play in this pathology. Our results suggest that β2-AR could be an interesting target at early stage of Dox-CM.


Introduction
Anthracyclines, like doxorubicin (Dox), epirubicin and daunorubicin, are among the most effective drugs used in chemotherapy for cancer patients. Since the late 60s, Dox is frequently used against a variety of cancers including Hodgkin's lymphoma [1], soft-tissue sarcomas [2], leukemia and solid tumors. However, Dox administration is limited due to severe cardiotoxic effects leading to dilated cardiomyopathy [3]. Prognosis of heart failure (HF) due to Dox-cardiotoxicity is poor and even worse than ischemic or idiopathic dilated cardiomyopathy. Although several mechanisms have been proposed to describe the mechanisms by which Dox induces cardiotoxicity (generation of free radicals, mitochondrial disruption, alteration of cellular energetic, and initiation of apoptotic cascades), these mechanisms are still not fully understood [4][5][6] and there is no specific treatment for Dox-induced cardiomyopathy (Dox-CM) [6,7]; treatments classically used for other HFs with systolic dysfunction induce only limited beneficial effects in Dox-CM.
As in different HF etiologies, Dox-CM is characterized by an alteration of adrenergic system [8]. However, at the present time, only few studies have examined the role of cardiac b 1 -and b 2adrenoceptor (b-AR) subtypes in the pathogenesis of Doxcardiotoxicity [9,10] and only one study, at late-onset Dox-CM, assessed b 3 -AR subtype [11], which is recently described as a new target for some b-blockers such as nebivolol [12,13]. Despite this lack in experimental data, some clinical studies investigated bblocker therapies in Dox-CM. Kalay et al., demonstrated that left ventricular (LV) diameters remained constant and diastolic function was better preserved after Dox-treatment in patients receiving carvedilol, compared to placebo [14]. However, Georgakopoulos et al., demonstrated that metoprolol, a b-blocker without antioxidative properties, failed to give cardioprotection in lymphoma-treated doxorubicin patients [15]. It was reported that an early start of treatment with angiotensin-converting enzyme inhibitors (ACEIs), in association or not with b-blockers, both improves myocardial contractility [16] and patients' prognosis [17]. Although the exact mechanism is still poorly understood, the authors highlighted the importance of an early diagnosis, because a delayed treatment (.6 months after the end of chemotherapy) is inefficient [17].
The aim of the present study was to identify a putative target (b-AR and/or its effectors) involved at the early phase of a chronic Dox-CM in a rat model. This study demonstrated for the first time that b 2 -AR expression was increased from 35 days after the first Dox-injection, this effect was maintained until 70 days after the first Dox-injection, resulting in an increase of b 2 -AR-induced contractility. In addition, b 1 -AR function was preserved, in spite of decreased b 1 -AR protein expression. This discrepancy could be explained by an increase of adenylyl cyclase (AC) expression and/ or activity as illustrated by an increased forskolin-induced response in Dox-CM rats.

Ethics Statement
All experiments were performed in accordance with the 1996 Guide for the Care and Use of Laboratory Animals published by the U.S. National Institute of Health. The protocol was approved by the Direction Départementale de la Protection des Populations (agreement number C-44 015) and all efforts were made to minimize suffering.

Animal Model
One hundred and fifty eight male Sprague-Dawley rats (225-250 g) were purchased from Janvier (Le Genest St Isle, France) and were housed under standard conditions of room temperature, humidity (40-60%) and 12 h light/dark cycle. Food and water were available ad libitum.
Doxorubicin (AdriblastineH 50 mg/25 mL, Pfizer, France) was administered intraperitoneally in six equal injections (each containing 2.5 mg.kg 21 ) over a period of two weeks, with a total cumulative dose of 15 mg.kg 21 body weight (Dox-CM: n = 100). Age-matched rats injected with saline were used as controls (Ctrl: n = 58). Rats were then bred in animal housing until three weeks after the last injection (day 35 (d35); Ctrl: n = 46, Dox-CM: n = 76) or until eight weeks after the last injection (day 70 (d70); Ctrl: n = 12, Dox-CM: n = 24). At d35, twenty Dox-CM rats and twenty Ctrl rats were randomly selected to be hemodynamically explored

Echocardiography-Doppler
Transthoracic echocardiography was performed using a commercially available ultrasound system (VIVID7, GE Healthcare, Horton, Norway) equipped with a 10 MHz sectorial probe. Rats were anaesthetized with a gas-mixture of 1% isoflurane (ForeneH, Abbott France, Rungis, France) in O 2 . The chest was shaved and the animal was positioned on a heating pad in a supine position. All recordings were monitored under a continuous single-channel electrocardiogram obtained on the imaging system by fixing the electrodes to the limbs. Using two-dimensional imaging, a short axis view of the LV at the level of the papillary muscles was obtained and the two-dimensionally guided M-mode recording through the anterior and posterior walls of the LV was taken as recommended by the American Society of Echocardiography [18]. Then, trans-mitral inflow in pulsed-wave Doppler from apical four chamber view and tissue Doppler imaging (TDI) on basal segments of septal and lateral walls in apical four chamber view were taken as previously described [19]. A cine-loop of LV parasternal short axis view with high frame rate was obtained. All acquisitions were performed by the same operator.
All images were digitally stored on hard disks for off-line analysis (EchoPac Q-analysis software, GE Healthcare). Measurements were made on five cardiac cycles and averaged for each data value. The following parameters were determined as recommended by the American Society of Echocardiography [18]: LV end diastolic and systolic diameters (LVEDD and LVESD), diastolic posterior wall thicknesses (dPWth). LV end diastolic and systolic volumes (LVEDV and LVESV) were calculated from the Teichholz method in order to assess LV ejection fraction (LVEF), whereas LV shortening fraction (LVSF) was calculated from LVEDD and LVESD previously measured. LV diastolic function parameters were derived from pulsed-wave trans-mitral inflow pattern and TDI off-line analyses as previously described [20]: the peak of E wave velocities, the isovolumic relaxation time (IVRT), the mean of peak velocities of basal septal and lateral walls (pulsed wave TDI) during systole (Sa) and in early diastole (Ea) to calculate E/Ea ratio. Radial 2D strain analyses were performed using the 2D speckle-tracking method on every medial myocardial segment [21].

Left Ventricle Catheterization
LV catheterization was performed, at d35, by a 2F microtip pressure catheter (SPR 838, Millar instruments Inc, Houston, Texas). Anaesthesia maintenance on spontaneously breathing rats was performed with an inhalational anaesthesia system for small animal (TEM anaesthesia, Lormont, France). Isoflurane was delivered through a nose mask at a concentration of 2% volume and 1 L.min 21 O 2 flow to limit hemodynamic repercussion. Body temperature was monitored by rectal probe and maintained constant (37uC) by warming-blanket. The right carotid artery was isolated, ligated at the proximal part and the pressure catheter was inserted in. Signals were recorded using an Analogic/Digital converter (EMKA Technologies, Paris, France), stored and displayed on a computer by the IOX1.5.7 Software System (EMKA Technologies). Data were analysed using Datanalyst software (EMKA Technologies). The following parameters were obtained: LV end diastolic pressure (LVEDP), LV contraction and relaxation velocities (dP/dt max and dP/dt min , respectively), the index of LV relaxation constant (Tau).
The animals were thereafter sacrificed by injection of a pentobarbital overdose.

Western-blot
The expressions of b 1 -AR, b 2 -AR, b 3 -AR, Gsa, Gia 2 and GAPDH were examined by western-blot, at d35 and d70. Hearts were rapidly isolated and placed in a cold Tyrode solution composed as followed (in mM): NaCl, 137; KCl, 5.4; MgCl 2 , 1.2; Na 2 HPO 4 , 1.2; Hepes, 20; CaCl 2 , 1.0; pH 7.4 (Sigma-Aldrich, St Quentin Fallavier, France). LV and septum free walls were carefully separated and freezed in liquid nitrogen and stored at 280uC until used. LV and septum free walls samples were homogenized in 3 mL of Ripa buffer for 1 g of tissue plus 1X PMSF (Interchim, Montluçon, France) and 2X protease inhibitor cocktail (Roche, Mannheim, Germany). Protein samples (25 mg) were submitted to electrophoresis on a 10% polyacrylamide/ sodium dodecyl sulfate gel and then were run 150 min at 20 mA per membrane in TG-SDS buffer (Interchim). Gels and nitrocellulose membranes (Hybond C super membrane, Amersham, Saclay, France) were equilibrated in TG-SDS buffer with 20% ethanol, and protein fractions were transferred using an electroblotting apparatus (Bio-Rad, Marnes La Coquette, France). Nonspecific  binding was blocked by incubating membranes in 5% non-fat dry milk in Tris-buffered saline (TBS) (200 mM Trizma base, 1.4 M NaCl, pH = 7.5) with 0.1% Tween 20 added (TBS-T) and then membranes were incubated with the primary antibody. Membranes were washed in 5% non-fat dry milk in TBS-T and hybridized with the secondary antibody in 1% non-fat dry milk in TBS-T. Finally, membranes were washed with 5% non-fat dry milk in TBS-T then TBS, and antibody complexes were revealed by the enhanced chemiluminescence detection process (Bio-Rad).
Chemiluminescence was visualized using an Amersham Image-Quant RT-ECL camera (GE Healthcare) and band signals were assessed by densitometry with ImageQuant TL software (GE Healthcare). For each lane, a ratio to the corresponding GAPDH band intensity was calculated; the use of GAPDH as reference was validated by checking the abundance stability of GAPDH protein between Ctrl and Dox-CM groups. Antibody references and conditions used in this study are summarized in Table 1.

Whole Heart Contractility
At d35, after removal, hearts were quickly flushed into a cold Tyrode buffer, and then were mounted on a cannula via the aorta on a Langendorff apparatus (EMKA Technologies) and perfused with a Krebs-modified solution (in mM: NaCl, 116; KCl, 5; MgSO 4 , 1.1; NaH 2 PO 4 , 0.35; NaHCO 3 , 27; glucose, 10; mannitol, 16; Na-Pyruvate, 2; CaCl 2 , 1.8) continuously oxygenated with a 95% O 2 , 5% CO 2 gas mixture to maintain a pH 7.4. Hearts were perfused at a constant flow-rate of 14 mL.min 21 at 37.060.4uC. After heart beatings became regular, left auricle was gently cut off to insert a small latex balloon, connected to a pressure transducer to record LV pressure, into the LV via mitral valves. Before starting record, the minimum pressure into the balloon was adjusted to 8-15 mmHg. Cardiac parameters were recorded using IOX1.5.7 software (EMKA Technologies): inotropism, lusitropism and chronotropism were respectively evaluated by measuring the contraction velocity (dP/dt max ), the relaxation velocity (dP/dt min ) and the heart rate (HR). After the stabilization period, specific b 1 -, b 2 -or b 3 -AR functions and AC stimulation were assessed by constructing concentration-response curves to different pharmacological agents. Thus, b 1 -and b 2 -AR functions were assessed by perfusing isoproterenol (a non selective b-AR agonist, 10 29 to 10 25 M) in the presence of 10 26 M L-748,337 (a selective b 3 -AR antagonist) associated to 10 26 M ICI-118,551 (a selective b 2 -AR antagonist) or to 10 26 M CGP-20712A (a selective b 1 -AR antagonist), respectively. b 3 -AR function was assessed by perfusing growing concentrations (10 29 to 10 25 M) of SR 58611A (a b 3 -AR agonist) and AC response was assessed by perfusing growing concentrations (3.10 28 to 10 25 M) of forskolin (an AC activator). In order to assess the role of Gi in the b 2 -AR pathway, we used pertussis toxin (PTX) (a selective Gi inhibitor). Classically, PTX is administered to rats two or three days before experiments [22]. However, due to high mortality in Dox-CM rats pre-treated with PTX, we were not able to apply this protocol and used another one consisting in pre-treating heart with PTX (4 mg.L 21 ) in a closed system during 30 min, as described by other groups [23]. Data were analysed using Datanalyst software (EMKA Technologies). After Langendorff experiments, hearts were weighed, as well as LV free walls which were carefully separated from the heart.

Data Analysis and Statistics
Data are presented as means 6 standard error to the mean (sem) of n experiments obtained from n different animals. In vivo and ex vivo cardiac parameters in baseline as well as protein expressions were compared using Student's t test for unpaired data. Concentration-response curves were compared with a twoway ANOVA (concentration, treatments) for repeated measures followed, when appropriated, by a Bonferroni's multiple comparisons test. Differences were considered significant if P,0.05.  France). All drugs were dissolved in distilled water, excepted isoproterenol which was solubilised in 1% acid ascorbic and L-748,337 and forskolin which were solubilised in dimethylsulfoxide (Sigma-Aldrich). The final concentration of the solvent in the organ bath was less than 0.1% v.v 21 and was used as controls for the effect of the active drug.

Mortality and General Characteristics of Animals
Dox-CM rats did not gain weight during the Dox-treatment, whereas the body weight of Ctrl rats increased regularly. After the   end of Dox-treatment, the rat body weight increased again, but remained significantly lower than those of Ctrl rats, and reached a plateau from 5 weeks after the end of Dox-treatment (day 49; Figure 1A). In addition, Dox-CM rats became lethargic compared to Ctrl one. The mortality rate was 25.0% at d35 and 58.3% at d70 in Dox-CM when no mortality was observed in Ctrl ( Figure 1B). At d35, all surviving Dox-CM rats presented a hepatomegaly and 79.4% of them presented ascites (30.865.8 mL), whereas at d70 no hepatomegaly was observed and ascites was present in 38.9% of Dox-CM rats (35.764.9 mL). Whole heart and LV weights from Dox-CM rats were significantly smaller than those of Ctrl animals both at d35 and d70, and they were still smaller after normalization to the tibia length. However, heart/body weight ratios and LV/body weight ratios were similar between Ctrl and Dox-CM rats at d35 when it was increased in Dox-CM group compared to Ctrl at d70 ( Table 2).

Cardiac Function Obtained by Echocardiography-Doppler
Echocardiography-Doppler analyses were performed at d35 and d70. Heart rate, measured before all echocardiographic acquisitions, was monitored in order to be similar between both groups (d35: Ctrl: 36466 bpm, Dox-CM: 37166 bpm, P = 0.427; d70: Ctrl: 36866 bpm, Dox-CM: 337617 bpm, P = 0.079). Representative images obtained by echocardiography in short axis view of a two-dimensionally directed M-mode at d35 are shown in Figure 2.

Cardiac Function Obtained by Left Ventricle Catheterization
As our study consists to identify molecular target at early stage of Dox-CM, we only performed LV catheterization at d35. Basal heart rate, measured before all pressure acquisitions, was increased in Dox-CM rats compared to Ctrl (P = 0.007) ( Table 3). Dox-CM rats presented a decreased Tau (210.062.2%; P = 0.009) albeit dP/dt max and dP/dt min were unchanged (Table 3). LVEDP was unchanged between both groups (P = 0.216).

b 3 -adrenoceptor Expression and Function
Using western-blot, we detected b 3 -AR protein at a band of 68 kDa (Figure 4A). At d35, the detected band presented a similar intensity in Dox-CM and Ctrl (P = 0.192), whereas, at d70, b 3 -ARs were down-expressed in Dox-CM hearts compared to Ctrl hearts (237.267.9%; P,0.001) (Figures 4A-C). In Ctrl isolated heart, at d35, b 3 -AR stimulation performed by cumulative increasing concentrations of SR 58611A produced no significant effect on heart rate and dP/dt max at any concentration (P = 0.239 and P = 0.385, respectively) ( Figures 4D, F) and decreased dP/dt min only at the higher concentration of SR 58611A (10 25 M) (P = 0.004) ( Figure 4E). Compared to Ctrl hearts, b 3 -AR stimulation in Dox-CM hearts produced a significant decrease in dP/dt max only at the higher concentration of SR 58611A (10 25 M) (P,0.001) ( Figure 4D) and had no effect on dP/dt min nor heart rate (P = 0.406 and P = 0.107, respectively) ( Figures 4E, F).

b 1 -adrenoceptor Expression and Function
By western-blot, the antibody directed against b 1 -AR detected a band at 72 kDa ( Figure 5A) whose intensity was significantly decreased in Dox-CM LV at d35 by 268.266.1% (P = 0.002 vs Ctrl) ( Figure 5B) and at d70 by 275.3614.0% (P,0.001 vs Ctrl) ( Figure 5C). In Ctrl heart, at d35, b 1 -AR stimulation induced a concentration-dependant increase in dP/dt max , dP/dt min and heart rate with a maximal effect observed at a concentration of 10 25 M (P,0.001, each) ( Figures 5D-F). The heart rate and the positive inotropic and lusitropic effects were unchanged in Dox-CM isolated hearts compared to Ctrl hearts (P = 0.544, P = 0.772 and P = 0.667, respectively) ( Figures 5D-F).

b 2 -adrenoceptor Expression and Function
By western-blot, the antibody directed against b 2 -AR detected a band at 55 kDa ( Figure 6A) whose intensity was significantly increased in Dox-CM LV at d35 by +57.9621.6% (P = 0.039 vs Ctrl) ( Figure 6B) and at d70 by +173.7635.1% (P,0.001 vs Ctrl) ( Figure 6C). At d35, b 2 -AR stimulation induced a similar concentration-dependant increase of heart rate between Ctrl and Dox-CM hearts (P = 0.695) ( Figure 6F). In Ctrl isolated hearts, b 2 -AR stimulation induced a concentration-dependant increase of dP/dt max and dP/dt min (P,0.001 and P = 0.005, respectively) ( Figures 6D, E) but at a lesser extent than b 1 -AR one. In Dox-CM hearts, the maximal effects induced by b 2 -AR stimulation on dP/ dt max and dP/dt min , compared to Ctrl, were both increased by +108.0618.2% (P,0.001) and +155.3626.8% (P,0.001), respectively ( Figures 6D, E), and reached a similar level than this produced by b 1 -AR stimulation.

Gs Protein Expression and Adenylyl Cyclase Stimulation
By western-blot, we detected two bands for Gsa protein expression: a short one at 45 kDa and a long one at 52 kDa ( Figure 7A). The expression of both forms was unchanged in Dox-CM hearts at both d35 and d70 (Figures 7B, C). In Ctrl and Dox-CM isolated hearts, at d35, forskolin induced a similar concentration-dependant increase of heart rate between Ctrl and Dox-CM hearts (P = 0.941) ( Figure 7F). In Ctrl group, forskolin induced a concentration-dependant increase of dP/dt max (P,0.001) and dP/dt min (P = 0.010) (Figures 7D, E). In Dox-CM isolated hearts compared to Ctrls, the forskolin effects on dP/dt max and dP/dt min were increased by +43.2610.3% (P = 0.004) and +61.1610.7% (P = 0.002), respectively ( Figures 7D, E).

Gi Protein Expression and its Involvement in b 2 -AR Cardiac Contractility
Gia 2 protein expression was detected by one band of 35 kDa ( Figure 8A) whose intensity was similar both in Ctrl and Dox-CM hearts at d35 (P = 0.184 vs Ctrl) ( Figure 8B), but was increased in Dox-CM hearts at d70 by +118.1637.6% (P = 0.009 vs Ctrl) ( Figure 8C). Ex vivo, at d35, Gi protein inhibition by PTX pretreatment produced no effect in heart rate in both groups (P = 0.346) ( Figure 8F). PTX-pretreatment induced a significant increase of b 2 -AR stimulation in Ctrl isolated hearts (dP/dt max : +64.469.8%; P,0.001; dP/dt min : +128.0620.8%; P,0.001) (Figures 8D, E) but induced no change of these two parameters in Dox-CM hearts (P = 0.173 and P = 0.451, respectively) ( Figures 8D, E).

Discussion
This study demonstrated for the first time in Dox-cardiotoxicity, b 2 -AR expression was increased at the early stage of the pathology, with an increase of b 2 -AR-induced contractility. Furthermore, b 1 -AR function was preserved in spite of decreased b 1 -AR protein expression and could be explained by an increase of AC expression and/or activity as illustrated by an increased forskolin-induced contractility in isolated hearts from Dox-CM rats.
In our study, the body weight of Dox-CM rats was smaller than in Ctrl one as previously described by other studies [8,24]. This has been related to a decrease in food consumption [25,26]. In our experimental conditions, we have, 35 days after the beginning of Dox-treatment, 25% of mortality and rats presented a large amount of ascites. Those observations were in agreement with other studies that reported a mortality rate fluctuating between 19% and 45% [27,28] and a volume of ascites generally varying from 30 mL to 140 mL [24,[28][29][30]. However, as Dox induced a multi-organ toxicity, it could be not excluded that mortality and ascites observed in rats could be due, at least in part, to liver and kidney damages. Seventy days after the beginning of Doxtreatment, mortality rate increased to 58% and ascites was still observed. Very few studies reported long term effects of Doxtreatment and data on mortality and ascites are very heterogeneous. Whereas a study observed 50% mortality rate 6 weeks after the last Dox-injection, associated with a very large amount of ascites [25] others did not observed mortality eight weeks after the last Dox-injection [11]. Also, others did not reported mortality and observed a decrease in ascites amount between days 40 and 70 [26].
During echocardiographic acquisitions, heart rate was monitored in order to have similar heart rate in both groups, allowing a comparison of echocardiographic parameters. At d35, Dox-CM rats presented a slight alteration of systolic function as illustrated by a mild decrease in LVEF and LVSF and in LV longitudinal deformation. Dox-CM rats presented also a mild diastolic dysfunction as suggested by the increased IVRT and Tau values. At d70, decreases in LVEF, LVSF and LV longitudinal deformation were exacerbated whereas diastolic dysfunction was unchanged. As diastolic dysfunction frequently occurs before systolic dysfunction [31][32][33], we could suggest that in our animal model, at d35, systolic function just began to be altered, meaning that this model corresponds to an early chronic Dox-cardiotoxicity. It is important to note that our results are very close to that observed in women who received anthracycline-treatment for breast cancer. Indeed, these patients had preserved LVEF compared to controls but a significantly reduced longitudinal strain with preserved radial contractility [34].
In human, Dox administered to adults (for breast cancer, for example) is well known to induce chronic cardiotoxicity where cardiac morphologic and functional alterations are close to those of dilated cardiomyopathy [3]. However, when administered in children, Dox can induce restrictive cardiomyopathy with normal LV dimensions [35]. Those findings, are consistent with those of other long-term follow-up studies conducted in other groups of anthracycline-treated childhood cancer survivors [36]. In our study, Dox was administered to young rats (7 week-old). At d35, we observed diastolic dysfunction with normal systolic function that could support the hypothesis of a restrictive-cardiomyopathy. At d70, the development of systolic dysfunction could suggest restrictive cardiomyopathy worsening with increase LV dimensions. Indeed, although LVEDDs were not statistically different between Ctrl and Dox-CM rats at d70, it is important to note that LV dPWth was significantly decreased in Dox-CM compared to Ctrl rats, strengthening the notion of LV morphologic changes in those rats. We could hypothesize that a period of 56 days posttreatment was a too short duration of the disease evolution to observe a significant LV dilation in Dox-CM rats. Indeed, in human HF, the dilation is considered as the last stage of ischemic or non ischemic cardiomyopathy evolution as in Dox-CM. In this latter disease, it is well known that cardiomyopathy and thus LV dilation occur up to decades after exposure [37].
In our Dox-CM model, we reported a remodeling in b-AR expression and function. Three weeks after last Dox-injection (d35), b 3 -AR protein expression was unchanged and a slight negative inotropic effect was obtained in Dox-CM hearts at the higher concentration of SR 58611A (10 25 M). Our data suggest that in early chronic Dox-cardiotoxicity, b 3 -AR could not play a major role in the cardiac alterations. At a later stage (d70), we observed a decreased b 3 -AR protein expression in Dox-CM hearts albeit another team reported an overexpression of b 3 -AR protein at the same stage [11]. Several hypotheses could explain those discrepancies. First, the rat strain used in the study is different (Wistar vs Sprague-Dawley) and several studies reported that protein expression could be different between Sprague-Dawley and Wistar rats [38,39] albeit no study compared b-AR subtypes expression between these two rat strains. Second, in the Sun's study, cardiac b-ARs expressions were assessed in Ctrl rats compared to Dox-treated rats which were also Sham-castrated (scrotum incision); no surgery being performed in Ctrl rats. Therefore, b-AR protein expressions were assessed in a Dox-CM model different from our one used in our study.
In our Dox-CM model, b 1 -AR protein expression was decreased at both stages. This result is in agreement with those reported for b 1 -AR protein expression both in rabbits [40] and in rats [8,11]. Surprisingly, albeit a decreased b 1 -AR expression at d35, the cardiac b 1 -AR responses were preserved. Several hypotheses could explain this feature. Firstly, a b 1 -AR reserve could be present and recruited in such conditions and, secondly, an increase in b 1 -AR signaling pathway can be evoked. We showed that Dox-treatment did not change Gsa protein expression, as it was reported in other study [8,11], but increased Figure 6. b 2 -AR expression and function are increased in Dox-CM hearts. A. Representative b 2 -AR immunoblotting at days 35 (d35) and 70 (d70). B. b 2 -AR protein quantification at day 35. C. b 2 -AR protein quantification at day 70. Protein levels were quantified using Amersham ImageQuant RT-ECL camera (GE Healthcare). The band signals were assessed by densitometry with ImageQuant TL software (GE Healthcare) and a ratio to the corresponding GAPDH band intensity was calculated. D. Cardiac inotropic (dP/dt max ), E. lusitropic (dP/dt min ) and F. chronotropic (HR) effects of increasing concentrations of isoproterenol (10 29  forskolin response, suggesting an increase in AC expression and/or activity. This surprising data contrasts with other studies which reported a decreased AC activity [41,42] or no change [43]. However, those differences seem to be due to the cardiotoxicity level since the Dox cumulative doses used were 24, 24.75 and 6 mg.kg 21 , respectively. Thus, the increased AC response to forskolin observed in our study could be a compensatory mechanism at early stage of Dox-cardiotoxicity to maintain cardiac contractility and could explain, at least in part, the preserved b 1 -AR response. However, as b 1 -ARs are involved in pro-apoptotic effects, b 1 -AR could be also responsible for some Dox cardiotoxic effects as previously suggested [9,10].
Finally, we reported, for the first time, in chronic Doxcardiotoxicity, a b 2 -AR protein overexpression. At d35, in Ctrl isolated heart, specific b 2 -AR stimulation induced positive inotropic and lusitropic responses but to a lesser extent than b 1 -AR one. This lower response could be explained by a more compartmentalized cAMP signaling of b 2 -AR [44]. Surprisingly, in Dox-CM hearts, b 2 -AR inotropic effect was increased and reached a similar level than b 1 -AR stimulation. This increase could be due to several mechanisms. Firstly, we show that b 2 -AR protein expression was increased in Dox-CM LV and could lead to an increased b 2 -AR response. Secondly, cardiac b 2 -AR could be linked both to Gs and Gi proteins [45,46]. In Dox-CM LV, Gia 2 protein expression was not modified at d35 but it was overexpressed at d70 of Dox-CM, as reported previously at this latter stage by Sun et al. [11]. Surprisingly, at d35, in Ctrl hearts, the Gi inhibition by PTX allowed to increase b 2 -AR response in order to reach the same level that obtained in Dox-CM in the absence of PTX treatment, whereas Gi inhibition in Dox-CM hearts slightly increased b 2 -AR response. These data suggest that, in Dox-CM, b 2 -AR was mainly linked to Gs.
To know whether enhanced b 2 -AR/Gs signaling observed in our study is a beneficial compensatory mechanism or a detrimental effect is still to be determined. To support the hypothesis of a beneficial effect of b 2 -AR agonist, it has been shown that b 2 -AR/Gi signaling can mediate cardiac protective effects by activating cell surviving pathway [10,47,48], and also in vivo studies reported beneficial effects of b 2 -agonist therapy in HF [49][50][51][52][53][54]. Concerning a putative detrimental effect of b 2 -AR, it could be proposed to use b 2 -AR blocker. However, it is important to note that b 2 -AR antagonists by blocking other physiological functions regulated by b 2 -AR, could also have detrimental effects: (i) a vasoconstriction because b 2 -AR is one of the most important vascular b-AR subtype whose activation induces a vasodilation in several vascular beds (including coronary vessels), (ii) a bronchoconstriction due to the reduction of bronchodilation induced by pulmonary b 2 -AR,… Thus, we rather suggest that further experiments must be performed to identify the most suitable effector of the cardiac b 2 -AR pathway. Then, according to the results of this next important study, we hope that we could propose a new therapeutic target for Dox-CM.
In conclusion, we have shown for the first time in rat Dox-CM, an increase of b 2 -AR protein expression at early stage of the pathology, associated to an increased contractility. Furthermore, we observed a preserved b 1 -AR function albeit b 1 -AR was reduced. Clinical studies have suggested no beneficial effect of metoprolol, a b 1 -AR blocker, in Dox-treated lymphoma patients [15] but a beneficial effect of nebivolol, a b 1 -AR blocker associated to a b 3 -AR agonistic properties against anthracycline-induced cardiomyopathy in patients with breast cancer [55]. However, the clinical roles of b 2 -AR remain to elucidate in order to determine whether b 2 -AR activation or blockade could prevent cardiac alterations due to anthracycline.