Introduction

Heart failure (HF) is the result of heart diseases and is a global threat to human health. Overactivation of the β1-adrenergic receptor (β1-AR) in cardiomyocytes is a core HF mechanism [1]. We and others have shown that in addition to the endogenous β1-AR ligand norepinephrine (NE), autoantibodies against β1-AR (β1-AA), harboring agonist-like effects, circulate in the sera of 30–40% of HF patients [2, 3]. Studies have demonstrated that β1-AA can directly lead to HF [4], suggesting the importance of β1-AA in HF development.

There are several potential approaches to protect against cardiac damage caused by β1-AA. β-blockers are widely used in treating chronic HF by blocking β1-AR activation [5]. However, the β1-AA-included pathological effects can only be partially abrogated by β-blockers, as suggested by animal [6] and clinical studies [7]. In vitro plasma exchange or immunoadsorption methods can increase the cardiac ejection fraction and restore cardiac function to a certain extent. However, these two schemes have poor specificity. Other approaches, such as peptide homologs mimicking the epitope(s) or aptamers, are still underway. One of the reasons for the shortcomings of treatment plans is our insufficient understanding of the characteristics of β1-AA-mediated activation of β1-AR.

In contrast to transiently activating β1-AR by NE, β1-AA can persistently activate β1-AR, which leads to the overactivation of β1-AR downstream signaling and cardiac injury. Existing studies have suggested that this mode of action of β1-AA may be related to decreased β1-AR endocytosis [8]. However, there is a lack of convincing evidence on the role of limited endocytosis in the sustained activation of β1-AR caused by β1-AA and the related mechanisms. β2-adrenergic receptor (β2-AR) is another important adrenergic receptor in cardiomyocytes [9]. Our previous studies unexpectedly found that although β1-AA does not directly bind to β2-AR, the sustained increase in NRCM-beating frequency induced by β1-AA can be neutralized by the β2-AR autoantibody-mediated activation of β2-AR [2], indicating cross-talk between β2-AR and β1-AR signals under the action of β1-AA. Similar to β1-AR, β2-AR can couple to Gs protein and stimulate adenylyl cyclase (AC) [9]. The difference is that β2-AR can also couple to Gi protein, which mediates an opposite effect [9]. For example, upregulated β2-AR expression negatively regulates the overcontraction of cardiomyocytes caused by β1-AR through the Gi-signaling pathway in HF [10]. Utilizing a hypothesis-driven approach, we speculated that the insufficient regulatory function of the β2-AR/Gi pathway might be involved in β1-AA-induced limited β1-AR endocytosis and persistent signaling. Strengthening the β2-AR/Gi pathway may be a new rescue strategy for β1-AA-positive HF.

Materials and methods

Animals

The animals used in this study were 6- to 8-week-old Balb/c mice and 0- to 3-day-old neonatal rats. The animals were fed in the SPF animal room of Capital Medical University. At the end of the experiment, the mice were euthanized with an intraperitoneal injection of 40 mg/kg sodium pentobarbital.

Patients

All patients gave informed consent. The detailed screening process is provided in the supplemental material. The information of patients was shown in Table S1.

Ethics statement

All animal experiments in this study complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. This study was approved by the Institutional Animal Care and Use Committee of Capital Medical University. The study complied with the journal’s applicable checklists for animal ethics (AEEI-2015-097, AEEI 2016-053).

The research study in patients complied with the Declaration of Helsinki and passed the approval of the Beijing Anzhen Hospital’s Ethical Review Committee.

NRCMs beating frequency and cyclic adenosine monophosphate measurements

Neonatal rats (0- to 3-day-old) were provided by the Capital Medical University animal laboratory. Neonatal rats were sacrificed by carbon dioxide inhalation. Hearts were immediately collected in iced-cold PBS. Ventricular cardiomyocytes were extracted by an enzymatic method and cultured in a cell incubator [11]. The spontaneous NRCMs contraction rate was monitored with a living cell station [2].

The cAMP content in NRCMs was measured by a cAMP [125I] radioimmunoassay (RIA) kit (RK-509, Hungary) [12]. The detailed experimental procedure was provided in the supplemental material.

Total internal reflection fluorescence microscopy

Total internal reflection fluorescence microscopy (TIRF) was utilized to determine β1-AR endocytosis in HL-1 cells transfected with β1-AR-GFP. The myocardial cell line HL-1 was cultured in confocal dishes. A TIRF inverted microscope (Olympus, Japan) equipped with an EMCCD camera and an oil immersion objective lens (×100 magnification, NA = 1.49) was utilized. The depth of field, 120-130 nm, was selected to observe β1-AR fluorescence upon the plasma membrane but not the cytoplasm. Changes in cell surface fluorescence were recorded at 0, 5, 10, and 30 minutes (min) after cell administration. Recordings were made every 30 seconds (s) using MetaMorph software version 7.8.8.0. Image fluorescence intensity was analyzed using ImageJ 1.51J8 [13]. To detect the recruitment of β1-AR-GFP and β-arrestin-RFP, we recorded images at 5 min after drug treatment and utilized ImageJ 1.53c for colocalization analysis. Before analysis, all images were adjusted with ImageJ 1.53c to the same brightness and color. Pearson’s R value (above threshold) was used as an indicator of the colocalization of β1-AR-GFP and β-arrestin-RFP [14].

Intracellular calcium ion detection

HL-1 cells were cultured at a density of ~30% in confocal dishes. The cells were incubated with the Ca2+ fluorescent probe Fluo-4 (10 μM, F14201, Thermo Fisher, USA) for 90 min before the experiment. At the time of the experiment, the cell culture medium was changed to FluoroBrite DMEM (A1896701, Thermo Fisher, USA). Intracellular Ca2+ levels were measured using a turntable confocal microscope (UltraVIEW VoX, USA) [13].

Antibodies

Please refer to supplementary table S2 for detailed information.

Establishment of a mouse heart dysfunction model by injecting β1-AA

A mouse monoclonal β1-AA (synthesized from AbMax Biotechnology, China) [15] was mixed with saline before use (concentration of 0.5 mg/ml). Heart dysfunction models were established in male 6 to 8-week-old Balb/c mice randomly by intraperitoneal β1-AA injection (10 μg/kg, every two weeks, i.p.). In β2-AR inverse agonist treatment group, 32.7 μg/kg ICI118551 was injected intraperitoneally every week for 2 months, accompanied by β1-AA injection every two weeks. Left ventricular function was determined in all mice by echocardiography, and morphological changes in the heart were observed by HE staining.

Statistical analysis

In this study, GraphPad Prism 8.0.2 software was used for statistical analysis and statistical graph production. The experimental results are expressed as the mean ± SEM. The differences in dynamic changes in NRCM-beating frequency, cAMP concentration, and β1-AR endocytosis in the different groups were analyzed by 2-way ANOVA followed by a Bonferroni post hoc test. One-way ANOVA followed by Dunnett’s test was used to compare multiple groups, and the t-test was used to compare the mean between two groups. P < 0.05 was considered statistically significant.

The detail of other material and methods was described in the supplementary material.

Results

Similar to β1-AA (+)-IgG from patients, mouse-derived monoclonal β1-AA induces sustained activation and limited endocytosis of β1-AR

β1-AA and negative IgG used in the current study originated from mouse-derived hybridoma cells and sera from healthy Balb/c mice, respectively. We first compared the effect of mouse-derived β1-AA with that from HF patients. As shown in Fig. 1, patients’ β1-AA (+)-IgG (β1-AA-positive total IgG, 10−7 M) elicited persistent activation of β1-AR, evidenced by the increasing NRCM-beating frequency for 1 hour, an effect partially inhibited by the β1-AR blocker metoprolol (Met, 10−5 M) and completely neutralized by the peptide corresponding to the extracellular second loop of β1-AR (β1-AR-ECII, 10−5 M) (Fig. 1A, Table S3). After preincubating NRCMs with phenoxybenzamine (Phe, 10−5 M) for 30 min to block the α1-adrenergic receptor (α1-AR), NE (10−5 M) increased the NRCM-beating frequency for a much shorter time, ~5 min (Fig. 1A). The mouse monoclonal β1-AA (10−7 M) also continuously increased NRCM-beating frequency through β1-AR, whereas negative IgG had no effect (Fig. 1B, Table S4). Moreover, monoclonal β1-AA increased the Ca2+ level in HL-1 cells, which lasted until the end of the observation (300 s). This effect could be blocked by the β1-AR-ECII peptide. In contrast, after NE stimulation, the intracellular Ca2+ increased rapidly and then decreased rapidly, maintaining for only 5 s (Fig. 1C, Table S5, Video S1, and S2). Met, Phe, and the β1-AR-ECII peptide alone had no effects on either NRCM-beating or intracellular Ca2+ levels (Fig. S1A, B, Table S7 and S8). These results indicated that the mouse monoclonal β1-AA and the β1-AA from HF patients acted very similarly, and the former was used in subsequent experiments.

Fig. 1: Monoclonal β1-AA purified from mouse-derived hybridoma cells caused long-term activation of β1-AR but restricted β1-AR endocytosis.
figure 1

A and B NRCMs were utilized to evaluate the effects of β1-AA (+)-IgG (β1-AA-positive total IgG) isolated from HF patients and mouse-derived monoclonal β1-AA on β1-AR activation. β1-AA (-)-IgG (β1-AA-negative total IgG) from HF patients and negative IgG from healthy mice were treated as the antibody controls. NE acted as a physiological ligand control. Phe and Met were α-adrenoceptor (α-AR) and β1-AR blockers, respectively. The peptide-based on β1-AR-ECII could specifically neutralize β1-AA. Data were analyzed by two-way ANOVA, followed by a Bonferroni post hoc test. n = 4–6/group, *P vs β1-AA, †P vs β1-AA (-)-IgG (A) or negative IgG (B), ‡P vs β1-AR-ECII, P < 0.05. C HL-1 cells were incubated with Fluo-4 (10 μM) for 90 min and then treated with different stimuli to dynamically measure the intracellular Ca2+ changes. n = 3/group. D–F TIRF was used to determine the endocytosis of β1-AR on the membrane of HL-1 cells, reflected by the decreased percentage of fluorescence value of β1-AR-GFP. D is the schematic diagram of TIRF detection. E and F are typical pictures and statistical charts of the effects of different stimuli on β1-AR-GFP endocytosis, respectively. Scale bar = 10 μm. Data were analyzed by two-way ANOVA, followed by a Bonferroni test. n = 3/group. *P vs β1-AA, P < 0.05.

Full size image

G protein-coupled receptor (GPCR) endocytosis can terminate the downstream signal over time. To investigate the potential role of endocytosis in long-term β1-AR activation, we transiently transfected the β1-AR-GFP plasmid into an HL-1 cell line (Fig. S2A, B, Fig. S14 A1, A2). TIRF was used to determine the degree of β1-AR endocytosis reflected by the reduced β1-AR-GFP fluorescence intensity on the cell membrane. TIRF detection showed that the fluorescence value was reduced by 28% at 5 min after treatment with NE plus Phe and by 44% and 67% after 10 min and 30 min (Fig. 1D–F, Fig. S20), respectively, indicating marked β1-AR endocytosis. However, the fluorescence intensity showed no significant change in the β1-AA treatment group or in the plasmid transfection control and negative IgG groups (Fig. 1F, Table S6).

Weakened β1-AR endocytosis is an important factor in the continuous activation of β1-AR

To clarify the causal relationship between endocytosis and β1-AR activation duration, we assessed the changes in β1-AR-mediated signaling and the effect after decreasing or increasing β1-AR endocytosis. As shown in Fig. 2A, after blocking α1-AR with Phe (10−5 M), NE (10−5 M) induced a short-term increase in NRCMs beating and intracellular cyclic adenosine monophosphate (cAMP) levels. In contrast, after preincubation with the GPCR endocytosis inhibitor Dynasore (10−4 M) for 30 min, NE increased the beating frequency of NRCMs for 60 min and the cAMP level for 30 min, similar to that caused by β1-AA (Fig. 2A, B, Tables S9 and S10). The result of Dynasore treatment alone was negative for the NRCMs beating (Fig. 2A).

Fig. 2: Inhibiting/promoting β1-AR endocytosis enhanced/weakened the sustained activation of β1-AR.
figure 2

A The effect of the receptor endocytosis inhibitor Dynasore on the NE-induced increase in NRCMs beating frequency. n = 5/group, *P vs NE + Phe, P < 0.05. B An radioimmunoassay kit was used to detect the cAMP concentration in NRCMs after treatment with different stimuli. n = 4/group. *P vs NE + Phe, †β1-AA vs negative IgG, P < 0.05. C–F After overexpressing β-arrestin1 (β-arr1 OE) or β-arrestin2 (β-arr2 OE), the effect of β1-AA on β1-AR-GFP endocytosis was measured via TIRF. HL-1 cells were transiently transfected with β1-AR-GFP, combined with or without transfection with β-arrestin1-RFP (C and D) or β-arrestin2-RFP (E and F). C and E are typical pictures, and D and F are the corresponding statistical charts. Scale bar = 10 μm. n = 3–4/group. ‡P vs β1-AA, P < 0.05. Data in A–F were analyzed by two-way ANOVA with a Bonferroni post hoc test. G and H The impacts of β1-AA on intracellular Ca2+ in HL-1 cells with or without β-arr1 OE or β-arr2 OE. Fluo-4 (10 μM) was used as a Ca2+ indicator. n = 3–4/group.

Full size image

β-arrestin1/2 are key scaffold proteins that initiate β1-AR endocytosis. Bioluminescence resonance energy transfer detection showed that it was difficult for β1-AR to recruit β-arrestin1 or β-arrestin2 upon β1-AA stimulation (Fig. S3 A-G, Fig. S14 B1-D2, Tables S15 and S16). Therefore, we overexpressed β-arrestin1-RFP (β-arr1 OE) or β-arrestin2-RFP (β-arr2 OE) in HL-1 cells transiently transfected with β1-AR-GFP to improve β1-AR endocytosis (Fig. S4 A, B, Fig. S15 A1-C2). As determined by TIRF, either β-arrestin1-RFP or β-arrestin2-RFP overexpression markedly enhanced β1-AA-induced β1-AR endocytosis. At 30 min after β1-AA administration (10−7 M), β1-AR-GFP fluorescence decreased by 42% in the β-arr1 OE group (Fig. 2C, D, Fig. S21 A, Table S11) and by 25% in the β-arr2 OE group (Fig. 2E, F, Fig. S21 B, Table S12). Moreover, both β-arrestin1-RFP and β-arrestin2-RFP overexpression significantly attenuated the sustained elevated levels of Ca2+ in HL-1 cells caused by β1-AA (Fig. 2G, H, Tables S13 and S14), suggesting that the prevention of enhanced β1-AR endocytosis led to the continuous activation of β1-AR. We concluded that β1-AA-initiated sustained activation of β1-AR could be at least partially attributed to insufficient β1-AR endocytosis.

Insufficient regulation of β2-AR is involved in the limited endocytosis and sustained activation of β1-AR caused by β1-AA

β2-AR regulates β1-AR signaling by interacting with it. To assess the possible role of β2-AR in β1-AA-restricted endocytosis, we first observed the effect of β1-AA on the binding of β1-AR to β2-AR. Purified human β1-AR and β2-AR proteins were isolated and used in the following experiment. The 300 nM purified β2-AR protein was coupled on the surface of a chip, and different concentrations (58.75 nM - 1880 nM) of purified β1-AR protein were added. Surface plasmon resonance (SPR) detection showed that purified β1-AR and β2-AR proteins could bind directly in a dose-dependent manner; the KD value was 2.147 μM (Fig. 3A, Table S17). β1-AA reduced the binding of β1-AR to β2-AR (Fig. 3B, Table S18) with an IC50 of 1.16 ± 0.7 μM (Fig. 3C, Table S19). Moreover, β1-AA did not bind directly to β2-AR. These results suggested that β1-AA might attenuate the regulatory effect of β2-AR on β1-AR.

Fig. 3: The ability of β2-AR to promote β1-AR endocytosis was reduced upon β1-AA stimulation.
figure 3

A The binding ability of purified β1-AR to β2-AR proteins was detected by SPR. The β2-AR protein was first coupled to the chip, and then different concentrations of β1-AR proteins were added. B SPR was used to observe the binding of β1-AA and β2-AR protein and the influence of β1-AA on the binding of β1-AR and β2-AR. C The IC50 of β1-AA on binding β1-AR and β2-AR. D, E TIRF was used to assess NE-induced β1-AR endocytosis when HL-1 cells were transfected with siRNA β2-AR or an siRNA control. Representative images and quantification of changes in β1-AR-GFP fluorescence density are shown in D and E, respectively. Scale bar = 10 μm. n = 3/group. *P vs NE, P < 0.05. F, G: Changes in β1-AR endocytosis caused by β1-AA in HL-1 cells with and without the overexpression of β2-AR-EYFP (β2-AR OE). Scale bar = 10 μm. n = 3/group. †P vs β1-AA, P < 0.05. Data in D–G were analyzed by two-way ANOVA with the Bonferroni test. H The effects of β1-AA on intracellular Ca2+ in HL-1 cells with and without the overexpression of β2-AR-EYFP (β2-AR OE). Fluo-4 (10 μM) acted as a Ca2+ indicator. n = 3/group.

Full size image

To investigate the influence of the insufficient regulatory function of β2-AR on β1-AR endocytosis, endogenous β2-AR expression in HL-1 cells was inhibited by transfection with siRNA β2-AR (Fig. S5A, B, Fig. S15 D1, D2). As observed with TIRF, after stimulation with NE (10−5 M) for 5 min, 10 min, and 30 min, the fluorescence intensities of β1-AR-GFP on the cell membrane were reduced by approximately 1%, 5%, and 11%, respectively, in the siRNA β2-AR group, which was markedly lower than that of the siRNA control group (reduced by 6%, 27%, and 29%). This indicated that the lack of β2-AR restricted β1-AR endocytosis (Fig. 3D, E, Fig. S22 A, Table S20).

To further clarify the influence of β2-AR on β1-AA-induced limited β1-AR endocytosis, β2-AR-EYFP was overexpressed in HL-1 cells transiently transfected with β1-AR-GFP (Fig. S6A, B, Fig. S16 A1, A2). As detected by TIRF, β1-AA (10−7 M) administration decreased the fluorescence intensities of β1-AR-GFP by 14%, 24%, and 27% at 5 min, 10 min and 30 min after β2-AR overexpression, respectively. These results indicated the promoting effect of the increased β2-AR expression on β1-AA-induced β1-AR endocytosis (Fig. 3F, G, Fig. S22 B, Table S21). We further observed the impact of β2-AR on the sustained activation of β1-AR signaling. The β2-AR-EYFP plasmid was transfected into HL-1 cells, and a Fluo-4 fluorescence probe was used to identify intracellular Ca2+ changes. As shown in Fig. 3H (Table S22), compared with the nonoverexpression group, β2-AR-EYFP overexpression significantly reduced the amplitude increase in Ca2+ induced by β1-AA and delayed the time of increase in Ca2+. Transfection of β2-AR-EYFP alone or the addition of negative IgG had no effect on β1-AR endocytosis or Ca2+ levels in HL-1 cells. The above results suggested that a lack of β2-AR regulation might be responsible for weakened endocytosis and the long-term activation of β1-AR caused by β1-AA.

Biased activation of β2-AR/Gi promotes β1-AR endocytosis and then inhibits the persistent activation of β1-AR elicited by β1-AA

Accumulating reports have demonstrated that β2-AR negatively regulates β1-AR by coupling with Gi. We demonstrated that as a specific β2-AR antagonist, ICI118551 (10−5 M) also acted as a biased agonist of the β2-AR/Gi-signaling pathway (Fig. S7A and B, Fig. S16 B1-B3). This effect could be inhibited by the Gi inhibitor pertussis toxin (PTX, 1.5 μg/ml, pretreatment for 13 h) (Figure S7C and D, Figure S17 A1 and A2). The cultured NRCMs were treated with β1-AA (10−7 M) with or without ICI118551 (10−5 M). As shown in Fig. 4, the β1-AA-induced increased beating frequency and the intracellular cAMP content of NRCMs were significantly relieved by ICI118551 supplementation, but ICI118551 alone had no effects on these parameters (Fig. 4A, B, Table S23 and S24). Furthermore, the continuous increase in intracellular Ca2+ induced by β1-AA disappeared after preincubation with ICI118551 in HL-1 cells (Fig. 4C, Table S25). Then, HL-1 cells transfected with β1-AR-GFP were used to detect β1-AR endocytosis via TIRF. After HL-1 cells were preincubated with ICI118551 for 5 min and subsequently stimulated with β1-AA, β1-AR-GFP fluorescence was reduced by 17%, 21%, and 24% at 5 min, 10 min, and 30 min, respectively, indicating marked β1-AR endocytosis (Fig. 4D, E, Fig. S22 C, Table S26). However, the above effects related to ICI118551 supplementation disappeared after interference with the expression of endogenous β2-AR (Fig. 5A–C, Table S27, Fig. S23A, Table S28) or with PTX (1.5 μg/ml) treatment (Fig. 5D–F, Table S29, Fig. S23B, Table S30) in HL-1 cells. We further treated HL-1 cells overexpressing β2-AR with PTX. The results showed that after PTX inhibited Gi, the effects of β2-AR overexpression on promoting β1-AR endocytosis and inhibiting the intracellular Ca2+ increase caused by β1-AA disappeared (Fig. 5G–I, Table S31, Fig. S23C, Table S32). To identify the role of β2-AR/Gi-signaling in NE-induced endocytosis of β1-AR, we tested the β1-AR endocytosis in HL-1 cells overexpressed Gi by TIRF. The results showed that overexpression of Gi did not enhance the endocytosis of β1-AR upon NE stimulation (Fig. S8A, B, Fig. S24 A, Table S33). In summary, activating the β2-AR/Gi-signaling pathway reversed β1-AA-induced weakened β1-AR endocytosis, leading to the termination of β1-AR signaling over time (Fig. 4F).

Fig. 4: By activating β2-AR/Gi, ICI118551 promoted β1-AA-induced β1-AR endocytosis and prevented continuous β1-AR activation.
figure 4

A, B The impact of ICI118551 on β1-AA-induced increased beating frequency and cAMP level of NRCMs, respectively. ICI118551 was added 10 min after administration of β1-AA. n = 3–6/group. *P vs β1-AA + ICI118551, P < 0.05 (two-way ANOVA with Bonferroni test). C The reversal effect of ICI118551 on the persistent rise in intracellular Ca2+ caused by β1-AA in HL-1 cells. HL-1 cells were pretreated with Fluo-4 for 90 min, ICI118551 was added, and then β1-AA was added after 5 min. n = 3/group. D and E The reversal effect of ICI118551 on the limited endocytosis of β1-AR caused by β1-AA. Scale bar = 10 μm. n = 3–6/group. †P vs β1-AA, P < 0.05 (two-way ANOVA with Bonferroni test). F illustrates that β2-AR/Gi-biased activation by ICI118551 could reverse the limited endocytosis and sustained activation of β1-AR induced by β1-AA.

Full size image
Fig. 5: siRNAβ2-AR or PTX blocked the endocytosis-promoting effect of β2-AR/Gi-signaling pathway for β1-AR.
figure 5

A The effect of ICI118551 on β1-AA-induced Ca2+ change in HL-1 cells after interfering with β2-AR (siRNA β2-AR). n = 3/group. Because the experiments in Fig. 4C and Fig. 5A were done at the same time, the same β1-AA (WT) group in Fig. 4C were re-used as control here. B and C The effect of ICI118551 on β1-AA restricting β1-AR endocytosis in HL-1 cells which were transfected with siRNA β2-AR. Scale bar = 10 μm. n = 3–6/group. C the β1-AA (WT) group represented that β1-AA stimulated β1-AR overexpressed HL-1 cells. Because the experiments in Fig. 4E and Fig. 5C were done at the same time, the same β1-AA (WT) group in Fig. 4E were re-used as control here. D The effect of ICI118551 on β1-AA-induced Ca2+ change in HL-1 cells pretreated with PTX (1.5 μg/ml, 13 hours). n = 3/group. E, F The effect of ICI118551 on β1-AA restricting β1-AR endocytosis in HL-1 cells, which were pretreated with PTX. Scale bar = 10 μm. n = 3/group. G The effects of β1-AA on intracellular Ca2+ when the HL-1 cells overexpressed β2-AR-EYFP (β2-AR OE) and pretreated with PTX. n = 3/group. H, I, The change of β1-AR endocytosis caused by β1-AA after the HL-1 cells overexpressing β2-AR-EYFP (β2-AR OE) and pretreated with PTX. Scale bar = 10 μm. n = 3/group. *P vs β1-AA + PTX, P < 0.05. Data of C, F, and I were analyzed by two-way ANOVA with Bonferroni test. B, E and H were the representative images. C, F and I were the quantitative graphs, respectively.

Full size image

β2-AR/Gi promotes β1-AR endocytosis by activating GRK2

β-arrestin1/2 initiates GPCR endocytosis after binding to receptors phosphorylated by intracellular protein kinases. G protein-coupled receptor kinase 2 (GRK2) and protein kinase A (PKA) are two predominant kinases that promote β1-AR phosphorylation in cardiomyocytes. To explore the mechanism of promoting β1-AR endocytosis after the biased activation of β2-AR/Gi, we tested the activities of GRK2 and PKA in HL-1 cells. As shown in Fig. 6, both β1-AA (10−7 M) and ICI118551 (10−5 M) costimulation and ICI118551 single stimulation increased GRK2 phosphorylation (Fig. 6A, Fig. S17 B1, B2), whereas negative IgG (10−7 M) stimulation showed no significant effect (Fig. 6A). However, neither β1-AA or ICI118551 alone nor their combination affected PKA phosphorylation (Fig. 6B, Fig. S18 A1, A2). These results indicated that GRK2 activation might promote the recruitment of β-arrestin and β1-AR, subsequently triggering β1-AR endocytosis. Therefore, we detected the recruitment of β1-AR-GFP and β-arrestin1/2-RFP on the surface of the HL-1 cell membrane by TIRF. The results showed that owing to the activity of ICI118551, β1-AA stimulation increased the recruitment of β1-AR and β-arrestin. During β1-AA stimulation, the Pearson coefficients (Pearson’s R values) that measured the colocalization of β1-AR with β-arrestin1 and β-arrestin2 were 0.160 ± 0.024 and 0.330 ± 0.083 as the mean ± SEM, respectively, whereas in the ICI118551 plus β1-AA stimulation group, the Pearson coefficient was increased to 0.687 ± 0.110 (β-arrestin1) and 0.753 ± 0.068 (β-arrestin2) as the mean ± SEM (Fig. 6C, D). In addition, after interference with GRK2 expression in HL-1 cells, the reversal effect of ICI118551 on β1-AA-restricted β1-AR-GFP endocytosis disappeared (Fig. S9 A-D, Fig. S18 B1, B2, Fig. S24 B, Table S34). These results suggested that after the biased activation of the β2-AR/Gi-signaling pathway, ICI118551 could activate GRK2, thereby increasing the recruitment of β1-AR and β-arrestin, promoting receptor endocytosis and terminating the continuous activation of β1-AR under the action of β1-AA.

Fig. 6: Activation of GRK2 enhanced the recruitment of β1-AR and β-arrestin stimulated by β1-AA.
figure 6

A The phosphorylation of GRK2 was detected after HL-1 cells were stimulated with ICI118551 alone or in combination with β1-AA for 5 min. n = 4/group. *P vs β1-AA, P < 0.05 (one-way ANOVA with Dunnett test). B The detection of PKA phosphorylation. n = 4/group. C, D TIRF was used to detect the recruitment of β1-AR and β-arrestin1/2 (above pictures), and Pearson’s coefficient (below tables) was calculated to reflect the colocalization between the two proteins. White arrows indicate the colocalization of β1-AR with β-arrestin1 (C) or β-arrestin2 (D). n = 3/group.

Full size image
Fig. 7: Work model.
figure 7

Under the action of β1-AA, the sustained activation of β1-AR is due to the decrease of endocytosis. Biased activation of β2-AR/Gi/GRK2 signaling pathway can promote the endocytosis of β1-AR that restricted by β1-AA, leading to the termination of β1-AA-induced continuous activation of β1-AR and improvement of cardiac structure and function.

Full size image

Biased β2-AR/Gi activation ameliorates cardiac structural damage and dysfunction in β1-AA-positive mice

To verify that activating the β2-AR/Gi-signaling pathway protected β1-AA-induced cardiac injury in vivo, wild-type Balb/c male mice (6–8 weeks old) without β1-AA were divided into four groups: saline (3.7 μg/kg, once a week, i.p.), β1-AA (10 μg/kg, every 2 weeks, i.p.), and therapy groups β1-AA combined with ICI118551 and ICI118551 alone (32.7 μg/kg, once a week, i.p.) (Fig. S10A). Eight weeks later, the β1-AA levels detected by enzyme-linked immunosorbent assay were significantly higher in the β1-AA and therapy groups than in the saline control group (Fig. S11, Table S35). M-mode echocardiography detection showed that the left ventricular ejection fraction (LVEF) decreased by 12.6% after 8 weeks of β1-AA treatment, and the left ventricular end-diastolic diameter (LVEDD) and left ventricular end systolic diameter (LVESD) increased by 15% and 42%, respectively. However, LVEF, LVESD, or LVEDD in the saline control group and the ICI118551 treatment group showed no significant change compared with preadministration measures (Fig. S10B, C, Table S36). Hematoxylin–eosin (HE) staining showed broken myocardial fibers in the β1-AA group, which were significantly alleviated after treatment with ICI118551 (Figure S10D). These results suggested that β1-AA caused cardiac dysfunction and structural damage in mice, and ICI118551, a biased agonist of β2-AR/Gi, had a certain therapeutic effect on cardiac injury caused by β1-AA.

Discussion

Previous studies have shown that β1-AA promotes the occurrence and development of HF by continuously activating β1-AR. However, the specific mechanism is not fully understood. Receptor endocytosis is one of the important mechanisms for terminating signal transduction. The main study findings were that β1-AA continuously activated β1-AR by restricting β1-AR endocytosis and that biased activation of the β2-AR/Gi pathway could activate GRK2, promote β1-AR endocytosis restricted by β1-AA in vitro and alleviate the morphologic and functional damage in the mouse heart in vivo Fig. 7.

β1-AA is produced against β1-AR-ECII [16]. In 1987, Wallukat et al. [17] found this autoantibody in dilated cardiomyopathy for the first time. Subsequently, high positivity rates of β1-AA have been detected in a variety of cardiovascular diseases, especially in HF [3, 18, 19]. Therefore, the relationship between β1-AA and HF has gradually attracted researchers’ attention. β1-AA elicits various pathological effects, such as triggering cardiomyocyte apoptosis [20], reducing the effective refractory period of the atrium and inducing atrial fibrillation [18], and inducing calcium overload [21], to cause cardiac damage. It is worth noting that continuously activating β1-AR is a common mechanism underlying these effects. Elucidating the related mechanism is conducive to drug development and clinical treatment. As an agonistic ligand control, NE was selected. It is an endogenous α1/β1-AR selective agonist belonging to the catecholamine category.

As a GPCR, the activation of β1-AR is a complicated process. The agonist ligand binds to β1-AR and induces its structural change, which in turn exposes the Gs protein-binding sites. The Gs protein undergoes the GTP cycle to convert to GTP-Gs and binds to β1-AR. Then, GTP-Gs activates AC and mediates signal transduction through cAMP-PKA [9, 22, 23]. To avoid the cytotoxicity caused by sustained GPCR activation, activated β1-AR needs to be desensitized and endocytosed over time. GPCR desensitization is divided into transient desensitization and long-term desensitization. Transient desensitization proceeds through the interaction of β-arrestin and GPCRs over a short period of time (minutes). Long-term desensitization proceeds through GPCR degradation in the lysosome over a long period of time (hours or days), during which the mRNA level of GPCR is reduced [24]. In this study, all experiments were completed in a short time; thus, receptor desensitization involved transient desensitization. The process is roughly as follows: GRK phosphorylates activated β1-AR, and then, β-arrestin1/2 in the cell recognizes and binds to phosphorylated β1-AR, blocking the binding between Gs protein and β1-AR and promoting the endocytosis of β1-AR [24, 25]. β-arrestin is a key protein in receptor endocytosis. Upon receptor activation, β-arrestin1/2 promotes receptor desensitization and endocytosis by binding to phosphorylated receptors. Overexpression of β-arrestin has been proved to be an effective way to enhance their recruitment to the receptor and promote receptor endocytosis [26]. In the present study, overexpressing β-arrestin promoted β1-AA endocytosis restricted by β1-AA in HL-1 cells.

To detect the β1-AR activation and endocytosis, NRCMs and HL-1 cells were used. NRCMs are rhythmic and their contraction profile just meets the needs of beating frequency experiments [27]. However, the rhythmic contraction will greatly disturb the TIRF observation and intracellular Ca2+ determination. Thus, a mouse myocardial HL-1 cell line, with good adhesion, was chosen as well. It has been studied that HL-1 cells can be used as a tool cell for β1-AR and β2-AR signaling pathways [28,29,30]. Here we confirmed that the proteins of β1-AR, β2-AR, β-arrestin1, and β-arrestin2 were expressed in HL-1 cells (Fig. S12A, Fig. S18 C1–D4). Furthermore, when given NE stimulation, the NRCMs need to be preincubated with Phe to block α1-AR, whereas HL-1 cells do not, because NE has a higher affinity for α1-AR than β1-AR in cardiomyocytes, whereas HL-1 cells do not express α1-AR (Fig. S12 B, Fig. S18 E1, E2).

In HL-1 cells, the change in cytosol Ca2+ upon NE stimulation was transient. However, a long time of elevated cytosol Ca2+ was observed under the β1-AA condition. The sustained high concentration of Ca2+ in the cytosol of cardiomyocyte was a fascinating phenomenon, which we cannot fully explain yet. We speculated that it may be owing to the change of activity of Ca2+ channels or calcium ATPase in sarcoplasmic reticulum when stimulated by β1-AA.

β2-AR is also important in cardiomyocytes because it can bind to either Gs or Gi protein [9, 31]. By coupling to the Gi protein, β2-AR inhibits AC activation to block β1-AR signaling and exerts cardioprotective effects [31]. In addition, β2-AR can target the C-terminus of β1-AR in a phosphodiesterase 4-dependent manner to limit signal diffusion to avoid cytotoxicity after β1-AR activation [32]. However, some studies found that the formation of the β1-β2-AR heterodimer strengthens the Gs signaling pathway in normal adult mouse cardiomyocytes and restricts β2-AR endocytosis [33, 34]. These reports point out the complexity of the relationship between β1-AR and β2-AR in different situations. Here, we verified that β2-AR could promote β1-AR endocytosis upon stimulation with either β1-AA or NE.

ICI118551 is used as a specific antagonist of β2-AR [35]. However, our study and other studies have found that ICI118551 preferentially activates β2-AR/Gi [10, 36]. ICI118551 is reported to have a protective effect on the cardiovascular system after activating the β2-AR/Gi pathway by reducing the contraction of myocytes in failing hearts [10] and reducing pulmonary artery tension [36]. We demonstrated that by activating the β2-AR/Gi pathway, ICI118551 could increase GRK2 activity, enhance the recruitment of β-arrestin1/2 to β1-AR, and improve β1-AA-induced β1-AR endocytosis. GRK2 is the predominant GRK and the primary regulator of β1-AR and β2-AR desensitization in the heart. GRK2 can phosphorylate activated β-adrenergic receptors, which are recognized by β-arrestin and initiate receptor endocytosis [37]. Contrary to expectation, we found that β1-AA upregulated β1-AR phosphorylation, whereas ICI118551 supplementation reduced it (Fig. S13A, B, Fig. S19 A1, A2). As reported, differences in receptor conformation stabilized by various ligands may expose different phosphorylation sites, which may elicit distinctive β-arrestin activation states and receptor endocytosis [38, 39]. We speculate that although β1-AR phosphorylation was decreased in the ICI118551 treatment group, the exposed phosphorylation sites were prone to coupling with β-arrestin. Of course, more work is needed to test this hypothesis.

Although the close relationship between β1-AA and HF is known, how to effectively block β1-AA remains difficult. β1-AR blockers are among the most commonly used medications in treating HF and the most convenient method because of their oral administration. However, β1-AR blockers cannot sufficiently block the continuous activation of β1-AR induced by β1-AA [40]. This problem was evident in this study, which may be owing to the differences in the binding sites of β1-AA and β1-AR blockers. Metoprolol recognizes the β1-AR ligand-binding pocket [23, 41, 42], whereas β1-AA binds to the second extracellular loop of β1-AR [16]. Animal studies have demonstrated that the combination of the β2-AR agonist fenoterol and β1-AR blocker metoprolol can improve the heart function of rats with HF. In addition, preferentially activating β2-AR/Gi can antagonize the apoptosis-promoting effect of β1-AR on cardiomyocytes through the PI3K-Akt pathway [43]. Here, we mainly found that activating β2-AR/Gi could exert a cardioprotective effect by promoting β1-AR endocytosis and terminating its signal in a timely manner. Therefore, biased activation of the β2-AR/Gi-signaling pathway may be a new strategy for the treatment of HF patients with β1-AA. However, the effectiveness of this program in clinical treatment needs to be further confirmed.