BDNF modulates heart contraction force and long-term homeostasis through truncated TrkB.T1 receptor activation

Abstract

Brain-derived neurotrophic factor (BDNF) is critical for mammalian development and plasticity of neuronal circuitries affecting memory, mood, anxiety, pain sensitivity, and energy homeostasis. Here we report a novel unexpected role of BDNF in regulating the cardiac contraction force independent of the nervous system innervation. This function is mediated by the truncated TrkB.T1 receptor expressed in cardiomyocytes. Loss of TrkB.T1 in these cells impairs calcium signaling and causes cardiomyopathy. TrkB.T1 is activated by BDNF produced by cardiomyocytes, suggesting an autocrine/paracrine loop. These findings unveil a novel signaling mechanism in the heart that is activated by BDNF and provide evidence for a global role of this neurotrophin in the homeostasis of the organism by signaling through different TrkB receptor isoforms.

Figure 1.

TrkB.T1 receptor isoform is expressed in adult mouse heart and cardiomyocytes. (A) Western blot analysis of brain, heart, and cardiomyocyte lysates from adult WT and TrkB.T1-deficient (T1^−/−) mice. Lysates were incubated with wheat germ lectin agarose to enrich for glycoproteins. The wheat germ agglutinated (WGA) precipitates were analyzed by Western blot analysis with an antibody directed against the extracellular domain of TrkB to detect all TrkB isoforms. Note that in whole heart and cardiomyocytes (Cardio) only a truncated TrkB isoform (80–90 Kd) is detected. The absence of the corresponding band in the TrkB.T1 knockout animals verifies the identity of the receptor. Brain lysates were used as a positive control. Right panel, input lysates. (B) Quantification of real-time PCR analysis as expressed by the number of PCR cycles at which full-length TrkB (TrkB.Kin) or TrkB.T1-specific PCR products are equal to GAPDH level (Δ Ct) from total cardiomyocyte RNA. Note that TrkB.T1 is expressed at a much higher level as PCR products appear after only 5 cycles of GADPH detection versus 11 cycles for TrkB.Kin. (C) Ethidium bromide agarose gel visualizing the size of the DNA fragments from heart RT-PCR analysis. Note that the size of the PCR reaction products corresponding to TrkB kinase (Kin) and TrkB.T1 (T1) are the same as those from brain used as a positive control. B, brain; H, heart.

Figure 2.

TrkB.T1 mediates BDNF-induced acute increase in cardiac contraction force and calcium transient increase evoked by direct stimulation in cardiomyocytes. (A and B) BDNF (1 ng in 50 µl Krebs solution) injected in the fluid streamline of a Langendorff-perfused mouse heart induces an increase in systolic pressure and a consequent decrease of the diastolic pressure (A) that is not affected by the TrkB kinase inhibitor K252a (B), Arrows indicate the time of BDNF injection and the broken line (B) indicates K252a presence throughout the 80-s duration of the experiment. (C) Representative traces showing the changes in LVPD before (baseline) and after BDNF injection. (D) Representative traces showing the changes in LVPD before (baseline) and after BDNF injection in the presence of 200 µM K252A. (E and F) Representative traces showing the increase in LVDP caused by BDNF in WT (E) but not in TrkB.T1 knockout (T1^−/−; F) hearts. A bolus of 50 µl of 5 mM caffeine in Krebs solution was injected 5 min after BDNF as a positive control. The BDNF and caffeine traces were overlapped using the injection time as the starting point (arrow). Note the lack of LVDP change in response to BDNF in the TrkB.T1 mutant mouse despite normal response to caffeine. (G and H) Quantification of data in E and F showing the percent change of baseline LVPD in response to BDNF (G) and caffeine (H) in WT and T1^−/− hearts. (I) Representative traces of Ca²⁺ transients elicited by 2-Hz stimulation in isolated adult cardiomyocytes. Ca²⁺ transient are increased by BDNF application only in WT (BDNF, light gray; no BDNF, black) but not in T1^−/− cardiomyocytes. (J) Effect of BDNF on transient amplitude is reversible as shown in typical time course from a WT and a T1^−/− cardiomyocyte. BDNF effect on Ca²⁺ release has a rapid onset and is completely reversed in <2 min upon BDNF removal. (K) Quantification of peak amplitude of calcium transient in WT and T1^−/− cardiomyocytes. The value for each group represents the transient change in percentage before (considered as 100%) and after BDNF application. 10 transients before and 10 transients after BDNF application were measured for each cardiomyocyte analyzed (n is shown within the bar). Values in G–K are indicated as the mean ± SEM. *, P < 0.05.

Figure 3.

BDNF signals independently of the catecholaminergic pathway and does not increase cardiac CamKII or PLB phosphorylation at the peak of its inotropic function. (A) Isoproterenol does not occlude the effect of BDNF in Langendorff-perfused hearts. Representative LVDP trace obtained from application of 100 nM isoproterenol (ISO) to a Langendorff-perfused heart. Once the heart reached stable LVDP a 50 µl (10 ng total) bolus of BDNF was applied followed by a second 50 µl (5 mM) bolus of caffeine as a positive control 2 min later. (B) Quantification of the effect of BDNF and caffeine on baseline LVDP in the absence or presence of ISO. The baseline for the values obtained with ISO is considered LVDP in the presence of ISO. Values are from four independent experiments. (C–F) BDNF does not increase cardiac CamKII or PLB phosphorylation at the peak of its inotropic function. (C) Western blot analysis of PLB phosphorylation at Ser16 (PLB P-S16) and Thr17 (PLB P-T17), CAMKII phosphorylation at Thr 286/287 (CAMKII P-S286), and Troponin I at Ser23/24. Langendorff-perfused hearts were injected with vehicle (negative control), BDNF, or Forskolin (positive control) as described in Material and methods. (D–F) Quantification of the bands intensity (±SEM) from C reported as the ratio between the phosphorylated form over the specific total protein. Means of the three samples ±SEM are shown. *, P < 0.05, calculated with two-tail t test.

Figure 4.

TrkB.T1 deletion induces cardiac pathological alterations coupled with an increase in L-type Ca²⁺ channel level and currents. (A) WT and TrkB.T1^−/− hematoxylin and eosin–stained transversal heart sections showing left ventricle dilation and reduced thickness of left ventricle posterior wall apparent in TrkB.T1-deficient animals. (B) Quantification of the thickness (arrow) of the posterior wall (PW) and left ventricle area (LV); RV, right ventricle; n = 4 for each genotype. (C) Representative echocardiography recording in M-mode obtained in WT and T1^−/− mice showing the left ventricle internal dimension at diastole (double head red arrows, LVID) and the left ventricle posterior wall (LVPW) thickness delineated by blue lines. (D) Quantification of LVID and LVPW thickness measured at diastole (dia), and fractional shortenings (FS) values obtained from the echocardiography recordings (P < 0.05). WT, n = 4; T1^−/−, n = 5. (E–J) TrkB.T1 deletion increases L-type calcium levels and currents in adult isolated cardiomyocytes. (E) I/V curves obtained in whole cell patch-clamp isolated adult cardiomyocytes showing similar activation of L-type calcium channels but higher current density in T1^−/− versus WT cardiomyocytes. (F) Channel kinetics and current decay recordings are comparable in T1^−/− and WT cardiomyocytes. (G) Measurements of plasma membrane capacitance are similar between WT and T1^−/− cardiomyocytes, suggesting no difference in cell size. (H) Graph showing similar calcium-induced L-type current inactivation calculated by paired pulse protocol in WT and T1^−/− cardiomyocytes. (I) Representative Western blot analysis of cardiac lysates from WT and T1^−/− hearts using an antibody specific for the α 1c subunit of the Cav1.2 channel. (J) Quantification of Western blot CaV1.2 a1c band intensity in relation to GAPDH; n = 6. Data are indicated as the mean ± SEM. *, P < 0.05.

Figure 5.

Specific deletion of TrkB.T1 in cardiomyocytes abolishes BDNF inotropic effect and leads to cardiomyopathy. (A–C) Representative traces of LVDP before (baseline considered as 100%) and after BDNF or caffeine injection (arrows) in control myosin 6–specific cre transgenic (myh6-cre) mice used as controls (A), double Myh6-cre, TrkB.T1 conditional (Myh6-cre/TrkB.T1^loxP; B), and endothelial-specific cre (Cdh5-cre) TrkB.T1 conditional (Cdh5-cre/TrkB.T1^loxP; C) mutant mice. Caffeine was injected 5 min after BDNF as a positive control as in Fig. 2. The BDNF and caffeine traces were overlapped using the injection time (arrow) as a starting point. (D–F) Quantification of LVDP recorded from the mice in A (D), B (E), and C (F). (G–I) Cardiomyocyte-specific deletion of TrkB.T1 or BDNF causes dilated cardiomyopathy. Representative hematoxylin and eosin–stained sections from 2–3-mo-old control Myh6-cre transgenic (G; myh6; n = 4), Myh6-cre/BDNF^loxP/loxP (H; Myh6-BDNF; n = 5), or Myh6-cre/TrkB.T1^loxP/loxP (I; Myh6-TrkB.T1; n = 5) mouse hearts showing dilated left ventricle and reduced left ventricle posterior wall thickness caused by BDNF or TrkB.T1 deletion. (J and K) Quantification of left ventricle (LV) area and posterior wall (PW) thickness was measured in heart transverse sections at the level and in between the papillary muscles (arrow). RV, right ventricle. Data are indicated as the mean ± SEM. *, P < 0.05.

Figure 6.

BDNF is expressed in the adult heart and is secreted by cardiomyocytes. (A) Heart (H) and brain (B) lysates obtained from a knock-in BDNF HA tagged and a WT mouse used as a control were immunoprecipitated and blotted with anti-HA antibodies. 15-kD molecular mass marker. Lysate inputs are on the right of the panel. (B and C) Embryonic cardiomyocytes secrete biologically active BDNF. (B) Culture medium from WT and BDNF-HA cultured embryonic cardiomyocytes was immunoprecipitated as in A. WT and BDNF-HA brains were used as negative and positive controls, respectively. (C) Cells stably expressing TrkB tyrosine kinase were used as a sensitive biological assay to test for the presence of secreted biologically active BDNF in the medium of cultured cardiomyocytes. After a 10-min incubation with the medium from different cardiomyocyte cultures, cells expressing TrkB were lysed and subjected to Western analysis with an anti–phospho-TrkB antibody. Anti-TrkB and GAPDH antibodies were used as controls for loading. The tested supernatants were from WT, BDNF heterozygous (HET), and BDNF knockout cultured cardiomyocytes. Unconditioned media supplemented with 1 ng/ml of recombinant BDNF (BDNF)F and 1 µg/ml TrkB-Fc BDNF blocking antibody (BDNF+α) were used as positive and negative controls, respectively. Ctrl, unconditioned media (DMEM). The molecular mass is shown on the right.