Adiponectin Deficiency, Diastolic Dysfunction, and Diastolic Heart Failure

Materials and Methods

Experimental animals

The Institutional Animal Care and Use Committee at Boston University School of Medicine approved all study procedures related to handling and surgery of the mice. Wild-type (WT) C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Adiponectin-deficient (APNKO) mice were generated as previously described (25).

Experimental model

Uninephrectomized mice received osmotic minipumps (Alzet, Durect Corp., Cupertino, CA) that delivered either saline (sham, n = 20) or d-aldosterone infusion (0.15 μg/h, n = 30) (Sigma-Aldrich Co., St. Louis, MO) for 4 wk (23,26). All mice were given 1% NaCl drinking water for 4 wk. The four groups studied were 1) WT-saline, n = 10; 2) WT-aldosterone, n = 15; 3) APNKO-saline, n = 10; and 4) APNKO-aldosterone, n = 15.

Physiological measurements after 4 wk

Heart rate and tail cuff blood pressure were measured using a noninvasive tail-cuff system (BP-2000; VisiTech, Apex, NC) as previously described (26,27).

Echocardiography determinants for LV dimensions

As previously described (27,28), transthoracic echocardiography was performed in conscious mice using an Acuson Sequoia C-256 (Siemens, Malvern, PA) echocardiograph machine and a 15-MHz probe. Total wall thickness was derived from an average of the interventricular septum and posterior wall thickness.

Echocardiography determinants for LV diastolic function

Mitral Doppler flow studies and pulse wave tissue Doppler imaging (TDI) was performed using the Vevo 770 high-resolution in vivo imaging system (VisualSonics, Toronto, Canada). Images were acquired using a high-resolution (30 MHz) transducer. Mice were anesthetized with an isoflurane face mask initially at 0.5% and then increased to 1.5% until a heart rate of around 350 beats/min was achieved because measures of diastolic function are sensitive to heart rate and loading conditions (29). Isoflurane (1.5%) has minimal effects on diastolic function (30). The images for each mouse were recorded for at least 5 sec (30–40 cardiac cycles) from which three to five representative cycles with the highest quality imaging were selected. Doppler measurements were made at the tips of the mitral leaflets for diastolic filling profiles in the apical four-chamber view. Peak early (E) and late (A) mitral inflow velocities, deceleration time of early filling (DT), and isovolumetric relaxation time (IVRT) were measured as previously described (31). TDI was used to determine the myocardial velocity of the septal side of the mitral annulus to derive e′. A more precise measure of diastolic dysfunction is the TDI including the transmitral flow velocity to annular velocity ratio, the E/e′ ratio.

In a subset of mice (n = 5), hemodynamic measurements were performed 4 wk after surgery using a 1.4F catheter tip micromanometer (ARIA; Millar Instruments, Houston, TX) as previously described (32). Chronic aldosterone resulted in an increased LV end-diastolic pressure (EDP). Increased LVEDP correlated with the increased E/A ratio (r = 0.95) and with the increased E/e′ ratio (r = 0.94; P < 0.05 for both).

Organ weight, biomarker, and tissue analysis

After 4 wk, mice were killed, at which time blood was obtained to determine serum creatinine (Diazyme, Poway, CA), plasma adiponectin levels (Otsuka Pharmaceutical Co. Ltd., Tokyo, Japan) and aldosterone levels (Diagnostic Products Corp., Los Angeles, CA).

Body weight (BW), heart weight, and LV weight were measured. Hearts were either 1) arrested in diastole by KCl (30 mmol/liter), weighed, perfused with 10% buffered formalin and sliced horizontally for histology or 2) snap-frozen in liquid nitrogen. To measure fibrosis, trichrome-stained sections (5 μm) were visualized by light microscopy, and the entire section was quantified using Bioquant Image analysis software (Memphis, TN). To determine cardiomyocyte cross-sectional area (area = πr^{), sections from the 30 hearts were studied and quantified using the Bioquant Image analysis software (}28,33). The wet-to-dry lung ratio an indicator of pulmonary congestion (26,34) and HF was determined.

Determinants of renal function

Spontaneously voided urine was collected from mice between 0900 and 1200 h. Urinary albumin was assayed with a murine albumin ELISA kit (Bethyl Laboratories, Montgomery, TX). To standardize urinary albumin excretion for glomerular filtration rate (GFR), albuminuria was expressed as milligrams of urinary albumin per gram of urinary creatinine using the creatinine liquid reagents assay kit (Diazyme). Kidney samples were imbedded in OCT compound (Miles Laboratories, Elkhart, IN) and snap-frozen in liquid nitrogen. Tissue slices (5 μm) were stained with hematoxylin and eosin. More than 20 consecutive sections in each mouse were examined, and the mean area and number of cells in the glomeruli were determined using computer-assisted pixel counting (Photoshop 7.0; Adobe) as described previously (35).

Western blot analysis

Aliquots of protein (30 μg) were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Amersham Pharmacia Biotech, Piscataway, NJ) as previously described (27). Membranes were probed with antibodies to matrix metalloproteinase (MMP)-9, MMP-2, tissue inhibitor of MMP (TIMP)-1, and TIMP-2, (Chemicon, Temecula, CA). Blots were also incubated with atrial natriuretic peptide (ANP; Bachem, Torrance, CA); interferon (IFN)-γ (Biosource, Camarillo, CA); TNF-α, IL-6, and TGF-β (R&amp;D Systems, Minneapolis, MN); and IL-1β and collagen I and III (Rockland, Gilbertsville, PA) for 18–20 h at 4 C. Blots were then incubated in horseradish peroxidase-conjugated secondary antibody, and the signal was detected by SuperSignal West Chemiluminescence (Pierce, Rockford, IL). Blots were normalized to GAPDH (Santa Cruz Biotechnology, Santa Cruz, CA) antibodies or with Coomassie Brillant Blue staining of the gels (Sigma-Aldrich). Chemiluminescence was quantified by densitometry (Molecular Analyst; Bio-Rad, Hercules, CA).

To assess the oligomeric state of plasma adiponectin, samples obtained at the time of killing (1.0 μl) were resolved by SDS-PAGE under non-denaturing conditions. After transfer to membranes, immunoblot analysis was performed with the antibodies to mouse adiponectin (R&amp;D Systems). The distribution of each oligomeric isoform of adiponectin (high molecular weight, hexamer, and trimer) was quantified by Image J software.

Statistical analysis

All results are presented as mean ± sem. Differences between WT and APNKO mice were tested for statistical significance by the two-tailed Student’s t test. When necessary, one- or two-way ANOVA (followed by Student-Newman-Keuls post hoc tests when appropriate) was applied. P values < 0.05 were considered statistically significant.

Results

Biomarkers

Plasma concentrations of adiponectin were not detectable in APNKO mice (Fig. 1A​1A)) but were significantly decreased in WT-aldosterone compared with WT-saline mice (9.2 ± 0.7 vs. 11.9 ± 0.6 μg/ml; P < 0.05). The distribution of the different oligomeric forms of adiponectin was no different between saline- and aldosterone-infusion groups (Fig. 1​1,, B and C). Aldosterone levels were significantly elevated, but there was no difference between APNKO-aldosterone (P < 0.001 vs. APNKO-saline) and WT-aldosterone mice (P < 0.01 vs. WT-saline; Fig. 1D​1D).

Figure 1

Adiponectin levels in aldosterone- or saline-infused WT mice. A, Plasma adiponectin (APN) levels are decreased in WT-aldosterone mice at the end of 4 wk; B and C, distribution of the three oligomeric forms of adiponectin in plasma as determined by Western blot analysis, with 1 μl plasma was loaded in each lane; C, representative Western blot; D, at 4 wk, aldosterone levels were no different between WT-aldosterone and APNKO-aldosterone mice; n = 5–10 per group. ND, Not detected.

Cardiac structure and systolic function

Four weeks of aldosterone infusion in WT mice increased LV end-diastolic diameter (LVEDD) and end-systolic diameter (LVESD) (P < 0.05 vs. WT hearts). These parameters were not significantly different in APNKO-aldosterone mice (Fig. 2​2,, A and B). LV ejection fraction was unchanged, and there was no difference between WT-aldosterone and APNKO-aldosterone hearts (63 ± 1 vs. 62 ± 1%; Fig. 2C​2C).). Consistent with LV hypertrophy (LVH) and cardiomyocyte hypertrophy, total wall thickness was increased in APNKO-aldosterone mice (1.03 ± 0.06 mm) vs. WT-aldosterone mice (0.85 ± 0.04 mm, P < 0.05; Fig. 2D​2D).). Aldosterone infusion increased ANP protein expression in WT-aldosterone hearts by 55 ± 1%, which was increased further in APNKO-aldosterone hearts by approximately 2.2-fold (Fig. 2​2,, E and F).

Figure 2

In vivo transthoracic echocardiography measurements and ANP expression in WT and APNKO mice 4 wk after surgery. A, LVEDD; B, LVESD; C, ejection fraction (EF); D, total wall thickness (TWT). *, P < 0.05 vs. WT-saline; †, P < 0.01 vs. APNKO-saline; n = 5–10 per group. E, ANP protein expression was increased in WT-aldosterone hearts (‡, P < 0.01 vs. WT-saline) and in APNKO-aldosterone (§, P < 0.05 vs. APNKO-saline). F, Representative immunoblot of myocardial ANP expression; n = 5–10 per group. Data are mean ± sem. NS, Not significant.

Myocardial fibrosis and renal function

Total fibrosis (interstitial and perivascular fibrosis) was increased, but there was no difference between APNKO-aldosterone (16 ± 3%) and WT-aldosterone (20 ± 4%) hearts (Fig. 3​3,, A and B). TGF-β and type I and III collagen protein expression was measured from the LV but was not different between APNKO-aldosterone and WT-aldosterone hearts (Fig. 3​3,, C–E). A cardiac-renal connection has been implicated in development of diastolic dysfunction in experimental models (36), and hypoadiponectinemia has been demonstrated to accelerate glomerular and tubulointerstitial injury (35). Therefore, renal function was measured 4 wk after aldosterone infusion. Serum creatinine was increased but not different between both groups of aldosterone-infused mice (Table 1​1).). Similarly, histology of the residual kidney demonstrated an increase in the glomerular cross-sectional area and the number of intraglomerular cells in both groups of aldosterone-infused mice (Fig. 3​3,, F and G). There were no differences between the APNKO-aldosterone and WT-aldosterone mice. Albuminuria (another measure of renal damage) was increased with aldosterone infusion but was not exacerbated with aldosterone infusion in adiponectin deficiency.

Figure 3

Cardiac fibrosis and renal function in aldosterone-infused APNKO and WT mice. A, Qualitatively, there is more fibrosis in both WT- and APNKO-aldosterone hearts vs. WT- and APNKO-saline hearts. B, Percentage of myocardial fibrosis is increased in both WT- and APNKO-aldosterone hearts. **, P < 0.001; †, P < 0.01 vs. respective saline for both. There was no difference in fibrosis between WT and APNKO-aldosterone hearts. Data are mean ± sem and reflects 10 measurements from three sections each for WT and APNKO (n = 3) hearts. C–E, TGF-β (C), type I collagen (D), and type I collagen (E) expression was not different between aldosterone-infused WT and APNKO hearts. F, Four weeks after uninephrectomy and aldosterone/saline infusion, histology of the remaining kidney in WT and APNKO were analyzed. Note the lack of differences in the glomeruli and tubules between WT-aldosterone and APNKO-aldosterone mice. G and H, Mean values of glomerular cross-sectional area (G) and the number of intraglomerular cells (H) were quantitatively measured. ‡, P < 0.01 vs. WT-saline; †, P < 0.01 vs. APNKO saline. I, Urinary excretion of albumin (Alb)/creatinine (Cr) (milligrams per gram Cr) was increased in both groups of aldosterone-infused mice. There was no difference between WT-aldosterone and APNKO-aldosterone mice. ‡, P < 0.01 vs. WT-saline; †, P < 0.01 vs. APNKO saline; n = 5 per group. NS, Not significant.

Myocardial MMP, TIMP, and cytokine protein expression

Myocardial MMP and TIMP protein expression was also measured (Fig. 4​4,, A–D). MMP-2 expression was significantly increased in WT-aldosterone hearts by 98 ± 2% (26,37) but was increased only 40 ± 2% in APNKO-aldosterone hearts. TIMP-2 expression (the inhibitor of MMP-2) was not increased in any of the groups. MMP-9 expression was increased in both APNKO-aldosterone and WT-aldosterone hearts (89 ± 1 and 79 ± 1%, respectively) but was not different between both groups. Similarly, aldosterone infusion resulted in a nonsignificant increase in TIMP-1 expression (inhibitor of MMP-9) in APNKO and had no effect on WT hearts. Thus, adiponectin deficiency in aldosterone-induced diastolic HF is associated with increased myocardial MMP-2 and MMP-9 expression. In the presence of adiponectin, there is more myocardial MMP-2 expression. Because aldosterone induces a proinflammatory phenotype in the heart (28,38), myocardial TNF-α, IFN-γ, IL-1β, and IL-6 protein expression was measured. Aldosterone infusion in WT hearts increased TNF-α expression 47 ± 2% (P < 0.001 vs. WT-saline) and IFN-γ expression 52 ± 22% (P < 0.05 vs. WT-aldosterone; Fig. 4​4,, E and F). Lack of adiponectin in aldosterone-induced diastolic HF augmented both myocardial TNF-α (74 ± 6%) and IFN-γ expression (260 ± 66%; P < 0.05 for both TNF-α and IFN-γ) even further. IL-1β and IL-6 myocardial proteins were not expressed in aldosterone-infused APNKO or WT hearts (data not shown).

Figure 4

Myocardial MMP, TIMP, and cytokine protein expression in WT and APNKO aldosterone-infused mice. A, MMP-2 expression was increased in WT-aldosterone hearts (‡, P < 0.01) and also increased in APNKO-aldosterone (§, P < 0.05) vs. respective controls. B, TIMP-2 expression was nonsignificantly increased in WT-aldosterone hearts and not expressed in APNKO-aldosterone vs. respective controls. C, MMP-9 expression was increased in WT-aldosterone hearts (**, P < 0.001) and in APNKO-aldosterone hearts (#, P < 0.001) vs. respective controls. D, TIMP-1 expression was nonsignificantly increased in WT-aldosterone and APNKO-aldosterone hearts vs. respective controls. E, TNF-α protein expression was increased in WT-aldosterone (**, P < 0.001) and in APNKO-aldosterone hearts (†, P < 0.01) vs. respective controls. There were no differences between aldosterone-infused hearts. F, IFN-γ protein expression was increased in WT-aldosterone (*, P < 0.05) and APNKO-aldosterone (†, P < 0.01) vs. respective controls. Additionally, aldosterone significantly increased IFN-γ expression in APNKO hearts vs. WT-aldosterone); data are mean ± sem; n = 5–10 per group. NS, Not significant.

Blood pressure and cardiac morphology (Table 1​1)

There were no deaths in the mice during the 4-wk period. Aldosterone increased tail cuff blood pressure in WT mice (132 ± 2 vs. 109 ± 3 mm Hg; P < 0.01), which was further augmented in APNKO than in WT-aldosterone mice (140 ± 3 mm Hg; P < 0.05). The heart rate was decreased in the aldosterone-infused APNKO vs. WT- aldosterone mice (613 ± 28 vs. 654 ± 26 beats/min; P < 0.01). The BW was no different in the APNKO and WT groups regardless of saline or aldosterone infusion. The LV/BW ratio, an indicator of cardiac hypertrophy, was increased in aldosterone-infused WT mice (4.8 ± 0.2) and further augmented in aldosterone-infused APNKO mice (6.0 ± 0.4; P < 0.05). Aldosterone-induced cardiomyocyte cross-sectional area was greater in APNKO- aldosterone than WT-aldosterone hearts (P < 0.0001). The increased wet/dry lung ratio, a measure of pulmonary congestion, was most marked in aldosterone-infused APNKO mice.

Biomarkers

Plasma concentrations of adiponectin were not detectable in APNKO mice (Fig. 1A​1A)) but were significantly decreased in WT-aldosterone compared with WT-saline mice (9.2 ± 0.7 vs. 11.9 ± 0.6 μg/ml; P < 0.05). The distribution of the different oligomeric forms of adiponectin was no different between saline- and aldosterone-infusion groups (Fig. 1​1,, B and C). Aldosterone levels were significantly elevated, but there was no difference between APNKO-aldosterone (P < 0.001 vs. APNKO-saline) and WT-aldosterone mice (P < 0.01 vs. WT-saline; Fig. 1D​1D).

Figure 1

Adiponectin levels in aldosterone- or saline-infused WT mice. A, Plasma adiponectin (APN) levels are decreased in WT-aldosterone mice at the end of 4 wk; B and C, distribution of the three oligomeric forms of adiponectin in plasma as determined by Western blot analysis, with 1 μl plasma was loaded in each lane; C, representative Western blot; D, at 4 wk, aldosterone levels were no different between WT-aldosterone and APNKO-aldosterone mice; n = 5–10 per group. ND, Not detected.

Cardiac structure and systolic function

Four weeks of aldosterone infusion in WT mice increased LV end-diastolic diameter (LVEDD) and end-systolic diameter (LVESD) (P < 0.05 vs. WT hearts). These parameters were not significantly different in APNKO-aldosterone mice (Fig. 2​2,, A and B). LV ejection fraction was unchanged, and there was no difference between WT-aldosterone and APNKO-aldosterone hearts (63 ± 1 vs. 62 ± 1%; Fig. 2C​2C).). Consistent with LV hypertrophy (LVH) and cardiomyocyte hypertrophy, total wall thickness was increased in APNKO-aldosterone mice (1.03 ± 0.06 mm) vs. WT-aldosterone mice (0.85 ± 0.04 mm, P < 0.05; Fig. 2D​2D).). Aldosterone infusion increased ANP protein expression in WT-aldosterone hearts by 55 ± 1%, which was increased further in APNKO-aldosterone hearts by approximately 2.2-fold (Fig. 2​2,, E and F).

Figure 2

In vivo transthoracic echocardiography measurements and ANP expression in WT and APNKO mice 4 wk after surgery. A, LVEDD; B, LVESD; C, ejection fraction (EF); D, total wall thickness (TWT). *, P < 0.05 vs. WT-saline; †, P < 0.01 vs. APNKO-saline; n = 5–10 per group. E, ANP protein expression was increased in WT-aldosterone hearts (‡, P < 0.01 vs. WT-saline) and in APNKO-aldosterone (§, P < 0.05 vs. APNKO-saline). F, Representative immunoblot of myocardial ANP expression; n = 5–10 per group. Data are mean ± sem. NS, Not significant.

Doppler echocardiography

To control for loading conditions, mitral Doppler flow measurements were made at comparable heart rates (of ∼350 beats/min) in all mice. The physiological heart rate in a mouse results in the loss of the A-velocity (29). E-velocity was increased in both aldosterone-infused APNKO and WT mice (P < 0.001 vs. respective saline controls). Aldosterone decreased A-velocity (late filling) in only APNKO hearts (see supplement Fig. 1S, A–C, published as supplemental data on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org). The resultant E/A ratio (a determinant of ventricular diastolic function) was greater in APNKO-aldosterone than in WT-aldosterone mice, indicative of a restrictive mitral inflow velocity profile and impaired diastolic function (Table 2​2).). Consistent with the elevated E/A ratio, there was a tendency for DT to shorten but was nonsignificant between both aldosterone-infused groups. Similarly, IVRT (the time interval between aortic valve closure and mitral valve opening) was no different between both aldosterone-infused groups. Aldosterone infusion in both WT and APNKO was associated with a lower early diastolic mitral annular motion (e′) velocity, determined by TDI, indicating diastolic dysfunction (Table 2​2).). The E/e′ ratio, a measure of increased left atrial pressure, was increased in WT-aldosterone but more marked in APNKO-aldosterone hearts (P < 0.05). The increased E/A and E/e′ ratios demonstrated worse diastolic dysfunction with adiponectin deficiency in aldosterone-induced diastolic HF.

Myocardial fibrosis and renal function

Total fibrosis (interstitial and perivascular fibrosis) was increased, but there was no difference between APNKO-aldosterone (16 ± 3%) and WT-aldosterone (20 ± 4%) hearts (Fig. 3​3,, A and B). TGF-β and type I and III collagen protein expression was measured from the LV but was not different between APNKO-aldosterone and WT-aldosterone hearts (Fig. 3​3,, C–E). A cardiac-renal connection has been implicated in development of diastolic dysfunction in experimental models (36), and hypoadiponectinemia has been demonstrated to accelerate glomerular and tubulointerstitial injury (35). Therefore, renal function was measured 4 wk after aldosterone infusion. Serum creatinine was increased but not different between both groups of aldosterone-infused mice (Table 1​1).). Similarly, histology of the residual kidney demonstrated an increase in the glomerular cross-sectional area and the number of intraglomerular cells in both groups of aldosterone-infused mice (Fig. 3​3,, F and G). There were no differences between the APNKO-aldosterone and WT-aldosterone mice. Albuminuria (another measure of renal damage) was increased with aldosterone infusion but was not exacerbated with aldosterone infusion in adiponectin deficiency.

Figure 3

Cardiac fibrosis and renal function in aldosterone-infused APNKO and WT mice. A, Qualitatively, there is more fibrosis in both WT- and APNKO-aldosterone hearts vs. WT- and APNKO-saline hearts. B, Percentage of myocardial fibrosis is increased in both WT- and APNKO-aldosterone hearts. **, P < 0.001; †, P < 0.01 vs. respective saline for both. There was no difference in fibrosis between WT and APNKO-aldosterone hearts. Data are mean ± sem and reflects 10 measurements from three sections each for WT and APNKO (n = 3) hearts. C–E, TGF-β (C), type I collagen (D), and type I collagen (E) expression was not different between aldosterone-infused WT and APNKO hearts. F, Four weeks after uninephrectomy and aldosterone/saline infusion, histology of the remaining kidney in WT and APNKO were analyzed. Note the lack of differences in the glomeruli and tubules between WT-aldosterone and APNKO-aldosterone mice. G and H, Mean values of glomerular cross-sectional area (G) and the number of intraglomerular cells (H) were quantitatively measured. ‡, P < 0.01 vs. WT-saline; †, P < 0.01 vs. APNKO saline. I, Urinary excretion of albumin (Alb)/creatinine (Cr) (milligrams per gram Cr) was increased in both groups of aldosterone-infused mice. There was no difference between WT-aldosterone and APNKO-aldosterone mice. ‡, P < 0.01 vs. WT-saline; †, P < 0.01 vs. APNKO saline; n = 5 per group. NS, Not significant.

Myocardial MMP, TIMP, and cytokine protein expression

Myocardial MMP and TIMP protein expression was also measured (Fig. 4​4,, A–D). MMP-2 expression was significantly increased in WT-aldosterone hearts by 98 ± 2% (26,37) but was increased only 40 ± 2% in APNKO-aldosterone hearts. TIMP-2 expression (the inhibitor of MMP-2) was not increased in any of the groups. MMP-9 expression was increased in both APNKO-aldosterone and WT-aldosterone hearts (89 ± 1 and 79 ± 1%, respectively) but was not different between both groups. Similarly, aldosterone infusion resulted in a nonsignificant increase in TIMP-1 expression (inhibitor of MMP-9) in APNKO and had no effect on WT hearts. Thus, adiponectin deficiency in aldosterone-induced diastolic HF is associated with increased myocardial MMP-2 and MMP-9 expression. In the presence of adiponectin, there is more myocardial MMP-2 expression. Because aldosterone induces a proinflammatory phenotype in the heart (28,38), myocardial TNF-α, IFN-γ, IL-1β, and IL-6 protein expression was measured. Aldosterone infusion in WT hearts increased TNF-α expression 47 ± 2% (P < 0.001 vs. WT-saline) and IFN-γ expression 52 ± 22% (P < 0.05 vs. WT-aldosterone; Fig. 4​4,, E and F). Lack of adiponectin in aldosterone-induced diastolic HF augmented both myocardial TNF-α (74 ± 6%) and IFN-γ expression (260 ± 66%; P < 0.05 for both TNF-α and IFN-γ) even further. IL-1β and IL-6 myocardial proteins were not expressed in aldosterone-infused APNKO or WT hearts (data not shown).

Figure 4

Myocardial MMP, TIMP, and cytokine protein expression in WT and APNKO aldosterone-infused mice. A, MMP-2 expression was increased in WT-aldosterone hearts (‡, P < 0.01) and also increased in APNKO-aldosterone (§, P < 0.05) vs. respective controls. B, TIMP-2 expression was nonsignificantly increased in WT-aldosterone hearts and not expressed in APNKO-aldosterone vs. respective controls. C, MMP-9 expression was increased in WT-aldosterone hearts (**, P < 0.001) and in APNKO-aldosterone hearts (#, P < 0.001) vs. respective controls. D, TIMP-1 expression was nonsignificantly increased in WT-aldosterone and APNKO-aldosterone hearts vs. respective controls. E, TNF-α protein expression was increased in WT-aldosterone (**, P < 0.001) and in APNKO-aldosterone hearts (†, P < 0.01) vs. respective controls. There were no differences between aldosterone-infused hearts. F, IFN-γ protein expression was increased in WT-aldosterone (*, P < 0.05) and APNKO-aldosterone (†, P < 0.01) vs. respective controls. Additionally, aldosterone significantly increased IFN-γ expression in APNKO hearts vs. WT-aldosterone); data are mean ± sem; n = 5–10 per group. NS, Not significant.

Discussion

In the present study, lack of adiponectin in aldosterone-induced cardiac remodeling exacerbated systolic blood pressure, LVH, diastolic dysfunction, pulmonary congestion, and diastolic HF. LV chamber size (39) and ejection fraction were not adversely affected by adiponectin deficiency. Renal dysfunction and aldosterone levels were comparable between both groups of aldosterone-infused mice as were cardiac fibrosis and myocardial MMP-9, TIMP-1 and TIMP-2 protein expression. This suggests that with adiponectin deficiency, alterations in the extracellular matrix is likely not a contributing factor in the worsening diastolic dysfunction and diastolic HF. ANP (a molecular marker of hypertrophy), IFN-γ, and TNF-α expression were also augmented in the presence of adiponectin deficiency in aldosterone-induced cardiac remodeling and diastolic HF.

Lack of adiponectin exacerbates diastolic dysfunction and HF

Hypoadiponectinemia occurs in obesity, type 2 diabetes, and hypertension (13,40), and administration of adiponectin ameliorates hypertension (14). Aldosterone infusion results in hypertension-induced cardiac remodeling and diastolic HF (26). There is not one single measurement that defines diastolic dysfunction; thus, several parameters were measured as shown in Table 2​2.. In the present study, hypoadiponectinemia worsened diastolic dysfunction in diastolic HF as demonstrated by the increased E-velocity and decreased A-velocity. This resulted in a further increase in the E/A ratio in APNKO-aldosterone mice and is consistent with a restrictive pattern of severe diastolic dysfunction. Similarly, e′ was decreased, also consistent with diastolic dysfunction. Importantly, Doppler measurements were performed at lower but comparable heart rates to prevent fusion of the E- and A-velocities (30). There are no effects of isoflurane anesthesia on LV relaxation and LV compliance (30,41). Determinants of chamber size were made in conscious mice because anesthesia may alter chamber dimensions and percent fractional shortening (42). DT and IVRT tended to decrease, presumably because this was not early/mild diastolic dysfunction but rather severe diastolic dysfunction. In early/mild diastolic dysfunction, DT may shorten to compensate for the relative differences between atrial and ventricular pressures, which offsets the slowing of LV relaxation (31). The IVRT represents the earliest phase of diastole and shortens when LV compliance is decreased and LV filling pressures are increased (as demonstrated in this study by the increased E/A and E/e′ ratios). These measures of diastolic dysfunction complement the findings of diastolic HF, i.e. preservation of LV systolic function and the presence of pulmonary venous congestion. Therefore, lack of adiponectin exacerbates diastolic dysfunction in aldosterone-induced cardiac remodeling and diastolic HF. We cannot, however, exclude direct effects of aldosterone infusion on the pulmonary vasculature in adiponectin-deficient mice. It is possible that the pulmonary congestion seen may not have been due to the impaired cardiac filling that is seen in diastolic dysfunction. It is plausible that aldosterone exerts direct effects on the pulmonary vasculature in hypoadiponectinemia. Similarly, experimental studies have suggested that renal dysfunction may mediate the cardiac effects seen in diastolic dysfunction (36). Because renal dysfunction was comparable between both groups of aldosterone-infused mice, the worsened renal function is not the cause of the exacerbation of diastolic HF and diastolic dysfunction seen in hypoadiponectinemia.

Adiponectin and structural remodeling

LV chamber dimensions were measured in conscious mice at physiological heart rates. LVESD and LVEDD were slightly increased in hypertension-induced diastolic HF (23,26). LV chamber size may increase in diastolic HF, and the chronic volume overload may contribute to the pathophysiology of the diastolic HF (36,43,44). However, LVED volume is usually normal/near-normal in clinical diastolic HF (39,45) and in this study was comparable in both groups of aldosterone-infused mice. It has been suggested that there are different pathophysiological findings in normal-LV-function HF with different mechanisms leading to the single entity of diastolic dysfunction (11,46); e.g. subjects with hypertrophic cardiomyopathy have diastolic dysfunction and diastolic HF that progress to LV chamber dilation and systolic HF. Various animals models of diastolic dysfunction and diastolic HF progress to systolic HF with initial changes in LV dimensions before LV function deteriorates (36,47,48). It, however, remains unclear in humans whether diastolic dysfunction always progresses to systolic dysfunction in HF (46). Our study demonstrates that adiponectin contributes to the preservation of chamber size in diastolic dysfunction and diastolic HF.

MMP and diastolic HF

Cardiac stress induces myocardial fibrosis and MMP-2 activity (26,49), and MMP-2 inhibition ameliorates adverse cardiac remodeling and profibrotic changes in the extracellular matrix (ECM) (49). In humans, diastolic dysfunction is associated with increased MMP-2 and MMP-9 plasma levels indicative of the ECM turnover (50). The amount, composition, and geometry of the ECM impairs cardiomyocyte relaxation and increases LV stiffness (51), thus leading to diastolic dysfunction. Our study showed comparable degrees of interstitial fibrosis in both groups of aldosterone-infused hearts, regardless of the presence or absence of adiponectin. This was a surprising finding because aldosterone increases MMP activation in vitro and in vivo (26,52), and myocardial interstitial fibrosis was increased in adiponectin-deficient mice subjected to myocardial infarction (32). We speculate that these disparate findings are due to different stressors/stimuli (e.g. myocardial infarction vs. aldosterone infusion) and the type of remodeling (eccentric vs. concentric). Additionally, other endogenous modifiers that were not studied, such as cathepsins, also play a role in effective matrix destruction (53). MMP-2 and MMP-9 may also mediate LVH (49,54). The presence of less MMP-2 expression despite augmented LVH in adiponectin deficiency in the presence of aldosterone suggests that MMP-2 does not play a role in the LVH that is seen with hypoadiponectinemia. Other causal mechanisms must be responsible for the interstitial fibrosis and the LVH seen with adiponectin deficiency in aldosterone-induced diastolic HF.

Cytokine expression and adiponectin deficiency

Cytokine expression is increased in hypertension and cardiac remodeling (55). Aldosterone-induced cardiac remodeling induces myocardial TNF-α and IFN-γ expression with even further increases in TNF-α and IFN-γ expression in adiponectin deficiency. Adiponectin mediates protective effects after pathological stimuli by modulating the proinflammatory responses of TNF-α (56,57). Adiponectin modulates the effects of TNF-α by several mechanisms; adiponectin induces IL-10, which is required for adiponectin-mediated reduction of lipopolysaccharide-induced TNF-α (58). In cardiomyocytes, adiponectin modulates the cyclooxygenase-2/prostaglandin E2-dependent pathway (59). Adiponectin also removes early apoptotic bodies and is associated with TNF-α down-regulation (60). Therefore, decreased adiponectin is associated with increased TNF-α. Additionally, it is unknown whether the effects of decreased adiponectin may be due to divergent effects of TNF-α, which are mediated by two cell surface receptors [TNF receptor 1 (TNFR1) and TNFR2] in cardiomyocytes. These receptors have opposing actions with TNFR1 being proinflammatory and TNFR2 activation ameliorating nuclear factor-κB activation and being antiapoptotic in cardiomyocytes (61).

The role of IFN-γ in cardiac remodeling is unknown. IFN-γ deficiency promotes myocardial inflammation, resulting in severe systolic and diastolic dysfunction (62). In contrast, overexpressing hepatic IFN-γ results in LV dilation and impairs systolic function (63). Similarly, the relationship between IFN-γ and adiponectin is also unknown. The finding that IFN-γ expression is markedly increased, when adiponectin is lacking in aldosterone- induced diastolic HF, suggests that IFN-γ may play a role in mediating myocardial adiponectin-induced changes in diastolic dysfunction and LVH. However, the present study does not establish a cause or an effect and only implicates IFN-γ in the pathogenesis of adiponectin-mediated diastolic dysfunction and diastolic HF. Likewise, cardiac IFN-γ expression may simply reflect comparatively worse LVH, hypertension, and adverse cardiac remodeling. The mechanisms by which adiponectin interacts with IFN-γ in cardiomyocytes are unknown and warrants further study.

In aldosterone-mediated cardiac remodeling and diastolic HF, plasma adiponectin levels were decreased. Our findings suggest that aldosterone negatively regulates the adiponectin produced by adipose tissue, thereby contributing to the reduction of adiponectin sensitivity in the heart. There is evidence of cross talk between adipocytokines and aldosterone secretion in obese hypertensive subjects (64). Our study suggests that adiponectin and aldosterone regulatory pathways occur together in the heart. In addition to secreting adiponectin, adipocytes also express RAAS receptors, such as angiotensin-II type 1 receptors (65) and mineralocorticoid receptors (66), suggesting that the RAAS plays a pathophysiological role in adipose tissue and adipocytokine signaling (64). Recent reports have suggested that angiotensin-II partially mediates its effect through aldosterone (67) and that angiotensin-II infusions lower adiponectin levels (65).

Limitations

Determinants of diastolic function are usually made with invasive techniques such as cardiac catheterization. Millar catheterization was performed in a subset of mice, and LVEDP correlated with the E/A ratio. Several measures of diastolic function were made with the Vevo 770 high-resolution system by Doppler echocardiography (mitral flow studies and pulse-wave TDI). These have been validated by other investigators (29,30,41) as reliable measures of diastolic dysfunction.

Our study demonstrates that adiponectin deficiency contributes to a worsening of diastolic dysfunction and diastolic HF in hypertension-induced diastolic HF. An adverse outcome of resistant hypertension is diastolic dysfunction and diastolic HF, which is exacerbated by hypoadiponectinemia (such as in obesity, insulin resistance, and coronary artery disease). Lowering blood pressure to comparable levels in the aldosterone-infused mice and measuring diastolic function or using gain-of-function studies (e.g. adiponectin transgenic mice) would test the hypothesis that adiponectin modulates the transition from LVH and hypertension to diastolic HF and diastolic dysfunction and requires further study.

In conclusion, the impaired relationship between adiponectin and the RAAS, specifically aldosterone, exacerbates diastolic HF in hypertension and obesity. Targeting underlying mechanism(s) remains complex, largely because of unexplored pathways. However this interaction between adiponectin and aldosterone in hypertension and obesity may be useful for future therapeutic interventions for severe and/or resistant hypertension and to prevent the progression to diastolic dysfunction and diastolic HF. Our data suggests that adiponectin effects are not purely salutary but represents an important pathophysiological role in hypertension-induced cardiac remodeling and diastolic HF that requires further exploration. The relationship between the adipose tissue and the heart is further demonstrated in this study and adds to the growing relationship between obesity and hypertension.

Supplementary Material