Hypoxia Attenuates Pressure Overload-Induced Heart Failure

Abstract

Background: Alveolar hypoxia is protective in the context of cardiovascular and ischemic heart disease; however, the underlying mechanisms are incompletely understood. The present study sought to test the hypothesis that hypoxia is cardioprotective in left ventricular pressure overload (LVPO)-induced heart failure. We furthermore aimed to test that overlapping mechanisms promote cardiac recovery in heart failure patients following left ventricular assist device-mediated mechanical unloading and circulatory support.

Methods and results: We established a novel murine model of combined chronic alveolar hypoxia and LVPO following transverse aortic constriction (HxTAC). The HxTAC model is resistant to cardiac hypertrophy and the development of heart failure. The cardioprotective mechanisms identified in our HxTAC model include increased activation of HIF (hypoxia-inducible factor)-1α-mediated angiogenesis, attenuated induction of genes associated with pathological remodeling, and preserved metabolic gene expression as identified by RNA sequencing. Furthermore, LVPO decreased Tbx5 and increased Hsd11b1 mRNA expression under normoxic conditions, which was attenuated under hypoxic conditions and may induce additional hypoxia-mediated cardioprotective effects. Analysis of samples from patients with advanced heart failure that demonstrated left ventricular assist device-mediated myocardial recovery revealed a similar expression pattern for TBX5 and HSD11B1 as observed in HxTAC hearts.

Conclusions: Hypoxia attenuates LVPO-induced heart failure. Cardioprotective pathways identified in the HxTAC model might also contribute to cardiac recovery following left ventricular assist device support. These data highlight the potential of our novel HxTAC model to identify hypoxia-mediated cardioprotective mechanisms and therapeutic targets that attenuate LVPO-induced heart failure and mediate cardiac recovery following mechanical circulatory support.

Keywords: cardiac hypertrophy; cardiac remodeling; hypoxia; left ventricular assist device; pressure overload.

Figure 1. Hypoxia attenuates adverse left ventricular remodeling and heart failure following transverse aortic constriction.

A, Experimental setup with gradual induction of alveolar hypoxia. B, Representative B‐mode transthoracic echocardiography images at end‐systole in long‐axis projection 4 weeks post surgery. The dashed line indicates left ventricular end‐systolic area (LVESA), scale bars: 2 mm. C and D, Contractile function determined by ejection fraction 2 days (#) and 4 weeks (#, &) post surgery. E and F, LVESA 2 days (#) and 4 weeks (#, $, &) post surgery. G, Heart weights normalized to tibia length 4 weeks post surgery (#, $). H through J, mRNA expression of heart failure markers Acta1, normalized to Rps16 (#) Myh7/Myh6 (#), and Nppa normalized to Rps16 (#) 4 weeks post surgery presented as fold change vs NxSham. K through M, Representative hematoxylin and eosin (H&E), wheat germ agglutinin (WGA), and Elastica van Gieson (EvG) stains (scale bars: 50 μm each) and stereological quantification of mean cross‐sectional area of cardiomyocytes (#, $) and fibrotic tissue (#) 4 weeks post surgery. Data are reported as mean±SEM. Two‐way ANOVA was performed to analyze differences after TAC by FiO₂, followed by Holm‐Šídák post hoc analysis (# P<0.05 for TAC, $ P<0.05 for FiO₂, and & P<0.05 for the interaction between TAC and FiO₂). FiO₂ indicates fraction of inspired oxygen; HxSham, hypoxia/sham surgery; HxTAC, hypoxia/transverse aortic constriction; NxSham, normoxia/sham surgery; NxTAC, normoxia/transverse aortic constriction; and TAC, transverse aortic constriction.

Figure 2. Hypoxia induces HIF‐1α expression and angiogenesis.

A and B, Representative immunoblots in left ventricular homogenates and densitometric quantification of HIF‐1α protein expression normalized to vinculin 3 days (&), 2 weeks (#, $), and 4 weeks post surgery. C, Vegfa mRNA expression normalized to Rps16 3 days (#), 2 weeks (&), and 4 weeks (#, $) post surgery. D and E, Quantification of capillary density per area and representative images 4 weeks post surgery (#, $; scale bars: 30 μm). Arrows indicate organized capillaries. Data are reported as mean±SEM and as fold change relative to NxSham at the same time point post surgery. Two‐way ANOVA was performed to analyze differences after transverse aortic constriction (TAC) by fraction of inspired oxygen (FiO₂), followed by Holm‐Šídák post hoc analysis (# P<0.05 for TAC, $ P<0.05 for FiO₂, and & P<0.05 for the interaction between TAC and FiO₂). HIF‐1α indicates hypoxia‐inducible factor‐1α; HxSham, hypoxia/sham surgery; HxTAC, hypoxia/transverse aortic constriction; NxSham, normoxia/sham surgery; NxTAC, normoxia/transverse aortic constriction; and WGA, wheat germ agglutinin.

Figure 3. Gene expression 3 days post surgery as determined by RNA sequencing.

A, Principal component analysis to visualize global gene expression clusters by surgery and fraction of inspired oxygen (FiO₂; n=6). Crosses indicate the center of the corresponding clusters. B, Venn diagram illustrating the number of altered transcripts (cutoff: |log₂ fold change| >0.6 and P<0.05). C, Heatmap of RNA sequencing count data corresponding to the 100 genes with the greatest variance across samples. Data are clustered by row after applying the regularized log transformation function in DESeq2. HxSham indicates hypoxia/sham surgery; HxTAC, hypoxia/transverse aortic constriction; NxSham, normoxia/sham surgery; and NxTAC, normoxia/transverse aortic constriction.

Figure 4. Hypoxia attenuates the transverse aortic constriction‐induced transcriptional response of pathological remodeling 3 days post surgery.

Data are presented as mean values for the comparisons as indicated (n=6). Blue and red bars indicate down‐ and upregulated, respectively. Black bars are not regulated (cutoff: false discovery rate <0.1). FC indicates fold change; HxSham, hypoxia/sham surgery; HxTAC, hypoxia/transverse aortic constriction; NxSham, normoxia/sham surgery; and NxTAC, normoxia/transverse aortic constriction.

Figure 5. Hypoxia attenuates the transcriptional response of metabolic genes 3 days post transverse aortic constriction.

A, Expression of transcripts involved in fatty acid oxidation, tricarboxylic acid (TCA) cycle, and glycolysis 3 days post surgery presented as mean values for the comparisons as indicated (n=6). Blue and red bars indicate down‐ and upregulated, respectively. Black bars indicate not regulated (cutoff: false discovery rate <0.1). B and C, mRNA expression of Ppargc1a and Pdk4 (#, $) normalized to Gapdh ($) 3 days post surgery as determined by quantitative reverse transcription‐polymerase chain reaction (qPCR) analysis. D and E, Mitochondrial state III oxygen consumption with pyruvate (#, $, &) and palmitoyl‐carnitine each combined with malate as substrates 3 days post ‐surgery. F, Citrate synthase enzyme activity in left ventricular (LV) tissue 3 days, 2 weeks (#), and 4 weeks (#) post surgery. G, Hydroxyacyl‐coenzyme A dehydrogenase (HADH) enzyme activity in LV tissue 3 days, 2 weeks (#, $) and 4 weeks (#, $) post surgery. H, Mitochondrial DNA content in LV tissue 3 days, 2 weeks (#) and 4 weeks post surgery. Data are presented as mean values ± SEM. Two‐way ANOVA was performed to analyze differences after TAC by fraction of inspired oxygen (FiO₂), followed by Holm‐Šídák post hoc analysis (# P<0.05 for TAC, $ P<0.05 for FiO₂, and & P<0.05 for the interaction between TAC and FiO₂). HxSham indicates hypoxia/sham surgery; HxTAC, hypoxia/transverse aortic constriction; NxSham, normoxia/sham surgery; and NxTAC, normoxia/transverse aortic constriction.

Figure 6. Commonly regulated transcripts in the HxTAC model and patients with cardiac recovery following left ventricular assist device‐mediated circulatory support.

A, Methodological approach for the identification of transcripts that are either protective or adverse in both LVAD‐mediated mechanical circulatory support and the HxTAC model. B and C, mRNA expression of Tbx5 ($, &) and Hsd11b1 (&) normalized to Gapdh 3 days post surgery as determined by quantitative reverse transcription‐polymerase chain reaction (qPCR) analysis. Data are reported as mean±SEM and as fold change relative to NxSham at the same time point post surgery. Two‐way ANOVA was performed to analyze differences after transverse aortic constriction (TAC) by fraction of inspired oxygen (FiO₂), followed by Holm‐Šídák post hoc analysis ($ P<0.05 for FiO₂ and & P<0.05 for the interaction between TAC and FiO₂). D and E, mRNA expression of TBX5 and HSD11B1 in preNR relative to preR as determined by RNA sequencing. Data are reported as mean±SEM. HxSham indicates hypoxia/sham surgery; HxTAC, hypoxia/transverse aortic constriction; LVAD, left ventricular assist device; NxSham, normoxia/sham surgery; NxTAC, normoxia/transverse aortic constriction; preR, LVAD responders (ie, recovery); and preNR, LVAD non‐responders (ie, no recovery post LVAD implantation).

Figure 7. Summary of potential cardioprotective mechanisms in HxTAC hearts.

Changes in HxTAC relative to NxTAC hearts are summarized. HIF‐1α indicates hypoxia‐inducible factor‐1α; HxTAC, hypoxia/transverse aortic constriction; and NxTAC, normoxia/transverse aortic constriction.