Col1a2-Deleted Mice Have Defective Type I Collagen and Secondary Reactive Cardiac Fibrosis with Altered Hypertrophic Dynamics

Abstract

Rationale: The adult cardiac extracellular matrix (ECM) is largely comprised of type I collagen. In addition to serving as the primary structural support component of the cardiac ECM, type I collagen also provides an organizational platform for other ECM proteins, matricellular proteins, and signaling components that impact cellular stress sensing in vivo.

Objective: Here we investigated how the content and integrity of type I collagen affect cardiac structure function and response to injury.

Methods and results: We generated and characterized Col1a2^-/- mice using standard gene targeting. Col1a2^-/- mice were viable, although by young adulthood their hearts showed alterations in ECM mechanical properties, as well as an unanticipated activation of cardiac fibroblasts and induction of a progressive fibrotic response. This included augmented TGFβ activity, increases in fibroblast number, and progressive cardiac hypertrophy, with reduced functional performance by 9 months of age. Col1a2-loxP-targeted mice were also generated and crossed with the tamoxifen-inducible Postn-MerCreMer mice to delete the Col1a2 gene in myofibroblasts with pressure overload injury. Interestingly, while germline Col1a2^-/- mice showed gradual pathologic hypertrophy and fibrosis with aging, the acute deletion of Col1a2 from activated adult myofibroblasts showed a loss of total collagen deposition with acute cardiac injury and an acute reduction in pressure overload-induce cardiac hypertrophy. However, this reduction in hypertrophy due to myofibroblast-specific Col1a2 deletion was lost after 2 and 6 weeks of pressure overload, as fibrotic deposition accumulated.

Conclusions: Defective type I collagen in the heart alters the structural integrity of the ECM and leads to cardiomyopathy in adulthood, with fibroblast expansion, activation, and alternate fibrotic ECM deposition. However, acute inhibition of type I collagen production can have an anti-fibrotic and anti-hypertrophic effect.

Keywords: cardiac hypertrophy; cardiomyocyte; extracellular matrix; fibroblast; type I collagen.

Figure 1

Germline deletion of Col1a2 leads to cardiomyopathy in mice. (A) Schematic representation of the targeted Col1a2 allele that was used for both constitutive and tissue-specific deletion of this gene in mice. Exons are numbered and shown as grey boxes, as are the positions of the FRT and LoxP sites. (B) Relative Col1a2 mRNA expression in both wild-type (WT, n = 3) and null (Col1a2^−/−, n = 4) mouse hearts at 3 months of age. (C,D) Representative immunohistochemistry from heart sections of type I collagen (red) staining intensity at 3 and 6 months of age in WT (n = 5) and Col1a2^−/− (n = 4) mice with DAPI (blue) costaining to show nuclei. Scale bar: 25 μm. (E) Ventricle weight to body weight (VW/BW) ratio in WT, Col1a2^+/− and Col1a2^−/− mice at 3 months of age (n = 13, 14, and 20, respectively) and (F) 9 months of age (n = 11, 7, and 12, respectively). (G) Echocardiographic measurement of cardiac fractional shortening (FS%) in WT (n = 7), Col1a2^+/− (n = 9), and Col1a2^−/− (n = 9) mice at 9 months of age. (H) Echocardiographic measurement of left ventricular (LV) volume at diastole in WT (n = 5), Col1a2^+/− (n = 9), and Col1a2^−/− (n = 9) mice at 9 months of age. (I) Representative baseline cardiac pressure-volume loops obtained by invasive hemodynamic measurements in hearts of WT and Col1a2^−/− mice at 8–10 weeks of age. Data are presented as mean +/− standard deviation (SD). For (B,D), Student’s t-tests were used. For (E–H), one-way ANOVA analysis was used p < 0.05, and then Tukey’s post hoc analysis was used for between-group comparisons. Adjusted p-values are given in the figures.

Figure 2

Hearts from Col1a2^−/− mice show fibrobotic response. (A) Change in stretch-induced force generation of decellularized left ventricle ECM strips from WT and Col1a2^−/− mice at 8–10 weeks of age. Data are presented as mean +/− SD. Multiple unpaired t-tests were used, p-values are stated in the figure. (B) Volcano plot of whole-ventricle mRNA microarray analysis of Col1a2^−/− mouse hearts compared to Col1a2^+/− at 2 months of age, n = 3 per genotype. (C) Mass spectrometry analysis of representative ECM proteins that were changed in Col1a2^−/− mouse hearts compared to Col1a2^+/− hearts at 3 months of age, n = 4 per genotype. (D) Representative immunofluorescence images of periostin staining in heart histological sections from Col1a2^+/− and Col1a2^−/− mice at 3 months of age. Scale bar: 25 µm. (E) Flow cytometric gating strategy and (F) analysis of cardiac fibroblasts (MEFSK4⁺/CD31⁻/CD45⁻) from dissociated hearts of Col1a2^+/− (n = 6) and Col1a2^−/− (n = 4) mice at 3 months of age. Data are presented as mean +/− SD. Student’s t-tests were used and p-values are as stated. (G) Representative immunofluorescence images of platelet-derived growth factor receptor (PDGFR)-α (purple) staining from heart histological sections from Col1a2^+/− and Col1a2^−/− mice at 3 months of age, marking cardiac fibroblasts (purple), and wheat germ agglutinin (WGA) in green to show cellular outlines. Scale bar: 100 µm.

Figure 3

Hearts from Col1a2^−/− mice have increased TGFβ activity and decreased cardiomyocyte contractility. (A) Western blot analysis of TGFβ isoforms in ECM fractions of hearts of 6-month-old Col1a2^+/− and Col1a2^−/− mice. (B) TGFβ activity in 3-month-old WT (n = 6) and Col1a2^−/− (n = 6) hearts measured by ELISA. (C) Representative images of immunofluorescent staining of phospho-Smad2/3 in heart histological sections from Col1a2^+/− and Col1a2^−/− mice at 3 months age. Scale bar: 25 µm. The dashed white box in the upper row show the regionsmagnified in the lower panels. (D) Langendorff-dissociated myocytes from Col1a2^+/− and Col1a2^−/− mice were assessed for cell contractility at 2–3 months and (E) 8 months of age (n = 4/group, via IonOptix system, as discussed in Section 2.12). (F) Dissociated cardiomyocytes were fixed, mounted, and imaged to determine average cell length and (G) width (n = 3 Col1a2^+/− and 4 Col1a2^−/−). For (B,D–G), data are presented as mean +/− SD. Student t-test was used and p-values are stated in each panel.

Figure 4

Col1a2^−/− mice hypertrophy in response to pressure overload, despite a lack of Col1a2. (A) Schematic representation of the experimental design in Col1a2^−/− mice that underwent TAC surgery at 8 weeks of age and tissue harvest 1 week later. Relative mRNA expression of (B) Col1a2, (C) Col1a1, (D) Postn, (E) Col3a1, and (F) Col5a1 from hearts of Col1a2^+/− and Col1a2^−/− mice after 1 week of TAC. Animals per group: n = 7/8 for sham Col1a2^+/− and Col1a2^−/−, respectively; n = 6/5 for TAC Col1a2^+/− and Col1a2^−/−, respectively. (G) Measurements of ventricle weight to body weight ratio (VW/BW) in the indicated groups of mice (n = 4 in sham groups, n = 6 for TAC groups). For panels (B–G), all error bars are SD, and two-way ANOVA analysis p < 0.05, and Tukey’s post hoc analysis for between-groups comparison was performed; (B–F), ^# p < 0.001, compared to sham of same genotype; * p < 0.007, Col1a2^+/− vs. Col1a2^−/− TAC. For (G), ^# p < 0.03 compared to a sham of the same genotype.

Figure 5

Fibroblast-specific Col1a2 deletion in adult mice alters fibrosis and the hypertrophic response following injury. (A) Schematic of cell-specific Cre induction to ablate Col1a2 in activated cardiac fibroblasts expressing the Postn^MCM allele. (B) Experimental timeline with MI (myocardial infarction) and TAM (tamoxifen) chow feeding followed by harvest. (C) Transmission electron microscope images from infarcted myocardium 2 weeks after MI injury. Red arrows indicate collagen fibers in the forming scar region. Scale bar: 800 nm. (D) Timeline for TAC surgery, TAM administration/chow, and tissue harvest in Postn^MCM/+; Col1a2^fl/fl and control mice starting at 8 weeks of age. (E) Representative immunohistological cardiac images of total type I collagen staining 1 week after TAC surgery in the 2 groups of mice shown. Scale bar: 25 μm. (F) Representative histological images of picrosirius red staining for fibrosis in control and Postn^MCM/+; Col1a2^fl/fl hearts after 6 weeks of pressure overload. Scale bar: 100 μm. (G) Measurements of ventricle weight to body weight ratio following 1, 2, or 6 weeks of TAC in Col1a2^fl/fl and Postn^MCM/+; Col1a2^fl/fl mice. For panel G, the number of mice per group is labeled on the graph and all error bars are SD. One-way ANOVA analysis p < 0.001, with Tukey’s post hoc analysis for between-group comparisons. ^# p < 0.001, compared to sham Col1a2^+/−; * p < 0.001, Col1a2^+/− vs. Col1a2^−/− TAC.