Loss of talin in cardiac fibroblasts results in augmented ventricular cardiomyocyte hypertrophy in response to pressure overload

Abstract

Pressure overload of the heart is characterized by concentric hypertrophy and interstitial fibrosis. Cardiac fibroblasts (CFs) in the ventricular wall become activated during injury and synthesize and compact the extracellular matrix, which causes interstitial fibrosis and stiffening of the ventricular heart walls. Talin1 (Tln1) and Talin2 (Tln2) are mechanosensitive proteins that participate in focal adhesion transmission of signals from the extracellular environment to the actin cytoskeleton of CFs. The aim of the present study was to determine whether the removal of Tln1 and Tln2 from CFs would reduce interstitial fibrosis and cardiac hypertrophy. Twelve-week-old male and female Tln2-null (Tln2^−/−) and Tln2-null, CF-specific Tln1 knockout (Tln2^−/−;Tln1^CF−/−) mice were given angiotensin-II (ANG II) (1.5 mg/kg/day) or saline through osmotic pumps for 8 wk. Cardiomyocyte area and measures of heart thickness were increased in the male ANG II-infused Tln2^−/−;Tln1^CF−/− mice, whereas there was no increase in interstitial fibrosis. Systolic blood pressure was increased in the female Tln2^−/−;Tln1^CF−/− mice after ANG II infusion compared with the Tln2^−/− mice. However, there was no increase in cardiac hypertrophy in the Tln2^−/−;Tln1^CF−/− mice, which was seen in the Tln2^−/− mice. Collectively, these data indicate that in male mice, the absence of Tln1 and Tln2 in CFs leads to cardiomyocyte hypertrophy in response to ANG II, whereas it results in a hypertrophy-resistant phenotype in female mice. These findings have important implications for the role of mechanosensitive proteins in CFs and their impact on cardiomyocyte function in the pathogenesis of hypertension and cardiac hypertrophy.

NEW & NOTEWORTHY The role of talins has been previously studied in cardiomyocytes; however, these mechanotransductive proteins that are members of the focal adhesion complex have not been examined in cardiac fibroblasts previously. We hypothesized that loss of talins in cardiac fibroblasts would reduce interstitial fibrosis in the heart with a pressure overload model. However, we found that although loss of talins did not alter fibrosis, it did result in cardiomyocyte and ventricular hypertrophy.

INTRODUCTION

Adverse myocardial remodeling in response to pressure overload is a leading cause of heart failure (1–3). The two principal components of myocardial remodeling in the context of pressure overload are cardiac hypertrophy and interstitial fibrosis (4, 5). These adaptations cause the heart to initially normalize left ventricle (LV) wall stress and maintain cardiac output (6, 7). However, as remodeling continues, interstitial fibrosis causes a stiffening of the heart walls, leading to impaired cardiomyocyte contraction and heart failure.

Cardiac fibrosis is a key feature of the remodeling response and is defined by the accumulation of excessive amounts of extracellular matrix (ECM) proteins, such as collagen and fibronectin (8). Fibrotic remodeling is driven by the phenotypic shift of cardiac fibroblasts (CFs) in the ventricular wall into activated myofibroblasts. Myofibroblasts secrete and compact ECM components such as collagen types I and III as they become contractile, indicated by their expression of α-smooth muscle actin (α-SMA), allowing for short-term adaptation to tissue injury (4, 9–13). These changes result in increasing ECM stiffness during pathological remodeling, which is transmitted to CFs by their focal adhesions. During prolonged pressure overload, increased ECM stiffness results in a positive feedback loop whereby CFs continue to differentiate into myofibroblasts, creating a stiff, noncompliant myocardium. These changes lead to impaired cardiomyocyte contraction, cardiac dysfunction, and heart failure (8, 14, 15).

Focal adhesions are protein complexes that link the cytoskeleton of CFs with the ECM via integrin receptors on the cellular membrane and mechanosensitive proteins. Integrins, through activation of their cytoplasmic tails, bind to actin via mechanosensitive proteins, including talin, vinculin, and α-actinin (16–18). As ECM rigidity increases, outside-in signaling through focal adhesions causes stress fibers within CFs to form (expression of actin fibers and α-SMA instead of depolymerized G-actin) as more mechanosensitive proteins are recruited to focal adhesions and these stress fibers allow CFs to contract (19). This contractility is transmitted to the ECM through inside-out signaling through focal adhesions and allows CFs to compact the ECM, leading to a more defined and rigid ECM in the heart (20–22).

Talins are a family of large, dimeric, cytoskeletal proteins that link the actin cytoskeleton via connections to the cytoplasmic domain of the integrin β-subunit (23, 24). The two talin genes of vertebrates, Tln1 and Tln2, encode very similar proteins with 74% amino acid sequence identity (25). In the adult heart, Tln1 and Tln2 are both highly expressed in CFs, whereas cardiomyocytes express Tln2 predominantly (26). Tln2 has a stronger affinity for F-actin, allowing it to make stronger bonds than Tln1 and, therefore, is expressed highly in cells under constant forces, such as cardiomyocytes (27). When Tln1 and Tln2 are removed in mouse fibroblasts in vitro, focal adhesions are unable to form, illustrating the necessity of talins in cellular mechanotransduction (28). Cardiomyocyte-specific deletion of Tln2 does not affect cardiac structure and function up to 1 yr of age, as Tln1 is upregulated and functionally replaces Tln2 (29). This expression pattern may be protectively redundant in adult mice; however, during development, Tln2 cannot replace Tln1 function in the entire embryo, as Tln1 knockout leads to an embryonically lethal phenotype by E8.5–9.0 (30). Likewise, when Tln1 and Tln2 are both deleted from cardiomyocytes in adult mice, dilated cardiomyopathy develops spontaneously and results in death by 24 wk of age, highlighting the need for a form of talin in cardiomyocytes to maintain function (29). Although deletion of talins in cardiomyocytes has been evaluated, the effect of talin deletion in CFs is unknown.

Here, we describe the effects of ANG II-induced cardiac injury on the hypertrophic and fibrotic response when Tln1 is deleted from CFs in the global Tln2 knockout mouse. We hypothesized that the absence of Tln1 and Tln2 from CFs would result in the attenuation of adverse myocardial remodeling in response to pressure overload.

MATERIALS AND METHODS

Animal Studies

All animal protocols were approved by the Institutional Animal Care and Use Committee at Vanderbilt University. Tln2-null, CF-specific knockout mice were created by crossing C57BL6/H Tln2^−/−;Tln1^flox/flox mice, provided by Dr. Roy Zent (Vanderbilt University Medical Center) (30–32), with the C57BL6 Tcf21-Cre mice, provided by Dr. Michelle Tallquist (University of Hawaii) (33). All mice were crossed with the Rosa26-stop-tdTomato reporter mice (Jackson Laboratory, Stock No. 007914) to visually verify Cre activation (34). The sequences for the primers used are listed in Supplemental Table S1 (https://doi.org/10.6084/m9.figshare.19337933.v1).

All mice were given tamoxifen injections (2 mg/day in PBS^−/−) for 5 days at 9 wk of age to activate the Tcf21-Cre. These mice were then randomly assigned to receive ANG II or saline through osmotic pumps. Chronic hypertension was imposed via ANG II infusion through mini osmotic pumps (Alzet Corp, 1004) on 12-wk-old male and female mice at a concentration of 1.5 mg/kg/day. Control mice were given saline pumps. This surgery was repeated in all mice at 16 wk of age (Fig. 1A). Mice were either euthanized with CO₂ exposure or via exsanguination followed by removal of the heart under isoflurane inhalation continuous at 1%–5% in accordance with Vanderbilt University Medical Center’s Division of Animal Care Guidelines. Littermates were used, and treatment groups were distributed throughout cages and litters.

Figure 1.

A: experimental approach. Nine-week-old mice were subjected to 5 consecutive days of tamoxifen injections. At 12 and 16 wk, pumps were surgically implanted in mice with angiotensin II-treatment (ANG II) or control (saline). Treatment was ceased 8 wk after initial injury. Echocardiography and systolic blood pressure (SBP) measurement were performed at 12 and 20 wk. B: SBP at 20 wk of age. C: quantitative polymerase chain reaction analysis of Nppa, the gene encoding the heart failure marker natriuretic peptide A, at 20 wk of age. B, C: two-way ANOVA, *P < 0.05, ****P < 0.0001 between groups noted with bar. #P < 0.05, ##P < 0.01, ###P < 0.001, ####P < 0.0001 between female groups noted with bar. ^^P < 0.01, ^^^^P < 0.0001 between male groups noted with bar.

Systolic Blood Pressure

Systolic blood pressure was measured using a noninvasive tail-cuff platform (Hatteras Instruments). Two rounds of 10 measurements were taken for each mouse and averaged for each mouse at each timepoint. Four mice were excluded from this study because their systolic blood pressure 8 wk post-ANG II infusion did not indicate hypertension injury, and one control mouse was excluded for having a systolic blood pressure higher than all ANG II mice, indicating a vascular defect.

Echocardiography

Blinded echocardiographic measurements were taken from short-axis cardiac M-mode images captured at the midpapillary level of nonanesthetized mice on a Vevo2100 small-animal ultrasound system (VisualSonics). Three independent measurements were analyzed per mouse for each timepoint.

Quantitative PCR

Quantitative PCR (qPCR) was performed on flash-frozen LV tissue dissected from experimental mice. The sequences for the primers used are listed in Supplemental Table S2. Gene expression was compared with the housekeeping gene Gapdh.

Histology

Upon euthanasia, hearts were perfused with PBS^−/−, excised, and submerged in 3 M potassium chloride to arrest hearts in diastole. Hearts were bisected along their transverse plane. The tissue was frozen and cryosectioned at 7 µm thickness. Picrosirius red staining (Fisher Scientific, No. 50–300-77) was used to identify ECM (red) and cytoplasm (yellow). Images were analyzed using a semiautomated image-processing pipeline that was developed based on color segmentation (34). A minimum of two images that encompassed the entire heart were analyzed per animal. Wheat germ agglutinin (WGA) staining (Invitrogen, No. W11261) was performed for 30 min at room temperature to quantify the cardiomyocyte area, which was calculated using ImageJ (35). A minimum of two LV images per animal were quantified.

RNA Sequencing

LVs from dissected hearts were homogenized in TRIzol reagent, and RNA was isolated with the Symo Direct-zol RNA Microprep Kit (Zymo, R2060). RNA integrity was measured with an Agilent Bioanalyser before library preparations (Supplemental Table S3). Sequencing and read alignment were performed by the Vanderbilt Technologies for Advanced Genomics (VANTAGE) Center, as described by Snider et al. (34), to an average depth of 57.9 M reads per sample. Differential expression analysis was performed with DEseq2 with Cook’s outliers to filter low gene counts (mean count of <6) and P_adj = 0.01 (34, 36). Protein-coding genes with an absolute log₂ fold change of >1 were analyzed. Visualizations were generated with ggplot2 in R. RNA sequencing (RNAseq) data have been deposited in the Gene Expression Omnibus (GEO) of the National Center for Biotechnology Information under accession code GSE189323.

Statistical Analysis

Data were compiled and shown as box and whisker plots with all individual points representing one mouse. The top of the box represents Q1, and the bottom represents Q3. The median is annotated by the center line on the box and whisker plot. Data were evaluated using a two-way ANOVA with a Sidak test to correct for multiple comparisons. All graphs were analyzed and created using GraphPad Prism software (GraphPad Inc., San Diego, CA). A P value < 0.05 was considered significant. In addition, * represents total group differences, # represents differences between females in different groups, ^ represents differences between males in different groups, and $ represents differences between males and females in the same group.

RESULTS

Global Deletion of Tln2 and CF-Specific Deletion of Tln1 in Adult Mice Do Not Result in a Cardiac Phenotype

Due to the ability of Tln2 to compensate for the loss of Tln1, we developed a strategy to delete both talin genes from CFs to determine the role of Tln1 in CFs during cardiac injury. We crossed the talin knockout mouse (Tln2^−/−;Tln1^flox/flox) with the fibroblast-specific Tcf21-Cre, referred to as Tln2^−/−;Tln1^CF−/− mouse. Tln2^−/− mice have a normal cardiac phenotype as compared with wild-type (WT) mice under basal conditions up to 1 yr of age, and therefore, these served as our control (29). Echocardiography (Supplemental Figs. S1 and S2), body weight, and systolic blood pressure (Supplemental Fig. S3) at 12 wk of age showed very little difference between the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice. There was, however, a significant decrease in the heart rate (HR) of the male Tln2^−/−;Tln1^CF−/− mice compared with the male Tln2^−/− mice (Supplemental Fig. S1) and a significant increase in the LVPW;s (left ventricular posterior wall thickness during systole) of the female Tln2^−/−;Tln1^CF−/− mice compared with the female Tln2^−/− mice (Supplemental Fig. S2). In addition, significant decreases in the echocardiographic measurements of heart weight and body weight were observed between the male and female Tln2^−/−;Tln1^CF−/− mice, indicating that the male mice have larger bodies and hearts than the females in this study (Supplemental Figs. S2 and S3).

We then randomized the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice into two groups and subjected them to ANG II or saline for 8 wk (Fig. 1A). At 20 wk of age, there was no increase in systolic blood pressure in either of the two saline groups, indicating that there was no pressure overload injury to the hearts (Fig. 1B). The saline Tln2^−/−;Tln1^CF−/− mice had no change in the expression of the myocardial injury marker atrial natriuretic peptide (Nppa) at 20 wk of age as compared with the saline Tln2^−/− mice (Fig. 1C). In addition, measurements of cardiac hypertrophy and interstitial fibrosis were not changed between the saline mice at 20 wk of age. Overall, these data suggest that under basal conditions, no overall morphological or functional changes to the heart occurred when Tln1 and Tln2 are deleted from CFs.

Female Mice with a Global Deletion of Tln2 and CF-Specific Deletion of Tln1 Develop Exaggerated Systolic Hypertension in Response to ANG II Infusion

Given the absence of significant changes at baseline and under saline conditions, we then analyzed the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice subjected to ANG II for 8 wk (Fig. 1A). Both ANG II groups had an increase in systolic blood pressure compared with their respective saline controls at 8 wk post-ANG II infusion (P < 0.0001) (Fig. 1B). This increase was seen in both male and female mice (both P < 0.0001) (Fig. 1B). In addition, the Tln2^−/−;Tln1^CF−/− mice had a significant increase in systolic blood pressure compared with the Tln2^−/− mice after ANG II infusion (P < 0.05). This increase was only reflected in the female mice (P < 0.01) (Fig. 1B).

Expression of Nppa was quantified in the LV to assess the myocardial injury. Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice had an increase in Nppa (both P < 0.0001) compared with their saline controls at 8 wk post-ANG II infusion (Fig. 1C). The increase in Nppa expression was seen in both male and female mice for both genotypes.

CF Deletion of Tln1 and Tln2 Does Not Affect Heart Hemodynamics during ANG II Infusion

Heart hemodynamics was assessed through echocardiographic analysis. An unchanging ejection fraction (EF) and heart rate (HR) with a decrease in stroke volume (SV) and cardiac output (CO) in all ANG II groups (both P < 0.001) indicated a decrease in end-diastolic volume. (Fig. 2, A–D). The decrease in CO and SV in the ANG II groups was only seen in the female Tln2^−/− mice (both P < 0.01) and the male Tln2^−/−;Tln1^CF−/− mice (both P < 0.01). Echocardiographic measurements of LV volume;d and LV diameter;d showed a decrease in Tln2^−/− mice (both P < 0.05) as compared with their saline controls, and the Tln2^−/−;Tln1^CF−/− mice had a trending decrease compared with their saline controls (Fig. 2, E and F). There were no changes seen in the systolic measurements (Fig. 2, G and H). Dry lung/tibia length ratio was not changed in the ANG II groups, which showed that no congestion occurred (Fig. 2I).

Figure 2.

A–H: echocardiographic analysis of ejection fraction (EF), heart rate (HR), stroke volume (SV), cardiac output (CO), left ventricular volume;d, left ventricular diameter;d, left ventricular volume;s, and left ventricular diameter;s at 20 wk of age, I: dry lung to tibia length ratio of mice at 20 wk of age. A–I: two-way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001 between groups noted with bar. #P < 0.05, ##P < 0.01 between female groups noted with bar. ^^P < 0.01 between male groups noted with bar.

CF Deletion of Tln1 and Tln2 in Female Mice Did Not Result in Cardiac Hypertrophy in Response to ANG II Infusion

Cardiac hypertrophy, in the setting of pressure overload, is a result of cardiomyocyte hypertrophy and interstitial fibrosis. To test if cardiomyocyte hypertrophy was occurring, echocardiographic analysis of LV wall thicknesses and LV mass were measured. The Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice under ANG II infusion, compared with their respective saline controls, had an increase in LVAW;s (left ventricular anterior wall thickness during systole; both P < 0.01), LVAW;d (left ventricular anterior wall thickness during diastole; both P < 0.0001), LVPW;d (left ventricular posterior wall thickness during diastole; both P < 0.0001), IVS;s (intraventricular septal thickness during systole; P < 0.001 and P < 0.01), and IVS;d (intraventricular septal thickness during diastole; both P < 0.00001), whereas LVPW;s was unchanged (Fig. 3, A–F). All changes between the Tln2^−/−;Tln1^CF−/− mice and their saline controls were only seen in the male mice, whereas changes between the Tln2^−/− mice and their saline controls were observed in both females and males.

Figure 3.

A–G: echocardiographic analysis of left ventricular anterior wall thickness during systole (LVAW;s) and diastole (LVAW;d), intraventricular septal thickness during systole (IVS;s) and diastole (IVS; d), left ventricular posterior wall thickness during systole (LVPW;s) and diastole (LVPW;d), and left ventricular mass (LV mass) at 20 wk of age. H: ventricles weight to tibia length ratio at 20 wk of age. A–H: two-way ANOVA, **P < 0.01, ***P < 0.001, ****P < 0.0001 between groups noted with bar. #P < 0.05, ##P < 0.01, ###P < 0.001 between female groups noted with bar. *P < 0.05, ^P < 0.05, ^^P < 0.01, ^^^P < 0.001 between male groups noted with bar. $P < 0.05 between male and female mice of that specific genotype.

LV mass was increased in both the Tln2^−/− mice (P < 0.05) and the Tln2^−/−;Tln1^CF−/− mice (P < 0.01) under ANG II infusion as compared with their respective saline controls (Fig. 3G). This difference was only present in the male mice (P < 0.05 and P < 0.01). No changes in ventricle/tibia ratio were found (Fig. 3H).

CF Deletion of Tln1 and Tln2 in Male Mice Results in Cardiomyocyte Hypertrophy in Response to ANG II Infusion

The increase in LV wall thickness in the ANG II animals was supported by measurements of cardiomyocyte hypertrophy. Histologic analysis showed that the Tln2^−/−;Tln1^CF−/− mice had an increase in cardiomyocyte area (P < 0.001) at 8 wk post-ANG II treatment, whereas the Tln2^−/− mice had no change (Fig. 4). This change was only seen in the male mice (P < 0.0001). In addition, the ANG II male Tln2^−/−;Tln1^CF−/− mice had a significant increase in cardiomyocyte area compared with the male ANG II Tln2^−/− mice. There was a significant difference between the male and female Tln2^−/−;Tln1^CF−/− saline (P < 0.05) and Tln2^−/−;Tln1^CF−/− ANG II-infused mice (P < 0.0001) (Fig. 4). This increase in cardiomyocyte area illustrates that at baseline, male Tln2^−/−;Tln1^CF−/− mice had larger cardiomyocytes under saline conditions than the female mice. It also showed that under ANG II infusion, male Tln2^−/−;Tln1^CF−/− cardiomyocytes increased in the area, whereas female mice showed no changes.

Figure 4.

A: representative images of wheat germ agglutinin (WGA) stain on the left ventricle of mice at 20 wk of age. B: quantification of WGA staining to determine cardiomyocyte area. B: two-way ANOVA, ***P < 0.001 between groups noted with bar. ^P < 0.05, ^^^^P < 0.0001 between male groups noted with bar. $P < 0.05, $$$$P < 0.0001 between male and female mice of that specific genotype.

CF Deletion of Tln1 and Tln2 Results in Minimal Changes in Cardiac Fibrosis Burden following ANG II Infusion

To measure the effects of CF Tln1 and Tln2 on fibrotic remodeling, we measured interstitial fibrosis and fibroblast activation following ANG II infusion. Picrosirius red staining was performed on transverse sections of the heart and quantified to determine fibrosis. Image quantification of the ratio of the fibrotic to nonfibrotic tissue showed that the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice both had a significant increase (P < 0.0001 and P < 0.05, respectively) in fibrosis after 8 wk of ANG II infusion compared with saline controls, but they were not significantly different from each other (Fig. 5, A and B). The increase in interstitial fibrosis was seen only in male and female Tln2^−/− mice (P < 0.05 and P < 0.01, respectively) and was not present in Tln2^−/−;Tln1^CF−/− mice.

Figure 5.

A: representative images of picrosirius red (PSR) stained hearts at 20 wk of age and their corresponding ×40 images indicated with black boxes. B: quantification of PSR staining of hearts. Collagen (red pixels) and heart tissue (yellow) were used to calculate the average collagen fraction in each heart. C–E: quantitative polymerase chain reaction analysis of Col1a1, the gene that encodes α1 type 1 collagen; Col4, the gene that encodes collagen IV; and α-SMA, the gene encoding the myofibroblast marker α-smooth muscle actin, in left ventricular tissue at 20 wk of age. B–E: two-way ANOVA, *P < 0.05, ****P < 0.0001 between groups noted with bar. #P < 0.05, ##P < 0.01, ###P < 0.001, ###P < 0.0001 between female groups noted with bar. ^P < 0.05, ^^P < 0.01, ^^^^P < 0.0001 between male groups noted with bar.

qPCR was performed on LV tissue to identify changes in collagen expression during fibrosis. α1 type 1 collagen (Cola1a1) was significantly increased in the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice compared with their saline controls (P < 0.05 and P < 0.0001, respectively) (Fig. 5C). The Tln2^−/− expression of Col1a1 was only significantly increased in the female mice (P < 0.05), whereas Col1a1 expression was significantly increased in both male and female Tln2^−/−;Tln1^CF−/− mice (both P < 0.01). Although not significant, ANG II-infused Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice had an upward trend of collagen IV (Col4) mRNA expression (Fig. 5D). αSMA, a marker of CF-to-myofibroblast transition, was increased in both the ANG II-infused Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice (both P < 0.001; Fig. 5E). RNAseq analysis of LV tissue did not provide further insight to changes in RNA expression after 8 wk of ANG II infusion (Supplemental Fig. S4).

DISCUSSION

The purpose of the present study was to evaluate the functional significance of talins in CFs during ANG II-induced injury. This was accomplished by developing a novel transgenic mouse: a Tln2-null mouse with Tln1 specifically deleted in CFs. The results illustrate that mechanotransduction in CFs through Tln1 and Tln2 may be an important mediator of cardiomyocyte hypertrophy and interstitial fibrosis and that these effects are discordant in male and female mice.

Under basal conditions, echocardiography showed limited differences in heart hemodynamics, LV chamber size, and LV chamber thickness between the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice. This suggests that talins in CFs are not needed to maintain cardiac function under basal conditions. Next, we subjected the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice to ANG II infusion for 8 wk to evaluate cardiomyocyte response under prolonged pressure overload of the heart. We found that the absence of Tln1 and Tln2 in CFs resulted in a hypertrophy-resistant phenotype in female mice, whereas male mice had an observed increase in cardiomyocyte hypertrophy with ANG II infusion.

Our data indicated that there was a sex-specific response to ANG II infusion in our study. This was seen primarily in the female mice, where the loss of Tln1 in CFs in the Tln2-null mouse resulted in an overall hypertrophy-resistant phenotype. This was seen despite an increase in systolic blood pressure in the female ANG II Tln2^−/−;Tln1^CF−/− mice as compared with the ANG II Tln2^−/− mice. When looking specifically at the female mice, markers of LV hypertrophy (chamber volume, diameter, LVAW, LVPW, and IVS) all increased in the Tln2^−/− mice as compared with saline controls, whereas there was no increase in these measurements in the Tln2^−/−;Tln1^CF−/− mice versus saline controls. In addition, interstitial fibrosis and Nppa mRNA expression were elevated in the Tln2^−/− mice compared with saline controls, whereas there was no change or less of a change in the Tln2^−/−;Tln1^CF−/− mice versus saline controls. These markers suggest that loss of talins results in a hypertrophy-resistant phenotype in female mice.

This is a specifically surprising finding, as the afterload of the female Tln2^−/−;Tln1^CF−/− mice under ANG II infusion, apparent in the increased systolic blood pressure, is increased as compared with the ANG II Tln2^−/− mice. The increase in systolic blood pressure seen in the Tln2^−/−;Tln1^CF−/− mice could be the result of using the Tcf21 as the driver for the Cre, instead of the increase in systolic blood pressure being a direct cause of increased pressure due to loss of both talins in cardiac fibroblasts. Tcf21 is highly expressed in resident CFs and myofibroblasts and is considered the best marker of all fibroblast populations in the heart (37). However, Tcf21 is also expressed in podocytes in the kidney (38). Integrin adhesions are necessary for podocytes to withstand the hydrostatic pressure in the kidney, and podocyte injury results in an increase in systolic blood pressure (39, 40). Therefore, systolic blood pressure could be increased in the female Tln2^−/−;Tln1^CF−/− mice due to Tln1 and Tln2 being removed from podocytes, resulting in podocyte injury. This increase in systolic blood pressure would result in more ventricular wall stress in the heart and should result in an increase in cardiac remodeling. Therefore, by not observing a resulting increase in cardiac remodeling in the female Tln2^−/−;Tln1^CF−/− mice under ANG II infusion, this further illustrates the hypertrophy-resistant effect of loss of talins in female mice.

Due to the increased systolic blood pressure in the Tln2^−/−;Tln1^CF−/− mice as compared with the Tln2^−/− mice after 8 wk of ANG II infusion, we expected to see an increase in interstitial fibrosis in the Tln2^−/−;Tln1^CF−/− mice since the hearts would have been under more pressure. Surprisingly, we observed the opposite, where the ANG II Tln2^−/− mice had an increase in interstitial fibrosis, whereas the Tln2^−/−;Tln1^CF−/− mice did not. However, this does not mean that the absence of Tln1 and Tln2 in CFs did not alter ECM deposition in these mice. The higher systolic blood pressure in the Tln2^−/−;Tln1^CF−/− mice would result in an increase in LV wall stress. Because there is no increased interstitial fibrosis in the female Tln2^−/−;Tln1^CF−/− mice but an increase in Col1a1 mRNA expression, this suggests that the absence of Tln1 and Tln2 in female CFs is either hampering the ability of CFs to compact collagen into fibrils and/or hampering the ability to translate collagen mRNA to protein. Due to the limitations of our current study, we cannot say if this is a protective mechanism that prevents the increase in interstitial fibrosis or merely delays the onset of interstitial fibrosis during constant ANG II infusion. To answer these questions, this study should be repeated over a longer time course and with a higher concentration of ANG II in female mice. Systolic blood pressure should be taken throughout the experiment to establish the time course of this phenotype and if an increase in collagen mRNA but with no resulting fibrosis occurs during a more progressive and/or acute injury.

All our mice had relatively high heart rates at baseline and at 20 wk post-ANG II or saline infusion for both genotypes. Manso et al. (29) showed that the Tln2^−/− mice have normal heart rates as compared with WT mice at 6 and 12 mo of age. Therefore, we concluded that the increased heart rate that we see at baseline and throughout our study is due to the tamoxifen injections that were given at 9 wk to all mice that were included in this study.

When looking specifically at the male mice, loss of Tln1 in CFs in the Tln2-null mouse resulted in an overall increase in cardiomyocyte hypertrophy. This was shown through an increase in markers of LV wall thickness and LV mass in the Tln2^−/−;Tln1^CF−/− mice as compared with their saline controls that was larger than the increase seen in the Tln2^−/− mice as compared with their saline controls. This was congruent with an increase in cardiomyocyte area in the Tln2^−/−;Tln1^CF−/− mice when compared with both their saline control and the Tln2^−/−;Tln1^CF−/− mice with ANG II infusion.

Although Col1a1 and Nppa were increased in the ANG II Tln2^−/−;Tln1^CF−/− mice, there was no observed increase in interstitial fibrosis. Limited changes in interstitial fibrosis may indicate that fibrosis is not the primary factor driving cardiomyocyte hypertrophy in the Tln2^−/−;Tln1^CF−/− mice and, instead, is the remodeling of the heart that occurs due to cardiomyocyte hypertrophy. If this were the case, this experiment should be repeated and the timeline extended to see if an increase in interstitial fibrosis occurred in the male Tln2^−/−;Tln1^CF−/− mice first and then appeared in the Tln2^−/− mice after they display an increase in cardiomyocyte hypertrophy.

However, an absence of increased cardiac fibrosis could also be driving the increase in cardiomyocyte hypertrophy that is observed. An absence of cardiac fibrosis may result in increased LV wall stiffness to compensate for the lack of increased collagen content in the Tln2^−/−;Tln1^CF−/− mice under ANG II infusion. This could lead to increased cardiomyocyte hypertrophy in the ANG II Tln2^−/−;Tln1^CF−/− mice that is seen in our study (41). The ability to measure the LV wall stiffness is a limitation of the current study, as we are only able to measure the collagen content present in the heart.

The ANG II infusions given to the mice to induce hypertension may be causing downstream signaling effects in the myocardium that are independent of the increase in systolic blood pressure of hypertension signaling that is occurring in the heart. To test if there was altered ANG II signaling in our treatment groups, as well as to elucidate changes between the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice, we performed RNAseq analysis of the LV of male and female mice after 8 wk of ANG II infusion. The results of our RNAseq analysis returned 10 genes that were differentially regulated between Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice. Due to only having 10 differentially expressed genes, we were unable to perform further analysis to determine specific groups of genes that were changing between our genotypes. What this indicates is that the differences in hypertrophy and interstitial fibrosis that we are observing in our mouse experiment are most likely due to posttranslational modifications in these mice, and therefore, they are not picked up in our RNAseq analysis. It was interesting that Agtr1a, the gene that encodes the ANG II receptor type 1a, was unchanged between the Tln2^−/− and Tln2^−/−;Tln1^CF−/− mice. This indicates that both genotypes can respond to ANG II and that if ANG II infusion was causing changes between these mice, then it is most likely during the ANG II signaling cascaded where posttranslational modifications could alter signaling. Posttranslational modifications have been shown to play a role in myocardial hypertrophy pathways that can cause modifications of proteins that can include but are not limited to protein regulation, diversity, localization, structure, and interactions (42). An example of this is the posttranslational modification of collagen 1 matrices, through hydroxylation of a distinctive lysyl residue to hydroxylysine, that results in collagen cross links that are less susceptible to matrix metalloproteinase-1 degradation. This results in an increase in overall collagen content, contributing to interstitial fibrosis in the heart (43, 44). To study this, future studies should focus on identifying global posttranscriptional modifications through immunoprecipitation and Western blot analysis.

In conclusion, this study demonstrates that global deletion of Tln2 and CF-specific deletion of Tln1 result in a hypertrophy-resistant phenotype following ANG II infusion in female mice by measurement of cardiac hypertrophy and interstitial fibrosis. Furthermore, the absence of Tln1 and Tln2 in CFs results in enhanced cardiomyocyte hypertrophy in male mice in the setting of ANG II-induced pressure overload.

SUPPLEMENTAL DATA

Supplementary Figures S1–S4 and Tables S1–S3: https://doi.org/10.6084/m9.figshare.19337933.v1.

GRANTS

This work was funded by National Heart, Lung, and Blood Institute Grants HL149168 (to N.A.N.), HL154596 (to L.A.R.), HL146951 (to M.R.B.), and HL135790 (to W.D.M.) and the Foundation Leducq (to W.D.M.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.A.N., R.Z., and W.D.M. conceived and designed research; N.A.N., C.S.M., and L.Z. performed experiments; N.A.N., L.A.R., L.Z., and M.R.B. analyzed data; N.A.N., L.A.R., L.Z., M.R.B., J.D.W., and R.Z. interpreted results of experiments; N.A.N. and L.A.R. prepared figures; N.A.N. drafted manuscript; N.A.N., L.A.R., C.S.M., M.R.B., J.D.W., R.Z., and W.D.M. edited and revised manuscript; N.A.N., L.A.R., C.S.M., L.Z., M.R.B., J.D.W., R.Z., and W.D.M. approved final version of manuscript.