Ubiquitin ligase Wwp1 gene deletion attenuates diastolic dysfunction in pressure-overload hypertrophy

Abstract

Heart failure with a preserved left ventricular (LV) ejection fraction (HFpEF) often arises from a prolonged LV pressure overload (LVPO) and accompanied by abnormal extracellular matrix (ECM) accumulation. The E₃ ubiquitin ligase WWP1 is a fundamental determinant ECM turnover. We tested the hypothesis that genetic ablation of Wwp1 would alter the progression of LVPO-induced HFpEF. LV echocardiography in mice with global Wwp1 deletion (n = 23; Wwp1^−/−) was performed at 12 wk of age (baseline) and then at 2 and 4 wk following LVPO (transverse aortic banding) or surgery without LVPO induction. Age-matched wild-type mice (Wwp1^+/+; n = 23) underwent identical protocols. LV EF remained constant and unchanged with LVPO and LV mass increased in both groups but was lower in the Wwp1^−/− mice. With LVPO, the E/A ratio, an index of LV filling, was 3.97 ± 0.46 in Wwp1^+/+ but was 1.73 ± 0.19 in the Wwp1^−/− group (P < 0.05). At the transcriptional level, mRNA for fibrillar collagens (types I and III) decreased by approximately 50% in Wwp1^−/− compared with the Wwp1^+/+ group at 4 wk post-LVPO (P < 0.05) and was paralleled by a similar difference in LV fibrillar collagen content as measured by histochemistry. Moreover, mRNA levels for determinants favoring ECM accumulation, such as transforming growth factor (TGF), increased with LVPO, but were lower in the Wwp1^−/− group. The absence of Wwp1 reduced the development of left ventricular hypertrophy and subsequent progression to HFpEF. Modulating the WWP1 pathway could be a therapeutic target to alter the natural history of HFpEF.

NEW & NOTEWORTHY Heart failure with a preserved left ventricular (LV) ejection fraction (HFpEF) often arises from a prolonged LV pressure overload (LVPO) and is accompanied by abnormal extracellular matrix (ECM) accumulation. It is now recognized that the ECM is a dynamic entity that is regulated at multiple post-transcriptional levels, including the E₃ ubiquitin ligases, such as WWP1. In the present study, WWP1 deletion in the context of an LVPO stimulus reduced functional indices of HFpEF progression and determinants of ECM remodeling.

INTRODUCTION

A chronic disease that afflicts millions of patients with a growing incidence is a specific form of heart failure (HF) termed HF with a preserved ejection fraction (HFpEF) (1–4). HFpEF results in a poor quality of life and is associated with significant morbidity and mortality (1). For example, following the first HFpEF related hospitalization, the one-year mortality rate is ∼25% and the 5-year mortality rate exceeds 54%. HFpEF can arise from a chronic left ventricular (LV) pressure overload (LVPO) stimulus, such as hypertension or vascular disease/stiffening, which can be exacerbated as a function of age (1, 4, 5). In this context, a process termed LV remodeling occurs, which is characterized by myocardial growth (LV hypertrophy) and an exuberant expansion of the extracellular matrix (ECM; generically termed myocardial fibrosis) (6–8). This unbridled growth of the ECM is a contributory factor to HFpEF development and progression because it causes increased LV chamber stiffness and thus impaired LV filling as well as increased LV filling pressures, such as pulmonary capillary wedge pressure and left atrial enlargement (4, 6, 8). Defining molecular pathways that contribute to the underlying pathophysiology of HFpEF will likely yield novel research directions and therapeutic strategies.

LV myocardial ECM accumulation, particularly of the fibrillar collagens, occurs due to dysregulation in a number of transcriptional and post-translational pathways. Thus, shifts in post-translational modification/turnover of both transcription factors as well as myocardial ECM proteins, such as the fibrillar collagens, would hold relevance in the context of HFpEF. One such post-translational modification is through ubiquitination, a process that relies on the sequential action of specific enzymes, in particular the E₃ ubiquitin ligase family, to regulate protein turnover (9–12). Shifts in the expression of a specific E₃ ubiquitin ligase, termed WWP1, have been implicated in the pathological remodeling processes, such as cancer (10–12). More specifically, increased levels of WWP1 facilitate invasion of cancer cells through the ECM, alter fibroblast form and function, and cause a shift in ECM protein expression (13–15). Our recent studies have identified that shifts in WWP1 expression in and of itself can alter myocardial ECM and influence LV function (16, 17). However, whether targeted alterations in WWP1 in the context of LVPO would alter the progression to HFpEF remained unexplored. Accordingly, the present study tested the hypothesis that the genetic ablation of WWP1 would alter the progression to HFpEF secondary to a LVPO overload stimulus and would be associated with shifts in the expression of determinants of ECM remodeling.

METHODS

Animals

The mice used in this study were of the C57BL/6J background strain in which global Wwp1 deletion was achieved (B6JNci.Cg-Wwp1^tm1Lmat/Mmnc, RRID:MMRRC_066735-UNC). For the purposes of nomenclature, the wild-type background strain will be referred to as Wwp1^+/+ and the knockout line as Wwp1^−/−. Mice were entered into the experimental protocol at 12 wk of age with a relatively balanced sex distribution (Wwp1^+/+: n = 11 females/n = 12 males, Wwp1^−/−: n = 9 females, n = 14 males). All mice were cared for in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals (8th ed., Washington, DC: 2011), and the University of South Carolina’s Institutional Animal Care and Use Committee approved all protocols.

Experimental Protocol

All mice first underwent a baseline LV function study by echocardiography as described previously and outlined in a subsequent section (17). First, the mice were anesthetized with isoflurane (2%) and following a 30-min equilibration period, transthoracic echocardiography was performed. Following which, mice were randomized to undergo LV pressure overload (LVPO) created by surgical constriction of the transverse aorta using previously described approaches (18, 19) or the surgical procedure without aortic constriction. Briefly, consistent body temperatures were maintained in all mice while a skin incision of 2–3 mm was made across the proximal sternum. The transverse aorta was banded with a 6-0 silk suture firmly tied around a 27-gauge needle, which was quickly removed after the ligature was placed. All echocardiographic and surgical procedures were performed with the genotypes coded and blinded until completion of the study.

LV Function Studies

Transthoracic echocardiography was performed on all mice to obtain parameters of LV function and geometry (VEVO 3100 ultrasound system/40-MHz transducer; VisualSonics, Toronto, Canada). Specifically, end systolic/diastolic volumes (ESV, EDV), left atrial (LA) area, and EF were computed using short- and long-axis views. LV mass was calculated using the following formula: 0.8 ×·(1.04·× {[(LVID + IVSd + LVPWd)3 − LVIDd3)]} + 0.6), where the variables are LV interior diameter (LVID), interventricular septum (IVS), LV posterior wall (LVPW), and diastole (d). Doppler velocities of LV filling (early E, late/atrial A; E/A ratio) were measured from quantifying respective pulse waves. All measurements were completed using three consecutive cardiac cycles of images captured on cine loops, where repeated measurements were performed by two individuals separately and were blinded to the genotype and banding status of the mice.

Quantitative Real-Time PCR

RNA was extracted from LV samples, converted to cDNA (RT² First Strand Kit, Qiagen, Cat. No. 330401, Valencia, CA), and quantitative real-time PCR (qPCR) performed (TaqMan Gene Universal PCR Master Mix Kit, Applied Biosystems, Cat. No. 4364338, Foster City, CA) with specific mRNA expression primers/probes (Qiagen, Valencia, CA) identified in Supplemental Table S1 (see https://doi.org/10.5281/zenodo.5557538). Briefly, a custom mRNA array was constructed that contained key determinants of ECM remodeling. In addition, Wwp1 mRNA levels were examined using murine specific primers (GenBank, NM 001276292.1) and described previously (17). The qPCR reactions (Bio-Rad CFX96, Bio-Rad, Hercules, CA) were carried out using cycling parameters including an activation step of one cycle for 10 min at 95°C followed by 40 cycles of 15 s at 95°C and 60°C for 1 min. Cycle time (Ct) values of the geometric mean for Gapdh and β₂-microglobulin were used as reference gene values for normalization. The final values were computed as a relative fold change using a modification of the ΔΔCt approach described previously (17).

Fibrillar Collagen Content

LV sections were fixed in optimal cutting temperature compound, frozen, and sliced into 55 µm sections and placed on the motorized stage of a multiphoton microscope (Leica TCS SP8 MP, Leica Microsystems). Short-pulsed infrared lasers were used to create nonlinear polarization effects in the form of second harmonic generation (SHG) signals at a magnification of ×20. SHG images were acquired for a fixed sample area of 0.2 mm² at depth intervals of 1 µm and analyzed as described previously (20). The images were virtually stacked in three-dimensions using the Leica LAS × 3-D analysis software. The collagen fibers were identified and computed as a function of the total sample field area. A minimum of 5 fields/sample were used.

Data Analysis

The LV function and geometry data were first subjected to multiway analysis of variance (MANOVA) whereby the main treatment effects were genotype, LVPO, and time. A secondary analysis was performed whereby sex was treated as a covariate in the MANOVA. The post hoc pairwise analyses were performed with a Bonferroni adjusted t test. In addition, relative changes in function and geometry variables were transformed as a function of respective no LVPO values, and hypothesis testing against a mean value of zero was performed using a t test. The trajectory of LA area with respect to time was subjected to linear regression and the regression coefficient subjected to t test. The fold change index from the PCR results was computed as described previously and subjected to a one-way analysis of variance or t test, where appropriate (17). Fibrillar collagen content was subjected to ANOVA with a Bonferroni adjusted t test. Data are presented as means ± SE.

RESULTS

The summary of LV function measurements with respect to strain and time, as well as final sample sizes are provided in Table 1. At the time of LV function measurements, there were no differences in heart rate between groups. For example, baseline heart rates were identical between Wwp1^+/+ and Wwp1^−/− groups (397 ± 7 vs. 400 ± 12 beats/min, respectively; P > 0.70). Representative LV echocardiographic images are shown in Fig. 1, in which the increased LV wall thickness was evident following 4 wk of LVPO, but wall motion appeared preserved. Indeed, LVPO did not cause a significant fall in LV EF from respective baseline values. LV EF was slightly but significantly lower in the Wwp1^−/− mice when compared with time-matched wild-type (Wwp1^+/+) mice. In both groups, LV end diastolic volume remained unchanged from baseline values, but LV mass robustly increased at 4 wk of LVPO. An index of LV filling, the E/A ratio changed in both groups at 4 wk of LVPO. LA area increased in both groups with LVPO. However, the magnitude of change in E/A ratio and LA area was greatest in the Wwp1^+/+ mice at 4 wk of LVPO.

Table 1.

Left ventricular function and geometry with LVPO: effects of Wwp1 deletion

	EF, %	EDV, µL	LV Mass, mg	E/A	LA Area, mm2
Wwp1+/+
Baseline	65 ± 2	59 ± 3	82.9 ± 2.1	1.86 ± 0.11	4.66 ± 0.11
Day 14
No LVPO	64 ± 3	57 ± 3	86.7 ± 4.5	2.01 ± 0.17	4.73 ± 0.22
LVPO	62 ± 4	51 ± 4	125.1 ± 4.5*†	2.65 ± 0.35	5.66 ± 0.23*†
Day 28
No LVPO	65 ± 2	59 ± 2	90.2 ± 3.3	1.92 ± 0.14	4.92 ± 0.15
LVPO	59 ± 2	59 ± 4	138.4 ± 9.0*†	3.97 ± 0.46*#†	5.69 ± 0.20*
Wwp1−/−
Baseline	56 ± 2†	65 ± 3	91.1 ± 3.5	1.74 ± 0.10	4.20 ± 0.13#
Day 14
No LVPO	61 ± 3	58 ± 3	87.0 ± 4.5	1.83 ± 0.16	4.72 ± 0.12
LVPO	58 ± 4	57 ± 5	108.2 ± 6.7*	2.34 ± 0.31	4.86 ± 0.28†#
Day 28
No LVPO	55 ± 2†	62 ± 3	90.9 ± 4.3	1.87 ± 0.14	4.52 ± 0.12†#
LVPO	52 ± 4	64 ± 6	122.1 ± 9.5*†	1.73 ± 0.19†#	4.90 ± 0.23†#

Values are means ± SE; n = 9 Wwp1^−/− LVPO, n = 11 Wwp1^+/+ no LVPO, n = 12 Wwp1^+/+ LVPO, and n = 14 Wwp1^−/− no LVPO. EDV, end-diastolic volume; EF, ejection fraction; LV, left ventricle; E/A, Doppler velocities: early/late atrial; LA, left atrium; LVPO, left ventricular pressure overload.

^*P < 0.05 vs. respective baseline value;

^†P < 0.05 vs. time-matched LVPO;

^# P < 0.05 vs. respective time-matched Wwp1^+/+.

Figure 1.

Left ventricular (LV) echocardiography. Representative LV echocardiographic studies from a Wwp1^+/+ and Wwp1^−/− mouse at baseline (before induction of LV pressure overload, LVPO) and at 4 wk of LVPO. The left panels are representative of two-dimensional short-axis (circumferential) views of the LV whereby LV dimensions and wall thickness were computed using M-mode (right). With LVPO, increased LV posterior wall (PW) and interventricular septal (IVS) wall thickness was evident, but LV shortening was preserved. The LV function and mass parameters determined from these echocardiographic studies are summarized in Table 1.

In addition to the absolute values for LV function as shown in Table 1, the relative changes in LV EF, mass, and E/A ratio as a function from respective baseline values are shown in Fig. 2. LV EF remained within baseline values in both groups at both 2 and 4 wk of LVPO. LV mass robustly increased by 14 days of LVPO in both groups, but the magnitude of the LV hypertrophic response was reduced in the Wwp1^−/− group when compared with respective Wwp1^+/+ mice. In contrast to the Wwp1^+/+ mice, the increase in the E/A ratio was blunted in the Wwp1^−/− group. The relative change in LA area is shown in Fig. 2. Although LA dilation occurred by 2 wk of LVPO in the Wwp1^+/+ group, the change in LA area was significantly attenuated in the Wwp1^−/− group. This difference in LA dilation persisted at 4 wk of LVPO.

Figure 2.

Changes in left ventricular (LV) function and left atrial (LA) geometry with LV pressure overload (LVPO). A: LV ejection fraction (EF) remained unchanged from baseline values and from no-LVPO values, consistent with the heart failure with a preserved LV EF (HfpEF) phenotype. Values presented as an absolute change in LVEF from baseline. B: LV mass quantified computed from measurements of LV wall thickness and volumes increased in both groups with LVPO, but LV mass was reduced in the Wwp1^−/− group. Values presented as an absolute change in LV mass from baseline. C: an index of LV diastolic function is the ratio of early and late filling (E/A ratio) determined by Doppler echocardiography. At 14 days post-LVPO, the E/A ratio increased in both groups. However, by day 28 post-LVPO the E/A ratio increased in the Wwp1^+/+ group and had returned to baseline in the Wwp1^−/− group. D: as a function from baseline values, the magnitude of LA dilation was highest in the Wwp1^+/+ group at both 14 and 28 days post-LVPO. Absolute values and sample sizes for these parameters are provided in Table 1 (*P < 0.05 vs. baseline; +P < 0.05 vs. respective Wwp1^+/+ time point).

With respect to the primary response variable LV mass, LV mass was identical between Wwp1^+/+ males and females at baseline (83 ± 3, 83 ± 3 mg, respectively, P > 0.90) and remained similar at 4 wk of LVPO (144 ± 12, 130 ± 14 mg, respectively, P > 0.50). In the Wwp1^−/− group, baseline LV mass was greater in the males compared with females (97 ± 4, 82 ± 5 mg, respectively, P < 0.05) and this trend continued at 4 wk of LVPO (130 ± 13, 106 ± 8 mg, respectively, P = 0.08). However, there were no sex-dependent differences in the indices of LV systolic and diastolic function (summarized in Table 1) at baseline or at 4 wk post-LVPO in either WT or Wwp1^−/− mice.

LV samples were collected for targeted PCR at 4 wk of LVPO and the findings for Wwp1 expression are summarized in Fig. 3. As expected, there was no positive signal for Wwp1 in the Wwp1^−/− group, however, Wwp1 significantly increased in the Wwp1^+/+ animals after 4 wk of LVPO. A summary of relative mRNA levels for the fibrillar collagens (type I and III), specific matrix metalloproteinases (MMPs), and the tissue inhibitor of MMP-1 (TIMP-1), as well as determinants of a profibrotic cascade (i.e., transforming growth factor β, TGF) are provided in Fig. 4. Overall, indices of fibrillar collagen expression increased with LVPO, but the mRNA levels for collagen type I were lower in the Wwp1^−/− group. Expression profiles for TGF and indices of cellular growth and viability, such as Myc, increased with LVPO and were lower in the Wwp1^−/− group. In light of the changes in TGF with LVPO and Wwp1^−/−, relative mRNA levels for the intracellular signaling proteins Smad 1–7 are shown in Fig. 5. There were divergent and variable changes in Smad mRNA levels with LVPO, with a reduction in Smad 4 in the Wwp1^+/+ LVPO group, but not in the Wwp1^−/− LVPO group. In both LVPO groups, Smad 5 levels were reduced from respective no LVPO values. Interestingly, Smad 7 levels fell in the Wwp1^+/+ LVPO group, with directionally opposite levels in the Wwp1^−/− LVPO group.

Figure 3.

Wwp1 expression increases with left ventricular pressure overload (LVPO). Left ventricular (LV) samples were subjected to PCR at the completion of the study interval (28 days) for mRNA levels of Wwp1. As expected, Wwp1 was nondetectable in Wwp1^−/− groups. In the Wwp1^+/+ LVPO group, a significant increase in Wwp1 occurred at 14 days. (+P < 0.05 vs. no LVPO). Sample sizes provided in Table 1.

Figure 4.

Differential mRNA levels for determinants of ECM remodeling with left ventricular pressure overload (LVPO). Steady-state mRNA levels for fibrillar collagens, TGF, and determinants of cell viability were examined in all groups. Significant shifts in several of these determinants of ECM remodeling changed at 28 days post-LVPO. Several determinants were reduced in the Wwp1^−/− LVPO group. Please see Supplemental Table S1 for definitions. Please see final sample sizes in Table 1. mRNA levels expressed as: (2^−ΔCt ×·100). (*P < 0.05 vs. Wwp1^+/+; +P < 0.05 vs. respective no LVPO).

Figure 5.

Differential changes in mRNA levels for Smads. Steady-state mRNA levels for Smad 1–7, which are components of the TGF signaling cascade. There was significant variation associated with mRNA levels, with no detectable differences between Wwp1^−/− and Wwp1^+/+ LVPO groups with respect to Smads 1–3. Smads 4, 5, and 7 fell in the Wwp1^+/+ left ventricular pressure overload (LVPO) group but a dissimilar pattern emerged in the Wwp1^−/− LVPO group, whereby Smad 4 did not change from no LVPO values, Smad 5 changed in a concordant direction to the Wwp1^+/+ LVPO group, and Smad 7 mRNA levels changed in a directionally opposite degree to that of the Wwp1^+/+ LVPO group. Please see final sample sizes in Table 1. mRNA levels expressed as: (2^−ΔCt·× 00). (*P < 0.05 vs. Wwp1^+/+; +P < 0.05 vs. respective no LVPO).

To determine if the shifts in mRNA expression profiles for determinants of collagen content observed with LVPO and with Wwp1^−/− would be translated to the structural level, fibrillar collagen content using quantitative SHG were performed and representative images and summary data are provided in Fig. 6. As expected, LV myocardial levels for fibrillar collagen were quite low with no LVPO, and similar between Wwp1^+/+ and Wwp1^−/− groups. With LVPO, however, a robust increase in LV collagen content occurred in both LVPO groups, but was lower in the Wwp1^−/− group.

Figure 6.

Fibrillar collagen content with left ventricular pressure overload (LVPO). Right: fibrillar collagen content using quantitative second harmonic generation (SHG) was performed and representative images are shown in the right. As expected, detectable collagen fibrils were quite low in the murine left ventricular (LV) myocardial samples with no LVPO and appeared similar between Wwp1^+/+ and Wwp1^−/− groups. With LVPO, however, a robust increase in LV collagen content occurred was easily detectable in both LVPO groups, but appeared lower in the Wwp1^−/− group. Left: quantitative results revealed that LVPO induced a robust increase in LV fibrillar collagen content, but was reduced in the Wwp1^−/− group. Sample sizes used in this analysis are shown in Table 1. Bar in left represents 50 µm. (*P < 0.05 vs. Wwp1^+/+; +P < 0.05 vs. respective no LVPO).

DISCUSSION

The more common causes for LV hypertrophy and the progression to HFpEF include a chronic LV pressure overload (LVPO), such as hypertension and vascular stiffening. However, despite adequate blood pressure control, the progression to HFpEF can continue, which is particularly true with aging (1–6). Shifts in protein turnover and stability occur within the extracellular matrix with LVPO, commonly causing an exuberant accumulation of the fibrillar collagens, such as type I and III, i.e., fibrosis (6–8). Abnormal ECM accumulation and shifts in critical protein turnover also occur as a function of aging and have been associated with shifts in the expression of the family of ubiquitin ligases (9, 10, 21). Our past studies using transgenic overexpression of the specific ubiquitin ligase, WWP1, was associated with changes in LV function and structure similar to the HFpEF phenotype (17). Accordingly, the present study examined the effects of Wwp1 deletion (Wwp1^−/−) in a mouse model of LVPO and progression to HFpEF. The significant and unique findings of the present study were twofold. First, although LVPO in Wwp1-null mice induced an adaptive LV hypertrophy, indices of LV diastolic dysfunction such as early relaxation and left atrial (LA) enlargement were significantly attenuated when compared with wild-type mice. Second, LVPO caused an induction of Wwp1 in wild-type mice and was associated with increased collagen type I mRNA levels, which was accompanied by increased LV fibrillar collagen content. However, LVPO in Wwp1^−/− mice was associated with reduced collagen type I expression, reduced LV collagen content, and reduced levels of the profibrotic signaling molecule TGF. These findings support the postulate that targeting specific ubiquitin ligases, such as WWP1, may have therapeutic potential in the context of LVPO and HFpEF.

Although the development and progression of HFpEF can arise from multiple etiologies, uniform LV structural and functional variables are the cornerstone for this form of HF (4, 6, 22–24). Specifically, LV hypertrophy with evidence of impaired filling (diastolic dysfunction), such as elevated filling pressures and LA enlargement, with a stable EF form the basis for the HFpEF definition (22–24). The present study used a murine LVPO stimulus to induce this HFpEF phenotype and identified that deletion of Wwp1 was sufficient to attenuate the magnitude of HFpEF progression. Specifically, an index of the rate of LV filling (E/A Doppler derived) ratio and LA enlargement were both reduced in Wwp1^−/− mice. The reduction in LA dilation is particularly relevant as this is a key factor in HFpEF prognosis (23). Thus, demonstrating that the deletion of Wwp1 did not significantly impair the adaptive LV myocardial growth response to a LVPO stimulus but did reduce these key functional indices associated with HFpEF progression is an important and translationally important finding. Although LVPO caused a robust increase in LV mass in the Wwp1-null mice, this degree of LV hypertrophy was lower than respective wild-type mice. How these differences in relative LV mass may have independently affected LV filling rates and LA dilation remain unclear. Nevertheless, the trajectory of LA dilation, a key index of HFpEF progression, was reduced in Wwp1^−/− mice with a prolonged LVPO stimulus. In the present study, steady-state mRNA levels for Wwp1 increased with LVPO in the wild-type mice, which was associated with a HFpEF phenotype. Although remaining associative, these findings implicate WWP1 as a contributory factor for the development and progression of HFpEF and potential marker of a tipping point between adaptive and maladaptive response to LVPO.

The ubiquitin ligase WWP1 has been associated with atrial fibrosis and atrial fibrillation and appeared to specifically cause an induction of TGF and the fibrillar collagens using ex vivo fibroblast assays (25). Our past studies have identified robust expression of WWP1 in isolated cardiac fibroblast cultures (17). Thus, WWP1 may contribute to increased synthesis and stability of ECM proteins through modifying fibroblast phenotype, potentially via regulation of key transcriptional regulators like KLF15, which has been shown to inhibit the transcription of connective tissue growth factor in cardiac fibroblasts (26). Past studies have identified that ubiquitin ligases can modify isolated fibroblast function, such as cellular senescence, by regulating the stability of transcription factors like p53, p21, and p27 (2). Whether and to what degree specific cell types (i.e., fibroblast, cardiac myocyte, etc.) are differentially affected by changes in WWP1 remain to be fully explored and will require future cell culture studies. One of the key bioactive signaling pathways in terms of stimulating fibroblast growth, proliferation, and ECM synthesis is TGF (27). In the present study, steady-state mRNA levels of TGF (β1, β2) robustly increased with LVPO and were paralleled by increased expression levels for the fibrillar collagens, consistent with activation of profibrotic pathways. Since fibrillar myocardial collagen accumulation contributes to increased LV stiffness properties and in turn diastolic dysfunction, then the relative reduction in TGF and the fibrillar collagen expression with Wwp1 deletion would hold physiological relevance. Specifically, reduced LV compliance (i.e., increased stiffness) would be reflected as shifts in diastolic filling dynamics and LA geometry, both quantified in the present study. Past studies identified that a similar ubiquitin ligase, WWP2, played a regulatory role in myocardial ECM accumulation and responsiveness to TGF-related signaling (14, 28). To examine this more carefully, the present study examined mRNA levels of Smads 1–7, the canonical intracellular signaling molecules for TGF (27). The results from this set of measurements were associated with a high degree of variation, which is likely due to the inherent limitations of mRNA measurements as well as the relative turnover of Smads. LVPO caused a relative reduction in Smad-4 and -5, which are part of a cluster of Smads formed secondary to TGF receptor activation (27, 29). Whether this is due to a feedback loop secondary to chronic TGF activation remains to be explored. One interesting observation from the Smad mRNA profiling was the directional changes in Smad-7 between the wild-type and Wwp1^−/−-null mice. It has been demonstrated that Smad-7 interferes in the TGF intracellular signaling cascade and is thus considered an inhibitory Smad (30, 31). With LVPO in wild-type mice, Smad-7 mRNA levels decreased, which would imply enhanced TGF intracellular signaling and the resultant profibrotic response. In the Wwp1^−/−-null mice with LVPO, Smad-7 increased, which would in turn inhibit TGF intracellular signaling and thus attenuate the downstream profibrotic response. Although it remains unclear from the present study whether alterations in WWP1 directly or indirectly affect TGF and Smad mRNA levels, the fact that Wwp1 deletion reduced TGF and increased Smad-7 expression likely contributed to the relative reduction in LV fibrillar collagen expression and content, and ultimately the attenuation of the HFpEF phenotype.

Although the MMPs were canonically considered to be responsible for ECM turnover and fibrillar collagen degradation, the action of MMPs is much more complex (32). For example, MMP-14 can cause a proteolytic-mediated release of TGF through degradation of the TGF latency binding protein (18, 19, 33). This would imply that increased levels of MMP-14 would not only contribute to destruction of the normal ECM architecture, but actually promote a profibrotic cascade. In the present study, LVPO caused an induction of MMP-14 in both strains of mice, but was reduced in the Wwp1-null mice. With LVPO, an increase in MMP-2 and TIMP-1 occurred, and this increase has been reported previously in patients with progressive HFpEF (34). In the Wwp1-null mice, MMP-2 and TIMP-1 levels were reduced with LVPO. Taken together, these shifts in specific MMP/TIMP levels would in turn potentially alter myocardial ECM synthesis/degradation pathways. However, it must be recognized that these measurements were taken at the transcriptional level and the structure and composition of the ECM is influenced by a number of post-transcriptional/translational events. Nevertheless, LV collagen content, as assessed by histochemical staining and SHG imaging, was increased with LVPO as expected, but was reduced in the Wwp1-null mice.

The mRNA levels for several transcription factors/adhesion molecules associated with growth regulation (Myc, Snail1, VCAM1) were increased with LVPO and reduced in the Wwp1-null mice. These shifts in determinants of myocardial growth may have contributed to the lower LV mass that occurred in the Wwp1-null mice. Whether deletion of Wwp1 directly or indirectly altered myocardial growth and adaptation to a LVPO stimulus remains to be established. In the Wwp1-null mice, while remaining within the normal range (>50%), was lower at baseline than in wild-type mice. The slightly lower EF in the Wwp1-null mice may be reflective of differences in intrinsic adaptation, such as myocardial growth. In the present study, LV mass was lower in the female LVPO Wwp1-null mice compared with LVPO Wwp1-null male mice. These findings suggest that a sex-dependent interaction with Wwp1 occurred in terms of LV myocardial growth and an LVPO stimulus. However, while the study design provided for a balance between male and female mice with LVPO, the statistical power of the present study prevented a meaningful examination of this potential interaction, such as a multiway analysis of variance. Past studies have identified sex-dependent responses in mice subsequent to LVPO, particularly with respect to profibrotic pathways (35, 36). These past findings and those of the present study notwithstanding, LV functional in particular diastolic function moved in an equivalent direction with LVPO in both male and female mice and thus clearly identified a potential role for WWP1 in terms of the progression to HFpEF.

Some of the limitations of the current study have been outlined in the previous paragraphs and in addition to these are the inherent limitations associated with a global gene deletion transgenic approach and the acute nature of the LVPO stimulus. Since WWP1 can regulate specific cell growth/viability pathways (11–14, 16), then the absence of WWP1 may have altered the initial myocardial adaptation to the acute induction of the LVPO stimulus. In terms of the LVPO stimulus and this murine model, caution must be exercised in terms of extrapolating these findings to the complex clinical presentation of HFpEF. These limitations notwithstanding, the unique findings have identified that deletion of Wwp1 modifies key physiological variables associated with HFpEF progression and reduced key determinants of myocardial collagen accumulation, i.e., fibrosis. Thus, strategies that can modify WWP1 expression and/or activity hold therapeutic relevance in the context of LVPO and HFpEF progression, which has, to this point, proven particularly difficult to treat.

SUPPLEMENTAL DATA

Supplemental Table S1: https://doi.org/10.5281/zenodo.5557538.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

L.E.M. and F.G.S. conceived and designed research; L.B.S., Y.L., H.D., L.A.F., V.K.L., A.M., and K.N.Z. performed experiments; L.B.S., H.D., L.A.F., V.K.L., A.M., K.N.Z., L.E.M., and F.G.S. analyzed data; L.B.S., H.D., L.A.F., A.M., L.E.M., and F.G.S. interpreted results of experiments; L.A.F., L.E.M., and F.G.S. prepared figures; L.E.M. and F.G.S. drafted manuscript; L.B.S., L.E.M., and F.G.S. edited and revised manuscript; L.B.S., H.D., L.A.F., V.K.L., A.M., K.N.Z., L.E.M., and F.G.S. approved final version of manuscript.