All experiments followed ethical and legal standards according to The Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication 8^th edition, update 2011) and the Institutional Animal Care and Use Committee of the University of South Dakota. The creation and validation of the GFPdgn transgenic mice for monitoring the dynamic changes in UPS performance in vivo was previously described (Kumarapeli et al., 2005). Littermate male GFPdgn transgenic and non-transgenic mice in the FVB/N inbred background at 6–7 months of age were randomly assigned to the CHOW group or the HFD+L-NAME group throughout the study. The CHOW (control diet) group had unrestricted access to a standard murine chow diet of 19.1% protein, 6.5% fat, and 47% carbohydrates (2020X from Teklad) and water. The HFD+L-NAME group had unrestricted access to an HFD where 60% of the calories were from fat (lard and soybean oil), 20% from proteins, and 20% from carbohydrates (D12492, Research Diets, Inc) and water spiked with L-NAME (0.5 g/L; Sigma Aldrich, Catalog #: N5751-10G). Body weight of each mouse was measured weekly for 18 weeks at the same time of day. The age of mice at the initiation time of this study (6–7 months of age), an age that should be more clinically relevant to HFpEF that naturally occurs in humans compared to the much younger age (8–10 weeks of age) used in the prior report of this double-hit method (Schiattarella et al., 2019).

A bolus intraperitoneal injection of 20% glucose solution (1 g/kg in water) was conducted after 6-h fasting. Blood from the tail was collected and measured immediately before (0 min) and 15, 30, 60, 90, and 120 min after the injection using Bayer Contour^® Next EZ Glucose Meter Kit and Contour^® Next Blood Glucose Test Strips.

Mice ran on an Exer-3/6 treadmill with Stimulus Detection (Columbus Instruments, Columbus, OH) on a 10-degree incline. The day before the test, animals were subjected to an acclimation at 5 m/min for 4 min, then increased to 10 m/min for 10 min. For the final test, mice were randomly grouped in a lane and perform at 5 m/min for 4 min then at 14 m/min for 2 min. Speed was increased by 2 m/min every 2 min until exhaustion. Exhaustion is defined as the animal getting shocked continuously for 5 s. The Treadmill Software (Columbus Instruments, Columbus, OH) measured running distance, time ran, and exhaustion.

The tail-cuff method was used to measure systolic blood pressure. Animals were trained in the instrument for at least 2 months before the final test. The mice were acclimated in the temperature-controlled chamber for 15 min, and results were measured, then another 15 min were allowed for resting data. Blood pressure was then recorded for 4 days, and readings averaged.

Echocardiography was performed as previously reported in a double-blinded manner (Zhang et al., 2019). In brief, mice were kept in light anesthesia with inhalation of Isoflurane (4% for induction and 1.5% for maintenance) via a face mask. Transthoracic echocardiography was performed using the Visual Sonics Vevo 3100 system and a 40-MHz probe (FUJIFILM Visual Sonics, Toronto, ON, Canada). A two-dimensional guided M-mode episode was acquired through the left ventricular (LV) anterior and posterior walls at short axis view. Parameters of LV were derived from primary measurements using Vevo LAB software.

Ventricular myocardial tissues were homogenized in 1x loading buffer (41 mM tris-HCI, 1.2% SDS, 8% glycerol). The homogenates were boiled for 5 min, then centrifuged for 20 min at 4°C at 14,000 rpm, and the supernatant was collected. Protein concentration of the samples were measured using Pierce BCA Protein Assay Reagent (Thermofisher, Rockford, Illinois). Equal amounts of proteins were loaded to each lane of a 10% SDS-PAGE gel and fractionated by electrophoresis, then electrically transferred to a PVDF membrane overnight in 4°C. The PVDF membrane was washed and incubated in blocking buffer for 60 min before being probed with the primary antibody overnight followed by incubation with HRP-conjugated secondary antibodies. The bound secondary antibodies on the PVDF were then detected with the SuperSignal West Pico PLUS Chemiluminescent Substrates (ThermoScientific) and imaged with a ChemiDoc MP imaging system (Bio-Rad). Densitometry quantification was performed using the ImageLab software (Bio-Rad).

Total RNA was extracted from left ventricular myocardium using TRI Reagent^® (Molecular Research Center Inc., Cincinnati, OH). Thermo Scientific™ NanoDrop 2000 UV Spectrophotometer was used to determine the concentration and purity of the RNA. Reverse transcription utilized the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems™) and a total of 900 ng of extracted RNA as template. qPCR reactions used 2 μL of 1:10 nuclease-free water diluted cDNA solution and PowerUp SYBR Green master mix (Applied Biosystems) for the following gene products: Atrial Natriuretic Factor (Nppa), Brain Natriuretic Peptide (Nppb), GFPdgn, Phospholamban (Pln), and Glyceraldehyde-3-phosphate dehydrogenase (GAPDH). All reactions for target genes (Nppa, Nppb, GFPdgn, and Pln) were performed in duplicate. The 2^−ΔΔCT relative quantification method using GAPDH as the normalization gene was used to compute the relative expression of the target genes. The primer sequences for the qPCR reactions are shown in Table 1.

GraphPad Prism software 9.2.0 (GraphPad Software, Inc., La Jolla, CA) was used for all statistical analysis. Results are presented as the mean ± SEM. Datum was deemed an outlier according to the Grubbs test (p < 0.05). Two-tailed unpaired Student’s t-test was used unless indicated otherwise. In addition, the Welches correction for unequal variance was used when analyzing gene expression results. Researchers were blinded to the individual genotypes during the data collection and data analysis. A p-value or adjusted p-value of <0.05 was considered statistically significant.

HFD can induce obesity and the latter inevitably increases body weight (BW). Thus, we measured mouse BW immediately before (week 0) and weekly after initiation of the HFD+L-NAME treatment. No significant changes in BW were observed in the Chow group throughout the 18 weeks of study (Figure 1). The difference in BW between the Chow and the HFD+L-NAME groups was not statistically significant at week 0 (p > 0.9999) but became significant from week 1 to week 17 (Figure 1A). Self-comparisons show significant BW increases in the HFD+L-NAME group as early as week 1 (p = 0.0014; Figure 1B) and the increases peaked at week 2 (p = 0.0001; Figures 1A,C) and plateaued thereafter (Figures 1A–D). Longitudinal BW comparisons between week 0 and week 1 (Figure 1B, p = 0.182 and 0.001 for the Chow and the HFD+L-NAME groups, respectively), between week 0 and week 2 (Figure 1C, p = 0.0891 and 0.0001), or between week 0 and week 18 (Figure 1D, p = 0.4778 and 0.0033) are also summarized in Figure 1. The BW to tibial length (TL) ratio (BW/TL) of the HFD+L-NAME group was greater than that of the Chow group (p = 0.0171, Figure 1E) at week 18 when the terminal experiments were performed.

Obesity can cause glucose intolerance; hence, we performed glucose tolerance tests on these animals at weeks 6 and 15. A bolus intraperitoneal injection of glucose induced increases in blood glucose levels in both groups, but the peak increase was significantly greater (week 6) and the increases lasted longer (weeks 6 and 15) in the HFD+L-NAME group compared with the Chow group (Figure 2), indicating that glucose intolerance was induced as early as week 6 by HFD+L-NAME. Administration of L-NAME is known to induce hypertension (Schiattarella et al., 2019); hence, we measured blood pressure using the tail-cuff method at week 17 and found that systolic blood pressure was significantly higher in the HFD+L-NAME group than in the Chow group (p = 0.0474, Figure 3). All three components of metabolic syndrome: hypertension, obesity, and glucose intolerance, are at least common comorbidities of HFpEF (Dunlay et al., 2017), and even can be an aspect of the multifactorial comorbidity driven pathology of the HFpEF syndrome (Tourki and Halade, 2021). The increased body weight, glucose intolerance, and hypertension observed in the HFD+L-NAME group are in alignment with recapitulating the metabolic syndrome central to HFpEF. and provide a catalyst for afterload induced cardiac remodeling of the left ventricle. Together, these pathologies formed the foundation for more advanced HFpEF manifestations.

Serial echocardiography was performed to evaluate the dynamic changes in cardiac morphometry and function in live mice. Consistent with an increase in systolic blood pressure observed in the HFD+L-NAME group, left ventricular (LV) concentric hypertrophy was detected at both week 7 and week 15. This is evidenced by significantly greater LV end-diastolic posterior wall thickness (LVPW; d, Figure 4) and a tendency of deceased LV end-diastolic chamber diameter (Figure 5) in the HFD+L-NAME mice, compared with the Chow group disclosing diastolic dysfunction.

At the terminal experiment, we detected that LV myocardial mRNA levels of fetal genes Nppb (p = 0.1092) and Nppa (p = 0.1327) tended to be higher and Pln mRNA levels (p = 0.1154) lower in the HFD+L-NAME group than in the Chow group (Figure 6) although the differences did not reach a statistical significance due to a large variation within each group.

Exercise tolerance tests performed at both weeks 7 and 16 revealed that the maximum distance and time of running to exhaustion by the HFD+L-NAME mice were markedly shorter than that by the Chow control group (Figure 7). At week 7 the Chow group ran an average of 269 m (18:07 min) while the HFD+L-NAME group ran an average of 168 m (13:30 min) prior to exhaustion. Similarly, at week 16 the Chow group ran 303 m (19:51 min) while the HFD+L-NAME group ran 167 m (13:21 min). These data indicate that exercise intolerance occurred as early as 7 weeks after HFD+L-NAME treatment. Reduced exercise tolerance including dyspnea and fatigue on mild exertion are common complaints seen in HFpEF patients clinically. The exercise intolerance in the HFD+L-NAME group indicate that the mouse model began to recapitulate the clinical aspects of HFpEF as early as week 7.

Despite the numerous characteristics of heart failure in the HFD+L-NAME treated mice, their LV ejection fraction (EF) and fractional shortening (FS) were not decreased throughout the study compared with those of the Chow control group (Figure 8), indicating that the systolic function was preserved in the HFD+L-NAME treatment group, one of the essential criteria for the diagnosis of HFpEF.

The GFPdgn transgenic mice, which were created and have been maintained in the FVB/N inbred background, have been extensively used to probe changes in UPS functioning in mice (Chen et al., 2005; Liu et al., 2006; Li et al., 2011); in absence of GFPdgn protein synthesis, the steady state level of GFPdgn protein inversely reflect UPS performance in the tissue or cell examined (Bence et al., 2001; Kumarapeli et al., 2005). We observed equivalent myocardial mRNA expression (p = 0.241) but significantly elevated GFPdgn protein levels (p = 0.0497) and greater GFPdgn protein to mRNA ratios (p = 0.0379) in the HFD+L-NAME treatment group compared with the CHOW group (Figure 9), indicating that elevated GFPdgn protein is not due to increased transcription, which provides compelling evidence that myocardial UPS performance is impaired in the HFpEF mouse.

The reactivation of the fetal gene program and a downregulation of PLN are often associated with and indicative of pathological hypertrophy and heart failure (Loffredo et al., 2014) (Man et al., 2018). Hence, these changes in gene expression are consistent with the significantly increased wall thickness revealed by echocardiography in the HFD+L-NAME group and the conclusion of induced cardiac remodeling in the HFD+L-NAME group. Clinically, diastolic dysfunction in HFpEF results in reduced cardiac output with maintained EF due to limited cardiac filling but at higher left ventricular end diastolic pressures. The increased systolic and diastolic wall thickness, while not essential, are characteristic of HFpEF as traditionally defined (Shah, 2013), while the reduction in PLN mRNA is indicative of the signature diastolic dysfunction of HFpEF (Grote Beverborg et al., 2021). Further, under the same “two-hit” hypothesis, diastolic dysfunction has been well described by invasively and non-invasively measured elevated LV filling pressures, histologically reported cardiac fibrosis and capillary rarefaction, pulmonary congestion, and reduction in coronary flow reserve. Although clear exercise intolerance was observed, the “two-hit” hypothesis is known to produce exercise intolerance without histopathological, molecular, or strength abnormalities in skeletal muscle indicating that decrement in exercise intolerance is not due to a skeletal muscle deficiency (Schiattarella et al., 2019). Taken together, the hypertensive afterload stress and metabolic syndrome pathologies, likely created an environment for pathologic hypertrophic cardiac remodeling and diastolic dysfunction to occur, a hallmark of HFpEF. Despite these changes, serial echocardiographic analyses displayed preserved ejection fraction across both groups through at least week 15 (Figure 8). Therefore, the present study demonstrates compellingly that HFpEF can be produced with the “two-hit” protocol in male FVB/N mice.

HFpEF is traditionally considered a disease of the elderly resulting in challenges studying the disease in vivo in a time and cost-efficient manner. To construct a model that most quickly emulates the complete severity and mortality of HFpEF, male mice were used exclusively (Patten, 2007) (Tromp et al., 2018). In an aim of striking a balance between time efficiency and model validity, mice six to 7 months of age were used. While the maiden voyage of the two-hit hypothesis successfully recapitulated the disease using mice 8–12 weeks of age (Schiattarella et al., 2019), mice 3 months senior likely accounted for the age-related etiology of HFpEF to a greater degree.

Conventionally, the “two-hit” HFpEF mouse model uses C57BL/6N wild-type mice (Schiattarella et al., 2019), a lineage traditionally renowned as the gold standard for producing diet induced obesity and metabolic syndrome. However, recent studies comparing FVB/N and C57BL/6J have begun to question this notion. FVB/N ob/ob and C57BL/6J ob/ob mice are used to study obesity and its comorbidities as both murine lineages are leptin deficient thus maximizing hunger and the effects of the HFD. Compared to the C57BL/6N ob/ob lineage, FVB/N ob/ob mice demonstrate increased hyperglycemia, whole body and muscle insulin resistance, and reduced clearance of circulating triglycerides (Haluzik et al., 2004), suggesting that the FVB/N murine lineage exhibits significant benefits over the C57BL/6J strain in modeling metabolic dysfunction, a quintessential comorbidity of HFpEF. Likewise, Nascimento-Sales et al. (2017), found that while C57BL/6N mice develop obesity more quickly than FVB/N mice, FVB/N mice actually show a significantly greater degree of metabolic intolerance after treatment with a HFD as evidenced by higher insulin resistance, greater liver steatosis, and a large degree of epididymal white adipose tissue induced inflammation. Notably, they conclude by recommending the FVB/N lineage as a new tool to uncover the complex multifactorial symptoms of obesity and metabolic syndrome of which HFpEF is inevitably intertwined.

Between 40%–50% of HFpEF patients are obese (Shah et al., 2013a). Similarly, 45% of HFpEF patients have diabetes (McHugh et al., 2019). However, due to the decreased prevalence of diabetes relative to obesity, HFpEF is more specific towards the diabetic population. This is perhaps because HF and metabolic syndrome share numerous underlying derangements that result in increased HF risk including insulin resistance, metabolic derangements, endothelial dysfunction, oxidative stress, mitochondrial dysfunction, and autonomic neuropathy (Dhingra and Vasan, 2012; Oktay et al., 2013). Type 2 diabetes mellitus (T2DM) shares the pathologies of sodium retention, impaired skeletal muscle function, and metabolic derangements with HFpEF, implying an integrated pathogenesis of both diseases, (Shah et al., 2013b). As more insight into comorbid HFpEF and T2DM is acquired, it is anticipated physicians will be better equipped to provide effective unique treatments for commonly comorbid HFpEF and T2DM.

Further, Reduced exercise tolerance and increased myocardial remodeling are critical features of HFpEF (Paulus and Tschope, 2013; Gevaert et al., 2019). FVB/N mice demonstrate higher capabilities and cooperativity of daytime exercise tolerance testing (Gibb et al., 2016). In response to exercise tolerance training, FVB/N mice display increased cardiac remodeling compared to C57BL/6J mice (Massett and Berk, 2005; Gibb et al., 2016). A higher level of myocardium hypertrophy in response to increased exercise stress is suggestive that FVB/N mice may also develop a greater degree of cardiac remodeling due to increased afterload stress elicited by HFpEF hypertension.

Mouse models emulating HFpEF generally originate from three different avenues of the most prominent risk factors of HFpEF: hypertensive models, diabetic (obesity) models, and aging models (Horgan et al., 2014). The novel “two-hit” hypothesis developed a HFpEF mouse model from both the hypertensive and diabetic avenues. The FVB/NJ mice strain has been used to illuminate the pathologies of HFpEF through the aging model avenue (Koch et al., 2013). We report a HFpEF mouse model that potentially utilizes all three avenues of the most significant HFpEF murine models (hypertensive, metabolic dysfunction, and aging) to emulate HFpEF. Finally, applying the “two-hit” protocol to the FVB/NJ mice may not only emulate the clinical symptoms of HFpEF but could also optimize the exercise tolerance measurements, essential cardiac remodeling, and circadian misalignments of the disease (Massett and Berk, 2005; Gibb et al., 2016). A deeper understanding of the “two-hit” hypothesis and HFpEF pathophysiology may be explored by expanding the protocol across murine lineages, namely, the FVB/NJ lineage.

The UPS is the primary pathway responsible for the targeted degradation of abnormal, misfolded, damaged, or oxidized proteins, in addition to for the degradation of normal but no longer needed proteins (Wang et al., 2013). In response to the recent reports of accumulated misfolded proteins (Gonzalez-Lopez et al., 2015) and altered unfolded protein response (UPR) playing a pivotal role in the pathogenesis of HFpEF (Schiattarella et al., 2019), we hypothesized that myocardial UPS performance could be impaired during HFpEF. Beginning to examine this hypothesis, the present study applied the “two-hit” protocol to the transgenic mice with ubiquitous expression of GFPdgn, a surrogate substrate of the UPS (Kumarapeli et al., 2005). In absence of changes in GFPdgn protein synthesis, changes in its steady-state protein levels reflect inversely UPS performance (Kumarapeli et al., 2005). Notably, the GFPdgn mRNA expression between CHOW and HFD+L-NAME groups was not significantly different, but GFPdgn protein levels were elevated in the HFD+L-NAME group (Figure 9), indicating diminished UPS function in HFpEF. Such impairment is consistent with the other preliminary studies suggesting that the accumulation and mismanagement of misfolded, damaged, and abnormal proteins play a vital role in the development of HFpEF (Wang et al., 2014; Gonzalez-Lopez et al., 2015; Schiattarella et al., 2019).

Evidence of impaired UPS functioning in HFpEF beckons for further investigation to provide a deeper understanding of the pathogenesis and devise new treatment strategies for HFpEF. For example, it will be important to elucidate the main factors that impair UPS performance during HFpEF. To this end, both diabetes and hypertension have been shown to impair myocardial UPS performance and proteasome functional insufficiency (Ranek et al., 2015; Li et al., 2017), but more detailed molecular mechanisms remain to be delineated. It is also conceivable that UPS impairment could affect myocardial proteostasis and thereby cardiac function through both slowing down the turnover of normal proteins and accumulating abnormal proteins, and both are quite detrimental to the heart in many ways. Ultimately, we report successful modeling of HFpEF in male FVB/N mice and, by taking advantage of a transgenic UPS reporter mouse, we have detected myocardial UPS functioning impairment during HFpEF, suggesting a pathogenic role for impaired protein degradation in the development and progression of HFpEF but insist on further exploration (Figure 9E).

This study has potential limitations. While the power of the study was sufficient for the primary readouts, the number of mice per group is not enough to decipher whether the GFPdgn transgenic expression has an impact on the occurrence of HFpEF. No female mice were included, leaving it untested whether the “double hit” method could cause HFpEF in female FVB/N mice or not. Additionally, mice were treated with HFD+L-NAME for up to 18 weeks. Despite both Schiattarella et al. and our results providing convincing evidence of HFpEF within this time frame, the pathogenesis of HFpEF is known to develop over the course of decades in humans (Lam et al., 2011). It begs the question, was HFpEF established in its completeness within the 18-week treatment course? Would longer treatment periods elicit models with grander disease severity, more confluent outcomes, and ultimately a clearer window into HFpEF etiologies? In response, the degree of HFpEF could be graded through measurement of the E/A ratio during echocardiology. Murine blood pressures were measured using tail-cuff methodology instead of telemetry. While tail-cuff methods have been validated against telemetry for measuring more course changes in blood pressure, telemetry allows for a longer period of measurement, giving greater power to the study so that fewer animals are needed (Fraser et al., 2001). Likewise, echocardiology was used to measure ventricular wall thickness as a surrogate for cardiomyocyte hypertrophy, fibrosis, and increasing filling pressures when histological data using wheat germ agglutinin, hematoxylin and eosin staining, and pressure volume loop data would have measured these more directly. Analysis of Nppa, Nppb, Pln, and GFPdgn may have been more suited for normalization with 36B4 rather than GAPDH as GAPDH has been shown to change during obesity and metabolic syndrome (Fan et al., 2020). Despite preliminary affirmation of the role of UPS malfunction in the pathogenesis of HFpEF, more in-depth deciphering the specific defect in UPS-mediated protein degradation pathway will provide concise guidance for therapeutic interventions should the UPS impairment play an important pathogenic role. It will be extremely important to address questions such as “does HFpEF drive UPS dysfunction or does UPS dysfunction drive HFpEF?”

In summary, the present study demonstrates for the first time that impaired UPS performance can occur in an animal model of HFpEF; thus, further investigation into a potential cause-effect relationship between UPS impairment and HFpEF is warranted.