We procured CBS^+/− mice from The Jackson Laboratory and bred these mice in the animal facility of the University of Nebraska Medical Center to obtain CBS^+/+ (WT) mice. All animal studies were performed following the guidelines of the National Institutes of Health and the protocol approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Nebraska Medical Center. The experimental design for the mice is shown in Figure 1A.

SG1002 is a proprietary slow release and orally active H₂S donor (Sulfagenix, Inc). We started 12-week-old male CBS^+/− mice and their WT littermates on an SG1002 or control diet (307 mg of SG1002/kg of feed; Research Diets, Inc.; New Brunswick, NJ). The SG1002 diet achieved a dose of 40 mg/kg/d and was chosen based on previous clinical and preclinical studies of SG1002, which demonstrated stable increase in H₂S from this dose (Kondo et al., 2013; Polhemus et al., 2015). Both diets contained a standard 15% kcal from fat. All groups were supplied with food and water ad libitum. Mice were selected randomly for the different treatment groups.

For genotyping, genomic DNA was extracted from ear punch tissue and was amplified using the protocol provided by The Jackson Laboratory. The forward primer sequence for CBS^+/+ and CBS^+/− was 5′GATTGCTTGCCTCCCTACTG3′. The reverse primer sequence for CBS^+/+ was 5′AGCCAACTTAGCCCTTACCC3′, and for CBS^+/− was 5′CGTGCAATCCATCTTGTTCA3′. cDNA was amplified in a BioRad cycler, and the PCR product was separated on a 2% agarose gel. WT mice had one band at 308 bp whereas CBS^+/− were characterized by two bands at 308 and 450 bp (Figure 1B).

We measured hemodynamic changes in the heart by pressure-volume (PV) loop study following the best practice guidelines for invasive hemodynamic measurement in mice described by Lindsey et al. (2018). Mice were first anesthetized with 1–2% isoflurane, intubated and connected to a ventilator (MidiVent, Harvard Apparatus, Cambridge, MA). After opening the chest, a hole was made in the apex area of the left ventricle (LV) using a 27-gauge needle to insert a 4mm 1.2F catheter probe (Transonic, Ithaca, NY). The probe was advanced slowly into the LV so that it was away from the ventricle and septal walls as observed by the phase measurements of the surrounding muscle derived from the catheter. We maintained body temperature at 37°C by a temperature controller (ATC2000, World Precision Instruments, Sarasota, FL). Because heart rate can affect hemodynamics, we kept the heart rate between 450 and 500 beats per minute during PV recording by adjusting anesthesia. Volume measurements were calibrated automatically by the probe using admittance by removing parallel conductance from overall conductance and converted to an actual volume using the Weis equation accounting for body weight (Wei et al. 2005). The calculated stroke volume was verified in one WT animal by echocardiography. Hemodynamic parameters of diastolic and systolic function were determined by averaging at least 20 cardiac cycles from a baseline scan after proper probe placement was confirmed using the LabChart Pro PV Loop Analysis Add-On.

After recording was completed, mice were euthanized, organs were harvested, and cryopreserved at −80°C for later analyses. Approximately 400 μl of blood was collected from the inferior vena cava and after clot formation centrifuged 10 min at 2000× g, and the resulting serum was stored at −80°C in a microcentrifuge tube.

H₂S levels were measured in serum using gas chromatography, as previously described (Predmore et al., 2012). 0.2 ml blood, previously frozen at −80°C, was sealed with 1 M sodium citrate buffer (pH 6.0) and incubated at 37°C for 10 min with shaking at 125 rpm to release H₂S gas. Then, 0.1 ml of the headspace gas was injected into a gas chromatograph (7890A GC System, Agilent, Santa Clara, CA) equipped with a dual plasma controller and chemiluminescence sulfur detector (355, Agilent) for H₂S quantification against a standard curve. The samples were blinded for this measurement.

Cysteine concentrations in serum were measured by a fluorometric enzymatic assay (211099, Abcam) following the manufacturer’s protocol. A standard curve with five standards from 0 to 10 nM of cysteine was prepared in the manufacturer’s assay buffer. About 10 μl of neat serum was added to each well and incubated with enzyme mix and detection reagent. Homocysteine concentrations in serum were similarly measured by a fluorometric enzymatic assay (228559, Abcam). About 10 μl of neat serum was added to each well and incubated with enzyme mix and detection reagent. Each serum sample for the homocysteine assay was also run without enzyme mix in separate wells to measure background fluorescence.

The assay plate was read using a fluorescence plate reader (Glomax Multi+, Promega, Madison, WI) at an excitation wavelength of 365 nm and emission at 450 nm (cysteine assay) or at an excitation wavelength of 625 nm and emission at 708 nm (homocysteine assay). Readings were taken in endpoint mode 30 min after addition of the detection reagent and the fluorescence was used to calculate sample concentration against the standard curve plotted using a linear regression. Standards and samples were all run in duplicate and calculated concentrations were averaged from each well to determine sample concentration. Assay performance was verified by running quality control samples alongside samples which were within 10% of the nominal concentration.

RNA was isolated from the left ventricle using the mirVana isolation kit (AM1560, Thermo Fisher) which isolates both total and small RNA. RNA was quantified and verified for quality (260/280 > 1.8) using a NanoDrop One (Thermo Fisher). We used 1 μg of total RNA, and iScript cDNA synthesis (Bio-Rad) to synthesize cDNA, which was then amplified using SYBR green in a CFX qPCR instrument (Bio-Rad). We used 18S rRNA as an endogenous control. The primer sequences were: 18S (NR_003278), forward 5′GTAGTTCCGACCATAAACGA3′ and reverse 5′TCAATCTGTCAATCCTGTCC3′, and CBS (NM_144855) forward 5′TGCGGAACTACATGTCCAAG3′ and reverse- 5′TTGCAGACTTCGTCTGATGG3′. The melting temperature for all reactions was 55°C. Results were analyzed by CFX Manager 3.0 software (Bio-Rad).

Protein was extracted from frozen portions of the LV by using RIPA buffer (BP-115D, Boston BioProducts, Ashland, MA) and a 2 ml glass homogenizer. Protein lysates were quantified using a BCA protein assay kit (23227, Thermo Fisher). About 30 μg of proteins loaded onto 10 or 12% SDS PAGE gels.

All membranes were blocked using 5% nonfat dry milk in Tris-buffered saline (TBS) at room temperature for 30 min. The primary antibodies used were: β-MHC (1:1,000, ab172967, Abcam), CSE (1:1500, H00001491-M02, Abnova), c-Fos (1:1,000, 2250 s, Cell Signaling), total c-Jun (1:1,000, 9165 s, Cell Signaling), MMP9 (1:1,000, ab119906, Abcam), and TGF-β1 (1:100, sc146, Santa Cruz) for 1 h. Secondary antibody staining was performed using: anti-mouse IgG-HRP, and anti-rabbit IgG-HRP (used at half the concentration of the respective primary antibody, Cell Signaling). Restore Stripping Buffer (46430, Thermo Fisher) was used as necessary. Protein expression was normalized against total protein measured by Ponceau staining. Bands were visualized with a ChemiDoc Imaging System (Bio-Rad), and band intensity was analyzed by ImageLab software (Bio-Rad).

Tissue sections of mouse hearts were made for visualizing cell size and collagen fibers. We made 5 µm transverse sections of hearts embedded in paraffin. Sections were fixed in 10% formalin. Staining was performed by the UNMC Tissue Sciences Facility using Sirius red for collagen staining and H&E staining following standard protocols. We also performed wheat germ agglutinin (WGA) staining on 5 μm cryosections of the heart. Sections were fixed in 4% paraformaldehyde and incubated in 5 μg/ml of WGA (W834, Thermo Fisher) for 10 min at room temperature.

Sections were imaged using a VWR bright field microscope and Motic Images Plus 2 imaging tool (Motic, Richmond, British Columbia). WGA sections were observed under a fluorescence microscope (EVOS, Life Technologies). We calculated perivascular and interstitial (INT) fibrosis by quantifying percentage red pixels/total pixels in a uniform area using Image J software. Exposure settings for imaging and thresholds for red pixel detection were kept consistent for all images. Two random areas of the LV were chosen and the same areas were imaged in all hearts. Fibrosis percentage was averaged across the two areas to determine interstitial fibrosis for each animal. For perivascular fibrosis, a similar diameter vessel was identified in all sections and two random areas around the vessel were quantified for percentage fibrosis and averaged. Hypertrophy was quantified by manually counting the number of cells in a uniform area of the image. Cells were counted in two randomly chosen areas of LV section, which were kept consistent in all hearts, and averaged for each animal.

MMP activity was visualized in tissue sections by in situ zymography. 5 μm transverse sections of the heart were made in a cryostat. Sections were incubated with fluorescein-conjugated DQ gelatin (D12054, Thermo Fisher), which yields highly fluorescent peptides when digested by MMPs, for 1.5 h at room temperature in the dark. After washing in phosphate buffered saline, sections were fixed with 4% paraformaldehyde, stained with DAPI for 5 min and mounted with a coverslip. Some sections were treated with 1 mM 1,10-phenanthroline monohydrate, an inhibitor of all MMPs, as a negative control.

All statistical analyses were performed using GraphPad Prism 8 software. One-way ANOVA and Tukey’s multiple comparisons were used to determine the statistical variance between the four groups. Unpaired student t-test was used for statistical analysis between two groups. p < 0.05 was considered statistically significant and individual p values are shown for all graphs.

We confirmed CBS^+/− mice by genotyping shortly after birth and by reduced gene expression of CBS measured by qPCR of left ventricle tissue at the end of the experiment (Figure 1). Serum concentrations of homocysteine in CBS^+/− mice were approximately twice the concentrations seen in WT mice confirming the presence of mild HHcy in CBS^+/− mice (Figure 1D). At 12 weeks of age, WT control and CBS^+/− mice were given a normal chow control diet or a diet with orally active SG1002. The diet continued for 14–16 weeks (Figure 1A) at which time the chest was opened to measure cardiac function, organs and serum were removed, and wet organ weight was measured (Table 1). Heart weight was significantly increased in CBS^+/− mice which were normalized in CBS^+/− mice that were given the SG1002 diet. Body weight and weight of other organs were unchanged between groups.

To determine if HHcy in CBS^+/− mice and SG1002 H₂S donor diet induces changes in the transsulfuration pathway, we measured serum concentrations of cysteine and H₂S and expression of CSE in the heart (Figure 2). Cysteine, which is produced by CBS from homocysteine, showed slightly, but not statistically significant, decreased concentrations in CBS^+/− mice compared to WT mice. H₂S concentrations were also measured in serum at the end of the experiment. H₂S concentrations were not reduced in CBS^+/− mice; however, the SG1002 diet in these mice did significantly increase H₂S concentrations (Figure 2B). Protein expression of the H₂S producing enzyme CSE was significantly increased in the left ventricle of CBS^+/− mice.

To investigate whether CBS^+/− mice develop cardiac fibrosis, we measured early markers of remodeling and histological and molecular markers of fibrosis in the left ventricle. C-Fos and c-Jun protein expression in the left ventricle were significantly elevated in CBS^+/− mice compared to WT mice (Figure 3). SG1002 administration did not significantly reduce this expression although there was a trend for decreased c-Jun expression compared to CBS^+/− mice.

We found increased collagen deposition in the left ventricle of CBS^+/− mice by picosirius red, which stains all types of collagen indicating increased cardiac fibrosis (Figure 4A). This increase was limited to the interstitial myocardium and not the perivascular area (Figure 4B).

In situ zymography, an assay for matrix metalloprotease (MMP) activity, was qualitatively increased in the left ventricle of CBS^+/− mice compared to WT activity (Figure 4C). Overlay of phase contrast images with DQ gelatin stain from sections demonstrated that DQ staining occurring mostly in the extracellular space and application of the MMP inhibitor 1,10-phenanthroline abolished all DQ signal confirming the specificity of the DQ staining for MMP collagenase/gelatinase fibrotic activity.

Finally, molecular markers and pro-fibrotic mediators were measured. TGF-β protein expression was increased in the CBS^+/− left ventricle (Figure 4D). MMP9 expression (Figure 4E), however, was not increased indicating increased collagen deposition is likely due to an increase in MMP activity rather than expression.

H₂S donor administration by diet mitigated the histological and molecular signs of cardiac fibrosis. Interstitial fibrosis normalized to WT levels in CBS^+/− mice. MMP activity also decreased in CBS^+/− + SG1002 mice to levels similar to WT. Finally, TGF-β protein expression was restored to WT levels in CBS^+/− mice treated with SG1002.

We subsequently measured histological and molecular markers of cardiac hypertrophy in CBS^+/− mice with and without SG1002 diet. First, we measured cardiomyocyte size in H&E and WGA stained sections of the left ventricle (Figures 5A,B). CBS^+/− mice sections showed fewer cells per uniform area measured indicating greater cell size and cardiac hypertrophy. Similarly, heart weight to body weight ratio was increased in CBS^+/− compared to WT hearts (Figure 5C). Finally, fetal/neonatal beta-cardiac myosin heavy chain (β-MHC) is commonly re-expressed in the adult heart during cardiac hypertrophy and remodeling. CBS^+/− mice hearts showed increased protein expression of β-MHC providing further confirmation of cardiac hypertrophy in HHcy mice (Figure 5D).

H₂S donor diet normalized all signs of cardiac hypertrophy in CBS^+/− mice. Cell size was reduced in H&E stained sections, heart weight to body weight ratio was reduced, and β-MHC protein expression was reduced to levels comparable to WT hearts.

Cardiac function and hemodynamics were evaluated at the end of the experiment using PV loop recordings to determine if the cellular remodeling alterations in HHcy mice resulted in altered cardiac function (Figure 6). We evaluated measures of systolic and diastolic function as cardiac remodeling affects diastolic function first (Swynghedauw, 1999). CBS^+/− mice exhibited increased end systolic pressure compared to WT mice but no change in end-diastolic or systolic volumes. Thus, stroke volume and cardiac output remained similar to WT mice. This depicts a scenario of increased afterload but no decrease in systolic function. SG1002 restored the increased afterload in CBS^+/− mice by reducing end systolic pressure. Also, end-diastolic volume was significantly increased in CBS^+/− + SG1002 mice (Figure 6G) while end-systolic volume was unchanged. This resulted in increased stroke volume in the CBS^+/− + SG1002 group and suggested an increase in preload and ventricular filling.

Measures of diastolic function such as Tau and dp/dt min were unchanged between all groups. Load independent measures of cardiac function such as the end-diastolic PV relationship were not evaluated in this study.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.