Among all patients, aged between 20 and 80 years old, with heart failure (HF) underwent guideline-directed medical therapy (GDMT) in a tertiary care hospital, 40 of them with stable clinical status greater than 4 weeks were sequentially enrolled in the study. Thery were further allocated into the high-intensity interval training (HIIT) and GDMT groups. The 20 HIIT participants underwent additional 36 sessions of exercise training with alternating 80% of peak oxygen consumption (VO2peak) and 40% of VO2peak for 30 mins, whereas the remaining 20 GDMT participants received regular medication treatement. Baseline clinical information was recorded. Cardiopulmonary exercise test (CPET), 2-D echocardiography, bioimpedance analysis, questionnaire for quality of life, serum b-type natriuretic peptide (BNP), circulating microRNA (miRNA) profile and serum metabolomics before intervention were assessed. Cardiomyocytes and/or cardiac fibroblasts were harvested in 10% patient serum for cell activities, western blot, and messenger RNA studies. All the above examinations were followed after completing the exercise training in HIIT participants and 3-4 months after initial visit in GDMT participants. Parametric anslysis was used to assess the differences of continuous parameters with normal distributions. Those without normal distributions were estimated with non-parametric analysis. 2x2 ANOVA was conducted to compare differences of the circulating miRNA between the two groups of participants before and after interventions. Non-continous parameters between the two groups were assessed by chi-square test.

Design: The Institutional Review Board of a tertiary care hospital approved the study (IRB No.: 201902170A3). Patients with heart failure (HF) were sequentially enrolled from August 1, 2019 to January 31, 2022. All enrolled participants received guideline directed medical therapy (GDMT) and provided informed consent after understanding the experimental procedures. They were then allocated to undergo additional 30 min of HIIT or regular medical treatment (GDMT) groups using a computer-generated, concealed allocation schedule. The HIIT participants underwent additional 36 sessions of HIIT in the hospital.

Participants: Patients with HF had stable clinical presentations for greater than 4 weeks and received individualized patient education under optimized GDMT by our HF care team. The HF care team provided individualized education related to HF and methods for self-monitoring, optimized guideline-based HF medication, and further laboratory assessments (Lee et al. Int Heart J. 2012;53:364-369). In the primarily included patients, we excluded individuals aged <20 years and >80 years, or with pregnancy, cardiac transplantation within the next 6 months, moderate to severe chronic obstructive pulmonary disease, decompensated HF, non-cardiac disease prohibiting cycling exercise, or absolute contraindications for exercise suggested by the American College of Sports Medicine (Pescatello et al. ACSM's guidelines for exercise testing and prescription. 9th ed. Philadelphia, PA.: Wolters Kluwer/Lippincott Williams & Wilkins; 2014). HF patients with estimated glomerular filtration rate of < 30 mL/min/1.73 m^2 were also excluded because most of them were under hemodialysis and could not have regular exercise training.

Clinical Assessment: We recorded baseline age, sex, body mass index, co-morbidities, incremental cardiopulmonary exercise test (CPET) findings, 2-D echocardiography measurements and bioimpedance analysis for body compositions. 25 mL of blood was sampled for serum b-type natriuretic peptide (BNP) levels in all participants. After completing the above study, the remaining blood sample was centrifuged at 2500 rpm for 5 min at room temperature for serum preparation. Circulating microRNA (miRNA) and metabolites profiles were also analyzed using the sampled serum. Thereafter, we used the prepared serum to harvest the cardiomyocyte and/or cardiac fibroblast to observe their cell activities.

Exercise Training: Among the included patients with HF, those who underwent additional 36 sessions (2-3 sessions weekly) of a supervised bicycle ergometer training as in the previous protocol (Fu et al. Int J Cardiol. 2013;167:41-50) were classified as HIIT participants. They exercised alternatively at 3-min intervals of 80% VO2peak and 3-min intervals of 40% VO2peak for 30-min in each session. The other participants without supervised HIIT were classified as GDMT participants.

Echocardiography: Baseline 2-D echocardiography images of all participants were acquired at end-expiration using a 2-5 MHz tightly curved-array ultrasound transducer to measure the left ventricular end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), LV ejection fraction (LVEF), LV mass (LVM), LVM index (LVMI), and early diastolic mitral inflow velocity/early diastolic mitral annular velocity (E/e') for all participants (Lang et al. J Am Soc Echocardiogr. 2005;18:1440-1463).

Cardiopulmonary Exercise Test: All participants underwent incremental CPET within 1 week before HIIT (Fu et al. Int J Cardiol. 2013;167:41-50). Minute ventilation (VE), CO2 ventilation (VCO2), and oxygen consumption (VO2) were measured breath-by-breath using a computer-based system. VO2peak, VE-VCO2 slope, oxygen uptake efficiency slope (OUES) and heart rate reserve (HRR) were defined as described in a previous protocol (Fu et al. Int J Cardiol. 2013;167:41-50; Fu et al. J Am Heart Assoc. 2025; 14: e040330) and the American College of Sports Medicine (ACSM) guidelines (Pescatello et al. ACSM's guidelines for exercise testing and prescription. 9th ed. Philadelphia, PA.: Wolters Kluwer/Lippincott Williams & Wilkins; 2014).

Follow-Up: HIIT participants were followed up within one week after completing 36 sessions of HIIT and GDMT participants were followed 3-4 months after the first visit. All participants underwent secondary CPET, QoL assessment, 2-D echocardiography, BNP, circulating microRNA profile, circulating metabolite profile and serum to process cell culture for cardiomyocytes and/or cardiac fibroblasts.

Circulating miRNA levels: The 5 microL total RNA, extracted from 200 microL serum, was used as input material for the small RNA sample preparations. Sequencing libraries were generated using miRNA Library Kit following the manufacturer's recommendations. Briefly, 3' and 5' adaptors were directly and specifically ligated to 3' and 5' end of small RNA, respectively. Then first-strand cDNA was synthesized using QIAseq miRNA NGS RT Enzyme and RT primer. After PCR amplification, the library was size-selected with 170 - 200 bp by QIAseq beads. The quality of purified libraries was assessed on the Qsep400 system. The qualified libraries were then sequenced on Illumina NovaSeq6000 platform with trimmed 75 bp single-end reads.

Messenger RNA RT-qPCR assay Messenger RNA (mRNA) reverse transcription (RT) kit, consisted of mRNA RT-qPCR primer/probe set, and 2×Probe qPCR mix. mRNA RT-qPCR assay enables the rapid, sensitive, and specific detection of mRNA amount for target miRNA in the study. A single cDNA preparation can be used with the mRNA RT-qPCR assay to rapidly quantify the mRNA expression of mature miRNAs. The mRNA RT-qPCR primer/probe set consists of specific primers and probes designed to detect and quantify mature miRNAs. This assay system uses a stemlooped primer to extend the cDNA product in RT and a probe-based quantitative polymerase chain reaction (qPCR) assay to increase messenger RNA (mRNA) detection accuracy. When used for mRNA analysis, the assay can discriminate mature miRNA sequences from their precursors and provides an efficient and accurate method for mRNA quantification.

Serum metaboite profile: The sample preparation procedure is summarized as follows. Briefly, 200 μL of acetonitrile (ACN) was added to 50 μL of serum. The mixture was vortexed for 30 seconds and subjected to ultrasonication for 15 minutes, followed by centrifugation at 10,000 ×g for 25 minutes at 4°C. The supernatant was collected and transferred to another glass tube. The precipitate was subsequently re-extracted with 200 μL of 50% methanol (MeOH), and the above procedures were repeated. The MeOH and ACN supernatants obtained from the two extraction steps were combined and evaporated to dryness under nitrogen gas. The dried samples were then stored at -80°C until subsequent mass spectrometry analysis. For metabolomic analysis, the dried samples were reconstituted in 100 μL of 5% ACN solution and centrifuged at 14,000 ×g for 5 minutes. The clear supernatant was collected for LC-MS analysis. Chromatographic separation was performed using a Waters ACQUITY UPLC System equipped with an Acquity C18 column (100 mm × 2.1 mm, 1.7 μm particle size). The column temperature was maintained at 45°C, and the flow rate was set at 0.5 mL/min. Linear gradient elution was performed using solvent A (0.1% formic acid in H2O) and solvent B (0.1% formic acid in ACN). The gradient conditions were as follows: 0-2.5 minutes, solvent B increased from 1% to 48%; 2.5-3 minutes, solvent B increased from 48% to 98%; 3-4.2 minutes, solvent B was maintained at 98%; and 4.3-6 minutes, solvent B was returned to 1% for column re-equilibration. Mass spectrometry analysis was conducted using a Waters SYNAPT G1 high-resolution time-of-flight mass spectrometer operated in positive electrospray ionization (ESI+) mode. Instrument parameters were as follows: desolvation gas flow, 700 L/h; desolvation temperature, 300°C; cone gas flow, 25 L/h; source temperature, 80°C; capillary voltage, 3,000 V; cone voltage, 35 V; and MCP detector voltage, 1,650 V. Mass spectrometry data were acquired in centroid mode with an acquisition rate of 0.1 s per scan, an interscan delay of 0.02 s, and a mass scan range of 20-990 m/z. Mass spectrometry data were processed using MassLynx V4.1 and MarkerLynx software. The intensity of each ion was normalized to the total ion count (TIC). Chromatographic retention time, mass-to-charge ratio (m/z value), and normalized peak area were retained for subsequent qualitative and quantitative metabolite analyses.

Cell proliferation assay: Prepared live cardiomyocytes and/or human cardiomyocyte and/or cardiac fibroblast cells lines for 5-10 passages were stained with Hoechst 33342 and were then separately treated with 10% FBS, 10% participant serum before HIIT, or 10% participant serum after HIIT. Cell numbers at 0, 24, and 48 h after harvesting with the three different culture media were estimated. The relative cell count (RCC) was calculated as the cell number measured at each time point divided by that at 0 h.

Cell migration assay: We used 10% patient serum before and after interventions, replacing 10% foetal bovine serum (FBS), to treat cardiac fibroblasts for 5-10 passages to observe serum effects on cell behaviours. The cardiac fibroblasts in different media were prepared. The migration speed was estimated from serial images according to the persistent random walk equation.

Protein analysis: Cardiomyocytes and/or cardiac fibroblasts of 5×10^5 were cultured in medium containing 1.5% FBS overnight for starvation. Then, cells were seeded in 10-cm culture dishes with 10 % FBS for 48 h. The prepared samples from both the GDMT (n=13) and HIIT (n=14) participants (before and after interventions) were washed with ice-cold RIPA cell lysis buffer containing protease inhibitor and were then scraped off each dish. Cell lysates were centrifuged at 12,000 rpm for 40 mins at 4 °C for three times and the supernatant was used for protein quantification. Protein samples of 20 microg in each lane were separated on a 7% SDS poly acrylamide gel for rabbit polyclonal anti-mTOR, mouse monoclonal anti-beta-catenin, rabbit polyclonal anti-CX43, rabbit monoclonal anti-N-cadherin, rabbit polyclonal anti-actin, rabbit polyclonal anti-TGF1beta and rabbit polyclonal anti-GAPDH as internal reference were transferred to polyvinylidene difluoride membranes. The horseradish peroxidase-conjugated secondary antibodies rabbit anti-mouse IgG or goat anti-rabbit IgG were then added for at least 1 h. The immunoreactive protein bands were visualized using enhanced chemiluminescence.

Immunofluorescent stain: Cells of 6×10^4 in 400 mL ordinary culture medium were inoculated in each well of 4-chamber coverglass (Lab-Tek, Nalge Nunc International, Naperville, IL, USA). Six pieces of coverglass were further incubated overnight at 37 ºC. They were separately incubated in 10% participant serum from GDMT and HIIT participants before (n=14) and after (n=14) interventions for 24 h. Prepared cells were loaded with 100 nM Vivid MitoTracker® probes for 30 mins and 1 μg/ml Hoechst 33342 for 15 mins at 37 °C. Then, live cells were imaged using Leica TCS SP8 Confocal microscope with the magnification of 40X and 630X. Mitochondria intensity was calculated from 6 low-power fields (40X) in each chamber and mean fluorescent intensity was estimated from respective 72 fields of HCFs incubated in pre- and post-HIIT serum. Similar cell numbers were loaded in each chamber of 4-chamber coverglass. Then, they were fixed with 4% paraformaldehyde for 5 mins and were permeabilized using 0.3% (wt/vol) Triton X-100 solution immediately after the cell attachment. Cells in coverglass were further incubated with serum before (n=14) and after (n=14) HIIT for 24 h at 37 ºC. These prepared cells were treated as described above and were stained with primary mouse monoclonal anti-actin and rabbit polyclonal anti-actin related protein 2 (Arp2) antibodies at 4 ºC overnight. They were then incubated with the secondary Alexa Fluor®-488 conjugated goat anti-mouse IgG and Alexa Fluor®-594 conjugated goat anti-rabbit IgG for 40 mins at room temperature to visualize actin filaments, N-cadherin and CX43 proteins, respectively.

Bioinformatics analysis: The differences in proteomic and DNA methylation profiling before and after HIIT were estimated by ingenuity pathway analysis, and signal transduction was analysed by the KEGG pathway database

Statistical analysis: The parametric test was used to assess the continuous parameters with normal districtions, whereas nonparametric test was used for parameters for values without normal distributions. Sutdent's t-test and paired t-test were conducted to estimate between and within groups differences of values with normal distributions, respectively. Those which were not normally distributed were estimated by the Mann-Whitney U test and Wilcoxon signed rank test, respectively. The 2x2 ANOVA was used to compared difference of microRNA profiles before and after interventions in the 2 groups. Non-continous parameters between the two groups were assessed by chi-square test.

Inclusion Criteria:

• Patients with heart failure, diagnosed according to the Framingham HF diagnostic criteria (McKee et al. N Engl J Med 1971, 285:1441-1446), who had stable clinical presentations greater than 4 weeks and received individualized patient education under optimized guideline-based management (Mao et al. J Cardiovasc Med (Hagerstown) 2015, 16:616-624), were initially surveyed.

Exclusion Criteria:

• Individuals aged <20 years and >80 years
• Pregnancy
• Cardiac transplantation within the next 6 months
• Moderate to severe chronic obstructive pulmonary disease
• Decompensated HF
• Non-cardiac disease prohibiting cycling exercise
• Estimated glomerular filtration rate of < 30 mL/min/1.73 m^2
• Absolute contraindications for exercise suggested by the American College of Sports Medicine (Pescatello et al. ACSM's guidelines for exercise testing and prescription. 9th ed. Philadelphia, PA.: Wolters Kluwer/Lippincott Williams & Wilkins; 2014).

The care provider did not know the classification of the included participants. The GDMT participants did not know the treatment plan of the HIIT participants and vice versa.