A tipping point in dihydroxyacetone exposure: mitochondrial stress and metabolic reprogramming alter survival in rat cardiomyocytes H9c2 cells

Abstract

Exogenous exposures to the triose sugar dihydroxyacetone (DHA) occur from sunless tanning products and electronic cigarette aerosol. Once inhaled or absorbed, DHA enters cells, is converted to dihydroxyacetone phosphate (DHAP), and incorporated into several metabolic pathways. Cytotoxic effects of DHA vary across the cell types depending on the metabolic needs of the cells, and differences in the generation of reactive oxygen species (ROS), cell cycle arrest, and mitochondrial dysfunction have been reported. We have shown that cytotoxic doses of DHA induced metabolic imbalances in glycolysis and oxidative phosphorylation in liver and kidney cell models. Here, we examine the dose-dependent effects of DHA on the rat cardiomyocyte cell line, H9c2. Cells begin to experience cytotoxic effects at low millimolar doses, but an increase in cell survival was observed at 2 mM DHA. We confirmed that 2 mM DHA increased cell survival compared to the low cytotoxic 1 mM dose and investigated the metabolic differences between these two low DHA doses. Exposure to 1 mM DHA showed changes in the cell's fuel utilization, mitochondrial reactive oxygen species (ROS), and transient changes in the glycolysis and mitochondrial energetics, which normalized 24 h after exposure. The 2 mM dose induced robust changes in mitochondrial flux through acetyl CoA and elevated expression of fatty acid synthase. Distinct from the 1 mM dose, the 2 mM exposure increased mitochondrial ROS and NAD(P)H levels, and sustained changes in LDHA/LDHB and acetyl CoA-associated enzymes were observed. Although the cells were exposed to low cytotoxic (1 mM) and non-cytotoxic (2 mM) acute doses of DHA, significant changes in mitochondrial metabolic pathways occurred. Further, the proliferation increase at the acute 2 mM DHA dose suggests a metabolic adaption occurred with sustained consequences in survival and proliferation. With increased exogenous exposure to DHA through e-cigarette aerosol, this work suggests cell metabolic changes induced by acute or potentially chronic exposures could impact cell function and survival.

Graphical Abstract

1. INTRODUCTION

Dihydroxyacetone (DHA) is more commonly known as the active ingredient in sunless tanning products (STPs), but it is also produced through the heating of the e-liquid in electronic cigarettes (e-cigarettes) [1-3]. DHA has also been used as a metabolic probe to monitor metabolic changes in cells and tissues [4-6]. Cytotoxic characterization of DHA has been largely focused on the skin. Initially, DHA was thought only to enter the stratum corneum, but studies have now shown that up to 30% ( ~30.5 mM, assuming a 5% w/v formulation) of the applied DHA enters the viable layers of the skin and approximately 0.5 % (~1.4 mM) enters the bloodstream [7].

Currently, the FDA has deemed DHA safe for external applications. However, the FDA has warned against inhalation of DHA through mucous membranes during spray tanning and recommended protective measures to reduce these exposures [8]. With DHA being identified in e-cigarette aerosol, the emerging exposure risks through inhalation and absorption have highlighted the current limitations in DHA research, which have focused on the genotoxic and cytotoxic effects of skin exposures [2, 9-12]. Therefore, DHA joins the growing list of chemicals found in e-cigarette aerosol in high amounts with no toxicological characterization in cells and tissues [3, 13, 14].

In e-cigarettes, DHA is produced from the main e-liquid components, propylene glycol, and glycerol [1]. Mass spectrometry and NMR studies of e-cigarette aerosol demonstrate that 40-55% of the e-liquid can be converted into DHA depending on the heating conditions and e-liquid composition [3]. For e-cigarette exposures, DHA is produced at ~ 0.16 to 4.16 μg per puff in a 100 mL puff volume, dependent on the wattage of the device [3]. E-cigarettes are highly variable in their puff volumes, with 96.81 to 133.92 ml per puff found in cigarette-like devices and 331.2 to 519.6 ml for tank devices [15, 16]. Most studies suggest a 10 puff per session duration with up to 24 smoking sessions per day for vapers [17, 18]. Using these estimates, DHA exposures range from 0.16 μg/puff * 240 puffs/day = 38.4 μg/day to ~ 1 mg/day [3, 17, 19].

Once inhaled or absorbed, cells can rapidly absorb DHA and convert it into its phosphorylated form, dihydroxyacetone phosphate (DHAP), with triose kinase/FMN cyclase (TKFC) [4, 5, 20]. The resulting DHAP enters multiple metabolic pathways, including glycolysis, gluconeogenesis, and glycerol metabolism [21]. DHAP and its isomer, glyceraldehyde-3-phosphate (GAP), are crucial metabolites, and imbalances in these triose sugars are linked with anemia, neurological disorders, diabetes, and cancer [22-24]. Imbalances in DHAP and GAP have also been noted after high fructose exposures, where increases in reactive oxygen species (ROS), generation of advanced glycation end products (AGEs), altered glycolysis, and mitochondrial dysfunction are also observed [4, 25, 26]. We previously proposed DHA exposures would mimic fructose exposures, inducing metabolic reprogramming and mitochondrial damage [25].

Given the potential of DHA to act in this manner during systemic exposures, we have expanded the characterization of DHA's cytotoxic effects by examining systemic exposure models like the epithelial embryonic kidney cell lines HEK293T and the metabolically active hepatocellular carcinoma C3A (HepG3) cell lines [27, 28]. In these model systems, we observed the induction of metabolic and mitochondrial stress, cytotoxicity at doses below 10 mM, and hybrid cell death mechanisms that included apoptosis and autophagy markers. Of particular note in both studies were changes in nutrient sensing mechanisms, suppression of glycolysis [27, 28]. Mitochondrial energetics were also altered in both cell lines with a decrease in oxidative phosphorylation found and a mitochondrial NAD⁺ levels increased in HEK293T resulting in mitochondrial-specific injury and altered mitochondrial function [27, 28].

The observed changes in NAD(P)H pools and mitochondrial reprogramming in these systemic models differ from the previous reports examining the cytotoxicity and genotoxicity in the skin [2, 9, 10]. These metabolic effects could strongly impact metabolic demanding tissues, so we have extended our studies to examine low-dose exposures of DHA in a cardiac model using rat cardiomyocytes, H9c2. E-cigarettes have toxic cardiac effects, including arterial stiffness, endothelial dysfunction, vascular injury, and oxidative stress, but the drivers of these effects are still being investigated [29, 30]. The overlap of effects observed in cardiac tissue from e-cigarette studies with those found after DHA exposure in our other systemic models suggests DHA may induce cardiac injury through metabolic and mitochondrial changes.

2. MATERIALS AND METHODS

2.1. Chemicals

Dihydroxyacetone (DHA) (CAS 96-26-4) (Cat No. PHR 1430) was purchased from Sigma-Aldrich (St. Louis, MO, USA).

2.2. Cell culture

The rat cardiomyocyte cell line, H9c2, was purchased from ATCC (CRL-1446 Manassas, VA). H9c2 cells were cultured in a 5% CO₂ incubator at 37°C and grown using Dulbecco's modified Eagle's medium (DMEM, Hyclone, Logan, UT, USA; 4.5 g/L) supplemented with 10% fetal bovine serum (FBS, Atlanta Biologicals, Flowery Branch, GA, USA), 2% Glutamax (Gibco, Carlsbad, CA, USA) and 1% sodium pyruvate (Gibco, Carlsbad, CA). The cells were tested monthly for mycoplasma contamination using the Lonza MycoAlert kit (Walkersville, MD, USA).

2.3. Cytotoxicity assays

H9c2 cells were plated at 5,000 cells per well in 6-well plates and incubated for two days in a 5% CO₂ incubator at 37°C to attach to perform a growth viability assay. Cells were dosed at increasing concentrations in triplicates (0, 1, 2, 5, 10, 14 mM) and grown for 5-7 days. After the exposure period, cells were collected by detaching with 0.25% trypsin-EDTA (Gibco, Carlsbad, CA, USA) for 15 sec and resuspended in 1x phosphate-buffered saline (PBS, VWR Life Sciences, Radnor, PA, USA). The Bio-Rad TC automated cell counter was used to count the cells (Bio-Rad, Hercules, CA, USA). The results were expressed as the percentage of cells ± standard error of the mean (SEM) of the three biological replicates. An IC₅₀ was calculated from the percentage survival relative to the control using a nonlinear regression model in GraphPad Prism.

A clonogenic assay was also used to confirm cytotoxicity. The same procedure used in the growth viability was used to seed and dose H9c2 cells. After exposure, plates were placed on ice, and the cell medium was removed. Cells were washed with ice-cold PBS once and fixed with ice-cold methanol for 15 min. Once fixation occurred, wells were stained with 1% crystal violet solution prepared in 15% methanol and incubated at room temperature (RT) ~20-22°C for 10 min. The excess solution was removed, and wells were washed twice with double-distilled wash (ddH₂O). The plates were left overnight (ON) to dry. Bio-Rad ChemiDoc XRS Imaging system was used to image the plates. Analysis was performed using the Nikon Elements software (Nikon, Nikon, Tokyo, Japan) to determine the average area fraction for each well. The final density was calculated over three technical triplicates and two biological repeats. An IC₅₀ was calculated from the average cell density compared to the control and graphed using a nonlinear regression model in GraphPad Prism.

2.4. Cell proliferation experiments

Cell proliferation changes were examined by measuring growth over 8 days to evaluate dose-dependent changes. H9c2 cells were seeded at 3,000 cells per well in 6-well plates and were left overnight (ON) to attach in a 5% CO₂ incubator at 37°C. After attachment, cells were dosed with 0.2, 1, 2, or 14 mM DHA and incubated with treatment. Every 2 days, wells to be counted were washed once with PBS, and cells were detached with 0.25% trypsin. The cell suspension was collected and counted using the Countess II FI (Thermo Fisher, Waltham, MA, USA). The process was continued for 8 days of treatment. The average growth and the SEM for three biological replicates are presented using GraphPad Prism. The proliferation rate was calculated using the following equation: Doublings per day = [log₂(final day 4 cell count /initial day 0 cell count)]/4 days [31].

2.5. Immunoblotting

H9c2 cells were seeded in 15 cm² dishes at 0.2 5x 10⁶ cells and left to attach in a 5% CO₂ incubator at 37°C for two days. After attaching, cells were dosed with either 1 or 2 mM DHA for 1, 4, and 24 h. After exposure, the cell medium was aspirated, and cells were scraped and then collected with PBS. Cells were spun down at 2,000 rpm for 5 min at 4°C, and the medium was aspirated. Cell pellets were stored at −80°C ON. The next day, cells were lysed with ice-cold lysis buffer, incubated on ice for 30 min, and spun at maximum speed for 25 min at 4°C. The supernatant was collected in a new tube to continue with protein quantification using the Bradford QuickStart protein assay kit. 4-20% Mini-Protein TGX precast gels (Bio-Rad) were loaded with 20 μg of protein, ran at 120V for 1 h, and transferred to a nitrocellulose membrane. The membrane was blocked for 1 h in 5% milk in Tris-buffered saline (TBS, VWR Life Sciences, Radnor, PA, USA) containing 0.1% Tween 20 (TBST), and primary antibodies were probed overnight in a cold room. A detailed list of the antibodies used is in Table 1.

Table 1:

Antibodies used for immunoblotting experiments

Antibodies by dilution	Source
1:5000
α-Tubulin (T9026)	Millipore Sigma, St. Louis, MO, USA
1:1000
Acetyl-CoA Carboxylase 1 (4190)	Cell Signaling, Danvers, MA, USA
Acetyl-CoA Carboxylase 2 (8578)	Cell Signaling
Aldolase A (3288)	Cell Signaling
Caspase 3 (GTX23585)	GeneTex, Irvine, CA, USA
Catalase (ab209211)	Abcam, Cambridge, UK
Complex I (ABN302)	Millipore Sigma
DRP1 (8570)	Cell Signaling
GAPDH (ab181602)	Abcam
Hexokinase-2 (ab209847)	Abcam
LC3B (PA1-46286)	Life Technologies, Carlsbad, CA, USA
LDHA (NBP2-67483)	Novus Biologicals, Littleton, CO, USA
LDHB (ab53292)	Abcam
Mitofusin 1 (ab126575)	Abcam
Mitofusin 2 (9482)	Cell Signaling
OPA-1 (80471)	Cell Signaling
OXPHOS Antibody (110411)	Abcam
p-Acetyl-CoA Carboxylase (11818)	Cell Signaling
p-DRP1 (S616) (PA5-64821)	Invitrogen, Carlsbad, CA, USA
p-DRP1 (S637) (6329)	Cell Signaling
p16 (92803)	Cell Signaling
PARP-1 (556494)	BD Biosciences, Franklin Lakes, NJ, USA
PDH (3205)	Cell Signaling
PKM1/2 (3190)	Cell Signaling
SOD2 (13141)	Cell Signaling
TKFC (HPA039486)	Sigma Aldrich, St. Louis, MO, USA
Tom20 (42406)	Cell Signaling
Total Acetyl-CoA Carboxylase (3676)	Cell Signaling
TPI (ab135533)	Abcam

The next day, membranes were washed three times with TBST for 5 min each and probed with the secondary antibody for 1 h at RT. After this period, membranes were washed three times for 5 min. Chemiluminescence (ECL, Advansta, San Jose, CA, USA) was detected by the Biorad ChemiDoc imagining system. Observed bands were quantified using the Bio-Rad Image Lab software. Three biological replicates were completed, and values for each protein band of interest were calculated relative to the loading control, then normalized to the control, untreated cells, and plotted, displaying the mean and SEM values in GraphPad Prism.

For long-time points, cells were dosed with 14 mM DHA for 24, 48, and 72 h and 100 μM of tert-butyl hydrogen peroxide (TBHP) for 6 h [32]. After the time point cell medium was collected in a 15 mL conical tube, cells were detached using cell scrapers and resuspended cell medium. The procedure remains the same as previously stated for the complete immunoblotting procedure. GraphPad Prism was used to plot the values calculated relative to the loading control and then relative to the untreated cells over the three biological replicates.

2.6. DHAP Fluorometric Assay

DHAP was measured using the dihydroxyacetone phosphate (DHAP) assay kit (Abcam, Cambridge, UK). H9c2 cells were seeded at 1 x 10⁶ in 15-cm dishes and left for two days to attach in a 5% CO₂ incubator at 37°C. After attachment, cells were dosed with 0.2, 1, or 2 mM for 1, 4, 24, or 48 h. After the exposure period, cells were harvested per the kit's instructions. Once collected, cells were resuspended in DHAP assay buffer and centrifuged according to the kit's instructions. Then, the samples completed a process of deproteinization using 1 M perchloric acid (PCA) and incubated in ice for 5 min. The samples were then centrifuged for 2 min at 23,000 x g, and excess PCA was precipitated using potassium hydroxide at a 1:1 ratio with the sample. The samples were centrifuged, and the supernatant was collected for the assay. The samples were prepared in a reaction mix per the kit's instructions, and fluorescence was measured at 535/587 nm using the BioTek Synergy HR microplate reader (Agilent Technologies, Santa Clara, CA, USA). The results were confirmed over three biological repeats and were graphed using GraphPad Prism to display the mean ± SEM relative to the control.

2.7. Lactate and Pyruvate Measurements

Lactate and Pyruvate levels were measured using the Lactate Glo and Pyruvate Glo kits, respectively, from Promega (Madison, WI, USA). To measure pyruvate levels, H9c2 cells were seeded at 0.75 x 10⁶ in 15-cm dishes and allowed to attach for two days in a 5% CO₂ incubator at 37°C. Once attached, cells were dosed with 0.2, 1, or 2 mM DHA for 1, 4, and 24 h. Once the exposure period was over, cells were washed with PBS, trypsinized with 0.25% Trypsin-EDTA for 1 min, and collected. H9c2 cells were counted for each condition, and 0.3 x 10⁶ cells were added to a microcentrifuge tube. Per the kit's instructions, cells were washed with PBS and glucose and diluted 1:10 in PBS to be added to a white 96-well plate. The samples were then treated with an inactivation solution to lyse the cells and a neutralization solution provided by the kit. A pyruvate detection reagent was added and incubated for 60 min at RT. A luciferin detection reagent was added and read in the BioTek Synergy HR plate reader (Agilent) after 10 min incubation at RT.

Similar steps were performed to measure lactate levels. After treatment, the cell medium was collected for extracellular lactate measurements. According to the kit's instructions, H9c2 cells were counted, and 20,000 cells per well were added to a 96-well plate. Cell medium was diluted 1:10 in PBS and added to the 96-well plate. Inactivation and neutralization solutions were added to the intracellular samples and incubated for 5 min. Then, the lactate detection reagent was added per the kit's instructions and incubated for 60 min. Finally, luminescence was read in the BioTek Synergy HR microplate reader. GraphPad Prism was used to graph the values over three biological repeats for the 1 and 2 mM and two biological replicates for the 0.2 mM DHA. The graph displays the mean ± SEM.

2.8. XF Seahorse Experiments

Glycolysis and Mito Stress Test Seahorse experiments were performed as previously described using a Seahorse XFe96 cell flux analyzer (Agilent Technologies) [27]. H9c2 cells were plated at 5,000 cells per well in a Seahorse cell culture 96-well microplate and left to attach for two days at 37°C in a 5% incubator. After attachment, cells were treated with either medium only or 1, 2, or 14 mM DHA for 24h. The sensor cartridge was hydrated and left ON in a CO₂-free incubator. On the day of the assay, Seahorse serum-free XF medium with required supplements was used to wash H9c2 cells. The Seahorse cell culture microplate was placed in a CO₂-free incubator at 37°C at least an hour before the experiments. Injections for each assay were prepared according to the manual's instructions and loaded into the sensor cartridge. The cartridge plate was then placed in the XFe96 Seahorse analyzer for calibration. After the 20 min calibration step, the cell culture microplate was replaced with a sensor cartridge inserted into the analyzer. The assay read either extracellular acidification rate (ECAR) or oxygen consumption rate (OCR) values, respectively. The values were normalized to cells per well using Hoechst staining read using the BioTek Synergy HR microplate reader. Assay graphs and parameters are presented as the mean values of three biological replicates relative to control ± SEM.

The same procedure as above was used for the Mito Fuel Flex test except for minor differences. The Seahorse medium used was the same for the Mito stress test. The injections prepared were obtained from the Mito Fuel Flex kit and prepared per the kit's instructions. Six parameters were calculated in this experiment: glucose dependency, glucose capacity, glutamine dependency, glutamine capacity, fatty acid dependency, and fatty acid capacity. Using the Wave software, a report is generated for the Mito Fuel Flex test and used to obtain the values for the percentage of dependency and capacity. Graphs are presented as the average of three biological replicates ± SEM using GraphPad Prism.

2.9. Intracellular NAD(P)H measurements

NAD(P)H levels were measured using the endogenous autofluorescence of these dinucleotides when they were excited with two-photon excitation from a 730 nm laser, similar to the techniques described in Blacker and Duchen 2016 [33]. H9c2 cells were plated at 0.1 x 10⁶ density in 60-mm dishes and were left to incubate for two days in a 5% CO₂ incubator at 37°C. After incubation, cells were dosed with either 1, 2, or 14 mM DHA or 100 μM NRH for 24 h as a positive control [28, 34]. Compounds were prepared in a phenol-free medium as the phenol interfered with the autofluorescence detection. Once the exposure time was completed, cells were left to equilibrate to room temperature for 5-10 minutes. Using the Nikon multi-photon microscope, NADH and NAD(P)H autofluorescence was excited at 325-390 nm, with emission collected between 425-500 nm. Experiments were performed using a 25× C-Apochromat (numerical aperture (NA) 1.1) immersion objective. Each condition was imaged over four areas with at least 50-100 cells captured. For analysis, images were thresholded to create a binary mask defining the NAD(P)H signal in the NIS elements software. Intensity values within the binary mask were measured and averaged over the cell fields for each biological replicate. The values are reported as the fluorescent intensity of the mean binary intensity ± SEM over three biological replicates.

2.10. Mitochondrial Reactive Oxygen Species Measurements

To determine mitochondrial superoxide and other reactive oxygen species, mitochondrial ROS was measured using MitoSOX [27]. H9c2 cells were plated at 20,000 cells per well in an 8-well chamber dish and left to attach for two days in a 5% CO₂ incubator at 37°C. Cells were treated with 1, 2, or 14 mM DHA for 24 h. After exposure, cells were stained using MitoSOX dye and incubated for 20 min at 37°C in a 5% CO₂ incubator. Then, the dye was removed, washed twice with PBS, and imaged using the Nikon A1rsci microscope with a 20X immersion objective. For analysis, a minimum of 50 cells per image were obtained, and over five images for each condition. The NIS Elements software created a binary mask over each image. Intensity values within the binary mask were measured and averaged over the cell fields for each biological replicate. At least three biological replicates were obtained, and GraphPad Prism was used to graph the fluorescent intensity for each condition.

2.11. Senescence β-Galactosidase Assay

Per the Senescence β-Galactosidase staining kit’s instructions, H9c2 cells were plated 2,000 cells per well in a 6-well plate and left to attach ON in a 5% CO₂ incubator at 37°C. The next day, the cells were dosed with 14 mM DHA for 72 and 220 h. After exposure, cells were washed with 1X PBS and fixed using the fixative solution in the kit. The cells were washed twice with 1X PBS and stained using the β-Galactosidase solution prepared with supplied reagents. H9c2 cells were incubated ON sealed using parafilm with the staining solution at 37°C in a CO₂-free incubator. After incubation, the staining solution was removed and replaced with 70% glycerol and stored at 4 °C until imaging was performed. Imaging was performed using the all-in-one fluorescence Keyence (BZ-X800) with a 10X objective (NA 0.45) (Keyence, Osaka, Japan). At least 50 cells were imaged over four different fields for each time point in three biological replicates.

2.12. DHA Absorption Levels

For NMR experiments, H9c2 cells were plated at 5,000 cells per well in a 12-well plate. The cells were left to attach in a 37 °C incubator with 5% CO2 for two days. Once attached, cells were treated with 1 or 2 mM DHA or medium only for 1 and 24 h. After exposure, the cell culture medium was collected and centrifuged at 1,000 rpm for 5 min. The supernatant was collected and flash-frozen for ¹H-NMR Analysis[20]. Each sample contains a matched control for each of the time points. The amount of DHA is calculated in millimolar (mM) concentration, measured over two biological replicates, and graphed as mean ± SD. The absolute concentration of DHA present in the solution was calculated relative to the concentration of an internal standard of DMSO (5mM). 50 μl of 50 mM DMSO solution in D₂O was added to 450 μl of supernatant. The resulting solution was analyzed on a Bruker 400MHz Avance NMR instrument for DHA content by ¹H NMR spectroscopy. The intensity of the four hydrogen atoms corresponding to the CH₂ peak of (HOCH₂COCH₂OH) relative to that of the 6 hydrogens present in DMSO allowed us to determine the concentration of DHA in the supernatant. The DHA levels in supernatants obtained from cell cultures were then compared to matched control for each time point. The values are displayed as the mM concentration of DHA for two biological replicates for 1 and 2 mM DHA.

2.13. Design and Analysis

Unless stated otherwise, the data is presented as the average ± SEM of three biological replicates. GraphPad Prism was used to plot the values calculated. Statistical analysis was performed using one-way analysis of variance (ANOVA) with Dunnett’s post hoc test for multiple comparisons while a student’s t-test was used to compare between two groups. Statistical significance was defined as *p< 0.05, **p<0.01, ***p< 0.001, ****p< 0.0001.

3. RESULTS

3.1. DHA promotes cytotoxic and apoptotic cell death in H9c2 cells

Given the cytotoxicity of DHA in other cell models, we first assessed the cytotoxicity of DHA to rat cardiomyocytes, H9c2, using a growth inhibition assay. Cells were acutely exposed to increasing DHA concentrations ranging from 0-14 mM. Cell survival assessed by cell growth decreased dose-dependently with an IC₅₀ value of 5.2 mM ± 0.44 (Fig 1). Interestingly, a small increase in cell viability was observed at 2 mM (93 ± 5.6%) compared to the surrounding 1 mM (84 ± 7.1%) and 4 mM doses (55 ± 2.8%). We further confirmed this dose response using a clonogenic assay and found similar trends to those observed in the growth inhibition assay, though the IC₅₀ value increased to 9.4 ± 0.77 mM (Suppl Fig 1). A plateau in survival was also observed above 8 mM for both assays, where no additional cell death or growth was observed. Approximately 30% of H9c2 cells remained even at DHA doses above 10 mM in the cell growth inhibition. Viable small colonies are also observed in the clonogenic assay. The plateaued growth impacts the inhibitor concentration curve fitting, resulting in changes in IC values between the two assays. However, both assays clearly show DHA is cytotoxic to a majority of cells above 5 mM.

Fig 1. DHA induces cytotoxicity in H9c2 cells.

H9c2 cells were exposed to a range of DHA concentrations ranging from 0-14 mM for 5 days. Cell viability was evaluated through counting, and the survival percentage was calculated relative to the control and displayed using the mean ± SEM of each dose. An IC₉₀ value of 5.2 ± 0.44 mM was calculated.

We previously observed a senescent cell population after the melanoma A375P cells were dosed with DHA [11]. Therefore, we used a β-galactosidase assay to confirm a senescent cell population was present here. H9c2 cells exposed to 14 mM DHA for 72 and 220 h showed increased senescence-associated β-galactosidase activity, confirming cellular senescence. Changes in cell morphology were also observed, with a decrease in cytoplasmic volume and an overall decrease in cell size in DHA-treated cells at 220 h (Suppl Fig 2).

With the cytotoxicity of DHA confirmed in the H9c2 cells, we examined the cell death mechanism responsible. Our previous work with DHA showed cell line-specific cell death mechanisms, including senescence, apoptosis, autophagy, and hybrid mechanisms of autophagy and apoptosis [11, 27, 28]. Therefore, we probed for these pathways in H9c2 cells dosed with 14 mM DHA (IC₇₀) for 24, 48, and 72 h (Fig 2). Confirming the β-galactosidase activity assay, we observed a significant increase in the senescent marker p16, starting at 24 h and continuing until 72 h (Fig 2A). We then examined apoptotic markers, cleaved PARP-1, and caspase 3 (Fig 2B and 2C). Total PARP-1 decreased starting at 48 h with a simultaneous significant increase in cleaved PARP-1 starting at 24 h and continuing to 72 h (Fig 2B). As a positive control, we treated H9c2 cells with tert-butyl hydrogen peroxide (TBHP), which induces the apoptotic cleavage of PARP-1. We probed for full-length and cleaved caspase 3 and found no detectable cleavage of caspase 3. A slight yet non-significant reduction in the full-length Caspase 3 was observed (Fig 2C), suggesting apoptosis through a non-caspase 3/7 dependent mechanism. We then evaluated the autophagy marker LC3B, which showed no increase in overall levels of LC3B or LC3BII (Suppl Fig 3). Although no increase in autophagy was observed, we did observe changes in the mitophagy marker Tom20, which increased at 48 h and significantly increased at 72 h (Fig 2D). These results suggest DHA induced mitophagy, leading predominantly to apoptosis with a small senescent population of cells in the H9c2 cells.

Fig 2. H9c2 cells activate various cell death mechanisms after DHA exposure.

Cells were dosed with 14 mM DHA for 24, 28, and 72, and cell death markers were examined. (A) the senescence marker, p16; (B) the apoptotic marker, cleaved PARP-1; (C) apoptotic marker, full-length and cleaved caspase-3; (D) mitophagy marker, Tom20. Expression levels quantified for three biological replicates reported as the mean ± SEM for each marker displayed relative to the control. The statistical significance is indicated relative to the control, *p < 0.05; **p < 0.01; ***p < 0.001.

3.2. Dose-dependent incorporation of DHA alters cell growth

We previously focused on the highly cytotoxic effects of DHA; however, exposure through e-cigarettes and sunless tanning product use results in lower exposures to DHA [17]. Therefore, we examined the dose-dependent effects of 0.2, 1, and 2 mM DHA on H9c2 cells, focusing on the spike in growth observed at the 2 mM DHA dose.

We first examined changes in cell growth, given that DHA is a carbohydrate fuel source. We acutely dosed the H9c2 cells with 0.2, 1, 2, or 14 mM of DHA, then counted cells every 2 days for 8 days to measure cell growth. Cells were dosed on day 0, and the medium was not changed over the 8 days (Fig 3). H9c2 cells dosed at 14 mM confirmed the senescent population measured (Fig 3A), and the expected decrease in cell number was found starting at day 4 (Fig 3A). At 0.2 and 1 mM DHA, no change in proliferation was observed compared to the untreated control. However, there was an increase in proliferation and doublings at 2 mM DHA, based on the final cell number at day 8 (Fig 3B) [31].

Fig 3. 2 mM DHA alters H9c2 cell proliferation.

(A) H9c2 cells were dosed with 0.2, 1, 2, and 14 mM DHA for 8 days and counted every 2 days to evaluate cell proliferation changes. The graph is displayed with the cell number counted at each time. (B) The proliferation rate was calculated for 0.2, 1, and 2 mM DHA and displayed as doublings per day. The statistical significance is indicated relative to the control, *p < 0.05.

Given the differences in growth, we examined the use of DHA as fuel through its conversion to DHAP by TFKC, which phosphorylates DHA to form DHAP. Using a DHAP fluorometric assay, we examined DHAP levels at 0.2, 1, and 2 mM doses at 1, 4, 24, and 48 h. No change in DHAP levels was observed at the 0.2 mM dose of DHA (Suppl Fig 4). At 1 mM of DHA, DHAP slightly increased at 1 h and remained slightly increased throughout the 48 h exposure period compared to control cells (Fig 4A). The 2 mM dose of DHA showed no increase in DHAP at 1 or 4 h, but a significant increase occurred at 24 h, and the levels of DHAP remained elevated at 48h (Fig 4A).

Fig 4. Dose-dependent conversion of DHA to DHAP in H9c2 cells.

(A) DHAP levels were measured using a fluorometric assay when cells were dosed with 0.2, 1, and 2 mM DHA for 1, 4, 24, and 48 h. DHAP levels were displayed relative to the control. (B) TKFC protein expression levels. (C) Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control. The statistical significance is indicated relative to the control, *p < 0.05.

DHA is only converted to DHAP in TKFC-expressing cells [6]. Immunoblotting of TKFC showed an overall increase in expression after 1 and 2 mM DHA exposure (Fig 4B and 4C). These results demonstrate DHA exposure increases the levels of TFKC in a dose-dependent manner. DHA measurements through NMR show absorption of DHA occurs at 1 h and remains stable through 24 h for both 1 and 2 mM (Suppl Fig 5A). Approximately 50% of the applied DHA dose is absorbed within the first 24 h of exposure. At 1 mM DHA, the average concentration of DHA remaining in the cell medium was 0.52 mM at 1 h and 0.58 mM at 24 h. 0.89 mM at 1 h and 1.1 mM at 24 h was found in the medium after 2 mM DHA exposure (Suppl Fig 5B).

The absorption of DHA and its conversion to DHAP shows a lag between exposure and its incorporation into the metabolic pathway. Further, the dose-dependent effect between 1 and 2 mM shows a boundary where disequilibrium in DHAP occurs 24 h post-acute exposure.

3.3. Low-dose DHA exposure only minimally impacts the glycolytic pathway

Excess DHAP generated through exposure to DHA can be incorporated into nine metabolic pathways [21]. However, its rapid incorporation into glycolytic pathways has been shown previously using 2-C¹³-hyperpolarized DHA [4, 5]. We previously showed IC₉₀ doses of DHA promoted metabolic changes in the glycolysis and oxidative phosphorylation in HEK293T and HepG3, respectively, with changes in the levels of lactate and lactic acid also observed specific to DHA exposure [20, 27, 28]. DHA's entry into the glycolytic pathway is shown in Figure 5A. To monitor changes in glycolysis at these lower doses, we probed for changes in triosephosphate isomerase (TPI), which interconverts GAP and DHAP, and examined proteins responsible for converting glycolytic metabolites upstream and downstream of the trioses to explore the dose-dependent changes between 1 and 2 mM DHA (Fig 5A).

Fig 5. DHA promotes fluctuations in glycolytic proteins.

(A) A diagram of the glycolysis pathway and DHA's incorporation into this pathway. (B) Glycolytic proteins up and downstream of DHA's incorporation were probed. Immunoblotting of triosephosphate isomerase-I (TPI), aldolase A, hexokinase-II, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was dosed with 1 and 2 mM for 1, 4, and 24 h. (C) Expression levels quantified for three biological replicates reported as the mean ± SEM for each marker displayed relative to the control. The statistical significance is indicated relative to the control, **p < 0.01.

TPI showed a gradual increase starting at 4 h and continuing at 24 h at 1 mM, while 2 mM had increased levels at 1 h, sustained at 4 h, and significantly increased at 24 h (Fig 5B and 5C). Aldolase A, which catalyzes the reversible conversion of fructose-1,6-bisphosphate to GAP and DHAP, levels were also examined, and no changes in cells dosed with 1 or 2 mM DHA were found (Fig 5B and 5C). We then looked at Hexokinase-II, the first step in glycolysis and the rate-limiting step in this pathway (Fig 5A). Hexokinase-II expression was slightly increased at 1 and 4 h when cells were dosed with 1 mM DHA. At 2 mM DHA, fluctuating levels between time points were observed but non-significant comparable to control (Fig 5B and 5C). We then examined glyceraldehyde 3-phosphate dehydrogenase (GAPDH) levels, which is the next step in glycolysis after DHAP/GAP, and observed no change in GAPDH protein levels at 1 and 2 mM DHA at any of the time points (Fig 5B and 5C).

Finally, we examined the pyruvate and lactate levels at the end of the glycolytic pathway. We assessed changes in intracellular pyruvate levels using a Pyruvate Glo assay from Promega. Cells were dosed with 0.2, 1, or 2 mM of DHA for 1, 4, and 24 h. When cells were dosed with 1 mM DHA, no change was observed at 1, 4, or 24 h compared to the control. At 2 mM, a slight increase was found at 1 h and a gradual decreasing trend to control levels at 24 h (Fig 6A). Changes in pyruvate levels were also measured at 0.2 mM DHA, which showed no changes compared to untreated control (Suppl Fig 6).

Fig 6. DHA alters pyruvate-lactate ratios in H9c2 cells.

H9c2 cells exposed to 1 and 2 mM DHA for 1, 4, and 24 h. (A) Intracellular pyruvate levels measured using Pyruvate Glo. (B) Intracellular lactate levels measured using Lactate Glo. The graphs display the mean ± SEM relative to the control for three biological replicates. The statistical significance is indicated relative to the control, *p < 0.05.

Changes in intracellular lactate were also measured using a Lactate Glo assay from Promega. At 1 mM, no changes were observed at any of the time points. At 2 mM, lactate levels significantly decreased at 1 h and gradually increased to 24 h compared to the control, consistent with intracellular pyruvate levels. Extracellular lactate levels were also measured and found to have no overall changes after DHA exposure, although levels were 50-fold higher in concentration than intracellular lactate levels (Suppl Fig 7).

We also examined pyruvate and lactate-related proteins using immunoblotting at these time points. DHA's integration into these downstream pathways is shown in Figure 7A. The pyruvate kinase muscle isozyme 1 and 2 (PKM1/2) converts phosphoenolpyruvic acid (PEP) to pyruvate. PKM1/2 remained unchanged at 1 and 2 mM dosed cells (Fig 7B and 7C). Pyruvate dehydrogenase (PDH), which converts pyruvate to acetyl CoA, was also probed. PDH showed no change at 1 and 4 h when dosed with 1 mM DHA, but a non-significant decrease occurred at 24 h. At 2 mM, no change was shown at 1 h, but PDH decreased non-significantly at 4 and 24 h (Fig 7B and 7C).

Fig 7. Pyruvate and lactate-related proteins were altered due to DHA exposure.

H9c2 cells exposed to 1 and 2 mM DHA for 1, 4, and 24 h. (A) A diagram of the downstream glycolysis pathway involving pyruvate and lactate and its incorporation into mitochondrial pathways. (B) Pyruvate-involved proteins, including pyruvate kinase muscle isozymes 1 and 2 (PKM1/2) and pyruvate dehydrogenase (PDH), probed at 1, 4, and 24 h. (C) Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control. (D) Lactate Dehydrogenase-A (LDHA) and Lactate Dehydrogenase-B (LDHB), lactate-related proteins were probed at 1, 4, and 24 h. (E) Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control. The statistical significance is indicated relative to the control, *p < 0.05; **p < 0.01; ***p < 0.001; ****p<0.0001.

Lactate-related proteins were also probed through immunoblotting (Fig 7D). Lactate dehydrogenase-A (LDHA) converts lactate to pyruvate, while lactate dehydrogenase-B (LDHB) converts pyruvate to lactate. At 1 mM, LDHA was found to have a significant decrease at 1 h and then increase at 4 and 24 h. The trends differ at 2 mM, where LDHA significantly decreases at all time points (Fig 7D and 7E). At the same time, LDHB was found to have no change at 1 and 4 h and a significant decrease at 24 h when dosed with 1 mM (Fig 7D and 7E). An opposite trend was found when cells were dosed with 2 mM, with a significant decrease at 1 h, with protein levels recovering at 4 and 24 h (Fig 7D and 7E).

Finally, we analyzed metabolic flux using an XF Seahorse glycolysis stress test, which measures changes in the glycolytic pathway. As expected from our previous studies, the highly cytotoxic doses of DHA (14 mM) show significantly decreased extracellular acidification rate (ECAR) and glycolytic parameters (Suppl Fig 8B). When dosed with 1 mM DHA for 24 h, an increase in ECAR was observed, while 2 mM DHA exhibited a slight decrease in ECAR (Suppl Fig 8A). A slight increase was also observed in both 1 and 2 mM in non-glycolytic acidification, while other glycolytic parameters remained constant (Suppl Fig 8B).

We observed little change in the glycolytic pathways at these low and non-cytotoxic doses. However, decreased LDH protein expression, increased pyruvate levels, and non-glycolytic acidification suggest changes in mitochondrial metabolic pathways.

3.4. Low-dose DHA exposure promotes mitochondrial energy flux

We examined mitochondrial metabolic flux using an XF Seahorse Mitofuel flex test. This assay uses various inhibitors to evaluate the dependency and capacity of the cells in mitochondrial oxidation of glucose, glutamine, or fatty acids when the other respective pathways are inhibited. To evaluate the mitochondrial glucose pathway, the UK5099 inhibitor blocks the conversion of pyruvate to acetyl-CoA in the mitochondria. To measure oxidation in the glutamine pathway, etomoxir inhibits the carnitine palmitoyltransferase-1 (CPT1) and prevents the conversion to acetyl-CoA. The glutamine pathway measured uses BPTES, a glutaminase inhibitor, to prevent α-ketoglutarate (α-KG) from entering the TCA cycle. H9c2 cells were treated with 1 and 2 mM for 24 h to evaluate fuel dependency and capacity. The glucose utilization changed when cells were dosed with 1 and 2 mM DHA. The glucose pathway utilization measured here differs from the glycolysis stress test because it is inhibited in this assay when pyruvate enters the mitochondria. After DHA treatment, glucose dependency remained constant in both 1 and 2 mM DHA (Fig 8A). At the same time, glucose capacity increased at 1 and 2 mM DHA 84.3 ± 9.2 % and 88.5 ± 22.5%, respectively, compared to a 60.9 ± 5.6% observed in untreated control cells (Fig 8B). Glutamine dependency also slightly decreased at 1 and 2 mM DHA, resulting in a dependency change after dosing with DHA. In cells dosed with 1 mM DHA, dependency shifted to 28.8 ± 22.2 %, while at 2 mM DHA was 17 ± 8.2 % (Fig 8C). In glutamine capacity, a slight increase was observed in 1 mM (68.2 ± 22.5 %) and a decrease in 2 mM DHA (44.2 ± 9.2 %) compared to control (59.9 ± 14.3 %) (Fig 8D). The fatty acid dependency remained unchanged at 1 and 2 mM DHA, while the capacity slightly increased (Fig 8E). At 1 mM DHA, capacity was observed at 71.7 ± 20.4 %, while cells dosed with 2 mM were 63.7 ± 11.5 % capacity (Fig 8F).

Fig 8. Cardiomyocytes use several mitochondrial fuel pathways after DHA exposure.

H9c2 cells were dosed with 1 and 2 mM DHA for 24 h, and a Mito Fuel Flex test was performed to determine changes in fuel utilization. (A and B) Glucose dependency and capacity percentages. (C and D) The glutamine dependency and capacity percentages. (E and F) Fatty acid dependency and capacity percentage. All fuel percentages are reported as the mean ± SEM of three biological replicates.

Since changes in the cell's fuel utilization were observed with DHA exposure, we probed key proteins involved in these mitochondrial metabolic pathways. Acetyl-CoA is involved in various pathways in the mitochondria, including the fatty acid synthase and TCA cycle (Fig 9A). Phospho-acetyl-CoA carboxylase is involved in the first step to converting acetyl-CoA to fatty acid and was found to remain constant at all time points with 1 mM DHA (Fig 9B and 9C). When cells were dosed with 2 mM DHA, a significant increase was found at 1 h and continued to increase significantly at 4 and 24 h (Fig 9C). Total acetyl-CoA carboxylase levels remained constant at 1 and 2 mM dosed cells (Fig 9B and 9C). Acetyl-CoA carboxylase 1 (ACC1) aids in converting acetyl-CoA to malonyl-CoA in fatty acid synthesis. Levels of ACC1 had a slight non-significant increase at 1 mM starting at 1 h and sustained until 24 h, while 2 mM had an initial spike at 1 h and continued to increase at 4 h (Fig 9B and 9D). Acetyl-CoA carboxylase 2 (ACC2) converts to acetyl-CoA from the β-oxidation pathway (Fig 9A). ACC2 had an increasing trend starting at 4 h and continued at 24 h at 1 mM, while 2 mM had an initial spike at 1 h and continued this trend until 24 h (Fig 9B and 9D).

Fig 9. DHA induces dose-dependent Acetyl-CoA imbalances in H9c2 cells.

(A) A schematic of the acetyl-CoA protein and its involvement in various metabolic pathways in the mitochondria. (B) Phospho-Acetyl-coA carboxylase, Acetyl-CoA carboxylase 1, Acetyl-CoA carboxylase 2, and total Acetyl-CoA carboxylase proteins were measured in cells dosed with 1 and 2 mM DHA for 1, 4, and 24 h. (C) Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control. (D) Acetyl-CoA carboxylase 1 and Acetyl-CoA carboxylase 2 protein expression levels. The statistical significance is indicated relative to the control, *p < 0.05; **p < 0.01; ***p < 0.001.

Finally, with the changes in fatty acid fuel utilization and increases in acetyl-CoA leading to fatty acid production, we examined fatty acid synthase (FASN) (Fig 10). A significant increase at 24 h was found in cells dosed with 1 mM. While at 2 mM, FASN significantly increased at 1 and 4 h and stayed elevated at 24 h (Fig 10B).

Fig 10. FASN increases after DHA exposure at low doses.

(A) Fatty acid synthase (FASN) in cells dosed with 1 and 2 mM DHA at 1, 4, and 24 h. (B) Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control. The statistical significance is indicated relative to the control, *p < 0.05; **p < 0.01.

DHA changes fuel utilization across several mitochondrial metabolic pathways, shifting toward fatty acid synthesis at 2 mM as an increase in fatty acid fuel capacity and increases in overall ACC and FASN proteins were found.

3.5. OXPHOS and ETC proteins are increased with low dose DHA

With altered mitochondrial metabolism, we measured changes in oxidative phosphorylation (OXPHOS) using the XF Seahorse Mito Stress test (Fig 11). H9c2 cells were treated with 1, 2, and 14 mM DHA for 24 h. As expected, the highly cytotoxic dose of 14 mM DHA decreased OCR, basal respiration, and ATP production consistent with our previous work (Suppl Fig 9) [27]. For the low and non-cytotoxic doses, OCR increased compared to control cells (Fig 11A). A significant increase was found in the proton leak parameter at 1 mM DHA, while only a slight increase was seen at 2 mM (Fig 11E). A slight non-significant increase in ATP production, basal respiration, and maximal respiration was observed at 1 and 2 mM DHA (Fig 11B, 11C and 11D). No changes were found in spare respitatory capacity or non-mitochondrial respiration (Fig 11F and 11G).

Fig 11. DHA exposure promotes changes in mitochondrial function.

(A) Mito Stress test was conducted in H9c2 cells, and the oxygen consumption rate (OCR) was measured in cells dosed with 1 and 2 mM DHA at 24 h. Several parameters were calculated in this experiment. Basal respiration (B), ATP production (C), maximum respiration (D), proton leak (E), spare respiratory capacity (E), and non-mitochondrial respiration (F) were graphed using three independent experiments and displayed as relative to control. The statistical significance is indicated relative to the control *p < 0.05.

Changes in OXPHOS suggest changes in the mitochondria electron transport chain (ETC) may occur after DHA exposure. Immunoblotting of ETC subunits showed changes in the various complexes when cells were dosed with DHA at various time points. Complex I levels significantly increased at 1 mM DHA for 1 and 4 h, while 2 mM DHA significantly increased at 1 h and continued at 4 h and 24 h (Fig 12A). A slight increase in complex II started at 1 h, and no change occurred at 4 and 24 h at 1 mM DHA (Fig 12B). For 2 mM DHA, a gradual decrease started at 4 h, and a significant decrease at 24 h was observed (Fig 12B). No changes in complex V were measured at 1 or 2 mM DHA (Fig 12C).

Fig 12. Electron chain subunit expression altered by DHA exposure.

Electron transport chain (ETC) subunits were probed by immunoblotting in cells dosed with 1 and 2 mM DHA for 1, 4, and 24 h. (A) Complex I, (B) Complex II, and (C) Complex V. Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control. The statistical significance is indicated relative to the control, *p < 0.05; **p < 0.01; ***p < 0.001; ****p<0.0001.

While the highly cytotoxic DHA dose significantly reduced OXPHOS, the low DHA doses showed elevated OXPHOS. The 2 mM dose has the strongest effect on ETC protein expression with elevated complex I and decreased complex II levels compared to 1 mM DHA.

3.6. Increased NAD(P)H and reductive stress result from 2 mM DHA exposure

Along with changing ETC complexes, DHA also impacts the essential cofactor pool for OXPHOS [28]. We used two-photon autofluorescence imaging to measure changes in NAD(P)H pools in the H9c2 cells. NADH and NADPH have the same excitation and emission spectrum, so we cannot separate these species, but we can detect increases or decreases in their pool within the cell. Dose-dependent NAD(P)H changes were observed after DHA exposure. A significant increase in NAD(P)H was measured 24 h after 2 mM DHA exposure (Fig 13B), while no NAD(P)H pool change was observed at 1 mM DHA (Fig 13). The NAD(P)H autofluorescence appeared more intense and punctate in the cytosol at 1 and 2 mM DHA doses. While we could not specifically colocalize these to the mitochondria, they are consistent with the mitochondria increases we observed in HEK293T [28].

Fig 13. NAD(P)H cofactor pools increase with DHA dose.

(A) H9c2 cells were exposed to 1 and 2 mM DHA for 24 h, and the NAD(P)H levels were measured using a multi-photon microscope. (B) NAD(P)H levels quantified and displayed as fluorescence intensity (a.u) using three biological replicates. The scale bar is 50 μm. Dihydronicotinamide riboside hydride (NRH) was dosed at 100 μM for 24 h and used as a positive control to increase NAD(P)H levels. The statistical significance is indicated relative to the control, **p < 0.01.

We also checked the highly cytotoxic 14 mM DHA dose, and no changes in NAD(P)H levels were measured (Suppl Fig 10). We confirmed changes are specific to the NAD(P)H pool by also dosing cells with 100 μM dihydronicotinamide riboside hydride (NRH), which directly boosts the NAD(P)H pool [28, 34]. Interestingly, the increase in NAD(P)H produced by NRH is equivalent to that produced by 2 mM DHA (Fig 13B).

Given these NAD(P)H changes, we expect changes in reactive species within the mitochondria. We measured ROS using MitoSOX, which reports on superoxide and hydroxyl radicals within the mitochondria. An increase in MitoSOX fluorescent intensity was measured in cells dosed with 1 and 2 mM DHA (Fig 14A). Thus, low-dose DHA exposure increases reactive species in the mitochondria. No change in mitochondrial ROS was observed when cells were dosed with 14 mM DHA for 24 h (Suppl Fig 11).

Fig 14. Mitochondrial ROS increased in H9c2 cells.

(A) Mitochondrial ROS was measured using MitoSOX dye in cells dosed with 1 and 2 mM DHA for 24 h and imaged using a confocal microscope. (B) Quantification of fluorescence intensity. Scale bar 50 μm. The statistical significance is indicated relative to the control, *p < 0.05; **p < 0.01.

Mitochondrial oxidative stress promoted by DHA may change antioxidant responses. We examined if there were changes in superoxide dismutase (SOD2) and catalase expression after DHA exposure. SOD2 scavenges oxygen radicals that specifically occur in the mitochondria. H9c2 dosed with 1 mM DHA showed a slight increase in SOD2 levels at all time points, while 2 mM DHA increased SOD2 at 1 h and remained elevated at 4 and 24 h (Fig 15A). Catalase serves as a catalyst protein that prevents oxidative damage. Catalase remained constant at 1 and 2 mM DHA compared to the control (Fig 15B).

Fig 15. Antioxidant balance altered by DHA exposure.

H9c2 cells were exposed to 1 and 2 mM DHA for 24 h. The following antioxidant proteins were examined: (A) superoxide dismutase 2 (SOD2) and (B) catalase. Expression levels quantified for three biological replicates reported as the mean ± SEM displayed relative to the control.

Low dose DHA increased mitochondrial ROS at 1 and 2 mM, but the strong NAD(P)H pool changes were only observed at 2 mM DHA. The changes in NAD(P)H and SOD2 contribute to more reductive stress at the 2 mM dose and contribute to the changes in complex I of the ETC observed.

3.7. Low dose DHA also impacts mitochondrial biogenesis

With the significant changes in mitochondrial function and cofactors observed, we examined fission/fusion markers, crucial for maintaining quality control in the mitochondria. Phosphorylation of DRP1 at S637 (pDRP1^S637) inhibits mitochondrial fission, while S616 (pDRP1^S616) activates it [35]. For 1 mM DHA, pDRP1^S637 remained constant at all time points (Fig 16A and 16B). At 2 mM, pDRP1^S637 levels remained constant at 1 and 4 h and significantly decreased at 24 h (Fig 16A and 16B). Conversely, pDRP1^S616 levels remained constant at 1 and 4 h, significantly decreased at 24 h when dosed with 1 mM, and slightly increased at 4 h and remained constant at 24 h with 2 mM DHA (Fig 16A and 16B). Fusion markers, mitofusin 1 (Mfn1) and mitofusin 2 (Mfn2) were also probed (Fig 16C). Mitofusin 1 and 2 function outside the membrane of the mitochondria as part of the fusion process [36]. Mitofusin 1 increased starting at 4 h and continuing until 24 h when dosed with 1 mM, while a two-fold increase was observed with 2 mM at 1 h and was maintained until 24 h (Fig 16C and 16D). Mitofusin 2 significantly decreased at 24 h with 1 mM DHA, and 2 mM showed slight increases at all time points (Fig 15C and 15D). Optic atrophy-1 (OPA-1) was also probed as it's required inside the mitochondria in the fusion process [36]. OPA-1 was found to have fluctuating levels in both 1 and 2 mM (Suppl Fig 12A and 12B).

Fig 16. DHA exposure promotes imbalances in fission-fusion dynamics.

(A) Fission proteins, p-DRP1 (S616), p-DRP1 (S637), and DRP1 were measured in cells exposed to 1 and 2 mM DHA for 1, 4, and 24 h. (B) Quantification of fission proteins graphed relative to the control. (C) Fusion proteins, mitofusin 1 and 2 in cells dosed with 1 and 2 mM DHA at 1, 4, and 24 h. (D) Quantification of protein levels displayed relative to control in the graphs using three independent experiments. The statistical significance is indicated relative to the control, *p < 0.05; ***p < 0.001.

DHA exposures altered mitochondrial biogenesis in a dose-dependent manner. At the 1 mM dose, H9c2 cells experienced imbalances in fission and fusion proteins, which can lead to dysregulation of mitochondrial dynamics. At 2 mM DHA, quality control of the mitochondrial biogenesis was observed as activation of both fission and fusion proteins was found with an increase in pDRP1^S616 and a significant decrease found in pDRP1^S637 levels and increases in both Mfn 1 and 2 was observed.

4. DISCUSSION

DHA is a triose sugar easily incorporated into metabolic pathways after phosphorylation. While its carbohydrate nature may appear harmless, we have previously hypothesized that exogenous exposures to DHA could mimic fructose exposures, increasing oxidative stress and mitochondrial dysfunction across exposed organs and tissues [25]. It is already known that fructose exposures increase DHAP and GAP levels, increasing reactive oxygen and reactive nitrogen species (ROS/RNS) while altering antioxidant balance, changing metabolic utilization, and reducing mitochondrial function [37-39]. Thus, exposure to DHA through inhalation or absorption can lead to similar effects [40]. Some studies of e-cigarette e-juice, focused on propylene glycol and glycerol, have identified mitochondrial changes and metabolic reprogramming similar to the effects we have observed for DHA alone, yet these studies have not consider the production of DHA from the aerosol [29, 30]. Additionally, the effects are not consistently observed in other works using propylene glycol and glycerol, suggesting the effects may be e-cigarette device-specific [29, 30, 41, 42]. Therefore, a consistent mechanism has not been attributed to these effects. However, given that DHA is produced by heating, which is variable from these devices, DHA might be a driver for these effects, but this has not been appreciated to date. With the strong mitochondrial effects we have observed in other cell models, highly energetic organs, like the heart, could be more affected by DHA exposures. Therefore, to understand the possible impact DHA exposures have on cardiac tissues, we examined the acute changes in metabolism and mitochondrial function associated with low and non-cytotoxic exposures to DHA using H9c2 rat cardiomyocytes. We selected H9c2 because it is commonly used to examine cardiac injury in fructose exposures, which we posit DHA mimics [43].

We first characterized the cytotoxicity of DHA to the cardiomyocytes. A unique spike in the growth inhibition curve was found at 2 mM DHA, which was not previously seen in other cell models [2, 9-11, 27, 28]. Additionally, the cytotoxic effects of DHA plateaued in the H9c2 after 8 mM, resulting in a senescent population (Fig 1A and Suppl Fig 1). The apoptotic cell death observed for ~70% of the cells was consistent with previous studies in skin models, including keratinocytes and melanoma [2, 9-11] (Fig 2). However, there was a sustained senescent population, which has not been previously reported after DHA exposures. Only one previous report noted a senescent population in the melanoma model A375P after 5 mM exposure to DHA [11]. Still, this population was only observed between 24 and 48 h after exposure, and the cell quickly transitioned to apoptotic cell death.

We then confirmed that 2 mM DHA increased the proliferation and growth of the H9c2 cells despite a low cytotoxic effect at 1 mM (IC₁₅). To understand the increase in cell proliferation, we examined the metabolic differences between the 1 and 2 mM. The 1 and 2 mM doses are on the higher end of expected exposures from spray tanning or e-cigarettes but are relevant for physiological exposures [3, 7, 10, 13, 14]. The dose-dependent effects observed may be attributed to the amount of DHA converted to DHAP by TKFC in the cells. TKFC increased at 1 and 2 mM doses, but DHAP levels only significantly changed at 2 mM 24 h after the DHA was applied (Fig 4). Changes in the equilibration of the excess DHAP through the glycolytic pathway could explain the differences between the two doses. However, the only significant changes observed along the glycolytic pathway were at the pyruvate-to-lactate conversion, where pyruvate levels increased 1 h after 2 mM exposure, and lactate levels decreased (Fig 6). At 2 mM DHA, LDHA and LDHB showed significant and sustained changes out to 24 h after 2 mM DHA (Fig 7). Suppressed LDHA and LDHB at 2 mM DHA likely contribute to the increase in DHAP measured at 24 h since glycolytic flux to pyruvate and lactate is suppressed from 4 to 24 h after exposure.

Increased pyruvate can change redox balance and increase NAD⁺/NADH levels in the cells and mitochondria [44, 45]. We confirmed these changes after DHA exposure by examining changes in the mitochondria relevant to pyruvate influx at 1, 4, and 24 h. We observed increased phospho-Acetyl-CoA at 1, 4, and 24 h after 2 mM DHA, which suggests the increased pyruvate at 1 h post-exposure is used by the citric acid cycle (TCA) (Fig 9). We also see increased ACC1 and ACC2 at 2 mM, 1, 4, and 24 h after exposure, consistent with the pyruvate equilibrating in other metabolic pathways (Fig 9). No significant changes in these proteins or pathways are observed at the 1 mM dose, consistent with the lack of change in pyruvate, lactate, or DHAP [46].

Altered metabolic flexibility was confirmed by the increase in mitochondrial glucose and fatty acid capacity for the 1 and 2 mM doses 24 h after exposure, as well as increased OXPHOS (Figs 8 and 11). However, the significant increase in FASN at the 2 mM dose and the Acetyl-CoA enzyme changes suggest the excess DHA at the 2 mM dose is channeled into usable fuel, while the 1 mM dose is not (Fig 10). Increases in the FASN pathway have increased cell proliferation and survival, much like we currently see with the 2 mM dose [47]. Also, FASN is a key fuel in cardiomyocytes. It has been shown that when a decrease in cardiomyocyte proliferation and survival was observed, it was attributed to FASN, and fatty acid-related genes were inhibited [48]. These changes would explain the increase in growth and proliferation for the 2 mM dose [49].

The altered metabolic profile produced by DHA increased mitochondrial ROS for 1 and 2 mM DHA (Fig 14). However, the 2 mM dose also sees changes in its NAD(P)H pools, significant increases in complex I expression, and an increase in SOD2 protein (Figs 12, 13, 15). These changes buffer the mitochondria injury from the 2 mM dose, supporting cell survival over the 1 mM dose. These changes can also be attributed to the increase in ACC proteins found at the 2 mM dose as they target cardiac repair and protection along with antioxidative response when increased due to a change in nutrient source [46]. Similarly, increased pDRP1^S616 and a significant decrease in pDRP1^S637 at 2 mM DHA allow the repair of any damaged mitochondria through the activated fission process. An increase in Mfn2 was also noted to promote cell survival (Fig 16) [50]. At 1 mM DHA, the significant decrease in Mfn2 and pDRP1^S616 disrupts the mitochondrial network balance, leading to further damage (Fig 16D and 16D) [51].

Altogether, these changes suggest a tipping point in DHA exposures exists in cardiomyocytes, which has not been observed in other cell models. Acute exposure to 1 mM DHA induced mitochondrial and metabolic stress, killing ~10-15% of the cell population, with the remaining population showing few lasting changes from the acute exposure. With 2 mM, H9c2 cells experienced a faster initial response and more prolonged signaling changes, resulting in metabolic shifts and reductive stress. As the cardiomyocytes attempt to compensate for the quick changes in metabolism due to the 2 mM DHA, they can prolong their survival and increase proliferation by using DHA as a fuel through their preferred energy source, fatty acids [52].

The increase in FASN and likely fatty acids in our model corresponds to other reports examining e-cigarettes and their systemic effects. Changes have been found in FASN, fatty acids, and circulating lipids which were observed after e-cigarette vapor exposure which have been linked to cardiovascular dysfunction [29, 53, 54]. Increases in free plasma fatty acids have been attributed to mitochondrial changes, and oxidative stress in these models [29, 55, 56]. Some of these effects have been attributed to nicotine in e-cigarettes and compared to similar effects observed with traditional cigarettes [29]. However, e-juices have also been shown to produce these effects, leading to one report suggesting e-cigarettes could cause more profound health effects than high fat diets [57]. The metabolic reprogramming we observe with DHA alone suggests it contributes to the observed effects in other e-liquid studies. Further work is needed to understand the fatty acid and potential inflammatory effects DHA produces.

The energetic demands of these cells may promote metabolic adaption to DHA between 1 and 2 mM, but the capacity for adaption is quickly overwhelmed by the imbalanced triose sugars, resulting in the cytotoxicity of higher millimolar doses after 2 mM. The cytotoxicity of DHA has been best studied in the skin. Still, we have demonstrated that systemic models like HEK293T and HepG3 show cytotoxicity at low millimolar doses, below the cytotoxic doses for the skin [27, 28]. While the cytotoxicity has been clearly demonstrated, the genotoxicity of DHA is much less certain. There are conflicting reports in the literature about DHA’s ability to induce strand breaks and DNA adducts, and the data reported seems to indicate the DNA damage observed is associated with cell death [2, 9, 10]. In 1979, Pham et al. demonstrated DHA was mutagenic in an Ames test and showed DNA damage was induced in Bacillus subtilis [58]. However, these studies have not been confirmed in mammalian systems. Our results suggest mitochondrial ROS occurs after DHA exposures, particularly at low millimolar doses, but the genotoxicity of DHA to genomic or mitochondrial DNA has not been confirmed.

While DHA exposures from sunless or spray tanning happen at higher concentrations less frequently, vaping, like smoking, can occur daily. The variety in vaping devices and volumes of e-liquids make measuring DHA exposure levels more difficult. However, DHA is potentially produced by all e-liquid-containing devices. Therefore, depending on the number of sessions per day or week, high micromolar doses to low millimolar doses of DHA may be inhaled and systemically distributed. The potential for frequent DHA exposures confirms that comparing DHA to fructose is not unrealistic. Like in fructose exposures, DHA shows changes in metabolic pathways through acetyl-CoA and FASN [59]. High fructose exposures also alter NAD(P)H pools and increase ROS levels while promoting changes in antioxidants in the heart, which we also observe with DHA [60]. Additionally, DHA produces significant metabolic reprogramming and mitochondrial changes consistent with inflammatory diseases seen after e-cigarette or traditional cigarette smoking [29, 53, 57]. While it is unclear if DHA acts as a genotoxin or carcinogen, the proliferation effects observed at 2 mM along with the mitochondrial reprogramming and reactive species suggest DHA may produce detrimental cellular effects with sustained or chronic exposures. More studies are needed in physiological models to map the metabolic and mitochondrial effects of chronically inhaling and absorbing DHA. Yet, the parallels between DHA and fructose exposures, as well as e-liquid exposures, are clear. DHA exposures impact cardiometabolic health, and sustained exposures need closer scrutiny with refined dose estimates and mapped accumulation effects.

5. CONCLUSIONS

DHA exposures in rat cardiomyocytes H9c2 promoted dose-dependent effects. At the cytotoxic DHA dose, H9c2 cells were found to promote apoptotic cell death and an overall decrease in metabolic function similar to the effects previously characterized in other cell models. While at the non-cytotoxic dose of 2 mM DHA, cells experienced an increase in cell survival and metabolic reprogramming through increases in ACC and FASN proteins compared to the low cytotoxic dose of 1 mM DHA. These changes are characterized by distinct changes in fuel utilization pathways in the mitochondria and reductive stress at 2 mM dose, resulting in metabolic adaption and mitochondrial protection.

Supplementary Material

1

Suppl Fig 1. DHA cytotoxicity confirmed in H9c2 cells. (A) H9c2 cells were exposed to increasing DHA concentrations (0-20 mM) for 5 days. Cells were counted for each concentration. (B) The survival percentages are displayed as percent survival. An IC₅₀ was calculated as 9.4 ± 0.77 mM DHA.

2

Suppl Fig 2. Senescence staining was confirmed in H9c2 cells. Senescent cells were stained using β-Galactosidase staining in cells dosed with 14 mM DHA for 72 and 220 h. Bright-field colorimetric images were taken, and representative images of three biological replicates are displayed. The scale bar is 25 μm.

3

Suppl Fig 3. Alternative cell death mechanisms investigated in H9c2 cells. (A) H9c2 cells were exposed to 14 mM DHA for 24, 48, and 72 h, and the autophagy marker, LC3B, was examined. (B) Expression levels of LC3B protein are displayed as the average of three biological replicates ± SEM.

4

Suppl Fig 4. DHAP levels were measured at the non-cytotoxic dose. H9c2 cells were dosed with 0.2 mM DHA for 1,4, 24, and 48 h. DHAP levels were measured using a DHAP fluorometric assay. The graph displays the mean ± SEM of treated values relative to control for three biological replicates.

5

Suppl Fig 5. DHA measurements were evaluated using NMR analysis. H9c2 cells were exposed to 1 and 2 mM DHA for 1 and 24 h. (A) The table shows DHA levels calculated using NMR Analysis. (B) The graph displays the mean ± SD of the millimolar concentration of DHA in two biological replicates.

6

Suppl Fig 6. Intracellular pyruvate measurements at the non-cytotoxic dose. H9c2 cells were dosed with 0.2 mM DHA for 1, 4, and 24 h, and pyruvate levels were measured using the Pyruvate Glo kit. The graph shows three biological replicates displayed relative to the control using the mean ± SEM.

7

Suppl Fig 7. Extracellular lactate levels remain unchanged by DHA exposure. Extracellular lactate levels were measured using the Lactate Glo kit at 1, 4, and 24 h for 1 and 2 mM DHA. The graph displays the mean ± SEM of three biological replicates relative to the control.

8

Suppl Fig 8. Dose-dependent changes in glycolysis after DHA exposure in H9c2 cells. (A) A glycolysis stress test was performed in H9c2 cells dosed with 1, 2, and 14 mM for 24 h, and the extracellular acidification rate (ECAR) was measured. (B) Several parameters were calculated: glycolysis, glycolytic rate, non-glycolytic acidification, and glycolytic reserved are displayed as the relative control of three individual runs and plotted using the mean ± SEM value. The level of statistical significance was marked as follows: *p < 0.05.

9

Suppl Fig 9. Mito Stress test revealed a suppression in OXPHOS at 14 mM DHA. (A) Mito Stress test was performed in H9c2 cells dosed with 14 mM DHA for 24 h, and the oxygen consumption rate (OCR) was measured. (B) Several parameters were calculated: basal respiration, ATP production, maximal respiration, proton leak, spare respiratory capacity, and non-mitochondrial respiration. The graphs display the mean ± SEM relative to the control of three independent experiments. The level of statistical significance was marked as follows: **p < 0.01, ***p < 0.01.

10

Suppl Fig 10. NAD(P)H levels at the cytotoxic dose of 14 mM DHA. H9c2 cells were dosed for 24 h with 14 mM DHA. NAD(P)H levels were measured using a multi-photon microscope and quantified as fluorescence intensity (a.u). The graph displays the mean ± SEM of three biological replicates.

11

Suppl Fig 11. Mitochondrial ROS measurements at the cytotoxic dose. Mitochondrial ROS was measured using MitoSOX dye in cells dosed with 14 mM DHA for 24 h. The graph is displayed as mean ± SEM of the fluorescence intensity (a.u) using three biological replicates. The scale bar is 50 μm.

12

Suppl Fig 12. OPA-1 levels remain unchanged after DHA exposure. (A) H9c2 cells were dosed with 1 and 2 mM DHA for 1, 4, and 24h and the fusion protein, optic atrophy-1 (OPA-1), was measured through immunoblotting. (B) Quantification of protein levels relative to control is displayed as the mean ± SEM of three independent experiments.

Highlights:

• Non-cytotoxic doses of DHA induce dose-dependent changes in mitochondrial metabolic pathways

• Increased cell proliferation occurs in H9c2 at 2 mM DHA

• Cytotoxic doses of DHA induce mitophagy and suppression of mitochondrial metabolic pathways

Data availability statement

All relevant data are included in the manuscript and its supporting information files.