Considerations for using isolated cell systems to understand cardiac metabolism and biology

Abstract

Changes in myocardial metabolic activity are fundamentally linked to cardiac health and remodeling. Primary cardiomyocytes, induced pluripotent stem cell-derived cardiomyocytes, and transformed cardiomyocyte cell lines are common models used to understand how (patho)physiological conditions or stimuli contribute to changes in cardiac metabolism. These cell models are helpful also for defining metabolic mechanisms of cardiac dysfunction and remodeling. Although technical advances have improved our capacity to measure cardiomyocyte metabolism, there is often heterogeneity in metabolic assay protocols and cell models, which could hinder data interpretation and discernment of the mechanisms of cardiac (patho)physiology. In this review, we discuss considerations for integrating cardiomyocyte cell models with techniques that have become relatively common in the field, such as respirometry and extracellular flux analysis. Furthermore, we provide overviews of metabolic assays that complement XF analyses and that provide information on not only catabolic pathway activity, but biosynthetic pathway activity and redox status as well. Cultivating a more widespread understanding of the advantages and limitations of metabolic measurements in cardiomyocyte cell models will continue to be essential for the development of coherent metabolic mechanisms of cardiac health and pathophysiology.

1. Introduction

Metabolism is an area of traditional biochemistry that, until recently, was considered a relatively mature field. The backbone of metabolic knowledge reflects the success of our scientific approaches and technological advances. New methods for understanding complex system behavior have further propelled the scientific community’s quest to understand metabolism, not only by increasing the breadth of exploration, but also by increasing the throughput and analysis of metabolic data. In addition, the development of cell models of cardiovascular disease, such as induced pluripotent stem cells (iPSCs), have expanded interest in understanding the metabolic properties of cultured cells, which could be used to predict disease or to devise actionable interventions. Experimental systems that integrate cell models with techniques for measuring metabolic activity have been used to generate volumes of data that must be carefully weighed for elements of rigor and for their meaningfulness for understanding in vivo biology. Ultimately, the value of data from isolated cell systems depends upon their usefulness to contribute to coherent explanations of biological phenomena. In this review, we discuss methodological considerations for measuring cardiomyocyte cell metabolism.

2. How does the heart’s metabolic machinery process substrates?

A general goal of many metabolic studies is to understand how living systems regulate the production of useable energy. Because the heart has an extremely high energetic demand, it is a model organ for understanding how cells control ATP levels (reviewed in [1]). Large quantities of ATP are required to maintain cardiac ion homeostasis and contractile function, and this ATP demand is met via catabolism of a variety of circulating hydrocarbon substrates [2–8]. For this reason, the heart has been compared to an engine, which utilizes oxygen to burn hydrocarbon and produce energy for mechanical work. In the context of heart failure, the heart has been suggested to be either an inefficient engine or an engine out of fuel [9, 10].

In addition to providing useable energy, the anabolic functions of metabolism are important for maintaining the fundamental structure of the heart, with numerous biosynthetic pathways contributing to lipid, nucleotide, and amino acid production [3]. Within the context of anabolism, the analogy of a heart to an engine breaks down—engines do not build or maintain themselves. This realization lays bare an intriguing question: How does the heart, or any tissue, know when to allocate resources for building block synthesis rather than for ATP production? Although this question has more theoretical than empirical answers, the idea that the heart functions as an autopoietic unit [11–15] seems to have conceptual value for understanding metabolite fate decisions (Fig. 1).

Fig. 1: Conceptual models of catabolism and anabolism in the heart.

The heart preserves or directs changes in its structure and function in part by taking free energy from the surroundings, e.g., carbon substrate, to produce ATP via catabolism and to synthesize building blocks (B; e.g., phospholipids, nucleotides, amino acids) via anabolism. In the context of catabolism, the heart is similar to an engine, in that hydrocarbon substrates are oxidized to provide energy for mechanical work. This engine analogy does not hold for anabolic processes, where substrates are used for production of the cellular building blocks that support structural integrity. The simplified equations provided above (right) describe how the generation (GEN) and decay of end products of anabolic reactions may affect tissue homeostasis, growth or hypertrophy, and atrophy or death. Anabolism appears to function as an autopoietic unit, which could describe the heart’s capacity to maintain its structure and function under a given set of conditions.

To obtain gainful knowledge related to catabolism, anabolism, and their interactions, we consider here types of data gathered from past and current methods. Methods to assess metabolism in cultured cells and tissues range from respirometry, where sometimes only a single readout (e.g., oxygen consumption) is measured, to metabolomics-based approaches that measure the abundances and characteristics of hundreds to thousands of metabolites. Nevertheless, there are numerous considerations for measuring metabolic processes as well as for how to interpret the vastly different types of data generated from distinct experimental systems.

3. Considerations for measuring metabolism in isolated mitochondria and cultured cells

3.1. Respirometry using isolated mitochondria:

Purified mitochondrial fractions can be used to understand specific aspects of mitochondrial biology, including respiratory efficiency and capacity [16–18]. Typically, respirometry uses oxygen electrode- or fluorophore-based systems to measure the consumption of O₂ as an index of catabolic activity in isolated mitochondria, cells, tissues, or whole organisms. Because the vast majority of O₂ in most cells is consumed at cytochrome c oxidase, the rate of O₂ consumption is used to understand the functionality of the electron transport chain in the context of ATP or heat production.

Common parameters derived from respirometry include State 3 and State 4 respiration rates as well as the respiratory control ratio (RCR; the ratio of the State 3 and State 4 rates) [16, 18, 19]. Respirometry experiments designed to obtain these values involve the ordered addition of substrates, ADP, and pharmacological compounds and define how mitochondria use particular substrates and how substrate utilization, proton transport, and O₂ consumption relate with coupled oxidative phosphorylation. Isolated mitochondria preparations can also be used to identify sites of mitochondrial damage [20] and to study redox signaling and calcium handling processes [21, 22]. In general, only freshly isolated mitochondria should be used to assess respiratory function from the heart, although recent protocols could bypass some loss of function caused by a freeze-thaw cycle [23].

Despite their many advantages, isolated mitochondrial preparations have several limitations. For example, standard techniques such as differential centrifugation may retrieve a low fraction of the total mitochondrial content. In highly active tissues (e.g., striated muscle), this shortcoming could lead to bias due to selective representation of the mitochondrial pool [24–26]. Moreover, the process of isolation could decrease mitochondrial integrity because it strips the mitochondria from their intracellular locale [24, 27], which can change mitochondrial morphology, alter mitochondrial respiration, sensitize mitochondria to permeability transition pore opening, and increase reactive oxygen species production [26, 27]. Potential loss of mitochondrial integrity during isolation is also important to consider given recent studies that suggest that membrane potential conduction via the mitochondrial reticular network is a major mechanism for energy distribution [28]. Last, respiration studies using isolated mitochondria typically involve millimolar concentrations of substrates, which are orders of magnitude above their levels in intact cells or tissues. Nonetheless, meaningful data that attest to mitochondrial health can be obtained after optimization of mitochondrial isolation and respirometry assays.

3.1.1. Instrumentation:

Instrumentation to measure respiration typically settles into two bins: high throughput respirometry platforms such as the Seahorse Extracellular Flux (XF) analyzer and lower throughput, but higher resolution platforms such as the Oroboros O2K FluoRespirometers. The fundamental advantage of XF analyzers is their capacity to examine respiration in up to 96 samples at a time; however, a biological “n” of 96 is usually not practical because of bottlenecks in mitochondrial isolation and the need for technical replicates. Unless facilitated by higher throughput mitochondrial isolation methods, mitochondrial extraction from tissue can be slow enough to limit the number of biological replicates per XF assay (e.g., to ~4-8 biological replicates per assay). Thus, despite the fact that mitochondria can be isolated in less than 1 h, the process is usually not automated enough to allow simultaneous isolation of numerous biological replicates per group. Protocols using commercial tissue dissociators may improve isolation throughput [29]. While there may not be a rigid rule on how long mitochondria retain their function after isolation, we typically use them within 4 h of isolation. In addition, technical replicates (e.g., at least 3 technical replicates per biological “n”) are usually built into the XF experimental design.

Classical mitochondrial function measurements have been improved upon in recent years. For example, Oroboros Instruments have the capacity to measure only two samples at a time; however, they have high resolution in that they not only measure oxygen consumption, but experiments can be designed to measure reactive oxygen species production, ATP production, pH, Ca²⁺, and mitochondrial membrane potential as well. As such, they are generally considered to have higher resolution than other respirometry platforms. Like XF analyzers, they can be used to measure bioenergetics in not only isolated mitochondria and intact cells, but in permeabilized cells as well.

3.1.2. Experimental considerations and data interpretation:

A mitochondria titration experiment to optimize the amount of mitochondria is an important first step for respiration assays. As shown in Fig. 2, such assays should include assessment of several respiration “states.” O₂ consumption in mitochondria measured in the presence of only inorganic phosphate is termed State 1 respiration (Fig. 2A). Little to no O₂ consumption will occur under this condition; a respiratory rate observed under this condition is usually thought to be caused by the presence of residual substrates in the isolation. Substrates such as pyruvate can then be added to measure ADP-independent respiratory activity, which is generally ascribed to proton leak and termed State 2 respiration. State 3 respiration occurs when ADP is added in the presence of respiratory substrates and is typically characterized by a rapid rate of oxygen consumption and the formation of ATP. Of note, State 2 respiration usually mirrors State 4 respiration, which occurs when ADP is depleted or when inhibitors of ATP synthase, such as oligomycin, are added (Fig. 2A). Using too much mitochondria could cause O₂ to become limiting in microplate-based assays, especially under an uncoupled condition (e.g., upon FCCP addition). For example, Fig. 2A shows that 5 μg of cardiac mitochondrial protein in an XF96e causes the O₂ concentration to fall below 50 mmHg, and depleting O₂ further could limit respiration [30]. As shown in Fig. 2B, we find that, for XF96e analyses, 1.0–2.5 μg protein is adequate for measuring respiration in isolated, adult cardiac mitochondria; however, this may differ slightly between laboratories and should be standardized empirically.

Fig. 2: Considerations for respiration measurements in isolated mitochondria.

Examples of data from isolated murine cardiac mitochondria experiments in a Seahorse XF96e analyzer: (A) O₂ traces in isolated mitochondria provided sequentially with no substrate (St. 1 respiration); pyruvate (5 mM), malate (2.5 mM), and ADP (1 mM) (PMA; St. 3 respiration); oligomycin (1 μM; St. 4 respiration); FCCP (1 μM); and antimycin A+rotenone (AA/Rot; 10 μM/1 μM); (B) oxygen consumption rate (OCR) in the absence (State 3) or presence of oligomycin (State 4) as a function of amount of mitochondrial protein used for the assay (GMA, glutamate+malate+ADP; PMA; SRA, succinate+rotenone+ADP; or OMA, octanoylcarnitine+malate+ADP); (C) pH traces after addition of PMA, Oligo, FCCP and AA/Rot; (D) State 3 and State 4 respiration rates in mitochondria provided with malate+ADP (MA), PMA, or OMA; (E) ECAR values from isolated mitochondria provided with the indicated substrates in the absence or presence of oligomycin; and (F) Schematic illustrating alkalinization of medium caused by pyruvate-supported respiration compared with respiration supported by fatty acyl substrates.

When pyruvate is used as a substrate, notable alkanization of the medium may occur during the assay (decreases ECAR) (Fig. 2C–2E). The alkalinization that occurs with pyruvate is thought to be due to H⁺ symport through the mitochondrial pyruvate carrier [31] (Fig. 2F). Thus, in addition to attention to O₂ levels, it is important to assess pH changes that may occur during the assay and make protocol adjustments should pH levels fall outside of the biological range.

Note that malate must be included in respiration assays, especially when pyruvate is used as a substrate. Adding relatively low concentrations of malate (e.g., 0.5 mM) allows for oxaloacetate production and condensation with acetyl CoA, thereby preventing acetyl CoA accumulation and pyruvate dehydrogenase inhibition [32]; however, mitochondria will not respire avidly on malate and ADP alone, as shown in Fig. 2D. It is advantageous to include malate also when glutamate or acylcarnitines are provided as substrates [33, 34]. If succinate is provided as the sole substrate for respiration, malate is not required, but it is advisable to include rotenone because it prevents reverse electron transport to Complex I from the ubiquinone pool, thereby limiting reactive oxygen species formation [35, 36]. A lower RCR is generally consistent with poorer coupling between proton translocation and ATP production.

3.2. Cellular measurements using extracellular flux analysis:

Extracellular flux (XF) analysis assesses mitochondrial and glycolytic metabolism in cultured cells or isolated tissues by measuring O₂ consumption and extracellular acidification in the medium. It has also been used to measure the NADPH oxidase-mediated respiratory burst in neutrophils [37–39]. In addition to Seahorse XF platforms, several commercially available plate-based assay systems can measure mitochondrial or glycolytic activity, and the assays can be complexed with pharmacological agents to define a metabolic profile [18, 40, 41]. Unlike traditional respirometry, XF analysis is typically used to measure metabolism in adherent cells [18, 19, 40, 41] but is adaptable to other preparations (e.g., spheroids [42], pancreatic islets [43], tissue explants (e.g., [44–46]) and small organisms such as C. elegans, planaria, and zebrafish embryos (e.g., [47–49]). Similar to traditional respirometry, XF analysis only provides information related to catabolism. It can be used with intact and permeabilized cell models, as outlined below.

3.2.1.1. Instrumentation:

The minimum instrumentation required for XF analysis is a plate reader, although now widely available are Seahorse XF analysis and RESIPHER assay systems. Whereas Seahorse XF systems measure mitochondrial activity and glycolysis in transient microchambers, the RESIPHER system measures dissolved oxygen concentration gradients formed by metabolically active, adherent cells in culture, which allows calculation of oxygen consumption rate. Oxygen consumption and extracellular acidification assays that require only a fluorescence plate reader (e.g., MitoXpress, pH-Xtra) usually require soluble phosphorescent or fluorescent probes that can be dispensed directly in the test sample (for extensive discussion, see [50, 51]).

3.2.1.2. Cell models:

Intact, cultured cells can provide a useful experimental system for understanding cardiomyocyte metabolism and the effects of genetic differences, gene manipulation, or drug treatment. An example is human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs), which provide a renewable supply of human-derived cells for studying cardiovascular disease mechanisms and treatments [52–54]. Because diffusion is adequate for nutrient supply, the experimenter has full control over the cell environment. This includes control over not only the levels of substrates in the culture medium, but also over hormone levels, oxygenation, and experimental interventions; however, a conspicuous disadvantage of using isolated cardiomyocytes as well as other cells from the heart (e.g., endothelial cells, fibroblasts, smooth muscle cells) is that they may not retain their in vivo metabolic characteristics. For example, adult primary cardiomyocytes typically lack contractile activity, which derives from the use of myosin inhibitors during and after isolation. Inhibition of contraction promotes a state of low energy demand which compromises understanding of the in vivo metabolic phenotypes of cardiomyocytes. Upon oligomycin addition to isolated adult cardiomyocytes, it is common to observe little or no change in the oxygen consumption rate (e.g., see [55, 56]), which indicates that the majority of O₂ consumption in isolated adult cardiac myocytes is attributable to proton leak. This disadvantage limits understanding of the extent to which oxidative phosphorylation meets the ATP demands of contracting adult cardiomyocytes. Nevertheless, adult cardiomyocyte respiratory values may be useful for delineating changes in maximal respiratory capacity and in bioenergetic responses to stress [19].

Another potential limitation of cardiac cell culture systems (e.g., neonatal cardiac myocytes, iPSC-CMs) is that they may present a relatively immature phenotype, which is associated with differences in sarcomere organization [57] and a higher reliance on glycolysis rather than fatty acid oxidation for energy, compared with adult cardiomyocytes [58, 59]. Transformed cell models, such as H9C2 and HL-1 cells, are further removed from the in vivo scenario. They are highly proliferative, have markedly different sarcomeric structure, demonstrate lower rates of fatty acid oxidation, and are more reliant on glycolysis for energy (Table 1). Although these model systems could reduce variability and the number of experiments required to address a scientific question, their limitations must be carefully weighed to prevent overgeneralizations to in vivo biological phenomena.

Table 1:

Common cell models used to examine cardiac metabolism.

Cell model	Tissue origin	Species	Type	Advantages	Limitation	General metabolic phenotype	References
H9C2	Ventricle cardiomyocyte (derived from embryonic BDIX rat heart)	Rat	Transformed	• Easy to propagate, passage, and revive from cryopreservation • Simple medium composition • Capable of limited differentiation	• Non-beating • Low mitochondrial content and capacity for fatty acid oxidation	• Undifferentiated cells: highly glycolytic • Differentiated cells: more metabolically mature (oxidative) than undifferentiated cells • Higher ATP content and more sensitive to hypoxia-reoxygenation injury compared with HL-1 cells	[157–161]
HL-1	Atrial cardiomyocyte (derived from AT-1 mouse atrial cardiomyocyte tumor cell line)	Mouse	Transformed	• Easy to propagate, passage, and revive from cryopreservation • Spontaneously contracting at optimal culture conditions	• Readily loses contractile activity • Requires highly specialized medium • Retains characteristics of tumor lineage	• Higher glycolytic and lower oxidative capacity than H9C2 and primary cardiomyocytes. • Lower mitochondrial mass and respiration than H9C2 cells • Less sensitive to hypoxia-reoxygenation injury compared with H9C2 cells • Can be rendered insulin resistant simila to primary cells	[159, 162–166]
NRCMs and NMCMs	Ventricle cardiomyocyte	Newborn rat/mouse	Primary	• Spontaneous beating (within 2-3 d of culture) • Easier isolation compared with adult cardiomyocytes • Responsive to electrical stimulation	• Limited propagation • Low fatty acid oxidation capacity and immature mitochondrial networks	• More glycolytic than adult myocytes • Electrical stimulation increases metabolic rate • Resistant to hypoxia-induced apoptosis, attributed to high glycolysis	[167–172]
Adult mouse CMs	Ventricle or atrial cardiomyocyte	Mouse	Primary	• More similar to human adult cardiomyocyte than NMCMs or transformed cells • Responsive to electrical stimulation • Helpful for studying some metabolic processes • Wide availability of genetic models	• Most preparations do not beat spontaneously • Loss of contraction-induced ATP demand • Short viability of rod-shaped cells culture • Isolation can be technically challenging	• More reliant on oxidative phosphorylation than NMCMs • Mature mitochondrial networks and dynamics • Electrical stimulation increases metabolic rate • Can be used to model fatty acid-induced metabolic inflexibility and insulin resistance	[167, 172–174]
Adult rat CMs	Ventricle or atrial cardiomyocyte	Rat	Primary	• Higher yield than mouse CM • More similar to human adult cardiomyocyte than neonatal rodent cardiomyocytes • Responsive to electrical stimulation • Helpful for studying some metabolic processes • Can be isolated from different areas of the heart	• Most preparations do not beat spontaneously • Loss of contraction-induced ATP demand • Short viability of rod-shaped cells culture • Isolation can be technically challenging	• Mature mitochondrial networks and dynamics and appreciable oxidative metabolism • Electrical stimulation increases metabolic rate • Can be used to mode fatty acid-induced metabolic inflexibility and insulin resistance	[167, 172, 175–181]
AC-16	Ventricular cardiomyocyte fused with SV40 transformed fibroblasts	Human	Transformed	• Easy to propagate, passage, and revive from cryopreservation • Retain a cardiac transcriptional program producing contractile and cardiac-specific proteins • Can be differentiated to a more non-proliferative state • Simple medium composition	• Non-beating • Lack some contractile components such as sarcomeres and T-tubules	• Limited literature addressing metabolism; however, AC-16 cells appear to rely on both glycolytic and oxidative metabolism	[182–184]
RL-14	Ventricular cardiomyocyte	Human (fetal)	Transformed	• Has been used in drug metabolism and hypertrophy studies, and in coculture studies • Can be useful as an immortalized human cell line for studying metabolism	• Non-beating • Limited literature addressing their metabolism	• Likely to have a fetal glycolytic metabolic profile	[185–188]
Primary hCMs	Cardiac ventricular	Human cardiac tissue biopsies	Primary	• Representative of the human heart	• Difficult to acquire • Yield may be low (depending on amount of starting material) • Biopsy may represent only distinct areas of the heart	• Few studies have addressed metabolism in primary human myocytes	[189, 190]
hESC-CMs	Stem-cell derived cardiac cells (embryonic cells)	Human	Stem cell-derived	• Human-derived • Spontaneous contraction upon differentiation • Expression of cardiac sarcomeric markers • Can be utilized to study insulin resistance	• Immature metabolic phenotype	• Likely to rely on glycolysis more so than myocytes in the adult heart	[191, 192]
iPSC-CMs	Stem-cell derived cardiac cells (Dedifferentiated somatic cells; fibroblasts)	Human	Stem cell-derived	• Retain patient-specific genetics • Initial collection of cells is minimally invasive (e.g., skin biopsy, blood) • Some properties similar to primary CM • Spontaneous beating • Responsive to electrical stimulation • Used widely in drug discovery and regenerative therapy studies	• Can be challenging to grow and differentiate • Depending on culture conditions and duration, may have a spectrum of metabolic phenotypes	• May display fetal CM-like characteristics (e.g., higher reliance on glycolysis) • Structural and metabolic maturation can be induced when provided oxidative substrates	[191, 193–196]

^*Abbreviations: NRCMs, neonatal rat cardiomyocytes; NMCMs, neonatal mouse cardiomyocytes; hESC-CMs, human embryonic stem cell-derived cardiomyocytes; iPSC-CM, human induced pluripotent stem cell-derived cardiomyocytes.

3.2.1.3. Experimental considerations:

Prior to bioenergetic profiling, it is critical to identify an optimal number of cells that provides oxygen consumption and extracellular acidification rate (OCR and ECAR, respectively) values in the measurable and interpretable range. Moreover, to interrogate metabolic differences, it is important to titrate pharmacological compounds to maximize their effectiveness and minimize off-target effects. Details regarding these steps are found below (see also: [16, 18, 19, 40, 41, 60]).

Step 1, Cell seeding:

An optimized seeding density ensures that OCR and ECAR values fall within an operative range. Initial experiments should include a range of cell densities and examination of relationships between cell number, O₂ concentration, and extracellular acidification (pH) (Fig. 3A, 3B). As shown in Fig. 3C, human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) show a linear OCR up to 25,000 cells per well in the XF96e instrument, whereas ECAR values increase progressively with cell number. The ratio of the OCR to ECAR values (OCR:ECAR) can provide an index of the reliance of cells on oxidative phosphorylation versus glycolysis. As shown in Fig. 3D, the lowest (5,000 cells/well) and the highest (50,000 cells/well) number of iPSC-CMs show relatively low OCR:ECAR ratios. Too few cells could render OCR values below the sensitivity level of the instrument, and too many cells could diminish local concentrations of O₂ or change cell phenotype in a manner that promotes higher reliance on glycolysis. For XF96e analyzers, it is generally advisable for basal OCR rates to be >20 pmol/min, and for XF24e analyzers, basal OCR rates are recommended to be >50 pmol/min.

Fig. 3: Experimental considerations for initial extracellular flux assays.

Examples of extracellular flux data from human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). Different numbers of iPSC-CMs were seeded in an XF96e plate, followed by measurements of: (A) O₂ concentration; (B) pH; (C) basal OCR and ECAR rates; and (D) OCR:ECAR ratio. To standardize oligomycin (E) and FCCP (F) concentrations, different concentrations of each compound were applied to iPSC-CMs followed by OCR measurements. Note the relatively narrow dose response curve of FCCP, which can shift depending on cell type or number and media composition.

Step 2, Titration of pharmacological compounds:

Pharmacological compounds that target specific aspects of metabolism are commonly used to interrogate mitochondrial activity. These drugs should be titrated to determine optimal concentrations, which can differ based on cell type or number. An optimal drug concentration is defined as the lowest concentration that promotes a maximum response, which should minimize off-target effects. Oligomycin A (a mitochondrial ATP synthase inhibitor) and carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP; an uncoupler of oxidative phosphorylation) are common compounds used in the mitochondrial stress test (see below). As shown in Fig. 3E and 3F, oligomycin and FCCP have different actions, with oligomycin inhibiting oxygen consumption and FCCP increasing oxygen consumption. As with most uncouplers, FCCP responses show a bell-shaped curve, indicating a relatively narrow operating range.

3.2.2. The mitochondrial stress test:

Once cell number and compound concentrations are optimized, a common next experiment is to develop a bioenergetic profile of mitochondrial metabolism. This mitochondrial “stress test” addresses five basic questions (Fig. 4A):

Fig. 4: Fundamentals of the mitochondrial stress test.

Strategic administration of inhibitors of mitochondrial activity and uncouplers can provide answers to several questions related to the bioenergetic activity of the cell. (A) Questions addressed by sequential administration of oligomycin, FCCP, and antimycin and rotenone. (B) Example of a bioenergetic profile in iPSC-CMs. (C) A simplified proton circuit that helps to understand how each inhibitor impacts mitochondrial activity.

1. What is the basal level of respiration in cells?

2. What is the compositional nature of cellular respiration, i.e., how much oxygen is linked to ATP production versus proton leak?

3. What is the maximal respiratory capacity of cells?

4. How much bioenergetic reserve (spare respiratory) capacity do cells have?

5. How much residual, non-mitochondrial oxygen consumption occurs in cells?

The strategic use of pharmacological inhibitors of oxidative phosphorylation and of uncouplers provide useful tests that address these questions and lead to the acquisition of a bioenergetic profile (Fig 4B). As shown in Fig. 4C, a proton circuit helps to visualize sites and processes that control oxygen consumption in cells and tissues. The basal OCR in intact cells represents the aggregate of all cellular processes capable of consuming O₂, with the majority of O₂ reduction occurring at cytochrome c oxidase. Upon addition of oligomycin, an inhibitor of mitochondrial ATP synthase, O₂ consumption is typically diminished, with the loss of OCR representing the proportion of electron transport chain activity used to generate ATP. The maximal respiration sustainable by cells can be assessed after the addition of a proton uncoupler. An important feature emerging from such assays is the mitochondrial reserve capacity (also called the spare respiratory capacity), which can be calculated by subtracting the basal OCR from the FCCP-stimulated rate [16, 18, 61]. Non-mitochondrial oxygen-consuming processes are then assessed by measuring OCR with inhibitors of the respiratory chain, e.g., rotenone and/or antimycin. The non-mitochondrial OCR can be subtracted from other OCR measurements to elucidate mitochondrial-based oxygen utilization. To make meaningful comparisons, OCR values should be normalized to cell number, total protein, or genomic DNA content at the end of the assay.

3.2.2.1. Interpreting the mitochondrial stress test:

Interpretation of the mitochondrial stress test can be facilitated via an interpretation chart (Fig. 5) and through the use of several resources (e.g., [16, 18, 40, 62]). Values from each metabolic parameter are products of several cellular processes. The compositional nature of basal OCR is provided from the ATP-linked and proton leak values, which are calculated using the oligomycin-sensitive rate. The decrease in OCR after oligomycin addition indicates the amount of oxygen consumed that is related to ATP turnover; a higher value compared with a control condition indicates higher cellular ATP demand whereas a lower value could indicate lower ATP demand, but also could arise from deficits in electron transport caused by mitochondrial damage or lower substrate availability.

Fig. 5: Interpretation tree for the mitochondrial stress test.

The basal rate of oxygen consumption is influenced by several factors, including differences in proton leak across the mitochondrial inner membrane, ATP turnover, non-mitochondrial oxygen consumption, and damage to the respiratory chain. Potential explanations for changes in each parameter can be assessed after sequential addition of oligomycin, FCCP, and antimycin A and/or rotenone.

The oligomycin-sensitive rate, after accounting for the non-mitochondrial OCR, provides an index of proton leak across the inner mitochondrial membrane. The presence of a basal level of proton leak is a general characteristic of all mitochondria [63–65] and is dependent on the polarization state of the mitochondrion. Because ATP synthase inhibition by oligomycin hyperpolarizes the mitochondrial inner membrane, the oligomycin-sensitive value slightly overestimates proton leak and underestimates ATP turnover; however, this small amount of over- and underestimation is typically ignored for semi-quantitative answers [16]. Relative differences in proton leak between control and experimental groups could be caused by differences in uncoupling protein (UCP) activity, mitochondrial integrity, or proton motive force (Δp).

The rate of respiration achieved after addition of uncouplers can be used to discern mechanisms that influence respiratory capacity. In intact cells, uncouplers promote proton translocation across the inner mitochondrial membrane, thereby promoting thermodynamic disequilibrium by dropping the mitochondrial membrane potential. The electron transport chain responds to the diminished proton motive force by increasing electron transport. Reasons for differences in the uncoupler-stimulated rate include differences in electron transport chain integrity, substrate availability, or mitochondrial mass.

There are several considerations for the use of uncouplers in bioenergetic assays. As alluded to earlier, uncouplers typically have a narrow dose-response curve (e.g., see Fig. 3F). Pharmacological uncouplers may also act as protonophores across other membranes, which could disrupt processes in the cytosol. Moreover, in intact cells, the artificial energy demand caused by the uncoupler may not be matched by substrate supply or by other factors controlling metabolism (e.g., cytoplasmic Ca²⁺). For all of these reasons, the maximal respiration rate provided by uncoupler use in intact cells may not provide a true maximal rate of respiration. Furthermore, in some cell types, inhibitors of ATP synthase such as oligomycin may diminish the respiratory rates observed upon addition of uncouplers, leading to an underestimation of the maximal respiratory capacity [66, 67]. Careful titration of uncouplers and attention to potential obfuscation of uncoupled respiration rates by oligomycin enable acquisition of a quantifiable index of the mitochondrial respiratory capacity.

Another index of mitochondrial health provided by the FCCP-stimulated rate is the bioenergetic reserve capacity, also called the spare respiratory capacity. This parameter is calculated by subtracting the rate of respiration achieved after addition of an uncoupler from the basal mitochondrial respiration rate. Conceptually, this parameter is useful for understanding how near their bioenergetic maxima mitochondria operate under experimental conditions. In general, the bioenergetic reserve capacity can be diagnostic of the ability of a cell to meet the energetic demands of a stressful condition. In isolated neonatal and adult cardiomyocytes, the bioenergetic reserve capacity is indicative of the ability to provide the increased energy required for maintenance of cellular function, detoxification of reactive species, and repair under conditions of stress [19, 61].

3.2.3. Beyond the mitochondrial stress test:

Although the mitochondrial stress test reports important bioenergetic information, it is not designed to provide information on differential fuel utilization or to address granular questions related to the causes of respiratory changes. To gain additional information, assays may be designed to provide information on substrate selectivity and fuel flexibility or to identify the causes of respiratory differences.

3.2.3.1. Mitochondrial fuel assessment assays:

Most XF experiments are designed with a standard medium containing glucose, pyruvate, and glutamine as substrates; however, substrates may be added one at a time to understand which fuels support basal and maximal respiratory capacity [68]. Fatty acids conjugated to albumin can also be included in the medium, which is important given the known reliance of the adult heart on fatty acids for ATP production. Combinations of substrates can be applied to provide additional information regarding fuel flexibility and catabolic substrate integration. For example, coordinate exposure of cells to first a mitochondrial substrate followed by addition of glucose can be used to discern Crabtree-like effects; the converse strategy, i.e., provision of glucose followed by mitochondrial substrates, could help delineate the effect of mitochondrial metabolism on glycolysis [68]. Inhibitors of key steps in metabolism can provide information on the reliance of mitochondria on glutamine, glucose, and fatty acid oxidation. For example, etomixir, a carnitine palmitoyltransferase inhibitor, can be used to delineate cellular reliance on glucose versus fatty acid oxidation [69]. Moreover, inhibitors of mitochondrial pyruvate uptake (e.g., UK5099) or glutaminolysis (e.g., BPTES) help identify the extent to which cells rely on pyruvate or glutamine to support respiration [70, 71]. Rates of substrate oxidation derived from metabolic tracing experiments (see below) can be used to complement XF fuel utilization assays.

3.2.3.2. Permeabilized cell assays:

Although intact cell systems provide useful information on how metabolism changes with disease or experimental interventions, these models may not provide the ability to identify site-specific changes such as those caused by damage to components of the respiratory chain. Isolated mitochondria provide this ability; yet, they have the potential limitations of relatively poor yield, deleterious changes to mitochondrial architecture and function, and selection bias. Permeabilized cell preparations bridge the gap between intact cell bioenergetic measurements and isolated mitochondria experiments. Such permeabilized cell or cardiac fiber assays, described in detail previously [32, 72–77], have been used to identify novel metabolic actions of thiazolidinediones [71], understand the influence of sex differences [78], diabetes [79], and cardiovascular disease [80, 81] on mitochondrial function, and to develop deeper insights into mitochondrial ion regulation [82–84]. The assay involves permeabilizing the plasma membrane using saponins, digitonin, or recombinant perfringolysin O, which when used at appropriate concentrations, permeabilize the plasma membrane without affecting mitochondrial membranes [32, 76]. Mitochondrial substrates can then be provided to assess mitochondrial respiration. In addition, other parameters, such as ADP sensitivity can be assessed by varying the concentration of ADP provided to the permeabilized cells.

Cell seeding density and permeabilizer concentrations are critically important for obtaining clean results, as outlined in Salabei et al. [32]. Because substrates are typically injected at saturating levels, it is common that, upon their application to permeabilized cells, oxygen consumption may reach levels considered too high for the XF platform and could limit O₂ availability during the measurement, which could affect the OCR. Permeabilizers such as saponin and digitonin are usually provided at ranges between 10-50 μg/ml, whereas recombinant perfringolysin O is typically useful in the 1-5 nM range. Optimization of permeabilizing agent concentrations should be standardized to ensure that mitochondria remain undamaged [32]. Respiratory responses to exogenous cytochrome c can be used to quickly assess permeabilizer-induced damage to mitochondrial membranes, if any (as in [60, 71]). Under optimal conditions, it is common to obtain RCR values that are double or triple that found with respirometry measurements in isolated mitochondrial preparations; inclusion of bovine serum albumin is important for maintaining well coupled mitochondria after permeabilization [32].

3.2.4. Glycolysis stress test:

Although the normal heart relies predominantly on fatty acid oxidation for energy, pathological conditions such as heart failure can increase reliance on glycolysis [3, 8, 85]. Basal glycolytic rates, maximal glycolytic capacity, and glycolytic “reserve” can be measured using a glycolysis stress test. Forming the basis of this assay is the extrusion of protons produced initially from reactions in the preparatory phase of glycolysis (Fig. 6A). The protons generated during glycolysis are then extruded via symport with lactate and are detected by extracellular flux analysis as a transient decrease in pH. A more detailed accounting of the reactions in Fig. 6B are described in Divakaruni et al. [40].

Fig. 6: Schematic of proton production during glycolysis.

Proton-producing and - consuming reactions in glycolysis: (A) The conversion of glucose to glucose 6-phosphate by hexokinase (Hk) and fructose 6-phosphate to fructose 1,6-bisphosphate by phosphofructokinase (PFK) promote proton formation. Importantly, protons are also released during the ATP hydrolysis reactions at these enzymatic steps. Protons are also generated at the step catalyzed by glyceraldehyde 3-phosphate dehydrogenase (GAPDH), where the phosphorylation of glyceraldehyde 3-phosphate to form 1,3-bisphosphoglycerate promotes release of a proton from inorganic phosphate. Reactions at pyruvate kinase (PK) and lactate dehydrogenase (LDH) are alkalinizing: they consume 4 protons, leaving 2 net protons produced during the generation of 2 lactate molecules from 1 glucose molecule. These protons are then extruded via symport with lactate and are detected by extracellular flux analysis as a transient decrease in pH. Of note, the formation of ATP during the phosphoglycerate kinase reaction does not consume protons. Although the same is true for the pyruvate kinase reaction, subsequent nonenzymatic tautomerization of pyruvate consumes two protons. (B) Net reactions related to specific steps in glycolysis. More detailed analysis of the proton-producing and -consuming steps of glycolysis can be found in Divakaruni et al. [40]

It is important to note that although glycolysis accounts for a major portion of protons extruded from the cell, protons are also generated during respiration via the production of CO₂, which is converted to HCO₃⁻ + H⁺. Thus, the total rate of extracellular acidification will exceed that caused by glycolysis alone; however, methods for delineating the contributions of these different sources of protons have been published [86] and implicate use of mitochondrial oxygen consumption rates to correct for CO₂-derived proton production. Furthermore, glycolytic rates can be verified by secondary methods, such as extracellular lactate measurements or by using [5-³H]-glucose tracer assays.

A common protocol for the glycolytic stress test involves switching assay medium to a medium without glucose and measuring basal glycolytic activity after injection of glucose. Then, oligomycin is added to inhibit mitochondrial ATP synthesis, which shifts the responsibility for ATP production solely to glycolysis. After recording oligomycin-stimulated rates, inhibitors of glycolysis such as 2-deoxyglucose are used to help delineate the contribution of glycolysis to ECAR (Fig. 7A). In some cell types, the rate after oligomycin addition can be deemed the maximal glycolytic capacity, with the difference between maximal capacity and basal glycolysis being a measure of the glycolytic reserve; however, as detailed in Mookerjee et al. [87], the oligomycin-responsive ECAR may not be indicative of the maximal glycolytic rate. Replacement of oligomycin with inhibitors of complex I (e.g., rotenone) and complex III (e.g., antimycin A) can sometimes increase glycolytic rates beyond that of oligomycin because it allows for an increased rate of ATP hydrolysis via the reversal of the F₁F_O-ATP synthase; the increase in ATP demand then could augment glycolytic rates beyond that caused by inhibition of mitochondrial ATP synthesis alone. Furthermore, ATP demand can be increased further by addition of monensin, which increases Na⁺ import, stimulates ATP hydrolysis via the Na⁺/K⁺-ATPase, and further increases glycolytic rate [87].

Fig. 7: Examples of glycolytic extracellular flux assays.

(A) Example of a glycolysis stress test in isolated neonatal rat cardiac myocytes. In this assay, the cells are deprived of glucose for 1 h followed by measurement of glucose-free ECAR readings. Glucose is then injected to help discern basal glycolytic rates. Then, oligomycin is injected to assess a putative maximal glycolytic rate, after which 2-deoxyglucose (2-DG) is injected to help discern non-glycolytic ECAR rate. (B) In some cell types, glycolysis may temporarily increase beyond that of the basal rates upon glucose re-introduction. It can be important to determine the kinetics of these changes prior to designing glycolysis stress tests, such as that shown in panel A. The cells shown in panel B derive from a subpopulation of mesenchymal cells isolated from murine heart.

Another potential confounding issue with glycolysis stress test assays is the acute nature of glucose addition. As shown in Fig. 7B, some cell types that are highly reliant on glycolysis, such as mesenchymal cell subpopulations from the heart [68], may require time to reach a stable rate of extracellular acidification following glucose addition. The period of glucose starvation prior to the assay (approximately 1 h) could promote a state of energy starvation in highly glycolytic cells that augments energy requirements. Thus, in certain cell types, it may be important to first examine the time-dependent changes in ECAR after glucose addition to determine the optimal time point for injection of inhibitors (e.g., oligomycin), or to use glucose-replete medium from the beginning of the assay.

4. Considerations for other metabolic assays in cultured cardiomyocytes

Other techniques in the investigator’s metabolic toolbox, discussed briefly below, complement XF data and provide additional information on catabolic pathway flux, redox state, and biosynthetic pathway activity (e.g., see: [88, 89]).

4.1. Targeted analyte measurements:

Assays that measure absolute abundances of metabolites are useful for assessing changes in catabolic and anabolic pathways. These assays may be conducted via spectrophotometric-, fluorometric-, ELISA-, NMR, LC, or mass spectrometry-based techniques and can provide data relevant to both catabolism and anabolism. The simplest example is extruded lactate, which can complement other glycolytic activity measurements. Measurements of end products of other pathways may also be useful. For example, glycogen is commonly measured to gain understanding of glucose utilization and storage under experimental conditions [90–92]. Moreover, measurements of adenine nucleotides (AMP, ADP, ATP) or metabolic intermediates and end products (e.g., lactate/pyruvate) can provide useful values for assessing the energy or redox state of cells or tissues.

4.1.1. Experimental considerations and data interpretation:

Although interpretation of extruded metabolite data is relatively uncomplicated, understanding how substrate levels change in the medium can be further helpful. For example, for assessing changes in glycolysis, determining the relative amount of glucose used in the medium complements extruded lactate measurements and can provide indications of how much glucose is catabolized via glycolysis and how much is likely fated for other metabolic pathways [93]; however, many cells, especially adult cardiac myocytes, also consume lactate [94–98]. Thus, it is possible that extruded lactate may reenter the cell and be utilized for energy. In general, metabolite measurements from cell extracts should be interpreted with the appreciation that intracellular metabolite levels are dependent on not only nutrient uptake and metabolite secretion rates, but also on nutrient contributions to biosynthetic pathways. Because differences in intracellular metabolite concentrations may be due to changes in the activity of not only metabolite-producing enzymes but also metabolite-consuming enzymes, changes in metabolite concentrations may not readily allow strong conclusions on metabolic fluxes. A relative increase or decrease in pool size does not necessarily indicate directionality in terms of a change in flux.

Another consideration is sample preparation. Measurements of metabolites with high turnover rates (e.g., ATP) or of metabolites in pathways with high flux rates (e.g., glycolysis) are especially sensitive to changes in substrate and oxygen availability. In cell culture, it is a common practice to collect cells by centrifugation; however, this may obfuscate metabolite quantification because substrate may become limiting and oxygen tension lower in the local environment of the cell pellet. This problem may be circumvented by rapid quenching of metabolism in the cell culture dish [88, 99–102].

4.2. Intracellular fluorescent metabolite reporters:

Advances in intracellular metabolite probes have provided new ways of visualizing changes in critical cofactors and metabolites in living cells. RNA-, fluorescence-, or FRET-based probes have been used to measure intracellular concentrations of several metabolites, including glucose, glucose-derived metabolites, adenosine, amino acids, and purine nucleotides (e.g., ADP, ATP, GDP, GTP) [103–112]. The addgene website (https://www.addgene.org/) is a useful resource for currently available metabolite reporters. In addition, recent advances in pyridine nucleotide probes have enabled quantification of intracellular levels of NAD(P)⁺ and NAD(P)H in the cytoplasm, nucleus, and mitochondrion. Genetically encoded NADH-sensitive (e.g., SoNar, Rex-YFP, Peredox-mCherry) [113–115], NAD⁺-sensitive (e.g., FiNad, LigA-cpVenus) [116–118], and NADPH-sensitive (e.g., iNAP) [119, 120] fluorescent biosensors introduced into cells via transfection or viral delivery provide the opportunity to track changes in pyridine nucleotides over time in the cytosol and in targeted organelles via fluorescence microscopy. Because fluorescent metabolite reporters measure metabolic intermediates and end products, as well as provide a relative understanding of energy charge and cofactors required for both energy provision (NADH, ATP) and building block synthesis (NADPH), they can be used to complement measurements of both catabolic and anabolic metabolism.

4.2.1. Experimental considerations and data interpretation:

Each biosensor has its own set of limitations and considerations. The reader is directed to the following reviews and protocols for a broader discussion of fluorescent metabolite probe applications and methodological tips [108, 120–123]. In general, many probes are ratiometric and thus require normalization to a fluorescent signal not perturbed by changes in the metabolite of interest. Moreover, many probes are subject to pH artifacts, which must be considered and addressed experimentally. In general, probes with the favorable characteristics of intense fluorescence, a large dynamic range, pH insensitivity, and quick responsiveness appear to be of most value. An issue to consider with contracting cultured cells (e.g., iPSC-derived cardiomyocytes, neonatal cardiomyocytes) is the potential movement of the imaged area in and out of the microscope focal plane, which can be overcome by ratiometric correction using values from co-expressed fluorescent controls.

4.3. Metabolic tracing:

Isotopic labeling is a common way to measure metabolic fluxes in cells and tissues. For cell culture studies, substrates containing rare isotopes (e.g., ³H, ¹³C, ¹⁴C) are provided in the culture medium or perfusate, which leads to isotopic labeling of metabolites and by products. The simplest of these assays uses radioactive tracers containing ³H and ¹⁴C. For example, ³H-glucose and ¹⁴C-glucose are commonly used to measure glucose utilization and can provide absolute measurements of glycolytic flux and glucose oxidation (or pentose phosphate pathway flux), respectively. In such experiments, the tracer is added to the nutrient mix and metabolic pathway activity is assessed after measuring levels of extruded or expelled ³H₂O or ¹⁴CO₂ by scintillation counting. Metabolic tracing may also be carried out with non-radioactive isotopes. Incorporation of the label (e.g., ¹³C) can be assessed in a targeted manner by mass spectrometry or NMR. For a more in-depth review of stable isotope labeling strategies to interrogate metabolic flux, the reader is directed to the following excellent reviews [124, 125].

4.3.1. Experimental considerations and data interpretation:

A metabolic steady state, where the intracellular metabolite concentrations and metabolic fluxes of a cultured cell population are relatively constant, is ideal for metabolic tracing. As long as nutrient supply is maintained, the exponential phase of growth is commonly assumed to reflect a near steady state condition [124, 126]. Control of nutrient supply can be achieved with a bioreactor, which adds fresh medium and removes waste products [127, 128]; however, experiments are often performed at a metabolic pseudo-steady state, in which changes in extracellular nutrient and intracellular metabolite concentrations are considered negligible during the experiment.

For radiolabeled tracers, it is important to understand the potential fates of the label and to account for those possibilities. Notable is the potential of the [5-³H]-glucose tracer to overestimate glycolytic flux because of ³H₂O release from the non-oxidative portion of the pentose phosphate pathway [129]; however, under most conditions it may be that non-glycolytic detritiation of [5-³H]-glucose in the PPP is insignificant [130]. Attention to label positioning is important for accurate interpretation, as exemplified from ¹⁴C-glucose tracer studies, where the origin of ¹⁴CO₂ or ¹⁴C-labeled metabolites can depend upon tracer label position [129, 131–133]. Knowledge of relative pathway activity is also important for designing experiments, especially with nonradioactive tracers such as ¹³C-glucose. For example, biosynthetic pathways such as the purine and pyrimidine biosynthetic pathways carry much lower flux than relatively rapid pathways such as glycolysis and the TCA cycle, and may require incubation with the labeled substrate for >12 h to achieve detectable nucleotide labeling [88, 134]. Pathways such as the hexosamine biosynthetic pathway also carry relatively low flux compared with fast pathways such as glycolysis; in the intact heart, its flux has been estimated using stable tracers to be only ~0.003-0.006% of glycolysis [135]. For stable isotope labeling experiments, the use of multiple time points is required to ascertain changes in flux with confidence [124, 125]. As with radiolabeled tracers, the positioning of labels in substrates is important for measuring the activity of distinct pathways [125]. Also, with stably labeled substrates, the detected ¹³C-metabolites must be corrected for natural ¹³C abundance. Assays that complement tracer data, or the addition of pharmacological controls that inhibit or activate the pathway of interest, provide additional confidence to data interpretation [88].

4.4. Unlabeled metabolomics:

The field of metabolomics evolved to gain a more global understanding of metabolism [125, 136–139]. With respect to the heart, metabolomic analyses have contributed to our understanding of the metabolic changes that occur in the heart with genetic manipulation and under conditions of health, disease, and environmental exposure (e.g., see: [140–143]). Typically, metabolomic measurements are made via NMR or mass spectrometry and can include targeted or untargeted approaches. The goal of targeted measurements is to identify and quantify a limited number of metabolites (e.g., 10-100 metabolites), whereas the goal of untargeted approaches is to provide relative quantification of as many metabolites as possible (e.g., hundreds to thousands of metabolites). Because of the breadth of metabolites measured in energy-providing and biosynthetic pathways, metabolomics is useful for assessing changes in both catabolic and anabolic metabolism.

4.4.1. Experimental considerations and data interpretation:

It is important to power the study to increase the likelihood to distinguish a biological effect, but also to prepare samples in a way that minimizes artifacts, quenches metabolic processes, and extracts metabolites. Cell culture studies require attention to media composition and the timing of media changes. It is advisable to quench cell metabolism in the culture dish rather than quenching the cell pellet after centrifugation. Additional guidance on the experimental design for metabolomics of cultured cells can be found in the following reviews [100, 144]. Metabolite extraction is an important consideration and usually involves use of solvents such as acetonitrile and methanol [88, 145, 146]. Metabolite separation, detection, identification, and quantification have their own set of considerations (reviewed in: [125, 137]). Although standardization of analytical procedures is an inherent challenge in metabolomics, published workflows, such as those reviewed in [125, 136, 137], can be used as guides. Moreover, open source software platforms for data processing (e.g., ElMaven [147], XCMS [148], MetSign [149]) and for metabolomic data analysis (e.g., MetaboAnalyst [150]) improve the speed, accuracy, and transparency of data handling, analysis, and interpretation. It is important to consider that the size of metabolite pools is not necessarily indicative of flux through the pool; however, computational models have integrated oxygen consumption data with metabolite abundances to help understand coordination between energy-providing and biosynthetic processes (e.g., [151]).

4.5. Stable isotope-resolved metabolomics (SIRM):

Although unlabeled metabolomics studies can provide evidence of altered metabolic pathways, snapshot concentrations do not clearly resolve pathway flux. The incorporation of stably labeled isotope tracers empowers metabolomics to resolve differences in relative metabolic pathway flux and can be used to assess the contribution of substrates to metabolite pools. For example, provision of ¹³C-glucose (or other labeled nutrients) to cells, tissues, or organisms leads to time-dependent incorporation of ¹³C into metabolic intermediates or biosynthetic end products. The pattern of metabolite labeling reflects shifts in metabolite mass. From such data, temporal differences in isotope enrichment, differences in the labeling pattern, and differences in the contribution of nutrients to catabolic and anabolic pathways can be assessed, which provide intricate knowledge of metabolism in cells or tissues [88, 124, 125, 152, 153]. Moreover, SIRM strategies can be adapted to incorporate label for much longer times, which may be required to measure the activity of relatively slow biosynthetic pathways. Because SIRM can be used to assess label incorporation in biosynthetic and energy-providing pathways, it can be used to gain insight into changes in both anabolic and catabolic metabolism. Nevertheless, SIRM experiments are labor-intensive and require numerous considerations, as discussed briefly below and extensively in [88, 124, 125].

4.5.1. Experimental considerations and data interpretation:

Major considerations for a SIRM experiment include: (1) identifying the pathway(s) of interest and choosing the appropriate stably labeled substrate for introduction into the cell system; (2) choosing appropriate controls, which can range from the required unlabeled controls to optional, but useful, pharmacological controls; (3) understanding the concepts of metabolic steady state and isotopic steady state, and including multiple time points of labeling; (4) adequately quenching and extracting metabolites from cells or tissues; (5) identifying the platform (e.g., GC/MS, LC/MS, NMR) that will be used to measure isotopologues or isotopomers; (6) using specialized software for spectral deconvolution, metabolite assignment, natural isotope abundance correction, and isotopologue/isotopomer assignment and quantification; and (7) interpretation of results and sometimes computational modeling [88, 124, 125]. Relative differences in the use of labeled substrate in particular metabolic pathways may be inferred from incorporation of the isotopic tracer into metabolites during the dynamic and steady state phases of labeling. To quantitate fluxes, more formal ¹³C flux analysis utilizing mathematical modeling can be applied [154–156]. Contributions of substrate to metabolite pools are assessed at isotopic steady state. The reader is directed to the following references for detailed discussions on SIRM [88, 124, 125, 152].

5. Summary:

Sensible application of metabolic assays can facilitate understanding of how metabolic changes influence cardiomyocyte biology. Initial experiments to optimize cell number, substrate media composition, and pharmacological compound concentrations are critical for deriving reproducible and interpretable results. Moreover, complementary metabolic assays can provide additional rigor and depth to experimental results and interpretation. Considerations for the advantages and limitations of each cardiomyocyte cell model and the basis of each metabolic assay can further improve data interpretation and help place the findings in context with coherent mechanisms of cardiac (dys)function and remodeling.

Abbreviations:

ECAR

extracellular acidification rate

FCCP

Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone

GC/MS and LC/MS

gas chromatography and liquid chromatography mass spectroscopy, respectively

hESC-CMs

human embryonic stem cell-derived cardiomyocytes

iPSC-CM

induced pluripotent stem cell-derived cardiomyocytes

TMPD

N,N,N′,N′-Tetramethyl-p-phenylenediamine

NMCMs

neonatal mouse cardiomyocytes

NMR

nuclear magnetic resonance

NRCMs

neonatal rat cardiomyocytes

OCR

oxygen consumption rate

RCR

respiratory control ratio

SIRM

stable isotope-resolved metabolomics

XF24e/XF96e systems

extracellular flux analysis systems for 24-well or 96-well microplates, respectively

Δp

proton motive force