Overview of methods that determine mitochondrial function in human disease

Abstract

Cellular metabolism has a key role in the pathogenesis of human disease. Mitochondria are the organelles that generate most of the energy needed for a cell to function and drive cellular metabolism. Understanding the link between metabolic and mitochondrial function can be challenging due to the variation in methods used to measure mitochondrial function and heterogeneity in mitochondria, cells, tissues, and end organs. Mitochondrial dysfunction can be determined at both the cellular and tissue levels using several methods, such as assessment of cellular bioenergetics, levels of mitochondrial DNA (mtDNA), mitochondrial membrane potential (MMP), mitochondrial reactive oxygen species (mito-ROS), and levels of mitochondrial enzymes. Recent advances involving novel radiotracers in combination with PET imaging have allowed for the determination of mitochondrial function in vivo with high specificity. Understanding the barriers in existing methodologies used to study mitochondrial function may help further establish the assessment of mitochondrial function as a biologically and clinically relevant biomarker for human disease severity and prognosis. Herein, we critically review the existing literature regarding the strengths and limitations of methods that determine mitochondrial function, and we subsequently discuss how emerging research methods have begun to overcome some of these hurdles. We conclude that a combination of techniques, including respirometry and mitochondrial membrane potential assessment, is necessary to understand the complexity and biological and clinical relevance of mitochondrial function in human disease.

1. Introduction

Mitochondria are the organelles that generate most of the energy needed for a cell to function and have additional critical biological functions [1,2]. The primary physiological function of mitochondria is to produce adenosine triphosphate (ATP) through oxidative phosphorylation. Additional functions include generating and catabolizing metabolites, creating and detoxifying reactive oxygen species (ROS), participating in apoptosis, and regulating calcium homeostasis [1,2]. Mitochondrial dysfunction instigates several human diseases, including cardiovascular, bone, and kidney disease [1,2]. The term “mitochondrial dysfunction” is widely used throughout cell biology, but a precise definition is challenging, as different studies use different metrics to delineate dysfunction from mitochondrial homeostasis [3].

Measuring mitochondrial dysfunction can be done at both the cellular and tissue levels. Muscle biopsies are currently the gold standard for determining mitochondrial function at the tissue level, but are invasive, muscle tissue is not readily accessible, and samples must be analyzed immediately [3]. Determination of mitochondrial function in skin grafts has not been extensively validated [4]. Such limitations have led to the development of several methods to assess the function of mitochondria in readily accessible whole-blood peripheral blood mononuclear cells (PBMCs) [5–7]. Measuring cellular oxygen consumption and mitochondrial reactive oxygen species (mito-ROS) in PBMCs in association with clinical outcomes has highlighted the importance of mitochondrial dysfunction in several experimental disease models [6,8,9]. Determinations of mitochondrial membrane potential (MMP), mitochondrial DNA (mtDNA) quantity, cardiolipin levels, and the activity of respiratory chain complexes have also been used to assess mitochondrial function [10–12]. Given that all methods that assess mitochondrial function have limitations, it is unclear if inconsistent data regarding mitochondrial function among different studies is partially related to variability in existing experimental approaches.

Herein, we review all available published methods to date that assess mitochondrial function in human biospecimens. This manuscript aims to critically review the existing literature regarding the strengths and limitations of processes that determine mitochondrial function. While no single method is without limitations, utilizing a combination of techniques such as respirometry, which remains the gold standard, and assessing mitochondrial membrane potential (MMP) can help elucidate mitochondrial function and dysfunction.

2. Overview of the role of mitochondrial dysfunction in human disease and pathology

Mitochondria are organelles with essential functions in almost all mammalian cells, including energy production, generation of reactive oxygen species (ROS), coordination of cellular Ca²⁺ signaling, autophagy, necrotic and apoptotic cell death. Maintenance of healthy mitochondria involves a complex system of quality control, involving degrading misfolded proteins, while damaged mitochondria are renewed by fusion or removed by autophagy. Mitochondria are central in several processes, including normal tissue homeostasis, metabolism, immunity, neurodegeneration, and infectious diseases [13]. Thus, altered mitochondrial function and cell signaling pathways lead to disordered cell function, which manifests as disease. Due to the key role of mitochondria in metabolism, mitochondrial dysfunctions manifest as heterogeneous multisystem diseases and mainly involve tissues dependent on OXPHOS, such as the brain, heart, and skeletal muscle.

Mitochondrial dysfunction can be inherited (primary) through alterations in proteins encoded by the mitochondrial or nuclear genome, or acquired through changes in cell physiology, mitochondrial biogenesis, autophagy, or extramitochondrial signals, such as energy substrates or oxygen, cytokines, intracellular Ca²⁺, or oxidative stress. Mitochondrial DNA mutations, infections, aging, and a lack of physical activity have been identified as major players in mitochondrial dysfunction in multiple diseases [13]. Mitochondrial dysfunction leads to disequilibrium of these cellular processes and ultimately alterations in tissue and end-organ functions, including liver and skeletal muscle metabolism, β-cell insulin production, and neuronal function [14]. Thus, mitochondrial dysfunction contributes to the pathogenesis of several diseases, such as genetic/primary mitochondrial diseases (reviewed elsewhere, [13]), neurodegenerative, muscle, and cardiovascular diseases, cancer, and diabetes [15,16]. Indeed, alterations in mitochondrial OXPHOS (oxidative phosphorylation) are linked to tumor formation, while changes in mitochondrial trafficking and mitochondrial mass through changes in turnover (mitophagy or altered biogenesis) are linked to neuropathic disease.

While mtDNA-encoded proteins are linked with disease, most mitochondrial proteins are encoded in nuclear DNA and have been increasingly linked with human disease [17]. Mitochondrial dysfunction has been implicated in almost all of the major neurodegenerative and neuroinflammatory diseases, including Alzheimer’s disease. Mitochondrial dysfunction is one of the hallmarks of aging [18]. However, most data linking defective mitochondria with aging are correlative, and the role of mitochondrial defects as a cause of aging remains controversial. Notably, animals that accumulate mitochondrial mutations at a high frequency appear to age prematurely, developing age-related diseases early, and losing their hair prematurely [19,20]. Mitochondrial dysfunction can also contribute to cardiovascular disease [21] in the heart [22] and vascular tissue, contributing to atherosclerosis [23,24]. Impaired mitochondrial permeability transition has been implicated in several human diseases such as ischemia-reperfusion injury, ischemia of the CNS [25], and pancreatitis [26]. Mitochondria are also involved in metabolic reprogramming through decreased OXPHOS, increased glycolysis and lactate production, leading to carcinogenesis and the development of hallmarks of cancer, such as limitless proliferation and evasion of apoptosis induced either by stress (hypoxia, oxidative stress, or nutrient depletion) or by chemotherapeutic agents [27–30]. Mitochondrial dysfunction in β-cells, liver, and skeletal muscle also contributes to the pathogenesis of diabetes [31]. Emerging technologies have also allowed studies of mtDNA heteroplasmy, noncoding RNA (ncRNA), and epigenetic modification of the mitochondrial genome.

Thus, studying mitochondrial function has elucidated new therapeutic strategies for many life-threatening diseases. However, despite the significant biological relevance of mitochondrial function in human disease, the clinical relevance of studying mitochondrial function remains limited, primarily due to a lack of consensus on the best approach to studying mitochondrial function in both blood (PBMC) and tissues.

3. Determination of mitochondrial function in PBMC

Determination of mitochondrial function in PBMCs has previously been described as a surrogate marker of overall mitochondrial function in human disease states such as diabetes mellitus [6,32–34], sepsis [35,36], heart failure [37], CNS disease [38,39], shock [40,41], and long COVID [42–46]. PBMCs are more available than tissues, and the determination of levels of mitochondrial function in PBMC may offer multiple advantages over tissue biopsies for determining mitochondrial function in human subjects [8]. Thus, assessing mitochondrial function in PBMC may be a valuable biomarker for studying cross-sectional correlates of overall mitochondrial function with specific disease phenotypes and the impact of clinical interventions on mitochondrial function in systemic disease.

However, given the variability in mitochondrial response between different tissues, PBMCs are often a poor proxy for tissue-specific mitochondrial dysfunction [6,38–40,47–53] (Table 1). Limited experimental studies have shown that the respiratory capacity of blood cells, monocytes, and platelets is related to contemporaneous respirometry assessments of skeletal and cardiac muscle mitochondria [54]. However, prior human studies have shown that mitochondrial respiration in circulating cells (PBMCs and platelets) does not reflect mitochondrial respiration of skeletal muscle fibers derived from the same subjects and can only reflect a general measure of overall metabolic function [54–56]. Differences in the experimental approach, such as tissue processing in animals compared to humans, variable protocols in mitochondrial assays, and limited sample size in human studies, may explain these discrepancies.

Table 1.

Limitations of assessment of overall mitochondrial function in peripheral blood immune cells as a surrogate measure of tissue specific mitochondrial function.

• Monocytes and lymphocytes differ in both mitochondrial function and metabolism [40,47]

• Mitochondria within platelets are overrepresented when compared to leukocytes [48]

• Variable contamination of PBMC with platelets [8]

• The composition of PBMCs can shift during systemic disease and may affect overall mitochondrial function in heterogeneous bulk immune cell population [47,49]

• Lack of linear mitochondrial response between PBMCs and different vital organs [50]

• Different tissues have different mitochondrial responses [40]

• PBMCs may be poorly representative of post-mitotic tissues that are more directly related to human tissue-specific mitochondrial disease [38,39]

• The eventual adaptation of PBMC to energy supply may differ from cells that exclusively rely on mitochondria to meet their energetic demands [6]

• Heterogeneity and heteroplasmy among mitochondria in cells and tissue [52,53]

• Unpredictable segregation during replication among mitochondria [51]

Mitochondrial function in cell lines may not fully predict mitochondrial function in differentiated primary cells [57–59], which maintain some tissue-specific characteristics [60]. Mitochondria are also highly heterogeneous organelles. Morphological and functional heterogeneity in mitochondria within the same tissue [61,62], heteroplasmy among mitochondria inside a single cell and in a given tissue, combined with the variability inherent in mitochondrial segregation during mitosis, further complicate the measurement of mitochondrial dysfunction. It is harder still to determine mitochondrial dysfunction at the whole tissue level or in a mixed cellular population like PBMCs [63]. Furthermore, sample contamination with platelets during PBMC isolation and the variation in mitochondrial activity produced by the threshold effect of mitochondrial DNA (mtDNA) regarding mitochondrial dysfunction may influence the determination of overall mitochondrial function in samples of PBMC that are likely to include both normal and abnormal cells [48].

After considering the above limitations when determining mitochondrial function in heterogeneous cellular populations and acknowledging the significant differences in mitochondrial function and metabolism between monocytes and T cells within PBMCs [64–66], mitochondrial function should ideally be studied separately for monocytes and T cells within PBMC samples. In healthy subjects, mitochondrial function differs substantially between monocytes, lymphocytes, and platelets [67,68].

Monocytes in peripheral blood have been used as surrogates for mitochondrial function based on specific characteristics, such as the presence of lipid metabolism genes that control monocyte cholesterol production and the high expression of the mitochondrial genome in these cells [66]. Monocytes may more closely resemble mitochondrial function at the tissue level since mitochondrial toxicity in patients manifests in resting or slowly dividing tissues. Indeed, a prior study showed that the maximal respiration of monocytes was positively correlated with the oxidative capacity of complex I and II in muscle fibers from the same animal [54].

Mitochondrial dysfunction, such as bioenergetic function, altered mitochondrial gene expression, and membrane potential in monocytes, is implicated in various human diseases, including kidney diseases [69,70], aging [71], metabolic disorders, atherosclerosis [70], and infections [72,73]. Mitochondrial dysfunction, such as altered mitochondrial bioenergetics, ROS, and membrane potential in lymphocytes, has been demonstrated in patients with autoimmune diseases such as systemic lupus erythematosus (SLE) [74], neurodegenerative diseases like Alzheimer’s disease [75], chronic HIV infection [76], schizophrenia [77], COVID-19 [78], and diabetes [79].

Further studies are therefore required to determine the suitability of this easily accessible cell population as a surrogate marker for mitochondrial dysfunction. Thus, cell-specific ROS generation and mitochondrial content should be considered when using monocytes or lymphocytes to assess the proliferation rate of PBMC mitochondrial function.

4. Measures of mitochondrial function in PBMC

4.1. Various measures of mitochondrial dysfunction in PBMC have been reported

Evaluation of cellular oxygen consumption [11], analysis of mtDNA content, the ratio of mtDNA to nuclear DNA [10,80,81], activities of mitochondrial respiratory chain complexes I, II (succinate dehydrogenase), III, IV (cytochrome c oxidase) [36,82], citrate synthase activity [9,36], cellular ATP content, increased mitochondrial synthesis of reactive oxygen species [6,12,83], oxidative stress, intracellular lipid accumulation, cytotoxicity/apoptosis [84], mitochondrial mass [84], intra-mitochondrial cardiolipin distribution [84], calcium concentration, and modified mitochondrial morphology and dynamics [34,83,85] in PBMCs are often used as markers of overall mitochondrial function in clinical studies (Table 2) (Fig. 1) [3,6,34,41,83,84,86–94]. Electronic paramagnetic resonance is the gold standard for measuring mitochondrial ROS production, followed by various chemical assays that use redox-sensitive fluorescent probes [56]. However, the usefulness of these methods for routine monitoring has not been established since all methods have limitations, and bulk methods are affected by the intrinsic heterogeneity in samples and tissues that limit the accurate determination of mitochondrial function [95,96].

Table 2.

Measures of mitochondrial dysfunction in PBMCs (cell level).

Method	Comments
Mitochondrial oxygen consumption (respirometry)	• Gold standard method [86] • Allows user to probe multiple parameters of respiration • Uses both the Oroboros and Seahorse method • Both methods lack standards of oxygen consumption • Seahorse method offers high resolution, high sensitivity, and requires little blood [87–89] • Requires specialized equipment and training that may not be readily available
mtDNA quantity	● Fairly simple to do with PCR [3] ● Widely used measure of mitochondrial function in bulk heterogeneous cell populations like PBMCs ● Has major limitations in heterogeneous samples like PBMCs ● May also be affected by the quality of full-length mtDNA [90] ● Mitochondrial dysfunction may occur even in the absence of mitochondrial DNA depletion [90] i.e. altered gene expression profiles, mitochondrial oxidative stress, mtDNA mutations, inhibition of mitochondrial enzymes, uncoupling of the electron transport, and induction of apoptosis
Fluorescent assessment of mitochondrial membrane potential (Δψm)	● Most common method for measuring mitochondrial function at the single cell level ● Measures energy state of the organelle [41,86] ● Self-quenching and non-specific binding may affect results ● Assessment at the single cell level resolves issue of heterogeneity ● Smaller amounts of fluorescent dyes limits toxicity ● Data must be interpreted carefully ● More well validated for cell culture experiments and not for PBMCs ● Cannot do fixation ● Semi-quantitative (more quantitative with flow cytometry) ● Better for assessment of relative (rather than absolute) differences in mitochondrial function between samples [86]
TPP+ electrode for assessment of MMP in isolated mitochondria	● Easy and precise tool to measure Δψm in isolated organelles and is largely used. ● Requires a lot of blood volume to study isolated mitochondria in immune cells ● Not applicable to large scale research
Fluorescent assessment of mitochondrial ROS (mito-ROS)	● Single cell level assessment of mitochondrial function (immunofluorescence or flow cytometry); overcomes the limitation of mitochondrial heterogeneity ● Key measure of systemic mitochondrial dysfunction [6,34,83] ● The fluorescent probe MitoSoX Red needs to be stained at specific concentrations since it can be unstable and prone to oxidation. Low amount of the fluorescent dyes avoids all the problems related to the toxicity of the probes, capable of affecting several mitochondrial functions ● Semi-quantitative (more quantitative with flow cytometry) ● Has been used to measure mitochondrial function in patients in relationship to clinical outcome [6,34,83,91]
Cardiolipin assay	● Cardiolipin is important for mitochondrial enzymes and apoptosis [92] ● Analyzes cardiolipin intramitochondrial distribution, using fluorochromes like [nonyl acridine orange (NAO)] ● Can use a cytofluorimetric method or flow cytometry ● Has been used in clinical studies [84]
Spectrophotometric determination of activities of respiratory chain complexes	● Uses substrates and inhibitors of the electron transport chain in combination with spectrophotometric analysis of related reactions. ● Resulting data is inconsistent due to the intrinsic heterogeneity in samples and tissues [93,94]

Fig. 1.

PBMCS can be used to study different aspects of mitochondrial structure and function at the cellular level. Respirometry, electrode measurements of Δψm, cardiolipin assays, spectrophotometry of the ETC, mtDNA quantification, and florescence analysis of Δψm and mito-ROS are all used to distinguish proper mitochondrial function from dysfunction.

4.2. The mitochondrial oxygen consumption rate is the best measure of mitochondrial function in PBMCs

When assessing mitochondrial function, the oxygen consumption rate (OCR) is more informative than the measurement of mitochondrial membrane potential (MMP) [86] (Table 3). Mitochondrial respiration has classically been assessed using the Clark-type polarographic electrode, which measures oxygen consumption in cells or isolated mitochondria in real time [97]. The Oroboros high-resolution respirometry system improved upon traditional Clark electrode technology to accurately measure oxygen consumption with maximal sensitivity and precision [98–100]. While the Clark electrode can measure the oxygen consumption rate in blood cells in suspension [101], the Seahorse method [102] uses fewer cells and has greater sensitivity [87–89] (Table 3). This method can assess mitochondrial function in platelets and leukocytes isolated from peripheral blood [67,103]. The XF analyzer determines the oxygen consumption rate and changes in pH, which can be attributed to mitochondrial respiration and glycolysis, respectively, in a high-throughput manner [86,104,105]. Determination of cellular bioenergetics in leukocytes and platelets can be accomplished with assays using inhibitors, uncouplers, and substrates of glycolysis and mitochondrial respiratory complexes in conjunction with the Oroboros or XF systems [32,67,106].

Table 3.

Measures of mitochondrial dysfunction at tissue level.

Muscle biopsies	● Gold standard for acquiring mitochondrial rich human tissue.
	● Highly invasive
	● Frozen or damaged tissue can hinder mitochondrial function results
Skin grafts	● Slightly less invasive methodology compared to muscle biopsy
	● Higher numbers of cells compared to muscle biopsy
	● Potential for variability in mitochondrial function in cell [58]
Near-infrared spectroscopy (NIRS)	● Relatively inexpensive
	● Data interpretation confounded
	● Measurement affected by adiposity and variable blood flow of the subject [87,102]
Phosphorus Magnetic Resonance Spectroscopy (31P-MRS)	● Snapshot of human tissue metabolism via phosphate group visualization, specifically phosphocreatine, the phosphate groups of ATP and inorganic phosphate [88]
	● Specialized equipment necessary
	● Requires extensive training
	● Expensive
Non-invasive positron emission tomography (PET)	● Sensitive measure of specific metabolism mechanisms based on the detection of injected positron-emitting radionuclide tracers [87,88]
	● Specialized equipment necessary
	● Requires extensive training
	● Expensive

4.3. Determining mitochondrial membrane potential is a standard method for observing mitochondrial function in PBMCs, but it has several challenges

The mitochondrial membrane potential (Δψ_m, MMP) indicates the organelle’s energization level. It is produced by the electric potential across the inner mitochondrial membrane during ATP synthesis [107]. It is worth noting that measuring the MMP of PBMCs could be a valuable tool for investigating mitochondrial dysfunction in vivo in different diseases [84,91,108]. Due to the simplicity and ease of access to imaging microscopes, a popular technique for monitoring the mitochondrial function of intact cells at the single-cell level is fluorescent monitoring of Δψ_m using redox-sensitive dyes [34].

However, MMP can be influenced by the overall energetic state of the cell, which depends on mitochondrial activity and oxygen consumption, cytosolic ATP production, and NADH:NAD1 (nicotinamide adenine dinucleotide) ratio [48,109]. Most of these assays are based on cationic lipophilic fluorochromes that have limitations, including low sensitivity, high cytotoxicity, specific inhibition of mitochondrial functions [110], the possibility of being pumped out of mitochondria by multidrug resistance pumps [111], or the need for multiple wash steps with long equilibrium times. Accurate design and analysis of experiments are challenging, and although accurate relative values of potential can be obtained, absolute calibration has only recently been accomplished [86]. Although Δψ_m describes bioenergetic status, it is generally less sensitive to deviations in function and less helpful than respiratory rate measurements. Combining MMP and respiration rates is much more valuable than measuring them alone.

4.4. Determination of mitochondrial DNA in PBMC

Mitochondrial DNA mutations and deletions contribute to mitochondrial dysfunction and can be easily measured by PCR [3]. Determination of mtDNA is widely used measure of mitochondrial function in heterogeneous cell populations like PBMCs However, this method may be affected by the quality of full-length mtDNA and several confounding factors such as nuclear-mitochondrial DNA cross-talk or the impact of heteroplasmy (presence of multiple mitochondrial DNA variants within the same cell or individual on disease phenotypes [90,112–114]. Mitochondrial dysfunction, such as mtDNA mutations, mitochondrial oxidative stress, altered gene expression profiles, inhibition of mitochondrial enzymes, uncoupling of the electron transport, and induction of apoptosis, may occur even without mtDNA depletion [90]. Thus, due to these limitations, the measurement of mtDNA alterations (deletions, mutations) is not considered a gold standard measure of mitochondrial function; it is only a general measurement of mitochondrial integrity and function.

4.5. Determination of mitochondrial ROS (mito-ROS) in PBMC

Fluorescent assessment of mito-ROS in PBMC allows single-cell level assessment of mitochondrial function by immunofluorescence or flow cytometry and overcomes the limitations of mitochondrial heterogeneity. Mito-ROS measurement using the fluorescent probe MitoSoX Red has been used to measure mitochondrial function in patients in relation to clinical outcome [6,34,83,91]. However, this method has several limitations, such as the lack of specificity of current probes and the dual role of ROS as both a damaging and signaling molecule that may not fully represent mitochondrial function [115,116].

4.6. Other assays of mitochondrial function in PBMC

Other functional mitochondrial processes that could be important in disease, such as mitophagy and mitochondrial dynamics, mitochondrial biogenesis, and morphology, can also be measured. The analysis of mitochondrial dynamics, such as the rate of mitochondrial fission and fusion within immune cells, allows us to understand how fundamental metabolism influences immune cell functions [117]. Mitochondrial dynamics can be determined by both the mRNA expressions of mitochondrial dynamics-related genes such as mitofusin 1 (MFN1), mitofusin 2 (MFN2), optic atrophy protein-1 (OPA1), fission protein-1 (FIS1), and dynamin-related protein 1 (DRP1) [118] as well as microscopy. Mitochondria morphology, size, and number can be determined by transmission electron microscopy (TEM), and the number, length, perimeter, and area of mitochondria are reduced in disease. Impaired mitochondrial morphology and dynamics in PBMCs were significantly correlated with human diseases such as frailty [118], diabetes [34,83] and vascular dysfunction [83,119]. Mitochondrial biogenesis can be assessed by measuring the mitochondrial biogenesis protein PGC1α and by the use of fluorescent probes that measure mitochondrial mass. Reduced mitochondrial mass in PBMC has been associated with diseases such as diabetes [34,120]. Increased mitophagy in PBMCs as measured by decreased mRNA and protein levels of PINK1 and Parkin and increased levels of LC3B and ubiquitination of the mitofusin proteins MFN1 and −2 has been linked to neurodegenerative diseases [121,122] and diabetes [120]. Patients with long COVID have been found to have increased serum levels of MFN2 and DRP1, which could signal increased mitophagy in response to increased oxidative stress [123]. Thus, assessing mitophagy, mitochondrial dynamics, mitochondrial biogenesis, and morphology can complement measurement of mito-ROS, MMP, and bioenergetics to provide a more comprehensive determination of mitochondrial function in PBMCs.

5. Addressing limitations in the determination of mitochondrial function in PBMC

Since all ex vivo studies have problems with the potential contamination of samples with unwanted cell types, purification methods, and single-cell analysis minimize such contamination and address significant limitations in determining mitochondrial function using human samples [124]. If platelets are not adequately removed from the cell samples, the reported changes in mitochondrial function could be due to variations in platelet quantity, mitochondrial content, sample processing, or storage. Some of the data presented in the literature, claiming to measure mtDNA content in leukocytes or other blood cells, must be interpreted carefully [48]. Platelet removal by sorting and purification can minimize platelet contamination [48] and allow precise and more accurate quantification of mtDNA in isolated cells.

Flow cytometry offers several advantages over the conventional measurements of bulk suspensions of organelles or cells. First, flow cytometry evaluates only fluorescence associated with single particles using low amounts of biological samples and dye, thus avoiding quenching and light-scattering variations. Second, the amount of fluorescent dyes is usually extremely low, which avoids all the problems related to the toxicity of the probes. Third, flow cytometry allows the quantification of functional characteristics of individual cells or organelles and addresses the crucial question of heterogeneity. Functional flow cytometry offers a quick, sensitive evaluation of the mitochondria’s overall function at the single-cell or even single-organelle level. Polychromatic flow cytometry quantifies multiple factors of mitochondrial activity and permits a better quantification of mitochondrial function in heterogeneous cell populations. Measures of mitochondrial function that can be determined by flow cytometry in PBMCs include mitochondrial membrane potential, mitochondrial mass, mitochondrial ROS/superoxide levels, intra-mitochondrial cardiolipin distribution, and early and late apoptosis [84,125].

6. Determination of mitochondrial function at the tissue level

Several experimental methods have been utilized to measure mitochondrial function at the tissue level (Fig. 2). Near-infrared spectroscopy measures tissue oxygenation by determining the optical absorption of the oxy- and deoxy-heme groups of hemoglobin, myoglobin, and cytochrome c [58,87,88,102,126] (Table 3). Although this method is relatively inexpensive, data interpretation is confounded by multiple heme-containing groups in tissue and further subject-dependent factors [127,128].

Fig. 2.

Mitochondrial function at the tissue level can be measured by a variety of experimental techniques. A diverse range of tissues from muscle to skin biopsies can be analyzed using techniques such as flow cytometry, respirometry, positron emission tomography (PET) imaging, levels of circulating mitochondrial proteins (i.e.. Peroxiredoxin), and Magnetic Resonance Spectroscopy (MRS) to determine mitochondrial function.

Furthermore, Phosphorus Magnetic Resonance Spectroscopy (MRS) can measure the metabolism of human tissues through visualization of crucial phosphate groups [129] (Table 3). This method can gather the energy status of static organs and is particularly useful for determining the kinetics of ATP generation in the working heart or skeletal muscle [90,130,131]. For example, recent studies utilized MRS to demonstrate muscle and brain alterations indicative of mitochondrial dysfunction in long COVID patients [132–134].

Lastly, noninvasive positron emission tomography can measure various aspects of metabolism by recognizing positron-emitting radionuclide tracers ([135]). ¹⁸F-fluorodeoxyglucose is most commonly used for imaging glucose metabolism ([136,137]) (Table 3).

Several studies showed reduced mitochondrial oxidative capacity, increased ROS production, and impaired mitochondrial morphology in the skeletal muscle of individuals with insulin resistance, a hallmark of diabetes [138,139]. High-resolution respirometry was a reproducible method to assess the function of mitochondria in native homogenates of fresh human heart muscle [140]. In a cross-sectional study of 27 patients older than 60 years with HFpEF and 45 healthy age-matched controls, high-resolution respirometry of skeletal muscles from patients with heart failure with preserved ejection fraction (HFpEF) revealed markedly reduced bioenergetic capacity associated with peak exercise oxygen consumption and exercise performance [141]. The authors concluded that these data may impact the development of therapeutic strategies that target mitochondrial dysfunction in patients with HFpEF [141]. Overall, there is limited evidence regarding the validation of measures of mitochondrial function in blood cells compared to tissues from the same persons.

7. Use of frozen biospecimens and respirometry

The mitochondrial electron transport chain requires intact mitochondrial membranes and electron transport chain proteins to function. Respirometry analysis is complex for stored samples and long-term studies since determining mitochondrial oxidative phosphorylation previously required immediate processing of tissue samples secondary to damage to the inner membrane from freeze-thaw [142]. A recent advancement in tissue sampling is a novel respirometry approach for frozen samples that corrects for variable permeabilization of mitochondrial membranes by restoring electron transfer components typically lost during freeze/thaw [142]. This method retains 90–95 % of the maximum respiratory capacity in frozen samples with high sensitivity and can be used on isolated mitochondria, permeabilized cells, and tissue homogenates [142]. This approach will allow the determination of the function of mitochondrial Complexes I to IV from various measurements, remote sides, or retrospectively from tissue biobanks. This advantage is incredibly valuable for age-related studies, as longitudinal experiments can monitor bioenergetic health over several years.

8. Determination of mitochondrial function in vivo

Although various tools exist to determine mitochondrial function in vitro, comparing differences in mitochondrial activity directly at the whole tissue level in vivo has proven more elusive. Recent advances involving novel radiotracers such as 4-[¹⁸F] fluorobenzyl triphenylphosphonium (¹⁸FBnTP) in combination with positron emission tomography (PET) imaging have allowed for the determination of mitochondrial function in vivo with high specificity [143]. ¹⁸FBnTP is a voltage-sensitive cationic radiotracer that localizes to the mitochondria due to its affinity for the negatively charged inner mitochondrial membrane. This allows for specific detection of differential levels of oxidative phosphorylation occurring in the mitochondria of heterogeneous cell populations [105]. This technique has been validated in non-small cell lung cancer but can be used in various in vivo settings to determine changes in the oxidative capacity of mitochondria in different disease states. ¹⁸FBnTP PET imaging has also been combined with microcomputed tomography and three-dimensional scanning block-face electron microscopy to create spatial maps of mitochondrial networks to help further relate changes in mitochondrial structure to changes in metabolic output [144].

9. Biomarkers of mitochondrial dysfunction

Measuring circulating biomarkers in patients is an indirect method to measure overall mitochondrial function during various disease processes. Although the use of complete blood count, creatine phosphokinase, transaminases, albumin, lactate and pyruvate, amino acids, and acylcarnitines, along with quantitative or qualitative urinary organic acids are useful biomarkers in the blood, urine, and spinal fluid to diagnose primary genetic mitochondrial diseases [145,146], they are not useful biomarkers in acquired mitochondrial alterations of complex diseases like diabetes and cancer with less extreme phenotypes of mitochondrial dysfunction. In this section, we will focus on biomarkers of mitochondrial function that have been validated specifically in acquired mitochondrial disorders such as the cardiometabolic syndrome and infection.

9.1. Cell-free mitochondrial DNA (cf-mtDNA)

The association between cfDNA and the severity of disease can be assessed by measuring the molecular characteristics of cfDNA such as integrity (fragment sizes of mtDNA), cellular origin and plasma concentration. An increase in plasma cfDNA levels is associated with organ dysfunction in diseases, such as inflammatory response in sepsis. Necrosis leads to larger fragment sizes of DNA, whereas apoptosis leads to shorter fragments. The assessment of cf-mtDNA may serve as a more sensitive diagnostic tool compared to cf-nDNA, since the lack of nucleosome-associated histone proteins in mitochondria in cf-mtDNA leads to shorter and more abundant fragments in the plasma [147]. Prior studies have shown the clinical relevance of these biomarkers since cfDNA can differentiate between patients with infection and those with sepsis and those with indications for intensive care [148]. Overall, elevated ccfmtDNA levels in plasma have been associated with acute respiratory distress syndrome, tumors, and inflammation [149].

9.2. Biomarkers of mitochondrial oxidative stress

Increased blood levels of biomarkers of mitochondrial oxidative stress, like SOD1, malondialdehyde, F2-isoprostane, and associated reduced levels of mitochondrial antioxidants like glutathione and coenzyme Q have been linked with several diseases such as aging, infection and cardiovascular disease. Malondialdehyde (MDA), a byproduct of lipid peroxidation that occurs when ROS react with fatty acids in the cell membrane, has been associated with aging and arterial stiffness [150]. Patients with long COVID have been found to have increased levels of biomarkers of mitochondrial stress, like SOD1, malondialdehyde, F2-isoprostane, and associated reduced levels of mitochondrial antioxidants like glutathione and coenzyme Q [123,151–154]. Deficiency of CoQ10 is present in several chronic and age-related disease including cardiovascular disease [155]. Levels of serum peroxiredoxin-3 (PRDX3), an antioxidant enzyme in the mitochondria that prevents ROS damage can suggest increased ROS damage in mitochondria and have been used to study mitochondrial dysfunction in long COVID [154] and have also been linked to insulin resistance [156].

9.3. Other biomarkers of mitochondrial function

The cytokine stress markers of mitochondrial dysfunction, growth differentiation factor 15 (GDF-15) and fibroblast growth factor 21 (FGF-21), have recently emerged as biomarkers of mitochondrial disease. GDF-15 is a member of the transforming growth factor β superfamily that is produced as a response to cellular stress and has immunoregulatory and anti-inflammatory functions, but unclear role in mitochondrial diseases. GDF-15 has previously been associated with numerous other conditions like cardiovascular disease, sepsis, cancer, and diabetes [157]. FGF-21 is produced in the liver and expressed in adipose tissue, muscle, and the pancreas, and regulates both glucose and lipid metabolism and has recently been linked with mitochondrial disease. Elevated FGF-21 levels have also been associated with several medical conditions in humans including obesity, metabolic syndrome, diabetes, cardiovascular disease, non-alcoholic fatty liver, and kidney disease [158].

However, all putative biomarkers of mitochondrial dysfunction must be rigorously tested to determine if a strong correlation exists between a biomarker of mitochondrial dysfunction and a corresponding disease state. Further standardization of biomarkers related to diseases in coordination with mtDNA analysis, respirometry, and MRS could allow for personalized medical care centered on mitochondrial function.

10. Clinical applicability of assaying mitochondrial parameters in disease

As discussed above, mitochondrial dysfunction plays a significant role in the pathogenesis of human diseases that involve metabolically active tissues enriched in mitochondria, such as the brain, heart, skeletal muscle, and liver. However, although the determination of mitochondrial function is biologically relevant for the pathogenesis of major diseases of public health, such as cardiovascular diseases, cancer, and diabetes [15,16], there is limited data on whether mitochondrial assays can be used to predict disease severity or diagnose a disease. Determination of mitochondrial function in PBMCs has previously been described as a surrogate marker of overall mitochondrial function in human disease states such as aging [159,160], diabetes mellitus [32–34], sepsis [35,36], heart failure [37], CNS disease [38,39], neurodegenerative diseases [75,159], shock [40,41], depression [161], fibromyalgia [162], autoimmune diseases [74], and Long COVID [42–46] This manuscript does not focus on primary genetic mitochondrial diseases, as they are rare and less relevant to most clinicians.

Current knowledge on using mitochondrial assays to predict human diseases like diabetes is mainly based on correlational data, where specific assays indicate the presence of disease. Several studies showed reduced mitochondrial oxidative capacity, increased ROS production, and impaired mitochondrial morphology in the skeletal muscle of individuals with insulin resistance, a hallmark of diabetes [138,139]. Other studies have demonstrated associations between mutations in mitochondrial DNA and forms of diabetes characterized by impaired insulin secretion [163]. It has been suggested that mtDNA mutations could lead to beta-cell dysfunction and insulin resistance, contributing to the pathogenesis of diabetes. A common mtDNA variant (the 16,189 variant) is positively correlated with blood fasting insulin and was associated with type 2 diabetes in a population-based case-control study [164]. Widlansky et al. showed that altered mitochondrial membrane potential, mass, and morphology in the mononuclear cells of humans with type 2 diabetes could discern mitochondrial perturbations in type 2 diabetes mellitus (T2DM) and be potentially clinically relevant [34]. Avila et al. performed a pilot study of mitochondrial respiratory function and proteomic changes in platelets extracted from insulin-sensitive (n = 8) and type 2 diabetic subjects (n = 7). They showed that platelet mitochondria had diminished mitochondrial respirometry, induction of the mitochondrial antioxidant enzymes and increased mitochondrial protein carbonylation as an oxidative stress index in diabetic vs. control subjects [32]. They suggested that platelets are accessible and may help measure the mitochondrial effects of antidiabetic treatments [32]. The study population comprised 182 diabetic patients and 50 body-composition- and age-matched controls. Impairment of mitochondrial function in diabetic patients, evidenced by a decrease in mitochondrial O₂ consumption, an increase in ROS production, a decreased GSH/GSSG ratio, a drop in glutathione (GSH) levels, and an undermining of the mitochondrial membrane potential and impairment of mitochondrial complex I [33]. Hartman et al., using freshly isolated PBMCs from 26 patients with type 2 diabetes mellitus and 28 non-diabetic controls and mitochondrial respirometry, showed that diabetes mellitus is associated with a pattern of mitochondrial oxygen consumption consistent with higher production of reactive oxygen species [6]. They also showed that a correlation between mitochondrial oxygen consumption and nitroglycerin-mediated dilation may suggest a link between mitochondrial dysfunction and vascular smooth muscle cell dysfunction [6]. Overall, there is limited evidence regarding the validation of measures of mitochondrial function in blood cells compared to muscle tissue in patients with diabetes.

Besides diabetes, there is limited evidence regarding using mitochondrial assays to predict disease phenotypes like aging, cardiovascular, and CNS disease. Tyrrel et al. showed that gait speed is associated both muscle mitochondrial respiratory control ratio and PBMC maximal and spare respiratory capacity in overweight/obese older adults [54]. In a study of 15 patients with congestive heart failure (CHF) and 9 healthy volunteers, CHF patients had higher mitochondrial ROS-positive arterial and venous white blood cells and platelets than controls, which may amplify oxidative stress in CHF [37]. In a study of 80 patients with migraine and 24 healthy control subjects, platelet mitochondrial enzyme activities such as NADH-dehydrogenase, citrate synthase, and cytochrome-c-oxidase activities were significantly lower in migraine patients than in controls [38]. In an uncontrolled study, platelet mitochondria purified from all 10 patients with Parkinson’s disease significantly reduced complex I (NADH: ubiquinone oxidoreductase) activity [39]. Although PCR methods to detect mtDNA deletions and mutations are used to diagnose genetic mitochondrial syndromes [165], there is limited evidence regarding mtDNA mutations and their link to other human diseases.

Overall, this evidence suggests that results from mitochondrial assays may not only provide associative data regarding the presence of mitochondrial dysfunction in human disease but could also be used to predict clinical disease phenotypes (aging, exercise tolerance, severity of insulin resistance) or potentially guide the development of therapeutics that target mitochondrial dysfunction. Mitochondrial high-throughput assays, such as the Bioenergetic Health Index (BHI), can be calculated to represent the patient’s composite mitochondrial profile for a selected cell type and have the potential to be a new biomarker for assessing patient health with both prognostic and diagnostic value [166]. However, there is limited evidence comparing mitochondrial assays in terms of sensitivity, specificity, and clinical applicability.

11. Future directions

Although mitochondrial dysfunction may be detected at various levels in cells, tissues, and end organs in human diseases, it may not always be the principal pathogenic instigator for the development and progression of the disease. Human diseases often have complex pathogeneses, and multiple pathophysiological alterations contribute to the progression of the disease. The complexity of human disease can be better studied at the single-cell level while considering the heterogeneity of cellular populations, tissues, and end organs. Thus, it is essential to use assays that predict challenges in the measurement of mitochondrial function, such as tissue heterogeneity.

Cutting-edge approaches such as single-cell metabolomics, combined with techniques like mass spectrometry, are increasingly used to study mitochondrial function at the single-cell level by analyzing metabolites involved in energy production and other mitochondrial pathways. [167,168] Studies using single-cell metabolomics have linked mitochondrial dysfunction to various diseases, including those affecting cardiovascular, neurological, and immunological systems [169]. Single-cell sequencing techniques, like single-cell RNA sequencing, are used to analyze mtDNA copy number and mutations, providing further insights into mitochondrial dysfunction [170]. Techniques like Förster Resonance Energy Transfer (FRET) sensor imaging allow for the visualization and quantification of mitochondrial flux in individual cells [171].

Emerging proteomic techniques have created a powerful tool for the characterization of static and dynamic proteomes, can detect protein-protein interactions and post-translation modifications that play key roles in mitochondrial function [172,173]. Transcriptomics approaches such as spatial transcriptomics can also further advance understanding of mitochondrial function in vivo in several diseases like neurogenerative diseases [173,174].

Microfluidic organ-on-a-chip technologies offer a powerful platform for studying mitochondrial function in a more physiologically relevant context than traditional cell culture methods. Replicate the complex microenvironment of organs, including tissue architectures, fluid flow, and mechanical cues, to simulate organ-level physiology [175,176]. These emerging methods allow real-time monitoring of metabolic function in organ-on-chip microdevices to track the dynamics of mitochondrial dysfunction [175]. Emerging Systems Biology systems also allow computational modeling of mitochondrial function to understand the control and regulation of mitochondrial energy metabolism and its interactions with cytoplasmic and other cellular compartments [177]. Integration of these emerging technologies has the potential to revolutionize the field, provide a more comprehensive understanding of mitochondrial function, and address current limitations in mitochondrial research, such as measurement of mitochondrial function in heterogeneous populations. These emerging technologies can be used in a translational setting for aiding both the diagnosis of mitochondrial disease and targeting of mitochondria for treatment.

12. Conclusion

Measuring mitochondrial dysfunction at cellular and tissue levels can provide valuable insight into human disease. Multiple methods have been used for measuring mitochondrial function in readily accessible whole-blood PBMCs [5–7]. Of note, measurements of cellular oxygen consumption (COC) and mito-ROS in PBMCs have been correlated with mitochondrial dysfunction in experimental disease models [6,8,9]. However, all methods using PBMCs as a surrogate marker of overall mitochondrial function have limitations. Still, using multiple tools, such as respirometry, flow cytometry, and measures of mitochondrial membrane potential, can lead to a more complete characterization of mitochondrial function. Given the limitations inherent in individual metrics, data from prior studies utilizing these flawed methods could have led to inconsistent findings. Assessment of mitochondrial function at the tissue level remains the gold standard, but tissue is often inaccessible. A less invasive and cost-effective method requiring less expertise would assess human mitochondrial function more practically and accurately. Future advancements could increase the measurement of mitochondrial function at the tissue level in translational research studies and clinics as a personalized medicine tool, especially in recent advancements in personalized genomic, metabolic, and proteomic data analysis. Recent improvements in respirometry have led to the development of a cost-effective and accurate method suitable for frozen samples [178]. This advancement is promising for age-related studies as it constructs a longitudinal design, bringing clinicians one step closer to measuring physiological respiration rates in frozen specimens. Emerging technologies such as single-cell metabolomics, single-cell sequencing, microfluidic organ-on-a-chip and emerging systems biology systems can provide a more comprehensive understanding of mitochondrial function, and address limitations of current mitochondrial assays, such as measurement of mitochondrial function in heterogeneous populations. These emerging methods can be used in a translational setting for aiding both the diagnosis of mitochondrial disease and targeting of mitochondria for treatment. With the emergence of mitochondrial dysfunction as an important marker and driver of diseases like diabetes, further interdisciplinary collaboration toward measurements of mitochondrial activity will lead to a better understanding of the pathophysiology of human disease.