Does hemoglobin affect measures of mitochondrial respiration in red blood cells? Comment to “Increased retention of functional mitochondria in mature sickle red blood cells is associated with increased sickling tendency, hemolysis and oxidative stress”

We are intrigued by the Connes group’s interesting study of the retention of mitochondria in mature red blood cells (RBC) in 61 patients with sickle cell anemia (SCA).¹ Using the sophisticated techniques of Image Stream (Amnis, MK II) and Mitotracker Red CMXRos Dye (Invitrogen) in combination with flow cytometry, they reported that a significant percentage of mature sickle cells in these patients retained mitochondria.¹ This presence of mitochondria in human erythrocytes is a unique phenomenon, heretofore considered to be a characteristic reserved for red cells in nonmammalian vertebrates (e.g., avians, fish, amphibians, reptiles).²

However, we were particularly fascinated by one of their primary aims; specifically, “to investigate the functionality of these mitochondria”.¹ Mitochondrial function was assessed with high-resolution respirometry, which according to the authors, resulted in “detectable mitochondrial oxygen consumption in sickle mature RBC, but not in healthy RBC.” The authors concluded that their data showed the presence of functional mitochondria in mature sickle RBC, “which could favor RBC sickling and accelerate RBC senescence, leading to increased cellular fragility and hemolysis”.¹ The presence of functional mitochondria in human erythrocytes is an interesting observation with important implications. For example, the Connes group noted that the propensity of RBC to sickle when deoxygenated was greater in the SCA subgroup with a high percentage (13%) of mitochondria retained in mature RBC.¹ These are provocative findings with great potential for both increased understanding of basic mechanisms and application to practice.

Oxygen consumption rate (J_O2) is, of course, a valid indicator of mitochondrial function, particularly when flux is modified by agents known to inhibit or accelerate various steps in the oxidative pathway, as was done by the Connes group here.¹ However, in such assessments, care must be taken to ensure the validity and accuracy of the J_O2 measurement itself. The Connes group used intact red cells in their polarographic measurement of RBC mitochondrial J_O2,¹ basically following the methods of Sjövall et al.³, and Stier et al.², with Stier being a co-author of this current study. Briefly, 100 μL of packed RBC was added to 1 mL of a potassium-based respiratory buffer (MiR05, see below), and transferred into an Oxygraph-2k high-resolution respirometer (Oroboros Instruments, Innsbruck, Austria) set at 37°C and containing another 1 mL of MiR05. Subsequently, a variety of standard mitochondrial respiratory measures were made, all based on J_O2. Of significant concern here is the O₂ associated with hemoglobin; i.e., oxyhemoglobin (HbO₂), in the red cells incubating in these respiration chambers. This store of O₂ was apparently ignored in the authors’ calculation of J_O2. The errors introduced by this omission are potentially quite large and variable depending on hemoglobin content, O₂ binding parameters (e.g., P₅₀), the oxygen tension (PO₂) at which the analysis is being done, and the duration of a given assay. The discussion below will be restricted to a simple presentation of the overall problem. Basic concepts and details related to how the Oroboros respirometer calculates J_O2 are clearly and completely described in the user’s manual.

Clark type O₂ electrodes such as that used in the Oroboros instrument generate an electrical signal proportional to the oxygen tension, PO₂ (mmHg), which, in turn, is proportional to the concentration of dissolved O₂ (nmolO₂ . mL-¹ or μM). As stated in Sander:⁴ “... the amount of dissolved gas is proportional to its partial pressure in the gas phase. The proportionality factor is called Henry’s law constant.” The value of the Henry constant (μmol . mmHg^-1); i.e., solubility, depends on temperature and other factors such as ionic strength. Pure water at 37°C exposed to (and in equilibrium with) room air at 1.0 atmosphere pressure contains about 213 μM of dissolved O₂ (213 nmolO₂ . mL-¹).⁴ If instead, MiR05 medium is the solution of interest, the higher ionic strength reduces O₂ solubility (“salting out”). In the Oroboros literature this is taken into account by the “FM” factor, which is 0.92 for MiR05. Accordingly, in our example: [O₂]=0.92 * 213 μM=196 μM at 1.0 atmosphere and 37°C. This value is roughly similar to the initial [O₂] shown in Figure 2 in Stier et al..² If we now multiply 196 nmol O₂ . mL-¹ by a total medium volume of 2.0 mL, we get 392 nmol O₂ in the respiratory chamber when the assay begins. Below we will compare this soluble O₂ mass to the O₂ mass bound to Hb.

When gas exchange with the environment is prohibited in the Oroboros, O₂ consumption decreases the PO₂ hence proportionally decreasing the electrode signal. The progressive fall in PO₂ across time (mmHg . min^-1), linked by the Henry constant (μmol . mmHg^-1) to the corresponding fall in [O₂] (μM or nmol O₂ . mL-¹), along with the medium volume (mL), allows the calculation of J_O2 (nmol O₂ . min-¹). Such calculations have been routinely made and reported in a vast literature, which began rapidly expanding shortly after Leland Clark developed his O₂ electrode with a cellophane covering in the 1950s.⁵

However, the presence of an O₂ buffering molecule such as Hb in the respiratory chamber complicates the above straightforward and linear relationships in two ways: i) the Hb mass available to bind O₂ in each assay must be determined, and ii) the HbO₂ binding parameters must be evaluated. Obviously, accounting for the utilization of this O₂ pool will involve a non-linear relationship between the electrode signal and O₂ concentration, and thereby J_O2. We can approximate the Hb-bound O₂ by assuming a normal mean corpuscular hemoglobin concentration of 34 g . dL-¹ (similar to the SCA patients in Figure 6C¹ if the 10-fold unit error is overlooked), and a Hb molecular weight of 67,000 g . mol-¹. With these values, the 100 μL of RBC in the assay contains a Hb mass of 0.507 μmol. At the onset of the assay, assuming 1.0 atmosphere barometric pressure, 0.2095 mole fraction of O₂ in dry air, and a 37°C assay temperature (thus 47.1 mmHg water vapor pressure at the air:water interface), gives a medium (MiR05) PO₂=(760-47.1)x0.2095=149 mmHg at the start of the assay. Thus, the Hb will initially be almost fully saturated, holding 4 O₂ . Hb-¹x0.507 μmol Hb=2,030 nmol O₂, which is 5.2 times greater than the soluble O₂ in the medium (see above). As J_O2 commences early in the assay the medium PO₂ will fall, drawing predominantly from the soluble pool. With time and ongoing J_O2, however, the continuous fall in PO₂ will elicit significant O₂ dissociation from Hb, which will dampen changes in PO₂ and result in an underestimation of the true JO₂ (if O₂ bound to Hb is ignored). Because in a given assay the PO₂ is progressively falling, the errors will confound not only the absolute rates but also the relative differences between what Stier et al.² call “routine, oxphos, leak, ETS, and non-mito” O₂ consumption. In a comparative study, like that of Esperti et al .¹, such errors may be amplified when the two RBC groups have different Hb contents and/ or O₂ binding kinetics. In any case, standard polarographic procedures used to measure JO2 by red blood cells must be viewed with extreme caution.

In practice, the mitochondrial respiration measurements may be largely made at high partial pressures of O₂ where the oxyhemoglobin dissociation curve is quite flat.⁶ However, a decrease of even 0.5% in percent saturation would release ≈5.2 nmol of O₂, an amount that would be about one-third of the decline in O₂ content occurring during “endogenous” mitochondrial function in king penguin RBC; see Figure 2 in Stier et al.² Of course, the further the PO₂ declines during the respiratory measurements, the closer the approach to a steeper portion of the HbO₂ dissociation curve. Accordingly, we propose that hemoglobin must be accounted for in quantitative measures of mitochondrial respiration in RBC.

We also noted that RBC were suspended in respiratory buffer MiR05 (0.5 mM EGTA, 3 mM MgCl₂, 60 mM K-lactobionate, 20 mM taurine, 10 mM KH₂PO₄, 20 mM Hepes, 110 mM sucrose, free fatty acid bovine serum albumin [1 g.L-¹], pH 7.1), a solution commonly used for studies of mitochondria, which are intracellular organelles.¹ Accordingly, we wondered i) whether a respiratory solution more closely resembling plasma might be more appropriate for RBC, and ii) whether glucose should be provided as a fuel.

Further, for quantitative mitochondrial respiratory analyses, it is important to be certain about the volume of RBC used in the assays. In this context, how an exact volume of RBC is assayed is unclear. If packed RBC are transferred, air displacement pipetting is error-laden, and if packed cells are resuspended, the amount of supernatant retained with prior centrifugation is uncertain.

Finally, we saw no mention of any pharmacotherapy for the SCA patients. Would any such treatments have potential ramifications for either the presence or function of mitochondria in the RBC of this group?

Once again, we acknowledge and appreciate the research of Esperti et al.,¹ but we also believe that their methods for mitochondrial respirometry deserve a candid discussion.

Footnotes

• Received June 20, 2024
• Accepted July 24, 2024

Correspondence

L.B. GLADDEN - gladdlb@auburn.edu

Disclosures

No conflicts of interest to disclose.