In vivo quantification of mitochondrial membrane potential

The paper by Momcilovic et al¹ presents important results on mitochondrial metabolism differences amongst various lung cancer subtypes in mouse. Unfortunately, the work of Momcilovic et al1 propagates critical misunderstandings and omissions about the underlying basis for application of voltage sensing tracers. The principle of mitochondrial membrane potential (ΔΨ_m) imaging with radiolabelled PET probes is based on benchtop in vitro methods. In vitro quantification of ΔΨ_m has been established decades ago using various lipophilic cations such as ³H-tetraphenylphosphonium (³H-TPP+).^2–4 In all cases, the Nernst equation is used to relate the probe’s equilibrium concentration on either side of a membrane to the electric potential across the membrane.⁵ Importantly, the Nernst equation is only valid when probe concentrations are at equilibrium, that is the concentration on either side of the membrane are not changing with time.

Momcilovic et al1 claim to “measure mitochondrial membrane potential in non-small cell lung cancer in vivo using a voltage sensitive positron emission tomography (PET) radiotracer”, the lipophilic cation 4-[¹⁸F]fluorobenzyl-triphenylphosphonium (¹⁸F-BnTP). It must be pointed out that, contrary to the authors’ claim, they do not measure membrane potential. Their work is based on the empirical endpoint, percent dose per gram-tissue (normalized by the %dose of the myocardium). %dose of ¹⁸F-BnTP is time dependent and a function of several physiological variables, including the level of ¹⁸F-BnTP in plasma, the fractional volumes of extracellular space (ECS), cellular membrane potential, as well as ΔΨ_m. The need to normalize the endpoint to that of some other organ which may evolve differently in time is also a problem for reproducibility and clinical translation. These crucial aspects of the way research and clinical translation of voltage sensing compounds can take place have already been studied by our group.⁶ We have shown that a unique absolute endpoint, in units of millivolts (mV), can be obtained during secular equilibrium of a radiotracer such as ¹⁸F-TPP⁺ and ¹⁸F-BnTP. Notably, we have shown that, at steady-state, the tracer concentration in tissue depends nearly linearly on the tracer concentration in plasma, fractional volumes of mitochondria and extracellular space but exponentially on the sum of the cellular and mitochondrial membrane potential:

$${\overline{C}}_T\approx {\overline{C}}_p(1-f_{ECS})·f_{\mathit{\text{mito}}}\cdot e^{-\beta (\mathrm{\Delta }{\mathrm{\Psi }}_c+\mathrm{\Delta }{\mathrm{\Psi }}_m)}$$ Eq. 1

Where ${\overline{C}}_T$ and ${\overline{C}}_p$ (with units of Bq/mL) indicate steady state tracer concentration in tissue and plasma respectively, ΔΨ_c and ΔΨ_m (mV) represent the cellular and mitochondrial membrane potential respectively, f_ECS and f_mito (unitless) represent the volume fraction of ECS and mitochondria respectively and β (mV⁻¹) is a constant term representing the ratio of known physical constants. It is also interesting to note that at steady-state, Eq.1 can be used to express the result in terms of percent dose.

$$\overline{\%D_T}\approx \overline{\%D_p}(1-f_{ECS})·f_{\mathit{\text{mito}}}\cdot e^{-\beta (\mathrm{\Delta }{\mathrm{\Psi }}_c+\mathrm{\Delta }{\mathrm{\Psi }}_m)},$$ Eq. 2

where we have expressed the result as a ratio of %dose fractions. We have found that the plasma tracer concentration C_p decreases monotonically after bolus injection of ¹⁸F-TPP⁺, meaning that the %dose index will not be time-invariant which may limit the reliability of this endpoint and its usefulness in research and clinical translation.

Although the authors did show similar mitochondrial density (f_mito) for both cancer subtypes with the pan-mitochondrial marker TOM12, they did not account for the other critical variables such as the volume fraction of ECS and the plasma tracer concentrations. Finally, adequate quantification of membrane potential requires measurements of equilibrium tissue and plasma concentrations, which cannot be done with a bolus radiotracer injection and delayed imaging protocol as performed in this study.⁶

Nevertheless, the results of Momcilovic et al¹ add to the body of evidence supporting the potential role of in vivo assessment of mitochondrial status in oncology. The authors rightly pointed out that imaging membrane potential might represent a valuable resource for the evaluation of mitochondrial activity in several areas of research including aging, physiology, and diseases. However, for successful translation of the methodology to human research and ultimately to the clinic, accurate and reproducible quantification is necessary. This can only be achieved with proper techniques and accounting for all critical variables.