Measuring Mitochondrial Oxygen Tension

From Basic Principles to Application in Humans

Abstract

The protoporphyrin IX-triplet state lifetime technique (PpIX-TSLT) has been recently introduced as the first method to measure mitochondrial oxygen tension (mitoPO2) in living cells and tissues. The current implementation of the technique is based on oxygen-dependent quenching of the delayed fluorescence lifetime of 5-aminolevulinic-acid–enhanced mitochondrial PpIX. It represents a significant step forward in our ability to comprehensively measure tissue oxygenation. PpIX-TSLT is feasible for application in humans and recently we have been able to measure for the first time mitoPO2 in humans. MitoPO2 in intact tissues reflects the balance between oxygen supply and demand at the cellular level. Administration of aminolevulinic acid induces measurable mitochondrial levels of PpIX. PpIX acts as a mitochondrially located oxygen-sensitive dye by emitting a red delayed fluorescence after excitation with a pulse of green light. The lifetime of the delayed fluorescence is inversely related to PO2 by the Stern–Volmer equation. In vivo measurements of mitoPO2 in liver, heart, and skin of rats have revealed surprisingly high values of typically several tens of mm Hg. Clinical measurements of mitoPO2 are possible as demonstrated by cutaneous measurements in healthy volunteers. Applications of PpIX-TSLT in anesthesiology and intensive care medicine might, e.g., be monitoring mitoPO2 as a resuscitation end point, targeting oxygen homeostasis in the critically ill, and assessing mitochondrial function at the bedside. PpIX-TSLT likely also has applications in other fields also, e.g., providing an oxygen-related feedback signal in photodynamic therapy of malignant tumors.

In Brief

Published ahead of print April 16, 2013

Molecular oxygen, dioxygen, or, in short, O₂ is the prime requisite of aerobic life on earth. Diatomic oxygen gas, mostly spoken of simply as “oxygen,” constitutes approximately 21% of the volume of air. The electron configuration of the oxygen molecule has 2 unpaired electrons with the same spin in degenerate orbitals. Therefore, oxygen is paramagnetic, and, in contrast to many other biochemically relevant molecules, the ground state of oxygen is a triplet state. Having a triplet state as ground state is very unusual, but it has the important implication that molecular oxygen does not react directly with many other molecules. In contrast, singlet oxygen is the electronically excited state of molecular oxygen that is highly reactive and a well-known oxygen radical. The amount of oxygen dissolved in a liquid depends on the solubility coefficient of oxygen in the solvent and the partial oxygen pressure (PO2) above the solution. Because the solubility coefficients in biological compartments are not well known, it is common to measure and report oxygen levels as PO2 value in kPa, mm Hg, or Torr (1 kPa = 7.5 mm Hg = 7.5 Torr).

The adequate supply of oxygen by inhalation and subsequent transport to tissues via the circulating blood is a conditio sine qua non for mammalian cells to sustain life. Molecular oxygen is the primary oxidant in biological systems, and its ultimate destination in vivo is the mitochondria where it is used in oxidative phosphorylation. Besides being indispensible for the energy production of our cells, oxygen plays a role in many other biochemical processes and mammalian tissue contains a large number of oxygen-consuming enzymes.¹ For example, oxygen is used for the production of reactive oxygen species that are important in signal transduction.^2,3

Because of the importance of adequate oxygen supply, many techniques have been developed for oxygen measurements in vivo^4,5 to gain insight into the mechanisms of oxygen delivery and use. However, the ultimate goal of measuring intracellular oxygen at the level of the mitochondria has long remained elusive. We introduced the protoporphyrin IX-triplet state lifetime technique (PpIX-TSLT) for measuring PO2 in mitochondria.⁶ The technique is based on the principle of oxygen-dependent quenching of the excited triplet state of 5-aminolevulinic acid (ALA)–induced mitochondrially located PpIX. Currently, PpIX-TSLT is using delayed fluorescence as the means to measure the PpIX triplet-state lifetime.

PpIX-TSLT is not only useful in cultured cells, but the technique is also applicable in tissues and organs in vivo. Unlike other porphyrin-based techniques⁷ used, e.g., for measurement of microvascular PO2,⁸ PpIX-TSLT does not rely on the injection of potentially toxic metalloporphyrin complexes. Therefore, it is applicable in humans, and in principle, PpIX-TSLT could be useful for monitoring mitochondrial oxygenation in the clinical setting. In this article, I will discuss the meaning of mitochondrial oxygen tension (mitoPO2) as a variable and its potential use as a clinically relevant marker in perioperative and intensive care. Furthermore, I will explain the background of the technology, our current technical implementation, the steps we took toward measurement in humans, and some potential clinical applications.

MITOCHONDRIAL OXYGEN TENSION

The functional capacity of tissues and organs relies on the ability of tissue cells to sustain integrity and perform their task-specific work. To do so, tissue cells need an unhampered ability to produce energy in the form of adenosine triphosphate (ATP). To sustain adequate levels of ATP, mammalian cells rely heavily on mitochondrial oxidative phosphorylation. Mitochondria are by far the largest oxygen-consuming system within cells, with mitochondrial respiration estimated to be approximately 90% of mammalian oxygen consumption.⁹ Therefore, mitochondrial respiration can be regarded as the demand side of the oxygen balance, i.e., the balance between oxygen supply and oxygen demand.

On the tissue level, the supply of oxygen to mitochondria is dependent on microvascular blood flow, the amount of hemoglobin, the level of oxygen saturation of hemoglobin, the hemoglobin dissociation characteristics in microvessels, and the diffusion barriers between red blood cells and the cytosol of tissue cells. Oxygen supply needs to meet oxygen demand to prevent hypoxia and cellular metabolic adaptation.¹⁰ However, oxygen supply exceeding oxygen demand leads to tissue hyperoxia and the risk of oxidative stress inducing a cellular adaptive response.¹¹ The regulation of oxygen homeostasis is a field of ongoing research with direct implications for the clinic, e.g., in newborn care.¹² Current clinical practice is aimed at safeguarding adequate oxygen supply to tissues by systemic oxygen administration and macrocirculatory optimization without knowing what actually is “adequate oxygen supply” at the tissue level.¹³ Concerns about negative effects of administering too much oxygen have recently led to the clinical concept of “permissive hypoxia.”¹⁴ However, this concept needs further evaluation before being put in clinical practice.¹⁵ Achieving oxygen homeostasis, i.e., maintaining a steady state of oxygenation and oxygen metabolism within a normal range, might be an ultimate goal in treating our patients, and therefore we need to gain insight in the oxygen balance at the tissue level.

Figure 1 shows a schematic representation of the oxygen balance and some factors influencing this balance at the tissue level. Since the mitochondria are the principal oxygen sink, the oxygen tension in the mitochondria (mitoPO2) is at the lowest end of the blood-to-tissue oxygen gradient. In fact, mitoPO2 is the direct resultant of the amount of oxygen diffusing into the mitochondria and the amount of oxygen consumed by the respiratory chain. In other words, mitoPO2 is a direct indicator of the oxygen balance, which is measured exactly at the place where it matters. Therefore, mitoPO2 is promising for assessing oxygen homeostasis at the tissue level and might be useful for guiding fluid management, red blood cell transfusion, and pharmacologic hemodynamic optimization.

Figure 1: The oxygen balance. Mitochondrial oxygen tension (mitoPO2) is a resultant of mitochondrial oxygen supply and demand. The respiratory chain occupies the demand side, and the biggest part of its oxygen consumption is coupled to synthesis of adenosine triphosphate (ATP). A small part is used for energy dissipation and heat generation by uncoupling via proton leak over the mitochondrial inner membrane.⁹ Oxygen supply is dependent on tissue blood flow (TBF), tissue blood volume (TBV), local hemoglobin saturation (HbO₂), microvascular oxygen tension (μPO2), and factors influencing the oxygen gradient between microvessels and mitochondria (ΔPO2).

MITOPO2 AS TREATMENT TARGET

Hemodynamic targets, such as cardiac output, mean arterial blood pressure, and central venous oxygen saturation, are only modest predictors of tissue perfusion and oxygen supply at the cellular level.¹⁶ This has led to an ongoing search for a perfusion target, or set of targets, to complement the hemodynamic targets in goal-directed resuscitation with, e.g., near-infrared spectroscopy, sidestream dark field video microscopy, and regional capnometry.¹⁷ Obviously, the direct assessment of tissue oxygenation by quantitative measurement of tissue oxygen levels deserves attention. For example, the subcutaneous partial pressure of oxygen (P_scO2) has been shown by Van Esbroeck et al.¹⁸ to be a valuable perfusion variable in the perioperative setting. However, such measurements are technically difficult to perform and are time consuming. In the words of the authors, “Peripheral tissue oximetry might become part of daily clinical monitoring procedures, but only if it is made simpler and quicker to perform.”

PpIX-TSLT as a technique to measure mitoPO2 has several advantages that make it highly suitable as a new approach for peripheral tissue oximetry. It provides quantitative information in a well-defined tissue compartment and gives insight into the oxygen balance at the cellular level. The technical details of PpIX-TSLT will be explained in detail below, but essentially it is a noninvasive optical technique that is quantitative, technically robust, and needs no recalibration. Instead of invading the tissue with an oxygen-sensing probe (like an oxygen electrode or optode), it uses the oxygen-dependent optical properties of the exogenously enhanced mitochondrial protein PpIX (priming) to measure mitoPO2 in intact tissue. PpIX-TSLT has been technically developed to a stage that enables application in humans, and recently our research group obtained permission by our IRB to start cutaneous mitoPO2 measurements in healthy volunteers. In practice, cutaneous mitoPO2 measurements need some planning and preparation for priming of the skin, but otherwise it really is the simpler and quicker to perform technology requested by Van Esbroeck et al.¹⁸

Priming of the skin can be performed easily by topical application of ALA,^19,20 but systemic administration of ALA is clinically feasible and in principle allows for mitoPO2 measurements in most tissues and organs. For the first time, a technology allows measurement of oxygen where it really matters, inside the tissue cells at the level of the mitochondria. It will be exciting to identify and evaluate clinical applications of this novel technology. For the time being, we will focus our research on the clinical usability of cutaneous mitoPO2 measurements as novel resuscitation end points and for guidance of oxygen administration and hemodynamic optimization. Other potential applications are discussed below, after the description of the background and technical implementation of mitoPO2 measurements.

PORPHYRINS

Porphyrins belong to the most abundant molecules on earth and are a group of organic compounds that appear in nature in a large variety. Porphyrins are essential for life as we know it, because of their key role in processes related to oxygen production, oxygen transport, and oxygen use.²¹ Porphyrins are cyclic macromolecules composed of 4 modified pyrrole subunits that form a highly conjugated system. The parent compound is porphine (Fig. 2A), and substituted porphines are called porphyrins.

Figure 2: Examples of porphyrins. A, Porphine, the parent porphyrin. B, Heme, an iron containing porphyrin. C, Chlorophyll, a magnesium-containing porphyrin. D, Palladium-meso-tetra(4-carboxyphenyl)porphine, commonly referred to as “Pd-porphyrin,” a commonly used phosphorescent oxygen-sensitive dye.

As ligands, porphyrins form easily into complexes with metallic ions like iron and magnesium. For example, some iron-porphyrin complexes are called hemes (Fig. 2B). Hemoglobin is an example of a heme-containing protein that binds oxygen and functions as an oxygen carrier in blood.^22,23 Chlorophyll is an example of a magnesium-containing porphyrin (Fig. 2C). Chlorophyll is the green compound in leaves that absorbs light for photosynthesis, providing us with oxygen and energy.²⁴ The metallic ion defines much of the specific photochemical properties of a porphyrin. For example, incorporation of palladium into the porphyrin ring can provide phosphorescent properties, like in palladium-meso-tetra(4-carboxyphenyl)porphine (Fig. 2D).

Porphyrins typically are extremely good absorbers of light in the visible range, and therefore porphyrin-metal complexes often have intense and dark colors. The absorption of a photon leads to photoexcitation of the porphyrin-metal complex. Photoexcitation can lead to population of an excited triplet state of which the energy can, e.g., be used for photosynthesis.²⁵ Oxygen, having a triplet state as ground state, is a very effective quencher of this excited triplet state. In the process of quenching, energy is transferred to oxygen leading to the formation of singlet oxygen, the basis of porphyrin-based photodynamic therapy.²⁶

TRIPLET-STATE QUENCHING

A so-called Jablonski diagram visualizes the different electronic states in an atom or molecule. Figure 3 shows the Jablonski diagram of porphyrin and oxygen and their interaction. After excitation of porphyrin to the first singlet state (S₁), population of the triplet state (T₁) can occur by intersystem crossing, a process in which the electron is relaxed to the triplet state by changing its spin orientation. This process occurs without emission of a photon. The probability that a porphyrin molecule in the T₁ state relaxes to the S₀ state by spontaneous relaxation is denoted by rate constant k_s. Relaxation to S₀ can also occur by collision with an oxygen molecule during which oxygen absorbs the energy from the porphyrin. This process is called “triplet-state quenching” and results in relaxation of the porphyrin without emission of a photon.

Figure 3: Jablonski diagram of states and state transitions of porphyrins and oxygen. Absorption (i), and fluorescence (ii), and radiationless transitions (iii), phosphorescence (iv), energy transfer (v), internal conversion (vi), intersystem crossing (vii), and S₀, S₁, S₂, and S₃ denote the groundstate and the first 3 excited singlet states of porphyrin, respectively. T₁ denotes the first excited triplet state of porphyrin.³O₂ and ¹O₂ denote the triplet ground state of oxygen and excited singlet oxygen, respectively.

Quenching of the triplet state is a process that makes the triplet-state lifetime dependent on the collision frequency, and thus on oxygen concentration. The collision frequency is determined by the amount of oxygen and the chance that a single oxygen molecule causes a quenching event, defined by the Smoluochowski equation:

in which k_q is known as the quenching constant, N is Avogadro’s number, γ the quenching efficiency, D_o and D_f are the diffusion coefficients of oxygen and the porphyrin, respectively, and R_o and R_f are the quenching radius of oxygen and the porphyrin, respectively.

The rate of relaxation of T₁ to S₀ is thus determined by both the transition probability by spontaneous relaxation (k_s) and the quenching probablility (k_q). The decay rate of porphyrin molecules in the excited triplet state after excitation with a pulse of light is given by the following differential equation:

where [T₁] denotes the amount of porphyrin molecules in the excited triplet state, and PO2 is the oxygen tension in the surrounding medium. Under ideal circumstances, i.e., when the excitation pulse duration is much shorter than the triplet decay time, the solution of this differential equation yields:

where [T₁]₀ denotes the amount of porphyrin molecules in the excited triplet state at t = 0, i.e. immediately after the excitation pulse. Equation (3) can be rewritten in the form:

in which

This last equation is known as the Stern–Volmer relationship, in which τ is the triplet-state lifetime and τ₀ = 1/k_s is the time constant of the decay of the triplet state in the absence of oxygen, i.e., the decay time of spontaneous relaxation.

ASSESSING TRIPLET-STATE LIFETIME

From the Stern–Volmer relationship (Equation (5)), it is clear that if we are able to measure the triplet-state lifetime, by any means, this actually allows measurement of the PO2 in the medium containing the porphyrin. Figure 4 shows an overview of optical modalities to measure state transitions that allow direct measurement of the triplet-state lifetime. These modalities are phosphorescence, delayed fluorescence, and triplet-triplet absorption.

Figure 4: Modalities to measure the triplet-state lifetime. Phosphorescence (A), delayed fluorescence after bidirectional intersystem crossing (B), and triplet-triplet absorption (C). S₀ and S₁ denote the groundstate and the first excited singlet states of porphyrin respectively. T₁ and T₂ denote the first and second excited triplet states of porphyrin respectively.

If the T₁ state relaxes directly to the S₀ state through emission of a photon, this radiation is called phosphorescence. In contrast to fluorescence (emission of a photon due to relaxation of S₁ to S₀), phosphorescence is relatively long-lived (typically several orders of magnitude longer than fluorescence), and its spectrum is shifted toward the red. The latter is due to the fact that, in case of phosphorescence, some energy is lost in the process of intersystem crossing, and therefore the energy of the T₁ state is lower than of the S₁ state. Because of this redshift, phosphorescence can be optically separated from prompt fluorescence. This makes measurement of phosphorescence the least complicated way to measure the T₁ lifetime. Unfortunately, not all porphyrins show detectable phosphorescence, but some metalloporphyrins have high phosphorescence yield which makes them excellent probes for oxygen measurements.⁷

If the T₁ state relaxes to the S₀ state via the S₁ state by a process called bidirectional intersystem crossing and the relaxation of S₁ to S₀ leads to emission of a photon, this radiation is called delayed fluorescence. In contrast to phosphorescence, delayed fluorescence has the same spectrum as prompt fluorescence and cannot be optically separated from fluorescence. Prompt fluorescence tends to overwhelm delayed fluorescence in intensity, and in practice, this easily leads to saturation of detection systems. Nevertheless, with some precautions delayed, fluorescence is a useable mode for measuring the T₁ lifetime. A practical example is the measurement of mitoPO2 by delayed fluorescence of PpIX,⁶ as discussed in detail below. Phosphorescence and delayed fluorescence are both examples of the physical phenomenon known as “delayed luminescence.”

A method that does not rely on the detection of emitted photons but on transient changes in absorption is triplet-triplet absorption. The idea is that absorption from T₁ to T₂ can only take place during population of the T₁ state. Therefore, an excitation pulse will temporarily increase absorption at the wavelength corresponding to T₁–T₂ transition, and the transient change in absorption will decay according to the T₁ lifetime. A drawback of this technique is that it requires an extra light source, making its application more cumbersome than the emission measurements. Nevertheless, triplet-triplet absorption is useful for nonradiating probes.²⁷

OXYGEN-DEPENDENT DELAYED LUMINESCENCE

The method of measuring PO2 in biological systems by means of oxygen-dependent quenching of phosphorescence was introduced by Vanderkooi et al.⁷ >2 decades ago. Complexes of porphyrins with certain heavy metals show high phosphorescence yield and are very efficient oxygen probes. Palladium-meso-tetra(4-carboxyphenyl)porphine (Pd-porphyrin), bound to albumin before injection, has become a standard phosphorescent dye for microvascular PO2 measurements in vivo.^28,29

Oxygen measurements by means of oxygen-dependent quenching of delayed luminescence lifetimes are based on the principle that delayed luminescence intensity is proportional to the amount of populated triplet states. As long as the excitation pulse is much shorter than the delayed luminescence lifetime, or after application of excitation pulse deconvolution,³⁰ the delayed luminescence signal in case of a homogenously distributed oxygen pressure is:

where I(t) is the delayed luminescence intensity over time, I₀ is the initial delayed luminescence intensity directly after the excitation pulse, and τ now denotes the delayed luminescence lifetime. The PO2 can be calculated from τ using the Stern–Volmer relationship (Equation (5)). From this relationship, it is clear that a higher PO2 results in more quenching and therefore a shorter lifetime (Fig. 5).

Figure 5: Principle of delayed luminescence lifetime measurements. Excitation with a pulse of light in a sample containing a suitable oxygen-sensitive porphyrin dye induces delayed luminescence. The delayed luminescence lifetime depends on the oxygen tension in the sample. The inserted equation is the Stern–Volmer relationship in which τ is the delayed luminescence lifetime, τ₀ is the lifetime at zero oxygen, and k_q is the quenching constant.

PROTOPORPHYRIN IX

Oxygen-dependent quenching of phosphorescence has proven itself to be a very useful technology for biological oxygen measurements, ranging from single cell microscopy³¹ to intravital microscopy^32,33 to fiber-based measurements on organs like heart³⁴ and kidney.³⁵ We have used it to measure microvascular oxygen tension in various animal models studying pathophysiological mechanisms in perioperative and intensive care.^36–39 The development of new porphyrin-based oxygen-sensitive dyes is a field of ongoing research.^40–43 However, the need for injection of the phosphors and concerns about potential toxicity of the metal-porphyrin complexes limits in vivo use of such probes to animal research.

This latter limitation could potentially be overcome by using an endogenous porphyrin as oxygen-sensitive dye. Several years ago, this was the idea behind our search for an endogenously synthesized porphyrin with easily measurable and calibratable oxygen-sensitive optical properties. PpIX appeared to be a very suitable candidate. PpIX is an endogenously present porphyrin that is synthesized inside the mitochondria.⁴⁴ PpIX is the final precursor of heme in the heme biosynthetic pathway, and the conversion of PpIX to heme is a rate-limiting step. Administration of the porphyrin precursor ALA to cells, tissues, and animals substantially enhances the PpIX concentration.⁴⁵ Furthermore, PpIX possesses a triplet state that reacts strongly with oxygen.⁴⁶

While PpIX shows no detectable phosphorescence,⁴⁷ we discovered that it does emit a weak oxygen-sensitive delayed luminescence (Fig. 6), which appeared to be delayed fluorescence.⁶ The delayed fluorescence lifetime relates quantitatively to the oxygen tension and can be calibrated. Since ALA-enhanced PpIX is located in the mitochondria, PpIX acts as an intramitochondrial oxygen-sensitive dye.

Figure 6: Spectra of prompt and delayed luminescence of protoporphyrin IX (PpIX). A, Prompt fluorescence. B, Delayed luminescence from an air-saturated sample measured at 30 and 100 µs after the excitation flash. C, Delayed luminescence measured in an anoxic sample at 30 µs, 100 µs, and 1 ms after the excitation flash. All spectra were measured in a 10 µM PpIX solution containing 2% bovine serum albumin and are shown in arbitrary units (a.u.). Reproduced from Mik et al.⁶

MEASURING MITOCHONDRIAL PO2

Oxygen-dependent quenching of delayed fluorescence of PpIX is the first technique that allows measurement of mitoPO2 in living cells.⁶ We have demonstrated that the technique is feasible for measuring mitoPO2 in vivo and have now validated and calibrated the technique in various organs and tissues.^19,48,49 A schematic outline of the technique showing the steps from ALA administration to detection of oxygen-sensitive delayed fluorescence is shown in Figure 7. The basic principle of the delayed fluorescence measurements is straightforward and very similar to phosphorescence lifetime measurements. However, the relatively weak delayed fluorescence signal is easily overwhelmed by intense prompt fluorescence, which leads to saturation of sensitive photodetectors and electronics. Pulsed excitation with a q-switched laser in combination with time-gated detection overcomes these problems.

Figure 7: Basic principle of mitochondrial oxygen tension (mitoPO2) measurements. Administration of the porphyrin precursor 5-aminolevulinic acid (ALA) enhances mitochondrial protoporphyrin IX (PpIX) levels (A) due to the fact that incorporation of iron in the PpIX molecule is a rate-limiting step. PpIX is the final precursor of heme, and its molecular structure resembles heme without the central iron atom (B). The use of fiber optics allows mitoPO2 measurements on the surface of tissues and organs without local destruction of the tissue (C). A short pulse of green light, with a pulse duration of 7 ns at a wavelength of 510 nm, provided by a pulsed laser causes excitation of PpIX in the excitation volume, and the emitted delayed fluorescence is oxygen-dependent (D). PBG = porphobilinogen; UPIII = uroporphyrinogen III; CPIII = coproporphyrinogen III; PpGIX = protoporphyrinogen IX.

Extensive descriptions of delayed luminescence setups that have been successfully used for mitoPO2 measurements have been published.^6,20,49,50 Here, I will give a short outline of our current setup and the methods we use for analysis of complex delayed fluorescence signals obtained in vivo.

Setup

Currently, we use a compact computer-controlled tunable laser (Opolette 355-I, Opotek, Carlsbad, CA), providing pulses with a specified duration of 4 to 10 nanoseconds and typically 2 to 4 mJ/pulse over the tunable range of 410 to 670 nm, as an excitation source. This laser is coupled into a multimode fiberoptic with a core diameter of 1000 µm by a Fiber Delivery System (Opotek). This fiber acts as interconnect between excitation light source and the actual measuring fiber (varies with specific task). For example, for measuring mitoPO2 at the surface of organs, we use a bifurcated reflection probe (FCR-7IR400-2-ME, Avantes b.v., Eerbeek, the Netherlands). Excitation light levels at the output of the excitation branch are typically in the order of 50 to 200 µJ/pulse.

The PpIX signal is detected by a gated microchannel plate photomultiplier tube (MCP-PMT R5916U series, Hamamatsu Photonics, Hamamatsu, Japan), having a quantum efficiency of 24% at 650 nm. The MCP-PMT is mounted on a gated socket assembly (E3059-501, Hamamatsu Photonics) and cooled to −30°C by a thermoelectric cooler (C10373, Hamamatsu Photonics). The MCP-PMT is operated at a voltage in the range of 2300 to 3000 V by a regulated high-voltage DC power supply (C4848-02, Hamamatsu Photonics). The emission branch of the reflection probe is coupled to the detector by in-house built optics. The MCP-PMT is configured for the “normally on” mode, and the PpIX emission light is filtered by a combination of a 590 nm longpass filter (OG590, Newport, Irvine, CA) and a broadband (675 ± 25 nm) bandpass filter (Omega Optical, Brattleboro, VT).

The output current of the photomultiplier is voltage-converted by an in-house built amplifier with an input impedance of 440 ohm, 400 times voltage amplification, and a bandwidth of approximately 20 Mhz. Data acquisition is performed by a PC-based data-acquisition system containing a 10 megasamples per second (MS/s) simultaneous sampling data acquisition board (NI-PCI-6115, National Instruments, Austin, TX). The data acquisition runs at a rate of 10 MS/s, and typically, a number of laser pulses (repetition rate 20 Hz) are averaged before analysis to increase signal-to-noise ratio. Control of the setup and analysis of the data are performed with software written in LabView (Version 8.6, National Instruments).

Signal Analysis

In case of a homogeneously distributed quencher concentration, the delayed fluorescence lifetime will decay monoexponentially according to Equation (6). However, the oxygen distribution in tissue is in general not homogenous. This is due to local differences in oxygen supply–demand and the presence of oxygen gradients. In case of a nonhomogenous mitoPO2, the delayed fluorescence signal can be described by an integral over an exponential kernel:

where f(λ) denotes the spectrum of reciprocal lifetimes that should be determined from the finite data set y(t). The underlying oxygen distribution can be recovered by assuming that the delayed fluorescence signal can be described by a sum of rectangular distributions with adequately small chosen width (2δ), resulting in the following fit equation:⁵¹

where Y(t) is the normalized phosphorescence data, k₀ is the first-order rate constant for delayed fluorescence decay in the absence of oxygen, k_q is the quenching constant, and w_i is the weight factor for the according bin with central PO2 Q_i and width 2δ (w_i ≥ 0 and Σw_i = 1). The basic principle is shown in Figure 8. This approach was used for the recovery of microvascular PO2^36–39 and mitoPO2^48,49 histograms.

Figure 8: Delayed fluorescence in vivo. Tissues are comprised of regions with different mitochondrial oxygen tension (mitoPO2; A), each with a different delayed fluorescence lifetime. The delayed fluorescence signal obtained from tissue is a summation of the individual contributions (B). The photometric signal therefore contains a lifetime distribution that can be retrieved by lifetime deconvolution. The retrieved lifetime distribution can be mathematically converted into a mitoPO2 histogram (C), which represents the original mitoPO2 distribution in the tissue.

As an alternative to the recovery of detailed lifetime distributions, especially for fast real-time signal analysis, it is possible to extract the mean mitoPO2 and an estimate of the variance in mitoPO2 directly from the photometric data. This can be achieved by fitting distributions of quencher concentration to the delayed luminescence data. The fitting function for a simple rectangular distribution with a mean PO2 Q_m and a PO2 range from Q_m – δ till Q_m + δ is:³³

where Y_R(t) is the normalized delayed fluorescence data, k₀ is the first-order rate constant for delayed fluorescence decay in the absence of oxygen, k_q is the quenching constant, and δ is half the width of the rectangular distribution. In terms of quenching constants and the Stern–Volmer relationship, Equation (9) can be rewritten as:

where <PO2> is the mean PO2 within the sample volume and τ₀ the lifetime in the absence of oxygen. The second moment of the assumed rectangular (or uniform) distribution, the variance (σ²), can be calculated from δ as:

Fitting Equation (10) to the photometric signal directly provides mean mitoPO2 and an estimation of the heterogeneity of mitoPO2 within the sample volume. Obviously, a rectangular distribution is only a rough first approximation to many heterogeneous systems. Nevertheless, retrieval of mean mitoPO2 has been proven reliable and robust. However, changes in standard deviation (square root of the variance) should be interpreted with care as long as it is uncertain whether the simple rectangular distribution is a feasible approximation to the real mitoPO2 distribution. This approach was used in both phosphorescence and delayed fluorescence lifetime measurements.^20,35,50

MITOCHONDRIAL PO2 IN VIVO

Oxygen-dependent quenching of delayed fluorescence of ALA-induced PpIX has now been calibrated and evaluated for in vivo mitoPO2 measurements in various tissues, i.e., liver,⁴⁹ heart,⁴⁸ and skin.¹⁹ Importantly, no significant differences in calibration constants have been found between the tissues. At a temperature of 37°C, the values for the quenching constants have been found to be approximately k_q = 830 mm Hg·s⁻¹ and τ₀ = 0.8 milliseconds, despite being determined in completely different types of tissues. Therefore, the technique likely is universally applicable to all tissues as long as delayed fluorescence is detectable.

Oxygen transport from the microvessels to the mitochondria is driven by a concentration gradient. Because this gradient is caused by the oxygen sink in the mitochondria, due to oxygen consumption, the mitochondrial compartment should be at the lowest end of oxygen tensions compared with other tissue compartments.⁵² Classically, a gradual decline in PO2 is anticipated from the vascular, interstitial, and cytosolic compartments. Most studies report rather low PO2 values ranging from 10 to 17 mm Hg for vascular and interstitial compartments.^53–55 As a consequence, mitoPO2 has been derived and estimated to be in the order of several mm Hg.^56–58 However, with the advent of new techniques with less invasiveness and increased accuracy, it appears that the levels of oxygen in tissue are much higher than originally thought.⁵⁹

For example, classical studies using oxygen electrode measurement in liver reported tissue oxygen tension to be approximately 10 to 20 mm Hg.⁶⁰ Due to the oxygen gradient induced by cellular respiration, the intracellular PO2 was assumed to be well below this value. However, more modern techniques like phosphorescence-based microvascular PO2 measurements show values around 60 mm Hg.⁵⁰ Such high PO2 values have recently been confirmed by even less-invasive 19F nuclear magnetic resonance studies.^61,62 The sinusoidal-to-cell oxygen gradient is small due to the hepatic architecture, a value of approximately 5 mm Hg has been reported,⁶³ and the intracellular oxygen gradient is only a few mm Hg.⁶ Taken together, this suggests that intracellular and mitochondrial PO2 levels in the liver should be in the order of several tens of mm Hg. In the heart, microvascular PO2 has also been reported to be in the order of 60 to 70 mm Hg.³⁴ While oxygen gradients in the heart can be expected to be steeper than in the liver due to different microvascular architecture, larger cells, and high oxygen consumption, extremely low mitochondrial oxygen levels are very unlikely. In normal hearts, the levels of hypoxia-inducible factor (HIF) are not enhanced whereas HIF degradation by prolyl hydroxylases becomes oxygen-dependent at PO2 levels as high as 20 to 40 mm Hg.^64,65 These more recent findings suggest that the classical estimations of mitoPO2 in vivo might be too low.

Indeed, our first direct measurements of mitoPO2 in vivo also indicated that mitochondrial oxygen levels might well exceed a few mm Hg. While mitoPO2 appeared to be highly heterogeneous (Fig. 9), average mitoPO2 in rat liver was around 45 mm Hg,⁴⁹ in rat heart around 35 mm Hg,⁴⁸ and in rat skin around 50 mm Hg.¹⁹ Also, straightforward measures to alter systemic oxygenation, e.g., changing fractional inspired oxygen levels, had direct and measurable effects on mitoPO2. It is yet unknown what the consequences are of these new insights, but our findings put in vivo mitoPO2 in the PO2 range associated with the concept of “oxygen conformance of metabolism.”^66,67

Figure 9: Examples of in vivo mitochondrial oxygen tension (mitoPO2) measurements. A, A typical example of a mitoPO2 histogram and average mitoPO2 measured in rat liver. B, Atypical example of a mitoPO2 histogram and average mitoPO2 measured in rat heart. Both examples were measured in anesthetized and mechanically ventilated rats.^{48 , 49}

Oxygen conformance of metabolism is a form of cellular metabolic adaptation in which cellular oxygen consumption is downregulated when cells are subjected to only moderate oxygen deprivation. Downregulation of respiration is associated with decreased ATP production, cellular work, and function. Oxygen conformance of metabolism is a phenomenon that occurs over a broad PO2 range starting below approximately 70 mm Hg and is therefore likely to be a form of physiological metabolic control by oxygen. Oxygen conformance is mediated by both HIF-dependent and HIF-independent pathways.⁶⁸ The role of oxygen in the concept of oxygen conformance is completely different than its role as a passive bystander in the classical view of cellular respiration. In the classical view, cellular respiration is unaffected by oxygen levels until PO2 decreases below 2 to 3 mm Hg,^69–72 because of the high affinity of the respiratory chain for oxygen. In the modern view, oxygen availability is directly influencing cellular metabolism and function, and therefore any therapeutic intervention that causes a change in PO2 at the tissue level potentially has such effects. Therefore, the actions of medical doctors to achieve “adequate tissue oxygenation” might in the end, unintentionally, alter cellular metabolism and function due to alterations in tissue oxygen levels, e.g., by superfluous oxygen administration or vasopressor therapy. The optimal upper and lower limits of oxygenation need to be defined and probably individualized to be able to achieve therapeutic oxygen homeostasis.¹² Direct measurement of the oxygen balance at the tissue level by means of mitoPO2 measurements might provide a useful tool in this respect.

THE STEP TOWARD HUMANS

Compared with other techniques to measure oxygen at the tissue level, oxygen-dependent quenching of delayed fluorescence of PpIX has some distinct advantages. It is quantitative, and, unlike oxygen electrodes, once calibrated it does not need recalibration at the time of usage. As opposed to oxygen saturation measurements based on near-infrared spectroscopy, the measurement site within the tissue is well defined. Therefore, the signal is not sensitive to changes in vessel density, like capillary recruitment. Furthermore, the technique relies on lifetime measurements instead of intensity measurements and therefore is highly insensitive to changes in tissue optical properties occurring, e.g., in case of venous congestion.

In the current form, mitoPO2 measurements need administration of ALA to induce enough mitochondrial PpIX for detection of the weak delayed fluorescence signal. ALA is a precursor in porphyrin synthesis, and its application induces the accumulation of PpIX inside mitochondria.⁷³ ALA is clinically used in photodynamic diagnosis and therapy of cancer.^45,74,75 While systemic administration of ALA in itself is safe, the transient photosensitization of the skin requires patients to limit exposure to (sun)light in the days after treatment. Obviously, this poses constraints to the use of systemically administered ALA in the general population. As an alternative to systemic administration, ALA can be topically applied.^76,77 Recently, we demonstrated that topical application of ALA cream to the skin induces oxygen-dependent delayed fluorescence in rats and humans.²⁰ The recent validation of the quenching constants in skin¹⁹ allowed for quantitative measurements of mitoPO2 in humans. Figure 10 shows an example of mitoPO2 measurements in skin overlying the sternum of a healthy volunteer.

Figure 10: Mitochondrial oxygen tension (mitoPO2) measurements in human skin. Examples of delayed fluorescence traces measured in the skin overlying the sternum of a healthy volunteer after topical application of 5-aminolevulinic acid (ALA) using an ALA-containing patch (Alacare, Photonamic GmbH & Co, Wedel, Germany). The traces represent measurements before and after 30 s of cessation of microvascular blood flow by applying local pressure with the measuring probe. The grey curves are the delayed fluorescence signals, and the solid lines are the results of a rectangular distribution fit procedure.

CLINICAL APPLICATION

While the need to use ALA for enhancing mitochondrial PpIX levels is a disadvantage of PpIX-TSLT in its current form, this does not preclude clinical use of the technique. ALA can be safely administered systemically and applied topically to humans and is in clinical use for, e.g., photodynamic diagnosis and therapy.⁷⁴ In dermatology, its use has been investigated, e.g., for the treatment of benign lesions like acne vulgaris.^78,79 In phototherapy, the treatment effect depends on the production of singlet oxygen by photoactivation of PpIX. Excitation of PpIX induces apoptosis in cells as a result of oxygen-radical formation.⁷⁵ In general, this requires illumination with continuous light and a high cumulative light dose. In contrast, PpIX-TSLT uses short-pulsed excitation and a total light dosage, i.e., orders of magnitude less than used for photodynamic therapy. Importantly, we did not find increased apoptosis using PpIX-TSLT in cell lines.⁶ While PpIX-TSLT is likely to be safe, the effects of phototoxicity after PpIX induction should always be considered a potential risk requiring a risk and safety assessment in any intended application.

One obvious application of PpIX-TSLT could be in PpIX-based photodynamic therapy. Because ALA is already administered to these patients and tumors are preferential sites for ALA uptake and PpIX conversion, PpIX-TSLT is potentially a powerful technique to measure oxygen levels in tumors. This is of relevance since oxygen is needed to produce singlet oxygen, making photodynamic therapy less effective at low oxygen levels. Measuring tumor oxygenation during photodynamic therapy using delayed fluorescence of PpIX is regarded as a potential means to guide the light dose and optimize treatment outcome.^80,81 Oxygen-dependent delayed fluorescence from ALA-induced PpIX has been confirmed in this context.⁸¹

For perioperative and intensive care medicine, PpIX-TSLT can likely be developed into a useful monitoring technology. While measuring mitoPO2 in itself might prove to be useful for optimizing, e.g., hemodynamic status or guiding resuscitation of the critically ill, PpIX-TSLT enables another unique possibility. Optical porphyrin-based oxygen measurements are nondestructive at the tissue level, and the measurement volume within the tissue can be chosen to be very small. This allows a functional optical biopsy in intact tissue in which not only mitoPO2 can be measured but also local cellular respiration.⁸² Currently, 1 of the aims of our research group is to see whether PpIX-TSLT can be used for in vivo respirometry.⁸³Figure 11 shows 3 examples of oxygen disappearance curves measured in skin during creation of microvascular stop-flow conditions by applying local pressure with the measuring probe. Appropriate analysis of such curves should allow assessment of variables of mitochondrial respiration. If this proves successful, we will aim to develop PpIX-TSLT into a monitor for mitochondrial function at the bedside. This would allow direct assessment of mitochondrial dysfunction in the critically ill.^84,85 Currently, it is difficult to distinguish clinically between problems of oxygen supply or oxygen use. For example, high blood lactate levels in sepsis can be caused by tissue hypoxia, mitochondrial dysfunction, or diminished liver clearance. The combination of measuring mitoPO2 and cellular respiration in tissues at the bedside could provide valuable information and hopefully aid in the development and implementation of novel strategies for the treatment of severe sepsis.

Figure 11: Examples of oxygen disappearance curves in skin. Oxygen disappearance curves measured by delayed fluorescence of protoporphyrin IX (PpIX) in skin of rat, humans, and pig. Measurements in rat and pig were performed after topical application of 5-aminolevulinic acid (ALA) cream. The measurement in a healthy volunteer was performed after topical application of ALA using an ALA-containing patch (Alacare, Photonamic GmbH & Co, Wedel, Germany). Applying local pressure with the measurement fiber induced microvascular stop-flow within the tissue directly under the fiber.

THE PATH FORWARD

With the discovery of the oxygen-dependent optical properties of PpIX and the subsequent development of PpIX-TSLT for measuring mitochondrial oxygen tension, we created a novel means to take a new look at old problems. Maintaining adequate tissue oxygenation remains 1 of the most significant tasks of anesthesiologists and intensive care physicians. That prolonged high levels of oxygen are toxic and cause, e.g., retrolental fibroplasia, and lung fibrosis have been recognized for many decades.⁸⁶ However, with the recent findings that even mild hyperoxia might negatively influence clinical outcome,^87,88 it became evident that we still do not know every aspect of altered tissue oxygenation and the subsequent response on the cellular level. We will aim at further developing and applying PpIX-TSLT for mitoPO2 measurements in preclinical animal models that are relevant to answering questions regarding perioperative and intensive care medicine.

Next to application in animal models, the usability of PpIX-TSLT in humans remains unanswered and has to be further explored. Our group recently finished a first clinical prototype of PpIX-TSLT and received approval of our IRB to start research in healthy volunteers. Hopefully, this will prove to be an important first step toward clinical application of the technique. If proven to be feasible and safe, it will be exciting to research the spectrum of clinical usability of mitoPO2 measurements. While several applications are anticipated outside our own field, our main interest of course lies in applying the technology for the benefit of patients in perioperative, acute, and intensive care medicine. Being able to accurately measure both mitoPO2 and mitochondrial oxygen consumption at the cellular level might be the basis for novel devices aimed at cellular monitoring at the bedside.

DISCLOSURES

Name: Egbert G. Mik, MD, PhD.

Contribution: This author wrote the manuscript.

Attestation: This author approved the final manuscript.

Conflicts of Interest: Egbert G. Mik is founder and shareholder of Photonics Healthcare B.V., a company aimed at making the delayed fluorescence lifetime technology available to a broad public. Photonics Healthcare B.V. holds the exclusive licenses to several patents regarding this technology, filed and owned by the Academic Medical Center in Amsterdam and the Erasmus Medical Center in Rotterdam, the Netherlands.

This manuscript was handled by: Dwayne R. Westenskow, PhD.

ACKNOWLEDGEMENTS

This work is financially supported by a Life Sciences Pre-Seed Grant (grant no 40-41300-98-9037) from the Netherlands Organization for Health Research and Development (ZonMW) and, in part, by the Young Investigator Grant 2009 (awarded to E.G.M.) from the Dutch Society of Anesthesiology.