OXYGEN DEPENDENT QUENCHING OF PHOSPHORESCENCE: A PERSPECTIVE D.F. Wilson Department of Biochemistry and Biophysics, Medical School University of Pennsylvania Philadelphia, PA 19104 U.S.A. SUMMARY Oxygen quenches phosphorescence by energy transfer from the phosphor when oxygen molecules collide with molecules of the phosphor in the excited triplet state. Thus increasing oxygen pressure causes an increase in the rate of decay of phosphorescence (shorter lifetimes) and a decrease in total phosphorescence intensity. Phosphors have been selected which decay with a single exponential and for which the relationship between phosphorescence lifetime and oxygen pressure is quantitatively described by the Stern-Volmer equation. The use of phosphorescence lifetime as the measure of oxygen pressure makes the method insensitive to the absorbance changes of other chromophores in the system. This method has permitted quantitative, rapid (less than 10 msec) and sensitive (to less than 10.8 Torr) measurements of oxygen pressure in suspensions of cells or subcellular organells. In tissues, oxygen pressure has been evaluated by measuring phosphorescence using an intensified CCD camera. Maps of oxygen pressure in the vasculature of the cortex of the brain and of other tissues demonstrate the method is limited only by the optics of the system and resolutions of a few microns are readily attained.
INTRODUCTION The oxygen dependent quenching of phosphorescence has been shown to accurately measure oxygen consumption by suspensions of mitochondria and cells from air saturation to less than 10.8 torr (see for example Vanderkooi et al, 1987; Wilson et al, 1988; Robiolio et al, 1989, Rumsey et al, 1990; Wilson et al, 1991). The oxygen pressure dependence of the rate of oxygen consumption was determined for suspensions of isolated mitochondria (Wilson et al, 1988) and cells (Robiolio et al, 1989; Rumsey et al, 1990). Video imaging of phosphorescence has been shown to be a valid method for obtaining two dimensional maps of oxygen pressure in tissues (see Rumsey et al, 1988; 1992; Wilson et al, 1991; 1992A; 1992B; Shonat et al, 1992).
Oxygen Transport to Tissue XIV. Edited by W. Erdmann and D.E Bruley, Plenum Press, New York, 1992
195
METHODS AND MATERIALS Oxygen dependent quenching of phosphorescence provides a very sensItIve measure of oxygen pressure in the environment of phosphorescent molecules. The Pdporphyrins are particularly useful because intersystem crossing is rapid enough to quantitatively convert the exCited singlet state to the triplet state. As a result, these compounds show little or no fluorescence and the phosphorescence quantum efficiency is high. Thus the phosphorescence lifetime can be readily measured not only using conventional photomultipliers but also using intensified video cameras. Photomultipliers are single detectors and all the light collected from the phosphorescent sample is concentrated on one detector. Intensified video cameras function as arrays of detectors, in the case of the camera used in the present studies this was a 512 x 480 array, and each of these must receive sufficient light energy to make a reliable measurement of light intensity. If the detectors are of equal sensitivity, the video camera requires 245,000 times (512 x 480) as much light as the photomultiplier for measurements of equal accuracy. Phosphorimeters appropriate for measuring phosphorescence lifetimes, and thereby oxygen pressure, in aqueous media including suspensions of biological material, as well as in local regions of tissue in vivo have been described in some detail by Green et al (1988) and Pawlowski and Wilson (1992). The present paper will, therefore, focus on the technology used to obtain two dimensional images of oxygen distribution in tissue. Oxygen dependent quenching of phosphorescence of selected compounds can be quantitatively described by the Stem-Volmer relationship: (1)
For the Pd complex of tetra-(4-carboxyphenyl) porphine bound to bovine serum albumin and at physiological pH and 38°C, for example, kQ has a value of 325 Torr-! sec-! and TO is 600 usee (see Wilson et al, 1991; Pawlowski and Wilson, 1992). 1° and 1 are the phosphorescence intensities and TO and T the phosphorescence lifetimes at zero oxygen pressure and at an oxygen pressure P02 respectively. Phosphorescence lifetime is a more accurate measure of oxygen pressure because the lifetime measurements are independent of probe concentration and of illumination light intensity. Moreover, lifetime measurements are essentially unaffected by changes in the absorbance of other chromophores in tissues, such as hemoglobin, myoglobin and the cytochromes. The measurements are therefore independent of the degree of oxygenation of hemoglobin and myoglobin as well as the state of reduction of cytochromes. Calibration of the oxygen probes If the values of the phosphorescence lifetime are determined both at a, known oxygen pressure and at zero oxygen pressure, the quenching constant, kQ can be calculated. In order to obtain the necessary measurements, the phosphorescence lifetimes have been measured in the samples either equilibrated with air or treated with glucose oxidase and glucose to remove the oxygen and sealed in glass chambers. In each case, the measurements have been made at temperatures from 21°C to 38°C (see Wilson et al, 1991; Pawlowski and Wilson, 1992).
196
Effect of temperature and pH on phosphorescence lifetime and quenching constant of selected oxygen probes The effects of temperature and pH on different probes have been reported (Wilson et al, 1991; Pawlowski and Wilson, 1992). The phosphorescence lifetime for Pd-meso-tetra (4-carboxyphenyl) porphine in the absence of oxygen was almost temperature independent (less than 0.5 %/degree), but increased significantly with increasing pH. There was little effect on TO at pH values more alkaline than 7.2 (about 6.7% increase from 7.2 to 7.75) but was somewhat larger at pH values more acidic than 7.2 (about 17% decrease from 6.2 to 6.8). The quenching constant (~) was independent of pH between pH 7.2 and 7.75 but increased on the acidic side of pH 7.2 (about 8% increase from pH 6.8 to pH 6.2). For most experimental conditions the value of pH and temperature can be measured and the correct values of TO and ~ used. Where this is not possible, it is useful to note that the pH induced changes in TO have little effect on the calculated oxygen pressure when the latter is above approximately 5 Torr (see Wilson et al, 1991). At a measured phosphorescence lifetime of 50 usec, for example, and using the value of TO for pH 7.2 the calculated oxygen pressures at pH 6.4,6.8, 7.2 and 7.75 are 47 Torr, 50.9 Torr, 56.5 Torr and 56.5 Torr respectively. Thus, at these relatively high oxygen pressures the effect of an uncertainity in r due to pH uncertainty is minimal. At a lifetime of 200 usec, the respective oxygen pressures are 7 Torr, 8.5 Torr, 10.1 Torr and 10.4 Torr. The pH effect increases with decreasing oxygen pressure but from pH 7.2 to 7.75 is very small until values are attained which in vivo would indicate severe hypoxia. Other probes, such as Pd-coproporphyrin and Pd-mesoporphyrin, have very little dependence on pH compared to Pd-meso-tetra (4-carboxyphenyl) porphine (see Wilson et al, 1991; Pawlowski and Wilson, 1992) and their use can avoid any effect of changing pH in the physiological range of values. The Pd-coproporphyrin is, however, substantially more expensive than the Pd-meso-tetra (4-carboxyphenyl) porphine and this can be a consideration. Pd-mesoporphyrin, while more reasonably priced, has not yet been extensively tested in in vivo experiments. Use of phosphorescent oxygen probes for in vivo experiments For measurements in tissue in vivo the Pd complex of meso tetra-(4carboxyphenyl) porphine (Porphyrin Products, Logan, Utah; approx. 20 mg/kg) has been infused through an artery or vein as a complex with bovine serum albumin (Fraction V, ICN ImmunoBiologicals, Costa Mesa, CA; 60-70 mg/ml) dissolved in physiological saline at pH 7.4. When the Pd-porphyrins are injected into the blood, there is no measurable decrease in phosphorescence over the course of several hours. Thus the probe is not removed from the blood at a significant rate and a single bolus injection is sufficient for experiments extending over periods of several hours. There have been no adverse reactions to the probes in our experience and the properties of the Pd-porphyrin probes suggest they should not have significant toxicity for animals. Method for obtaining maps of phosphorescence lifetime and oxygen pressure In our laboratory, observations are generally made using a Wild Macrozoom microscope with an epifluorescence attachment. The phosphorescence is imaged using a
197
Xybion intensified CCD camera (Xybion Electronics Systems Corp., San Diego, CA). The camera is focused on the surface of the tissue prior to injection of the Pd-porphyrin. The phosphorescence can be measured using excitation at either near 400 nm (the Soret band) or in the visible region near 530 nm (alpha and beta bands). The depth of tissue sampled is dependent on absorbance of the excitation light, and the blue (about 400 nm) light penetrates only 0.1-0.2 mm whereas the green light penetrates about 1 mm. Phosphorescence emission is generally at wavelengths greater than 630 nm where there is little absorption of light by the tissue. For the tissues examined to date, there has been no detectable phosphorescence before the probe was injected. The illuminating light for the epifluorescence attachment was a EG&G 45 watt xenon flashlamp (EG&G, Salem, MA) with a flash duration of less than 5 usec mounted in a Leitz lamp housing. The flash lamp was controlled by a 16 MHz 80386 microcomputer (Spear Technology, Northbrook, IL) which determined the timing of the flashes and the gating of the video camera intensifier. A typical protocol was as follows: Number of flashes averaged for each delay time, 8; Delay times after the flash, 20 USec' 40 usec, 80 usec, 160 usec, 300 usec, 600 usec, and 2,500 usee; gate width in all cases, 2,500 usec. The image processor averaged the 8 frames for each delay time and this averaged image was displayed and recorded on the hard disk. Between 1 and 1.5 seconds was used to acquire the image for each delay time and each set of 7 images was acquired in total of about 40 seconds, including saving the images to the hard disk. It was assumed during analysis of the data that the brain did not move significantly during collection of a set of images. Direct comparison of the positions of prominent features of the images in the sequence confirmed that this was a reasonable assumption. Analysis of the phosphorescence data Two different methods have been used to analyze the data: 1. The regions of interest (small rectangular areas), are identified and the average intensity in that region determined for each image. These are plotted as the logarithm of phosphorescence intensity against delay time. The fit to a straight line is determined and where the decay is biphasic (usually indicating a fit to a straight line with a correlation coefficient of less than 0.97) fits to two or more straight lines can be used. In most cases the fit to a straight line is satisfactory (decay follows a single exponential). The measured decay constants (T) are substituted into the Stem-Volmer equation (equation 1), and the oxygen pressure in each region of interest is calculated from known values of ~ and TO. This approach is useful for following oxygen pressure in discrete anatomical features, particularly when there is reason to suspect the oxygen pressures may not be homogeneous (the phosphorescence decay will not follow a single exponential). 2. The images of phosphorescence are digitized as 512 x 480 pixel arrays of data. These data arrays are filtered and used to calculate a phosphorescence lifetime for each pixel of the image set. The computation involved passing a filter over each image and then subtracting the background (the image taken with a 2,500 usec delay) from each of the other images of the set. The decrease in intensity at each pixel of the array which occurs with increasing delay time is then fitted to a single exponential decay curve. The result is two dimensional maps of the distribution of phosphorescence lifetimes in the observed area of the tissue. Albumin as an aid in using Pd-pomhyrins as oxygen probes It is important to use the probes in a homogeneous environment, since the
198
phosphorescence lifetime at zero oxygen and access of the excited triplet state to oxygen can be influenced by the local environment. For most physiological conditions it is very convenient to have the probes in a medium with an excess of bovine serum albumin. The latter binds porphyrins and thus provides: 1. The same environment for all the molecules of probe, assuring that the population of probes have the same lifetimes in the absence of oxygen and the same quenching constant. 2. Self-quenching, which occurs when molecules of probe in the excited triplet state collide with and transfer energy to molecules of probe in the ground state, is suppressed. This is due to a combination of shielding of the probe by the surrounding albumin and thereby decreasing the efficiency of energy transfer and the fact that the diffusion constants of the probe-albumin complexes are smaller than those for the probe alone (this decreases collisional frequency). 3. The environment provided by the albumin binding site includes restricted access to oxygen, decreasing the quenching constant for oxygen by approximately a factor of 10 (see Vanderkooi et al, 1987; Wilson et al, 1991). The result is a quenching constant optimally suited for measurements of oxygen in vivo i.e. for Pd-meso-tetra (4carboxyphenyl) porphine the phosphorescence lifetime at air saturation is approximately 30 usec and that at zero oxygen is approximately 600 usec. 4. Last, but not least, the currently used Pd-porphyrins have very limited water solubility and albumin is required to keep them from precipitating in physiological media or at least binding to the cells in the blood and to vessel walls. When the Pd-porphyrins are used as solutions in the presence of albumin and the oxygen pressure is homogeneous, the phosphorescence decay curves have been found to follow a single exponential (correlation coefficient of greater than 0.99). Representative data obtained by ima&in& of phosphorescence of oxy&en probes in the blood in vivo The images of phosphorescence intensity of the brain cortex (Wilson et al, 1991; 1992A,B) show that under control conditions the veins as relatively bright, well defined vessels on a background of lower phosphorescence intensity. The oxygen probe is dissolved in the blood and phosphorescence intensity is dependent primarily on: 1. The amount of probe exposed to the excitation light (this is proportional to the concentration of blood in the approximately 1 mm thickness of surface tissue which is penetrated by the excitation light). 2. The oxygen pressure in the blood in the region of observation (phosphorescence intensity increases with decreasing oxygen pressure as described by the Stem-Volmer equation). 3. The intensity of the illuminating light (phosphorescence intensity increases in proportion to the illuminating light intensity). To the extent that the illuminating light flash provides uneven illumination, the phosphorescence intensity is uneven. These factors combine to provide an image in which the brightest regions are the veins because of their high blood concentration and low oxygen pressure. The capillary beds have only 10% or less of their volume as blood, the rest being cells and interstitial space. Therefore the capillary beds have a lower phosphorescence than veins although the oxygen pressure is similar to that in the veins. Arteriols are not seen without using experimental conditions emphasizing short phosphorescence. lifetimes. Although the concentration of blood, and therefore of probe, in the arteriols is high, the vessels are small in diameter and the oxygen pressure is high. The latter' quenches the
199
phosphorescence emission of the blood in the arteriols to below that of the capillary bed. The excitation light penetrates to the capillary bed below the arteriol, and the greater phosphorescence from the capillaries dominates the images and the calculated phosphorescence lifetimes. The phosphorescence lifetimes are independent of the concentration of phosphorescent probe (amount of blood in the observed tissue). The oxygen pressures in the veins and capillary beds are normally not very different, and therefore the phosphorescence lifetimes are not very different. The two dimensional maps of phosphorescence lifetime and oxygen pressure thus often do not show the pattern of vessels in the surface of the brain and these morphological features are lost. The phosphorescence characteristics associated with any observable structure, including veins, can, however, be determined by reading the values of the pixels at the corresponding positions of the initial intensity, phosphorescence lifetime, and oxygen pressure maps. Oxygen dependent quenching of phosphorescence is generally applicable to study of the distribution of oxygen in living tissue. The types of tissue that have been examined and the questions addressed is rapidly expanding. Initial measurements were made of isolated perfused rat liver (Rumsey et al, 1989; Wilson et al, 1989) but more recent studies have included the cortex of the brain of newborn piglets (Wilson et al, 1991; 1992B) and adult cats (Wilson et al, 1992A), the carotid body of the cat in vitro (Rumsey et al, 1991) and in vivo (Rumsey et al, 1992), the retina of the cat eye in vivo (Shonat et al, 1992), rat heart in vivo (Ince et al, 1992), and subcutaneous tumors in the rat in vivo (Wilson and Cerniglia, 1992). In each case the method has been able to provide new and valuable data on the delivery and utilization of oxygen in situ, suggesting it will make a major contribution to research in this important area of biochemistry and physiology. Although initial indications are that it will also become a powerful new tool for the diagnosis and treatment of medical problems such as tumors, wound healing, vascular disease and eye disease, there remain many barriers to overcome before this future can be realized. Acknowledgements: This work was supported in part by a grant NS-10939 from the National Institutes of Health and NOOOI4-89-J-1243 from the Office of Naval Research. BIBLIOGRAPHY Green, TJ., Wilson, D.F., Vanderkooi, J.M., and DeFeo, S.P. (1989) Phosphorimeters for analysis of decay profiles and real time monitoring of exponential decay and oxygen concentrations. Analy. Biochem. 174: 73-79. Ince, C., Ashruf, J., Sanderse, B.A. Pierik, E.GJ.M., Coremans, J.M.C.C., and Bruining, H.A. (1992) In vivo NADH and Pd-porphyrin video fluori/phosphorimetry. Adv. Expti. Med. BioI. these proceedings. Pawlowski, M. and Wilson, D.F. (1992) Monitoring of the oxygen pressure in the blood of live animals using the oxygen dependent quenching of phosphorescence. Adv. Exptl. Med. BioI. in press. Robiolio, M., Rumsey, W.L., and Wilson, D.F. (1989) Oxygen diffusion and mitochondrial respiration in neuroblastoma cells. Amer. J. Physioi. 256: C1207C1213.
200
Rumsey, W.L., Iturriaga, R., Spergel, D., Lahiri, S., and Wilson, D.F. (1991) Optical measurements of the dependence of chemoreception on oxygen pressure in the cat carotid body. Amer. J. Physiol. 261: in press. Rumsey, W.L., Lahiri, S., Iturriaga, R., Mokashi, A., Spergel, D., and Wilson, D.F. (1992) Optical measurements of oxygen and electrical measurements of oxygen chemoreception in the cat carotid body. Adv. Exptl. Med. Biol. These proceedings. Rumsey, W.L., Robiolio, M., and Wilson, D.F. (1989) Contribution of diffusion to the oxygen dependence of energy metabolism in human neuroblastoma cells. Adv. Exptl. Med. Biol. 248: 829-833. Rumsey, W.L., Schlosser, C., Nuutinen, E.M., Robiolio, M., and Wilson, D.F. (1990) Cellular energetics and the oxygen dependence of respiration in cardiac myocytes isolated from adult rats. J. Biol. Chem. 265: 15392-15399. Rumsey, W.L., Vanderkooi, J.M., and Wilson, D.F. (1989) Imaging of phosphorescence: a novel method for measuring oxygen distribution in perfused tissue. Science, Wash. DC 241: 16491-1651. Shonat, R.D., Wilson, D.F., Riva, C.E., and Pawlowski, M. (1992) Oxygen distribution in the retinal and choroidal vessels of the cat as measured by a new phosphorescence imaging method. Applied Optics in press. Vanderkooi, J.M., Maniara, G, Green, TJ., and Wilson, D.F. (1987) An optical method for measurement of dioxygen concentration based on quenching of phosphorescence. J. BioI. Chem. 262: 5476-5482. Vanderkooi, J.M., and Wilson, D.F. (1986) A new method for measuring oxygen in biological systems. Adv. Exptl. Med. Biol. 200: 189-193. Vanderkooi, J.M., Wright, W.W., and Erecinska, M. (1991) Oxygen gradients in mitochondria examined with delayed luminescence from excited-state triplet probes. Biochemistry 29: 5332-5338. Wilson, D.F. and Cerniglia, GJ. (1992) Localization of tumors and evaluation of their state of oxygenation by phosphorescence imaging. J. Cancer Res., in press. Wilson, D.F., Gomi, S., Pastuszko, A., and Greenberg, J.H. (1992A) Oxygenation of the cortex of the brain of cats during occlusion of the middle cerebral artery and reperfusion. Adv. Exptl. Med. BioI. These proceedings. Wilson, D.F., Pastuszko, A., DiGiacomo, J.E., Pawlowski, M., Schneiderman, R., Delivoria-Papadopoulos, M. (1991) Effect of hyperventilation on oxygenation of the brain cortex of newborn piglets. J. Appl. Physiol. 70(6): 2691-2696. Wilson, D.F., Pastuszko, A., Schneiderman, R., DiGiacomo, LE., Pawlowski, M. and Delivoria-Papadopoulos, M. (1992B) Effect of hyperventilation on the oxygenation of the brain cortex of neonates. Adv. Exptl. Med. BioI. In press. Wilson, D.F., Rumsey, W.L., Green, T.L, and Vanderkooi, LM. (1988) The oxygen dependence of mitochondrial oxidative phosphorylation measured by a new optical method for measuring oxygen concentration. J. BioI. Chem. 263: 27122718. Wilson, D.F., Rumsey, W.L., and Vanderkooi, J.M. (1989) Oxygen distribution in isolated perfused liver observed by phosphorescence imaging. Adv. Exptl. Med. BioI. 248: 109-115. Wilson, D.F., Vanderkooi, J.M., Green, TJ., Maniara, G., DeFeo, S.P., and Bloomgarden, D.C. (1987) A versatile and sensitive method for measuring oxygen. Adv. ExptI. Med. Biol. 215: 71-77.
201
INTRODUCTION The oxygen dependent quenching of phosphorescence has been shown to accurately measure oxygen consumption by suspensions of mitochondria and cells from air saturation to less than 10.8 torr (see for example Vanderkooi et al, 1987; Wilson et al, 1988; Robiolio et al, 1989, Rumsey et al, 1990; Wilson et al, 1991). The oxygen pressure dependence of the rate of oxygen consumption was determined for suspensions of isolated mitochondria (Wilson et al, 1988) and cells (Robiolio et al, 1989; Rumsey et al, 1990). Video imaging of phosphorescence has been shown to be a valid method for obtaining two dimensional maps of oxygen pressure in tissues (see Rumsey et al, 1988; 1992; Wilson et al, 1991; 1992A; 1992B; Shonat et al, 1992).
Oxygen Transport to Tissue XIV. Edited by W. Erdmann and D.E Bruley, Plenum Press, New York, 1992
195
METHODS AND MATERIALS Oxygen dependent quenching of phosphorescence provides a very sensItIve measure of oxygen pressure in the environment of phosphorescent molecules. The Pdporphyrins are particularly useful because intersystem crossing is rapid enough to quantitatively convert the exCited singlet state to the triplet state. As a result, these compounds show little or no fluorescence and the phosphorescence quantum efficiency is high. Thus the phosphorescence lifetime can be readily measured not only using conventional photomultipliers but also using intensified video cameras. Photomultipliers are single detectors and all the light collected from the phosphorescent sample is concentrated on one detector. Intensified video cameras function as arrays of detectors, in the case of the camera used in the present studies this was a 512 x 480 array, and each of these must receive sufficient light energy to make a reliable measurement of light intensity. If the detectors are of equal sensitivity, the video camera requires 245,000 times (512 x 480) as much light as the photomultiplier for measurements of equal accuracy. Phosphorimeters appropriate for measuring phosphorescence lifetimes, and thereby oxygen pressure, in aqueous media including suspensions of biological material, as well as in local regions of tissue in vivo have been described in some detail by Green et al (1988) and Pawlowski and Wilson (1992). The present paper will, therefore, focus on the technology used to obtain two dimensional images of oxygen distribution in tissue. Oxygen dependent quenching of phosphorescence of selected compounds can be quantitatively described by the Stem-Volmer relationship: (1)
For the Pd complex of tetra-(4-carboxyphenyl) porphine bound to bovine serum albumin and at physiological pH and 38°C, for example, kQ has a value of 325 Torr-! sec-! and TO is 600 usee (see Wilson et al, 1991; Pawlowski and Wilson, 1992). 1° and 1 are the phosphorescence intensities and TO and T the phosphorescence lifetimes at zero oxygen pressure and at an oxygen pressure P02 respectively. Phosphorescence lifetime is a more accurate measure of oxygen pressure because the lifetime measurements are independent of probe concentration and of illumination light intensity. Moreover, lifetime measurements are essentially unaffected by changes in the absorbance of other chromophores in tissues, such as hemoglobin, myoglobin and the cytochromes. The measurements are therefore independent of the degree of oxygenation of hemoglobin and myoglobin as well as the state of reduction of cytochromes. Calibration of the oxygen probes If the values of the phosphorescence lifetime are determined both at a, known oxygen pressure and at zero oxygen pressure, the quenching constant, kQ can be calculated. In order to obtain the necessary measurements, the phosphorescence lifetimes have been measured in the samples either equilibrated with air or treated with glucose oxidase and glucose to remove the oxygen and sealed in glass chambers. In each case, the measurements have been made at temperatures from 21°C to 38°C (see Wilson et al, 1991; Pawlowski and Wilson, 1992).
196
Effect of temperature and pH on phosphorescence lifetime and quenching constant of selected oxygen probes The effects of temperature and pH on different probes have been reported (Wilson et al, 1991; Pawlowski and Wilson, 1992). The phosphorescence lifetime for Pd-meso-tetra (4-carboxyphenyl) porphine in the absence of oxygen was almost temperature independent (less than 0.5 %/degree), but increased significantly with increasing pH. There was little effect on TO at pH values more alkaline than 7.2 (about 6.7% increase from 7.2 to 7.75) but was somewhat larger at pH values more acidic than 7.2 (about 17% decrease from 6.2 to 6.8). The quenching constant (~) was independent of pH between pH 7.2 and 7.75 but increased on the acidic side of pH 7.2 (about 8% increase from pH 6.8 to pH 6.2). For most experimental conditions the value of pH and temperature can be measured and the correct values of TO and ~ used. Where this is not possible, it is useful to note that the pH induced changes in TO have little effect on the calculated oxygen pressure when the latter is above approximately 5 Torr (see Wilson et al, 1991). At a measured phosphorescence lifetime of 50 usec, for example, and using the value of TO for pH 7.2 the calculated oxygen pressures at pH 6.4,6.8, 7.2 and 7.75 are 47 Torr, 50.9 Torr, 56.5 Torr and 56.5 Torr respectively. Thus, at these relatively high oxygen pressures the effect of an uncertainity in r due to pH uncertainty is minimal. At a lifetime of 200 usec, the respective oxygen pressures are 7 Torr, 8.5 Torr, 10.1 Torr and 10.4 Torr. The pH effect increases with decreasing oxygen pressure but from pH 7.2 to 7.75 is very small until values are attained which in vivo would indicate severe hypoxia. Other probes, such as Pd-coproporphyrin and Pd-mesoporphyrin, have very little dependence on pH compared to Pd-meso-tetra (4-carboxyphenyl) porphine (see Wilson et al, 1991; Pawlowski and Wilson, 1992) and their use can avoid any effect of changing pH in the physiological range of values. The Pd-coproporphyrin is, however, substantially more expensive than the Pd-meso-tetra (4-carboxyphenyl) porphine and this can be a consideration. Pd-mesoporphyrin, while more reasonably priced, has not yet been extensively tested in in vivo experiments. Use of phosphorescent oxygen probes for in vivo experiments For measurements in tissue in vivo the Pd complex of meso tetra-(4carboxyphenyl) porphine (Porphyrin Products, Logan, Utah; approx. 20 mg/kg) has been infused through an artery or vein as a complex with bovine serum albumin (Fraction V, ICN ImmunoBiologicals, Costa Mesa, CA; 60-70 mg/ml) dissolved in physiological saline at pH 7.4. When the Pd-porphyrins are injected into the blood, there is no measurable decrease in phosphorescence over the course of several hours. Thus the probe is not removed from the blood at a significant rate and a single bolus injection is sufficient for experiments extending over periods of several hours. There have been no adverse reactions to the probes in our experience and the properties of the Pd-porphyrin probes suggest they should not have significant toxicity for animals. Method for obtaining maps of phosphorescence lifetime and oxygen pressure In our laboratory, observations are generally made using a Wild Macrozoom microscope with an epifluorescence attachment. The phosphorescence is imaged using a
197
Xybion intensified CCD camera (Xybion Electronics Systems Corp., San Diego, CA). The camera is focused on the surface of the tissue prior to injection of the Pd-porphyrin. The phosphorescence can be measured using excitation at either near 400 nm (the Soret band) or in the visible region near 530 nm (alpha and beta bands). The depth of tissue sampled is dependent on absorbance of the excitation light, and the blue (about 400 nm) light penetrates only 0.1-0.2 mm whereas the green light penetrates about 1 mm. Phosphorescence emission is generally at wavelengths greater than 630 nm where there is little absorption of light by the tissue. For the tissues examined to date, there has been no detectable phosphorescence before the probe was injected. The illuminating light for the epifluorescence attachment was a EG&G 45 watt xenon flashlamp (EG&G, Salem, MA) with a flash duration of less than 5 usec mounted in a Leitz lamp housing. The flash lamp was controlled by a 16 MHz 80386 microcomputer (Spear Technology, Northbrook, IL) which determined the timing of the flashes and the gating of the video camera intensifier. A typical protocol was as follows: Number of flashes averaged for each delay time, 8; Delay times after the flash, 20 USec' 40 usec, 80 usec, 160 usec, 300 usec, 600 usec, and 2,500 usee; gate width in all cases, 2,500 usec. The image processor averaged the 8 frames for each delay time and this averaged image was displayed and recorded on the hard disk. Between 1 and 1.5 seconds was used to acquire the image for each delay time and each set of 7 images was acquired in total of about 40 seconds, including saving the images to the hard disk. It was assumed during analysis of the data that the brain did not move significantly during collection of a set of images. Direct comparison of the positions of prominent features of the images in the sequence confirmed that this was a reasonable assumption. Analysis of the phosphorescence data Two different methods have been used to analyze the data: 1. The regions of interest (small rectangular areas), are identified and the average intensity in that region determined for each image. These are plotted as the logarithm of phosphorescence intensity against delay time. The fit to a straight line is determined and where the decay is biphasic (usually indicating a fit to a straight line with a correlation coefficient of less than 0.97) fits to two or more straight lines can be used. In most cases the fit to a straight line is satisfactory (decay follows a single exponential). The measured decay constants (T) are substituted into the Stem-Volmer equation (equation 1), and the oxygen pressure in each region of interest is calculated from known values of ~ and TO. This approach is useful for following oxygen pressure in discrete anatomical features, particularly when there is reason to suspect the oxygen pressures may not be homogeneous (the phosphorescence decay will not follow a single exponential). 2. The images of phosphorescence are digitized as 512 x 480 pixel arrays of data. These data arrays are filtered and used to calculate a phosphorescence lifetime for each pixel of the image set. The computation involved passing a filter over each image and then subtracting the background (the image taken with a 2,500 usec delay) from each of the other images of the set. The decrease in intensity at each pixel of the array which occurs with increasing delay time is then fitted to a single exponential decay curve. The result is two dimensional maps of the distribution of phosphorescence lifetimes in the observed area of the tissue. Albumin as an aid in using Pd-pomhyrins as oxygen probes It is important to use the probes in a homogeneous environment, since the
198
phosphorescence lifetime at zero oxygen and access of the excited triplet state to oxygen can be influenced by the local environment. For most physiological conditions it is very convenient to have the probes in a medium with an excess of bovine serum albumin. The latter binds porphyrins and thus provides: 1. The same environment for all the molecules of probe, assuring that the population of probes have the same lifetimes in the absence of oxygen and the same quenching constant. 2. Self-quenching, which occurs when molecules of probe in the excited triplet state collide with and transfer energy to molecules of probe in the ground state, is suppressed. This is due to a combination of shielding of the probe by the surrounding albumin and thereby decreasing the efficiency of energy transfer and the fact that the diffusion constants of the probe-albumin complexes are smaller than those for the probe alone (this decreases collisional frequency). 3. The environment provided by the albumin binding site includes restricted access to oxygen, decreasing the quenching constant for oxygen by approximately a factor of 10 (see Vanderkooi et al, 1987; Wilson et al, 1991). The result is a quenching constant optimally suited for measurements of oxygen in vivo i.e. for Pd-meso-tetra (4carboxyphenyl) porphine the phosphorescence lifetime at air saturation is approximately 30 usec and that at zero oxygen is approximately 600 usec. 4. Last, but not least, the currently used Pd-porphyrins have very limited water solubility and albumin is required to keep them from precipitating in physiological media or at least binding to the cells in the blood and to vessel walls. When the Pd-porphyrins are used as solutions in the presence of albumin and the oxygen pressure is homogeneous, the phosphorescence decay curves have been found to follow a single exponential (correlation coefficient of greater than 0.99). Representative data obtained by ima&in& of phosphorescence of oxy&en probes in the blood in vivo The images of phosphorescence intensity of the brain cortex (Wilson et al, 1991; 1992A,B) show that under control conditions the veins as relatively bright, well defined vessels on a background of lower phosphorescence intensity. The oxygen probe is dissolved in the blood and phosphorescence intensity is dependent primarily on: 1. The amount of probe exposed to the excitation light (this is proportional to the concentration of blood in the approximately 1 mm thickness of surface tissue which is penetrated by the excitation light). 2. The oxygen pressure in the blood in the region of observation (phosphorescence intensity increases with decreasing oxygen pressure as described by the Stem-Volmer equation). 3. The intensity of the illuminating light (phosphorescence intensity increases in proportion to the illuminating light intensity). To the extent that the illuminating light flash provides uneven illumination, the phosphorescence intensity is uneven. These factors combine to provide an image in which the brightest regions are the veins because of their high blood concentration and low oxygen pressure. The capillary beds have only 10% or less of their volume as blood, the rest being cells and interstitial space. Therefore the capillary beds have a lower phosphorescence than veins although the oxygen pressure is similar to that in the veins. Arteriols are not seen without using experimental conditions emphasizing short phosphorescence. lifetimes. Although the concentration of blood, and therefore of probe, in the arteriols is high, the vessels are small in diameter and the oxygen pressure is high. The latter' quenches the
199
phosphorescence emission of the blood in the arteriols to below that of the capillary bed. The excitation light penetrates to the capillary bed below the arteriol, and the greater phosphorescence from the capillaries dominates the images and the calculated phosphorescence lifetimes. The phosphorescence lifetimes are independent of the concentration of phosphorescent probe (amount of blood in the observed tissue). The oxygen pressures in the veins and capillary beds are normally not very different, and therefore the phosphorescence lifetimes are not very different. The two dimensional maps of phosphorescence lifetime and oxygen pressure thus often do not show the pattern of vessels in the surface of the brain and these morphological features are lost. The phosphorescence characteristics associated with any observable structure, including veins, can, however, be determined by reading the values of the pixels at the corresponding positions of the initial intensity, phosphorescence lifetime, and oxygen pressure maps. Oxygen dependent quenching of phosphorescence is generally applicable to study of the distribution of oxygen in living tissue. The types of tissue that have been examined and the questions addressed is rapidly expanding. Initial measurements were made of isolated perfused rat liver (Rumsey et al, 1989; Wilson et al, 1989) but more recent studies have included the cortex of the brain of newborn piglets (Wilson et al, 1991; 1992B) and adult cats (Wilson et al, 1992A), the carotid body of the cat in vitro (Rumsey et al, 1991) and in vivo (Rumsey et al, 1992), the retina of the cat eye in vivo (Shonat et al, 1992), rat heart in vivo (Ince et al, 1992), and subcutaneous tumors in the rat in vivo (Wilson and Cerniglia, 1992). In each case the method has been able to provide new and valuable data on the delivery and utilization of oxygen in situ, suggesting it will make a major contribution to research in this important area of biochemistry and physiology. Although initial indications are that it will also become a powerful new tool for the diagnosis and treatment of medical problems such as tumors, wound healing, vascular disease and eye disease, there remain many barriers to overcome before this future can be realized. Acknowledgements: This work was supported in part by a grant NS-10939 from the National Institutes of Health and NOOOI4-89-J-1243 from the Office of Naval Research. BIBLIOGRAPHY Green, TJ., Wilson, D.F., Vanderkooi, J.M., and DeFeo, S.P. (1989) Phosphorimeters for analysis of decay profiles and real time monitoring of exponential decay and oxygen concentrations. Analy. Biochem. 174: 73-79. Ince, C., Ashruf, J., Sanderse, B.A. Pierik, E.GJ.M., Coremans, J.M.C.C., and Bruining, H.A. (1992) In vivo NADH and Pd-porphyrin video fluori/phosphorimetry. Adv. Expti. Med. BioI. these proceedings. Pawlowski, M. and Wilson, D.F. (1992) Monitoring of the oxygen pressure in the blood of live animals using the oxygen dependent quenching of phosphorescence. Adv. Exptl. Med. BioI. in press. Robiolio, M., Rumsey, W.L., and Wilson, D.F. (1989) Oxygen diffusion and mitochondrial respiration in neuroblastoma cells. Amer. J. Physioi. 256: C1207C1213.
200
Rumsey, W.L., Iturriaga, R., Spergel, D., Lahiri, S., and Wilson, D.F. (1991) Optical measurements of the dependence of chemoreception on oxygen pressure in the cat carotid body. Amer. J. Physiol. 261: in press. Rumsey, W.L., Lahiri, S., Iturriaga, R., Mokashi, A., Spergel, D., and Wilson, D.F. (1992) Optical measurements of oxygen and electrical measurements of oxygen chemoreception in the cat carotid body. Adv. Exptl. Med. Biol. These proceedings. Rumsey, W.L., Robiolio, M., and Wilson, D.F. (1989) Contribution of diffusion to the oxygen dependence of energy metabolism in human neuroblastoma cells. Adv. Exptl. Med. Biol. 248: 829-833. Rumsey, W.L., Schlosser, C., Nuutinen, E.M., Robiolio, M., and Wilson, D.F. (1990) Cellular energetics and the oxygen dependence of respiration in cardiac myocytes isolated from adult rats. J. Biol. Chem. 265: 15392-15399. Rumsey, W.L., Vanderkooi, J.M., and Wilson, D.F. (1989) Imaging of phosphorescence: a novel method for measuring oxygen distribution in perfused tissue. Science, Wash. DC 241: 16491-1651. Shonat, R.D., Wilson, D.F., Riva, C.E., and Pawlowski, M. (1992) Oxygen distribution in the retinal and choroidal vessels of the cat as measured by a new phosphorescence imaging method. Applied Optics in press. Vanderkooi, J.M., Maniara, G, Green, TJ., and Wilson, D.F. (1987) An optical method for measurement of dioxygen concentration based on quenching of phosphorescence. J. BioI. Chem. 262: 5476-5482. Vanderkooi, J.M., and Wilson, D.F. (1986) A new method for measuring oxygen in biological systems. Adv. Exptl. Med. Biol. 200: 189-193. Vanderkooi, J.M., Wright, W.W., and Erecinska, M. (1991) Oxygen gradients in mitochondria examined with delayed luminescence from excited-state triplet probes. Biochemistry 29: 5332-5338. Wilson, D.F. and Cerniglia, GJ. (1992) Localization of tumors and evaluation of their state of oxygenation by phosphorescence imaging. J. Cancer Res., in press. Wilson, D.F., Gomi, S., Pastuszko, A., and Greenberg, J.H. (1992A) Oxygenation of the cortex of the brain of cats during occlusion of the middle cerebral artery and reperfusion. Adv. Exptl. Med. BioI. These proceedings. Wilson, D.F., Pastuszko, A., DiGiacomo, J.E., Pawlowski, M., Schneiderman, R., Delivoria-Papadopoulos, M. (1991) Effect of hyperventilation on oxygenation of the brain cortex of newborn piglets. J. Appl. Physiol. 70(6): 2691-2696. Wilson, D.F., Pastuszko, A., Schneiderman, R., DiGiacomo, LE., Pawlowski, M. and Delivoria-Papadopoulos, M. (1992B) Effect of hyperventilation on the oxygenation of the brain cortex of neonates. Adv. Exptl. Med. BioI. In press. Wilson, D.F., Rumsey, W.L., Green, T.L, and Vanderkooi, LM. (1988) The oxygen dependence of mitochondrial oxidative phosphorylation measured by a new optical method for measuring oxygen concentration. J. BioI. Chem. 263: 27122718. Wilson, D.F., Rumsey, W.L., and Vanderkooi, J.M. (1989) Oxygen distribution in isolated perfused liver observed by phosphorescence imaging. Adv. Exptl. Med. BioI. 248: 109-115. Wilson, D.F., Vanderkooi, J.M., Green, TJ., Maniara, G., DeFeo, S.P., and Bloomgarden, D.C. (1987) A versatile and sensitive method for measuring oxygen. Adv. ExptI. Med. Biol. 215: 71-77.
201
