Search

Living cells may be modified in diverse ways by the combined action of visible light and photosensitizing molecules. The effects appear most frequently as disruptions of subcellular structure, changes in surface membrane function or inhibition of mitotic ability. This review concentrates on the four most thoroughly studied cell types--yeast cells, nerve cells, erythrocytes, and cultured tumor cells. Research on these cells indicates that potency of sensitization depends at least as much on the factors affecting an association between sensitizer and cell prior to illumination as on photochemical properties. While sensitizers which permeate may lead to altered DNA, it appears that surface membrane modification occurs simultaneously and may be critical in the inactivation mechanism. There is much circumstantial evidence suggesting that excited singlet molecular oxygen acts as an intermediate between photoexcited sensitizer and target alteration. Proteins, lipids, and nucleic acids are all susceptible to photosensitized attack, but the correlation between cellular and molecular modification remains ill-defined. The use of the photodynamic process as a therapeutic technique, particularly in the treatment of malignant tumors, holds great promise, but awaits further research to develop greater selectivity of action.

The numerical assessment of sensitizing potency of different species of photosensitizer molecules has historically been impracticable because of the difficulty of measuring all of the factors that must be taken into account. This paper describes a new method for quantifying relative sensitizer potency that obviates most of the past difficulties. The physical data needed to assess photosensitizing potency are the absorption spectrum of a sample of medium containing sensitizer, the illumination spectrum impinging on the preparation, the duration of illumination, and the amount of light-induced modification. The measurement of these quantities is relatively straightforward, but it is easy to overlook a number of artifacts that can distort the values for potency. Avoiding error requires careful attention to the optical properties of the system, particularly in the light path through the medium containing sensitizer.

Ignorance of phototoxicity by the general biomedical community constitutes a health hazard. Certain food colorants and dyes proposed for use in angiography produce lethal effects in cells and whole organisms on light exposure but not in the dark. Yet-light-relatedness is usually neglected in toxicity studies. We urge that the specific involvement of light be considered whenever the toxicity of a light-absorbing substance is assessed.

Eosin derivatives that bind primarily to lipid or protein sites in erythrocyte membranes were studied in solution and as sensitizers of erythrocyte membranes. In 50% ethanol-water mixtures eosin maleimide (EYMA) and 5-N-hexadecanoyl amino eosin (E16) had nearly identical absorption spectra. Higher ethanol concentrations did not change peak absorbances. In the presence of neutral detergent both sensitizers had equivalent absorbance at all ethanol concentrations. In water, EYMA was more effective than E16 at bleaching RNO, probably because of E16 aggregation into micelles, while in ethanol-water mixtures E16 was slightly more effective at bleaching DPBF, indicating equivalent singlet oxygen generation when the sensitizers are in monomeric form. In water with neutral detergent, azide in the 20 microM range inhibited the majority of RNO bleaching with both sensitizers; in 50% ethanol-water mixtures azide at 1 mM showed a 50% inhibition of DPBF bleaching with both sensitizers. Iodide in the 30 mM range reduced DPBF bleaching by 50% in 50% ethanol-water mixtures. When matched for amount loaded in erythrocyte membranes these sensitizers were about equally effective at sensitizing induction of cation permeability, assayed as rate of delayed photohemolysis, while E16 was slightly more effective at sensitizing loss of cholinesterase (AchE) activity. The relation of lysis rate to load was somewhat steeper for E16 than EYMA. For both sensitizers lysis rate increased at about the 1.5 power of light dose. Deoxygenation of the reaction media with argon totally blocked detectable photomodification. Ghost membranes made from sensitizer-treated cells were effective generators of singlet oxygen, assayed by RNO bleaching. However, when mixtures of EYMA-treated and untreated cells were illuminated together, only the EYMA-treated cells showed evidence of photomodification. Azide at 5 mM slowed the initial rate of AchE loss by about 75% with E16 and EYMA. Azide partially slowed photohemolysis. Azide decreased RNO bleaching by sensitizer-treated ghosts as it did in water with detergent micelles. A deuterium oxide solvent increased photohemolysis rate with E16 by 41%, but did not increase photohemolysis rate with EYMA. Deuterium oxide had a positive, but statistically insignificant effect on loss of AchE with both sensitizers. Deuterium oxide following illumination slowed lysis sensitized by both sensitizers more than 50%. Iodide exerted a modest inhibition of photohemolysis and loss of AchE sensitized by E16, but had virtually no influence on sensitization by EYMA. The results in solution indicate that EYMA and E16 have nearly identical photochemical properties when in monome

A kinetic model of colloid osmotic hemolysis for cation-permeable cells has been developed. The model consists of three essential components. The first is a set of flux equations, under the assumption that the membrane potential is equal to the chloride equilibrium potential and that cation fluxes are described by the Goldman flux equation. The second is the osmotic equilibrium model of Freedman and Hoffman that takes into account the non-ideal osmotic behavior of erythrocytes. The third is an empirical relation between hemolysis and cell volume, developed from the lysis behavior in hypoosmotic media. Model simulations are compared with lysis experiments using the antibiotic nystatin to raise cation permeability. The form of the kinetics and inhibition of lysis by sucrose are described well by the model. In additional lysis experiments at different external pH the small pH dependence is accounted for by the model.

The double sucrose gap technique for the study of lobster giant axons has been reexamined. The leakage behavior of the system cannot be successfully modeled by conventional sucrose gap theory, but is accounted for by the McGuigan-Tsien model that takes into account the cable properties of membrane under sucrose. The facts of high-leakage conductance and the ability to maintain large resting potentials in the face of low sucrose gap resistance lead to a hypothesis that membrane resistance under sucrose is very low because of a large negative surface potential. Computer simulations of the leakage behavior of the conventional gap model and the McGuigan-Tsien model were compared with experimental measurements on lobster axons using normal sucrose or sucrose doped with Na+, Ca2+ or La3+ ions. As the concentration of doping ion increased, the leakage rose, but the species of doping ion had more influence on leakage than gap resistance. At equal gap resistance, leakage decreased with an increase in valence of the doping species. Leakage was even lower in La-doped sucrose at 20 M omega gap resistance than in normal sucrose at 200 M omega gap resistance. Resting potentials decreased with decreasing gap resistance and increasing valence of the doping species. Resting potential behavior was successfully simulated with a hybrid model consisting of a point node flanked by infinite cables and a shunt between ground and the voltage-measuring pool. The data support the hypothesis that the membrane resistance under sucrose is low and that it can be raised by doping the sucrose with multivalent cations, with La3+ being particularly effective. Both the leak conductance and resting potential are influenced more by membrane under sucrose than membrane in the node. The experiments also demonstrate that doping with La3+ vastly improves the stability and longevity properties of the lobster axon preparation.

The pH dependence for sensitized photochemical block of sodium channels in lobster giant axons was determined and compared with direct channel block by protons. Isolated axons were studied in a double sucrose gap voltage clamp arrangement and the pH of the external bath was varied over the range 4.1--11.0. Irreversible photochemical block was achieved by illumination with visible light in the presence of eosin Y or acriding orange. The rate constant for photochemical block of sodium channels was depressed at both high and low pH relative to that at neutral pH, revealing the existence of two receptors involved in the process with pK values of 4.8 and 10.4. A direct reversible channel-blocking receptor titrates with a pK of 4.8, the same as one of the receptors involved in the photochemical block, and senses about 9% of the electric field as determined by a Woodhull analysis. Lowering the pH from 8.2 to 4.6 shifted the sodium conductance versus voltage relation in the depolarizing direction. It is proposed as a hypothesis that the low and high pK receptors are histidine imidazole and primary amino groups, photooxidation of which leads to channel block via cross-linking of channel proteins.

Many of the known features of photohemolysis have been organized in a kinetic model that simulates the lytic time-course in a variety of conditions. The model combines Nernst-Planck flux principles, the osmotic equilibrium model of Freedman and Hoffman, equations relating illumination parameters to ion permeability, and an empirical relation between cell volume and lysis. Model simulations are compared with experiments showing the dependence of lysis kinetics on sensitizer concentration and on the osmotic content of the reaction medium. Additional experiments demonstrate that the inherent osmotic fragility of erythrocytes is not altered by illumination conditions that cause major delayed lysis 23 h later. The successful simulations support the hypothesis that photohemolysis is a colloid osmotic lysis occurring in cells behaving as imperfect osmometers.