Photodynamic therapy (PDT) involves the use of photosensitizing agents for the treatment of malignant disease. These photosensitizing agents are infused intravenously and are selectively retained within tumor cells. When exposed to light of the proper wavelength, the photosensitizing agent is activated. This leads to the formation of toxic oxygen radicals, which result in cell death.
Until recently, PDT for lung cancer had primarily been a research tool. Several institutions worked with PDT in the 1980s, but its use in the United States remained limited to the research setting. Photodynamic therapy using the first FDA-approved photosensitizing agent, porfimer sodium (Photofrin), is now available to US clinicians.
Photodynamic therapy has the potential to complement and improve the approach to a variety of clinical problems. Potential applications include carcinoma in situ, early-stage lung cancer in nonsurgical candidates, advanced lung cancer, and tumors metastatic to the lung. To properly assess the role of PDT in these settings requires careful analysis of the mechanism of action and technique of PDT, a review of clinical studies using PDT, and an exploration of how PDT may be integrated into a multimodality approach.
The concept of photochemical sensitization and subsequent cell death is not a new one. Light was used for healing by the ancient Greeks, and by 1900 photochemical reactions were used to kill paramecia.
Photodynamic therapy was initially used to treat skin cancers in the early 1900s, leading to a variety of chemicals being developed to promote photochemical cytotoxicity. Among the agents used, hematoporphyrin subsequently demonstrated the ability to be selectively concentrated or preferentially
retained within malignant cells.[2,3] Hematoporphyrin derivatives were later shown, first by Lipson et al and then by Gregorie et al, to be retained in a large percentage of squamous cell carcinomas and adenocarcinomas.[4-8] This led to the development of newer hematoporphyrin derivatives with experimental applications, including breast cancer and bladder cancer, and the use of PDT in an even wider spectrum of malignancies.[9-13]
These pioneering studies highlighted the importance of several factors related to the mechanism of PDT: membrane injury, delivery of oxygen, the role of the immune and vascular systems, the importance of the photosensitizer, and light dosimetry. Each of these factors is critical to the mechanism of action of PDT and has potential clinical implications.
The basic mechanism of the cellular cytotoxicity of PDT seems to be membrane damage. The plasma and mitochondrial membranes, in particular, are targets of PDT because of the water-lipid partition coefficient of the photosensitizing agents.[14-19] Porphyrin uptake studies demonstrate initial binding of the photosensitizing agent with the plasma membrane, with subsequent extension to the internal cellular membranes. Damage occurs after light activation and is visible immediately. Initially, the damage is characterized by the formation of multiple areas of membrane injury, or blebs. These progress to form larger balloon-like areas.[16,21] Cellular division and normal functions cease; this is followed by cell lysis.
Concurrent with damage to the plasma membrane is injury to other internal cellular membranes, including the mitochondrial membrane, nuclear membrane, Golgi apparatus, and endoplasmic reticulum.[18,22] With mitochondrial injury comes inhibition of oxidative phosphorylation and generation of adenosine(Drug information on adenosine) triphosphate (ATP), followed by a decrease in cellular ATP.[22,23] Using phosphate-31 spectroscopy to assess metabolic response to PDT, it can be shown that the fall in ATP is dramatic, becoming virtually undetectable 2 to 4 hours after treatment.[24,25]
Although PDT results in nuclear membrane injury and DNA strand breaks, it does not appear that DNA injury plays an important role in cell death. Importantly, PDT has not been shown to be mutagenic in vitro.[16,19,26]
Delivery of Oxygen
In vitro, when oxygen is not present or is present in a concentration of less than 2%, cells become resistant to PDT.[27,28] The cytotoxicity of PDT is free radicalmediated via a type II photooxidation reaction.[16,19] In type II reactions, light energy excites and activates the photosensitizer, energy transfer occurs from the sensitizer in its excited state to molecular oxygen, and singlet oxygen species are generated. The resulting free radical generation leads to the membrane injury described above.
The importance of free radicalmediated injury is underscored by the observation that scavengers of singlet oxygen, such as 1,3-diphenylisobenzofuran, reduce PDT-mediated cytotoxicity. Low oxygen concentrations have also been shown to lead to decreased sensitivity to PDT in vitro, and tissue hypoxia models in animals support this observation. Other investigators have suggested that local tumor hypoxia may account for some cases of nonresponsiveness to PDT.
Thus, all of the membrane injury described above is predicated on the availability of an adequate concentration of oxygen to generate free radicals.
Role of the Immune and Vascular Systems
Part of the in vivo tumor destruction results from the effect of PDT on the vasculature. The neovasculature of tumors may be a target for PDT since these venous-derived vessels may not have sufficient strength to remain patent in the face of high extravascular pressures. Decreased flow occurs, leading to arteriolar and venular stasis, arteriolar vasoconstriction, thrombosis of venules, and increasing interstitial edema.[31,32]
In addition, other investigators have hypothesized and demonstrated varying degrees of increased coagulation in the vascular bed associated with PDT. Studies with nuclear magnetic resonance imaging using in situ fluorine have shown that damage to the tumor vasculature occurs prior to actual tumor necrosis. Ben-Hur and Orenstein demonstrated increased coagulation associated with injury to the endothelium from PDT, with resultant red blood cell agglutination and thrombus formation.
Associated with this complex picture of free radical injury, vascular injury, and coagulation is a concurrent immune response, characterized by platelet and neutrophil activation.[16,35] This immune activation results in the release of vasoactive compounds, including arachidonic acid derivatives, such as prosta-glandins E2 and I2 and thromboxane. The potential contribution of these mediators to cell death and tumor injury has been demonstrated by the observation that cyclooxygenase inhibitors reduce the effect of PDT on arterioles.
Thus, in addition to oxygen-mediated free radical membrane injury and ATP depletion, both direct vascular damage and immune-mediated injury may contribute to cell death from PDT.
A wide variety of photosensitizing agents have been studied and developed, including the chlorins, phthalocyanines, tetraphenylporphine sulfate, porphines, rhodamine-123, and porphyrin-based agents. The clinically significant aspects of photosensitizers include their relative concentration in tumors, their yield of singlet oxygen, the amount of tissue penetration allowed, and their photostability/lability.
The only photosensitizing agent that has been approved by the FDAporfimer sodium (Photofrin)belongs to the porphyrin-based family of agents. Consequently, this discussion will focus on porfimer sodium, although many of the comments apply to other agents currently under investigation.
Selective retention or uptake of the photosensitizer by tumors allows for a relatively high tumor-tissue concentration ratio. Most sensitizers have a concentration ratio ranging from 2:1 to 5:1. The selective tumor retention associated with the porphyrin-based agents was initially reported by Figge in 1948. Lipson and colleagues subsequently demonstrated that derivatives of hematoporphyrin were associated with an even higher tumor-tissue concentration ratio.[4-8] This tumor-tissue ratio is highest at 24 to 48 hours after intravenous injection.
However, Lipson and colleagues also demonstrated that fluorescence could be detected within 3 hours of intravenous injection. This has clinical importance in terms of how photoactive agents are used. When employed for lung cancer treatment (ie, PDT), the time interval between injection and light application is typically 48 to 72 hours. When used in investigational studies for tumor detection and localization, the 3-hour time interval is more useful.
The mechanism of selective retention and uptake of the porphyrins is based on studies using murine models, which demonstrated that these photosensitizers are accumulated and retained by endocytosis via the vascular endothelium.[16,35,37] This is due, in part, to the lipophilic nature of the compounds, previously described in relation to their propensity to bind to cell membranes based on their partition coefficient.
The distribution pattern of porphyrins after intravenous infusion mirrors that of low-density lipoprotein receptors in the various organs, with the greatest amount found in the liver, followed, in descending order, by the adrenal glands, urinary bladder, pancreas, kidney, spleen, stomach, bone, lung, heart, muscle, and brain. The serum half-life of photosensitizing agents in humans is 20 to 30 hours, but these compounds may persist in the skin at low levels of 2% to 5% for up to 4 to 6 weeks.
The exact mechanism of selective retention of photosensitizing agents remains an area of investigation. Possible theories include: increased tumor uptake of low-density lipoproteinassociated sensitizers; increased uptake by tumors due to their lower pH, and the associated increase in water solubility of the sensitizer at low pH; poor lymphatic drainage of tumors; tumor angiogenesis factors: and changes within the stromal cells of tumors that increase uptake.[16,35]
As described above, the cytotoxicity of PDT is mediated via a complex set of interactions, including photooxidative reactions. The photosensitizer must be able to absorb photons of appropriate wavelengths so as to become a triplet species. The light-excited triplet state of the sensitizer transfers its energy to endogenous state triplet oxygen to produce an excited state of molecular oxygen, singlet oxygen. This is the type II photooxidation reaction described above.
Importantly, singlet oxygen can be generated with quantum energies as low as 0.98 MeV, corresponding to a wavelength of 1,220 nm. However, most available sensitizers work efficiently only at wavelengths up to 850 nm, with a quantum yield of 0.2 to 0.6 MeV.
The differences in quantum yield appear to be related, or at least affected by, the location of the sensitizer in the cells. In addition, for any given quantum yield, lipophilic sensitizers, such as porfimer sodium, seem to be more efficient than hydrophilic sensitizers at similar quantum yields. Thus, while these considerations may seem esoteric, they have direct clinical impact in terms of efficacy.
Furthermore, the quantum energy used relates to the important clinical issue of tissue penetration. The wavelength, as determined by the quantum energy used, directly affects the maximum absorbance capacity and depth of tissue penetration of the photosensitizing agent.
With respect to the porphyrin family of photosensitizers, the absorption spectra demonstrate a peak at 405 nm. Use of this peak allows for fluorescence of tumors to be used as a tumor marker. However, for treatment (PDT), this wavelength of light is suboptimal, since it is nearly completely absorbed within 1 mm of the surface. Thus, a wavelength of 630 nm is used for PDT. This allows for penetration to a depth of approximately 5 mm.
Obviously, the development of future photosensitizers will be greatly affected by consideration of the quantum energies used, their absorption spectra, and the resultant clinical consequences, in terms of depth of penetration.
Finally, all photosensitizers can be destroyed by light, which may have an impact on cytotoxicity. When exposed to light, a sensitizer may generate enough energy to destroy tumor tissue but be destroyed and lose its cytotoxic potential within normal cells, a process described as photobleaching. This effect is an important consideration, since it allows minimal damage to adjacent normal tissue while selectively destroying tumor cells.
The efficacy of PDT depends on accurate delivery of light to the area to be treated. It is useful to think of PDT in terms of three components: (1) the photosensitizers characteristics and concentration at the tumor site (described above), (2) the rate of energy delivery (power), and (3) the total energy delivered
With respect to the delivery of light, any source of light with the appropriate characteristics could be used. In practice, laser light is typically used because it offers the advantage of a uniform spectrum and coherence. For lung cancer, the argon dye pump or excimer laser is often used, although, in theory, any laser with the proper wavelength and power could be employed. Important in vivo considerations include the effect of dose rate delivery and total energy delivery.
Although in vitro evidence has suggested that a high dose rate (power) is associated with improved cytotoxicity, in vivo data have demonstrated that lower dose rates may be more effective.[41,42] In a study involving the treatment of human mesothelioma allografts in nude mice, decreasing the light intensity from 200 to 50 mW/cm² actually improved response. The investigators hypothesized that a reduction in the fluence rate or fractionation
may paradoxically enhance the effect of treatment because of an increase in singlet oxygen in regions of poor capillary flow.
Regardless of whether this is the only mechanism involved, what is clearly important is the empiric observation that light dosimetry has a clinical impact on response and side effects. Thus, controlled, reproducible light dosimetry and the technology used to deliver it are important considerations for PDT.