PHOTODYNAMIC THERAPY OF A SOLID TUMOR

Photodynamic therapy (PDT) is a technique that combines a light-activated drug with nonthermal light to achieve the selective, singlet-oxygen (¹O₂) photochemical destruction of a tumor with minimal effect on surrounding normal tissue. The basis of PDT depends thus on the distribution of the photosensitizer, the propagation of the photoactivating light, and the amount of oxygen in the tumor [[1]]. PDT has recently become an established modality of cancer treatment, but it can be further improved through a better understanding of the determinants affecting its therapeutic efficiency. Among the determinants are (1) photochemical and photophysical properties of a given photosensitizer; (2) the uptake of the photosensitizer by a tumor; (3) the localization pattern of the photosensitizer in the tumor; (4) the relationships between the chemical properties of the photosensitizer and the characteristics of its intratumoral distribution (uptake and localization); and (5) the relationships between the intratumoral localization patterns of the photosensitizer and the actual target sites of its PDT effect. Since ¹O₂, the main cytotoxic product, formed through reaction between excited sensitizer molecules and oxygen has lifetime less than 0.05 µs in cells, it can diffuse less than 0.02 µm from the site of production [[2]], and the targets of PDT are thus the sites where the photosensitizer is localized.

An enormous amount of information is available on the effects of PDT on neoplastic cells of tumor [[1], [3]] and thus is not reviewed here in detail. In brief, ¹O₂ produced by dyes used for PDT can mediate oxidation of lipids, amino acid, and cross-linkage of protein components of cell/organellae membranes. The modifications of these biomolecules in cellular membranes can cause alterations in membrane permeability, loss of fluidity, and inhibition/inactivation of membrane-associated enzyme systems and receptors, and finally result in cellular necrosis and apoptosis [[4–6]]. Such direct PDT effects are induced by dyes that localize in neoplastic cells of tumor [[7–10]]. However, the direct effects are very limited, with less than 2 logs of neoplastic cells being killed, far short of the 6–8 logs needed for tumor cure [[1]]. This indicates that tumor stroma plays a crucial role in the destruction of tumor.

STROMA OF A SOLID TUMOR

A solid tumor consists of 2 distinct compartments: parenchyma and stroma. Tumor stroma, partially induced by tumor cells themselves, is essential for tumor growth. It interposes connective tissue between tumor cells and normal host tissues and provides vascular supply that carries nutrients, gas exchanges, and waste disposal. All solid tumors, no matter what type or cellular origin, require stroma if they are to grow.

Carcinoma, that originates from epithelial cells, often contains a basal lamina that separates clumps of tumor cells from stroma. However, in other types of tumors, such as sarcoma, tumor cells may intermingle with or abut on stromal components.

Tumor stroma is composed of interstitial connective tissue that is formed from elements derived from both circulating blood and adjacent connective tissues. The main components of tumor stroma include (1) plasma protein-rich interstitial fluid; (2) structural proteins, such as collagens (types I, III, and IV), elastin, fibronectin, and laminin; (3) ground substances, such as proteoglycans and hyaluronic acid; (4) blood vessels (endothelial cells and pericytes); and (5) fixed connective tissue cells (fibroblasts, histiocytes/macrophages, mast cells) and inflammatory cells (lymphocytes, macrophages and granulocytes) from the blood.

Tumors vary greatly in quantity and quality of stroma. In some carcinomas stroma contributes up to 90% of the total tumor mass, whereas other tumors have only minimal amounts of stroma. Even within a single tumor significant variations can be seen in stromal compositions from one area to another.

DISTRIBUTION PATTERN OF A PHOTOSENSITIZER IN TUMOR

The fine structure of biological samples has been probed by fluorescence of dyes, but the quality of the images had been severely limited in conventional fluorescence microscopy until about 15 years ago when highly light-sensitive, charge-coupled device (CCD) cameras were introduced. Such CCD cameras can be employed to produce bright images from very weakly illuminated microscopic specimens, thus permitting observation of cell and even tissue materials that otherwise are not seen or rapidly photobleached by microscopic light exposure. These advanced systems make it possible to study intratumoral localization patterns of photosensitizers used for PDT.

Although Policard and Leulier found preferential accumulation of porphyrins in tumors 80 years ago [[11]], the mechanism of such selective localization is still not fully understood. This is at least partly because hematoporphyrin derivative (HpD) and its purified version, Photofrin, contain several porphyrin components with various hydrophobicity and probably with different intratumoral localization patterns as well. Chemical properties of a dye affect its intratumoral localization. Pure dyes with a well-defined lipophilicity are thus better suited for such investigation. For example, sulfonated aluminium phthalocyanines (AlPcS_ns) are second-generation photosensitizers with several advantages over HpD/Photofrin. The number of sulfonate groups determines the hydrophobicity of AlPcS_ns. The intratumoral localization patterns of AlPcS_ns have been studied by use of confocal laser scanning microscopy in human LOX melanoma xenografts in nude mice [[7], [12]]. Hydrophilic AlPcS₄ and AlPcS₃ are mainly localized in the stromal tissue of the LOX tumor, while more hydrophobic AlPcS₂ and AlPcS₁ tend to localize intracellularly in the tumor. The intracellular localization of the dyes is particularly pronounced 24 h or later after the administration of the dyes [[13]]. Similar results have also been found in the family of sulfonated meso-tetraphenyl porphines [[13]].

In general, when a photosensitizer is administered into the blood the photosensitizer binds different serum proteins depending on its chemical properties. This binding affects further its transport in the blood circulation [[14], [15]]. The localization pattern of a dye in tumor tissue is, to a great extent, determined by its hydrophobicity. Relative hydrophilic dyes are largely transported by albumin and globulins and mainly localized in the vascular stroma of tumor tissue; while more hydrophobic drugs are preferentially incorporated into lipoproteins, particularly low-density lipoproteins (LDL), and localized in the parenchyma cells of tumor tissue [[16]]. The amphiphilic dyes may differ in their hydrophobicity as well as pharmacokinetic patterns. It should be pointed out that, a drug delivery system, such as LDL, liposomes, oil emulsions, and inclusion complexes, can significantly change the localization pattern of a drug in the components of tumor [[14]].

STRUCTURAL PROPERTIES OF TUMOR STROMA AFFECTING TUMOR DISTRIBUTION OF A DYE

Hematoporphyrin derivative is taken up in vitro by normal and cancer cells with varying oncogenic potential to similar extents [[17], [18]]. This indicates that structural properties of tumor stroma rather than tumor cells may account for selective distribution of a photosensitizer in tumor tissue. The abnormal pathohistological structure of tumor stroma includes hyperpermeable microvasculature with a poor lymphatic network, resulting in a high interstitial fluid pressure [[19], [20]]. A large interstitial space [[21]] and a high amount of tumor-associated macrophages present in some cases [[22]] also favor accumulation of a photosensitizer in tumor stroma. In addition, low pH values [[23–26]] and a large number of LDL receptors, as well as secretion of vascular endothelial growth factor in some types of tumors [[13], [27–29]], may contribute to such a selective localization.

TUMOR STROMAL PROTEINS INVOLVED IN TUMOR LOCALIZATION OF A DYE

Most types of tumors have a high amount of collagenous stroma. This is generally believed to be a defensive response mechanism in which host fibroblasts are provoked to synthesize collagen to prevent tumor invasion, although some studies have also shown that tumor cells are able to produce collagen [[30], [31]]. It has been demonstrated that porphyrins have a high affinity for fibrin, collagen, and elastin [[32–34]]. Fibrin and collagen are essential elements of tumor stroma, while collagen and elastin are components of vascular wall. We have noted that Photofrin and AlPcS₄ localize primarily where collagenous proteins are found, i.e., basal lamina, vascular wall, and collagenous connective tissue [[12]]. These observations indicate that tumor retention of photosensitizers may be related to a strong affinity of the dyes for certain types of collagenous proteins, in particular newly synthesized proteins usually seen in neoplastic tissue [[34], [35]].

Certain enzymes in tumor stroma may be involved in tumor localization of a photosensitizer. For example, 5-aminolevulinic acid (ALA) is a precursor to heme biosynthetic pathway. Addition of exogenous ALA can lead to a selective accumulation of porphyrins in tumor tissue that, when activated by light, causes the photosensitizing effect [[36–38]]. Significant differences in the activities of key enzyme profiles in the heme pathway between normal and tumor tissues contribute this selective property [[36–38]]. ALA-mediated PDT has successfully been applied in several clinical indications [[38]]. Due to its low lipid solubility, however, ALA has a limited ability to cross certain biological barriers, such as cellular membranes. A number of lipophilic ALA esterified derivatives have thus been synthesized and improved production and selectivity of ALA ester-induced porphyrins observed in several in vitro and in vivo biological systems [[39–42]]. This may be related to the fact that the activity of esterase in tumors is greater than that in surrounding normal tissues [[43]].

Another approach studied by Kennedy et al. is to design a prodrug molecule that consists of a photosensitizer, a solubility modifier, and a linker (polypeptide). Such a molecule can be specifically cleaved by the proteolytic enzymes present in tumor stroma, so that the sensitizer is accumulated in the tumor [[44]].

EFFECTS OF PDT ON TUMOR MICROVASCULATURE

Vulnerable and superpermeable vasculature of tumor stroma is sensitive to PDT. This has been shown as initial functional disturbance (such as vasoconstriction, vasodilatation and aggregation of blood cells) of arterioles, capillaries, and postcapillary venules followed by slowing down the blood flow and final complete stasis (Figure 1, part a) [[45–48]]. As a result, hypoxia/anoxia and subsequent death of neoplastic cells occur [[1]]. It is generally accepted that vascular damage is the major pathway of the destruction of tumor by PDT, although the exact mechanism is still not fully understood.

Effects of Photodynamic Therapy on Tumor Stroma

All authors

Qian Peng & Jahn M. Nesland

https://doi.org/10.1080/01913120490515586

Published online:

10 July 2009

Fig. 1 Transmission electron micrographs of CaD2 mouse tumors taken 1 (a) and 24 (b) hours after m-THPC-mediated PDT. The tumor-bearing mice were given an IP injection of 1 mg/kg m-THPC, and 24 h later, the tumors were exposed to laser light (652 nm, 200 mW/cm² for 100 s). (a) A vessel (V) packed with red blood cells, indicating stasis (originalmagnification : × 4, 840); (b) both endothelial cells and subendothelial zone of the vascular wall (VW) severely damaged as well as fragmented, and blood elements leaking out (original magnification: × 8, 600).

Display full size

Thromboxane, a potent vasoconstrictor, can be formed from damaged cell membrane lipids [[48], [49]] and histamine, a powerful vasodilator, can be released from degranulation of mast cells [[50]]. Moreover, injured endothelial cells can produce von Willebrand factor (vWF), a glycoprotein known to mediate thrombosis and ischemia [[51], [52]].

PDT can significantly slow the rate of tumor regrowth [[9], [53]], a similar effect to so-called tumor bed effect (TBE) induced by ionizing radiation. TBE is due to a destruction of the endothelium of tumor bed, so that the endothelium is unable to proliferate to support the growth of the tumor. Furthermore, destruction of the subendothelial zone of the capillary wall, which consists of collagen fibers and other connective tissue elements, seems to be crucial for effective PDT (Figure 1b) [[54]]. This is consistent with the findings that collagen, elastin, and fibrin have a high affinity for porphyrins and phthalocyanines [[12], [34]]. This may explain why ALA-mediated PDT of solid tumors is much less efficient than AlPcS_n-based PDT, since endogenous porphyrins induced from ALA are mostly localized in endothelial cells and ALA-PDT targets mainly the endothelium rather than the subendothelial layer. As a result, the basic vascular structure, perhaps also function, may not be severely affected. One of the strategies for ALA-PDT is thus to increase vascular destruction. We have tried PDT with ALA and Photofrin (vessel-localizing) in several tumor models and found that such a combination can improve PDT efficiency [[55]]. Similarly, antiangiogenic drugs may improve tumor response to PDT. Gomer's group has recently confirmed that efficacy of Photofrin-PDT can be significantly enhanced by using antiangiogenic agents (EMAP-II and IM862) and cycloxygenase-2-selective inhibitor (NS-398) [[56], [57]].

An effective PDT condition may be obtained from destruction of both stromal and parenchymal compartments of tumor. Although earlier studies have shown that efficiency of PDT correlated with the uptake of porphyrins by tumor tissue [[58]], our data have recently indicated that the effect of PDT is dependent not only on the absolute amount of a dye in tumor tissue, but also on the intratumoral localization pattern of the dye. For example, water-soluble AlPcS₄ and hydrophobic AlPcS_2a have similar yields of ¹O₂ and extents of uptake by the mouse CaD2 tumor, but the efficiency of AlPcS₄-mediated PDT is significantly less than that of AlPcS_2a-PDT in the tumor model. This is probably due to the fact that AlPcS₄ was mainly localized in the stroma of the tumor when it was activated by light, whereas AlPcS_2a distributed in both parenchymal and stromal elements [[9]]. The results suggest that such a dual localization pattern is a favourable factor for high PDT efficiency [[9], [10]]. Although we proposed in 1990 that a mixture of the two dyes may augment PDT effect [[7]], PDT with ALA and Photofrin may be a better example of this proposal [[55]]. In principle, one may manipulate various PDT protocols with more than one light exposure and/or drug administration to achieve both neoplastic and stromal effects. For example, Dolmans et al. have recently found that the sensitizer MV6401 was given at 15 min and 4 h before light irradiation to enable the dye to localize in both vascular and tumor cell compartments of tumor and PDT with this protocol produced more effectively tumor growth delay than a single dose (but the same total drug dose) either at 15 min or 4 h for light activation [[59]].

EFFECTS OF PDT ON STROMAL FIXED TISSUE CELLS

Tumor invasion is a complicated balance between tissue destruction and the synthesis of stromal support. The ability of neoplastic and stromal cells to synthesize and secrete extracellular proteases is a prerequisite for the invasion. Such proteases can digest components of basement membranes and extracellular matrix (laminin, fibronectin, and collagen). Metastasis is the spread of tumors to distant sites that includes the ability of tumor cells to gain access to the vessel, to survive there, and to exit from there as well as grow in a foreign organ.

Matrix metalloproteinases (MMPs) are not only known as collagen degrading enzymes, but also involved in tumor angiogenesis, growth, invasion, and metastatic potential [[60], [61]]. They are often produced by stromal cells rather than by parenchymal ones of a solid tumor. Karrer et al. have reported that a sublethal dose (LD₁₀) of ALA-mediated PDT in vitro of both normal and scleroderma fibroblasts could result in a time-dependent induction of MMP-1 and MMP-3, while the levels of MMP-2, collagen type III and tissue inhibitor of MMP-1 were not altered. This effect was significantly inhibited by sodium azide, a ¹O₂ quencher [[62]]. Gomer's group has found in vivo that endothelial cells and infiltrating inflammatory cells of the BA mouse mammary carcinoma can produce both latent and active forms of MMP-9 after Photofrin-PDT, an enzyme associated with tumor angiogenesis and growth [[63]]. Further, the use of Prinomastat, a synthetic MMP inhibitor, can improve PDT efficiency of the tumor [[63]].

It is still not entirely clear whether PDT has the potential for inducing tumor metastasis, although there is no explicit clinical evidence. Gomer et al. have reported that in C57BL/mice Photofrin-mediated PDT of subcutaneously transplanted mouse Lewis lung carcinoma reduced metastatic rates of the mouse lung tissues [[64]], while Hasan et al. have found that in Copenhagen rats benzoporphyrin derivative-based PDT of orthotopic MatLyLu prostate tumor led to a mean number of lung metastases per animal 9 and 34 times more than untreated animals or animals treated with surgery and PDT, respectively [[65]]. It has been found that PDT can induce a reduction of tumor cell adhesion [[66]], severe vascular destruction, synthesis of MMPs, a massive infiltration of inflammatory cells that can secrete proteases and growth factors, and upregulation of vascular endothelial growth factor and cycloxygenase-2, both of which are potent angiogenic factors [[56], [57]]. All these alterations may have implications of tumor invasion and metastasis.

Haylett et al. have compared the effect of hematoporphyrin ester (HpE)-based PDT on collagen secretion with those of hyperthermia, ionizing radiation, and chemotherapeutic agents in a murine fibroblast and a human fibroblast cell lines in vitro. They have found that, while other therapies known to be associated with scarring gave rise to an increase in collagen levels in the cell lines, HpE-PDT did not enhance collagen synthesis, suggesting that PDT may not induce scar formation [[67]].

EFFECTS OF PDT ON STROMAL INFLAMMATORY CELLS DERIVED FROM THE BLOOD

PDT-induced degradation products of cellular membrane phospholipids (e.g., lysolipids and arachidonic acid metabolites) can trigger a strong inflammatory response with a massive infiltration of inflammatory cells, such as neutrophils, mast cells, monocytes, macrophages, and lymphocytes. This locally nonspecific immune response causes destruction of tumor cells. Phagocytosis and processing of dead and damaged tumor cells and debris by macrophages that are specialized antigen presenting cells gives rise to the tumor antigen presentation in the context of MHC class II molecules to T lymphocytes to produce systemically specific antitumor immunity [[68], [69]]. This immune response is involved in eliminating neoplastic cells, but it may be particularly effective for small foci of tumor cells that have survived initial PDT effects. Furthermore, this reaction may include disseminated and metastatic lesions [[1]]. Although the mechanism of this antitumor immunity is not fully known yet, an increasing amount of evidence suggests that upregulation of chemokines, cytokines, and adhesion molecules, which are responsible for PDT-induced initial activation and accumulation of neutrophils, may play a crucial role in the long-term tumor control [[70]]. Direct interaction between tumor cells and immune cells, which results from PDT-mediated destruction of extracellular elements, may also play an important role in this process [[71]]. Gollnick et al. have recently observed that mouse tumor-cell lysates previously treated with Photofrin-PDT in vitro can be used to prevent the tumor growth in vivo in animals. The PDT-generated tumor-cell lysates can maturate dendritic cells to express interleukin-12, a cytokine associated with the development of a cellular immune response [[72]]. These findings indicate that PDT-induced systemic immune response may be employed as a way of vaccination for treatment and prevention of remaining of primary tumors in certain cases and distant metastatic ones. One of the future challengens in this field is to combine PDT with various types of immunotherapy that allow immune cells to recognize treated tumor cells so effectively that it can enhance tumor-specific immunity [[73–75]].

PDT has, however, been shown to inhibit the activity of natural killer cells [[76]] and cause systemic immune suppression [[77], [78]]. The suppression can be correlated with the increased expression of the inflammatory cytokine interleukin-10, which can suppress cellular immunity [[79], [80]]. Such PDT-induced immunosupression may have a negative effect on tumor destruction; on the other hand, it may be exploited for the treatment of certain autoimmune disorders (rheumatoid arthritis, psoriasis) and organ transplant rejection with a low PDT dose [[81]].

EFFECTS OF PDT ON STROMAL ELEMENTS FOR NONONCOLOGIC DISORDERS

Although PDT started with the treatment of tumor, the most successful application of this modality today is age-related macular degeneration (AMD), a disorder associated with choroidal neovascularization of eyes. The principle of the treatment is to utilize the selective effect of PDT on microvasculature of the disease. Approvals for using PDT of AMD have been granted worldwide since 1999 [[82]].

Demidova and Hamblin have recently shown that macrophages can be selectively (up to 3 logs) targeted by PDT using a sensitizer attached to ligands against scavenger receptor that is upregulated in the cells [[83]]. The impact of this selective PDT effect on macrophages may be significant. For example, animal studies have shown that PDT is effective at destroying macrophages in atherosclerotic plaque of arteries largely via apoptotic mechanism [[84], [85]]. Preliminary clinical data from such patients have indicated that PDT is safe, well tolerated, and nontraumatic, suggesting that this modality may have the potential for atherosclerotic vascular diseases [[86]].