Photodynamic therapy of cancer — Challenges of multidrug resistance

Zheng Huang*†‡, Yih-Chih Hsu, Li-Bo Li¶, Luo-Wei Wang||, Xiao-Dong Song**, Christine M. N. Yow, Xia Lei, Ali I. Musani, Rong-Cheng Luo¶ and Brian J. Day *MOE Key Laboratory of OptoElectronic Science and Technology for Medicine, Fujian Normal University Fuzhou, P. R. China College of Engineering and Applied Science University of Colorado Denver, CO, USA Department of Medicine National Jewish Health, CO, USA Department of Bioscience Technology Center for Nanotechnology and Institute of Biomedical Technology Chung Yuan Christian University, Taoyuan, Taiwan ¶TCM-Integrated Cancer Center Southern Medical University, Guangzhou, P. R. China ||Department of Gastroenterology & Endoscopy Changhai Hospital, Second Military Medical University Shanghai, P. R. China


Introduction
Photodynamic therapy (PDT) of cancer involves the administration of a photosensitizer (PS) followed by illumination of the cancer site with visible light.This process can lead to generation of reactive oxygen species (e.g., singlet oxygen) in the presence of oxygen molecules via photon-induced energy and/or electron transfer.PDT-mediated oxidation can cause local cytotoxicity leading to cancer cell death through apoptosis and/or necrosis pathways.Unlike other oxidant-based cancer therapies, in addition to the rapid direct oxidationdriven cytotoxic e®ects on cancer cells, PDTinduced damage to the tumor vasculature, acute in°ammatory reaction and systemic immunity also play signi¯cant roles in the anticancer e®ectiveness of PDT. 1 As a clinically approved and minimally invasive therapeutic modality, PDT has been used for curative or palliative management of various premalignant (e.g., actinic keratoses, Barrett's Esophagus) and malignant diseases (e.g., obstructive lung cancer and esophageal cancer) worldwide. 2,3lthough it is generally believed that cancer has no signi¯cant resistance toward PDT and can be ultimately ablated through maximizing PDT drug and/or light dose, a better understanding of cancer biology suggests that similar to many other cancer therapies, PDT is inevitably subject to intrinsic cancer resistance at the cellular and molecular level via drug e®lux, hypoxia, levels of pigmentation and damage reversal mechanisms.Serious considerations have to be taken to overcome these defense mechanisms in order to make anticancer PDT more e®ective and acceptable by mainstream medicine.
This review paper will primarily discuss the challenges of cancer resistance involved in the cellular uptake of PS.The strategies that can be used in the clinical setting to overcome this particular type of resistance will also be discussed.

Photosensitizer 2.1. Classi¯cation of PS
PS is one of three critical elements in PDT (i.e., PS, light and oxygen).The majority of PSs possess a heterocyclic ring structure (e.g., tetrapyrrole ring) similar to that of hematoporphyrin or chlorophyll.In general, they can be divided into three broad families: (i) porphyrin-based PS (e.g., Photofrin or Por¯mer Sodium, benzoporphyrin derivative (BPD), hematoporphyrin monomethyl ether (HMME, or Hemo-por¯n)), (ii) chlorophyll-based PS (e.g., chlorins, bacteriochlorins, purpurins) and (iii) dye-based PS (e.g., phtalocyanine, napthalocyanine).Both hematoporphyrin and chlorine PSs are also classi¯ed as porphyrins.Many PSs possess distinct and strong °uorescence that can be used for their in vivo detection and quanti¯cation, which provides a unique tool for photodynamic diagnosis and theranostics. 4raditionally, the porphyrins and those PSs developed in the 1970s and early 1980s are called the ¯rst generation PSs (e.g., Photofrin).Porphyrin derivatives or synthetics of known chemical structures made since the late 1980s are called the second generation PSs (e.g., m-tetrahydroxyphenylchlorin (mTHPC), BPD-MA, HMME, hexyloxyethyl pyropheophorbide-a (HPPH)).5-Aminolaevulinic acid (ALA) and its ester derivatives are also often called PS but they are the prodrug of protoporphyrin IX (PpIX). 5The third generation PSs generally refers to the modi¯cations such as biologic conjugates (e.g., antibody conjugate) and built-in photo quenching or bleaching capability.Target-speci¯c PDT refers to the use of the antibody-or antisenseconjugated PSs, which combines the speci¯city to an over-expressed cellular marker with the phototoxic properties of the conjugated PDT PS.The targeted cellular marker can be a cancer-associated or noncancer-associated marker.The conjugation may or may not necessarily enhance the internalization process, although the internalization might enhance PDT-induced cytotoxicity.

Mode of actions
PSs are currently administrated systemically (mainly intravenously) or topically (mainly for prodrugs such as heme precursor) in clinical settings.The cancerlocalizing properties of PS might include the preponderance of leaky and tortuous blood vessels due to neovascularization and the absence of lymphatic drainage known as the enhanced permeability and retention e®ect of PS in tumor tissues.Some of the most e®ective PSs bind preferentially to low-density lipoprotein (LDL), suggesting that upregulated LDL receptors in cancer cells could be an important factor in PS uptake. 1 PSs can be covalently attached to various biomolecules that have some a±nity for neoplasia or to receptors expressed on cancer cells and result in unique subcellular localization patterns.In general, water soluble porphyrin ethers (e.g., Photofrin) have variable localization patterns mostly associated with lipid membranes.Mono-L-aspartyl chlorin e6 (NPe6, talapor¯n) targets the lysosomes.BPD targets the mitochondria.mTHPC can target the mitochondria, endoplasmic reticulum (ER), or both.Phthalocyanine Pc4 has a broad spectrum of a±nity to di®erent organelles. 6Nevertheless, it should be noted that speci¯c patterns of cellular localization may vary among di®erent cell types and therefore trigger di®erent photocytotoxicity.The presence of competitive binding agents can also in-°uence the subcellular localization and/or binding of PSs. 7 As one can expect, the presence of prooxidant factors in cancer cells might scavenge PDT-induced oxygen species and therefore have a negative e®ect on PDT e±cacy. 8reat variations in PS uptake can be found between individual cell lines, resulting in even more pronounced di®erences in photocytotoxicity.It is important to be able to predict PS uptake pro¯les in the clinical setting in order to adjust the dose for e®ective and complete cancer elimination.Meanwhile, one should be aware that at equivalent cellular PS levels, there are many other factors (e.g., level of oxygenation, presence of antioxidant) that might a®ect the sensitivity of cancer cells to PDT.

MDR transporters
Multidrug resistance (MDR) is a phenomenon where resistance to one anticancer chemotherapy drug is accompanied by resistance to drugs with di®erent chemical structures and mechanism of action. 9MDR is often attributed to the over-expression of certain members of ATP-binding cassette (ABC) transporter proteins (also known as e®lux pumps) including p-glycoprotein (P-gp/ABCB1/MDR1), multidrug resistance proteins (MRPs, e.g., MRP1-9), and breast cancer resistance proteins (BCRP/ABCG2/ MXR/ABCP). 10ABC transporters form one of the largest protein superfamilies encoded in the human genome, and more than 48 human ABC protein genes have been identi¯ed.

Interaction of MDR and PS
Early studies in the 1990's showed that PDT-resistant variants obtained from multiple PDT treatments did not exhibit MDR phenotype nor did they have altered the uptake properties of porphyrinbased PS. 11,12 PDT-resistant variants did not display a broad cross-resistance to di®erent types of PSs, suggesting that the mechanism of PDT resistance may, to some extent, depend upon the physical nature of the PS molecule. 13It was generally believed then that chemotherapy-resistance or radiotherapy-resistance was not signi¯cantly crossresistant to porphyrin-based PDT nor did PDT induce resistance to chemotherapy or radiotherapy.Some PSs might even act as a MDR reverser, e.g., methylene blue (MB) on P-gp. 146][17][18][19] showed that PDT-resistant squamous carcinoma cells had a more ¯broblastic morphology, higher number of stress ¯bers, more expression of cell-substrate adhesion proteins and higher expression of phospho-survivin but few di®erences in intracellular PpIX content after incubation with ALA methyl derivative.20 Topical ALA PDT has been successfully used for some resistant cases of cutaneous T-cell lymphoma.21 Chu and Yow showed that hexyl ALA-mediated PDT could sig-ni¯cantly provoke an up-regulation of phosphorylated p38MAPK and c-Jun N-terminal kinase (JNK) proteins in P-gp expressing doxorubicin-resistant human uterine sarcoma cells (MES-SA/Dx5).22 Interestingly, a recent study suggests that pheophorbide-a mediated PDT could inhibit the MDR activity by down-regulating the expression of P-gp via JNK activation.23 However, the heterogeneity of PS uptake has been demonstrated in various in vitro and in vivo models.24,25 The mechanisms of resistance to PDT ascribed to the PS may be shared with the general mechanisms of MDR, and are related to altered PS uptake and e®lux rates or altered intracellular tra±cking within cancer cells.26

P-gp and PS uptake
In the mid-1990's, Luna et al. demonstrated that certain cellular receptors (e.g., alpha-2 macroglobulin receptor/low density lipoprotein receptorrelated protein) could modulate PS uptake and a®ect PDT sensitivity of targeted cells. 27For the ¯rst time, Purkiss et al. reported that the P-gp export mechanism may have an e®ect on the cytotoxicity of PDT by reducing the concentration of hematoporphyrin derivative (HpD) within human colorectal cancer cells (HRT 18) since P-gp mediated resistance to PDT could be reversed through modulation with verapamil (an antagonist of Pgp). 28An in vitro study demonstrated that the addition of verapamil could increase the intracellular levels of PSD-007 (a mixture of hematoporphy derivative) and DNA content in colon cancer cells, meanwhile decreasing S and G1 phase cells. 29lthough this is contradictory to early studies showing that over-expression of P-gp in a mouse ¯broblast cell line (3T3 cells) had no signi¯cant e®ect on the cellular concentration of chloroaluminum tetrasulfonate phthalocyanine (AlSPc) nor did over-expression of P-gp in human breast adenocarcinoma cells (MCF-7 cells) a®ect chlorin e6 accumulation. 30,31Early studies also suggested that the intra and extracellular PpIX accumulation mediated by ALA were not subjected to the level of P-gp expression, 32 but the rapid e®lux of PpIX had also been demonstrated in DT-resistant variants. 7avitskiy et al. showed that the speci¯c cellular protein P-gp 170 did not appear to alter the intracellular accumulation of chlorins. 33Later Saczko et al. showed that P-gp appeared to play a role in the intracellular accumulation of Photofrin but not hypericin in doxorubicin-resistant human colon cancer cell lines (LoVo cells). 34Horibe et al. demonstrated that cisplatin resistance in A549 cells that had high level of P-gp mRNA had no signi¯cant in°uence on accumulation and photodynamic activity of chlorin e6. 35Most transporters have transmembrane domains (TMD).P-gp are con¯ned to membrane loci associated with the transporter and it was believed that it might have little e®ect on the migration of cytotoxic photo-products. 36evertheless, the in°uence of P-gp on PS uptake in cancer cells remains inconclusive.

BCRP and PS uptake
In addition to early interest in the e®ect of P-gp expression and e®lux mechanism on PS uptake, for the ¯rst time Robey et al. demonstrated in the early 2000's that PSs with similar structure to that of pheophorbide-a were a substrate of BCRP but not to other major drug e®lux transporters such as P-gp or MRP1. 37,38They reported that BCRP-transfected human embryonic kidney cells (HEK-293 cells) were 11-fold, 30-fold, 4-fold, and >7-fold resistant to PDT mediated with pheophorbide a, pyropheophorbide a methyl ester, chlorin e6 and ALA, respectively.BCRP, a member of the phase III system of xenobiotic metabolism, is responsible for protecting the body from toxic xenobiotics and for removing toxic metabolites, including the transport of porphyrin and chlorophyll metabolites.Liu et al. demonstrated that the use of tyrosine kinase inhibitors (e.g., imatinib mesylate) can block the function of BCRP and increase accumulation of HPPH, PpIX and BPD-MA from 1.3-to 6-fold in BCRPþ cells and consequently enhance PDT e±cacy in RIF-1 tumor model. 39Jendzelovský et al. showed that Proadifen, an inhibitor of cytochrome P450 enzymes, could a®ect the function of BCRP and MRP1 leading to increased hypericin content in colon cancer cells (HT-29 cells). 40BCRP-mediated PpIX e®lux was also a major factor that prevented PpIX accumulation in human urothelial carcinoma cells (T24 cells). 41Bebes et al. showed that the PpIX extrusion ability of keratinocytes (HaCaT cells) was correlated with their BCRP expression which was higher in proliferating cells than in di®erentiated cells. 42The speci¯c inhibition of BCRP enhanced the sensitivity of keratinocytes to ALA PDT which improved the topical PDT of skin lesions.In addition to keratinocytes, it also enhanced the sensitivity of human esophagus cells (OE19 adenocarcinoma) and bladder cells (HT1197 carcinoma) to ALA/MAL PDT. 43Along with BCRP, peptide transporter PEPT1 has been identi¯ed as an ALA in°ux transporter and participates in the regulation of intracellular PpIX levels in human gastric cancer cells. 44 large number of single nucleotide polymorphisms (SNP) were identi¯ed for a variety of drug transporters, which provides a useful means to determine the relationship between nonsynonymous polymorphisms and the substrate speci¯city of drug transporter proteins.Based on SNP data, Tamura et al. demonstrated in insect Spodoptera frugiperda Sf9 cells that the amino acids at position 431, 441 and 489 located in TMD were critically involved in substrate recognition and/or transport of drugs. 45ore speci¯cally, the S441N variant of BCRP completely lost transport activity for both hematoporphyrin and methotrexate but the F431L and F489L variants maintained hematoporphyrin transport but lost the activity of methotrexate transport.Later they showed that Flp-In-293 cells containing S441N and F489L variants exhibited high levels of both cellularly accumulated pheophorbide-a and photosensitivity.The accumulation of PpIX from ALA and pheophorbide-a in the cytoplasm compartment was maintained at low levels in Flp-In-293 cells expressing ABCG2 WT, V12M or Q141K.However, in the presence of BCRP inhibitor imatinib or novobiocin, those cells became sensitive to light. 45,46They further demonstrated that the planar structure of inhibitors was an important factor for interactions with the active site of BCRP.These results suggested that certain genetic polymorphisms and/or inhibition of BCRP could enhance porphyrin-mediated photosensitivity.The over-expression of BCRP in glioma stem-like cells (GSC) isolated from U251 glioma cells can result in e®lux of Photofrin, which can be reversed by a pretreatment of GSC with a speci¯c BCRP inhibitor fumitremorgin C (FTC). 47The DNA damage reversal mechanisms may have important functions in Photofrin PDT resistance through the activation of alkylation repair homologue 2 by tumor protein TP53 in glioma cells. 48he PS selectivity of BCRP was still unclear at that time.Usuda et al. demonstrated that BCRPoverexpressing human epidermoid carcinoma cells (A431 cells) were more resistant to Photofrin-PDT and FTC could reverse such resistance.However, the cell line did not show cross-resistance toward NPe6. 49What was more signi¯cant was that they further examined 81 tumor specimens obtained from patients with centrally located early lung cancers that underwent PDT treatment.All specimens were BCRP-positive.The expression of BCRP signi¯cantly a®ected the e±cacy of Photofrin-PDT in cancer lesions 10 mm in diameter.On the other hand, NPe6-PDT exhibited a strong antitumor e®ect, regardless of the expression status of BCRP in the lung cancers.They suggest that Photofrin may be a substrate of BCRP and be pumped out from cancer cells, therefore, the PS selectivity of BCRP may be a molecular determinant of the outcome of PDT.This translational study represents a milestone work in addressing the profound impact of cancer MDR on PS selectivity and PDT outcome.
PSs without substitutions including pyropheophorbides (e.g., HPPH) and purpurinimides are general substrates for BCRP.Morgan et al. demonstrated in BCRP-expressing HEK-293 cells that carbohydrate groups conjugated at positions 8, 12, 13 and 17 but not at position 3 could abrogate BCRP a±nity regardless of structure or linking moiety. 50Yet, they showed that in the murine mammary tumor (4T1) HPPH but not the galactose conjugate of HPPH selectively preserved a small ABCG2-expressing side population (SP) which is believed to be primarily responsible for tumor regrowth.A PDT-resistant SP may be responsible for recurrences observed both preclinically and clinically.The SP could be targeted by addition of imatinib mesylate and ultimately preventing PS e®lux.Tracy et al. demonstrated in tumor/stroma cocultures derived from lung cancers that HPPH was lost from ¯broblastic cells more rapidly than from epithelial cells, even under low BCRP expression, facilitating selective eradication by PDT of epithelial over ¯broblastic cells. 51Enhanced BCRP expression led to the selective PDT survival of tumor cells in tumor/stroma co-cultures.This survival pattern was reversible through no-substrate HPPH derivatives or imatinib mesylate.They concluded that PS retention, not di®erences in subcellular distribution or cell signaling responses, was determining cell type selective death by PDT.Therefore, up-front knowledge of cancer characteristics, speci¯cally MDR status, could be helpful in individualizing PDT treatment design.

Possible Strategies in Clinical Setting 4.1. Inhibition of MDR transporters
To prevent MDR-mediated resistance, in addition to utilizing nonsubstrate PS and various PS conjugates, administering an MDR inhibitor alongside a substrate PS is another straightforward strategy to reduce the PS e®lux rate in MRP expressing cells.Although numerous in vitro and in vivo studies show that the co-administration of a MRP inhibitor could reverse the e®ect of certain MRP transporters on PS accumulation, such combination has not yet been tested in clinic.
The utilization of MRP inhibitor (or modulator or antagonist) in the treatment of solid tumor is under clinical investigation worldwide.Preliminary data suggest that for P-gp, a single dose of verapamil 120 mg or 80 mg three times daily (total daily doses of 240 mg) for 6 days could improve bioavailability of P-gp substrates or chemotherapy drugs. 52,53It is noteworthy that clinical trials indicated that the Pgp-modulating agent Valspodar did not improve the treatment outcome of refractory multiple myeloma and P-gp inhibitors may not be a practical solution to MDR.Ceramide, the central molecule of sphingolipid metabolism, generally mediates antiproliferative and proapoptotic functions.Changes of the bioactive sphingolipid ceramide in various types of cancer cells have been observed in response to PDT.Korbelik et al. demonstrated that combining ceramide analog treatment with PDT could enhance intracellular calcium release in cancer cells and strongly promote apoptosis after PDT treatment. 54Since the mechanism the drug resistance which develops with increased glucosylceramide expression is associated with P-gp overexpression, 55 it can be expected that combining P-gp inhibitors with PDT might enhance the ability to generate intracellular ceramide and amplify apoptotic death.Nevertheless, Wagner et al. demonstrate that the third-generation P-gp modulator tariquidar could inhibit P-gp function at the human blood-brain barrier (BBB), which might be a useful approach for PDT of brain tumor. 56ecently, Sun et al. demonstrated that pretreatment of human glioma cells with Ge¯tinib, a BCRP inhibitor, could enhance intracellular accumulation of PpIX through the inhibition of BCRP expression and BCRP-mediated PpIX e®lux, ultimately improving the e®ectiveness of ALA-PDT. 57n addition to chronic myelogenous leukemia, the ¯rst indication for imatinib mesylate (Gleevec or Glivec) for the treatment of solid tumor was approved by the US Food and Drug Administration (FDA) in 2001.For advanced tumor, the recommended dose of imatinib mesylate as a molecular targeted cancer drug is 400 or 600 mg daily.However, chronic exposure to imatinib was shown to result in upregulation of P-pg and BCRP transporters in Caco-2 cells, but this phenomenon was not reproducible in the hepatic and intestinal compartments in mice. 58,59ese preclinical and clinical studies suggest that the co-administration of a MRP inhibitor and PS might be a feasible strategy for PDT.Preclinical studies show that co-administration of imatinib mesylate and PS could indeed inhibit PS e®lux although their combination in clinical setting still needs to be optimized and validated.Nevertheless, it should be noted that the intervals between the administration of inhibitor and chemotherapy drugs (or PS) is critical for the inhibitor to be e®ective.Dosing and scheduling of co-administration of inhibitor and PS should be carefully explored due to di®erent pharmacokinetics of PS and inhibitor.Moreover, the use of inhibitor might increase systemic toxicity of chemotherapy.This could be a concern for PS since the increase of PS in°ux and peripheral PS accumulation might alter the skin photosensitization.

Antivascular PDT
Cancer treatment can be exerted by targeting both cancer cells and the vasculature supplying solid tumors.Antivascular PDT or vascular-targeting PDT (VTP or vPDT) represents the recent progress in PDT that can meet the paradigm shift in cancer treatment.vPDT is characterized by a short drug to light interval (DLI), typically 0-30 min after the completion of intravenous (iv) injection of PS.The PS used in vPDT should have fast clearance and therefore might not selectively accumulate in cancer cells.In vPDT, light irradiation takes place while the PSs are still circulating in the vascular compartment and, therefore, cause vascular damage and lead to thrombosis and micro-vessel occlusion. 60vPDT has been used primarily for the management of the neovascularization lesions (e.g., wet age-related macular degeneration, AMD) and cutaneous capillary malformations (e.g., port wine stain birthmarks, PWS). 61,62For MDR cancers, vPDT can target nonmalignant vascular networkthe lifeline of cancer and therefore bypass MDR transporters and o®er a novel approach to treat MDR expressing solid tumors.Although vPDT might change the traditional criteria of PS selection, longer wavelength and rapid clearance might be the key criteria for designing a PS for antivascular PDT.
Pd-bacteriochlorophyll based PSs have a high extinction coe±cient in the near-infrared (IR) spectrum and rapid clearance from the blood circulation and skin after iv injection.Preise et al. showed that P-gp expressing human HT29/MDR colon carcinoma cells were resistant in vitro to PDT medicated with Pd-bacteriopheophorbide (TOO-KAD, also known as WST09), however, the vPDT with iv injection of TOOKAD and immediate light irradiation (0 min DLI) induced tumor necrosis with equal e±cacy in HT29/MDR-derived xenografts and their wild-type counterparts. 63These results are ascribed to the rapid antivascular e®ects of vPDT, suggesting that MDR cancers can be successfully eradicated by indirect approaches that bypass their inherent drug resistance.Moreover, targeting tumor vessels and angiogenesis might reduce the risk of metastasis. 64ermanent occlusion of feeding arteries and draining veins in solid tumor have been demonstrated in vPDT of mouse model. 657][68] The massive shutdown of pathological and normal vessels in the tumor can deprive the supply of oxygen and nutrients and subsequently achieve tumor ablation.To ablate a bulky solid tumor it might require combining both cellular-targeting and vascular-targeting approaches.Co-administration of anti-angiogenic agent (e.g., inhibitor of pro-angiogenic factor, endogenous inhibitor) and use of nanocarriers consisting of vasculature targeting agent (e.g., NRP-1 peptide) and PS represent some new developments in targeted therapy. 69

Photochemical internalization
Photochemical internalization (PCI) is a novel site-speci¯c drug and gene delivery method developed to improve the intracellular release of macromolecules and hydrophilic chemotherapeutic agents from endosomes and lysosomes. 70PCI is based on the combination of endosomal and lysosomal localizing amphiphilic PSs and light, therefore it could be considered as a form of intracellular PDT with a primary goal of time-and space-controlled and lighttriggered drug delivery.After activating PS by light, photodynamic reactions result in destruction of endocytic vesicle membranes and subsequently release the entrapped drugs into the cytosol of targeted cells.Although PCI might reverse or bypass the MDR phenotype by endo-lysosomal release of the MDR substrate drug, PCI of macromolecular therapeutic agents that are not targets of MDR transporters represents another therapeutic strategy to treat MDR cancer. 71,724][75] PCI of numerous macromolecules has been demonstrated in vitro and in vivo.Disulfonated tetraphenyl chlorin (TPCS 2a ) is found to be a clinically suitable PCI PS for photochemical activation of molecules that do not readily penetrate the cellular plasma membrane.It is currently subject to a ¯rst clinical trial in patients with various cancers.Preliminary results suggest that PCI seems to be a promising treatment modality for MDR cancer.PCI may exert direct cytotoxic e®ects on endothelial cells, which lays a foundation for utilizing the PCI technology as an antivascular strategy to ablate tumors. 76

Intratumoral injection
A systemic administration of certain PSs can cause prolonged skin photosensitization and result in poor tumor selectivity.These drawbacks might be overcome by intratumoral injection of PS.For small and localized cancers, there is still a need to explore intratumoral delivery approaches, this is particularly true since many cancers are detected at an early stage at a small size.
8][79] The e®ectiveness and safety of intratumoral injection of PS can be a®ected by drug formulation, injection volume, velocity and site.The use of lipid-based PS (e.g., Foscan) and nanocarriers might improve the PS distribution and retention in targeted tumor. 80n e®ective control of cancer might require the selective destruction of parenchymal and/or stromal tissue.It is well known that the generation of a reactive stroma environment can promote tumorigenesis.Fibroblast-activation protein (FAP) is a membrane-bound serine protease that is expressed on the surface of reactive stromal ¯broblasts present within the majority of human epithelial cancers but is not expressed by normal tissues.Therefore, FAP represents a potential pan-tumor target whose enzymatic activity can be exploited for the intratumoral activation of prodrugs and protoxins. 81This represents a paradigm shift in cancer therapy and inspires people to develop PS suitable for intratumoral injection that can bypass MDR and generate a high PS concentration inside tumor stromal tissues.Lo et al. developed a novel FAP-triggered PDT beacon which could induce signi¯cant photocytotoxicity in FAP-expressing cells and be activated in FAP-expressing stromal ¯broblasts in vivo. 82Although many PS-conjugates are unsuitable for systemic administration for many reasons (e.g., systemic toxicity, high cost), they may be an ideal candidate for intratumoral delivery.

Nanodelivery
Drug delivery is a key determinant of drug e±cacy in cancer chemotherapy.Because of unique physicochemical properties of nanomaterials, such as small size, large surface area to mass ratio and high reactivity, nanocarrier-based approaches have shown great promise for carrying, protecting and delivering potential therapeutic molecules with diverse physiological properties.Current nanotechnology is revolutionizing drug delivery by improving pharmacokinetics, biodistribution, speci-¯city and molecular targeting of cancer therapeutics.Nanocarrier-based approaches not only can circumvent limitations in the delivery of cancer therapeutics, related to their poor aqueous solubility and toxicity issues with conventional vehicles, the use of nanocarriers (e.g., liposomes, nanoemulsions, nanoparticles, carbon nanotubes) can also overcome MDR in cancer therapy. 83Some of nanopreparations have advanced to clinical trials.
It is highly likely that nanotechnology will modify and alter both the basic science and clinical applications of PDT. 84,85Noticeably, most of the current studies are aimed at either improving existing formulations of clinically approved PS or focused on the development of targeted delivery vehicles. 86ome of the PS currently used in clinics are in fact nanosized materials, e.g., liposomal formulations (e.g., Foscan, Visudyne), which have shown improved PS distribution and retention in targeted tumor. 80Actively targeted liposomes can be developed by conjugating ligands (e.g., glycoproteins, peptides, oligonucleotide aptamers, antibodies) to the liposomal surface which allow speci¯c targeting to certain cancer cells.
In addition to the potentials of nanodelivery of PS in vasculature targeting and PCI approaches, 69,[73][74][75] the combination nanocarrier for dual modalities has also been used to overcome drug resistance.Khdair et al. showed that a combination of doxorubicin and MB bound to Aerosol OT alginate nanoparticles had signi¯cant therapeutic potential against tumors expressing P-gp. 87he most advanced nanocarrier systems combining disease diagnosis with therapy (theranostics) is also noteworthy.Due to the dual functions of °uorescence and photosensitization, PS is a good candidate for developing cancer theranostics. 88

Enhancing antitumor immunity
Drug resistance strongly argue for innovative strategies to treat and manage cancer.Stimulating the power of cancer patient's own immune defense is a highly attractive strategy to complement the activity of standard chemotherapy.Moreover, several immunotherapy approaches could be used to combat cancer MDR.For instance, direct immune attack against MDR transporters, using MDR as an immune target to deliver cytotoxic agents, conditional immunotoxins expressed under MDR control, and modulating immunogenic potential of some cytotoxic agents. 89reclinical studies have shown that PDT could enhance local and systemic antitumor immunity and increased expression of proin°ammatory cytokines play a key role in initiating speci¯c cellular and humoral antitumor immunity.The implications of PDT-induced antitumor immunity and e±cacious PDT-generated vaccines provide a possibility for using PDT in the treatment of metastatic disease and as an adjuvant in combination with other modalities for treating MDR cancers. 1,90,91On the other hand, the presence of an intact adaptive immune system could bene¯t the long-term e±cacy of antitumor PDT since both the direct cancer cell killing and the control of cancer cells revival after treatment are equally crucial.
Although antitumor immunity is able to speci¯cally target cancer cells, the existence of a variety of immune escape mechanisms can be involved in minimizing the overall e®ectiveness of cancer therapy.Therefore, the elimination of immunosuppressive activities in tumor microenvironment is another attractive strategy for enhancing the e®ectiveness of cancer immunotherapy.Immune-suppressive cells include a heterogeneous population of immature myeloid cells expanded systemically as a consequence of a profound tumor-associated pro-in-°ammatory milieu.Recently, Barth et al. demonstrated in a mouse model that PDT might overcome immunosuppressive cells via the regulation of immature myeloid cells and the in°ammatory milieu critical to their expansion during tumor progression. 92Reginato et al. showed that the depletion of T-regulatory cells could potentiate PDT-mediated immunity. 93

Conclusive Remarks
PS is a critical element in PDT.Although to a certain extent the quantity and location of PS can predict the nature of photodynamic reactions and determine the consequence of anticancer e®ect, it should be aware that at equivalent cellular PS levels, there are many other factors that might a®ect the sensitivity as well as phototoxicity of cancer cells to PDT.Mounting evidence suggests that many PSs are substrates of MDR transports and PS e®lux mediated by P-gp and BCRP can negatively a®ect anticancer e±cacy.Therefore, the screening of cancer MDR pro¯le can be helpful in individualized PS selection and PDT treatment design.
To meet this signi¯cant paradigm shift and prevent MDR mediated resistance, in addition to utilizing nonsubstrate PS or PS conjugates, administering an MDR inhibitor alongside a substrate PS is a feasible strategy to reduce the PS e®lux in MDR expressing cells.However, dosing and scheduling of co-administration of MDR inhibitor and PS are yet to be investigated since PS and inhibitor could have di®erent pharmacokinetics.
Much progress has been seen in both basic research and clinical application in recent years.The majority of approved PDT clinical protocols have primarily been used for the treatment of super¯cial lesions of both malignant and nonmalignant diseases.The implication of antivascular PDT, PCI, nanodelivery and immunotherapy in bypassing the MDR transports represents novel approaches in anticancer PDT.It can be expected that in conjunction with PS of longer excitation wavelengths, these approaches might provide an e®ective alternative for the treatment of deep-seated tumors.