Photodiagnosis and Photodynamic Therapy (2004) 1, 279—293

REVIEW

Mechanisms in photodynamic therapy:

part one—–photosensitizers, photochemistry
and cellular localization
Ana P. Castano a,b, Tatiana N. Demidova a,c,
Michael R. Hamblin PhD a,b,∗

a Wellman Center for Photomedicine, Massachusetts General Hospital, 40 Blossom Street,

Bartlett 3, Boston, MA 02114, USA
b Department of Dermatology, Harvard Medical School, USA
c Department of Cellular, Molecular and Developmental Biology, Tufts University, USA

Available online 8 March 2005

KEYWORDS Summary The use of non-toxic dyes or photosensitizers (PS) in combination with
Photodynamic therapy; harmless visible light that is known as photodynamic therapy (PDT) has been known
Photosensitizers; for over a hundred years, but is only now becoming widely used. Originally developed
Photochemistry; as a tumor therapy, some of its most successful applications are for non-malignant
Photophysics;
disease. In a series of three reviews we will discuss the mechanisms that operate in
the field of PDT. Part one discusses the recent explosion in discovery and chemical
Tissue optics;
synthesis of new PS. Some guidelines on how to choose an ideal PS for a particular
Subcellular localization
application are presented. The photochemistry and photophysics of PS and the two
pathways known as Type I (radicals and reactive oxygen species) and Type II (singlet
oxygen) photochemical processes are discussed. To carry out PDT effectively in vivo,
it is necessary to ensure sufficient light reaches all the diseased tissue. This involves
understanding how light travels within various tissues and the relative effects of
absorption and scattering. The fact that most of the PS are also fluorescent allows
various optical imaging and monitoring strategies to be combined with PDT. The most
important factor governing the outcome of PDT is how the PS interacts with cells in
the target tissue or tumor, and the key aspect of this interaction is the subcellular
localization of the PS. Examples of PS that localize in mitochondria, lysosomes,
endoplasmic reticulum, Golgi apparatus and plasma membranes are given. Finally
the use of 5-aminolevulinic acid as a natural precursor of the heme biosynthetic
pathway, stimulates accumulation of the PS protoporphyrin IX is described.
© 2005 Elsevier B.V. All rights reserved.

* Corresponding author. Tel.: +1 617 726 6182; fax: +1 617 726 8566.
E-mail address: [email protected] (M.R. Hamblin).

1572-1000/$ — see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/S1572-1000(05)00007-4
280 A.P. Castano et al.

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Photochemistry and photophysics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Photosensitizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Photophysics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Photochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
Light delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Subcellular localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Lysosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Mitochondria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Plasma membrane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Golgi apparatus and endoplasmic reticulum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
ALA-induced PPIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290

Introduction on developing new PS [12,13], and at the present

time there is such a great number of potential PS
Photodynamic therapy (PDT) can be defined as the for PDT that it is diffcult to decide which ones are
administration of a nontoxic drug or dye known as suitable for which particular disease or application.
a PS either systemically, locally, or topically to a Some PS can easily be prepared by partial syntheses
patient bearing a lesion (frequently but not always starting from abundant natural starting materials,
cancer), followed after some time by the illumi- such as heme, chlorophyll and bacteriochlorophyll.
nation of the lesion with visible light (usually long This route leads to both economical and environ-
wavelength red light), which, in the presence of mental advantages compared to complicated total
oxygen, leads to the generation of cytotoxic species chemical synthesis [14].
and consequently to cell death and tissue destruc- In parallel with the advances in chemistry there
tion. The concept dates from the early days of the has also been much activity in developing new
twentieth century when workers used dyes such light sources. These include user-friendly lasers fre-
as eosin together with light to treat skin cancer quently based on solid state laser diodes, as well
[1]. Hematoporphyrin (HP) was also first used at as inexpensive light-emitting diodes and filtered
this time [2] and sporadic reports [3] of both se- broad-band lamps [15]. Advances in knowledge of
lective localization of porphyrins in tumors and re- tissue optics has allowed great improvements to be
gression after exposure to visible light appeared un- made in treatment planning and predicting how the
til the 1960s. The modern explosion of interest in light is distributed within the target tissue or organ
PDT dates from the discovery of hematoporphyrin and therefore to optimizing the clinical outcome.
derivative (HPD) by Lipson and Baldes in 1960 [4], The recognition that different tissue types have
and was fueled by pioneering studies in both basic very different optical properties, and even that the
science and clinical application [5—7] by Dougherty same tissue type or organ can vary markedly be-
et al. (notable among many groups). A semi-purified tween individuals in how light is absorbed and scat-
preparation of HPD known as Photofrin® (PF) was tered. The fact that most PS are also fluorescent as
the first PS to gain regulatory approval for treat- well as photochemically active, means that imaging
ment of various cancers in many countries through- and detection strategies can be applied in PDT pro-
out the world, including the United States. After ex- tocols. These techniques are sometimes known as
perience of treating tumors with HPD—PDT was ac- photo(dynamic) detection or diagnosis. They may
cumulated, it was realized that this compound had be carried out to detect otherwise hidden disease
disadvantages, including prolonged skin sensitivity such as dysplasia, to delineate tumor borders, or to
necessitating avoidance of sunlight for many weeks visualize disease in inaccessible areas such as the
[8], sub-optimal tumor selectivity [9], poor light esophagus, bronchus or colon that can, however,
penetration into the tumor due to the relatively be reached endoscopically. Another application of
short wavelength used (630 nm) [10], and the fact fluorescent imaging and quantification is in its abil-
that it was a complex mixture of uncertain struc- ity to improve PDT dosimetry. For instance, fluores-
ture [11]. In recent times much work has been done cence measurements can be made to quantify the
Mechanisms in photodynamic therapy 281

actual amount of PS in the patient’s lesion before determine its chemical structure and to identify
deciding on the appropriate illumination parame- its components [11,16—19]. There was significant
ters. Fluorescence measurements can be also be variation between batches and attempts to frac-
made to measure photobleaching (see later) of the tionate it into its individual component molecules
PS in the tissue, which under some circumstances, frequently yielded mixtures as complicated as the
can be a surrogate marker for optimal completion starting material [20,21]. Although there is good
of the treatment. Although PS are usually selected evidence for the presence of hematoporphyrin
based on photochemical and pharmacokinetic con- oligomers, it is uncertain whether they are pre-
siderations, in the future there may also be an addi- dominantly ethers or esters, and the whether the
tional factor to be taken into consideration involv- side chains are predominantly vinyl or hydroxy
ing the need for fluorescence imaging. ethyl groups [22]. When these uncertainties were
In order to make rational choices from among combined with other significant deficiencies of the
the myriad available PS and light sources available, preparation, enthusiasm for its widespread use was
it is necessary to understand some of the mech- decreased. These deficiencies include a long-lasting
anistic aspects of how PS behave upon illumina- skin photosensitivity so that patients may have to
tion and what happens to the PS when they are avoid sunlight for as long as eight weeks, the lack of
put in contact with mammalian cells in tissue cul- a reasonably-sized absorption band > 650 nm, and
ture. In recent years there has been an explosion the fact that its tumor-localizing properties were
of knowledge in cell biology about signal transduc- not as pronounced as first thought. These consid-
tion pathways, transcription factors and regulation erations spurred a large effort amongst organic
of cell cycle control, inflammation and cell death. chemists to develop novel PS that could in the-
Some of the mechanisms involved after PDT of cells ory be candidates for mediating PDT. The net result
in culture will be covered in the second part of is a collection of probably hundreds of compounds
this review series. The precise way that PDT influ- and it can be bewildering to try to choose among
ences these cellular pathways, is largely governed them.
by where in the cell the PS is located. This subcel- The characteristics of the ideal PS have been
lular localization in turn is governed by the chem- discussed in recent reviews [23,24]. They should
ical nature of the PS (molecular weight, lipophilic- have low levels of dark toxicity to both humans
ity, amphiphilicity, ionic charge and protein binding and experimental animals and low incidence of
charactersitics), the concentration of the PS, the administrative toxicity (i.e. hypotension or aller-
incubation time, the serum concentration and the gic reaction). They should absorb light in the red
phenotype of the target cell. or far-red wavelengths in order to penetrate tis-
One of the chief attractions of PDT as a therapy sue (see later). Absorption bands at shorter wave-
is the concept of dual selectivity. Collateral dam- lengths have less tissue penetration and are more
age to normal tissue can be minimized by increas- likely to lead to skin photosensitivity (the power in
ing the selective accumulation of the PS in the tu- sunlight drops off at  > 600 nm). Absorption bands
mor or other diseased tissue, and by delivering the at high wavelengths (> 800 nm) mean that the pho-
light in a spatially confined and focused manner. tons will not have sufficient energy for the PS triplet
Nevertheless PDT can have side effects including state to transfer energy to the ground state oxygen
long-lasting skin photosensitivity, occasional sys- molecule to excite it to the singlet state (see later).
temic and metabolic disturbances, and excessive They should have relatively high absorption bands
tissue destruction at the treated site. It is hoped (>20,000—30,000 M−1 cm−1 ) to minimize the dose
that advances in mechanistic understanding of PDT of PS needed to achieve the desired effect. Syn-
will minimize the risk reward ratio and extend the thesis of the PS should be relatively easy and the
number of disorders both serious and more minor starting materials readily available to make large-
that can be treated. scale production feasible. The PS should be a pure
compound with a constant composition and a sta-
ble shelf life, and be ideally water soluble or solu-
ble in a harmless aqueous solvent mixture. It should
Photochemistry and photophysics not aggregate unduly in biological environments as
this reduces its photochemical efficiency. The phar-
Photosensitizers macokinetic elimination from the patient should be
rapid, i.e. less than one day to avoid the necessity
Hematoporphyrin derivative or Photofrin was the for post-treatment protection from light exposure
first PS to be studied in detail. However, it proved and prolonged skin photosensitivity. A short interval
highly frustrating for scientists who attempted to between injection and illumination is desirable to
282 A.P. Castano et al.

facilitate outpatient treatment that is both patient- ally has a respectable magnitude. As can be imag-
friendly and cost-effective. Pain on treatment is un- ined the presence of four phenyl groups (or even
desirable, as PDT does not usually require anesthe- worse four naphthyl groups) causes solubility and
sia or heavy sedation. Although high PDT activity is aggregation problems. PCs are frequently prepared
thought to be a good thing, it is possible to have with sulfonic acid groups to provide water solubil-
excessively powerful PS that are unforgiving. With ity and with centrally coordinated metal atoms. It
limitations in both PS and light dosimetry, highly ac- was found that the assymetrically substituted di-
tive PS may easily permit treatment overdosage. It sulfonic acids acted as the best PS (compared to
is at present uncertain whether it is better to have mono-, symmetrically di-, tri- and tetra-substituted
a PS ‘‘tailored’’ to a specific indication and to have sulfonic acids) in both the zinc [25] and aluminum
families or portfolios of PS for various diseases or [26] series of PC derivatives.
patient types, or to seek one PS that works against Another broad class of potential PS includes com-
most diseases. Lastly a desirable feature might be pletely synthetic, non-naturally-occurring, conju-
to have an inbuilt method of PS dosimetry monitor- gated pyrrolic ring systems. These comprise such
ing and following response to treatment by measur- structures as texaphyrins [23], porphycenes [27],
ing in vivo fluorescence and its loss by photobleach- and sapphyrins [28]. A last class of compounds
ing. that have been studied as PS are non-tetrapyrrole-
The majority of PS used both clinically and ex- derived naturally occurring or synthetic dyes. Ex-
perimentally, are derived from the tetrapyrrole amples of the first group are hypericin (from St
aromatic nucleus found in many naturally occur- Johns wort) [29] and from the second group are
ring pigments such as heme chlorophyll and bac- toluidine blue O [30] and Rose Bengal [31]. As yet
teriochlorophyll. Tetrapyrroles usually have a rela- these compounds have perhaps been more often
tively large absorption band in the region of 400 nm studied as agents to mediate antimicrobial photoin-
known as the Soret band, and a set of progres- activation [32] rather than as PS designed to kill
sively smaller absorption bands as the spectrum mammalian cells for applications such as cancer.
moves into the red wavelengths known as the Q- Examples of PS from some of these structural
bands. Naturally occurring porphyrins are fully con- groups that have either received regulatory ap-
jugated (non-reduced) tetrapyrroles and vary in the proval or are undergoing clinical or advanced pre-
number and type of side groups particularly car- clinical testing are given in Fig. 1. The tetrapyrrole
boxylic acid groups (uroporphyrin has eight, copro- nucleus frequently holds a co-coordinated metal
porphyrin has four and protoporphyrin has two). atom, but it has been found that only diamag-
Porphyrins have the longest wavelength absorption netic metals such as (Zn, Pd, In, Sn, Lu) allow the
band in the region of 630-nm and it tends to be tetrapyrrole to retain its photosensitizing ability,
small. Chlorins are tetrapyrroles with the double while paramagnetic metals such as (Fe, Cu, Gd) do
bond in one pyrrole ring reduced. This means that not [33]. Many of these compounds are lipophilic
the longest wavelength absorption band shifts to and some are even insoluble in water. These com-
the region of 650—690 nm and increases several- pounds must either be delivered in an emulsion or
fold in height; both these factors are highly de- else incorporated in liposomes. Although much re-
sirable for PDT. Bacteriochlorins have two pyrrole search has been undertaken in this area, the struc-
rings with reduced double bonds, and this leads to tural features of the molecule necessary to make
the absorption band shifting even further into the the ideal PS, i.e. with both a high selectivity for
red, and increasing further in magnitude. Bacteri- tumors or other target lesions, and a high level of
ochlorins may turn out to be even more effective phototoxicity towards cells and tissue in the illumi-
PS than chlorins, but with relatively few candidate nated area, are still unknown.
molecules and some questions about the stability
of these molecules upon storage this remains to be
seen. There are a set of classical chemical deriva- Photophysics
tives generally obtained from naturally occurring
porphyrins and chlorins that include such structures Fig. 2 graphically illustrates the processes of light
as purpurins, pheophorbides, pyropheophorbides, absorption and energy transfer that are at the heart
pheophytins and phorbins some of which have been of PDT. The ground state PS has two electrons with
studied (a few extensively) as PS for PDT. opposite spins (this is known as singlet state) in
A second widely studied structural group of PS the low energy molecular orbital. Following the ab-
is the phthalocyanines (PC), and to a lesser extent, sorption of light (photons), one of these electrons
their related cousins the naphthalocyanines. Again is boosted into a high-energy orbital but keeps its
their longest absorption band is in > 650 nm and usu- spin (first excited singlet state). This is a short-lived
Mechanisms in photodynamic therapy 283

Figure 1 Structures of PS either clinically approved or in trials.

Figure 2 Graphical illustration of the photophysical and photochemical mechanisms of PDT.

284 A.P. Castano et al.

(nanoseconds) species and can lose its energy by gen peroxide (H2 O2 ) into the hydroxyl radical (HO• ).
emitting light (fluorescence) or by internal conver- This reaction is called the Fenton reaction, and was
sion into heat. The fact that most PS are fluores- discovered over a hundred years ago. It is impor-
cent has led to the development of sensitive assays tant in biological systems because most cells have
to quantify the amount of PS in cells or tissues, and some level of iron, copper, or other metals, which
allows in vivo fluorescence imaging in living animals can catalyze this reaction. The reduced metal (fer-
or patients to measure the pharmacokinetics and rous iron or Fe2+ ) then catalyzes the breaking of
distribution of the PS. The excited singlet state PS the oxygen—oxygen bond of hydrogen peroxide to
may also undergo the process known as intersys- produce a hydroxyl radical (HO• ) and a hydroxide
tem crossing whereby the spin of the excited elec- ion (HO ). Superoxide can react with the hydroxyl
tron inverts to form the relatively long-lived (mi- radical (HO• ) to form singlet oxygen, or with ni-
croseconds) excited triplet-state that has electron tric oxide (NO− ) (also a radical) to produce perox-
spins parallel. The long lifetime of the PS triplet- ynitrite (OONO− ), another highly reactive oxidizing
state is explained by the fact that the loss of energy molecule.
by emission of light (phosphorescence) is a ‘‘spin- Like H2 O2 , HO• passes easily through membranes
forbidden’’ process as the PS would move directly and cannot be kept out of cells. Hydroxyl radi-
from a triplet to a singlet-state. cal damage is ‘‘diffusion rate-limited’’. This highly
reactive radical can add to an organic (carbon-
containing) substrate (represented by R below),
Photochemistry this could be, for example, a fatty acid which would
form a hydroxylated adduct that is itself a radical.
The PS excited triplet can undergo two kinds of The hydroxyl radical can also oxidize the organic
reactions (Fig. 2). Firstly, in a Type 1 reaction, it substrate by ‘‘stealing’’ or abstracting an electron
can react directly with a substrate, such as the cell from it. The resulting oxidized substrate is again it-
membrane or a molecule, and transfer a proton or self a radical, and can react with other molecules in
an electron to form a radical anion or radical cation, a chain reaction. For example, it could react with
respectively. These radicals may further react with ground-state oxygen to produce a peroxyl radical
oxygen to produce reactive oxygen species. Alter- (ROO• ). The peroxyl radical again is highly reac-
natively in a Type 2 reaction, the triplet PS can tive, and can react with another organic substrate
transfer its energy directly to molecular oxygen (it- in a chain reaction. This type of chain reaction is
self a triplet in the ground state), to form excited- common in the oxidative damage of fatty acids and
state singlet oxygen. Both Type 1 and Type 2 reac- other lipids, and demonstrates why radicals such as
tions can occur simultaneously, and the ratio be- the hydroxyl radical can cause so much more dam-
tween these processes depends on the type of PS age than one might have expected.
used, the concentrations of substrate and oxygen. These ROS, together with singlet oxygen pro-
Type 1 pathways frequently involve initial pro- duced via Type 2 pathway, are oxidizing agents that
duction of superoxide anion by electron transfer can directly react with many biological molecules.
from the triplet PS to molecular oxygen (monova- Amino acid residues in proteins are important tar-
lent reduction) [34,35]. Superoxide is not particu- gets that include cysteine, methionine, tyrosine,
larly reactive in biological systems and does not by histidine, and tryptophan [36,37]. Due to their re-
itself cause much oxidative damage, but can react activity, these amino acids are the primary target
with itself to produce hydrogen peroxide and oxy- of an oxidative attack on proteins. The reaction
gen, a reaction known as ‘‘dismutation’’ that can mechanisms are rather complex and as a rule lead
be catalyzed by the enzyme superoxide dismutase to a number of final products. Cysteine and me-
(SOD). Hydrogen peroxide is important in biologi- thionine are oxidized mainly to sulfoxides, histidine
cal systems because it can pass readily through cell yields a thermally unstable endoperoxide, trypto-
membranes and cannot be excluded from cells. Hy- phan reacts by a complicated mechanism to give
drogen peroxide is actually necessary for the func- N-formylkynurenine, tyrosine can undergo phenolic
tion of many enzymes, and thus is required (like oxidative coupling. Unsaturated lipids typically un-
oxygen itself) for health. Superoxide is also impor- dergo ene-type reactions to give lipid hydroperox-
tant in the production of the highly reactive hy- ides (LOOHs derived from phospholipids and choles-
droxyl radical (HO• ). In this process, superoxide ac- terol) [38—41].
tually acts as a reducing agent, not as an oxidizing DNA can be oxidatively damaged at both the nu-
agent. This is because superoxide donates one elec- cleic bases (the individual molecules that make up
tron to reduce the metal ions (such as ferric iron the genetic code) and at the sugars that link the
or Fe3+ ) that act as the catalyst to convert hydro- DNA strands by oxidation of the sugar linkages, or
Mechanisms in photodynamic therapy 285

cross-linking of DNA to protein (a form of damage eter necessary to define tissue optical properties is
particularly difficult for the cell to repair). Although the anisotropy factor that measures the direction of
all cells have some capability of repairing oxidative scattering of light. It is possible to use mathemat-
damage to proteins and DNA, excess damage can ical approaches such as diffusion theory or Monte
cause mutations or cell death. Of the four bases in Carlo modeling to predict how light will travel into
nucleic acids guanine is the most susceptible to ox- target tissue and the illumination parameters (flu-
idation by 1 O2 . The reaction mechanism has been ence, fluence rate, wavelength, angle of incidence)
extensively studied in connection with oxidative may then be adjusted to maximize the light dose.
cleavage of DNA [42]. The first step is a [4 + 2] cy- The combination of absorption of lower wavelength
cloaddition to the C-4 and C-8 carbons of the purine light by the important tissue chromophores (oxy and
ring leading to an unstable endoperoxide [43]. The deoxyhemoglobin and melanin) together with re-
subsequent complicated sequence of reactions and duced light scattering at longer wavelengths and
the final products depend on whether the guanine the occurrence of water absorption at wavelengths
moiety is bound in an oligonucleotide or a double- greater than 1300 nm has led to the concept of the
stranded DNA [44]. ‘‘optical window’’ in tissue (see Fig. 3). In terms
Because of the high reactivity and short half- of PDT the average effective penetration depth (in-
life of singlet oxygen and hydroxyl radicals, only tensity reduced to 37%) is about 1—3 mm at 630 nm,
molecules and structures that are proximal to the the wavelength used for clinical treatment with PF,
area of its production (areas of PS localization) are while penetration is approximately twice that at
directly affected by PDT. The half-life of singlet 700—850 nm [46,47].
oxygen in biological systems is < 40 ns, and, there- The increased penetration depth of longer wave-
fore, the radius of the action of singlet oxygen is of length light is a major incentive for the develop-
the order of 20 nm [45]. ment of PS absorbing at such wavelengths, and a
naphthalocyanine (776 nm) [48] and bacteriochlo-
rin (780 nm) [49] fall into this category. The ab-
Light delivery sorption of light by the PS itself can limit tis-
sue light penetration. This phenomenon has been
In PDT it is important to be able to predict the termed ‘‘self-shielding’’ and is particularly pro-
spatial distribution of light in the target tissue. nounced with PS that absorb very strongly at the
Light is either scattered or absorbed when it enters treatment wavelength [50]. Many PS are prone to
tissue and the extent of both processes depends photo-destruction during light exposure; a process
on tissue type and light wavelength. Tissue optics called ‘‘photobleaching’’ [51]. This thought to hap-
involves measuring the spatial/temporal distribu- pen when the singlet oxygen or other ROS produced
tion and the size distribution of tissue structures
and their absorption and scattering properties. This
is rather involved because the biological tissue is
inhomogeneous and the presence of microscopic
inhomogenities (macromolecules, cell organelles,
organized cell structure, interstitial layers, etc.)
makes it turbid. Multiple scattering within a turbid
medium leads to spreading of a light beam and loss
of directionality. Absorption is largely due to en-
dogenous tissue chromophores such as hemoglobin,
myoglobin and cytochromes. Complete character-
ization light transport in tissue is a formidable
task; therefore, heuristic approaches with differ-
ent levels of approximations have been developed
to model it. An effort for modeling light transport
also requires accurate values for the optical proper-
ties of the tissue. Scattering is generally the most
important factor in limiting light penetration into
most tissues and is measured by s (which for soft
tissues is in the range 100—1000 cm−1 ). Absorption Figure 3 Optical window in tissue. Absorption spectra
is usually of lesser importance and measured by a of important tissue chromophores such as water, oxy- and
(values in the range of 0.1—5 cm−1 for most tissue deoxyhemoglobin and melanin are plotted on a logarith-
at green and longer wavelengths). The third param- mic scale.
286 A.P. Castano et al.

upon illumination reacts with the PS molecule itself PS also tend to have the greatest uptakes into cells
in a manner that reduces its efficiency for further in vitro, especially when present in relatively low
photosensitization processes. PS of different chem- concentrations in the medium (<1 ␮M). Those PS
ical structures have widely varying photobleaching which are less hydrophobic and have >2 negative
rates and in some cases (particularly that of PPIX) charges tend to be too polar to diffuse across the
the first product of photobleaching is actually a plasma membrane, and are therefore taken up by
better PS than the starting molecule. Nevertheless endocytosis.
photobleaching usually means loss of PDT reactiv- Some PS distribute very broadly in vari-
ity but this may still have beneficial effects regard- ous intracellular membranes. An example is
ing treatment differential. These are based on the pyropheophorbide—–a methyl ester that was re-
following considerations: there exists a threshold ported to be localized in endoplasmic reticulum,
PDT dose to produce tissue necrosis [48]. If pho- Golgi apparatus, lysosomes and mitochondria, in
tobleaching occurs (which does not have such a NCI-h446 cells [55].
threshold) before this threshold is achieved, no tis-
sue damage is incurred. This is desirable for normal
tissue exposed to therapeutic light but not for the Lysosomes
tumor tissue to be treated. Thus, the net result is
that one can achieve greater depth of tumor necro- In 1993 lysosomes were proposed to be a critical in-
sis while sparing the normal skin. tracellular target for localization of PS [56]. How-
ever, succeeding studies [57] have found that al-
though lysosomally localized PS can lead to cell
killing upon illumination, the relative efficacy is sig-
Subcellular localization nificantly lower than that seen with PS localized in
mitochondria and other organelles [58]. This may
PS uptake by cancer or other cells is crucial for ef- be due to the tendency of PS with greater degrees
fective PDT. ROS have a short half-life and act close of aggregation to accumulate in lysosomes.
to their site of generation, therefore to a certain Woodburn et al. [59] studied intracellular local-
degree the type of photodamage that occurs in cells ization, in V79 Chinese hamster lung fibroblasts and
loaded with a PS and illuminated depends on the C6 glioma cells, of a series of porphyrins derived
precise subcellular localization of the PS within the from HP and PPIX with side chains chemically mod-
cell. The understanding of PS localization principles ified to give hydrophobic and anionic or cationic
is therefore important for choosing the most effec- residues at physiological pH. Compounds were se-
tive PS for each application. Confocal laser scan- lected to represent all combinations of these char-
ning fluorescence microscopy has made the deter- acteristics and it was found that those with a net
mination of intracellular location of PS much easier, cationic character localized in mitochondria, while
and gives more sensitivity and better spatial res- those with net anionic character localized in lyso-
olution than earlier non-confocal techniques. Co- somes. As the anionic porphyrins all carried two
localization of subcellular organelle specific probes negative charges, these results are in accord with
with differing fluorescence emission maxima to that previous work suggesting that sensitizers with a net
of the PS can be used to more closely identify the charge of −2 or greater accumulate in lysosomes.
site of localization [52] and these probes can also Nagata et al. [60] showed that the chlorin-based PS,
be used to identify sites of damage after illumina- ATX-S10 (Na) had a primary site of accumulation in
tion [53]. Fluorescence resonance energy transfer lysosomes but cells underwent apoptosis upon illu-
(FRET) [54] can also be used to determine intracel- mination doses leading to 70% cell death, suggesting
lular location of PS. that apoptotic pathways may be activated via mi-
Intracellular distributions in cultured cells have tochondrial destabilization following the damage of
been determined for a range of PS with widely dif- lysosomes by PDT.
fering structures. The important structural features A technique termed photochemical internaliza-
are (a) the net ionic charge which can range from tion was developed by Berg et al. [61,62]. This pro-
−4 (anionic) to +4 cationic, (b) the degree of hy- cedure relied on co-incubating cells with a macro-
drophobicity expressed as the logarithm of the oc- molecule that needed to be delivered into the cell
tanol/water partition coefficient, (c) the degree of cytoplasm and a PS such as aluminum phthalocya-
asymmetry present in the molecule. PS which are nine disulfonate (AlPCS2). Both of these molecules
hydrophobic and have two or less negative charges were incorporated into lysosomes with the PS local-
can diffuse across the plasma membrane, and then izing in the lysosomal membrane. On delivery of the
relocate to other intracellular membranes. These correct amount of light the lysosome was ruptured
Mechanisms in photodynamic therapy 287

by photochemical damage to its membrane thus Many workers have proposed to deliver PS to cells
releasing the intact macromolecule into the cyto- by targeting the PS molecules to a cell-type specific
plasm. This technology has been used to increase receptor. This is often accomplished by covalently
the intracellular delivery of genes [63], viruses [64], attaching one or more PS molecules to the specific
peptide nucleic acids [65], and ribosome inactivat- ligand of the receptor. When most of these recep-
ing proteins [66]. tors bind their natural ligand to which is covalently
The initial intracellular localization of PS in lyso- attached the PS, the entire construct is internalized
somes may redistribute due to photodynamic ac- into the cell and follows the endosomal-lysosomal
tion after only a small amount of light has been pathway discussed above. In many cases PS deliv-
delivered. It was found that exposure of cells pre- ered by this method gives the most specific lysoso-
incubated with anionic porphyrins, to light doses mal fluorescence seen in PDT. An example of this
that inactivated 20% of the cells resulted in relo- is given in Fig. 4A and B where the class A scav-
calization of the sensitizers from the lysosomes to enger receptors whose expression is specific to ma-
the cytoplasm in general, and, more specifically, ture macrophages is targeted by a conjugate be-
the nucleus [67,68]. This behavior was attributed tween maleylated albumin and the PS chlorin (e6)
to photodynamic permeabilization of the lysosomal delivered to the J774 mouse macrophage tumor cell
membrane, thus allowing small molecules, includ- line [69]. The maleylated albumin is endocytosed
ing the PS to leak out into the cytoplasm. after binding to its receptor and the PS ends up

Figure 4 Scanning laser confocal fluorescence microscope images of cells loaded with PS (red fluorescence) and
organelle specific green fluorescent probes. (A and B) J774 macrophage cells that have been incubated with a lysosomal-
targeted photosensitizer—protein conjugate with (A) lysotracker green and (B) mitotracker green. (C and D) J774 cells
that were incubated with BPD and (C) lysotracker green and (D) mitotracker green. Scale bar is 10 ␮m.
288 A.P. Castano et al.

specifically localized in lysosomes. There is good dioxolane)kryptocyanine, caused marked, light-

red and green overlap (orange) with the lysoso- dependent killing of human bladder, squamous, and
mal marker in Fig. 4A while the separation of red colon carcinoma cell lines after 30-min incubations
and the green mitochondrial marker is complete in at 1—0.01 ␮M but was minimally toxic to human
Fig. 4B. keratinocytes and to normal monkey kidney ep-
ithelial cells. Dummin et al. [74] prepared cationic
zinc (II) phthalocyanines with lipophilic side-chains
Mitochondria and showed they specifically accumulated in the
inner mitochondrial membranes. On irradiation of
Mitochondria have been found to be a very im- the incubated HeLa cells, the cristae were af-
portant subcellular target for many PS used in fected and finally completely destroyed. The respi-
PDT [70]. This is related to the tendency of many ration stopped and the energy metabolism was shut
PS to produce apoptosis by mitochondrial dam- down.
age after illumination (see Section 5). Benzopor- It was known previously that Pc4 localized in mi-
phyrin derivative (BPD) is one of the well-studied tochondria and Golgi complexes and ER [77]. At
mitochondrial-localized PS [71], however, cellular early times (0—1 h) after introduction of Pc 4 to
localization depends on cell type and BPD for- LY-R cells, the dye was found in the mitochondria,
mulation (free BPD, liposomal or encapsulated in lysosomes and Golgi apparatus, as well as other cy-
polycationic liposomes) used [71,72]. Some en- toplasmic membranes, but not in the plasma mem-
dothelial cells (ECV304) preferentially accumulated brane or the nucleus. Over the next 2 h, there was
BPD in perinuclear region, others (HUVEC) in cyto- some loss of Pc 4 from the lysosomes but an accumu-
plasm; polycationic liposomal BPD was mostly de- lation in the Golgi apparatus and the mitochondria.
posited in the mitochondria while free BPD was The exact binding site of Pc4 was discovered only
also found in perinuclear region [71,72]. Our re- recently. Pc 4-PDT photodamaged Bcl-2 and Bcl-xL,
sults with liposomal BPD are shown in Fig. 4C and D. antiapoptotic proteins interacting with the perme-
Co-incubation with green Lysotracker (Fig. 4C) ability transition pore complex that forms at con-
showed that BPD may localize in lysosomes to a tact sites between the inner and outer mitochon-
small extent but the majority was extralysosomal. drial membranes. These complexes and the inner
Co-incubation with Mitotracker (Fig. 4D) showed membrane are unique in containing the phospho-
that a part of the BPD was in mitochondria, but lipid cardiolipin. Nonyl-acridine orange (NAO) is a
that there was a significant amount of BPD that specific probe of cardiolipin and Morris et al. [54]
was in neither lysosomes nor mitochondria and showed evidence for fluorescence resonance energy
hence probably in the endoplasmic reticulum or transfer from NAO to Pc 4, defining a binding site
Golgi. for the photosensitizer.
Two meso-tetraphenylporphyrin derivatives be-
aring adjacent: 5,10-di[4-N-trimethylaminophe-
nyl]-15,20-diphenylporphyrin (DADP-a) or opposite: Plasma membrane
5,15-di[4-(N-trimethylaminophenyl)-10,20-diphen-
ylporphyrin (DADP-o) cationic-N-(CH3)3+ groups on Compounds that localize in plasma membranes of
two of the para-phenyl positions were compared cultured cells are relatively uncommon in the PDT
in study by Kessel et al. [73]. DADP-a localized field. Aveline and Redmond [78] used confocal fluo-
in mitochondria, while DADP-o (a much more rescence microscopy to show that deuteroporphyrin
symmetric molecule) localized in lysosomes, and IX (DP) and its monobromo and dibromo deriva-
led to extensive lysosomal photodamage after tives localized preferentially in the plasma mem-
irradiation. brane of L1210 cells. PF shows a dynamic dis-
PS with cationic charges and which are also hy- tribution in human carcinoma cells: the plasma
drophobic can localize in mitochondria [74]; this membranes are the main target sites of PF af-
is thought to be due to the influence of the mi- ter a brief (3 h) incubation, while the Golgi com-
tochondrial membrane potential as well as the plex is affected after prolonged (24 h) incubation
lipid bilayer of the membrane [75]. It is known [79]. The effects of PDT on cells with plasma
that carcinoma cell mitochondria preferentially ac- membrane-localized PF was found to be a cessa-
cumulate and retain certain cationic dyes to a tion of proliferation post PDT at Photofrin dose
much greater extent than most normal cells. Os- less than 7 ␮g/ml, and at higher dose (28 ␮g/ml)
eroff et al. [76] evaluated 10 rhodamine and cya- plasma membrane disruption and cell swelling were
nine dyes as carcinoma-specific mitochondrial PS observed immediately after PDT. Characteristics
in vitro. The most effective, N,N -bis(2-ethyl-1,3- of typical apoptosis such as phosphatidylserine
Mechanisms in photodynamic therapy 289

externalization and DNA fragmentation were not ALA-induced PPIX

detected in the cell death process caused by this
PDT regime. Recently there has been much interest in a different
approach to PDT where, instead of the PS being ad-
Golgi apparatus and endoplasmic reticulum ministered in a pre-synthesized form, a metabolic
precursor is administered and the PS is synthesized
Teiten et al. [80] studied Foscan subcellular lo- in situ in tumors or cells of the target tissue [81].
calization in the MCF-7 human adenocarcinoma The precursor is 5-aminolevulinic acid (ALA) which
cell line by means of confocal microscopy and interacts with the heme biosynthetic pathway as
microspectrofluorometry. The fluorescence topo- shown in Fig. 5. Almost all types of cells of the hu-
graphic profiles recorded after cells co-stained with man body, with the exception of mature red blood
Foscan and organelle-specific fluorescent probes cells, are equipped with this metabolic machinery.
revealed that Foscan presents low localization in In the first step of the pathway ALA is formed from
lysosomes and a weak accumulation in mitochon- glycine and succinyl CoA. The synthesis of ALA by
dria. However, the Foscan fluorescence topographic ALA-synthetase is under feedback regulation by the
profile turned out to co-localize perfectly with that amount of heme in the cell. The last step in the
obtained for the endoplasmic reticulum (ER) and pathway is incorporation of iron into PPIX catalyzed
the Golgi apparatus. The patterns of fluorescence by the enzyme ferrochelatase and this is rate-
derived from confocal microscopy studies were con- limiting. By adding exogenous ALA, the feedback
sistent with predominant localization of Foscan in inhibition is bypassed, and PPIX will accumulate be-
these organelles. Furthermore, evaluation of enzy- cause of the limited capacity of ferrochelatase to
matic activity of selected organelles immediately transform PPIX to heme. PPIX is formed in the mi-
after laser light irradiation (650 nm) indicated the tochondria of cells, but rapidly diffuses to other
Golgi apparatus and ER as the primary damaged intracellular membrane sites. Gaullier et al. [82]
sites resulting from Foscan-mediated PDT in the found early staining in mitochondria but at later
MCF-7 cell line. time points the plasma membrane showed strong

Figure 5 ALA-induced PPIX. Schematic illustrating the interaction of the heme biosynthesis pathway with exoge-
nous ALA to give intracellular PPIX. Abbreviations are ALA-D: ALA dehydratase; ALA-S: ALA synthetase; Coprogen III:
coproporphyrinogen III; CPO: coproporphyrinogen oxidase; FCH: ferrochelatase; HMB: hydroxymethylbilane, PBG-D:
porphobilinogren deaminase; protogen III: protoporphyrinogen; PPO: protoporphyrinogen oxidase; Urogen III: uropor-
phyrinogen III; UCS: uroporphyrinogen cosynthase, UGD: uroporphyrinogen decarboxylase.
290 A.P. Castano et al.

staining, and fluorescent spots (shown to be lyso- chosen sets of candidate PS molecules may allow
somes by co-localization experiments with lysoso- more or less general conclusions to be deduced
mal probes) were observed within the cytoplasm [93,94]. These experiments may have to be car-
especially in the perinuclear region. Fluorescence ried out both in vitro and in vivo. There have been
spectra suggested that the PPIX microenvironments instances when PS that were highly active against
were quite different at short and long incubation cancer cells in culture, were completely inactive
times. against tumors in animal models. There may be
In vivo the ALA may be administered orally [83], an analogy with the pharmaceutical industry that
intravenously [84], or topically [85]. The reasons has begun to develop computer software that can
why cancer cells tend to synthesize more PPIX than predict drug performance from chemical structure
normal cells, has been much investigated. Hypothe- and target identity using such benchmarks as the
ses include greater expression of heme biosynthesis druggability rules [95] and pharmacophore anal-
enzymes, porphobilinogen deaminase [86], copro- ysis [96]. It may not be too futuristic to expect
porphyrinogen oxidase [87], or reduced expression similar developments in the PDT field, especially
of ferrochelatase [88], but increased delivery of as more clinical approvals are obtained and more
ALA to the tumor may play a role especially in top- money becomes available for this type of research.
ical application [89]. Recent attempts to increase The growth of non-cancer applications of PDT to
the efficacy of ALA-mediated PDT include the use include ophthalmology, cardiovascular, immunology
of iron chelators to decrease the amount of PPIX and anti-infection.
converted to heme by ferrochelatase by removing
the free iron that is the necessary for the enzyme to
work [90]. Another approach is to administer ALA as Acknowledgements
one of various alkyl esters (methyl, pentyl, hexyl or
benzyl) in order to increase cellular uptake by mak-
Ana P. Castano was supported by U.S. Department
ing the molecule more lipophilic [91]. Since ALA is
of Defense CDMRP Breast Cancer Research Grant.
frequently applied topically to the skin, the ALA-
Tatiana N. Demidova was supported by a Wellman
methyl ester that penetrates through the skin’s nat-
Center of Photomedicine graduate student fellow-
ural permeability barrier much better than the po-
ship. Michael R. Hamblin was supported by the U.S.
lar ALA, recently received approval to treat basal
National Institutes of Health (R01-CA/AI838801).
cell skin cancers [92].
We are grateful to Anna Yaroslavsky and Tayyaba
Hasan for help and support.

Conclusion
References
The recent explosion in the synthesis and devel-
opment of novel PS has led to a ‘‘logjam’’ in the [1] Jesionek A, von Tappenier H. Zur behandlung der hautcarci-
investigation of the pros and cons of which com- nomit mit fluorescierenden stoffen. Muench Med Wochneshr
1903;47:2042.
pound should be used for which application. PDT [2] Hausman W. Die sensibilisierende wirkung des hematopor-
is an inherently complex technology that depends phyrins. Biochem Z 1911;30:276.
on multiple variables including the chemical and [3] Figge FH, Weiland GS, Manganiello LO. Affinity of neoplas-
photochemical properties of the PS, the PS dosage tic, embryonic and traumatized tissues for porphyrins and
and delivery vehicle, the drug—light time interval, metalloporphyrins. Proc Soc Exp Biol Med 1948;68:640.
[4] Lipson RL, Baldes EJ. The photodynamic properties of
the wavelength, energy dose, power density and a particular hematoporphyin derivative. Arch Dermatol
pulse structure of the light, and the oxygenation 1960;82:508.
state of the tissue. Therefore, it is unreasonable [5] Dougherty TJ. Activated dyes as antitumor agents. J Natl
to expect a comprehensive study of all these pa- Cancer Inst 1974;52:1333—6.
rameters for each of several hundred candidate [6] Dougherty TJ, Kaufman JE, Goldfarb A, Weishaupt KR, Boyle
D, Mittleman A. Photoradiation therapy for the treatment
molecules. The way around this otherwise insur- of malignant tumors. Cancer Res 1978;38:2628—35.
mountable problem is to carry out rational mecha- [7] Dougherty TJ, Lawrence G, Kaufman JH, Boyle D, Weishaupt
nistic studies that would allow predictions of effi- KR, Goldfarb A. Photoradiation in the treatment of recur-
cacy and preferred clinical application to be made rent breast carcinoma. J Natl Cancer Inst 1979;62:231—
based on the PS chemical structure and simple spec- 7.
[8] Baas P, van Mansom I, van Tinteren H, Stewart FA,
troscopic and photophysical measurements. There van Zandwijk N. Effect of N-acetylcysteine on Photofrin-
are signs that this is starting to happen. Well- induced skin photosensitivity in patients. Lasers Surg Med
designed structure—function studies on rationally 1995;16:359—67.
Mechanisms in photodynamic therapy 291

[9] Orenstein A, Kostenich G, Roitman L, et al. A comparative [29] Agostinis P, Vantieghem A, Merlevede W, de Witte PA. Hy-
study of tissue distribution and photodynamic therapy se- pericin in cancer treatment: more light on the way. Int J
lectivity of chlorin e6, Photofrin II and ALA-induced pro- Biochem Cell Biol 2002;34:221—41.
toporphyrin IX in a colon carcinoma model. Br J Cancer [30] Stockert JC, Juarranz A, Villanueva A, Canete M. Photody-
1996;73:937—44. namic damage to HeLa cell microtubules induced by thi-
[10] Spikes JD. Chlorins as photosensitizers in biology and azine dyes. Cancer Chemother Pharmacol 1996;39:167—
medicine. J Photochem Photobiol B 1990;6:259—74. 9.
[11] Kessel D, Thompson P. Purification and analysis of hemato- [31] Bottiroli G, Croce AC, Balzarini P, et al. Enzyme-assisted cell
porphyrin and hematoporphyrin derivative by gel exclusion photosensitization: a proposal for an efficient approach to
and reverse-phase chromatography. Photochem Photobiol tumor therapy and diagnosis. The rose bengal fluorogenic
1987;46:1023—5. substrate. Photochem Photobiol 1997;66:374—83.
[12] Boyle RW, Dolphin D. Structure and biodistribution rela- [32] Hamblin MR, Hasan T. Photodynamic therapy: a new antimi-
tionships of photodynamic sensitizers. Photochem Photobiol crobial approach to infectious disease? Photochem Photo-
1996;64:469—85. biol Sci 2004;3:436—50.
[13] Gomer CJ. Preclinical examination of first and second gen- [33] Rosenthal I, Murali Krishna C, Riesz P, Ben-Hur E. The role
eration photosensitizers used in photodynamic therapy. of molecular oxygen in the photodynamic effect of phthalo-
Photochem Photobiol 1991;54:1093—107. cyanines. Radiat Res 1986;107:136—42.
[14] Nyman ES, Hynninen PH. Research advances in the use of [34] Bilski P, Motten AG, Bilska M, Chignell CF. The photooxi-
tetrapyrrolic photosensitizers for photodynamic therapy. J dation of diethylhydroxylamine by rose bengal in micellar
Photochem Photobiol B 2004;73:1—28. and nonmicellar aqueous solutions. Photochem Photobiol
[15] Brancaleon L, Moseley H. Laser and non-laser light sources 1993;58:11—8.
for photodynamic therapy. Lasers Med Sci 2002;17:173— [35] Ma J, Jiang L. Photogeneration of singlet oxygen (1O2 ) and
86. free radicals (Sen•− , O2 •− ) by tetra-brominated hypocrellin
[16] Kessel D. Probing the structure of HPD by fluorescence spec- B derivative. Free Radic Res 2001;35:767—77.
troscopy. Photochem Photobiol 1989;50:345—50. [36] Grune T, Klotz LO, Gieche J, Rudeck M, Sies H. Pro-
[17] Kessel D, Thompson P, Musselman B, Chang CK. Probing the tein oxidation and proteolysis by the nonradical oxi-
structure and stability of the tumor-localizing derivative of dants singlet oxygen or peroxynitrite. Free Radic Biol Med
hematoporphyrin by reductive cleavage with LiAlH4. Cancer 2001;30:1243—53.
Res 1987;47:4642—5. [37] Midden WR, Dahl TA. Biological inactivation by singlet
[18] Kessel D, Thompson P, Musselman B, Chang CK. Chemistry oxygen: distinguishing O2(1 delta g) and O2(1 sigma g+).
of hematoporphyrin-derived photosensitizers. Photochem Biochim Biophys Acta 1992;1117:216—22.
Photobiol 1987;46:563—8. [38] Girotti AW. Mechanisms of lipid peroxidation. J Free Radic
[19] Kessel D. Photosensitization with derivatives of haemato- Biol Med 1985;1:87—95.
porphyrin [review]. Int J Radiat Biol Relat Stud Phys Chem [39] Girotti AW. Mechanisms of photosensitization. Photochem
Med 1986;49:901—7. Photobiol 1983;38:745—51.
[20] Kessel D. On the purity and definition of oligomeric HPD [40] Bachowski GJ, Korytowski W, Girotti AW. Characterization
formulations. J Photochem Photobiol B 1989;3:637—8. of lipid hydroperoxides generated by photodynamic treat-
[21] Kessel D. Components of hematoporphyrin deriva- ment of leukemia cells. Lipids 1994;29:449—59.
tives and their tumor-localizing capacity. Cancer Res [41] Bachowski GJ, Pintar TJ, Girotti AW. Photosensitized
1982;42:1703—6. lipid peroxidation and enzyme inactivation by membrane-
[22] Kessel D. Proposed structure of the tumor-localizing frac- bound merocyanine 540: reaction mechanisms in the ab-
tion of HPD (hematoporphyrin derivative). Photochem Pho- sence and presence of ascorbate. Photochem Photobiol
tobiol 1986;44:193—6. 1991;53:481—91.
[23] Detty MR, Gibson SL, Wagner SJ. Current clinical and pre- [42] Buchko GW, Wagner JR, Cadet J, Raoul S, Wein-
clinical photosensitizers for use in photodynamic therapy. J feld M. Methylene blue-mediated photooxidation of 7,8-
Med Chem 2004;47:3897—915. dihydro-8-oxo-2 -deoxyguanosine. Biochim Biophys Acta
[24] Allison RR, Downie GH, Cuenca R, Hu X-H, Childs CJ, Sibata 1995;1263:17—24.
CH. Photosensitizers in clinical PDT. Photodiag Photodynam [43] Buchko GW, Cadet J, Ravanat JL, Labataille P. Isolation and
Ther 2004;1:27—42. characterization of a new product produced by ionizing irra-
[25] Fingar VH, Wieman TJ, Karavolos PS, Doak KW, Ouellet R, diation and type I photosensitization of 2 -deoxyguanosine
van Lier JE. The effects of photodynamic therapy using in oxygen-saturated aqueous solution: (2S)-2,5 -ANHYDRO-
differently substituted zinc phthalocyanines on vessel con- 1-(2 -deoxy-beta-d-erythro-pentofuranosyl)-5-guanidin
striction, vessel leakage and tumor response. Photochem ylidene-2-hydroxy-4-oxoimidazolidine. Int J Radiat Biol
Photobiol 1993;58:251—8. 1993;63:669—76.
[26] Peng Q, Moan J, Nesland JM, Rimington C. Aluminum ph- [44] Ravanat JL, Cadet J. Reaction of singlet oxygen with
thalocyanines with asymmetrical lower sulfonation and 2 -deoxyguanosine and DNA. Isolation and characteriza-
with symmetrical higher sulfonation: a comparison of lo- tion of the main oxidation products. Chem Res Toxicol
calizing and photosensitizing mechanism in human tumor 1995;8:379—88.
LOX xenografts. Int J Cancer 1990;46:719—26. [45] Moan J, Berg K. The photodegradation of porphyrins in cells
[27] Szeimies RM, Karrer S, Abels C, et al. 9-Acetoxy-2,7,12,17- can be used to estimate the lifetime of singlet oxygen. Pho-
tetrakis-(beta-methoxyethyl)-porphycene (ATMPn), a novel tochem Photobiol 1991;53:549—53.
photosensitizer for photodynamic therapy: uptake kinet- [46] Svaasand LO. Optical dosimetry for direct and interstitial
ics and intracellular localization. J Photochem Photobiol B photoradiation therapy of malignant tumors. Prog Clin Biol
1996;34:67—72. Res 1984;170:91—114.
[28] Kral V, Davis J, Andrievsky A, et al. Synthesis and bi- [47] Wilson BC, Jeeves WP, Lowe DM. In vivo and post mortem
olocalization of water-soluble sapphyrins. J Med Chem measurements of the attenuation spectra of light in mam-
2002;45:1073—8. malian tissues. Photochem Photobiol 1985;42:153—62.
292 A.P. Castano et al.

[48] Firey PA, Rodgers MA. Photo-properties of a silicon naph- [67] Berg K, Madslien K, Bommer JC, Oftebro R, Winkelman
thalocyanine: a potential photosensitizer for photodynamic JW, Moan J. Light induced relocalization of sulfonated
therapy. Photochem Photobiol 1987;45:535—8. meso-tetraphenylporphines in NHIK 3025 cells and effects
[49] van Leengoed HL, Schuitmaker JJ, van der Veen N, Dubbel- of dose fractionation. Photochem Photobiol 1991;53:203—
man TM, Star WM. Fluorescence and photodynamic effects 10.
of bacteriochlorin a observed in vivo in ‘sandwich’ obser- [68] Peng Q, Farrants GW, Madslien K, et al. Subcellular localiza-
vation chambers. Br J Cancer 1993;67:898—903. tion, redistribution and photobleaching of sulfonated alu-
[50] Dougherty TJ, Potter WR. Of what value is a highly ab- minum phthalocyanines in a human melanoma cell line. Int
sorbing photosensitizer in PDT? J Photochem Photobiol B J Cancer 1991;49:290—5.
1991;8:223—5. [69] Hamblin MR, Miller JL, Ortel B. Scavenger-receptor
[51] Spikes JD, Bommer JC. Photobleaching of mono-l-aspartyl targeted photodynamic therapy. Photochem Photobiol
chlorin e6 (NPe6): a candidate sensitizer for the pho- 2000;72:533—40.
todynamic therapy of tumors. Photochem Photobiol [70] Morgan J, Oseroff AR. Mitochondria-based photodynamic
1993;58:346—50. anti-cancer therapy. Adv Drug Deliv Rev 2001;49:71—
[52] Wilson BC, Olivo M, Singh G. Subcellular localization of 86.
Photofrin and aminolevulinic acid and photodynamic cross- [71] Runnels JM, Chen N, Ortel B, Kato D, Hasan T. BPD-MA-
resistance in vitro in radiation-induced fibrosarcoma cells mediated photosensitization in vitro and in vivo: cellular
sensitive or resistant to Photofrin-mediated photodynamic adhesion and beta1 integrin expression in ovarian cancer
therapy. Photochem Photobiol 1997;65:166—76. cells. Br J Cancer 1999;80:946—53.
[53] Kessel D, Luo Y, Deng Y, Chang CK. The role of subcellu- [72] Takeuchi Y, Kurohane K, Ichikawa K, Yonezawa S, Nango M,
lar localization in initiation of apoptosis by photodynamic Oku N. Induction of intensive tumor suppression by antian-
therapy. Photochem Photobiol 1997;65:422—6. giogenic photodynamic therapy using polycation-modified
[54] Morris RL, Azizuddin K, Lam M, et al. Fluorescence liposomal photosensitizer. Cancer 2003;97:2027—34.
resonance energy transfer reveals a binding site of a [73] Kessel D, Luguya R, Vicente MG. Localization and photody-
photosensitizer for photodynamic therapy. Cancer Res namic efficacy of two cationic porphyrins varying in charge
2003;63:5194—7. distributions. Photochem Photobiol 2003;78:431—5.
[55] Sun X, Leung WN. Photodynamic therapy with [74] Dummin H, Cernay T, Zimmermann HW. Selective photosen-
pyropheophorbide-a methyl ester in human lung car- sitization of mitochondria in HeLa cells by cationic Zn (II)
cinoma cancer cell: efficacy, localization and apoptosis. phthalocyanines with lipophilic side-chains. J Photochem
Photochem Photobiol 2002;75:644—51. Photobiol B 1997;37:219—29.
[56] Geze M, Morliere P, Maziere JC, Smith KM, Santus R. [75] Rashid F, Horobin RW. Interaction of molecular probes with
Lysosomes, a key target of hydrophobic photosensitizers living cells and tissues. Part 2. A structure-activity analysis
proposed for photochemotherapeutic applications. J Pho- of mitochondrial staining by cationic probes, and a discus-
tochem Photobiol B 1993;20:23—35. sion of the synergistic nature of image-based and biochem-
[57] Berg K, Moan J. Lysosomes as photochemical targets. Int J ical approaches. Histochemistry 1990;94:303—8.
Cancer 1994;59:814—22. [76] Oseroff AR, Ohuoha D, Ara G, McAuliffe D, Foley J, Cin-
[58] MacDonald IJ, Morgan J, Bellnier DA, et al. Subcellu- cotta L. Intramitochondrial dyes allow selective in vitro
lar localization patterns and their relationship to photo- photolysis of carcinoma cells. Proc Natl Acad Sci USA
dynamic activity of pyropheophorbide-a derivatives. Pho- 1986;83:9729—33.
tochem Photobiol 1999;70:789—97. [77] Trivedi NS, Wang HW, Nieminen AL, Oleinick NL, Izatt JA.
[59] Woodburn KW, Vardaxis NJ, Hill JS, Kaye AH, Phillips Quantitative analysis of Pc 4 localization in mouse lym-
DR. Subcellular localization of porphyrins using con- phoma (LY-R) cells via double-label confocal fluorescence
focal laser scanning microscopy. Photochem Photobiol microscopy. Photochem Photobiol 2000;71:634—9.
1991;54:725—32. [78] Aveline BM, Redmond RW. Can cellular phototoxicity be ac-
[60] Nagata S, Obana A, Gohto Y, Nakajima S. Necrotic and apop- curately predicted on the basis of sensitizer photophysics?
totic cell death of human malignant melanoma cells follow- Photochem Photobiol 1999;69:306—16.
ing photodynamic therapy using an amphiphilic photosensi- [79] Hsieh YJ, Wu CC, Chang CJ, Yu JS. Subcellular localization
tizer ATX-S10(Na). Lasers Surg Med 2003;33:64—70. of Photofrin determines the death phenotype of human epi-
[61] Berg K, Selbo PK, Prasmickaite L, Hogset A. Photochemical dermoid carcinoma A431 cells triggered by photodynamic
drug and gene delivery. Curr Opin Mol Ther 2004;6:279—87. therapy: when plasma membranes are the main targets. J
[62] Berg K, Prasmickaite L, Selbo PK, Hellum M, Bonsted A, Cell Physiol 2003;194:363—75.
Hogset A. Photochemical internalization (PCI)—–a novel [80] Teiten MH, Bezdetnaya L, Morliere P, Santus R, Guillemin F.
technology for release of macromolecules from endocytic Endoplasmic reticulum and Golgi apparatus are the prefer-
vesicles. Oftalmologia 2003;56:67—71. ential sites of Foscan localisation in cultured tumour cells.
[63] Prasmickaite L, Hogset A, Berg K. Evaluation of different Br J Cancer 2003;88:146—52.
photosensitizers for use in photochemical gene transfec- [81] Peng Q, Warloe T, Berg K, et al. 5-Aminolevulinic acid-based
tion. Photochem Photobiol 2001;73:388—95. photodynamic therapy. Clinical research and future chal-
[64] Hogset A, Ovstebo Engesaeter B, Prasmickaite L, Berg K, lenges. Cancer 1997;79:2282—308.
Fodstad O, Maelandsmo GM. Light-induced adenovirus gene [82] Gaullier JM, Geze M, Santus R, et al. Subcellular localization
transfer, an efficient and specific gene delivery technology of and photosensitization by protoporphyrin IX human ker-
for cancer gene therapy. Cancer Gene Ther 2002;9:365—71. atinocytes and fibroblasts cultivated with 5-aminolevulinic
[65] Folini M, Berg K, Millo E, et al. Photochemical internaliza- acid. Photochem Photobiol 1995;62:114—22.
tion of a peptide nucleic acid targeting the catalytic subunit [83] van den Boogert J, van Hillegersberg R, de Rooij FW, et al.
of human telomerase. Cancer Res 2003;63:3490—4. 5-Aminolaevulinic acid-induced protoporphyrin IX accumu-
[66] Selbo PK, Sandvig K, Kirveliene V, Berg K. Release of gelonin lation in tissues: pharmacokinetics after oral or intravenous
from endosomes and lysosomes to cytosol by photochemical administration. J Photochem Photobiol B 1998;44:29—
internalization. Biochim Biophys Acta 2000;1475:307—13. 38.
Mechanisms in photodynamic therapy 293

[84] Svanberg K, Liu DL, Wang I, Andersson-Engels S, Stenram U, ing hydroxypyridinone iron-chelating agents. Br J Cancer
Svanberg S. Photodynamic therapy using intravenous delta- 1998;78:1278—82.
aminolaevulinic acid-induced protoporphyrin IX sensitisa- [91] Kloek J, Akkermans W, Beijersbergen van Henegouwen GM.
tion in experimental hepatic tumours in rats. Br J Cancer Derivatives of 5-aminolevulinic acid for photodynamic ther-
1996;74:1526—33. apy: enzymatic conversion into protoporphyrin. Photochem
[85] Calzavara-Pinton PG. Repetitive photodynamic therapy Photobiol 1998;67:150—4.
with topical delta-aminolaevulinic acid as an appropri- [92] Morton CA. Methyl aminolevulinate (Metvix) photody-
ate approach to the routine treatment of superficial namic therapy - practical pearls. J Dermatolog Treat
non-melanoma skin tumours. J Photochem Photobiol B 2003;14(Suppl 3):23—6.
1995;29:53—7. [93] Graham A, Li G, Chen Y, et al. Structure-activity relation-
[86] Gibson SL, Cupriks DJ, Havens JJ, Nguyen ML, Hilf R. A ship of new octaethylporphyrin-based benzochlorins as pho-
regulatory role for porphobilinogen deaminase (PBGD) in tosensitizers for photodynamic therapy. Photochem Photo-
delta-aminolaevulinic acid (delta-ALA)-induced photosen- biol 2003;77:561—6.
sitization? Br J Cancer 1998;77:235—43. [94] Potter WR, Henderson BW, Bellnier DA, et al. Parabolic
[87] Ortel B, Chen N, Brissette J, Dotto GP, Maytin E, Hasan T. quantitative structure-activity relationships and photody-
Differentiation-specific increase in ALA-induced protopor- namic therapy: application of a three-compartment model
phyrin IX accumulation in primary mouse keratinocytes. Br with clearance to the in vivo quantitative structure-
J Cancer 1998;77:1744—51. activity relationships of a congeneric series of py-
[88] Van Hillegersberg R, Van den Berg JW, Kort WJ, Terpstra ropheophorbide derivatives used as photosensitizers for
OT, Wilson JH. Selective accumulation of endogenously pro- photodynamic therapy. Photochem Photobiol 1999;70:781—
duced porphyrins in a liver metastasis model in rats. Gas- 8.
troenterology 1992;103:647—51. [95] Schuffenhauer A, Popov M, Schopfer U, Acklin P, Stanek J,
[89] Szeimies RM, Sassy T, Landthaler M. Penetration potency Jacoby E. Molecular diversity management strategies for
of topical applied delta-aminolevulinic acid for photody- building and enhancement of diverse and focused lead dis-
namic therapy of basal cell carcinoma. Photochem Photo- covery compound screening collections. Comb Chem High
biol 1994;59:73—6. Throughput Screen 2004;7:771—81.
[90] Curnow A, McIlroy BW, Postle-Hacon MJ, Porter JB, Mac- [96] Sirois S, Wei DQ, Du Q, Chou KC. Virtual screening for SARS-
Robert AJ, Bown SG. Enhancement of 5-aminolaevulinic CoV protease based on KZ7088 pharmacophore points. J
acid-induced photodynamic therapy in normal rat colon us- Chem Inf Comput Sci 2004;44:1111—22.