Far-red-absorbing photosensitizes their use in the photodynamic therapy of tumours

Giulio Jori Department of Biology, University of Padova, Ha Tneste 75, 1-35121 Padova (Italy)

Abstract

Light in the 60>1000 nm spectral region is scattered to a relatively small extent by most mammalian tissues and 
is poorly absorbed by endogenous chromophores such as melanin, cytochromes and haemoglobin. As a 
consequence, red light possesses a high penetration power into human tissues and can be selectively absorbed 
by photosensitizing agents (e.g. porphyrins, chlorins, phthalocyanines, naphthalocyanines) localized in 
predetermined sites of the organism. Recently developed procedures allow for the specific loading of tumour 
tissues by several red-light-absorbing photosensitizers; this property is the basis of a novel phototherapeutic 
modality for the treatment of a variety of solid tumours. The efficacy of the light plus photosensitizes 
combination in inducing tumour regression (the technique is often defined as "photodynamic therapy") is 
dependent on the photophysical properties of the photosensitizes and its affinity for malignant tissues.

1. Introduction

The steadily increasing knowledge of the optical properties of mammalian tissues (for a recent review, see ref. 
1) has introduced new possibilities for the utilization of visible light for phototherapeutic purposes. It is now 
established that light in the 600-1000 nm spectral region (the "phototherapeutic window") possesses maximum 
penetration power into most human tissues owing to the low absorptivity of the normal cell constituents in this 
region and the relatively inefficient scattering of red light by cell organelles [2]. The penetration depth (defined 
as the depth in the tissue at which the intensity of incident light is reduced to 1/e) can be as large as 2 cm for 
lightly pigmented tissues irradiated with light above about 700 nm, and it reaches approximately 1.5 cm at 900 
nm for heavily pigmented tissues, such as nevi or melanotic melanomas [1]. As a consequence, if a given tissue 
is loaded with an exogenously introduced redlight-absorbing photosensitized its selective photodamage can be 
achieved with a minimal risk of non-specific photoinduced alterations in adjacent tissues not containing the 
photosensitized
The recently developed photodynamic therapy (PDT) of tumours takes advantage of the ability of some 
porphyrins to be accumulated in significant amounts and to be retained for prolonged periods of time by tumour 
tissues; tumour necrosis can thus be obtained by irradiation of the neoplastic area with light of approximately 
630 nm, i.e. corresponding to the longest wavelength absorption band of porphyrins [3]. Until now, a few 
thousand patients have been treated by PDT worldwide, and a large majority of them have responded in an 
objectively favourable way to the treatment. Therefore, PDT is now in an advanced experimental phase in 
several centres, especially
for the endoscopic irradiation of lung, bronchial and oesophageal tumours, and for the photosterilization of the 
tumour bed after surgical resection of large neoplasias [4]. However, there is now general agreement that the 
scope and potential of PDT can be substantially improved only if new photosensitizers are found with 
photophysical and tumour-localizing properties markedly superior to those typical of the presently used 
porphyrin, i.e. haematoporphyrin derivative (HpD) or its partially purified form known under the commercial 
name of Photofrin II [4].

2. Limitations of Photofrin II as a phototherapeutic agent for tumours

Some physicochemical, photosensitizing and pharmacokinetic properties of Photofrin II are summarized in 
Table 1. An overall examination of the data demonstrates the following main limitations of this porphyrin as a 
tumour phototherapeutic agent.
(i) The high degree of chemical heterogeneity makes it difficult to control its behaviour in vivo, in terms of both 
the photosensitizing properties and the distribution between the tumour and the normal tissues. Oligomeric 
components of Photofrin II exhibit the greatest hydrophobicity, and hence ability to cross cell membranes, but 
they have short-lived triplet states and are inefficient generators of the cytotoxic species singlet oxygen [6].
(ii) The stability of Photofrin II in tissues is questionable. Ester bonds linking haematoporphyrin units are 
susceptible to esterase-catalysed hydrolysis in body fluids;
moreover, the non-covalent aggregates can be split into monomers in subcellular loci having a low dielectric 
constant, such as the lipid regions of the cytoplasmic, mitochondrial and Iysosomal membranes [7]. These 
processes would lead to a population of porphyrins characterized by widely different photosensitizing activities.
(iii) The small extinction coefficient of Photofrin II around 630 nm lowers the probability of its 
photoexcitation, thus requiring the administration of relatively large amounts of dye (2-5 mg kg-l body weight) 
in order to obtain a satisfactory phototherapeutic response. Since some components of Photofrin II are retained 
in amounts as large as several micrograms per gram of tissue by liver and spleen [7], the possible onset of 
long-term toxic effects must be taken into consideration.
(iv) The selectivity of tumour loading by Photofrin II is particularly large in brain owing to the inability of 
porphyrin compounds to cross the blood-brain barrier; however, this parameter may be very low in other cases, 
such as in skin tumours. This would obviously increase the risk of undesired photodamage in healthy tissues, 
especially when PDT is applied for the treatment of infiltrating neoplasias.
For these reasons it appears essential to develop new tumour photosensitizers in order to overcome the present 
limitations of Photofrin II.

3. Second generation tumour photosensitizers: ground state properties

The investigations to identify better tumour photosensitizers are mainly directed towards the synthesis of 
porphyrin analogue whose absorption properties in the farred spectral region are enhanced through a suitable 
manipulation of the electronic conjugation. Thus it is well known that chlorins (which differ from porphyrins 
by the partial hydrogenation of one pyrrole ring) exhibit an intense (extinction coefficient above 105 M-l cm-l) 
absorption band in the 650-700 nm interval. Bacteriochlorins display a significant absorption around 780 nm, 
which might be particularly useful for the PDT of pigmented tumours [8].
Other approaches are based on the extension of the delocalization of the 7r electron cloud either through the 
insertion of conjugated double bonds into the ring system [9], leading to the so-called 227r or 347r porphyrins, 
or by the condensation of benzene or naphthalene rings with the pyrrole moieties, such as occurs with 
phthalocyanine and naphthalocyanine derivathes. The absorption properties of these compounds are given in 
Table 2.
In general, the structural modifications described above increase the hydrophobic character of the 
photosensitizers, and hence their potential affinity for tumours [11]. However, the drop in water solubility may 
cause serious problems for the direct injection of these porphyrinoids into the bloodstream. For this reason, 
polar substituents, such as sulphonate, carboxylate or hydroxyl, are often introduced into the peripheral 
positions of the macrocycle. The relationship betveen the chemical structure and the tumour-localizing activity 
of photosensitizers has been investigated in detail using cultured cells [11] and different experimental tumours 
[5,11]. It appears that optimal tumour-localizing efficiency is imparted to be porphyrin-type macrocycle by the 
presence of two substituents in adjacent rings, yielding an amphiphilic molecule which retains a hydrophobic 
matrix to facilitate the interaction with cells [12, 13]. However, the most frequently used synthetic procedures 
generally yield mixtures of monosubstituted and polysubstituted derivatives, as well as positional isomers; the 
isolation of a specific derivative is very laborious. Alternatively, chemically pure unsubstituted compounds can 
be injected in vivo by incorporating the photosensitizer into suitable carriers,
including liposomes, oil emulsions or cyclodextrin inclusion complexes [14]. Under these conditions, the apolar 
environment provided by the carrier interferes with the intermolecular forces between the flat aromatic moieties 
of the photosensitizer which favour the formation of aggregated species; hence the photosensitizes molecules 
are present in a purely monomeric state up to at least millimolar endoliposomal concentrations [15]; it is 
known that the monomeric porphyrinoids have an appreciably higher photosensitizing efficiency than the 
corresponding oligomeric dyes.
The large extinction coefficients in the 680-800 nm range, which are typical of most second generation 
photosensitizes allow for the injection of significantly lower doses with excellent photoresponses of irradiated 
tumours. Thus the eradication of different experimental tumours has been obtained [16, 17] on red-light 
irradiation after administration of phthalocyanine or naphthalocyanine doses as low as 0.2-0.5 mg kg- l
Lastly, phthalocyanines and naphthalocyanines have the additional advantage of showing an appreciable light 
absorption only in the near-UV and red spectral regions, being essentially transparent between 400 and 650 nm. 
Thus these dyes are substantially less efficient in inducing skin photosensitivity to sunlight compared with the 
porphyrins, which absorb all visible light wavelengths.

4. Second generation tumour photosensitizers: excited state properties

The mechanism(s) by which red-light-absorbing photosensitizers cause tumour necrosis are not completely 
understood. Several lines of evidence suggest that singlet oxygen, generated by energy transfer from triplet 
photosensitizer to ground state dioygen, plays a major role [3]. However, the possibility of type I 
photosensitization pathways, involving the generation of radical species via electron transfer between the 
photoexcited sensitizer and suitable substrates, cannot be ruled out. The formation of singlet oxygen requires a 
minimum energy level of 22.5 kcal mold for the triplet photosensitized i.e. the energy of the 14 state of oxygen, 
unless endoergonic energy transfer takes place. This energy corresponds to a wavelength of 1270 nm, as 
deduced from the luminescence decay of 102 [18]. However, in selecting the longest wavelength absorption 
maximum of a dye, we must take into account the energy gap between the lowest excited singlet and triplet 
states.
This problem has been systematically addressed by Rodgers and coworkers [17-19] who analysed the triplet 
state properties and 102 generation efficiency of several phthalocyanines and naphthalocyanines with 
absorption maxima between 750 and 850 nm. According to these workers, a triplet state energy of at least 26 
kcal mold is necessary for irreversible energy transfer to oxygen and formation of 102. The bimolecular rate 
constant for the energy transfer reaction is close to one-ninth of the diffusioncontrolled encounter rate constant 
in the given medium. On the other hand, silicon(IV)naphthalocyanine, whose triplet energy is about 0.5 kcal 
mold below the lAg oxygen level, can generate 102 in a reversible process. This does not prevent silicon(IV)-
naphthalocyanine from having a good tumour-photosensitizing efficiency in vivo [17]. The situation can be less 
favourable for dyes with even lower triplet energies; thus 347r porphyrin, whose longest wavelength absorption 
maximum is located at 970 nm, appears to be devoid of any detectable photoactivity in vivo [9]. As a rule of 
thumb, for predicting the ability of a photosensitizes to generate 102 on the basis of its 0-0 band, an energy gap 
between the lowest excited singlet and triplet states of around 15 kcal mold is required [5].
The quantum yields for triplet state and 102 generation by selected potential second generation tumour 
photosensitizers are shown in Table 3. The data reported in the table refer to monomeric dyes either dissolved 
in homogeneous organic solutions or embedded in organized assemblies. Aggregation of the dyes, such as often 
occurs in aqueous media, almost invariably causes a shortening of the triplet lifetime and a drastic reduction of 
the overall photosensitizing efficiency [4].
In this connection, the presence of axial ligands to the centrally coordinated metal ion is often advantageous, 
since it generates some degree of steric hindrance to intermolecular aggregation without hopefully impairing 
the photophysical properties of the dye. As shown in Table 3, the presence of zinc(^) and silicon(IV) ions, both 
of which can give hexacoordination through d2sp3 hybrid orbitals, guarantees a satisfactory yield of 102 
generation; however, the photosensitizing activity is quenched by the insertion of copper(II) ions; this is true of 
all transition metal ions with a d-electron configuration which have unpaired electrons [21]. The presence of 
peripheral substituents (Table 3) has only a minor effect on the photoproperties of phthalocyanines and 
naphthalocyanines, apart from limiting the tendency to undergo aggregation to some extent. Thus recent 
investigations [22-24] have emphasized the optimal photosensitizing properties in vitro and phototherapeutic 
efficiency in vivo of zinc(II)-phthalocyanine
bearing two pyridyl or N-methyl-pyrrolidone axial ligands and silicon(IV)-naphthalocyanine coordinated with 
two alkyl-syloxy groups. The nature of the axial ligands can also be used to modulate the degree of 
hydrophilicity or lipophilicity of the photosensitized and hence to control its biodistribution between the various 
compartments of cells and tissues. For example, a derivative of silicon(IV)-naphthalocyanine axially ligated 
with two polyethylene glycol chains has been synthesized [19]. The water solubility of the naphthalocyanine 
increases with increasing length of the glycol chain and the derivative bearing more than 1000 monomeric units 
per chain can be directly injected into the bloodstream in a homogeneous aqueous solution without the need for 
specific delivery systems, such as liposomes or oil emulsions.
Therefore, the presently emerging trend is focused on the development and in vitro and in vivo testing of 
metalloporphyrinoids devoid of peripheral substituents (which might cause serious problems for the isolation of 
a single compound) and complexed with different axial ligands.

5. Approaches for selective delivery of photosensitizers to tumour tissue

The optimal utilization of the PDT of tumours can be obtained by taking full advantage of the two possibilities 
for selective tumour damage: (i) selectivity of photoexcitation, through the choice of irradiation wavelengths 
which are specifically absorbed by the photosensitizer compared with other constituents of the tumour tissue; 
(ii) selectivity of tumour localization by the photosensitized in particular with regard to the tumour-adjacent 
healthy tissues. Several investigations have aimed to exploit differences which exist between normal and 
tumour cells in order to improve the selectivity of tumour targeting. The most promising approaches use 
tumour-oriented carriers of the injected photosensitized
Both porphyrins and chlorins have been covalently coupled to monoclonal antibodies directed against antigens 
specifically present at the surface of neoplastic cells [25, 26]. Antibody-bound dyes fully retain their 
photophysical properties, including a good efficiency of 102 generation, and display a much larger selectivity of 
tumour targeting compared with the corresponding unbound photosensitizers. Such specificity of tumour 
targeting disappears if the antibody-photosensitizer complexes are administered to animals having transplanted 
tumours which lack the required antigens. This approach deserves further exploration, although some problems 
may arise due to the relatively small number of photosensitizer molecules which can be associated with the 
antibody without impairing its ability to interact with the antigens in the malignant cells.
In addition, since the antibody (and the bound photosensitizer) will remain outside the cell, we must assess the 
efficiency with which 102 generated in close proximity, but externally, to the cell can attack endocellular 
targets.
A different approach takes advantage of the fact that malignant cells possess a significantly larger number of 
surface receptors for low-density lipoproteins (LDLs) [27], which are natural carriers of systemically injected 
hydrophobic dyes in the serum [28]. LDLs and the photosensitizer possibly embedded in the lipid moiety of the 
protein are taken into malignant cells through a receptor-mediated endocytotic process; the recognition of the 
LDL molecule by the receptor involves some amino acid residues of apoprotein E, and hence there is no 
interference from the lipid-bound dye. Several hundred molecules of porphyrinoid compounds can be 
incorporated into the LDL molecules. The dye taken into the cell via an endocytotic process is released f29] 
inside the tumour cell and, on photoexcitation, mainly destroys the cell membrane.
The evidence supporting a major role of LDLs as tumour-specific carriers of injected photosensitizers has 
recently been reviewed [30]. Thus it is essential to devise photosensitizer administration procedures which 
favour its preferential interaction with LDLs rather than with other serum proteins or other components of the 
lipoprotein class. Studies performed in our laboratory indicate that the pre-incorporation of the dye into 
unilamellar liposomes prepared from phospholipids, which are in a quasisolid state at the physiological 
temperature of 37  C (eg. dipalmitoylphosphatidylcholine), leads to about 25%-30% association of a variety of 
phthalocyanines and naphthalocyanines with LDLs; this percentage can be further increased by adding a 
suitable concentration of cholesterol to the phospholipid bilayer [20] owing to the large affinity of LDLs for 
circulating cholesterol.
While the full potential of these strategies for tumour targeting by photosensitizers has not yet been defined, it 
is probable that a markedly better control of the in vivo behaviour of photosensitizing drugs will be developed 
and refined in the next few years, thus allowing for a more widespread application of PDT for tumour 
eradication or palliation and the treatment of a variety of other diseases [10].

Acknowledgment

This work received financial support from Consiglio Nazionale delle Ricerche (Italy) under the special project 
"Tecnologie Elettroottiche", contract No. 90.00227.PF65.

References

1 s. C. Wilson, M. S. Patterson, S. T. Flock and D. R. Wyman, in B. Chance (ed), Photon
Migration in Tissue, Plenum, New York, 1990, p. 25.
2 W. M. Star, J. P. Marijnissen and M. J. C. van Gemert, J. Photochem. Photobiol. B: Biol.,
I (1987) 149.
3 T. J. Dougherty, Photochem. Photobiol., 45 (1987) 879.
4 H. Kostron, D. Bellnier, C. W. Lin, M. R. Schwartz and R. Martuza, J. Neurosurg., 64 (1986)
768.
5 C. Zhou, J. Photochem. Photobiol. B: Biol., 3 (1989) 299.
6 D. Kessel, Int. J. Padiat. Biol., 49 (1986) 901.
7 D. Kessel, Photochem. Photobiol., 44 (1986) 489.
8 C. F. Borland, D. J. McGarvey, T. G. Truscott, R. J. Cogdell and E. J. Land, J. Photochem.
Photobiol. B: Biol., I (1987) 93.
9 B. Franck, G. Fulling, M. Gosmann, G. Knubel, H. Mertes and D. Schroder, in T. Hasan
(ed.), Advances in Photochemotherapy, Proc. SPIE, 1987, Society of Photo-Optical Instrumentation Engineers, 
Washington, 1988, p. 115. 10 J. D. Spikes and G. Jori, Lasers Med Sci., 2 (1987) 3. 11 T. J. Dougherty, W. R. 
Potter and K. R. Weishaupt, Adv. Exp. Med. Biol., 170 (1984) 301. 12 J. E. van Lier, N. Brasseur, B. Paquette, 
H. Amli, R. Langlois and J. Rousseau, in G.
Moreno, R. Pottier and T. G. Truscott (eds.), Photosensit~ation: Molecular, Cellular and Medical Aspects, 
Springer Verlag, Berlin, 1988, p. 435. 13 H. Barr, C. J. Tralau, P. B. Boulos, A. J. McRobert, M. R. Lewis, D. 
Phillips and S. G.
Bown, Gastroenterology, 9S (1990) 1532.
14 G. M. Garbo and A. R. Morgan, J. Photochem. Photobiol. B: Biol., I (1988) 494.
15 G. Valduga, S. Nonell, E. Reddi, G. Jori and S. sraslavsky, Photochem. Photobiols, 48 (1985)
1. 16 E. Reddi, C. Zhou, R. siolo, E. Menegaldo and G. Jori, Br. J. Cancer, Cl (1990) 407.
17 V. Cuomo, G. Jori, B. Rihter, M. E. Kenney and M. A. J. Rodgers, Br. J. Cancer, 62 (1990) 966.
18 M. A. J. Rodgers, in R. V. Bensasson, G. Jori, E. J. Land and T. G. Truscott (eds.), Primary Photoprocesses 
in Biology and Medicine, Plenum, New York, 1985, p. 1.
19 W. E. Ford, B. D. Rihter, M. E. Kenney and M. A. J. Rodgers, Photochem. Photobiol., 50 (1989) 277.
20 G. Jori, Lasers Med. Sci, 5 (1990) 115.
21 M. Gouterman, in D. Dolphin (ed.), The Porphyrins, Vol. 111, Academic Press, New York, 1978, p. 1.
22 P. A. Firey, W. E. Ford, J. R. Sounik, M. E. Kenney and M. A. J. Rodgers, l. Am. Chem. Soc., 110 (1988) 
7626.
23 B. D. Rihter, M. E. Kenney, W. E. Ford and M. A. J. Rodgers, J. Am. Chem. Soc., 95 (1991) 598.
24 G. Jori, in G. Bock and S. Hamett (eds.), Photosensitizing Compounds: Their Chemistry, Biology and 
Clinical Use, Wiley, Chichester, 1989, p. 78.
25 A. M. Richter, B. Kelly, J. Chow, D. J. Lui, G. H. N. Towers, D. Dolphin and J. G. Levy, J. Natl. Cancer 
Inst., 79 (1987) 1327.
26 T. Hasan, C. W. Lin and A. Lin, Immun. Cancer, 2 (1989) 471.
27 J. L. Goldstein, R. G. W. Anderson and M. S. Brown, Nature, 279 (1979) 679.
28 J. P. Reyftmann, P. Morliere, S. Goldstein, R. Santus, L. Dubertret and D. Lagrange, Photochem. Photobiol., 
40 (1984) 721.
29 J. C. Maziere, R. Santus, P. Morliere, J. P. Reyftmann, C. Candide, L. Mora, S. Salmon, C. Maziere, S. Gatt 
and L. Dubertret, J. Photochem. Photobiol. B: Biol., 6 (1990) 61.
30 G. Jori and E. Reddi, in D. Kessel (ed.), Photodynamic Therapy of Neoplastic Disease, Vol. 11, CRC Press, 
Boca Raton, 1990, p. 117.