In vitro demonstration of the heavy-atom effect for photodynamic therapy

Article abstract of DOI:10.1021/ja047649e

Photodynamic therapy (PDT) is an emerging treatment modality for a range of disease classes, both cancerous and noncancerous. This has brought about an active pursuit of new PDT agents that can be optimized for the unique set of photophysical characterist

Full text of DOI:10.1021/ja047649e

Published on Web 08/10/2004
In Vitro Demonstration of the Heavy-Atom Effect for
Photodynamic Therapy
Aoife Gorman,† John Killoran,† Caroline O’Shea,‡ Tony Kenna,‡
William M. Gallagher,‡ and Donal F. O’Shea*,†
Contribution from the Centre for Synthesis and Chemical Biology, Conway Institute of
Biomolecular and Biomedical Research, Department of Chemistry and Department of
Pharmacology, UniVersity College Dublin, Belfield, Dublin 4, Ireland
Received April 23, 2004; Revised Manuscript Received June 11, 2004; E-mail: donal.f.oshea@ucd.ie
Abstract: Photodynamic therapy (PDT) is an emerging treatment modality for a range of disease classes,
both cancerous and noncancerous. This has brought about an active pursuit of new PDT agents that can
be optimized for the unique set of photophysical characteristics that are required for a successful clinical
agent. We now describe a totally new class of PDT agent, the BF2-chelated 3,5-diaryl-1H-pyrrol-2-yl-3,5-
diarylpyrrol-2-ylideneamines (tetraarylazadipyrromethenes). Optimized synthetic procedures have been
developed to facilitate the generation of an array of specifically substituted derivatives to demonstrate how
control of key therapeutic parameters such as wavelength of maximum absorbance and singlet-oxygen
generation can be achieved. Photosensitizer absorption maxima can be varied within the body’s therapeutic
window between 650 and 700 nm, with high extinction coefficients ranging from 75 000 to 85 000 M-1
cm-1. Photosensitizer singlet-oxygen generation level was modulated by the exploitation of the heavy-
atom effect. An array of photosensitizers with and without bromine atom substituents gave rise to a series
of compounds with varying singlet-oxygen generation profiles. X-ray structural evidence indicates that the
substitution of the bromine atoms has not caused a planarity distortion of the photosensitizer. Comparative
singlet-oxygen production levels of each photosensitizer versus two standards demonstrated a modulating
effect on singlet-oxygen generation depending upon substituent patterns about the photosensitizer. Confocal
laser scanning microscopy imaging of 18a in HeLa cervical carcinoma cells proved that the photosensitizer
was exclusively localized to the cellular cytoplasm. In vitro light-induced toxicity assays in HeLa cervical
carcinoma and MRC5-SV40 transformed fibroblast cancer cell lines confirmed that the heavy-atom effect
is viable in a live cellular system and that it can be exploited to modulate assay efficacy. Direct comparison
of the efficacy of the photosensitizers 18b and 19b, which only differ in molecular structure by the presence
of two bromine atoms, illustrated an increase in efficacy of more than a 1000-fold in both cell lines. All
photosensitizers have very low to nondeterminable dark toxicity in our assay system.
Introduction
resulting in oxidative cellular damage, leading to cell death via
apoptosis and/or necrosis.2 The photosensitizer returns to its
Photodynamic therapy (PDT) is a noninvasive technique for
the treatment of a variety of cancer tumor types by the combined
use of visible or near-visible light with a photosensitizing drug.1
The therapeutic strategy is as follows: a photosensitizer of
minimal dark toxicity is introduced into the body, which
accumulates to some extent preferentially within the tumor. The
tumor is then selectivity irradiated with low-energy light of
wavelength that can pass through the body’s therapeutic window
(650-900 nm), resulting in excitation of the photosensitizer.
The light-activated photosensitizer then transfers its excited-
state energy to surrounding biological tissue via singlet oxygen,
ground state and the process can repeat as long as the light
source remains and molecular oxygen is present. After light
treatment, the photosensitizer is allowed to clear from the body
and the treatment is complete. Due to the short half-life (0.6 ×
10-6 s) and diffusion distance (0.1 µm) of singlet oxygen in
aqueous media, PDT can be considered a highly selective form
of cancer treatment, as only the irradiated areas are affected,
provided that the photosensitizer is nontoxic in the absence of
light.3 This combination of light/photosensitizer/oxygen as a
mode of disease treatment has expanded from an initial focus
on cancer tumors to include application in certain nonneoplastic
diseases including age-related macular degeneration (AMD),4
coronary heart disease,5 periodontal diseases,6 and microbial
infections.7
† Department of Chemistry.
‡ Department of Pharmacology.
(1) For recent reviews, see: (a) Dolmans, D. E. J. G. J.; Fukumura, D.; Jain,
R. K. Nature ReViews Cancer 2003, 3, 380. (b) MacDonald, I. J.; Dougherty,
T. J. J. Porphyrins Phthalocyanines 2001, 5, 105. (c) Sharman, W. M.;
Allen, C. M.; van Lier, J. E. Drug DiscoVery Today 1999, 4, 507. (d)
Dougherty, T. J.; Gomer, C. J.; Henderson, B. W.; Jori, G.; Kessel, D.;
Korbelik, M.; Moan, J.; Peng, Q. J. Natl. Cancer Inst. 1998, 90, 889. (e)
Bonnett, R. Chem. Soc. ReV. 1995, 24, 19.
(2) Oleinick, N. L.; Morris, R. L.; Belichenko, I. Photochem. Photobiol. Sci.
2002, 1, 1.
(3) Henderson, B. W.; Dougherty, T. J. Photochem. Photobiol. 1992, 55, 145.
9
10.1021/ja047649e CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 10619-10631
10619

A R T I C L E S
Gorman et al.
amples from the porphyrin class are protoporphyrin IX 2, which
therapeutically is administered as its biosynthetic precursor,
5-aminolevulinic acid, for the treatment of Barrett’s esophagus
and skin cancers.9 The benzoporphyrin derivative 3 (Visudyne)
is widely used for the treatment of AMD.4 Examples from the
dihydroporphyrin or chlorin class are m-tetrahydroxyphenyl-
chlorin 4 (Foscan), which has gained regulatory approval in
Europe for the treatment of head and neck cancer,10 and tin
etiopurpurin (Purlytin) 5, which is in a clinical trial for the
treatment of AMD.11 The palladium-bacteriopheophorbide 6
(TOOKAD),12 the expanded porphyrin 7 (Motexafin Lute-
tium),13 and the phthalocyanines14 8 are all in various stages of
clinical trials for a number of cancerous tissue targets.
The investigation of non-porphyrin photosensitizers for the
development of novel PDT agents has been considerably less
extensive.15 To date, the focus has been primarily on the cationic
photosensitizers such as methylene blue 9,16 nile blue analogue
10,17 and the chalcogenopyrylium class of sensitizer 1118 (Figure
2). Significant drawbacks of these classes appear to be their
high inherent dark toxicity, perhaps due to the nuclear localiza-
tion of these cationic compounds, and their relatively short
absorption maxima of less than 650 nm in aqueous solu-
tions.16b,c,17a,18d Even though many of the nonporphyrinic PDT
agents investigated predate the development of second-genera-
tion photosensitizers, none have been approved for use in vivo.
The most successful has been methylene blue 9, which is used
ex vivo for the inactivation of extracellular enveloped viruses
in blood plasma.19
Despite its achievements, PDT remains in its developmental
stages with a marked need to progress alternative photosensi-
(9) (a) Overholt, B. F.; Panjehpour, M.; Halberg, D. L. Gastrointestinal
Endoscopy 2003, 58, 183. (b) Gossner, L.; Stolte, M.; Sroka, R.; Rick, K.;
May, A.; Hahn, E. G.; Ell, C. Gastroenterology 1998, 114, 448. (c) Barr,
H.; Shepherd, N. A.; Dix, A.; Roberts D. J. H.; Tan, W. C.; Krasner, N.
Lancet 1996, 348, 584.
Figure 1. Polypyrrole macrocyclic PDT agents.
(10) (a) Copper, M. P.; Bing Tan, I.; Oppelaar, H.; Ruevekamp, M. C.; Stewart,
F. A. Arch. Otolaryngol. 2003, 129, 709. (b) Fan, K. F. M.; Hopper, C.;
Speight, P. M.; Buonaccorsi, G. A.; Bown, S. G. Int. J. Cancer 1997, 73,
25.
Photofrin was the first, and at the present time remains the
most common, clinically used PDT agent. Although it has been
approved for use in the United States, Canada, Japan, and
Europe for the treatment of esophageal, endobroncheal, bladder,
lung, stomach, cervical, and skin cancers,1b it is widely
recognized that it is far from being an ideal drug for use in
PDT.8 It is an undefined mixture of dimeric and oligomeric
compounds derived from the acidic treatment of hematopor-
phyrin 1 followed by a semipurification (Figure 1). Over the
past 20 years attempts to develop alternatives to Photofrin have
predominantly focused on various types of polypyrrole macro-
cycles generally termed second-generation PDT agents. Ex-
(11) Mang, T. S.; Allison, R.; Hewson, G.; Snider, W.; Moskowitz, R. Cancer
J. Sci. Am. 1998, 4, 378.
(12) (a) Koudinova, N. V.; Pinthus, J. H.; Brandis, A.; Brenner, O.; Bendel, P.;
Ramon, J.; Eshhar, Z.; Scherz, A.; Salomon, Y. Int. J. Cancer 2003, 104,
782. (b) Chen, Q.; Huang, Z.; Luck, D.; Beckers, J.; Brun, P. H.; Wilson,
B. C.; Scherz, A.; Salomon, Y.; Hetzel, F. W. Photochem. Photobiol. 2002,
76, 438.
(13) Sessler, J. L.; Miller, R. A. Biochem. Pharmacol. 2000, 59, 733 (review).
(14) (a) Bench, B. A.; Beveridge, A.; Sharman, W. M.; Diebold, G. J.; van Lier
J. E.; Gorun, S. M. Angew. Chem., Int. Ed. 2002, 41, 748. (b) Ball, D. J.;
Mayhew, S.; Wood, S. R.; Griffiths, J.; Vernon, D. I.; Brown, S. B.
Photochem. Photobiol. 1999, 69, 390. (c) Hu, M.; Brasseur, N.; Yildiz, S.
Z.; van Lier, J. E.; Leznoff, C. C. J. Med. Chem. 1998, 41, 1789. (d)
Allemann, E.; Brasseur, N.; Kudrevich, S. V.; La Madeleine, C.; van Lier,
J. E. Int. J. Cancer 1997, 72, 289.
(15) Wainwright, M. Chem. Soc. ReV. 1996, 351 (review).
(4) Hooper, C. Y.; Guymer, R. H. Clin. Exp. Ophthalmol. 2003, 31, 376
(review).
(16) (a) Wainwright, M.; Crossley, K. B. J. Chemother. 2002, 14, 431 (review).
(b) Mellish, K. J.; Cox, R. D.; Vernon, D. I.; Griffiths, J.; Brown, S. B.
Photochem. Photobiol. 2002, 75, 392. (c) Wainwright, M.; Phoenix, D.
A.; Rice, L.; Burrow, S. M.; Waring, J. J. Photochem. Photobiol. B 1997,
40, 233. (d) Orth, K.; Ruck, A.; Stanescu, A.; Berger, H. G. Lancet 1995,
345, 519.
(5) (a) Kereiakes, D. J.; Szyniszewski, A. M.; Wahr, D.; Herrmann, H. C.;
Simon, D. I.; Rogers, C.; Kramer, P.; Shear, W.; Yeung, A. C.; Shunk, K.
A.; Chou, T. M.; Popma, J.; Fitzgerald, P.; Carroll, T. E.; Forer, D.;
Adelman, D. C. Circulation 2003, 108, 1310. (b) Hayase, M.; Woodbum,
K. W.; Perlroth, J.; Miller, R. A.; Baumgardner, W.; Yock, P. G.; Yeung,
A. CardioVascular Res. 2001, 49, 449. (c) Mansfield, R.; Bown, S.;
McEwan, J. Heart 2001, 86, 612 (review).
(17) (a) Tong, Z.; Singh, G.; Rainbow, A. J. Photochem. Photobiol. 2001, 74,
707. (b) Frimberger, A. E.; Moore, A. S.; Cincotta, L.; Cotter, S. M.; Foley,
J. R. Clinical Cancer Res. 1944, 4, 2207. (c) Cincotta, L.; Foley, J. W.;
Maceachern, T.; Lampros, E.; Cincotta, A. H. Cancer Res. 1994, 54, 1249.
(18) (a) Detty, M. R.; Merkel, P. B. J Am. Chem. Soc. 1990, 112, 3845. (b)
Leonard, K. A.; Nelen, M. I.; Anderson, L. T.; Gibson, S. L.; Hilf, R.;
Detty, M. R. J. Med. Chem. 1999, 42, 3942. (c) Leonard, K. A.; Hall, J.
P.; Nelen, M. I.; Davies, S. R.; Gollnick, S. O.; Camacho, S.; Oseroff, A.
R.; Gibson, S. L.; Hilf, R.; Detty, M. R. J. Med. Chem. 2000, 43, 4488. (d)
Brennan N. K.; Hall, J. P.; Davies, S. R.; Gollnick, S. O.; Oseroff, A. R.;
Gibson, S. L.; Hilf, R.; Detty, M. R. J. Med. Chem. 2002, 45, 5123.
(19) (a) Williamson, L. M.; Cardigan, R.; Prowse, C. V. Transfusion 2003, 43,
1322 (review). (b) Jockusch, S.; Lee, D.; Turro, N. J.; Leonard, E. F. Proc.
Natl. Acad. Sci. U.S.A. 1996, 93, 7446.
(6) (a) Ko¨merik, N.; Nakanishi, H.; MacRobert, A. J.; Henderson, B.; Speight,
P.; Wilson, M. Antimicrob. Agents Chemother. 2003, 47, 932. (b) Ko¨merik,
N.; Wilson, M. J. Appl. Microbiol. 2002, 92, 618. (c) Wilson, M. J. Appl.
Bacteriol. 1993, 75, 299.
(7) (a) Hamblin, M. R.; Zahra, T.; Contag, C. H.; McManus, A. T.; Hasan, T.
J. Infect. Diseases 2003, 187, 1717. (b) Romanova, N. A.; Brovko, L. Y.;
Moore, L.; Pometun, E.; Savitsky, A. P.; Ugarova, N. N.; Griffiths, M. W.
Appl. EnViron. Microbiol. 2003, 69, 6393. (c) Hamblin, M. R.; O’Donnell,
D. A.; Murthy, N.; Contag, C. H.; Hasan, T. Photochem. Photobiol. 2002,
75, 51.
(8) Ochsner, M. Drug Res. 1997, 47, 1185.
9
10620 J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004

Heavy-Atom Effect for Photodynamic Therapy
A R T I C L E S
Figure 3. Simplified Jablonski diagram for the PDT process.
would require different excitation wavelengths dependent upon
tissue location (deep-seated tumors versus skin lesions), size,
density, and pigmentation. The depth of light penetration through
tissue is directly related to wavelength;20 thus, if an array of
photosensitizers could be generated containing only minor
peripheral structural modifications, which had different maxima
of absorption, this would allow for the matching of a specific
PDT agent to tissue type.
As singlet oxygen is the key cytotoxic agent in the PDT
therapeutic process, it is essential that control can be exerted
over this component of the therapy. The quantity of singlet
oxygen generated by a photosensitizer is regulated by the
efficiency of a spin-forbidden electronic transition from a singlet
to a triplet state (ISC). The introduction of a heavy atom into a
molecule is known to have an influence over the rates of the
ISC and is termed the heavy-atom effect.21 The ability to
synthetically incorporate a heavy atom within the photosensitizer
could be a mechanism for exerting the desired control over levels
of singlet-oxygen generation. It would be of significant clinical
benefit if a range of photosensitizers were available with
programmed levels of singlet oxygen production capability that
only differed in chemical structure by minor peripheral modi-
fications to a single core photosensitizer. This could facilitate
the building of an array of varying singlet oxygen producers to
suit different therapeutic needs and profiles.
Herein we will show how we can exert control over several
key stages of a PDT process by the judicious positioning of
substituents around a common photosensitizer scaffold. The
goals of this work were to determine the peripheral substitution
patterns that would facilitate a stepwise bathochromatic shift
of absorption maximum from 650 to 700 nm and deliver
incremental controlled singlet oxygen production levels for an
array of photosensitizers all constructed from a common
scaffold. These compounds would then be tested in vitro to
determine if our controls would be viable in a living cellular
system.
Figure 2. Non-porphyrin PDT agents.
tizers with improved efficacy and side effect profiles. To further
advance this novel form of treatment, it has become apparent
that the development of new PDT compounds is needed,
together with a more thorough and integrated understanding of
the multitude of targets/actions so far ascribed to PDT. New
PDT agents have largely come from the class of polypyrrole
macrocycles. As a class of compound they have, to varying
degrees, many of the inherent photophysical properties that
would be required for a successful PDT agent, such as cancer
tissue localization, therapeutic window absorption, and capacity
for singlet-oxygen generation. The drawbacks associated with
this class of compound can be their challenging synthesis and
purification, which can restrict the fine-tuning of their photo-
physical and biological properties.
As a result of the rapidly increasing number of divergent
therapeutic targets ascribed to PDT, it has become imperative
that new classes of agents be explored. Our approach was to
develop a completely new structural platform from which a
sequentially modified array of PDT agents could be developed.
This would facilitate the refinement of the essential therapeutic
parameters of a PDT agent to match the specific therapeutic
target of choice by straightforward functional group manipula-
tion of a core photosensitizer scaffold. To achieve this goal it
would be essential to develop a robust synthetic route to our
scaffold that would enable us to incorporate specific functional
groups either before or after scaffold synthesis, which in turn
would allow for optimization of the photophysical and biological
properties necessary for a successful therapeutic agent.
PDT agents have an atypical mode by which their efficacy
can be optimized. Conventional therapeutics, which operate
through modes of action such as enzyme inhibition or receptor
interaction, are often discovered by a structure-based design
approach, centered upon molecular features such as 3-D binding,
electrostatic, hydrophobic, and hydrogen-bonding interactions
of the drug with the site of interaction. In contrast, the efficacy
of a PDT agent can be governed by three factors: (i) the
photosensitizer localization, (ii) the extent of light activation
of the photosensitizer in vivo, and (iii) the efficiency of an
intersystem crossing (ISC) from the first excited singlet state
to the triplet state of the photosensitizer (Figure 3).
To have two photochemical processes (absorption and ISC)
as pivotal controlling points determining a treatment efficacy
is completely unique to this mode of therapy. Thus, any
approach to influence the therapeutic effectiveness by exerting
direct control over these processes warrants investigation.
It would be advantageous to have control over absorbance
wavelength, as it could be envisaged that different tissue types
Figure 4. Azadipyrromethenes.
Our starting point was a neglected class of chromophore, the
azadipyrromethenes 12 (Figure 4). These compounds were first
described in the 1940s and have remained unstudied for any
application since that time.22 As compounds of type 12 are not
macrocyclic, we anticipated that it would be necessary to convert
(20) Wilson, B. C.; Patterson, M. S.; Lilge, L. Lasers Med. Sci. 1997, 12, 182
(review).
(21) Turro, N. J. In Modern Molecular Photochemistry; University Science
Books: Sausalito, CA, 1991; pp 191-195.
(22) Rogers, M. A. T. J. Chem. Soc. 1943, 596.
9
J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004 10621

A R T I C L E S
Gorman et al.
Scheme 1
Table 1. Reaction Optimization for 16a-d
entry compd
R1
R2
reagent/solvent
time/h
% yielda
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16a
16a
16a
16a
16a
16b
16b
16b
16b
16c
16c
16c
16d
16d
16d
H
H
H
H
H
H
H
H
H
OCH3
OCH3
OCH3
Br
H
H
H
H
H
NH4CO2H/none
NH4OAc/none
NH4CO2H/EtOH 24
NH4OAc/EtOH
NH4OAc/BuOH
1
1.5
34
47
33
24 (48) 35 (42)
24
1.5
1.5
24
24
0.5
48
48
1.5
39
33
45
24
47
46
23
48
15
OCH3 NH4CO2H/none
OCH3 NH4OAc/none
OCH3 NH4OAc/EtOH
OCH3 NH4OAc/BuOH
H
H
H
H
H
H
NH4OAc/none
NH4OAc/EtOH
NH4OAc/BuOH
NH4OAc/none
NH4OAc/EtOH
NH4OAc/BuOH
Br
Br
24 (48) 16 (22)
24 24
a Isolated purified yield.
discovered the reaction to be successful with an improved
isolated yield of 47% (entry 2). The same trend of improved
yields was also observed for 16b when substituting ammonium
acetate for ammonium formate under solventless conditions
(entries 6 and 7).
12 into a more structurally constrained architecture if they were
to be exploited for the development of PDT agents. This we
proposed could be achieved by conversion into their corre-
sponding BF2 chelates 13, which would prevent the isomeriza-
tion around the pyrrole carbon to bridging nitrogen bonds.23
We now report the synthesis, photophysical properties, test
for singlet oxygen production, cellular localization, and in vitro
light-induced cytotoxicity assays for this new class of PDT
agent. The majority of PDT agents investigated to date are based
on polypyrrole macrocycles from which it can be synthetically
difficult to generate an array of sequentially modified deriva-
tives. However, our proposed class of non-porphyrin sensitizers
would be a good starting point, as they are amenable to
modification around the periphery of the sensitizer, which could
allow for optimization of aspects of their photophysical and
therapeutic properties. This sensitizer class differs from previ-
ously investigated non-porphyrin-related photosensitizers, as
they are overall charge neutral compounds.
To replace the harsh solventless methodology and develop a
more synthesis-friendly reaction procedure, we investigated the
reaction in various solvents. Encouragingly, we found that the
reaction could be readily achieved by heating under reflux in
either ethanol or butanol for time periods of 24-48 h. The
reaction of 15a with ammonium acetate in ethanol under reflux
for 24 h gave a 35% isolated yield of 16a. This yield could be
increased to 42% by lengthening the reaction time to 48 h (Table
1, entry 4). Substituting butanol for ethanol as reaction solvent
gave an improved yield of 16a of 39% over a 24 h reaction
period (entry 5). A similar trend was observed for 16b-d in
that higher isolated yields were obtained in butanol than ethanol
(entries 9, 12, and 15). In general, the reaction gave higher yield
in butanol or the solventless conditions in comparison to ethanol.
As each method gave the desired product, some of the
advantages or drawbacks of each procedure are associated with
product purification. The solventless reactions require the
harshest conditions, but have an advantage of the shortest
reaction time. Upon completion of these reactions the product
is a solid mass in the reaction flask and requires silica gel
chromatography for purification. When ethanol or butanol was
used as the solvent, the product precipitated during the reaction
and could be isolated by simple filtration in each case. The
overall facile synthesis and purification of the tetraarylaza-
dipyrromethenes 16 makes this class of chromophore very
attractive as a starting point for the development of an array of
novel PDT therapeutics.
We next sought an efficient synthetic method to introduce a
heavy atom onto 16 and chose to exploit the high reactivity of
the pyrrole ring to electrophilic aromatic substitution to ac-
complish this. Bromination of both the unsubstituted â-pyrrole
positions of 16a-c was readily achieved by their reaction with
bromine at room temperature, yielding 17a-c in excellent yields
(Scheme 2). Conversion of 16a-d and 17a-c into their
corresponding BF2 chelates 18a-d and 19a-c was achieved
by stirring with BF3.etherate using diisopropylethylamine
(DIEA) as base in dichloromethane at room temperature for
24 h, giving isolated purified yields ranging from 62 to 90%
(Scheme 2). Purification of products was carried out by column
Results and Discussion
Synthesis and Purification. The starting point of our
synthesis was the diaryl R,â-unsaturated ketones (chalcones)
14a-d of which many are commercially available or are readily
made by an aldol/dehydration reaction of the corresponding
aldehyde and acetophenone (Scheme 1).24 The Michael addition
of nitromethane to the R,â-unsaturated ketones, with diethyl-
amine (DEA) as base, yields the 1,3-diaryl-4-nitrobutan-1-ones,
15a-d in high yields. These two classical organic synthesis
reactions provide the precursors to the tetraarylazadipyr-
romethenes very efficiently.
Our initial approach for the generation of tetraarylazadipyr-
romethenes 16a-d was based on the method first reported by
Rogers but following this we undertook a detailed study of the
reaction to refine the methodology.22 The previously reported
synthesis of tetraphenylazadipyrromethene 16a used a solvent-
less reaction at 180 °C between 15a and an excess of ammonium
formate (Scheme 1). We repeated these conditions as described
and for 16a achieved a comparable yield of 34% to that first
reported (Table 1, entry 1). We tested ammonium acetate as an
alternative ammonium source for the synthesis of 16a and
(23) (a) Killoran, J.; Allen, L.; Gallagher, J. F.; Gallagher, W. M.; O’Shea, D.
F. Chem. Commun. 2002, 1862. (b) O’Shea D.; Killoran J.; Gallagher, W.
PCT, WO 03/080627 A1.
(24) Wattanasin, S.; Murphy, W. S. Synthesis 1980, 647.
9
10622 J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004

Heavy-Atom Effect for Photodynamic Therapy
A R T I C L E S
Scheme 2
Figure 5. UV-visible spectra of 18a (green), 18b (blue), 18d (brown),
and 19b (red) in H2O/CrEL.
Table 2. Spectroscopic Absorbance Properties of 18a-d and
19a-ca
λmax/nm
fwhm/nm
H O/
H O/
ꢀb/
2
2
entry compd CrEL toluene ethanol chloroform CrEL chloroform
M
-1 cm-1
1
2
3
4
5
6
7
18a 658
18b 696
18c 671
18d 666
19a 651
19b 683
19c 655
655
693
666
662
652
683
655
647
686
660
655
645
675
646
650
688
664
658
650
679
653
53
57
57
51
57
67
66
49
55
57
47
47
57
57
79 000
85 000
78 000
83 000
79 000
75 000
80 000
chromatography on silica gel prior to spectroscopic measurement
and assaying.
An inherent advantage of our photosensitizer system is that
the synthesis utilizes classical high-yielding organic synthetic
reactions to generate their precursors, which can be easily
transformed into the tetraarylazadipyrromethenes 16 in one facile
step. Synthetic manipulation of 16 to facilitate the inclusion of
the heavy-atom bromine is readily achievable, and conversion
into their final chelates is routine. The compounds are polar
but nonionic and are soluble in organic solvents such as THF,
dichloromethane, or toluene.
Photosensitizer Formulation. To obtain aqueous solutions
of 18a-d and 19a-c for spectroscopic analysis and in vitro
assays, the emulsifier Cremophor EL (CrEL) was used. CrEL
is a nonionic surfactant, frequently used in vivo as a delivery
agent for poorly water soluble anticancer drugs such as
Paclitaxel (taxol).25 The formulation procedure required dis-
solving the photosensitizer in the minimum quantity of THF
and treating the solution with a mixture of CrEL and 1,2-
propanediol (10:3 v/v). This was sonicated for 1 h followed by
removal of the THF under vacuum. The resulting blue/green
oil was dissolved in phosphate-buffered saline (PBS) solution
followed by filtration through a 0.2 µm membrane filter. To
ensure the accuracy of the concentration, which may vary due
to partial precipitation of the photosensitizer, the exact concen-
tration was confirmed by UV-visible spectral analysis of the
solutions prior to assaying.
a Concentration 5 × 10-6 M, rt. b Chloroform.
substituents on the aryl rings. An aqueous solution of the
tetraphenyl-substituted “parent” molecule 18a has a λmax within
the therapeutic window spectral region at 658 nm (Figure 5,
Table 2). Introduction of an electron-donating methoxy group
onto the para position of the aryl rings R to the pyrrole nitrogen
in 18b (R2 ) OCH3) results in a significant bathochromic shift
of 38 nm (Figure 5, Table 2, entry 2). Interestingly, para-
substitution with a methoxy moiety on both of the aryl rings in
the â-pyrrole positions of 18c (R1 ) OCH3) had significantly
less impact, with a bathochromic shift of 13 nm (Table 2, entry
3). In comparison, 18d, which contains an electron-withdrawing
group (R1 ) Br) in the same position, resulted in a minor
bathochromic shift of 8 nm.
Comparisons of 18a-c with their corresponding dibrominated
derivatives 19a-c show that introduction of the heavy atoms
gives rise to a moderate hypsochromic shift, ranging from 7 to
16 nm. Remarkably, the substitution of bromines onto the
pyrrole rings results in only minor changes in the maxima or
the shape of the absorption bands. For example, a comparison
of 18a and 19a in aqueous solution show only a variance of
7 nm in their λmax values (Table 2, entries 1 and 5). This
demonstrates that the heavy atom can be introduced without
diminishing the advantageous absorption characteristics of our
photosensitizers. In general, the absorbance characteristics of
this class of photosensitizer are relatively insensitive to solvent
effects with small hypsochromic shifts (6-9 nm) observed in
the polar solvent ethanol when compared to a nonpolar solvent
such as toluene (Table 2). The spectral bandwidth of the
principal absorption band can be analyzed by the full width at
half-maximum (fwhm) values, which vary from 51 to 67 nm in
aqueous solutions. This sharp absorption band is indicative of
Photosensitizer Spectroscopic Properties. The absorption
spectra of 18a-d and 19a-c in aqueous formulated solution
all show a strong S0 f S1 transition with wavelength of
maximum absorbance varying between 651 and 696 nm,
depending upon substituents (Figure 5, Table 2). We have
observed a striking dependence of absorption maxima on
(25) Gelderblom, H.; Verweij, J.; Nooter, K.; Sparreboom, A. Eur. J. Cancer
2001, 37, 1590.
9
J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004 10623

A R T I C L E S
Gorman et al.
emission for 18a is 676 nm in toluene and 669 nm in ethanol
(Table 3, entry 1). The compounds 18a-d showed a range of
high fluorescence quantum yields (Φf) measured in chloroform
from 0.23 to 0.36 (Table 3). Interestingly, the quantum yield
for the bromoaryl-substituted derivative 18d did not show a
significant decrease in fluorescence quantum yield when
compared to 18a, but as the bromine substituents are distant
from the central core of the photosensitizer, it could be
anticipated that a significant heavy-atom effect may not be
observed. In comparison, the introduction of bromine directly
into the core of the photosensitizer gave rise in each case to
substantial reduction in fluorescence quantum yields for 19a-c
(Table 3, entries 5-7). This would indicate that when bromine
is directly substituted onto the central core of the photosensitizer,
a larger heavy-atom effect can be induced that, depending upon
other possible competing photophysical pathways, may translate
into increased singlet oxygen production.
Figure 6. Fluorescence spectra of 18a (green), 18b (blue), 18d (brown),
and 19b (red) in H2O/CrEL.
Heavy-Atom Effect and Singlet-Oxygen Generation. An
electronic transition from a singlet to a triplet excited state within
a molecule is a spin-forbidden process and as such occurs
inefficiently for many compounds. In order for a transition
between states of different spin multiplicities to occur ef-
fectively, a spin-orbit perturbation is generally required.27
Enhanced spin-orbit perturbations can be achieved by the
attachment of a heavy atom directly onto the molecule28 (internal
heavy-atom effect) or placing the molecule in a surrounding
environment containing heavy atoms29 (external heavy-atom
effect). We chose to exploit the internal heavy-atom approach,
but in an attempt to effect varying degrees of spin-orbit
coupling, we placed the heavy atoms at two different positions
in the molecule. This we anticipated would allow us to modulate
the degree of spin-orbit coupling and hence the efficiency of
singlet oxygen production by three modes: (i) no heavy atom
within the sensitizer, thereby relying on the inherent spin-orbit
coupling of the molecule; (ii) positioning of bromines on the
aryl rings (not directly on the sensitizer), which could give rise
to an intermediate singlet-oxygen generation level (termed an
intramolecular external heavy-atom effect);30 and (iii) position-
ing of the bromine directly onto the sensitizer, thereby fully
exploiting the internal heavy-atom effect and providing the most
efficient singlet-oxygen producers (Figure 7).31
The final step of the singlet-oxygen generation process is an
energy transfer from photosensitizer triplet state to ground-state
oxygen. An additional consequence of introducing heavy-atom
substituents can be to give rise to nonradiative internal back-
conversion to the ground state or inhibiting the photosensitizer
triplet to ground-state oxygen energy transfer. These competing
pathways would give rise to loss of the excited-state energy
without the generation of singlet oxygen. This makes the
position of the heavy-atom within the sensitizer critical, as the
atom(s) must be positioned to have a substantial effect on the
degree of spin-orbit coupling but not give rise to competing
excited-state energy loss pathways. As such a heavy atom can
Table 3. Spectroscopic Fluorescence Properties of 18a-d and
19a-ca
λ
maxb/nm
Stoke
shiftb/nm
H O/
CrEL
2
entry
compd
toluene
ethanol
chloroform
Φfc
1
2
3
4
5
6
7
18a
18b
18c
18d
19a
19b
19c
25
31
30
22
28
36
38
683
727
701
688
679
719
693
676
717
693
683
672
714
683
669
715
697
677
666
712
680
672
715
695
680
673
714
679
0.34
0.36
0.23
0.34
0.01
0.1
<0.01
a Concentration 2 × 10-7 M, rt. b H2O/CrEL. c Chloroform.
a nonaggregated system in these solutions, as the observed
values are comparable to those in chloroform, which vary from
47 to 57 nm (Table 2).
The extinction coefficients of this class of photosensitizer are
in the range of 75 000-85 000 M-1 cm-1, which is a significant
improvement over substituted porphyrins (3000-5000) or
chlorins (30 000-40 000) and is only matched in this region
of the absorption spectrum by phthalocyanine or cyanine dyes
(150 000-250 000).1c This relatively high extinction coefficient
within the therapeutic window is an advantage over other PDT
agents such as Photofrin or protoporphyrin IX 2, as the most
efficiently excited photosensitizers are those that have strong
absorption bands in this region of the spectrum. This strong
absorbance is one of the factors that can facilitate a high singlet-
oxygen generation when using therapeutic irradiance conditions
of wavelengths greater than 650 nm. Conversely, it is also of
some advantage that the extinction coefficient is not too high,
as this can limit light penetration in vivo due to the self-shielding
effect of a sensitizer.26
The fluorescence properties of the sensitizers were examined
in aqueous formulated solutions, toluene, ethanol, and chloro-
form. Excitation of the compounds 18a-d and 19a-c in
aqueous solutions at 630 nm all gave fluorescence bands that
were mirror images of the absorbance spectra with Stoke shifts
in the range of 22-38 nm (Figure 6, Table 3). Overall, the
emission spectra were relatively insensitive to solvent effects,
with small hypsochromic shifts observed in ethanol when
compared to toluene. For example, wavelength maximum of
(27) Lower, S. K.; El-Sayed, M. A. Chem. ReV. 1966, 66, 199.
(28) (a) Yuster, P.; Weissman, S. I. J. Chem. Phys. 1949, 17, 1182. (b) McClure,
D. S. J. Chem. Phys. 1949, 17, 905.
(29) (a) Koziar, J. C.; Cowan, D. O. Acc. Chem. Res. 1978, 334 (review). (b)
McGlynn, S. P.; Reynolds, M. J.; Daigre, G. W.; Christodoyleas, N. D. J.
Phys. Chem. 1962, 66, 2499. (c) Kasha, M. J. Chem. Phys. 1952, 20, 71.
(30) Chandra, A. K.; Turro, N. J.; Lyons, A. L., Jr.; Stone, P. J. Am. Chem.
Soc. 1978, 100, 4964.
(31) For this description, we regard the fused five-six-five-membered rings
(26) Dougherty, T. J.; Potter, W. R. J. Photochem. Photobiol. B 1991, 8, 233.
as the central sensitizer and the aryl rings as peripheral substituents.
9
10624 J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004

Heavy-Atom Effect for Photodynamic Therapy
A R T I C L E S
Figure 7. Positioning of the bromine heavy atoms around the core sensitizer
(blue).
promote an S1 f T1 ISC, though it may not necessarily result
in enhanced singlet-oxygen production. This effect has been
reported for the pyrylium class of PDT agents 11, which
following the replacement of the ring oxygen with the heavier
atom tellurium did not give rise to an increase in singlet-oxygen
quantum yields, even though a significant decrease in the
fluorescence quantum yields was observed (Figure 2).18b
Similarly, substitution of two pyrrole nitrogen atoms of meso-
tetraarylporphyrins for selenium atoms also gave reduced
singlet-oxygen quantum yields when compared to the parent
porphyrin.32 In both cases this has been explained by the fact
that the larger atomic radii of the heavy atoms led to structural
distortions of the pyrylium and porphyrin rings, giving rise to
loss of sensitizer planarity, which promoted nonradiative decay
to the ground state and as a result a decrease in singlet-oxygen
quantum yields. This effect may also be attributed to a
shortening of the triplet lifetime as a result of the internal heavy-
atom effect.21,29a Recently polyhalogenated meso-tetraarylpor-
phyrins and bacteriochlorins have been shown to modulate their
singlet-oxygen quantum yields in organic solutions, depending
upon the position and number of halo substituents.33
X-ray Structure of 19a. As the introduction of a heavy atoms
with large atomic radii into a photosensitizer can result in
structural deformation to the planarity of the molecule, we
sought to gain structural evidence for the effects, if any, on our
sensitizer structure imparted by the bromine atoms in 19a-c.
19a crystallized, by the slow room-temperature evaporation of
a toluene solution, in the monoclinic space group Cc (#9) with
four molecules in the unit cell. A thermal ellipsoid drawing of
19a is shown in Figure 8.
Figure 8. Perspective drawing of 19a. Thermal ellipsoids drawn at the
80% probability level.
Table 4. Selected Bond Lengths for 19a
bond
length/Å
bond
length/Å
C(1)-N(3)
C(1)-N(1)
C(1)-C(2)
C(2)-C(3)
C(3)-C(4)
C(3)-Br(1)
C(4)-N(1)
N(1)-B
1.322(3)
1.395(3)
1.439(3)
1.375(4)
1.429(4)
1.880(2)
1.355(3)
1.568(4)
1.372(4)
C(17)-N(3)
C(17)-N(2)
C(17)-C(18)
C(18)-C(19)
C(19)-C(20)
C(19)-Br(2)
C(20)-N(2)
N(2)-B
1.322(3)
1.393(3)
1.443(4)
1.382(4)
1.421(4)
1.870(2)
1.359(3)
1.564(3)
1.379(3)
B-F(1)
B-F(2)
molecule was preserved. The bromine bond angles from pyrrole
ring 1 (N1/C1/C2/C3/C4) are -0.126(1) Å for Br(1) and
0.109(1) Å for Br(2) from planarity and for pyrrole ring 2
(N2/C17/C18/C19/C20) Br(1) is -0.415(1) Å and Br(2) is 0.004
Å. The angle of intersection of pyrrole ring 1 and pyrrole ring
2 is small at 4.2(2)°. The 12-atom plane of the central tricyclic
structure shows the greatest deviation from this plane of 0.100(2)
Å for the atom N(1). This compares favorably with our pre-
viously reported structure of nonbrominated 18b, which has a
comparable intersect angle of pyrrole ring 1 and pyrrole ring 2
of 4.1(3)°.23a Additionally, the central tricyclic 12-atom plane
of 18b shows the greatest deviation from this plane of 0.087(3)
Å for the atom N(2). This lends weight to the supposition that
introduction of the bromines at the â-pyrrole position of 19a-c
could give rise to a more efficient population of the triplet state
without causing an increase in nonradiative decay to the ground
state.
The molecular structure confirmed the conjugated nature of
this class of sensitizer with comparable bond lengths for the
bridging nitrogen (N3) to both pyrrole rings (C1, C17) and for
the pyrrole nitrogen (N1, N2) to boron bond lengths. A
comparison of all bond lengths of both pyrrole rings of 19a
showed them to be equivalent (Table 4).
Despite the introduction of two bromine atoms onto the
sensitizer 19a, the planarity of the central 12-atom core of the
Comparative Study of Singlet-Oxygen Generation in
Solution with Light > 600 nm. To test our ability to gain
control over singlet-oxygen production by exploiting the heavy-
atom effect, we carried out a comparative singlet-oxygen
generation analysis. The study was undertaken by monitoring
the reaction of the known singlet oxygen acceptor 1,3-
diphenylisobenzofuran (DPBF) with photosensitizer-generated
singlet oxygen.34 This was achieved experimentally by following
the disappearance of the 410 nm absorbance band of DPBF at
(32) (a) You, Y.; Gibson, S. L.; Hilf, R.; Davies, S. R.; Oseroff, A. R.; Roy, I.;
Ohulchanskyy, T. Y.; Bergey, E. J.; Detty, M. R. J. Med. Chem. 2003, 46,
3743. (b) Hilmey, D. G.; Abe, M.; Nelen, M. I.; Stilts, C. E.; Baker, G. A.;
Baker, S. N.; Bright, F. V.; Davies, S. R.; Gollnick, S. O.; Oseroff, A. R.;
Gibson, S. L.; Hilf, R.; Detty, M. R. J. Med. Chem. 2002, 45, 449.
(33) (a) Pineiro, M.; Rocha Gonsalves, A. M. d’A.; Pereira, M. M.; Formosinho,
S. J.; Arnaut, L. G. J Phy. Chem. A 2002, 106, 3787. (b) Azenha, E. G.;
Serra, A. C.; Pineiro, M.; Pereira, M. M.; de Melo, J. S.; Arnaut, L. G.;
Formosinho, S. J.; Rocha Gonsalves, A. M. d’A. Chem. Phy. 2002, 280,
177.
9
J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004 10625

A R T I C L E S
Gorman et al.
Figure 9. Comparative singlet-oxygen generation of 18a (green),
18b (blue), 18c (yellow), 18d (brown), and hematoporphyrin 1 (black) at
5 × 10-6 M concentration.
Figure 10. Comparative singlet-oxygen generation of 19a (orange),
19b (red), 19c (white), and methylene blue 9 (black) at 5 × 10-8
M
concentration.
initial concentration of 5 × 10-5 M over a time period of 1 h.
We used a filtered light source of wavelength > 600 nm, which
would mimic those used for our in vitro assays or those used
in a clinical therapeutic setting. Each of the sensitizers 18a-d
was examined at a concentration of 5 × 10-6 M and compared
to hematoporphyrin 1 as a reference sensitizer. The brominated
pyrrole derivatives 19a-c were examined at a lower concentra-
tion of 5 × 10-8 M and compared to the reference sensitizer
methylene blue 9. Relative rates of oxygenation of DPBF by
18a-d and 19a-c versus the standards 1 and 9 were estimated
by comparison of the rates of consumption of DPBF at the initial
stages of each experiment.35 The standard sensitizers 1 and 9
both have high singlet-oxygen quantum yields of 0.65 and 0.50,
respectively, in methanol, though these values are not necessarily
an accurate indicator for the efficiency of singlet-oxygen
generation under therapeutic irradiance conditions, as the
extinction coefficient of the sensitizer at the irradiation wave-
lengths (above 600 nm) is also a key contributing factor.36,18a
The experimental results showed that 18a and 18b had moderate
singlet-oxygen production over the 1 h time period, whereas
18c gave a better production of singlet oxygen comparable to
that of reference 1 (Figure 9). This confirmed that our new class
of photosensitizer has some inherent spin-orbit coupling. 18d,
which has two bromo-substituted aryl rings, did show an
enhanced singlet-oxygen generation over the other derivatives
18a-c and 1, as would be expected if the intramolecular external
heavy-atom effect contributed to spin-orbit coupling. These
results prove that we have achieved a range of varying singlet-
oxygen generation profiles at a single photosensitizer concentra-
tion with a greater than 4-fold rate distribution from least
effective 18b to most active 18d (Figure 9).
lower concentration of 5 × 10-8 M (Figure 10). Each of our
tested photosensitizers performed better than 9, with 19b being
the best, having a 7.7-fold rate enhancement when compared
to the reference (Figure 10). This is evidence that the heavy-
atom effect can be exploited to deliver controlled levels of the
key cytotoxic agent for our photosensitizer class. The dramati-
cally enhanced singlet-oxygen production levels of 19a-c, when
contrasted with 18a-c, show that the inclusion of the heavy
atom as a substituent directly onto the central core of the
photosensitizer has achieved our desired aim and has not given
rise to loss of excited-state energy by internal radiationless
transitions. No significant photobleaching of the sensitizers was
observed during these experiments (see the Supporting Informa-
tion).
As the incorporation of the bromine substituents can be
achieved by one facile high-yielding synthetic transformation,
this approach lends itself to tailoring a singlet-oxygen quantum
yield of photosensitizer to the specific therapeutic requirements
that are being addressed.
Photosensitizer Cellular Localization. To ascertain the
cellular localization of our photosensitizers, we chose 18a as a
representative compound, due to its inherent high fluorescence
quantum yield, which would enable facile visualization with
fluorescence microscopy. Incubation of HeLa cervical carcinoma
cells with 1 × 10-5 M aqueous 18a (CrEL formulated) for 1 h
at 37 °C gave rise to an efficient cellular uptake of photosen-
sitizer as determined by fluorescence microscopy (Figure 11).
The imaged cells displayed an uptake of 18a that appeared to
be localized to the cytoplasm (red color), with the nuclei
remaining free of photosensitizer (dark area).
To confirm cytoplasmic localization of our photosensitizers,
a confocal laser scanning microscopy (CLSM) study was carried
out. CLSM allows for the 3-D imaging of fluorescent com-
pounds within a single cell.37 The 3-D reconstruction of the
cellular distribution of 18a in HeLa cells was determined by
the combination of 20 focal plane sections of 0.5-1 µm apart
through a single cell. CLSM confirmed that the subcellular
It would be anticipated that to achieve the greatest internal
heavy-atom effect the bromines should be incorporated directly
on the central core of the photosensitizer. This was borne out
for 19a-c, which all showed an increased efficiency of singlet-
oxygen generation in comparison to 18a-c, even at a 100-fold
(34) Gollnick, K.; Griesbeck, A. Tetrahedron 1985, 41, 2057.
(35) Kochevar, I. E.; Redmond, R. W. Methods Enzymol. 2000, 319, 20.
(36) Reddi, E.; Rodgers, M. A. J.; Spikes, J. D.; Jori, G. Photochem. Photobiol.
1984, 40, 415.
(37) Shotton, D. M. J. Cell Sci. 1989, 94, 175.
9
10626 J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004

Heavy-Atom Effect for Photodynamic Therapy
A R T I C L E S
Figure 11. Fluorescence microscope (rhodamine filter) imaging of HeLa
cervical carcinoma cells following incubation with 18a (red color); dark
areas are the cell nuclei.
Figure 13. CLSM image of a HeLa cell following incubation with 18a
(red color) and nuclear stain SYTOX green (green color) (see the Supporting
Information).
light dose of either 8 or 16 J cm-2 40
. Following irradiation, the
cells were incubated for a further 48 h at 37 °C, after which
time the percentage cell viability was determined using a
tetrazolium chlorimetric reduction assay. Dark toxicity of
photosensitizers was determined by carrying out an identical
experiment as described above except that the light irradiation
step was omitted (0 J cm-2). All assay experiments were carried
out in triplicate and an average of the three individual runs are
presented. Hematoporphyrin 1 was used as a comparative
standard control and was assayed according to previously
documented procedures.41
Figure 12. CLSM image of a HeLa cell following incubation with 18a
(red color); the dark area is the cell nucleus (see the Supporting Information).
In our assay system, MRC5-SV40 cells displayed no deter-
minable dark toxicity with 18a, 18b, 19b, or 1 up to a
concentration of 10-4 M, whereas for 18d, dark toxicity was
observed at high concentrations with a calculated EC50 value
of 7.1 × 10-4 M (Figure 14a, Table 5). In contrast, irradiation
with 8 J cm-2 light dose showed a significant light-induced
toxicity with EC50 values determined for 18a, 18b, and 19b as
3.1 × 10-6, 1.1 × 10-4, and 3.7 × 10-8 M, respectively (Table
5). The exceptional light-induced toxicity of 19b was very
encouraging, as this molecule contained the two bromine heavy
atoms directly substituted onto the core of the photosensitizer.
Surprisingly 18d, which has two bromine substituents on the
benzene rings and was better that 18a-c in the singlet-oxygen
generation tests, performed poorly under these assay conditions,
with an EC50 value of 2.3 × 10-4 M.
localization of 18a was exclusive to the cytoplasm with no
nuclear localization detected (Figure 12) (see the Supporting
Information).
To prove the perinuclear localization of our photosensitizers,
a dual staining experiment was carried out. HeLa cells were
incubated for 1 h with 18a, followed by the treatment of the
cells with the green nucleic acid stain SYTOX green.38 18a and
SYTOX green were excited with a 1 mW green helium neon
(HeNe) 543 nm laser and a 50 mW argon 488 nm laser,
respectively. Simultaneous detection of cell-bound 18a and the
nuclear probe SYTOX green was achieved using 640-730 and
505-550 nm band-pass filters, respectively. The results gave
excellent contrast imaging between the green nuclear region and
the red cytoplasm, proving that 18a localizes to areas of the
cell other than the nucleus (Figure 13) (see the Supporting
Information). The absence of nuclear localization suggests that
this class of compound would not induce direct nuclear DNA
damage but mediate their actions at extranuclear targets. This
selective localization to the cytoplasm could be attributed to
some extent to the fact that these compounds are charge neu-
tral.
As phototoxicity should be dependent upon light dose as well
as photosensitizer concentration, we repeated this assay series
with a higher light irradiation of 16 J cm-2. The higher light
dose resulted in an improved EC50 value for each of our studied
photosensitizers, with values obtained for 18a at 5.0 × 10-7
M, 18b at 1.7 × 10-5 M, 18d at 2.8 × 10-4 M, and 19b at
1.4 × 10-8 M (Figure 14b, Table 5). Each of our tested
photosensitizers, apart from 18d, performed better at this light
dose than the standard control hematoporphyrin 1 (Table 5).
These data clearly portray how we have achieved a remark-
able spectrum of activity (from the micro- to nanomolar range)
across these structurally related sensitizers. In the case of 19b,
exploitation of the heavy-atom effect in vitro is seen to be a
viable method to control the excited triplet-state population and
Light-Induced Cytotoxicity Assay. Two different cell types
were examined in the assay, MRC5-SV40 transformed fibroblast
cells and HeLa cells. Varying concentrations of CrEL formu-
lated39 aqueous solutions of the photosensitizers 18a, 18b, 18d,
and 19b were incubated with the cells in the dark for 3 h.
Subsequently, the culture medium was removed and fresh
culture medium added to each well. The plates were irradiated
using a light source of wavelength 600-750 nm delivering a
(40) The light source was either a Waldmann PDT 1200 L instrument with an
emission spectrum of 600-750 nm or a 500 W halogen lamp combined
with red and water filters with an emission spectrum of 600-750 nm. Both
systems gave comparable assay data.
(41) Sol, V.; Blais, J. C.; Carre´, V.; Granet, R.; Guilloton, M.; Spiro, M.; Krausz,
P. J. Org. Chem. 1999, 64, 4431.
(38) Suzuki, T.; Matsuzaki, T.; Takata, K. Acta Histochem. Cytochem. 1998,
31, 297.
(39) No light or dark toxicity was observed for CrEL alone when tested in our
assay system.
9
J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004 10627

A R T I C L E S
Gorman et al.
Figure 15. (a) Dark toxicity of 18a (green), 18b (blue), 19b (red), and
Figure 14. (a) Dark toxicity of 18a (green), 18b (blue), 19b (red), and
hematoporphyrin 1 (black) for HeLa cells. (b) Light-induced (16 J cm-2
)
hematoporphyrin 1 for MRC5-SV40 cells. (b) Light-induced (16 J cm-2
)
toxicity of 18a (green), 18b (blue), 19b (red), and hematoporphyrin 1 (black)
for He La cells.
toxicity of 18a (green), 18b (blue), 19b (red), and hematoporphyrin 1 (black)
for MRC5-SV40 cells.
Table 6. In Vitro EC50 Assay Data for HeLa Cellsa
Table 5. In Vitro EC50 Assay Data for MRC5-SV40 Cellsa
EC50 (M)
EC50 (M)
entry compd
0 J cm-2
8 J cm-2
16 J cm-2
entry compd
0 J cm-2
8 J cm-2
16 J cm-2
(6.3 ( 3) × 10-5
(3.1 ( 3) × 10-6
(1.1 ( 1) × 10-4
(3.7 ( 0.3) × 10-8 (1.4 ( 0.1) × 10-8
(3.7 ( 1) × 10-5
(5.0 ( 1) × 10-7
(1.7 ( 1) × 10-5
(2.8 ( 2) × 10-4
1
2
3
4
5
1
ND
(3.3 ( 2) × 10-5
(1.9 ( 0.4) × 10-5
(3.1 ( 1 × 10-5
(3.4 ( 1) × 10-5
(4.1 ( 3) × 10-8
1
2
3
4
5
1
18a
18b
ND
ND
ND
18a (2.7 ( 1) × 10-5 (2.8 ( 0.7) × 10-6 (9.2 ( 1) × 10-7
18b (1.1 ( 1) × 10-4 (7.2 ( 5) × 10-5
18d (6.7 ( 2) × 10-4 (1.5 ( 1) × 10-4
18d (7.1 ( 2) × 10-4 (2.3 ( 2) × 10-4
19b
ND
(6.3 ( 2) × 10-8
19b
ND
a Standard deviation in brackets.
a Standard deviation in brackets.
singlet-oxygen quantum yields, as well as to translate that control
into significantly greater in vitro efficacy. It should be noted
that in structural terms 18b and 19b only differ by the two
bromine substituents, which gives rise to a divergence in efficacy
by over a 1000-fold.
the least active in our assay, followed by significantly improved
results for 18a (9.2 × 10-7 M) and the best being 19b (4.1 ×
10-8 M) (Figure 15b, Table 6). In this series both 18b and 18d
were marginally out-performed by the standard hematoporphyrin
1. An improvement in efficacy for each photosensitizer was
observed on increasing the light dose from 8 to 16 J cm-2, again
showing the expected light-dose response behavior. Again, the
in vitro heavy-atom effect was clearly observed; when compar-
A second identical study was carried out using the HeLa cell
line. Photosensitizer dark toxicity was only observed at high
concentrations for 18a,b,d, with our most active compound 19b
showing no observable dark toxicity in the tested concentration
range (Figure 15a, Table 6). A broad range of light-induced
cytotoxicity was also observed for the photosensitizer series in
HeLa cells that was comparable to that observed in the MRC5-
SV40 cell line. The determined EC50 values for the series with
a light dose of 16 J cm-2 showed that 18b (3.1 × 10-5 M) was
ing EC50 data for 18b and 19b at a light dose of 8 J cm-2
,
there was over a 1000-fold efficacy increase while at 16 J cm-2
there was greater than a 750-fold increase (Table 6).
An overview of the compounds studied shows that we can
clearly control the degree of light induced cytotoxicity of our
new PDT agent class by minor structural modifications about
9
10628 J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004

Heavy-Atom Effect for Photodynamic Therapy
A R T I C L E S
reaction, the product precipitated from the reaction mixture. The reaction
was cooled to room temperature and filtered and the isolated solid
washed with ethanol (2 × 10 mL) to yield the product 16a as a blue-
black solid (0.28 g, 33%).
Method D. A 100 mL round-bottomed flask was charged with 15a
(1.0 g, 3.71 mmol), ammonium acetate (10.0 g, 0.13 mol), and ethanol
(40 mL) and heated under reflux for 24 h. During the course of the
reaction, the product precipitated from the reaction mixture. The reaction
was cooled to room temperature and filtered and the isolated solid
washed with ethanol (2 × 10 mL) to yield the product 16a as a blue-
black solid (0.30 g, 35%).
the core of the sensitizer. The low activity of 18d in both tested
cell lines is a reminder of the difficulty of prediction of how a
photosensitizer will perform in a cellular system, as for this
case the promising singlet-oxygen production tests did not
translate into improved efficacy. On the basis of the in vitro
assay data, we conclude that for the photosensitizers 18a and
19b, there is a clear potential for their development into new
PDT agents, whereas 18b would have more practical applica-
tions in tumor imaging, due to its high fluorescence quantum
yield and low light and dark toxicity.
Method E. A 100 mL round-bottomed flask was charged with 15a
(1.0 g, 3.71 mmol), ammonium acetate (10.0 g, 0.13 mol), and butanol
(40 mL) and heated under reflux for 24 h. The reaction was cooled to
room temperature, the solvent concentrated to 10 mL and filtered, and
the isolated solid washed with ethanol (2 × 10 mL) to yield the product
Conclusions
We have described an optimized synthesis and the photo-
physical characteristics of a new class of photosensitizer that
has a significant potential to deliver PDT agents that could be
individually fine-tuned for an exact therapeutic requirement.
Minor structural modification about the periphery of the core
photosensitizer allows for control over key features such as
absorption maximum and singlet-oxygen generation. These
properties are translated into control over light-induced toxicity
in vitro. We are currently developing new modifications that
will facilitate tumor cell selectivity and specific subcellular
localization of the PDT agents. On a wider basis, the phthalo-
cyanine-like spectral properties of these compounds and their
unchelated precursors, coupled with their ease of synthesis and
structural manipulation, may elicit attention from researchers
concerned with a wide range of applications, such as nonlinear
optical media, chemical sensors, fluorescent biochemical as-
saying, optical storage media, and catalysis.
1
16a as a blue-black solid (0.33 g, 39%), mp 287-289 °C. H NMR
(CDCl3) δ: 8.04-8.08 (m, 4H), 7.94-7.97 (m, 4H), 7.33-7.57 (m,
12H), 7.21 (s, 2H) (NH not observed). 13C NMR (CDCl3) δ: 155.1,
149.6, 142.7, 133.7, 132.2, 130.1, 129.2, 129.1, 128.3, 128.0, 126.6,
114.9. IR (KBr disk) cm-1: 1598. λmax (CHCl3) nm: 598. EI-MS: m/z
449. HRMS Calcd for C32H24N3 [M + H]+: 450.1970. Found:
450.1948. Anal. Calcd for C32H23N3: C, 85.50; H, 5.16; N, 9.35.
Found: C, 85.21; H, 5.12; N, 9.41.
[5-(4-Methoxyphenyl)-3-phenyl-1H-pyrrol-2-yl][5-(4-methoxy-
phenyl)-3-phenylpyrrol-2-ylidene]amine (16b). 15b (1.25 g, 4.17
mmol) and ammonium acetate (11.27 g, 0.146 mol) were dissolved in
butanol (50 mL) and heated under reflux for 24 h. The reaction was
cooled to room temperature, the solvent concentrated to 25 mL and
filtered, and the isolated solid washed with ethanol (2 × 10 mL) to
yield the product 16b as a black-green solid (0.51 g, 47%), mp 238-
240 °C. 1H NMR (CDCl3) δ: 8.04-8.07 (m, 4H), 7.90 (d, J ) 8.8 Hz,
4H), 7.33-7.44 (m, 6H), 7.13 (s, 2H), 7.04 (d, J ) 8.8 Hz, 4H), 3.91
(s, 6H) (NH not observed). 13C NMR (CDCl3) δ: 161.8, 154.2, 149.4,
142.4, 134.2, 129.2, 128.4, 128.3, 127.8, 125.5, 114.9, 114.7, 55.9. IR
(KBr disk) cm-1: 1600. λmax (CHCl3) nm: 621. EI-MS: m/z 509.
HRMS Calcd for C34H28N3O2 [M + H]+: 510.2182. Found: 510.2165.
Anal. Calcd for C34H27N3O2: C, 80.13; H, 5.34; N, 8.25. Found: C,
79.89; H, 5.20; N, 8.28.
[3-(4-Methoxyphenyl)-5-phenyl-1H-pyrrol-2-yl][3-(4-methoxy-
phenyl)-5-phenylpyrrol-2-ylidene]amine (16c). 15c (1.0 g, 3.34 mmol)
and ammonium acetate (9.0 g, 0.117 mol) were dissolved in butanol
(50 mL) and heated under reflux for 24 h. The reaction was cooled to
room temperature, the solvent concentrated to 25 mL and filtered, and
the isolated solid washed with ethanol (2 × 10 mL) to yield the product
16c as a brown solid (0.40 g, 48%), mp 290-292 °C. 1H NMR (CDCl3)
δ: 8.03 (d, J ) 8.6 Hz, 4H), 7.92-7.95 (m, 4H), 7.42-7.54 (m, 6H),
7.10 (s, 2H), 6.97 (d, J ) 8.6 Hz, 4H), 3.88 (s, 6H) (NH not observed).
13C NMR (CDCl3) δ: 159.7, 155.0, 149.5, 142.3, 132.4, 130.4, 129.9,
129.1, 126.7, 126.5, 113.8, 113.6, 55.4. IR (KBr disk) cm-1: 1605.
Experimental Section
(3,5-Diphenyl-1H-pyrrol-2-yl)(3,5-diphenylpyrrol-2-ylidene)am-
ine (16a). Method A. A 50 mL round-bottomed flask was charged
with 15a (1.5 g, 5.57 mmol) and ammonium formate (12.3 g, 195 mmol)
and set up to carry out a distillation. The reagents were heated with
stirring on an isomantle at 180 °C until the distillation was complete
(1-1.5 h). During the course of the reaction, the mixture turned an
intense blue color, ammonia was evolved, and aqueous formic acid
was distilled out of the reaction flask at 101-102 °C. The reaction
mixture in the round-bottomed flask was allowed to cool to room
temperature, dissolved in CH2Cl2 (200 mL), and washed with water
(2 × 50 mL) and the organic layer dried over sodium sulfate. The
reaction mixture was preabsorbed onto silica. The preabsorbed silica
was placed on top of a bed of silica (150 g) in a sintered glass funnel
and eluted with methanol. The silica was then eluted with CH2Cl2 to
yield the product 16a as a blue-black solid (0.43 g, 34%).
Method B. A 50 mL round-bottomed flask was charged with 15a
(2.5 g, 9.28 mmol) and ammonium acetate (25.05 g, 0.325 mol) and
set up to carry out a distillation. The reagents were heated with stirring
on an isomantle at 180 °C until the distillation was complete (1-1.5
h). During the course of the reaction, the mixture turned an intense
blue color, ammonia was evolved, and liquid distilled out of the reaction
flask over a temperature range of 125-140 °C. The reaction mixture
in the round-bottomed flask was allowed to cool to room temperature,
dissolved in CH2Cl2 (200 mL), and washed with water (2 × 50 mL)
and the organic layer dried over sodium sulfate. The reaction mixture
was preabsorbed onto silica. The preabsorbed silica was placed on top
of a bed of silica (150 g) in a sintered glass funnel and eluted with
methanol. The silica was then eluted with CH2Cl2 to yield the product
16a as a blue-black solid (1.0 g, 47%).
λmax (CHCl3) nm: 608. EI-MS: m/z 509. HRMS Calcd for C34H28N3O2
[M + H]+: 510.2182. Found: 510.2188. Anal. Calcd for C34H27N3O2:
C, 80.13; H, 5.34; N, 8.25. Found: C, 79.98; H, 5.38; N, 7.99.
[3-(4-Bromophenyl)-5-phenyl-1H-pyrrol-2-yl][3-(4-bromophenyl)-
5-phenylpyrrol-2-ylidene]amine (16d). 15d (0.3 g, 0.86 mmol) and
ammonium acetate (2.3 g, 30.2 mmol) were dissolved in butanol (30
mL) and heated under reflux for 48 h. The reaction was cooled to room
temperature, the solvent concentrated to 15 mL and filtered, and the
isolated solid washed with ethanol (2 × 10 mL) to yield the product
16d as a purple solid (0.062 g, 24%), mp 343-345 °C. 1H NMR
(CDCl3) δ: 7.88-7.95 (m, 8H), 7.48-7.60 (m, 10H), 7.19 (s, 2H) (NH
not observed). 13C NMR (CDCl3) δ: compound was too insoluble to
record a spectrum. IR (KBr disk) cm-1: 1542. λmax (CHCl3) nm: 603.
EI-MS: m/z 607. HRMS Calcd for C32H22Br2N3 [M + H]+: 606.0180.
Found: 606.0192. Anal. Calcd for C32H21Br2N3: C, 63.28; H, 3.49;
N, 6.92. Found: C, 62.96; H, 3.34; N, 6.69.
Method C. A 100 mL round-bottomed flask was charged with 15a
(1.0 g, 3.71 mmol), ammonium formate (8.20 g, 0.13 mol), and ethanol
(40 mL) and heated under reflux for 24 h. During the course of the
9
J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004 10629

A R T I C L E S
Gorman et al.
(4-Bromo-3,5-diphenyl-1H-pyrrol-2-yl)(4-bromo-3,5-diphenylpyr-
rol-2-ylidene)amine (17a). 16a (0.2 g, 0.44 mmol) was dissolved in
dry benzene (20 mL) and treated with bromine (0.67 mL, 1.34 mmol)
and the reaction stirred at room temperature for 2 h. The solvent was
evaporated to dryness, yielding a black-green solid 17a as the
Found: 558.2188. Anal. Calcd for C34H26BF2N3O2: C, 73.26; H, 4.70;
N, 7.54. Found: C, 73.16; H, 4.67; N, 7.45.
BF2 Chelate of [3-(4-Methoxyphenyl)-5-phenyl-1H-pyrrol-2-yl]-
[3-(4-methoxyphenyl)-5-phenylpyrrol-2-ylidene]amine (18c). 16c
(0.13 g, 0.24 mmol) was dissolved in dry CH2Cl2 (30 mL), treated
with diisopropylethylamine (0.41 mL, 2.41 mmol) and boron trifluoride
diethyl etherate (0.41 mL, 3.37 mmol), and stirred at room temperature
under N2 for 24 h. The mixture was washed with water (2 × 50 mL),
and the organic layer was dried over sodium sulfate and evaporated to
dryness. Purification by column chromatography on silica eluting with
CH2Cl2/hexane (1:1) gave the product 18c as a metallic red solid
1
hydrobromide salt (0.3 g, 97%), mp 320-322 °C. H NMR (CDCl3)
δ: 13.70 (bs, 2H), 8.21 (bs, 4H), 7.57-7.59 (m, 4H), 7.30-7.41 (m,
8H), 7.13-7.18 (m, 4H). 13C NMR (CDCl3) δ: 154.5, 146.7, 133.3,
132.0, 130.3, 129.8, 128.2, 127.7, 127.3, 126.7, 125.7, 107.3. IR (KBr
disk) cm-1: 1597. λmax (CHCl3) nm: 577. EI-MS: m/z 607. Anal. Calcd
for C32H22Br3N3: C, 55.84; H, 3.22; N, 6.11. Found: C, 55.44; H,
3.19; N, 5.80.
1
(0.11 g, 88%), mp 288-290 °C. H NMR (CDCl3) δ: 8.05 (d, J )
[4-Bromo-5-(4-methoxyphenyl)-3-phenyl-1H-pyrrol-2-yl][4-bromo-
5-(4-methoxyphenyl)-3-phenylpyrrol-2-ylidene]amine (17b). 16b (0.2
g, 0.392 mmol) was dissolved in dry benzene (15 mL) and treated with
bromine (0.06 mL, 1.17 mmol) and the reaction stirred at room
temperature for 2 h. The solvent was evaporated to dryness, yielding
a black-green solid 17b as the hydrobromide salt (0.28 g, 96%), mp
286-288 °C. 1H NMR (CDCl3) δ: 13.5 (bs, 2H), 8.31 (d, J ) 8.8 Hz,
4H), 7.25-7.34 (m, 6H), 7.07-7.17 (m, 8H), 3.90 (s, 6H). 13C NMR
(CDCl3) δ: 163.9, 154.1, 147.2, 134.3, 133.7, 131.0, 130.8, 129.2,
127.9, 119.5, 114.6, 108.1, 55.9. IR (KBr disk) cm-1: 1604. λmax
(CHCl3) nm: 600. EI-MS: m/z 667. Anal. Calcd for C34H26Br3N3O2:
C, 54.57; H, 3.50; N, 5.62. Found: C, 54.15; H, 3.22; N, 5.41.
[4-Bromo-3-(4-methoxyphenyl)-5-phenyl-1H-pyrrol-2-yl][4-bromo-
3-(4-methoxyphenyl)-5-phenylpyrrol-2-ylidene]amine (17c). 16c (0.2
g, 0.39 mmol) was dissolved in dry benzene (20 mL) and treated with
bromine (0.06 mL, 1.17 mmol) and the reaction stirred at room
temperature for 2 h. The solvent was evaporated to dryness, yielding
a blue-black solid 17c as the hydrobromide salt (0.28 g, 96%), mp
306-307 °C. 1H NMR (CDCl3) δ: 14.04 (bs, 2H), 8.24-8.26 (m, 4H),
7.53-7.59 (m, 6H), 7.20-7.31 (m, 4H), 6.68 (d, J ) 8.6 Hz, 4H),
3.85 (s, 6H). 13C NMR (CDCl3) δ: 159.9, 153.9, 146.1, 133.0, 131.7,
131.6, 130.2, 127.5, 125.9, 122.2, 112.1, 105.4, 54.2. IR (KBr disk)
cm-1: 1604. λmax (CHCl3) nm: 585. EI-MS: m/z 667. Anal. Calcd for
C34H26Br3N3O2: C, 54.57; H, 3.50; N, 5.62. Found: C, 54.17; H, 3.65;
N, 5.25.
8.9 Hz, 4H), 7.98-8.04 (m, 4H), 7.44-7.48 (m, 6H), 7.0 (d, J )
8.9 Hz, 4H), 6.92 (s, 2H), 3.90 (s, 6H). 13C NMR (CDCl3) δ: 161.0,
159.1, 145.5, 143.8, 131.9, 130.9, 130.6, 129.5, 128.5, 125.4, 117.5,
114.2, 55.5. IR (KBr disk) cm-1: 1602. EI-MS: m/z 558.2. HRMS
Calcd for C34H27BF2N3O2 [M + H]+: 558.2164. Found: 558.2174.
Anal. Calcd for C34H26BF2N3O2: C, 73.26; H, 4.70; N, 7.54. Found:
C, 72.98; H, 4.62; N, 7.41.
BF2 Chelate of [3-(4-Bromophenyl)-5-phenyl-1H-pyrrol-2-yl][3-
(4-bromophenyl)-5-phenylpyrrol-2-ylidene]amine (18d). 16d (0.156
g, 0.26 mmol) was dissolved in dry CH2Cl2 (50 mL), treated with
diisopropylethylamine (0.22 mL, 1.28 mmol) and boron trifluoride
diethyl etherate (0.22 mL, 1.80 mmol), and stirred at room temperature
under N2 for 24 h. The mixture was washed with water (2 × 50 mL),
and the organic layer was dried over sodium sulfate and evaporated to
dryness. Purification by column chromatography on silica eluting with
CH2Cl2/hexane (1:1) gave the product 18d as a metallic brown solid
1
(0.105 g, 62%), mp 265-266 °C. H NMR (CDCl3) δ: 8.04-8.07
(m, 4H), 7.93 (d, J ) 8.3 Hz, 4H), 7.63 (d, J ) 8.3 Hz, 4H), 7.51-
7.53 (m, 6H), 7.06 (s, 2H). 13C NMR (CDCl3) δ: 159.9, 145.4, 142.9,
131.9, 131.3, 131.2, 130.6, 129.6, 128.7, 124.3, 119.3, 107.8. IR (KBr
disk) cm-1: 1519. EI-MS: m/z 654. HRMS Calcd for C32H21BBr2F2N3
[M + H]+: 654.0163. Found: 654.0190. Anal. Calcd for C32H20-
BBr2F2N3: C, 58.67; H, 3.08; N, 6.41. Found: C, 58.40; H, 2.96; N,
6.23.
BF2 Chelate of (4-Bromo-3,5-diphenyl-1H-pyrrol-2-yl)(4-bromo-
3,5-diphenylpyrrol-2-ylidene)amine (19a). 17a·HBr (0.14 g, 0.20
mmol) was dissolved in dry CH2Cl2 (50 mL), treated with diisopro-
pylethylamine (0.38 mL, 2.22 mmol) and boron trifluoride diethyl
etherate (0.38 mL, 3.11 mmol), and stirred under N2 for 24 h. The
mixture was washed with water (2 × 25 mL), and the organic layer
was dried over sodium sulfate and evaporated to dryness. Purification
by column chromatography on silica eluting with CH2Cl2/hexane
(2:1) gave the product 19a as a metallic brown solid (0.12 g, 90%),
BF2 Chelate of (3,5-Diphenyl-1H-pyrrol-2-yl)(3,5-diphenylpyrrol-
2-ylidene)amine (18a). 16a (0.2 g, 0.45 mmol) was dissolved in dry
CH2Cl2 (80 mL), treated with diisopropylethylamine (0.8 mL, 4.6 mmol)
and boron trifluoride diethyl etherate (1 mL, 8.13 mmol), and stirred
at room temperature under N2 for 24 h. The mixture was washed with
water (2 × 50 mL), and organic layer was dried over sodium sulfate
and evaporated to dryness. Purification by column chromatography on
silica eluting with CH2Cl2/hexane (1:1) gave the product 18a as a
metallic brown solid (0.19 g, 86%), mp 231-232 °C. 1H NMR (CDCl3)
δ: 8.02-8.08 (m, 8H), 7.40-7.53 (m, 12H), 7.03 (s, 2H). 13C NMR
(CDCl3) δ: 159.6, 145.7, 144.2, 132.4, 131.6, 130.9, 129.7, 129.6,
129.4, 128.7, 128.6, 119.2. IR (KBr disk) cm-1: 1514. EI-MS: m/z
497.1. HRMS Calcd for C32H23BF2N3 [M + H]+: 498.1953. Found:
498.1963. Anal. Calcd for C32H22BF2N3: C, 77.28; H, 4.46; N, 8.45.
Found: C, 77.0; H, 4.38; N, 8.29.
1
mp 278-281 °C. H NMR (CDCl3) δ: 7.86-7.90 (m, 4H), 7.71-
7.74 (m, 4H), 7.42-7.50 (m, 12H). 13C NMR (CDCl3) δ: 157.5, 143.3,
141.9, 129.8, 129.7, 129.5, 129.3, 128.6, 128.4, 127.0, 126.9, 109.8.
IR (KBr disk) cm-1: 1518. EI-MS: m/z 653. HRMS Calcd for C32H21-
BBr2F2N3 [M + H]+: 654.0163. Found: 654.0414. Anal. Calcd for
C32H20BBr2F2N3: C, 58.67; H, 3.08; N, 6.41. Found: C, 58.62; H, 3.04;
N, 6.29.
BF2 Chelate of [5-(4-Methoxyphenyl)-3-phenyl-1H-pyrrol-2-yl]-
[5-(4-methoxyphenyl)-3-phenylpyrrol-2-ylidene]amine (18b). 16b
(0.2 g, 0.39 mmol) was dissolved in dry CH2Cl2 (60 mL), treated with
diisopropylethylamine (0.7 mL, 3.93 mmol) and boron trifluoride diethyl
etherate (0.7 mL, 5.5 mmol), and stirred at room temperature under N2
for 24 h. The mixture was washed with water (2 × 50 mL), and the
organic layer was dried over sodium sulfate and evaporated to dryness.
Purification by column chromatography on silica eluting with CH2Cl2/
hexane (1:1) gave the product 18b as a metallic red solid (0.16 g, 73%),
BF2 Chelate of [4-Bromo-5-(4-methoxyphenyl)-3-phenyl-1H-
pyrrol-2-yl][4-bromo-5-(4-methoxyphenyl)-3-phenylpyrrol-2-yli-
dene]amine (19b). 17b·HBr (0.13 g, 0.174 mmol) was dissolved in
dry CH2Cl2 (50 mL), treated with diisopropylethylamine (0.2 mL, 1.15
mmol) and boron trifluoride diethyl etherate (0.3 mL, 2.43 mmol), and
stirred under N2 for 24 h. The solution was washed with water (2 × 25
mL), and the organic layer was dried over sodium sulfate and evaporated
to dryness. Purification by column chromatography on silica eluting
with CH2Cl2/hexane (3:1) gave the product 19b as a metallic brown
1
1
mp 200-202 °C. H NMR (CDCl3) δ: 8.04-8.09 (m, 8H), 7.40-
solid (0.11 g, 88%), mp 271-273 °C. H NMR (CDCl3) δ: 7.84-
7.48 (m, 6H), 7.01-7.25 (m, 4H), 6.99 (s, 2H), 3.87 (s, 6H). 13C NMR
(CDCl3) δ: 162.0, 158.1, 145.4, 143.1, 132.5, 131.8, 129.5, 129.4,
128.6, 124.2, 118.7, 114.3, 55.5. IR (KBr disk) cm-1: 1600. EI-MS:
m/z 557. HRMS Calcd for C34H27BF2N3O2 [M + H]+: 558.2164.
7.87 (m, 4H), 7.75 (d, J ) 9.0 Hz, 4H), 7.40-7.46 (m, 6H), 6.90 (d,
J ) 9.0 Hz, 4H), 3.85 (s, 6H). 13C NMR (CDCl3) δ: 161.7, 157.1,
144.0, 142.2, 132.7, 131.0, 130.9, 129.7, 128.2, 122.0, 113.8, 110.2,
55.5. IR (KBr disk) cm-1: 1603. EI-MS: m/z 715. HRMS Calcd for
9
10630 J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004

Heavy-Atom Effect for Photodynamic Therapy
A R T I C L E S
C34H25BBr2F2N3O2 [M + H]+: 714.0375. Found: 714.0359. Anal.
Calcd for C34H24BBr2F2N3O2: C, 57.10; H, 3.38; N, 5.88. Found: C,
57.30; H, 3.52; N, 5.87.
either 18a (5 × 10-6 M) or 19b (5 × 10-8 M) gave no reduction in
intensity of the 410 nm absorption band over 1 h.
Cytotoxicity Studies of 18a, 18b, 18c, and 19b. Stock photosen-
sitizer solutions were diluted with minimum essential medium (MEM)
containing 10% fetal calf serum (FCS). Cells were seeded at 5000 cells/
well in 96-well plates and incubated for 24 h. Cells were incubated
with photosensitizer in the dark for 3 h at 37 °C. The culture medium
was removed, the cells were washed three times with PBS, and fresh
culture medium was added to each well. The plates were irradiated
with a light dose of either 8 or 16 J cm-2 from a filtered light source
(wavelength 600-750 nm). Following irradiation, the cells were
incubated for a further 48 h before being assessed for cell viability.
The dark toxicity of each photosensitizer was also assessed in every
experiment. Each assay was carried out in triplicate, and results are
presented as an average.
BF2 Chelate of [4-Bromo-3-(4-methoxyphenyl)-5-phenyl-1H-
pyrrol-2-yl][4-bromo-3-(4-methoxyphenyl)-5-phenylpyrrol-2-yli-
dene]amine (19c). 17c·HBr (0.18 g, 0.24 mmol) was dissolved in dry
CH2Cl2 (100 mL), treated with diisopropylethylamine (0.30 mL, 1.72
mmol) and boron trifluoride diethyl etherate (0.30 mL, 2.44 mmol),
and stirred under N2 for 24 h. The solution was washed with water (2
× 25 mL), and the organic layer was dried over sodium sulfate and
evaporated to dryness. Purification by column chromatography on silica
eluting with CH2Cl2/hexane (1:1) gave the product 19c as a metallic
1
red solid (0.13 g, 76%), mp 297-298 °C. H NMR (CDCl3) δ: 7.91
(d, J ) 8.7 Hz, 4H), 7.68-7.72 (m, 4H), 7.43-7.47 (m, 6H), 6.99 (d,
J ) 8.7 Hz, 4H), 3.89 (s, 6H). 13C NMR (CDCl3) δ: 159.9, 156.9,
143.1, 141.3, 131.4, 129.6, 129.2, 128.5, 126.9, 122.4, 112.6, 107.6,
54.3. IR (KBr disk) cm-1: 1600. EI-MS: m/z [M + H]+ 714. HRMS
Calcd for C34H25BBr2F2N3O2 [M + H]+: 714.0375. Found: 714.0352.
Anal. Calcd for C34H24BBr2F2N3O2: C, 57.10; H, 3.38; N, 5.88.
Found: C, 56.92; H, 3.33; N, 5.66.
Acknowledgment. This research was supported by grants
from the Association for International Cancer Research, UK
(AICR), and the Centre for Synthesis and Chemical Biology,
University College Dublin, Ireland. Thanks to Dr. J. Gallagher
for helpful crystallographic discussions, Dr. H. Mueller-Bunz
of the UCD Crystallographic Centre and Dr. D. Rai of the CSCB
Mass Spectrometry Centre for analysis.
Fluorescence Quantum Yield Measurements. Fluorescence quan-
tum yields for 18a-d and 19a-c were determined in spectroscopic-
grade chloroform at a concentration of 2 × 10-7 M. Mg-tetra-tert-
butylphthalocyanine was used as a standard with a literature value for
Φf of 0.84.42
Supporting Information Available: General experimental and
experimental details for the synthesis of 14d and 15a-d;
experimental details for CrEL formulation of 18a-d and 19a-
c; cell culture, fluorescent microscopy, and confocal laser
scanning microscopy methods; procedure for determining cy-
totoxicity of 1; assay procedure for measurement of cell
viability; UV-visible spectra for comparative singlet oxygen
generation by 18a; 1H and 13C NMR spectra for 16a-d, 17a-
c, 18a-d, 19a-c; UV-visible for 16a-d, 17a-c, 18a-d,
19a-c; fluorescence spectra for 18a-d, 19a-c; X-ray crystal-
lographic data of 19a as a CIF file with a view of the crystal
packing; complete focal plane CLSM image data for Figures
12 and 13 and movie of Figure 12. This material is available
free of charge via the Internet at http://pubs.acs.org.
Comparative Singlet-Oxygen Generation Measurements. An
aerated 100 mL isopropyl alcohol solution of photosensitizer (5 × 10-6
or 5 × 10-8 M) and 1,3-diphenylisobenzofuran (5 × 10-5 M) was
irradiated with a 250 W filtered red light source (>600 nm) at 25 °C
for 1 h. Aliquots (2 mL) were removed from the reaction mixture at 5
min intervals, and a UV-visible spectrum was recorded. Reaction of
1,3-diphenylisobenzofuran with singlet oxygen was monitored by the
reduction in intensity of the absorption band at 410 nm over time.
Irradiation of a 100 mL isopropyl alcohol DPBF solution (5 × 10-5
M) in the absence of photosensitizer gave no reduction in intensity of
the 410 nm absorption band over 1 h. Irradiation of a deoxygenated
100 mL isopropyl alcohol DPBF solution (5 × 10-5 M) containing
(42) Stiel, H.; Teuchner, K.; Paul, A.; Freyer, W.; Leupold, D. J. Photochem.
Photobiol. A 1994, 80, 289.
JA047649E
9
J. AM. CHEM. SOC. VOL. 126, NO. 34, 2004 10631