In vitro and in vivo photocytotoxicity of boron dipyrromethene derivatives for photodynamic therapy

Article abstract of DOI:10.1021/jm901823u

To understand the effects of substitution patterns on photosensitizing the ability of boron dipyrromethene (BODIPY), two structural variations that either investigate the effectiveness of various iodinated derivatives to maximize the "heavy atom effect" o

Full text of DOI:10.1021/jm901823u

J. Med. Chem. 2010, 53, 2865–2874 2865
DOI: 10.1021/jm901823u
In Vitro and In Vivo Photocytotoxicity of Boron Dipyrromethene Derivatives for Photodynamic
Therapy
†
Siang Hui Lim, Cliferson Thivierge, Patrycja Nowak-Sliwinska, Junyan Han, Hubert van den Bergh,
Georges Wagni ꢀe res, Kevin Burgess, and Hong Boon Lee*
‡
§
‡
§
‡
,†
†
Cancer Research Initiatives Foundation (CARIF), Sime Darby Medical Centre, 47500 Subang Jaya, Selangor, Malaysia,
Department of Chemistry, Texas A&M University, College Station, Texas 77842, Institute of Bio-Engineering, and
‡
§
Institute of Chemical Sciences and Engineering, Medical Photonics Group, Swiss Federal Institute of Technology (EPFL),
015 Lausanne, Switzerland
1
Received December 10, 2009
To understand the effects of substitution patterns on photosensitizing the ability of boron dipyrro-
methene (BODIPY), two structural variations that either investigate the effectiveness of various
iodinated derivatives to maximize the “heavy atom effect” or focus on the effect of extended conjugation
at the 4-pyrrolic position to red-shift their activation wavelengths were investigated. Compounds with
conjugation at the 4-pyrrolic position were less photocytotoxic than the parent unconjugated com-
pound, while those with an iodinated BODIPY core presented better photocytotoxicity than com-
pounds with iodoaryl groups at the meso-positions. The potency of the derivatives generally correlated
well with their singlet oxygen generation level. Further studies of compound 5 on HSC-2 cells showed
almost exclusive localization to mitochondria, induction of G /M-phase cell cycle block, and onset of
2
apoptosis. Compound 5 also extensively occluded the vasculature of the chick chorioallantoic
membrane. Iodinated BODIPY structures such as compound 5 may have potential as new photo-
dynamic therapy agents for cancer.
Introduction
blue, and Nile red analogues and the chalcogenopyrylium
class of photosensitizers. These classes of compounds, how-
ever, suffer from a major drawbacks due to their inherent
dark cytotoxicity.
a
Photodynamic therapy (PDT ) is now a well-recognized
modality for cancer treatment, but to date, only a small
number of PDT drugs, namely, porfimer sodium, temopor-
fin, and aminolevulinic acid, have been approved mainly
for treatment of skin, gynecological, gastrointestinal, and
some head and neck cancers. PDT involves site-specific
activation of an administered photosensitizer using light of
One alternative class of non-porphyrin photosensitizers
that has emerged recently is the BODIPY chromophore.
BODIPYs have many characteristics of an ideal photosen-
sitizer including high extinction coefficients, high quantum
efficiencies of fluorescence, relative insensitivity to environ-
ment, and resistance to photobleaching. Nagano and co-
workers reported the synthesis of a simple diiodo-substi-
tuted BODIPY with possible applications that included
PDT. The diiodo-substituted BODIPY at the 4-pyrollic
position produces the supposed internal heavy atom effect
and subsequently enhanced the intersystem crossing effi-
ciency from a singlet to a triplet state that controls the
1
2
,3
a wavelength matched to the λmax of the photosensitizer
in order to generate cytotoxic reactive oxygen species
1
(
vasculature damage or by recruiting members of the infla-
e.g., O ) that eradicate tumors via cellular damage, via
2
2
,3
mmatory and immune response system. These photo-
sensitizers, as well as the majority of those currently being
investigated in clinical trials, share a common cyclic tetra-
4
pyrrole structure, probably due to the fact that modern
5
singlet oxygen production. O’Shea and co-workers pre-
pared a series of azadipyrromethenes with high absorption
in the far-red wavelengths and demonstrated their efficacy
PDT has evolved from naturally derived porphyrins such as
hematoporphyrin. Aside from cyclic tetrapyrrole struc-
tures, a number of naturally occurring and synthetic dyes
that are non-porphyrin have also been evaluated for their
photosensitizing ability against cancer. The focus has been
primarily on cationic structures such as methylene blue, Nile
6
in light-induced toxicity in a panel solid tumor cell line.
One of the azadipyrromethenes was later shown to effec-
tively eradicate the subcutaneously xenografted MDA-231
7
breast tumors in nude mice. Akkaya and co-workers
introduced another class of water-soluble BODIPY dyes
8
with extended conjugation at the 5-pyrollic positions.
*
To whom correspondence should be addressed. Phone: þ603-5639-
1
874. Fax: þ603-5639-1875. E-mail: hongboon.lee@carif.com.my.
These photosensitizers were shown to have strong absorp-
tions in the 650-680 nm therapeutic window and good
photoinduced cytotoxicity in K562 leukemia cells at sub-
micromolar concentrations even under low fluence rate
LED irradiation.
a
Abbreviations: BODIPY, boron dipyrromethene; PDT, photodyna-
mic therapy; CAM, chorioallantoic membrane; EDD, embryo develop-
ment day; Rh123, rhodamine 123; DPBF, 1,3-diphenylisobenzofuran;
MTT, methylthiazolyldiphenyltetrazolium bromide; DMSO, dimethyl
sulfoxide.
r
2
010 American Chemical Society
Published on Web 03/03/2010
pubs.acs.org/jmc

2
866 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Lim et al.
In the first variation, functionalizations such as meso-sub-
stitution and sulfonation to improve hydrophilicity were
tested to fine-tune the activity of iodinated BODIPY-based
structures such as Nagano’s compound (6 in Figure 1).
Because photosensitizers that absorb in the longer wave-
lengths may be activated deeper in the tissues and may there-
fore be clinically favored, a second variation consisting of
compounds with extended conjugation at the 4-pyrollic posi-
tion, which is a structural variation that has not been studied
priortothis, was included in this study. First, thePDT efficacy
of these BODIPY-based photosensitizers on a panel of leuke-
mia and solid tumor cell lines, with particular attention on
their structure-activity relationships, was studied. The effect
ofthe mostactivederivative on anoral cancer cell line, HSC-2,
was further investigated for intracellular localization, cell
cycle arrest, and onset of apoptosis experiments, as well as
in a preclinical in vivo model using the chick chorioallantoic
This study investigates the effects of two structural varia-
tions of BODIPY on their photocytotoxicity in terms of
photophysical properties and in vitro and in vivo efficacies.
Figure 1. Structures of BODIPY.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2867
membrane (CAM), in ovo, for PDT efficacy in terms of
9
vascular occlusion.
Table 1. Damage Score of PDT-Induced Vasculature Network Occlu-
sion
occlusion
score
findings
0
1
2
No occlusion
Partial closure of capillaries of diameter of <10 μm
Closure of capillary system, partial closure of blood vessel
of diameter of <30 μm, and size reduction of larger blood
vessels
3
4
5
Closure of vessels of diameter of <30 μm and partial
closure of larger blood vessels
Results and Discussion
Total closure of vessels of diameter of <70 μm and partial
closure of larger vessels
Total occlusion of vessels in the irradiated area
Structural Variations and Photophysical Properties. Two
structural variations around the BODIPY core of compound
1
were investigated in this study (Figure 1). The first
(compounds 2-6) investigated the effectiveness of various
iodinated derivatives in order to maximize the “heavy atom
effect”. To fine-tune the activity of iodinated BODIPY-
based structures, additional functionalizations such as
meso-substitution with alkyl or aryl groups as well as
sulfonation to improve hydrophilicity were tested. Com-
pound 5 contains a carboxylic acid handle and could be
easily attached to other molecules later if required. For the
second variation (compounds 7-15), the effect of extended
conjugation at the 4-pyrollic position was examined. Ex-
tended conjugation at the 4-pyrrolic positions increases the
absorption wavelength of the compounds to the red, permit-
ting the use of longer excitation wavelength that penetrates
deeper into biological tissues for effective treatment.
Table 2. Photophysical Properties of BODIPY Derivatives
λ
abs max
ε
(M cm
λ
em max
ref
-
1
-1
a
b
compd
(nm)
)
(nm)
Φfl
dye
solvent
EtOH
1
2
3
4
5
6
7
8
9
505
504
537
498
525
534
494
531
529
527
559
528
560
517
530
83 000
82 000
89 000
100 000
93 000
110 000
81 000
27 000
c
516
510
552
509
540
548
504
570
560
549
580
551
580
527
539
0.80
0.64
0.05
0.34
0.03
0.02
0.95
0.42
0.25
0.73
0.51
0.73
0.52
0.78
0.92
1
1
2
1
2
1
1
2
2
2
3
2
3
2
2
b
CHCl
CH Cl
3
2
2
H O
2
EtOH
MeOH
EtOH
EtOH
EtOH
EtOH
EtOH
EtOH
EtOH
EtOH
EtOH
1
1
1
1
1
0
1
2
3
4
63 000
13 000
62 000
30 000
63 000
72 000
Table 2 summarizes the photophysical data of compounds
1
-15. The absorption and emission wavelengths of com-
pounds 1-15 range from 500 to 600 nm. As expected for
compounds 8-15, extending the conjugation with an acry-
late shifts the λabs max to the red by 20-30 nm while attaching
two acrylates red-shifts the λabs max by 50-60 nm, compared
to compound 1. The λabs max values for iodinated compounds
15
a
Φ represents quantum efficiency of fluorescence. Reference dyes
fl
used for quantum yield determination (solvent, Φ ): (1) fluorescein (0.1
fl
M NaOH, 0.92); (2) rhodamine 6G (EtOH, 0.94); (3) rhodamine B
(
c
EtOH, 0.97). Value not determined.
3
, 5, and 6 are also red-shifted compared to compound 1 but
not for compounds 2 and 4, which are aryl-iodinated at the
meso position. In addition, all compounds have high extinc-
tion coefficients and high quantum efficiencies of fluores-
cence except for a few structures. Among the exceptions,
compounds 3, 5, and 6 have much lower quantum efficiency
of fluorescence and a correspondingly higher singlet oxygen
generation rate (Table 3) compared to the other BODIPYs
studied here, probably as a result of the enhanced intersys-
tem crossing efficiency from the lowest singlet excited state to
the triplet state contributed by the internal heavy-atom
effect. As in the case of the λabs max values above, compounds
irradiation were also carried out to determine cytotoxicity in
the dark. Results were expressed as IC , the concentration
5
0
of compound (in μM) that inhibits proliferation rate by 50%
compared to control untreated cells. The parent compound,
denoted as 1, was also prepared and tested for compari-
son. From the assay, all compounds had negligible or
undeterminable unirradiated cytotoxicity up to 100 μM.
2
Upon irradiation with 4.1 J/cm of light, compounds 1-6
and 10-15 demonstrated photosensitized cytotoxicity with
IC50 values in the submicromolar to tens of micromolar
range. Compounds 5 and 6, which had two iodine atoms
directly attached to the BODIPY core, showed the highest
activity among the analogues, with IC50 values that are up
to 100 times lower than that of compound 1 (0.045 μM vs
4.4 μM in HL-60). In contrast, compounds 7-9 displayed
poor activity with undeterminable IC50 up to 100 μM.
The influence of the iodine atom on the photocytotoxic
activity of the compounds was evident from studying com-
pounds 1-6. Meso-substitution with a para-iodoaryl group
in compound 2 does not alter the photocytotoxicity signifi-
cantly compared to compound 1, while further substitution
with iodine atoms on the two pyrrolic 4-carbon to yield either
compound 3 from 2 or compound 6 from 1 improved the
activity by 10- and 100-fold, respectively, in all three cell
lines, alluding to the importance of iodine atom substitution
on the pyrrollic carbons rather than on the meso-aryl posi-
tion. Compound 5, which has an additional carboxylic acid
2
and 4 which contain a para-iodoaryl group at the meso
position did not demonstrate the same loss in fluorescence
yields or the same increase in singlet oxygen generation rate
as compounds 3, 5, and 6 probably because in the former
group, first, the iodine atom is not directly attached to the
BODIPY core and, second, the iodoaryl plane is twisted
1
0
relative to the BODIPY plane, overall causing the aryl
iodine atom to only, at most, elicit an intramolecular external
6
heavy-atom effect.
In Vitro Photocytotoxic and Comparative Singlet-Oxygen
Generation. The in vitro photocytotoxic activity of com-
pounds 1-15 against a promyelocytic leukemia cell line
(HL-60), an oral squamous carcinoma cell line (HSC-2),
and a nasopharyngeal carcinoma cell line (HK1) following
2
irradiation with 4.1 J/cm of a broad spectrum light was
determined using a modified methylthiazolyldiphenyltetra-
zolium bromide (MTT) assay. Parallel assays without light

2
868 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Lim et al.
Table 3. Comparative Singlet Oxygen Generation and in Vitro Photo Cytotoxicity Induced by BODIPY
activity IC50 (μM)
b
HL-60
HSC-2
HK1
singlet oxygen generation,
a
relative rate
2
2
2
2
2
2
BODIPY
0 J/cm
4.1 J/cm
0 J/cm
4.1 J/cm
0 J/cm
4.1 J/cm
1
2
3
4
5
6
7
8
9
0.48
0.15
13.9
0.07
24.6
23.9
0.07
0.01
0.00
0.57
0.41
0.73
0.17
0.24
0.29
>100
>100
10
4.4 ( 0.4
2.7 ( 1.2
0.42 ( 0.06
54.3 ( 8.3
0.045 ( 0.004
0.062 ( 0.011
>100
>100
>100
10
8.7 ( 2.0
5.1 ( 0.8
0.5 ( 0.1
59.4 ( 6.2
0.10 ( 0.06
0.64 ( 0.06
>100
76.8 ( 10.6
>100
6.2 ( 1.2
5.7 ( 0.1
0.69 ( 0.08
59.0 ( 7.1
0.57 ( 0.1
0.57 ( 0.06
>100
95.5 ( 7.9
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
55.8 ( 0.8
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
10
11
12
13
14
15
4.8 ( 0.7
5.2 ( 0.6
0.49 ( 0.07
57.7 ( 6.6
4.9 ( 0.6
3.8 ( 0.7
5.4 ( 0.6
11.1 ( 7.4
0.6 ( 0.1
37.7 ( 15.7
4.1 ( 0.2
>100
>100
5.3 ( 0.6
8.5 ( 0.5
1.1 ( 0.6
>100
99.6 ( 0.6
>100
>100
5.0 ( 0.6
4.8 ( 0.9
>100
a
b
Comparative singlet oxygen generation of photosensitizers relative to methylene blue. IC50, the concentration of compound that inhibits the
proliferation rate by 50% compared with control untreated cells. Values represent the mean ( SD of three determinations assessed 24 h using standard
MTT assay. Cells were incubated with compound for 2 h prior to irradiation with 4.1 J/cm .
2
tether at the meso position, has similar to marginally better
activity than compound 6 in all three cell lines. An attempt to
improve the water solubility of compound 2 by substitut-
ing with sodium sulfonate to yield compound 4 resulted in
over a period of 1 h. A light source filtered at 510 nm
wavelength was used to minimize the photobleaching of
DPBF. As a result, this would have caused an underestima-
tion of singlet oxygen generation rate of the compounds with
λabs max lower than 510 nm, as the light transmission through
the filter begins to drop below 510 nm. Each of the com-
pounds was tested at an equivalent concentration of refer-
ence sensitizer methylene blue. The results from this study
ranged from 0.01- to over 24-fold of singlet oxygen genera-
tion rate compared to that of methylene blue. Importantly,
the rate of reactive oxygen generation measured generally
correlated with the potency of these compounds and there
may be a main factor affecting the photocytotoxicity of these
BODIPY compounds.
Photosensitizer Cellular Localization. To ascertain the
intracellular localization of the BODIPYs, compound 5,
owing to its good potency data, was chosen and examined
by spinning disk confocal microscopy using dual staining
techniques (Figure 2). Costaining images and topographic
profiles of HSC-2 cell line loaded with compound 5 and a
mitochondria-specific dye rhodamine 123 (Rh123) revealed
an almost identical overlap, suggesting that compound 5
localized particularly well in mitochondria (Figure 2A,B). In
comparison, compound 5 displayed only partial colocali-
zation with endoplasmic reticulum and lysosomes, accord-
ing to the confocal images and topographic profiles of
compound 5 with ER-Tracker (Figure 2C,D) and with
LysoTracker (Figure 2E,F), respectively. Staining of the
cytoplasmic or nuclear membrane by compound 5 was not
detected, indicating that it does not react nonspecifically
with biological membranes. Furthermore, the nucleus re-
mained free of compound 5 (dark nuclear area), signifying
that this class of compounds would not be expected to
directly damage DNA.
1
0 times loss in activity.
For the effect of extended conjugation at 4-pyrrolic posi-
tions on photocytotoxic activity of BODIPYs, compounds
-15 were studied. Extending the 4-pyrollic carbon with a
single aldehyde (7) or the hydrophilic allylic carboxylic acid
8) and allylic suphonic acid (9) resulted in loss of activity
of greater than 100 μM IC50 values. Single extensions at the
-pyrrolic position with acrylate esters affected the activity
7
(
4
differently depending on the length of the alkyl ester group,
where the methyl compound 10 showed no change in activity
while n-butyl compound 12 demonstrated 10-fold improve-
ment in activity compared to compound 1. For compounds
with double extension with the same groups at the 4-pyrollic
positions, the methyl compound 11 remained unaltered in
its activity compared to compound 1, but interestingly, the
n-butyl compound 13 showed 10-fold loss in activity com-
pared to 1 or 100-fold loss compared to the singly extended
counterpart 12. The explanation for the reversal in struc-
ture-activity relationship of the n-butyl acrylate esters
compared with the methyl acrylate esters from the single-
to the double-extension (compounds 10 and 12 compared
with compounds 11 and 13) is not obvious and could just be
due to the poorer solubility of the doubly extended n-butyl
derivative 13. Compounds 14 and 15, which were failed
attempts to prepare the analogous ring-constrained conju-
gated structures, were also tested and showed similar acti-
vities compared to compound 1, further suggesting the minor
role that 4-pyrrolic extended conjugation or the lack of it
plays in modulating the photocytotoxicity of the BODIPY
structures.
Subsequently, the relative rate of singlet oxygen genera-
tion was measured for compounds 1-15 by monitoring the
reaction of known singlet oxygen acceptor 1,3-diphenyliso-
benzofuran (DPBF) with photosensitizers generated singlet
oxygen. This was achieved by following the loss of DPBF
absorbance at 410 nm at an initial concentration of 50 μM
Mitochondria perform vital cellular functions and are
involved in multiple signaling cascades in regulation of
metabolism, cell cycle control development, and cell death.
In PDT, mitochondria are an important target. During
mitochondria-photosensitization in PDT, cytochrome c is
released from mitochondria to directly effect rapid cell death
12
1
1

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2869
Figure 2. Intracellular localization of compound 5 in HSC-2 cells. Spinning disk confocal images (A, C, E) and fluorescence topographic
profiles (B, D, F) of HSC-2 cells double-stained with 100 nM compound 5 and respective organelle probes. (A, B) Mitochondria were labeled
with 100 nM Rh123 and excited at 494 nm. (C, D) Endoplasmic reticulum were labeled with 100 nM ER-Tracker and excited at 365 nm. (E, F)
Lysosomes were labeled with 500 nM of LysoTracker and excited at 365 nm. Compound 5 was excited at 575 nm. Line indicates the longitudinal
transcellular axis analyzed to generate the topography fluorescence profiles. Objective magnification is ꢀ63.
through downstream effector pathways of apoptosis, by-
passing other upstream apoptotic signaling pathways that
require synthesis of new proteins. This mechanism of action
is particularly useful in the treatment of cancer types that
are chemoresistant because of mutations in upstream pro-
apoptotic signaling pathways.
Cell Cycle Arrest and Apoptosis. To study the mechanism of
action of compound 5 that contributed to its photocytoxicity,
the effect of compound 5 on cell cycle and apoptosis of HSC-2
cells was investigated using flow cytometric method. The cell
cycle profile of HSC-2 cells treated with compound 5 was
analyzed in a time course experiment. At 0.25 μM, compound
5was found to induce G /M arrest in HSC-2 cells as early as 2 h
following light irradiation (Figure 3). HSC-2 cells in G /M
2
1
3
2
phase gradually increased from 26.1% in the control group to
39.8%, 48.4%, 55.5%, respectively, at 2, 4, and 6 h after light

2
870 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Lim et al.
Figure 3. Effects on HSC-2 cell cycle phase at various intervals postirradiation analyzed using flow cytometry after treatment with 0.25 μM
compound 5 and irradiated with a light dose of 4.1 J/cm . DC represents unirradiated dark control.
2
Figure 4. Representative histograms of the event of annexin V-fluorescein isothiocyanate (FITC) binding to phosphatidylserine as an
2
indicator of apoptosis in HSC-2 cells treated with 0.5 μM compound 5 and irradiated with a light dose of 4.1 J/cm . M1 represents viable cell
population, M2 represents apoptotic cell population, and DC represents unirradiated dark control.
irradiation. Concomitant with the increased proportion of
G /M phase cells, HSC-2 cells in the G phase were reduced
from 57.7% to 46.3%, 33.8%, and 26.4% while the proportion
of HSC-2 cells in S phase remained fairly constant. At 8 and
recovery of the cell cycle profile to one with a reduced G2/M
peak may be due to the redistribution of cells to sub-G1,
indicating the onset of apoptosis of arrested cells. A similar
2
1
G /M arrest was also observed in hypericin-based PDT, a
2
1
2 h, the proportion of cells arresting in the G /M phase was
2
naturally occurring photosensitizer currently undergoing
1
4,15
reduced to 48.8% and 26.0%, respectively, after light irradia-
tion. Following the reduction of the proportion of cells arrest-
ing in G /M phase, an increase of sub-G cells population from
8
of cells treated with compound 5 without irradiation remains
unchanged compared to control.
In the present study, the maximal level of G /M cell-
cycle arrest was observed at 6 h after PDT. Thereafter, the
research.
with hypericin resulted in phosphorylation of mitochondria
Photosensitization of Hela cervical cancer cells
Bcl-2 that correlated with G /M cell cycle arrest, followed by
2
2
1
1
4
to 12 h (from 4.4% to 10.6%) was observed. The proportion
the onset of apoptosis.
In addition to cell cycle analysis, the onset of apopto-
sis was also quantified in flow cytometry experiments by
measuring the externalization of membrane phosphatidyl-
serine through annexin V-FITC staining, an event that is
2

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2871
Figure 5. Effects of drug concentration and light dose on vascular
occlusion efficacy of compound 5 in the CAM model. Error bars
represent standard error mean from at least 10 embryos.
considered characteristic of cells undergoing apoptosis
(
Figure 4). Flow cytometric analysis of HSC-2 cells treated
with 0.5 μM compound 5 showed the onset of apoptosis by
2 h following irradiation with 20.7% of the cells stained
1
positive for annexin V compared to less than 10% at 0 or 4 h
time points. The proportion of cells undergoing apoptosis
continued to increase rapidly to 47.4% by 16 h, and at 24 h
the apoptotic cell proportion was at 89.9%.
PDT-Induced Vascular Occlusion. One of the ways PDT
causes damage during cancer treatment is by shutdown
1
6
of blood vessels feeding the tumor. Hence, the ability of
compound 5 to exert in vivo vasculature disruption was
investigated using the in ovo CAM model. In the present
study, the ability of compound 5 to induce occlusion of blood
vessels in the CAM by PDT was performed at 3.5-7.0 nmol/
2
embryo with a light dose of 20-40 J/cm . The degree of
vascular occlusion was scored 24 h after treatment according
to Table 1. The score for photosensitizer-mediated PDT-
induced vascular occlusion is shown in Figure 5, and the
angiograms representing the vascular occlusion score are
shown in Figure 6. As expected, the degree of vascular
occlusion increased with drug and light dose. The average
Figure 6. Representative angiographies of blood vessels supplying
the CAM at beginning and 24 h after PDT, illustrating the vascular
occlusion efficacy induced by compound 5 at 3.5-7 nmol/embryo.
2
Irradiation was performed at 510-560 nm of excitation with 20-
CAM vasculature damage score when irradiated at 20 J/cm
on embryos treated with 3.5 and 7.0 nmol/embryo of com-
pound 5 was at approximately 2 and 3, respectively. When
2
0 J/cm light dose. Cremophor EL (2.5%) and EtOH (2.5%) in
4
saline were used as a control vehicle. Objective magnification isꢀ4.
2
the light dose increased to 40 J/cm , the damage score
approved such as palladium bacteriopheophorbide, porfi-
mer sodium, lutetium texaphyrin, 5-aminolevulinic acid, and
increased accordingly to 3.5 and 5. Meanwhile, the control
eggs that received 20 μL of vehicle (cremophor EL 2.5%,
EtOH 2.5% in saline) and exposed to a similar light dose
showed no detectable vascular alteration in the treated area.
This indicated that the vascular occlusion observed in em-
bryos treated with compound 5 was neither caused by the
vehicle components nor was it thermally induced. Finally,
treatment with compound 5 alone without irradiation did
not induce any vascular occlusion, as nonirradiated areas
remained perfused after 24 h.
1
8,19
verteporfin.
For example, a PDT experiment in the
CAM model using verteporfin at a dose that is similar to
the recommended clinical dosage for the treatment of age-
related macular degeneration has been shown to cause
complete occlusion of the large neovessels, an observation
9
that correlates well with clinical setting. In addition, in a
BALB/c mouse model where PDT treatment by verteporfin
resulted in suppression of tumor growth, antivascular effects
were also observed, whereby a decrease in the blood volume
As the main respiratory organ of the chick embryo, CAM
is a well-vascularized membrane that is suitable as a model
for PDT. It is easily accessible, inexpensive, and relatively
easy to handle for photosensitizer administration, light
irradiation, fluorescence analysis of administered photo-
sensitizer, and microscopy examination of PDT-induced
2
0,21
at the tumor site was noted.
The results from the study
demonstrated that compound 5 was able to induce complete
closure of larger vessels, which is considered favorable in
PDT treatment of cancer.
9
,17
Conclusions
vascular damage.
CAM is a viable model that has been
successfully used to evaluate the photodynamic-induced
vascular occlusion efficacy of some photosensitizers that
are in clinical trials as well as those that are already clinically
We have demonstrated the in vitro photocytotoxic activity
of BODIPY derivatives against cell lines from leukemia
and two types of solid tumor. Structure-activity relationship

2
872 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Lim et al.
a
study indicated the importance of having iodine atoms di-
rectly substituted on the BODIPY pyrollic carbon-4 position
rather than at the meso-aryl position, in accordance with the
high singlet oxygen generation rate of these compounds.
Extended conjugation at the 4-pyrollic positions shifts the
λmax abs to the red but did not confer extra potency except for
the compound with a single n-butyl acrylate ester (12).
Hydrophilic analogues substituted with groups such as car-
boxylic acid, sulfonic acid, or sodium sulfonate generally
drastically diminished the activity. An exception here was
the doubly iodinated compound5 with an aliphatic carboxylic
acid which showed up to 100-fold lower IC50 value in HL-60
compared to the parent compound 1, perhaps because of the
structural difference where the carboxylic acid group is not
directly conjugated with the BODIPY core. Fluorescence
microscopy studies showed that compound 5 localized exclu-
sively within the mitochondria. This, together with data from
cell cycle analysis and onset of apoptosis studies, suggests that
compound 5 probably induced cell death through mitochon-
dria-dependent apoptosis rather than through damage to
nucleic materials. In addition, an emulsion of compound 5
was able to occlude the vasculature network in the CAM
in vivo model, further showing its potential as an effective
PDT agent.
Scheme 1
a
Although photosensitizers with >600 nm excitation wave-
lengths allow deeper access into biological tissues for effective
treatment of abnormal tissues of bigger volume, the shorter
absorption wavelength of the BODIPY-based photosensiti-
zers studied here, such as compound 5, may be preferred in the
treatment of superficial tissues. In a clinical PDT of cancer in
the esophagus and bronchi with porfimer sodium II, treat-
ment regime with light corresponding to the 514 nm excitation
wavelength exhibited similar effectiveness as that with 630 nm
in terms of tumor eradication but, significantly, with less
damage of the deep tissue which could cause perforation in
3
Reagents and conditions: (i) POCl , DMF, 1,2-dichloroethane,
85 °C, 17 h; (ii) I
BF , toluene, 80 °C, 5 h; (iv) BF
3 OEt
HIO
, MeOH, 25 °C, 30 min.
2
, HIO
3
, EtOH, 60 °C, 20 min; (iii) succinic anhydride,
3
2
3
3 OEt , Et N, 20 °C, 16 h; (v) I
2
3
2
,
3
25
using chlorosulfonic acid, respectively. Compound 5 was
prepared from 2,4-dimethylpyrrole and succinic anhydride in
the presence of BF3 OEt (to yield compound A), followed by
3
2
iodination of the resultant BODIPY. Compound 6 was obtained
from diiodination of compound 1. Compound 7 was obtained
from treating compound 1 with dry dimethylformamide and
5
POCl
catalyzed activation of the parent compound 1 using Heck-type
. Compounds 8-15 were prepared using palladium-
3
2
2
the esophagus. Overall, our results suggest that BODIPY
structures, especially compound 5, have potential to be ex-
plored as a clinically useful agent for PDT of cancer.
2
6
coupling with the respective acrylate structures. The experi-
mental details for measuring the photophysical properties and
synthesis of compounds 3, 5, and 7 are included in the Support-
ing Information. All compounds were dissolved in dimethyl
sulfoxide (DMSO) at 10 mM, aliquoted, and stored at -20 °C
prior to use.
Experimental Section
Materials. ER-Tracker Blue-White DPX, LysoTracker Blue
DND-22, rhodamine 123, SYTOX Green were purchased from
Molecular Probes, Invitrogen (OR). Annexin V-FITC apop-
tosis detection kit 1 was purchased from BD Biosciences (CA).
All cell culture reagents were purchased from Gibco, Invitrogen
Cells. HL-60 human promyelocytic leukemia cells were
obtained from American Type Culture Collection (VA) and
maintained in RPMI 1640 medium supplemented with 10%
FBS. HSC-2 oral cavity human squamous carcinoma cells were
obtained from Health Science Research Resources Bank (Japan
Health Sciences Foundation, Japan). HK1 nasopharyngeal
epithelial carcinoma cell line was a gift from University of
Hong Kong Culture Collection (University of Hong Kong,
Hong Kong). Both cell lines were grown in MEM medium
supplemented with 10% FBS.
(
Auckland, New Zealand). RNase A, dimethyl sulfoxide, pro-
pidium iodide, methylene blue, 1,3-diphenylisobenzofuran, and
methylthiazolyldiphenyltetrazolium bromide were purchased
from Sigma (St. Louis, MO). Fluorescein isothiocyanate dex-
tran (FITC-dextran, 20 kDa) was purchased from Sigma-
Aldrich (Buchs, Switzerland). Indian ink was purchased
from Pebeo S.A. (Romanel-sur-Morges, Switzerland). Cremo-
phor EL and absolute ethanol (Fluka Biochemika, Buchs,
Switzerland) were used as dosing vehicles. NaCl, 0.9%,
pH 7.4 (saline), was purchased from B Braun Medical AG
Photoinduced Cytotoxicity Assay. Approximately 15 000 HL-
60 cells/well or 3000 cells/well for HSC-2 and HK1 cells in
phenol-red-free culture medium containing 10% fetal bovine
serum were seeded in a 96-well plate. HSC-2 and HK1 cells were
allowed to adhere overnight before test compounds were intro-
duced. Photosensitizer stock solutions (10 mM in DMSO) were
diluted with medium, and concentrations varying from 0.001 to
100.0 μM were tested on the cells. The control wells received
0.01% of DMSO, equivalent to the highest amount of DMSO
used as vehicle in the compound-treated wells. Following 2 h of
(
Emmertbr u€ cke, Switzerland).
Photosensitizers. The synthesis of many of the compounds
studied here has already been reported elsewhere except for
compounds 3, 5, and 7 which is shown in Scheme 1. Compound 1
was prepared using a new method whereby 3,5-dimethylpyrrole-
2
by addition of BF
-carbaldehyde was treated with 1.2 equiv of POCl followed
3
2
3
2
3
3
OEt
2
.
Compound 2 was prepared from
treatment, cells were irradiated with a light dose of 4.1 J/cm
from a broad spectrum light source and fluence rate of 6.8 mW/
condensation of 2,4-dimethylpyrrole with 4-iodobenzene acid in
POCl and reaction of the resultant dipyrromethene intermedi-
ate with BF3 OEt2. Compounds 3 and 4 were prepared from
2
cm and were further incubated for 24 h before cell viability was
3
2
4
assessed using MTT assay. Following incubation, 15 μL of MTT
solution (5 mg/mL) was added into each well and incubated for
3
compound 2 through iodination using I /HIO
2
3
and disulfonation

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2873
4
h at 37 °C. The medium was then removed, and 100 μL of
development day (EDD) 3, an opening on the eggshell, about
4 mm in diameter, was bored at the apex and sealed with
adhesive tape to avoid contamination and desiccation of the
egg contents. The eggs were further incubated in stationary
position with the apex upright until EDD-9.
DMSO was added to dissolve the formazan crystal formed.
Absorbance, as a measure of viable cell number, was read at
5
Labsystems, Chantilly, VA). The dark toxicity of each photo-
sensitizer was also determined in every experiment.
Comparative Singlet-Oxygen Generation Measurements. An
amount of 8 mL of aerated isopropanol containing 50 μM
of DPBF and the photosensitizer (0.5 μM or 5 μM) in a 6-well
70 nm with an OpsysMR microplate spectrometer (Thermo-
Microscopic observation of CAM vasculature and the light
irradiation during PDT were performed with an epifluorescence
Eclipse 600 FN microscope equipped with a CFI Achromat
4ꢀ/0.1 objective (Nikon, Japan). Illumination was provided by a
100 W mercury arc lamp (Osram, GmbH, Augsburg, Germany).
Light doses were adjusted with neutral density filters and measured
with a calibrated Field-Master GS power analyzer (Coherent,
Santa Clara, CA). For exciting and detecting compound 5, the
microscope was equipped with a G-2A filter set (excitation, 510-
560 nm) (Nikon, Japan). For detecting FITC, a B-2E/C filter set
(excitation, 465-495 nm) was used (Nikon, Japan). Fluorescence
angiograms (1280 ꢀ 1024 pixels with 4095 gray level, i.e., 12 bits)
were acquired with an F-view II 12-bit monochrome Peltier-cooled
digital CCD camera driven with analySIS DOCU software from
Soft Imaging System (M u€ nster, Germany). The fluorescence
images were stored in a 16-bit TIF file.
2
plate was irradiated at 6.8 mW/cm of filtered light source
of >510 nm wavelength with a Roscolux medium yellow no.
1
2
0 filter (Rosco, NY) at room temperature for 1 h. Aliquots of
00 μL were removed from the mixture at various fixed inter-
vals, and the absorbance was measured at 410 nm. The rate
of singlet oxygen production was determined from the reduction
in intensity of absorbance recorded over time. Irradiation of
DPBF-isopropanol solution in the absence of photosensitizer
as a negative control and solution containing methylene blue as
a comparative control was also carried out. The relative singlet
oxygen generation rate for each of the photosensitizers was
determined by using methylene blue as a reference.
Cellular Localization. HSC-2 cells grown on round glass
coverslips in 12-well plates were co-incubated with 100 nM
photosensitizer together with organelle-specific fluorescence
probes. The endoplasmic reticulum was labeled with 100 nM of
ER-Tracker Blue-White DPX, the lysosomes were stained with
On EDD-9, the egg opening was extended to ∼30 mm in
diameter. Embryo was intravenously administered with a single
bolus of 3.5-7 nmol/embryo of photosensitizer in dosing vehi-
cle (cremophor EL 5%, EtOH 5% in saline) at the CAM main
vasculature. A minute after injection, a site with vessels of
diameter between 5 and 100 μm was irradiated at a light dose
5
00 nM of LysoTracker Blue DND-22, and the mitochron-
2
dria were tracked with 100 nM Rh123, each for 15-30 min of
incubation at room temperature. After incubation, cells were
gently rinsed in PBS to remove free dyes, and the stained cells
were observed using Olympus DSU spinning disk confocal
microscope configured with a PlanApo 63ꢀ oil objective
of 20-40 J/cm filtered at 510-560 nm with an irradiation area
2
2
of 0.02 cm and fluence rate of 40 mW/cm . The site was
photographed at the beginning and at the end of irradiation.
Subsequently, the egg opening was sealed with parafilm and the
embryo was further incubated for 24 h before assessing the PDT
damage induced.
(
Olympus Optical Corp. Ltd., Tokyo, Japan) and iXon EM þ
Fluorescence angiograms were performed in order to assess
the PDT-induced vasculature damage. Blood vessels were per-
fused with 10 μL of 25 mg/mL FITC-dextran followed by
injection of Indian ink in order to decrease the embryo’s
interfering fluorescence from deeper located vessels. This inter-
fering luminescence may change rapidly with time because of the
embryo’s movement. Prior to injection, the Indian ink was
filtered using a sterile cellulose acetate membrane (0.2 μm pores,
Renner GmbH, Darmstadt, Germany). The vasculature net-
work at the site of irradiation was illuminated by exciting the
FITC at 465-495 nm wavelengths on the epifluorescence
microscope. The vasculature network was imaged, and the
damage induced by PDT was scored according to the criteria
digital camera (Andor Technology, South Windsor, CT). Fluo-
rescent images of XY sections at 0.2 μm were collected sequentially
using Olympus Cell software. Organelle-specific fluorescence
probes were respectively excited at 330-385 nm wavelengths
to illuminate ER-tracker and LysoTracker, at 460-490 nm for
Rh123 and at 520-550 nm for the photosensitizer.
Annexin V-FITC Apoptosis Analysis. HSC-2 cells grown in
0 mm dishes at 50% confluency were treated with 0.5 μM
6
compound 5. Following 2 h of incubation, cells were irradiated
2
with 4.1 J/cm of broad spectrum light. At various treatment
intervals, floating cells in the medium were pooled together with
the adherent cells after trypsinization and were washed twice
with cold phosphate buffered saline (PBS). The cells were
8
6
as defined by Lange et al. (Table 1). At least 10 embryos were
assessed for each treatment group.
resuspended with 1ꢀ binding buffer at 1ꢀ10 cells/mL. A 100
μL of cell suspension was transferred to a flow cytometry tube
followed by 5 μL of annexin V-FITC and 5 μL of 200 μg/mL
propidium iodide in PBS. The cells were gently mixed and
incubated for 15 min at room temperature in the dark before
analysis on a FACSCalibur flow cytometer equipped with a 488
Acknowledgment. This work was supported by grants
from the Cancer Research Initiatives Foundation. We thank
David Lyn (Matrix Optics, Malaysia) for help with the
cellular localization experiments.
4
nm argon laser. The fluorescence data of 1 ꢀ 10 cells were
collected with the FL1 detector with 530/30 band-pass filter to
collect annexin-FITC fluorescence and with the FL3 detector
with a 630 nm long-pass filter to collect propidium iodide
fluorescence.
Note Added after ASAP Publication. This paper was pub-
lished on March 3, 2010 with an error in the O’Shea structure.
The revised version was published on March 8, 2010.
Cell Cycle Analysis. HSC-2 cells were treated with 0.25 μM
compound 5 and collected as above. Cells were then fixed in
Supporting Information Available: General procedures for
spectroscopic measurements, synthesis procedures, and charac-
terization data for compounds 1-15. This material is available
free of charge via the Internet at http://pubs.acs.org.
7
0% ice-cold ethanol (v/v in PBS) overnight at 4 °C. Following
fixation, the cells were washed twice in cold PBS. The pellet was
then resuspended in PBS solution containing 20 μg/mL RNase
A and 1 μM SYTOX Green for 30 min. The cells were analyzed
on a FACSCalibur flow cytometer with 488 nm argon laser.
4
The DNA-SYTOX Green fluorescence of 1 ꢀ 10 cells were
References
collected with the FL1 detector with 530/30 band-pass filter.
PDT on the CAM Vasculature (in Ovo). Freshly obtained
fertilized chicken eggs were incubated with the narrow apex
down in a 90° swinging incubator (Savimat MG 200, Chauffry,
France) at 37 °C and 65% relative humidity. On embryo
(
1) Allison, R. R.; Downie, G. H.; Cuenca, R.; Hu, X. H.; Childs,
C. J. H.; Sibata, C. H. Photosensitizers in clinical PDT. Photodiagn.
Photodyn. Ther. 2001, 1, 27–42.
(2) Dolmans, D. E.; Fukumura, D.; Jain, R. K. Photodynamic therapy
for cancer. Nat. Rev. Cancer 2003, 3, 380–387.

2
874 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Lim et al.
(
3) Hopper, C. Photodynamic therapy: a clinical reality in the treat-
ment of cancer. Lancet Oncol. 2000, 1, 212–219.
4) Nyman, E. S.; Hynninen, P. H. Research advances in the use of
tetrapyrolic photosensitizers for photodynamic therapy. J. Photo-
chem. Photobiol., B 2004, 73, 1–28.
5) Yogo, T.; Urano, Y.; Ishitsuka, Y.; Maniwa, F.; Nagano, T. Highly
efficient and photostable photosensitizer based on BODIPY chro-
mophore. J. Am. Chem. Soc. 2005, 127, 12162–12163.
(15) Lee, H. B.; Ho, A. S. H.; Teo, S. H. p53 status does not affect
photodynamic cell killing induced by hypericin. Cancer Chemother.
Pharmacol. 2005, 58, 91–98.
(16) Chen, B.; Pogue, B. W.; Hoopes, P. J.; Hasan, T. Vascular and
cellular targeting for photodynamic therapy. Crit. Rev. Eukaryotic
Gene Expression 2006, 16, 276–305.
(17) Vargas, A.; Zeisser-Labou ꢁe be, M.; Lange, N.; Gurny, R.; Delie, F.
The chick embryo and its chorioallantoic membrane (CAM) for the
in vivo evaluation of drug delivery systems. Adv. Drug Delivery
Rev. 2007, 59, 1162–1176.
(18) R u€ ck, A.; B o€ hmler, A.; Steiner, R. PDT with TOOKAD studied in
the chorioallantoic membrane of fertilized eggs. Photodiag. Photo-
dyn. Ther. 2005, 2, 79–90.
(
(
(
(
6) Gorman, A.; Killoran, J.; O’Shea, C.; Kenna, T.; Gallagher, W. M.;
O’Shea, D. F. In vitro demonstration of the heavy-atom effect for
photodynamic therapy. J. Am. Chem. Soc. 2004, 126, 10619–10631.
7) Byrne, A. T.; O’Connor, A. E.; Hall, M.; Murtagh, J.; O’Neill, K.;
Curran, K. M.; Mongrain, K.; Rousseau, J. A.; Lecomte, R.;
McGee, S.; Callanan, J. J.; O’Shea, D. F.; Gallagher, W. M.
Vascular-targeted photodynamic therapy with BF2-chelated tetra-
aryl-azadipyrromethene agents: a multi-modality molecular ima-
ging approach to therapeutic assessment. Br. J. Cancer 2009, 101,
(19) Hammer-Wilson, M. J.; Cao, D.; Kimel, S.; Berns, M. W. Photo-
dynamic parameters in the chick chorioallantoic membrane
(CAM) bioassay for photosensitizers administered intraperitone-
ally (IP) into the chick embryo. Photochem. Photobiol. Sci. 2002, 1,
721–728.
1565–1573.
(
8) Atilgan, S.; Ekmekci, Z.; Dogan, A. L.; Guc, D.; Akkaya, E. U.
Water soluble distyryl-boradiazaindacenes as efficient photosensi-
tizers for photodynamic therapy. Chem. Commun. 2006, 42, 4398–
(20) Ichikawa, K.; Takeuchi, Y.; Yonezawa, S.; Hikita, T.; Kurohane,
K.; Namba, Y.; Oku, N. Antiangiogenic photodynamic therapy
(PDT) using Visudyne causes effective suppression of tumour
growth. Cancer Lett. 2004, 205, 39–48.
4400.
(
9) Lange, N.; Ballini, J. P.; Wagni ꢀe res, G.; van den Bergh, H. A new
drug screening procedure for photosensitizing agents used in
photodynamic therapy for CNV. Invest. Ophthalmol. Visual Sci.
(21) Kurohane, K.; Tominaga, A.; Sato, K.; North, J. R.; Namba, Y.;
Oku, N. Photodynamic therapy targeted to tumour-induced an-
giogenic vessels. Cancer Lett. 2001, 167, 49–56.
2
001, 42, 38–46.
(22) Grosjean, P.; Wagnieres, G.; Fontolliet, C.; van den Bergh, H.;
Monnier, P. Clinical photodynamic therapy for superficial cancer
in the oesophagus and the bronchi: 514 nm compared with 630 nm
light irradiation after sensitization with photofrin II. Br. J. Cancer
1998, 77, 1989–1995.
(23) Wu, L.; Burgress, K. A new synthesis of symmetric boraindacene
(BODIPY) dyes. Chem. Commun. 2008, 40, 4933–4935.
(24) Tahtaoui, C.; Thomas, C.; Rohmer, F.; Klotz, P.; Duportail, G.;
M ꢁe ly, Y.; Bonnet, D.; Hibert, M. Convenient method to access new
4,4-dialkoxy- and 4,4-dialkoxy-diaza-s-indacene dyes: synthesis
and spectroscopic evaluation. J. Org. Chem. 2007, 72, 269–272.
(25) Li, L.; Han, J.; Nguyen, B.; Burgess, K. Synthesis and spectral
properties of functionalized, water-soluble BODIPY derivatives.
J. Org. Chem. 2008, 73, 1963–1970.
(
10) Burghart, A.; Kim, H.; Welch, M. B.; Thoresen, L. H.; Reibenspies,
J.; Burgess, K. 3,5-Diaryl-4,4-difluoro-4-bora-3a,4a-diaza-s-inda-
cence (BODIPY) dyes: synthesis, spectroscopic, electrochemical,
and structural properties. J. Org. Chem. 1999, 64, 7813–7819.
11) Kochevar, I. E.; Redmond, R. W. Photosensitized production of
singlet oxygen. Methods Enzymol. 2000, 319, 20–28.
(
(
(
(
12) McBride, H. M.; Neuspiel, M.; Wasiak, S. Mitochodria: more than
just a powerhouse. Curr. Biol. 2006, 16, R551–R56.
13) Morgan, J.; Oseroff, A. R. Mitochondria-based photodynamic
anti-cancer therapy. Adv. Drug Delivery Rev. 2001, 49, 71–86.
14) Vantiegham, A.; Xu, Y.; Assefa, Z.; Piette, J.; Vandenheede, J. R.;
Merlevede, W.; de Witte, P. A. M.; Agostinis, P. Phosphorylation
2
of Bcl-2 in G /M phase-arrested cells following photodynamic
therapy with hypericin involves a CDK1-mediated signal and
delays the onset of apoptosis. J. Biol. Chem. 2002, 277, 37718–
(26) Thivierge, C.; Bandichhhor, R.; Burgess, K. Spectral dispersion
and water solubilization of BODIPY dyes via palladium-catalyzed
C-H functionalization. Org. Lett. 2007, 9, 2135–2138.
37731.