Phorochembrry und Photobiology Vol. 53, No. 4, pp. 549-553. 1991 0031-8655/91 $03.00 +O.

W
Printed in Great Britain. All rights rcservcd Copyright @ 1991 Pergamon Prcss plc

RESEARCH NOTE

THE PHOTODEGRADATION OF PORPHYRINS IN CELLS

CAN BE USED TO ESTIMATE THE LIFETIME OF
SINGLET OXYGEN
JOHAN MOAN*and KRISTIAN BERG
Institute for Cancer Research, 0310 Montebello, Oslo 3, Norway
(Received 25 June 1990; accepted 3 October 1990)

Abstract-NHIK 3025 cells were incubated with Photofrin I1 (PII) and/or tetra (3-
hydroxypheny1)porphyrin (3THPP) and exposed to light at either 400 or 420 nm, i.e. at the wavelengths
of the maxima of the fluorescence excitation spectra of the two dyes. The kinetics of the photo-
degradation of the dyes were studied. When present separately in the cells the two dyes are photode-
graded with a similar quantum yield. 3THPP is degraded 2-6 times more efficiently by light quanta
absorbed by the fluorescent fraction of 3THPP than by light quanta absorbed by the fluorescent
fraction of PI1 present in the same cells. The distance diffused by the reactive intermediate, supposedly
mainly lo2, causing the photodegradation was estimated to be on the order of 0.01-0.02 pm, which
corresponds to a lifetime of 0.01-0.04 ps of the intermediate in the cells. PI1 has binding sites at
proteins in the cells as shown by an energy transfer band in the fluorescence excitation spectrum at
290 nm. During light exposure this band decays faster than the Soret band of PI1 under the present
conditions. Photoproducts (lo2etc.) generated at one binding site contribute significantly in the
destruction of remote binding sites.

INTRODUCTION larger when cells are exposed t o light in a D20-

Porphyrins are degraded or modified by light (Cox
et af.,1982; Krieg and Whitten, 1984). Considerable
interest has been focused on such photodegradation
(often called photobleaching), since it can be taken
advantage of in photodynamic therapy of cancer
(Moan, 1986; Dougherty, 1987, Mang et af.,1987).
In fact, the concept of photobleaching has been
successfully applied in human clinical trials (Mang
et af., 1990). The mechanisms of photodegradation
are different for porphyrins in solution and for por-
phyrins in biological systems (Krieg and Whitten,
1984). In the latter case the rate constants are sig-
nificantly larger and different products are formed.
To a first approximation the photodegradation fol-
lows first order kinetics and the rate constants are
independent of the initial porphyrin concentration,
both in cells and in tissues (Mang er af., 1987).
This indicates that an excited porphyrin molecule is
degraded either by products formed by itself or by
direct interaction with its surrounding. However,
this is a crude approximation since the kinetics devi-
ate from first order after extended (but still clinically
relevant) exposure times. Furthermore, the rate
constants for degradation of Photofrin I1 (PII)? are
*To whom correspondence should be addressed.
tAbbreviations: HPLC, high performance liquid chroma-
tography; MEM, minimal essential medium, NCS, new-
born calf serum, PII, Photofrin 11; PBS, Dulbeccos
phosphate buffered saline; 3THPP, tetra (3-
hydroxyphenyl) porphyrin.
PAP 53:4-I
549
buffer than when they are exposed in a H 2 0 buffer,
indicating that '02-production plays a role (Moan
et af., 1988).
In order to elucidate the photobleaching mechan-
isms further we have incubated cells with two differ-
ent porphyrins, Photofrin I1 (PII) and tetra (3-
hydroxylphenyl) porphyrin (3THPP). which have
different fluorescence excitation and emission spec-
tra. Thus, even when they are present in the same
cell they can be excited and monitored with some
selectivity. Both are negatively charged lipophilic
dyes which, according to fluorescence microscopic
studies (data not shown), localize almost identically
in the cells, presumably in membrane structures. In
such a system it is possible to study the degradation
of one dye caused by excitation of the other dye,
and to estimate the distance travelled by reactive
intermediates before deactivation.
MATERIALS AND METHODS
Chemicals. Photofrin I1 was obtained from Photomed-
ica, Raritan, NJ. The drug was stored frozen in small vials
for up to 8 months before being used in the present
experiments. It was thoroughly checked by HPLC, and by
cell survival experiments that the drug did not change in
its characteristics during the time of storage. The pro-
cedures of HPLC and cell survival experiments are
described elsewhere (Moan et a / . , 1982; Moan and
Sommer, 1983). The 3THPP was obtained from Porphyrin
Products, Logan, UT. Stock solutions (0.1 mg/mL) were
prepared in 0.05 M NaOH as earlier described (Moan et
al., 1987).
Cell cultivation. Cells of the line NHIK 3025 (derived
550 JOHANMOANand KRlsnAN BERG

from a human carcinoma in situ) were cultivated in MEM

containing 10% NCS. For fluorescence experiments 1W
cells were incubated in 25 cm2 Falcon tissue culture dishes.
Six hours later the medium was changed to MEM with
3% NCS with either 25 pg PII/mL and/or 1 pg
3THPPhL. These concentrations were chosen on the
basis of pilot experiments, in view of optimal analysis of
the fluorescence spectra of samples containing both drugs.
At these concentrations the photosensitivity and the flu-
orescence yield were about a factor 2 larger for cells with
PI1 than for cells with 3THPP, when exposed to equal
fluence rates at the optimal wavelengths (400 nm for PI1
and 420 nm for 3THPP).
After an incubation period of 18 h with the dyes the
cells were washed five times in icecold PBS, brought into
suspension by trypsinization, washed once more with PBS
and finally suspended in PBS at a concentration of approx.
loh cellshl-as determined by Biirker chamber counting.
Irradiation of the samples. The light source was a 900 W
Osram high pressure xenon lamp fitted to a Bausch &
Lomb grating monochromator. The bandwidth of the light
was 15 nm as measured with a second monochromator
(Jarrel Ash) with narrow slits (Ah = 1 nm) and a UDT
1la detector (United Detector Technology, Santa Monica,
CA). This detector is calibrated and was also used to
determine the fluence rates at the position of the cells:
30 mW/cm2 at both wavelengths (400 and 420 nm). The
samples were irradiated in a fluorescence cuvette
(3 x 3 x 20 mm) and were gently stirred during the light
exposure.
Fluorescence measurements. Fluorescence spectra were
recorded by means of a Perkin-Elmer LS 5 spectrofluor-
imeter equipped with a Hamamatsu R928 red sensitive
photomultiplier tube. A 580 nm cut-off filter was used to
remove scattered light from the light entering the emission
slit, which was usually set to give AX = 5 nm. The distor-
tion of the spectra resulting from such slit widths were ml Loo m
taken into consideration in the evaluation of the data but Wavelength hm)
found to be of minor importance.
Using a microcuvette with a cross section of only
3 x 3 mm, inner filter effects played no significant role. Figure 1. Fluorescence excitation spectra of PI1 and
3THPP in NHIK 3025 cells 18 h incubation in MEM with
3% NCS containing 25 pm PII/mL and/or 2 pg/mL
3THPP. The dotted line in the spectrum of PI1 shows the
RESULTS AND DISCUSSION spectrum in methanol in the region 250-329 nm.
We have earlier shown that the fluorescence
quantum yields of PI1 and 3THPP in NHIK 3025 comparable to those of cell membranes but without
cells are similar to within 30% (Moan et al., 1987). proteins, it is possible to separate the energy trans-
Therefore, and since it is difficult to record the fer peak from the direct excitation of PI1 as indi-
absorption spectra of dyes in cells with accuracy, we cated by the dotted line on the spectrum for PI1
decided to base our calculations on the fluorescence (Fig. 1, Moan et al., 1988). In the case of 3THPP
measurements. In fact, it is more relevant to use such a separation was not attempted, since this dye
fluorescence spectra than absorption spectra for has a peak in its own excitation spectrum in this
investigations of this type, since the aggregated frac- wavelength region due to the phenyl rings. One can
tion of the dyes in the cells is both non-fluorescent easily separate the spectra of samples containing
and photochemically inactive. This has been shown both dyes (Fig. l ) , both manually and by means of
by means of absorption, fluorescence and action a simple computer program.
spectroscopy for both dyes considered in the present To a first approximation and for exposure times
work (Western and Moan, 1988; Moan and shorter than 6 min, the decay kinetics of the Soret
Sommer, 1984). band is of first order for both dyes (Figs. 2 and 3).
Figure 1 shows the fluorescence excitation spectra The same is true for the energy transfer band at
of cells with PII, 3THPP and with the mixture of 290 nm (Fig. 2). However, the rate constant for the
the two dyes. There is a peak in the spectra at about decay of the energy transfer band is larger than that
290 nm, which, in the case of PII, is entirely due to for the decay of the Soret band. Thus, the decay of
energy transfer from proteins containing aromatic the energy transfer band is due to at least two
amino acids to porphyrin molecules (Moan et al., factors: the photodegradation of PI1 and the
1988). Thus, using the known shape of the excitation destruction of binding sites at proteins. Since the
spectra of PI1 in liposomes or lipids with properties rate constant for the decay of the energy transfer
 Research Note
Figure 2. The decay of the Soret band of PI1 in NHIK
3025 cells monitored with the excitation and emission
wavelengths set at 400 and 625 nm, respectively; and of
the energy transfer band (Acxc = 290 nm, Acm = 625 nm),
the contribution from direct excitation being subtracted.
The wavelength of the exposure light was 400 nm.
band is dependent on the concentration of PI1 in
the cells, the binding sites at proteins can be
degraded by photoproducts generated remotely
from these binding sites (Moan ef al., 1988).
From Fig. 3 rate constants for the first part of
the decay of the Soret band of the dyes can be
determined. In the calculations we have taken into
account only the two first exposure times (i.e. 3 and
Figure 3. Decay curves for the Soret bands of PI1 and
3THPP separately or simultaneouslypresent in NHIK 3025
cells exposed to light at either 400 or 420 nm. Incubation
conditions, see legend of Fig. 1.
55 1
6 min) as indicated by the regression lines drawn in
Fig. 3.
Table 1 shows that the quantum yield of photo-
degradation of PI1 by light quanta absorbed by PI1
itself is comparable to the corresponding yield for
3THPP or slightly lower. However, the relative
value of the quantum yield of photodegradation of
3THPP caused by quanta absorbed by 3THPP is
3-6 times larger than the relative value for photo-
degradation of the dye caused by quanta absorbed
by PII. (Only the fluorescent and photosensitizing
fraction of the dyes are considered). Similarly, the
relative value of the quantum yield of photo-
degradation of PI1 caused by light absorption in PI1
is more than twice as large as the relative value
for photodegradation caused by light absorption in
3THPP. This is in agreement with our earlier con-
clusion that the photodegradation of porphyrins in
cells is predominantly a first order process (Mang
et al., 1987). Therefore, the present results indicate
that a photoproduct, like lo2,generated by the
absorption of a quantum of light in a porphyrin
molecule causes damage mainly to that molecule
and not to other porphyrin molecules in the vicinity.
The average intracellular concentrations of 3THPP
can be estimated to 40 p M under the present con-
ditions, and that of PI1 can be estimated to 80 p M
-
of porphyrin rings (mol. wt 600) (Moan ef al.,
1987). In these estimations it is assumed that the
amount of dye in the nucleus is negligible (Moan er
al., 1989). The nucleus constitutes about 30% of
the volume of these cells (Moan and Boye, 1981).
Thus, spheres of radius 0.02 pm located randomly
in the cytoplasm contain on the average 1 fluor-
escent molecule of 3THPP. It is, of course, a very
crude approximation that the dye molecules are
homogeneously distributed in the cytoplasm, but it
may serve in a first approximative calculation of
the distance travelled by the reactive intermediates
generated by the photoexcitation of the dye mole-
cules. If we assume that the membranes constitute
about 10% of the cytoplasmic volume (White et
al., 1978) and that the present lipophilic dyes are
localized mainly in membrane structures, their local
concentrations are a factor of 10 larger than esti-
mated above and the radius of a sphere containing
on average 1 fluorescent molecule of 3THPP is
about 0.01 pm. Singlet oxygen (lo2) is almost cer-
tainly a reactive intermediate generated under the
present conditions (Moan et al., 1987). The lifetime
rA of lo2under the present conditions can be esti-
mated by use of the formula 6 = ( 6 0 T~), where
6 is the distance diffused by lo2 before it is
quenched and D is the diffusion coefficient of 0 2 .
If we assume that D = 1.4 x cm2 s-l (Moan
and Boye, 1981) and 6 is within the range
0.01-0.02 p , rA is within the range 0.01-0.04 ps.
Such a short lifetime is consistent with the lack of
a D 2 0 effect in many photosensitized processes in
which lo2is very likely to be involved. Using the
552 JOHAN MOANand KRISTIAN
BERG

Table l(a). Photodegradation of 3THPP

~ ~~~~~~

Concentrations hCxp
WmL) (nm) A(3THPP) A(P1I) k(min-') @(3THPP33THPP) Q(3THPP.PII)

3THPP PI1
1 0 400 0.2 0 0.014 0.07
1 0 420 1.0 0 0.10 0.10
1 25 400 0.2 2 0.07 0.022
1 25 420 I .0 2 0. 14 0.016
~ ~~~~~

Table l(b). Photodegradation of PI1

Concentrations Lip
()lLg/mL) (nm) A(PI1) A(3THPP) k(min-') @(PII,PII) (D(P11.3THPP)

PI1 3THPP
25 0 400 2.0 0 0. 13 0.M5
25 0 420 2.0 n 0.11 0.055
25 1 400 2.0 0.2 0.13
25 1 420 2.0 1 .0 0.11 < 0.03'
~~ ~~~ ~~

Aexp is the wavelength of the photodegrading light.

A(PI1) and A(3THPP) are relative numbers of light quanta absorbed by fluorescing and photoactive molecules
of PI1 and 3THPP in the samples, approximated in relative values by the convolution integrals of the
fluorescence excitation spectra and the spectra of the photodegrading light at 400 and 420 nm rcspcctively,
with AA = 15 nm. In this approximation it is assumed that the fluorescence quantum yield of the
fluorescing molecules of 3THPP in the cells is similar to that of fluorescing molecules of PII, which is
suggested by our earlier work (Moan et al., 1988). @(3THPP, 3THPP) is the yield of degradation of
3THPP resulting from absorption of light by the fluorescent fraction of 3THPP. Correspondingly.
@(3THPP. PII) is the yield of degradation of 3THPP resulting from light absorption by thc tluorcsccnt
fraction of PI1 in the cells. @(PII, PII) and @(PII, 3THPP) are defined analoguously. The yields arc
given in relative values.
"0.03 is the upper limit of @(PII, 3THPP) since the maximal error in the relative valucs of k (thc rate
constants for photodegradation of the dyes as estimated by the fluorescence experiments) is 30% in ihc
present data as estimated from two parallel experimental series.

value T,, = 0.04 ps water will account for only 1%
of the quenching of '02in a cell and no D 2 0 effect
can be expected. This is consistent with experiments
indicating that T A < 0.6 ps in cells (Firey et al.,
1988). Thus, the time resolution of experiments
intended to determine in cells directly would
have to be improved by more than a factor of 10
from what can be achieved at present.
It should be noted that detection of the 1272 nm
lo2phosphorescence is probably the only way to
prove that is generated in a cell, since scavenger
experiments can always be questioned because of
inhomogeneity of the dye- and scavenger molecules
and reactions of most scavengers with other possible
intermediates. The possibility that scavenger mole-
cules and sensitizer molecules are present in differ-
ent micro-compartments of the cells should always
be considered, even in the case that both types of
molecules have a similar lipophilicity. Since most
efficient scavenger molecules are only weakly flu-
orescent, it is difficult to study their interacellular
localization pattern. The use of two different lipo-
philic, fluorescent and photosensitizing dyes may
give valuable information about the processes of
photosensitization in cells as indicated by the pre-
sent results. The conclusion that the main reactive
intermediated diffuse of the order of 0.01-0.02 pm
in the cells is in agreement with our observation
that photoexcitation of a porphyrin present at or
outside the cell wall of an Escherichia cofi cell,
whose thickness is approx. 0.03 pm, does not result
in any photodamage to DNA in the bacteria (Boye
and Moan, 1980) and that uroporphyrin, which
produces with a high quantum yield (0.7-0.8)
when photoexcited (Blum and Grossweiner, 1985)
is inefficient in sensitizing cells when present in
the medium outside the cells during light exposure
(Nelson et al., 1986; Madslien, K., unpublished
data).
Acknowledgement-The authors want to express their
thanks to professor Claude Rimington, F.R.S. for valuable
comments and advice during the preparation of the manu-
script.
REFERENCES
Blum, A. and L. I. Grossweiner (1985) Singlet oxygen
generation by hematoporphyrin IX.uroporphyrin I and
hematoporphyrin derivative at 546 nm in phosphate
buffer and in the presence of egg phosphatidylcholine
liposomes. Phorochem. Photobiol. 41, 27-32.
Cox, G. C., C. Bobillier and D. G. Whitten (1982) Photo-
oxidation and singlet oxygen sensitization by protopor-
phyrin IX and its photooxidation products. Phorochem,
Phorobiol. 36, 401-407.
Dougherty, T. J. (1987) Photosensitizers: Therapy and
detection of malignant tumors. Photochem. Photobiol.
45, 879-889.
 Research Note
Fiery, P. A , , T. W. Jones, G. Jori and M. A. J. Rodgers
(1988) Photoexcitation of zinc phthalocyanine in mouse
myeloma cells: the observation of triplet states but not
of singlet oxygen. Photochem. Photobiol. 48, 357-360.
Krieg, M. and D. Whitten (1984) Self-sensitized photo-
oxidation of protoporphyrin IX and related porphyrins
in erythrocyte ghosts and microemulsions: A novel
photo-oxidation pathway involving singlet oxygen. J .
Photochem. 25, 235-252.
Mang, T. S., T. J. Dougherty, W. R. Potter, D. G. Boyle,
S. Sommer and J. Moan (1987) Photobleaching of por-
phyrins used in photodynamic therapy and implications
for therapy. Photochem. Photobiol. 45, 501-506.
Mang, T. S., B. D. Wilson and S. Kahn (1990) An evalu-
ation of the role of photobleaching concepts of Photofrin
I1 in photodynamic therapy to human clinical trials.
Photochem. Photobiol. 51s. 72s.
Moan, J. (1986) Effect of bleaching of porphyrin sensitiz-
ers during photodynamic therapy. Cancer Lett. 33,
45-53.
Moan, J., K. Berg, E. Kvam, A. Western, Z. Malik,
A. Ruck and H. Schnekenburger, (1989). Intracellular
localization of photosensitizers. In Photosensitizing
Compounds: Their Chemistry, Biology and Clinical Use.
(Edited by S . Hernet), pp. 95-111. Ciba Foundation
Symp. No. 146. Wiley, New York.
Moan, J. and E. Boye (1981) Photodynamic effect on
DNA and cell survival of human cells sensitized by
hematoporphyrin. Photobiochem. Photobiophys. 2 ,
301-307.
553
Moan, J., T. Christensen and S. Sommer (1982) The
main photosensitizing components of hematoporphyrin
derivative. Cancer Lett. 15, 161-166.
Moan, J., 0.Peng, J. F. Evensen, K. Berg, A. Western
and C. Rimington (1987) Photosensitizing efficiencies,
tumor- and cellular uptake of different photosensitizing
drugs relevant for photodynamic therapy of cancer. Pho-
tochem. Phorobiol. 46, 713-721.
Moan, J., C. Rimington and Z. Malik (1988) Photo-
induced degradation and modification of Photofrin I1 in
cells in vitro. Photochem. Photobiol. 47, 363-367.
Moan, J. and S . Sommer (1983) Uptake of the components
of hematoporphyrin derivative by cells and tumors. Can-
cer Lett. 21, 167-174.
Moan, J. and S . Sommer (1984) Action spectra for hema-
toporphyrin derivative and photofrin I1 with respect
to sensitization of cells in vitro to photoinactivation.
Photochem. Photobiol. 40,631-634.
Nelson, J. S., C. H. C. Sun and M. W. Berm (1988) Study
of the in vivo and in virro photosensitizing capabilities
of Uroporphyrin I compared to Photofrin 11. Lasers
Surg. Med. 6 , 131-136.
Western, A. and J. Moan (1988) Action spectra for photo-
inactivation of cells in the presence of tetra (3-
hydrooyphenyl) porphyrin, chlorine e,, and aluminium
phthalocyanine tetrasulfonate. In Light in Biology and
Medicine (Edited by R. H. Douglas, J. Moan and F.
Dall'Acqua), vol. 1, pp. 5689. Plenum, New York.
White, A. P. Handler, E. L. Smith, R. L. Hill and 1.
R. Lehman (1978). Principles of Biochemistry, p 301.
McGraw-Hill Kogakusha.