Laser Phys. Lett. 8, No. 3, 239246 (2011) / DOI 10.1002/lapl.

201010119 239
Abstract: We present experimental evidence of the existence
of cell variability in terms of threshold light dose for Hep G2
(liver cancer cells) cultured. Using a theoretical model to de-
scribe the effects caused by successive photodynamic therapy
(PDT) sessions, and based on the consequences of a partial
response we introduce the threshold dose distribution concept
within a tumor. The experimental model consists in a stack of
asks, and simulates subsequent layers of a tissue exposed to
PDT application. The result indicates that cells from the same
culture could respond in different ways to similar PDT induced-
damages. Moreover, the consequence is a partial killing of the
cells submitted to PDT, and the death fraction decreased at each
in vitro PDT session. To demonstrate the occurrence of cell
population modication as a response to PDT, we constructed
a simple theoretical model and assumed that the threshold dose
distribution for a cell population of a tumor is represented by a
modied Gaussian distribution.
Microscopic image (100) of a necrosis caused by PDT in a
rat liver. The edge is the depth of necrosis
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
Experimental evidence and model explanation for cell
population characteristics modication when applying
sequential photodynamic therapy
L.G. Sabino,
1
L.M.V. de Negreiros,
1
J.D. Vollet-Filho,
1
J. Ferreira,
2
D.P.C. Tirapelli,
3
P.C. Novais,
3
L.F. Tirapelli,
3
C. Kurachi,
1
and V.S. Bagnato
1,
1
Instituto de Fsica de S ao Carlos, Universidade de S ao Paulo, 13566-590, S ao Carlos, S ao Paulo, Brazil
2
Instituto de Pesquisa e Desenvolvimento, Universidade Vale do Paraba, 12244-000, S ao Jos e dos Campos, S ao Paulo, Brazil
3
Departamento de Cirurgia e Anatomia da Faculdade de Medicina de Ribeir ao Preto, Universidade de S ao Paulo, 14049900, Ribeir ao
Preto, S ao Paulo, Brazil
Received: 15 October 2010, Revised: 29 October 2010, Accepted: 1 November 2010
Published online: 25 January 2011
Key words: photodynamic therapy; theoretical model; cancer cells variability; Hep G2 cell culture; threshold dose distribution
1. Introduction
Photodynamic therapy (PDT) is an attractive treatment op-
tion for many types of cancer. Since the 1980s, skin can-
cer, head and neck cancer, brain tumors, and others have
been treated by PDT with satisfactory results. Many stud-
ies propose adequate protocols to improve PDT outcome
and present positive results in the treatment of supercial
neoplastic lesions. At the same time, several clinical cases
have reported cancer recurrence after PDT application, and
in some cases an increase in tumor resistance to posterior
PDT applications are related [14]. This has also been ob-
served by us in many clinical applications. A reasonable
hypothesis is that the surviving cells are more resistant to
PDT, for reasons not yet understood. It is possible to infer
that PDT applications after the rst one would be less ef-
cient; resulting in a regrowth of the tumor more difcult
to be treated [5].
PDT consists of a photosensitizer (PS) administration
followed by the illumination of the tumor with light at an

Corresponding author: e-mail: vander@[Link]

c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
240 L.G. Sabino, L.M.V. de Negreiros, et al. : Evidence for cell population characteristics
appropriate wavelength. The light excitation leads to the
generation of triplet excited photosensitizer (PS

), which
interacts with and transforms the molecular oxygen from
its ground state into a metastable state, generating singlet
oxygen. Such reactive oxygen species are powerful oxida-
tive molecules responsible for cell machinery destruction
[612]. In order to kill the tumor cells, a minimum damage
must be induced. In this context, an important concept was
introduced by L.I. Grossweiner et al., the threshold light
dose (D
th
), which is the minimal light dose able to pro-
duce an irreversible cell injury and subsequent cell death
[7].
Microscopic images of PDT treated normal liver show
a sharp edge between the necrotic region and the non-
necrotic region [13,14]. This observation associated with
a model of light-tissue interaction shows that cell death
depends exclusively on the amount of cell toxicity, gen-
erated by the local photoproducts. Inside the tissue, when
the photoproducts exceed a minimum amount (generated
by reaching a threshold light dose in the tissue) necrosis
is achieved. The existence of a well dened line of necro-
sis implies well determined threshold dose, as previously
analyzed by J. Ferreira et al. and by M.S. Patterson et al.
[15,16]. Contrary to healthy tissues, where a high degree
of homogeneity is expected, the cell population within a
tumor is very distinct, showing a broader cell distribution.
Instead of a homogeneous cells response, a large variance
on neoplastic cell response to PDT is expected for those
tissues [1720]. This may result in differences in cell re-
sponse to PDT and consequent variations in the nal out-
come [2124]. In this case, it is possible that tumor char-
acteristics may change after successive PDT applications,
due to cell selectivity based on their values of D
th
. The
variation of tumor cells in terms of D
th
has not been con-
sidered for the PDT outcome.
An interesting observation by Y. Shimada et al. [25]
reported the survival of 31 cultured human esophageal tu-
mor cells using a colony-forming assay after PDT; they
found that each cell group formed multiple morphologi-
cally distinct colonies in a single dish. In addition, it has
been reported that the onset of chemo-resistance in tu-
mor cells might be associated with changes in the cy-
toskeleton components [26,27]. Previous reports showed
that PDT strongly affects cell-cell and cell-substrate adhe-
sion, which consequently causes changes in the cytoskele-
ton [26,27].
In the present study, we rst demonstrated, using a set
up constructed to simulate subsequent layers of a human
tissue, an idea presented as a theoretical model to examine
a hypothetical tumor response under successive PDT ap-
plications, when a distribution of values for threshold light
doses is veried in the tissue. The cells of this tumor model
are considered to have a non-negligible distribution of the
threshold. The consequence of this distribution is that only
a part of the cells are killed, with further consequences af-
ter the PDT. This hypothesis is a way to understand the
possible limitations of PDT, and to develop new interven-
tions for the improvement of PDT technique when multi-
ple applications are required, leading to a more effective
treatment of tumors and a long-term survival of patients.
2. Materials and methods
Experiment was made with a multilayer distribution of cell
cultures of Hep G2.
2.1. Photosensitizer and light source
A rst test to observe evidence of PDT variations of out-
come during a sequential realization of illumination was
performed using a cell culture arrangement. The photo-
sensitizer used was Photogem

(PG Photogem, Rus-

sia) at 1.5 m/ml in Dulbeccos Modied Eagles Medium
(DMEM, GIBCO

invitrogen) supplemented with 10%

calf serum at 37

C in a humidied atmosphere with 5%

CO
2
.
The asks containing the cell culture were illumi-
nated using a LED-based device at 63010 nm, called
BioTable

. The irradiance of the illumination surface used

in all experiments was 405 mW/cm
2
.
2.2. Hep G2 cancer cell culture and treatment
In this sequence of experiments we used a hepatocarci-
noma cell line (ATCC, HEPG2/2.2.1). Cells were seeded
in 25 cm
2
asks (Techno Plastic Products, TPP, Trasadin-
gen. Swiss) containing 5 ml of Dulbeccos minimum es-
sential medium (DMEM) supplemented with 10% of fetal
calf serum (FCS Cultilab

, Campinas, Brazil), and in-

cubated until reaching the cultures conuence. We have
placed a stack of asks with the purpose to simulate the
light penetration through many layers of tissue.
Five 25 cm
2
asks of HepG2 in conuence were used
in this experiment. PG in a concentration of 1.5 g/ml of
media and completed a volume of 5 ml with DMEM, with-
out FCS was used.
After drug incubation of 4 hours, the cell cultures were
washed again, a fresh medium added and the illumination
started. All ve asks were positioned in the LED-based
device (Fig. 1). The light uence at the rst stack ask
was of 364 J/cm
2
, decreasing as a function of distance
of each ask to the illumination surface. After irradiation,
the cell cultures were incubated at 37

C for 2 hours, and

then the cell counting and cell viability evaluation was
performed. The cells were detached by trypsinization and
counted in a Newbauer chamber with trypan blue. Surviv-
ing cells were seeded in 5 new asks, and incubated un-
til the conuence to form the new generation of cells, the
second generation. This procedure was repeated for 5 cell
generations.
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
[Link]
Laser Phys. Lett. 8, No. 3 (2011) 241
Flask
4
3
2
5
1
LED-based device
Light fluence,
J/cm
2
Hep G2
cultured
(a)
(b)
Figure 1 (online color at [Link]) (a) schematic dia-
gram showing the arrangement of the illumination of the asks
containing Hep G2 cell cultures. The goal is to stack the asks to
simulate several slices of a tissue. The graph shown is an illustra-
tion of the light uence decay as a function of the distance of the
illumination surface. (b) image showing the asks during illu-
mination by the LED-based device. The light uence at the rst
stack ask was of 364 J/cm
2
, decreasing at each ask stacked
3. Results
It is not clear that all cells in the same tissue could present
similar values of threshold concerning PDT. The PDT
damage depends on the light uence, PS concentration and
presence of molecular oxygen. In this way, the parameters
used in this sequence of experiments were chosen to sim-
ulate a partial PDT effect caused by inferior light uence.
An analysis of the result in the graph in Fig. 2 is an in-
dicator that cells from the same culture could respond in
different ways to similar PDT induced-damages. The rea-
son is not yet understood, but the consequence is a partial
killing of the cells submitted to PDT, as shown in the fol-
lowing graph.
The rst PDT application caused 50% of death in
ask 1, while the death cell fraction decreased gradually
D
e
a
t
h

f
r
a
c
t
i
o
n
,

%
50
40
30
20
10
0
5 4 3 2 1
Cell culture flask
Treatment 1
Treatment 2
Treatment 3
Treatment 4
Treatment 5
Figure 2 (online color at [Link]) Death cell fraction
of 5 asks containing HepG2 cultured submitted to 5 succes-
sive PDT application using Photogem

and a LED-based device

(630 nm). The count of the death fraction was performed using
a Newbauer chamber and trypan blue, 2 hours after the illumina-
tion
for ask 2 to ask 5. In this case, we noted the death frac-
tion dependence on light uence. As the amount of light
applied decreases, the cell death decreases. Using only
light or only PS, the death fraction resulting was inferior to
10%, which conrms that the cell death results in all asks
shown in Fig. 2 occurred primarily due to the PDT effect.
After that, using the same parameters, the second PDT
application resulted in 40% of death fraction in ask 1. A
decrease of the death fraction for ask 2 to ask 5 was
observed. In addition, the result observed for the second
PDT was inferior for all asks compared to the result of
the rst PDT. It means that somehow the cell population
was selected by the PDT, which created cultures present-
ing increased resistance to PDT.
Finally, the third, the fourth and the fth PDT applica-
tions resulted in values of death fraction similar to the val-
ues obtained in the control groups (only drug without light
and only light). We cannot afrm those cells increased
their resistance to PDT effect. However the cell popula-
tion characteristics changed, decreasing the death fraction
at each cycle of PDT. The explanation for this experimen-
tal observation is the selection of the high resistant cells by
PDT. Those cells presenting low resistance to the photody-
namic effect were killed and eliminated from the culture.
On the other hand, more resistant cells survived and the
culture regrowth with different characteristics occurred.
Selection of more resistant cells occurs in a population
presenting diversity in terms of threshold. If those cells
were equal, the PDT response using the same parameters
would be similar and reproducible after each application.
In this study, a cell population presenting a natural
broad threshold distribution is described as an extreme
case to illustrate the fact that the nal outcome of a se-
[Link]
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
242 L.G. Sabino, L.M.V. de Negreiros, et al. : Evidence for cell population characteristics
Figure 3 (online color at [Link]) Microscopic image
(100) of a necrosis caused by PDT in a rat liver. The edge is
the depth of necrosis
quential PDT application, with intermediate regrowth of
the tumor, may vary due to cell selection during the pro-
cess. This is strongly dependent on the light dose used.
This effect would be minimized if a lower threshold dose
would be characteristic for all cells.
4. Theoretical model
In order to better visualize the idea of a broad threshold
distribution and its effects on PDT outcome, we have or-
ganized a hypothetical model for the system. In the PDT
action for a homogeneous threshold system, cell death oc-
curs since the amount of toxic photoproducts exceeds a
minimum value necessary to produce an irreversible dam-
age to the cell. Supposing a constant and sufcient pho-
tosensitizer and oxygen concentration, the cell death will
depend only on the local light dose. In an animal model
using a normal liver, a homogeneous distribution of D
th
is observed, where there is a well-dened deep edge be-
tween the necrotic region and the normal region. If we
dene g(D
th
) as the threshold dose distribution of a cell
population of a normal tissue, in this case g(D
th
) is well
represented by a Dirac function, described as:
g(D
th
) = A

D(z) D
th

, (1)
where A is a constant, D(z) is the local light at z depth,
and D
th
is the threshold dose of the tissue before the rst
PDT session. Integration of Eq. (1) in the interval [,]
produces the identity connected to the probabilistic inter-
pretation of g(D
th
):

g(D
th
)dD
th
= 1. (2)
If we assume a homogeneous distribution of PS and
oxygen in tissue, the death fraction f(D(z)), caused by
PDT, as a function of light dose should be described as:
f

D(z)

0, D(z) < D
th
,
1, D(z) D
th
.
(3)
This means that cells receiving a dose above D
th
will
be killed while cells receiving a dose below D
th
will sur-
vive. In general, the model of light interaction with biolog-
ical tissues introduced by L.I. Grossweiner et al. is based
on the Beer Lambert law for the light attenuation, and it
depends on the optical properties of the tissues and on the
light dose. In this approximation the light dose at each po-
sition D(z) [7] is
D(z) = D
0
exp(z) , (4)
where D
0
is the supercial dose and represents the effec-
tive attenuation coefcient. In this case, for a generic dis-
tribution of threshold g(D
th
), after illumination the death
fraction is a function of the position and can be determined
through an integration of the distribution function
f(z) =
D(z)

0
g(D
th
)dD
th
, (5)
which has the following interpretation: at some position z,
where the light dose is D(z), the cells with D
th
<D(z)
will be killed. Based on Eq. (5), the depth where the light
dose reaches D
th
is
z
th
=
1

ln
D
0
D
th
,
and the death fraction should follow a spatial distribution
dened by a delta function distribution:
f(z) =

1, z < z
0
th
,
0, z z
0
th
.
(6)
where
z
0
th
=
1

ln
D
0
D
0
th
.
Cells within the extension going from surface (z =0)
and z
th
are killed. These conditions imply in a well-
dened depth boundaries between the necrotic and the nor-
mal cell region, as observed in experiments for normal
tissues Fig. 3. Hence, successive PDT applications could
cause equal necrosis in terms of depth and tissue charac-
teristics. This simple light dosimetry model has been used
with good acceptance in cases of tumor tissue, without any
explanation or certication of validity. However, there are
many reports that showed the cell variability in a tumor in
terms of the resistance to PDT, so that the interest in study-
ing the effect of this variability in the PDT outcome is well
justied [27,28].
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
[Link]
Laser Phys. Lett. 8, No. 3 (2011) 243
A
m
o
u
n
t

o
f

c
e
l
l
s
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
3.5 3.0 2.5 2.0 1.5 1.0 0.5 0
Threshold dose, sigma units
Figure 4 (online color at [Link]) Modied Gaussian
distribution shape; amount of cells versus the threshold dose
(D
th
). This kind of distribution has a standard deviation and a
mean value that characterize the threshold dose of the tumor cell
population
In the PDT action in a heterogeneous threshold system
the cellular response would be quite different. It is well
known that cells in a tumor tissue are heterogeneous in
terms of their morphology and differentiation status, even
though the tumor mass may be derived from a single neo-
plastic cell [25,29]. This heterogeneity suggests that the
tumor-cell population, in the theoretical model, could be
represented by a distribution similar to a Gaussian. Taking
into account that no cells die by PDT without light, in a
tumor volume there are a certain number of cells present-
ing a specic threshold dose which can be represented by
a modied Gaussian distribution:
g(D
th
) = AD
2
th
exp

D
2
th

. (7)
In Eq. (7) D
th
represents the threshold light dose val-
ues, is the standard deviation characterizing the width of
the distribution, and Ais a constant that depends on the tis-
sue type. Fig. 4 shows the distribution shape. The amount
of cells/volume is given by the integral of the function
g(D
th
), and the fraction of cells with a threshold dose be-
tween (D
th
) and (D
th
+dD
th
) can be calculated through
the ratio R
R =
g(D
th
)dD
th

0
g(D
th
)dD
th
. (8)
The normalized g(D
th
) is obtained using principles of
the probability distribution [30], determining the constant
A

0
g(D
th
)dd
th
= 1 (9)
and the D
th
distribution function is rewritten in a normal-
ized form as
g(D
th
) =
4

3
D
2
th
exp

D
2
th

. (10)
Furthermore, the average of the cells distribution (D
th
)
can be derived by using the probability distribution theory
by integration of Eq. (10),
D
th
=

0
D
th
g(D
th
)dD
th
resulting in
D
th
=
2

. (11)
The average value for the threshold dose D
th
is a
macroscopic quantication for the resistance of the tumor
to PDT. Higher D
th
implies a higher level of light neces-
sary to promote cell death.
The resulting death fraction after the rst PDT applica-
tion was simulated using some assumptions. For simplic-
ity, the theoretical model only considers the behavior of
the threshold dose at the z axis, as shown in Fig. 4. Here,
we assumed a full and homogeneous PS distribution in tu-
mor cells, while PS is absent of healthy cells; therefore,
the PS was owed 100% of cellular selectivity. Hence, at a
position z, cells that present D
th
<D(z) will die. On the
other hand, cells with D
th
>D(z) will survive and prolif-
erate; recovering the tissue, but with a different character-
istic concerning the threshold dose distribution.
Before the rst light application, we assume the pop-
ulation density (number of cells/volume) and the distri-
bution of threshold (D
th
) to be the same at any depth z.
Therefore, D(z) determines the death fraction for each po-
sition. After illumination, cells will die, and the death frac-
tion at each position is determined by
f
1
(z) =
D(z)

0
g(D
th
)dD
th
(12)
with g(D
th
) as in Eq. (10). The solution of Eq. (12) is
f
1
(z) =
4

2
exp

4
erf()

, (13)
where =D(z)/.
The death rate function f
1
(z) represents the percent-
age of dead cells after illumination, as a function of the
depth z. The quantity f
1
(z) as a function of z is shown in
Fig. 4, when different supercial light doses are considered
in the simulation. Because of the distribution of threshold
the death cell fraction will also be spatially dependent.
[Link]
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
244 L.G. Sabino, L.M.V. de Negreiros, et al. : Evidence for cell population characteristics
D
0
z= 0 z
Figure 5 (online color at [Link]) Supercial illumina-
tion of the tumor tissue model; the depth is represented by the z
axis; D
0
is the supercial dose homogeneously applied
After the rst PDT, part of the original neoplastic cells
survives and tumor tissue regrowth will take place, assum-
ing the re-grown tumor to preserve the geometry of the ini-
tial tumor model. The biological characteristics, consider-
ing the cell population, of the re-grown tumor will be dis-
tinct from the original one. The new threshold distribution
is now spatially dependent, which is a property not present
before the rst application. The new distribution function
for the threshold dose will now be spatially dependent and
given by:
g
(1)
(D
th
, z) = (14)
=

0 , D
th
D(z) ,
A

D
2
th
exp

D
2
th

2
(z)

D
th
> D(z) .
Indeed, this threshold distribution function for the tu-
mor leads to an interpretation that a small subsets of cells
have D
th
towards innity, in this case, survival cells will
exist after the PDT session. The average for the survival
cell population (D
(1)
th
(z)) can be calculated by Eq. (15):
D
(1)
th
(z) =

0
D
th
g
(1)
(D
th
)dD
th

0
g
(1)
(D
th
)dD
th
(15)
and it is, however, a function of the depth. The previous
equation can be rewritten as
D
(1)
th
(z) = (16)
=

0
D
th
g(D
th
)dD
th

D(z)

0
D
th
g(D
th
)dD
th

0
g(D
th
)dD
th

D(z)

0
g(D
th
)dD
th
and, after applying Eq. (9), Eq. (11), and Eq. (12), we ob-
tain:
D
(1)
th
(z) =
2

I(z)
1 f
1
(z)
, (17)
where
D(z)

0
D
th
g(D
th
)dD
th
= I(z)
is the average of the dead fraction cells after the rst PDT
application, which is once again determined by using the
distribution probability theory [30]. The difference, how-
ever, is that average and standard deviation are now func-
tions of the depth z
f
2
(D
th
, z) =
D(z)

0
g
(1)
(D
th
)dD
th
. (18)
The solution for Eq. (18) determines the death fraction
f
2
(z) for the second PDT application, as follows:
f
2
(z) =
4

2
exp

4
erf()

, (19)
where =D(z)/
1
(z) and
1
(z) is the standard variation
function, determined by using the implications of Eq. (5)

1
(z) =

2
D
(1)
th
(z) . (20)
The growth of D
th
(z) implies a macroscopic increase
for the tumor resistance to further PDT application.
This model is still quite far from reality, but it serves
well to illustrate the ideas presented in this study. As ob-
served from the graphics generated by the model, the over-
all behavior observed obeys the general lines of the exper-
iments.
5. Discussion
The existence of more than 100 types of neoplastic le-
sions in many stages of growth represents an important
challenge to develop models to predict PDT response. In
this way, the concept of a threshold light dose was created
to classify the resistance of specic types of cancer cells
and to evaluate reliable parameters. However, the concept
of threshold does not consider the existence of a variety
of cells in the same tissue. The results of our experiment
show that cells derived from a single culture are different
in terms of threshold light dose.
Several studies have produced the tumor eradication by
multiple sessions of PDT. However, in a fewcases the PDT
is not efcient enough to completely eliminate some can-
cer types, which then cause the cancer recurrence. As an
example, a recent application by our group in cholangio-
carcinoma has clearly produced a tumor harder to respond
to PDT as applications followed. We have analyzed if the
cell variety could be involved in the tumor resistance to
PDT. As obtained through this simple model, the results of
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
[Link]
Laser Phys. Lett. 8, No. 3 (2011) 245
D
e
a
t
h

f
r
a
c
t
i
o
n
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
3.5 3.0 2.5 2.0 1.5 1.0 0.5 0
Depth, mm
Dose 1
Dose 2
Dose 3
Dose 4
Dose 5
Figure 6 (online color at [Link]) Simulation of the
death fraction in function of the depth after 1

PDT applica-
tion using doses: Dose 1 =1.0, Dose 2 =2.0, Dose 3 =3/0,
Dose 4 =4.0, and Dose 5 =5/0. To simplify the dose is ex-
pressed in sigma units (). These curves were obtained using
MatLab

. The death fraction is 1 to 100% cell death and 0 (zero)

to 100% cell survival
the rst PDT application are plotted as a function of depth
in Fig. 6 for different used supercial doses. To simplify
we consider the dose in units of . Particularly, the light
dose equal to 1.0 do not even cause complete damage
in the tumor surface. The result suggests that in order to
achieve 100% of tumor cells killing in a xed depth, dam-
age should be induced beyond this depth. Otherwise, the
cancer will not be eradicated, and tumor regrowth will oc-
cur. The remaining fraction of cells resistant to PDT might
help to promote a modication of the tumor characteris-
tics, and after a re-grow process the remaining tumor bulk
would present a different distribution of D
th
, and would
therefore be more resistant to PDT. This resistance may
not be associated with an increase of the tumors aggres-
siveness, but only with its response to PDT.
This idea of modifying tumor characteristics is already
established for others modalities of treatment. It is a well
known fact that recurrent tumors are usually more resistant
to further chemotherapy or radiotherapy [31]. Equally, we
had also observed this increase in resistance after multi-
session PDT treatment of bile duct cancer. It is possible
that light and/or PS dosimetry might contribute to this un-
desirable effect, as well as the heterogeneity of the tumor
tissue [26].
In our simulation, the rst PDT application caused tis-
sue damage and changes in its characteristics. For tumor
regions where the death rate was below 100%, resistant
cells survived, and a new tumor re-grew from them. A rea-
sonable assumption is that those surviving cells are more
resistant to PDT; since the PDT left only those cells alive,
it selects more resistant cells. Hence, future PDT applica-
tions would be less efcient at each new section [5,26].
The theoretical model describes the new threshold dose
S
i
g
m
a

f
u
n
c
t
i
o
n
30
25
20
15
10
5
0
1.5 1.0 0.5 0
Depth, mm
Dose 1
Dose 2
Dose 3
Figure 7 (online color at [Link]) Sigma function ((z))
after rst PDT application, using Dose 1 =1.0, Dose 2 =1.3,
and Dose 3 =1.6
D
e
a
t
h

f
r
a
c
t
i
o
n
0.1
0
4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0
Depth, mm
Dose 1
Dose 2
Dose 3
Dose 4
Figure 8 (online color at [Link]) Simulation of the
death rate fraction after 2

PDT application for Dose 1 =1.0,

Dose 2 =2.0, Dose 3 =3.0, and Dose 4 =4.0. The 2

PDT-
application was performed repeating the same doses of 1

PDT-
application
distribution for the re-grown tumor by probability distri-
butions as a function of depth (
1
(z)). Fig. 7 shows the
evolution for the main characteristic of the threshold dis-
tribution (width of the Gaussian on position) as a function
of depth for three different PDT applications using doses
equal to 1.0, 1.3, and 1.6, as described in Fig. 7. The
main conclusion from these results is the fact that differ-
ent light doses may cause different changes in tumor resis-
tance after subsequent PDT. Higher doses select the more
resistant cells to survive, producing a considerably more
resistant tumor.
Moreover, the cells of the tissue surface became more
resistant than the cells inside the tumor. These results
are consistent with common observations in the literature
[Link]
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
246 L.G. Sabino, L.M.V. de Negreiros, et al. : Evidence for cell population characteristics
[5,28], and highlight the relevance to perform more ac-
curate dosimetry studies. Death rate simulations after the
second PDT application using the same light doses re-
sulted in less tissue damage, as PDT was unable to in-
duce total necrosis even at the surface. Plots shown in
Fig. 8 reveal that a second generation of tumor cells is
more resistant than the rst one. According to the theoreti-
cal model, insufcient illumination can also result in more
resistant tumors and perhaps even with higher aggressive-
ness. The successive sub-threshold dose for PDT applica-
tions can enhance those effects; and should, however, be
done with necessary care. We are presently preparing an
animal model to test the basis of the model here presented.
6. Conclusion
We have studied the effects of subsequent PDT applica-
tions when a tumor presents a distribution for the values
of threshold doses instead of a single value. The result of
the sequential photodynamic therapy using Hep G2 cells is
an indicative of the existence of the threshold dose distri-
bution. Increasing resistance to sequential PDT is a conse-
quence of the selectivity resulting from this broad distribu-
tion. There are many open questions that still require more
investigation and the theoretical model still needs to be
conrmed by experimental models. Further improvements
in the clinical applications of PDT for heterogeneous tu-
mor cells can be expected. Because of the characteristics
of tumor cells in general, it seems realistic to expect a dis-
tribution of the threshold and effects like the ones pointed
out in this paper. Finally the main idea here presented is
not a cell modication as it takes place during the sequen-
tial PDT. Rather, what we have is selectivity of more resis-
tant cells with the overall result of a regrowing tumor with
different characteristics.
Acknowledgements This study was supported by the Brazil-
ian National Board for Scientic and Technological Develop-
ment (Conselho Nacional de Desenvolvimento Cientco e Tec-
nol ogico, CNPq) and the S ao Paulo State Research Supporting
Fund (Fundac ao de Amparo ` a Pesquisa do Estado de S ao Paulo,
FAPESP/CEPID Program). There are no potential conicts of in-
terests in the performed study.
References
[1] F. Stewart, P. Baas, and W. Star, Radiother. Oncol. 48, 233
(1998).
[2] F. Calzavara, L. Tomio, L. Corti, P.L. Zorat, I. Barone, A.
Peracchia, L. Norberto, R.F. DArcais, and F. Berti, J. Pho-
tochem. Photobiol. B 6, 167 (1990).
[3] M. Schaffer, L. Corti, H. Hollenhorst, C. Boso, P.M. Schaf-
fer, M. Busch, and E. D uhmke, Eur. J. Cancer 33, S29
(1997).
[4] A. Maier, F. Tomaselli, F. Gebhard, P. Rehak, J. Smolle,
and F.M. Smolle-J uttner, Ann. Thorac. Surg. 69, 1006
(2000).
[5] P.G. Calzavara-Pinton, J. Photochem. Photobiol. B 29, 53
(1995).
[6] T.J. Dougherty, B.W. Henderson, C.J. Gomer, G. Jori, D.
Kessel, M. Korbelik, J. Moan, and Q. Peng, J. Natl. Cancer
Inst. 90, 889 (1998).
[7] L.I. Grossweiner, J. Photochem. Photobiol. B 38, 258
(1997).
[8] M. Atif, M. Fakhar-e-Alam, S. Firdous, S.S.Z. Zaidi, R.
Suleman, and M. Ikram, Laser Phys. Lett. 7, 757 (2010).
[9] C.S. Xu, A.W.N. Leung, L. Liu, and X.S. Xia, Laser Phys.
Lett. 7, 544 (2010).
[10] M. Atif, S. Firdous, A. Khurshid, L. Noreen, S.S.Z. Zaidi,
and M. Ikram, Laser Phys. Lett. 6, 886 (2009).
[11] M.C. Milanetto, H. Imasato, and J.R. Perussi, Laser Phys.
Lett. 6, 611 (2009).
[12] Y. Tan, C.S. Xu, X.S. Xia, H.P. Yu, D.Q. Bai, Y. He, and
A.W.N. Leung, Laser Phys. Lett. 6, 321 (2009).
[13] A. Khursid, M. Atif, S. Firdous, S.S.Z. Zaidi, R. Salman,
and M. Ikram, Laser Phys. 20, 1673 (2010).
[14] J. Ferreira, P.F.C. Menezes, C.H. Sibata, R.R. Allison, S.
Zucoloto, O. Castro e Silva, and V.S. Bagnato, Laser Phys.
19, 1932 (2009).
[15] J. Ferreira, L.T. Moriyama, C. Kurachi, C. Sibata, O. Cas-
tro e Silva, Jr., S. Zucoloto, and V.S. Bagnato, Laser Phys.
Lett. 4, 469 (2007).
[16] M.S. Patterson, B.C. Wilson, and R. Graff, Photochem.
Photobiol. 51, 343 (1990).
[17] D. Mitton and R. Ackroyd, Photochem. Photobiol. Sci. 3,
839 (2004).
[18] Z. Huang, Technol. Cancer Res. Treat. 4, 283 (2005).
[19] A.K. DCruz, M.H. Robinson, and M.A. Biel, Head Neck
26, 232 (2004).
[20] M.A. Biel, Photochem. Photobiol. 83, 1063 (2007).
[21] J. Ferreira, C. Kurachi, C.A.S. Melo, L.T. Moriyama, S.
Zucoloto, and V.S. Bagnato, Laser Phys. 14, 209 (2004).
[22] G. Singh, M. Espiritu, X.Y. Shen, J.G. Hanlon, and A.J.
Rainbow, Photochem. Photobiol. 73, 651 (2001).
[23] C.F. Oliveira, F.G. Basso, E.C. Lins, C. Kurachi, J.
Hebling, V.S. Bagnato, and C.A. de Souza Costa, Laser
Phys. 20, 1659 (2010).
[24] D.Q. Bai, C.M.N. Yow, Y. Tan, E.S.M. Chu, and C.S. Xu,
Laser Phys. 20, 544 (2010).
[25] Y. Shimada, M. Maeda, G. Watanabe, S. Yamasaki, I. Ko-
moto, J. Kaganoi, T. Kan, Y. Hashimoto, I. Imoto, J. In-
azawa, and M. Imamura, Clin. Cancer Res. 9, 243 (2003).
[26] A. Casas, G. Di Venosa, S. Vanzulli, C. Perotti, L.
Mamome, L. Rodriguez, M. Simian, A. Juarranz, O. Pon-
tiggia, T. Hasan, and A. Batlle, Cancer Lett. 271, 342
(2008).
[27] A. Casas, F. Sanz-Rodriguez, G. Di Venosa, L. Rodriguez,
L. Mamone, A. Bl azquez, P. Ja en, A. Batlle, J.C. Stockert,
and A. Juarranz, Cancer Lett. 270, 56 (2008).
[28] T. Kushibiki, M. Sakai, and K. Awazu, Cancer Lett. 268,
244 (2008).
[29] P.C. Nowell, Science 194, 23 (1976).
[30] E. Butkov, Mathematical Physics, 1st ed. (Addison-
Wesley, Reading, 1968).
[31] [Link].M. Capella and L.S. Capella, J. Biomed. Sci. 10, 361
(2003).
c 2011 by Astro Ltd.
Published exclusively by WILEY-VCH Verlag GmbH & Co. KGaA
[Link]