-
tu
rk, PriJ 0847d Stgy, A
anthederimt caconsistent with temperature measurements obtained
from the experimental
of deatsecond only to cardiovascular diseases [1]. Bas
from cancer in 2015 and 11.4 million in 2030. At this rate,
itmay surpass have shown that heat affects nuclear function through
the inhibition
Materials Science and Engineering C 32 (2012) 22422249
Contents lists available at SciVerse ScienceDirect
Materials Science a
.e lcardiovascular disease as the leading cause of death by 2012
[2].Cancer is a generic term for a large group of diseases that can
affect
any part of the body. Current scientic evidence suggests that
cancercan be triggered by environmental and genetic factors [3].
Regardlessof the trigger, unless it is diagnosed in the early
stages, the prognosisfor patients is often poor [46]. Current
treatment modalities include:radiotherapy, chemotherapy, hormonal
therapy and surgical removal[1]. When possible, surgical removal,
in combination with other treat-ment modalities, often offers the
best prognosis for patients [7].
However, common treatment modes such as radiotherapy
andchemotherapy are known to induce multiple side effects that
canhave long-lasting impact on a patient's quality of life [8] and
hormonal
of RNA [1215], DNA [16,17], and protein synthesis [1618]. In
addition,hyperthermia causes delay or arrest in cell cycle
progression [19]; chief-ly through mitotic arrest [2023] and
inhibiting S phase entry from G2[24,25]. Groups have also reported
reduced cell metabolism followinghyperthermia [18,26,27].
Additionally, cells up-regulate heat shock pro-teins (HSPs) in
response to heat treatment immediately following aprior heat
treatment. This up-regulation of HSPs leads to a
well-knownphenomenon, known as thermo-tolerance [11]. Consequently,
ndingsfrom prior work justify investigating new pulsed heating
regimentsthat would allow sufcient time for the thermo resistance
to subsidewhile taking advantages of cell cycle disruptions
[8,13,14].
In recent times, the rst ofcial clinical use of hyperthermia was
in
therapy is only available to patients with cerConsequently,
there has been increasing intea treatmentmodality because it has
minimal
Corresponding author at: Princeton Institute for Scirials
(PRISM), Princeton University, Princeton, NJ 0854258 5609; fax: +1
609 258 5877.
E-mail address: [email protected] (W.O. Sobo
0928-4931/$ see front matter 2012 Elsevier B.V.
Alldoi:10.1016/j.msec.2012.06.010ed on projections, cancerd 9
million people dying
treatment for cancer is based on its direct effect on cells.
Heat causes acellular stress which triggers a cascade of molecular
events. Studiesdeaths will continue to rise with an
estimateHyperthermiaBreast cancer cellsCell cytoskeletonHeat shock
proteinsLocalized cancer treatment
1. Introduction
Cancer is the second leading causeis shown to cause signicant
physical changes in the cell cytoskeleton. Finally, the paper
explores the effectsof hyperthermia on cell growth and cell death
under isothermal and cyclic conditions. The underlying effectsof
heat shock protein expression are elucidated before discussing the
implications of the results for cancertreatment via localized
hyperthermia.
2012 Elsevier B.V. All rights reserved.
h throughout the world;
synergistic effects when used in combination with radiotherapy
andchemotherapy [911].
The biological rationale for the use of hyperthermia as a
potentialKeywords: set-ups. The paper also examines the effects of
isothermal heating on the cell morphology. Isothermal
heatingAvailable online 16 June 2012temperature variations areAn
in-vitro study of the effects of temperaand models
C. Theriault a,b, E. Paetzell a,b, R. Chandrasekar c, C. Baa
Princeton Institute for Science and Technology of Materials
(PRISM), Princeton Universityb Department of Mechanical and
Aerospace Engineering, Princeton University, Princeton, Nc
Department of Electrical and Computer Engineering, Purdue
University, West Lafayette, INd Department of Mechanical
Engineering, University of Delaware, Newark, DE 19716, Unitee
Department of Materials Science & Engineering, African
University of Science & Technolo
a b s t r a c ta r t i c l e i n f o
Article history:Received 1 March 2011Received in revised form 30
April 2012Accepted 11 June 2012
This paper presents an implthermia. Heating, accompliswhich are
tested under expto 45 C as studied on breas
j ourna l homepage: wwwtain types of cancer [9].rest in
hyperthermia asside effects and potential
ence and Technology of Mate-4, United States. Tel.: +1 609
yejo).
rights reserved.re on breast cancer cells: Experiments
ey d, Y. Oni a,b, W.O. Soboyejo a,b,e,nceton, NJ 08544, United
States544, United States907, United Statesatesbuja, Nigeria
able biomedical device for the localized killing of cancer cells
through hyper-via resistive heating, is modeled using numerical
heat transfer techniques,ental conditions. The effect of
temperature in the therapeutic domain of 37ncer cell line
MDA-MB-231 is also reported. The results show the predicted
nd Engineering C
sev ie r .com/ locate /msecthe early part of the 20th century,
when it was used as a treatment forcervical cancer [11]. However,
it was not until the 1970s that the mod-ern discipline of
thermotherapy really emerged beyond the regime ofexperimentation
[2830]. Nevertheless, due to limits in technologicaladvances, very
few clinical studies were performed before the 1990s.However, by
the turn of the 21st century, there was a renewed interestin
hyperthermia research and clinical applications in local and
regionalhyperthermia [3134].
-
Similar to other treatment modalities, hyperthermia's clinical
ob-jective is to achieve the localized death of tumorigenic tissue
withoutdamaging the surrounding normal tissue. In contrast to bulk
systemictreatments, thermotherapy can be administered locally,
regionally, orsystemically [11,3537]. In any case, it is possible
to achieve a treat-ment that is locally administered at tumorigenic
sites with minimaldamage to surrounding tissues.
Within the last decade, researchers have developed multiple
deliverymodalities for hyperthermia and tested them in both invitro
and invivoconditions as previously described [12,17,38,39]. The
most common isthe use of ferromagnetic uids in combination with an
oscillating mag-netic eld [4042]. Magnetic rods or needles have
also been exploredas excitable sources of heat [43,44]. Other
radio-frequency deliverymodalities that have beendeveloped
includehigh intensity focusedultra-sound and high-frequency eddy
currents. Furthermore, the use of intra-peritoneal surgeries as
treatment modalities for both hyperthermia
applied modalities, the new device induces hyperthermia by the
cyclicapplication of heat to an implantable device. Short heat
pulses (15 min
2.2. Design and testing of a biocompatible implant for inducing
hyperthermiausing resistive heating
The hyperthermia implant was designed to have two largesurfaces
that are implanted perpendicular to the tumor growth axis.A 10102
mm geometry was used in order to efciently distributethe resistive
heating element within the implant (Fig. 1). When con-nected to a
standard AA battery and in dry air conditions, the
implant'stemperature increased from room temperature (23 C) to 125
C in50 s.
To simulate the potential effect of in-vivo implantation, a
similartest was run in a 50 mL test tube lled with 10 mL of aqueous
solu-tion. The test tube was immersed in a water bath at 37 C and
allowedto equilibrate before turning on the hyperthermia inducing
device.The results showed that it took an average of 4 min for the
device toincrease the temperature at the interface of the water
bath and thesample to 41 C under static conditions (Fig. 2).
cell growth, leading to cell cycle arrest and eventually
inducing cell
2243C. Theriault et al. / Materials Science and Engineering C 32
(2012) 22422249in duration) at 45 C for 45 min are followed by a 12
h relaxation period.The in-vitro responses of breast cancer cells
are examined and the mea-sured temperatures are compared with
predictions from a nite differ-ence model. The implications of the
results are discussed along withsuggestions for future work.
2. Device design and fabrication
2.1. Device fabrication
The device that was used to apply heat shock treatments to
thecell samples consisted of 100 m thick insulated copper wire
embed-ded in PDMS (Elastomer Sylgard 184 from Corning with 1:10
curingagent:PDMS mass ratio). Approximately 0.8 m of wire was
wrappedaround a PDMS block with dimensions 110.2 cm. The woundblock
was then embedded in a PDMS package that was cured at80 C for 2.5
h. The wire was wound uniformly within the PDMS, asshown
schematically in Fig. 1.
Fig. 1. Schematic of the hyperthermia inducing device. The
hyperthermia inducing de-vice was fabricated out of isolated copper
wire (d=0.2 mm) and PDMS. The copperwire was rst wound to cover an
area of 1 cm2 and then placed in a mold. Freshly pre-pared PDMS was
then poured into the mold and allowed to cure for 2.5 h at 80 C.
Onceand/or high-dose chemotherapy is becomingmore prevalent in the
treat-ment of cancers of the gastrointestinal tract and female or
male internalreproductive organ cancers [45]. Apart from the
peritoneal surgeries,there is a lack of post-treatment follow-ups
and many physicians haveexpressed concerns about the long-lasting
effects of hyperthermic-inducing modalities [5].
This paper presents the results of a combined experimental and
com-putational study of the effects of temperature on the structure
and deathof breast cancer cells. Heating is achieved via an
implantable device thatuses resistive heating to induce cell death
in the surrounding cells/tissue.The device was fabricated from
polydimethylsiloxane(PDMS), an FDA-approved biocompatible polymer
[46]. Unlike the current externally-un-molded, the device had the
following dimensions: 110.2 cm.death after multiple cycles.
3. Experimental procedures
3.1. Cell culture experiments
All the cell culture experiments performed in this study were
con-ducted on the breast cancer cell line MDA-MB-231 (ATCC,
ManassasVA). The cells were grown at 37 C in a humidied environment
withatmospheric CO2 levels. They were grown in Leibovitz's L-15
mediumsupplemented with 10% fetal bovine serum and 2% penicillin.
The cellswere seeded for 12 h prior to the rst heat shock
treatmentwith an ini-tial conuence of ~50%, resulting in a ~70%
conuence sample at thestart of cyclic treatment. For all cell
counting samples, the sampleswere counted and then seeded 12 h
prior to heat shock treatment.
Fig. 2. Simulation of in-vivo implantation. Hyperthermia
inducing device was im-mersed in 10 mL of DiH2O at 37 C and was
connected to a 1.5 V AA battery for15 min. The average temperature
of the 10 mL sample of DiH2O was recorded. Resultsshowed a rapid
ramp-up followed by a period of equilibrium in the thermal
diffusion.For the nal experimental conditions, a voltage of 1.5 V
was appliedto the wire and the temperature within the samples was
maintainedusing an Omega CN7533 Proportional Integral Differential
(PID) con-troller. Temperature was measured using T thermocouples
(Omega,Stamford, CT). In this way, hyperthermic heating was applied
for45 min, followed by cooling to 37 C (relaxation) to simulate
actual hy-perthermic cycles.
Since current hyperthermia regimens include heat exposure
from45min to 2 h with relaxation periods lasting from 3 days to 2
weeks[3537], the pulsed-hyperthermia regimen presented here
introducesthe idea of frequent short-duration heat exposures, more
specicallypulse durations of 45 min followed by 12 h relaxation
periods. The ra-tionale behind this new regimen system would be to
induce multiplecytoskeletal reorganizations due to short heat
pulses that should inhibitThe power to the device was turned off
after 15 min.
-
2244 C. Theriault et al. / Materials Science and Engineering C
32 (2012) 224222493.2. Cell exposure to continuous hyperthermia
The effects of continuous hyperthermia treatments were
assessedusing clonogenic and trypan blue exclusion assays. First, a
trypan blueexclusion assay was used to analyze cellular growth
under hyperther-mic conditions. Briey, multiple samples of
MDA-MB-231 breast cancercell linewere cultured. Each sample set
contained aminimumof four cellculture asks. Once the sample sets
had reached ~70% conuence, theywere exposed to one of four
experimental conditions for up to 72 h.The four experimental
conditions were: (i) 37 C; (ii) 41 C; (iii) 43 C;and (iv) 45 C. The
number of cells in each sample was counted at leastfour times using
a hemocytometer and a trypan blue exclusion assay[47]. This was
done at regular intervals of 24 h.
In parallel, clonogenic assays were conducted to assess the
cellviability/colony-forming ability of the cells immediately after
treatment[48]. MDA-MB-231 cells were grown at 37 C until they
reached ~70%conuence. The cells were then placed under constant
temperaturesof 37, 41, 43, or 45 C for up to 72 h. The cells were
harvested immedi-ately after 24, 48 and 72 h (or 6 and 12 h for 45
C) of heat treatment,and counted using a hemocytometer.
Subsequently, the cells were plat-ed into 6-well petri dishes, in
order to plate 150 cells/well. They werethen allowed to grow into
colonies for 710 days at 37 C, before xingthemwith a 6%
glutaraldehyde and 0.5% crystal violet mixture. The col-onieswere
then countedwith a colony counter pen. All the experimentswere
performed in triplicates, and the results are reported as
percent-ages of the colonies (% of control).
Insight into the effectiveness of continuous hyperthermia on
cellkilling was given by the trypan blue exclusion assay, while
cell repro-ductive death/viability was characterized with the
clonogenic assays.Overall, trends in cell numbers were elucidated
through both assays.
3.3. Cyclic heat shock procedure
A cyclic heat shock procedurewas used to stimulate the potential
ef-fects of multiple heat shock therapy. Each cycle of treatment
consistedof 45 min of exposure to hyperthermic temperatures with
in- andbetween-cycle temperatures of 45 and 373 C, respectively.
Thiswas followed by 12 h of incubation. During the heat exposure, a
tem-perature gradient was achieved within the sample, with the
center ofthe dish at 4548 C and the edge of the dish at 3941 C, as
discussedabove. The experiments were carried out for 5, 10 and 20
cycles. Thecell populations were then analyzed using both
hemocytometry andpropidium iodine staining.
3.4. Propidium iodine assay for both cell death and apoptotic
bodyformation
Propidium iodine (PI) assayswere used to determine the amount
andlocation of cell death within a sample. Cell death was assumed
to be afunction of the local temperature, and thus a guide for the
heat diffusion.PI, at a concentration of 1 mg/mL in
L15mediumsupplementedwith 10%of FBS, was added directly to the
samples, after a 12 h incubation periodfollowing the last heat
treatment. Subsequently, the samples were incu-bated for 15 min,
allowing for the PI to stain and label the nuclei of thedead cells.
The cells were then xed with 3.7% formaldehyde for15 min. Finally,
the samples were observed and imaged using a NikonEclipse 50i with
a medium band blue excitation lter (Tokyo Japan).
3.5. Cell cytoskeleton imaging
The cells were grown on glass cover slips for 24 h and then
exposedto heat shock in a circulating water bath for 15, 30, 45 and
60 min. Thesamples were then rinsed in PBS, and xed in 3.7%
formaldehyde solu-tion in PBS for 15 min. After washing, the cells
were stained and incu-bated with DRAQ5 (Biostatus Limited,
Shepshed, Leicestershire, UK)
for nuclear staining for 5 min. The samples were then stained
andincubated with Oregon Green 488 paclitaxel (Invitrogen,
Carlsbad, CA)for tubulin labeling for 1 h. Subsequently, the cells
were rinsed withPBS and labeled for actin with FITC-conjugated
phalloidin (Sigma, St.Louis, MO) for 20 min. After several washes
with PBS, the sampleswere mounted on slides using a mounting
medium, Aqua Poly/Mount(Polysciences, Warrington, PA). A RS3
Spinning Disk Confocal micro-scope (Perkin Elmer, Waltham MA) with
a 60 objective was thenused to examine the immuno-uorescence of the
cytoskeleton proteins.
3.6. Effects of temperature (37 C, 41 C, 43 C, and 45 C) on heat
shockprotein expression
To assess the effects of cyclic heating on heat shock protein
expres-sion, 2.5 millionMDA-MB-231 cellswere cultured into 6-well
plates at aconcentration of 0.5105 cells per mL. The cells were
then grown for24 h at 37 C and fed fresh medium prior to heat shock
treatment. Foreach heat shock treatment, cultured plate was heated
at either 41 C,43 C or 45 C for 30 min. This was accomplished by
direct immersionin a circulatingwater bath. The uniformbath
temperatures serve as con-trol experiments to simulate the effects
of constant temperatures. Thetemperatures in the medium were
monitored with a PID controller. Inall cases, the culture plates
reached the desired temperatures within5 min. A 6-well plate was
kept in the incubator at 37 C to representthe control group. Medium
was harvested from duplicate wells fromeach plate immediately after
treatment and 12 h after treatment. Thisprocedurewas repeated every
12 h for 5 cycles to investigate the effectsof cyclic heating.
The number of cells was counted at each temperature to adjust
theresults of the ELISA by the viable cell number. Two separate
wellswere used exclusively for counting (to avoid collecting any
viable cellsin the ELISA assay). These wells were harvested every
48 h and themean cell count and viability were determined by direct
cell count ina hemocytometer using the trypan blue exclusion
assay.
To determine the concentration of inducible HSP70 in cells, an
ELISAassay was conducted following the manufacturer's instructions
(R&DSystems, Minneapolis, MN). Briey, a standard curve was
createdusing HSP70 standard solution. 100 L of standards and
samples wasadded to a 96-well plate and incubated at room
temperature for 2 h.Eachwell waswashedwith awash buffer three times
and then incubat-ed with 100 L mAb against HSP70 at a concentration
of 0.25 g/mL for2 h at room temperature. Wells were again washed
three times, and100 L of Streptavidin-HRP (1:200 dilution from
stock solution in buff-ered protein base) was added and incubated
for 20 min at room tem-perature. Wells were washed three times and
100 L of substratesolution (1:1 ratio of hydrogen peroxide to
tetramethylbenzidine)was added and incubated for 20 min at room
temperature. Then, the re-action was stopped with 50 L of 2 N
sulfuric acid, and the optical den-sity at 450 nm was measured
using an ELISA plate reader.
3.7. Statistical analysis
All statistical analyses were performed using S-PLUS (Ver.
7.0,TIBCO Software Inc., Palo Alto, Ca). A one-sided student's
t-test wasused to determine the statistical signicance between two
samplesmeasuring the same variable. When comparing multiple
treatmentswith multiple factors, a multiple sample analysis of
variance (ANOVA)test was performed using Tukey's method. A p-value
lower than 0.05was considered a signicant difference, and condence
intervals weremade using an alpha of 0.05.
4. Modeling
4.1. Numerical modeling of heat transfer
A nite difference model was written in MATLAB (The MathWorks
Inc., Natick, MA) to calculate the heat diffusion prole from the
device
-
to the cell culture medium. The model used Fourier's equation
and anite difference scheme to predict the temperature gradient
withina plane of the implant. Briey, a 25 cm2 square mesh was
constructedand divided uniformly into 10,000 nodes. An initial
condition of 37 Cwas imposed on the entire array at t=0. Four
boundary conditionswere considered: two in the x, two in the
y-axis.
The boundaries were the contour of the implants and of the
sys-tem. By assuming that the distance to the boundary of the
systemwas much larger than the size of the implant, heat diffusion
in the fur-thest boundary region becomes negligible, allowing the
coil boundarytemperatures to be set to 37 C. The steady-state
temperature of eachnode was then calculated using central
difference and forward differ-ence approximations of solutions to
Fourier's laws.
4.2. Two-dimensional analysis of the heat diffusion from the
hyperthermiabiocompatible device into a uid lled environment
This expression can be rearranged to solve for the temperature
atnode Ti,j at the time n+1. This gives:
Tn1i;j Tni;j tTni1;j Tni1;j Tni;j1 Tni;j14Tni;j
2
!3
where =x=y for a square mesh.This second order partial
differential equation in Eq. (3) requires
that there must be 4 boundary conditions (two in the x
directionand two in the y direction) as well as 1 initial condition
at t=0. Therst heat treatment for each sample was started after an
incubationperiod of about 12 h, and therefore, the initial
temperature of the en-tire sample can be set at 37 C. Assuming a
small temperature gradi-ent between the implant and the initial
temperature and largeimplant-to-boundary distance, all nodes on the
outer edges of thegrid can also be set at an incubation temperature
of 37 C.
5. Results
5.1. Isothermal hyperthermia
The results of the in-vitro continuous hyperthermia
experimentsare presented in Fig. 4(a) and (b). Both gures show cell
culturedata that was obtained fromMDA-MB-231 breast cancer cell
incubat-ed for up to 72 h at constant hypothermic conditions. Fig.
4(a) showsthe growth rate of this particular cell line at 37 C, as
well as the se-vere decrease in growth, as the temperature
increased. Necrosis is ob-served above 43 C, which is similarly
reported for other cell lines
-50%N
2245C. Theriault et al. / Materials Science and Engineering C 32
(2012) 22422249To determine the temperature gradient in the area
surrounding theimplant, the region was rst divided into an evenly
distributed squaremesh. The grid consisted of 100 nodes dened along
both x and y-axes, for an overall distribution of 10,000 total
nodes. Each node wasrepresented by an area of x by y, as shown in
Fig. 3.
The implant with square cross-sectional area of 1 cm2 was
consid-ered to occupy 2020 nodes in the center of the array. The
nodes oc-cupied by the implant were given initial conditions
ranging from 45to 55 C, while those outside of the system were set
at 37 C. TheMATLAB programwas then used to move forward in time,
performingiterations until steady-state temperature conditions were
reached.
An energy balance was performed on each node using the
nitedifference method (implementing the explicit method) in
conjunc-tion with the linear heat diffusion equation. The
derivation for eachtemperature equation is governed by a
two-dimensional form ofFourier's law of conduction. This is given
by:
d2Tdx2
d2T
dy2 1
kg 1
dTdt
1
where g is dened as the heat generation term, and k and are
thethermal conductivity and thermal diffusivity of the medium,
respec-tively. Neglecting the heat generation term and applying a
forward-time, central difference discretization for an interior
node with noheat generation yields:
Tni1;j2Tni;j Tni1;jx2
Tni;j12Tni;j Tni;j1
y2 1Tn1i;j Tni;j
t2
Fig. 3. Graphical representation of algorithm used for modeling
temperature gradient.A square grid pattern was used to model the
heat diffusion in static conditions. These
conditions were very similar to the actual in-vitro testing
conditions.0
25
50
75
100
-100%
0 24 48 72Num
ber o
f Col
onie
s (%
of C
ontro
l)
Time (Hours)
Time (hr)
41C
43C
45C
43C45C
b
Fig. 4. Effect of different temperatures of continuous
hyperthermia on MDA-MB-231Cell Population. Breast cancer cells
MDA-MB-231 were incubated continuously at ei-ther 37, 41, 43, or 45
C for 72 h or until the population collapsed. Results are
shown[4952]. Treatment at 43 C is shown to be the optimal
temperaturefor hyperthermia treatments, which is consistent with
reports in theliterature [53]. Fig. 4(b) shows similar trends for
colony growth.
120 24 36 48 60 72
100%
50%
0%
orm
aliz
ed C
ell G
row
th
37C
41C
ausing a) trypan blue exclusion assay and b) clonogenic
assay.
-
5.2. Comparison of measured and predicted temperatures proles
ofdevice mediated hyperthermia
In-vitro experimental observation showed that steady-state
tem-perature conditions were achieved in a very short period of
time(less than 4 min). It was, therefore, determined that the
transient re-sponse was not very signicant. Consequently, only the
steady-statetemperature distributions were examined. Typical
results from thetemperature modeling are presented in Fig. 5(a).
These show thatthe temperatures decrease with increasing distance
from the centerof the device. The modeling results were supported
by the in-vitro ex-periments. These were found to be in close
agreement with thepredicted temperatures from the MATLAB
simulation, as shown inFig. 5(b).
5.3. Staining of cell death and cell morphology under
isothermalhyperthermia
The MDA-MB-231 cells were cultured in 60 mm Petri dishes
(BD,Franklin Lakes, NJ) for 12 h and then isothermally heated (~2
h) toevaluate the effects of the heat treatment on the cell
populations.Using our implantable device, the center of the sample
was broughtto 55 C, while the outside edge remained below 41 C.
Visual inspec-tion of PI staining of these samples showed the
presence of both ne-crotic and apoptotic cell death, with necrosis
prevalent in the region
a
40
50
60
70
80
90
100
44
46
48
50
52
54
2246 C. Theriault et al. / Materials Science and Engineering C
32 (2012) 22422249Fig. 5. Temperature variation radiating from
hyperthermia implant. Results are shownfrom (a) modeling the
thermal diffusion from the hyperthermia device into its
sur-roundings using MATLAB. The implant is set at 55 C and the
outside extremitieswere set to 37 C (scale: 10 unit=0.5 cm). The
results from the model (blue) are com-10 20 30 40 50 60 70 80 90
100
10
20
30
38
40
42
bpared with the average experimental measurements (red) in
(b).of high-hyperthermia (>45 C) and the apoptotic cell death in
thelower-temperature regime (Fig. 6). This result is consistent
with pre-vious studies linking necrotic cell death with
temperatures in theupper-hyperthermic range and apoptotic cell
death with tempera-tures in the lower-hyperthermic range
[11,52,54,55].
Confocal microscopy images of cell samples exposed to 15, 30,
45and 60 min of hyperthermia at 43 C are presented in Fig. 7. The
sam-ples were xed and stained for cytoskeletal structures including
actin,microtubules, and nucleus. No signicant difference was
observed be-tween the cytoskeletons of the control and the samples
exposed for15 min at 43 C. The most signicant changes were observed
in themicrotubule network. Specically, after 30 min of exposure,
changesin the microtubule networks were observed. Some condensation
ofthe actin cytoskeleton was also observed. After 45 min of
exposure,both the microtubules and the actin network appeared to be
affected.Furthermore, aggregation of the microtubule network was
observedat 45 and 60 min. Actin networks were more disorganized
after60 min, compared to those that were observed after 45 min.
Nuclear changes were present in approximately one third of
thesamples exposed for 60 min. For those that did exhibit some
nuclearchanges, mitotic body formation could be observed;
presumably indi-cating that these cells had entered apoptosis.
These results clearlyshow that after 30 min of exposure to 43 C,
the cells started to expe-rience signicant deterioration of their
cytoskeleton, which contin-ued with increased heat exposure.
5.4. Cyclic hyperthermia
The results of the cyclic hyperthermia experiments are presented
inFig. 8(a) and (b). There is a statistically signicant difference
betweenthe control and hyperthermia treatment at all cycles
(pb5105). Al-though there is a decline in cell population after 5
hyperthermia treat-ment cycles, statistical analysis revealed that
it was not signicant. Infact, the analysis showed that there was no
signicant difference be-tween the mean number of cells at the
beginning of the experimentsand after 5 or 10 cycles. However,
after 20 cycles, a statistically signi-cant decrease was observed
in the cell populations, when comparedto the other treatment
cycles.
5.5. In-vitro cyclic heating and heat shock protein
expression
The PI staining showed clearly that cell death propagated
radiallyoutwards from the device after cyclic hyperthermia (Fig.
8(b)). Simi-lar trends were observed in the cell death proles after
isothermalheating at 43 C (Fig. 6). The most prominent form of cell
deathseems to be apoptosis, as demonstrated by the numerous
apoptotic-bodies present in the samples. These results also show
that at highertemperatures, fewer cycles are needed to induce cell
death. More-over, the results also suggest that cell-cycle arrest
was achieved with-in a large range of temperatures, ranging from 41
to 47 C. The resultsshow that heat shock protein expression
increases with increasingtemperature between 41 and 47 C (Fig.
9).
Prior work has shown that heat shock proteins are proteins that
areover-expressed during hyperthermia. In most tumor cells, they
are al-ready over-expressed when compared to the basal level in
normal cells[52,5658]. Recent work suggests that HSPs have the
ability to serve ascarriers of tumor antigens and inammatory
agents. Specically, multi-ple studies have shown that HSP70
interacts with receptors on antigenpresenting cells and can mediate
T-lymphocytes to trigger an immuneresponse against the cells that
are expressing HSP70 [57,59,60]. Amajor determinant of immune
response is the interaction between theantigen-presenting cells and
the T lymphocytes [61] due to its vasodila-tation properties. The
current results (Fig. 9) show that the extent of HSPexpression
increases with increasing temperature. The above interac-tions are,
therefore, likely to increase with increasing temperature, for
temperatures between 41 and 47 C.
-
6. Discussion
6.1. Morphological changes in the cell cytoskeleton as a
potential cause
cycle arrest, MDA-MB-231 cells start to enter the apoptotic
pathway,most likely as a direct consequence of the cell cycle
arrest or of mitot-ic block. Although more work is needed, our
immuno-urescence re-
Fig. 6. Propidium iodine staining of 2 h continuous temperature
exposure with the hyperthermia device. MDA-MB-231 cells were
exposed to propidium iodine (PI) and xed. DNA uo-rescent
labelingwas assayedusing 488 nmexcitation light source. Shown above
are images from three regions of interest: the center (A),
theposition (B), and the edge (C). Scale bars inbottom right corner
correspond to 100 m; positions A, B, and C were approximately 0.5,
2, and 4 cm from the edge of the hyperthermia device. A1, B1, and
C1 show the xed sample asviewedunder regular lightmicroscopy. All
three images show a similar cell distribution pattern, indicating a
homogeneous cell distribution. A2, B2, and C2 show the PI stain
indicating theamount of cells with a ruptured nuclear membrane in
each region (red labeling). A clear radial pattern emerges from the
pictures with higher concentration of stained nucleus, a markerfor
cellular death, towards the middle of the sample.
2247C. Theriault et al. / Materials Science and Engineering C 32
(2012) 22422249for both the disturbance of cellular function and
for the cell's thermo-sensitive properties
Prior research has shown that heat shocks can have negative
ef-fects on the cell cycle, often causing a temporary arrest that
can lastfrom 2 to 14 h [52,62]. Our results suggest that, after 5
days of cellFig. 7. Cytoskeleton of MDA-MB-231 cells exposed to
hyperthermia for different amount of tand 60 min. No signicant
differences could be observed in the samples that were
exposedsignicant aggregation of the microtubule network could be
observed which increased wit30 min of exposure. At 60 min no clear
actin network organization could be observed in thabove) was only
observed in about one third of cells after 60 min of exposure.sults
do support the hypothesis of mitotic arrest as the primarytrigger
for apoptosis in our cell populations (Fig. 7).
The above observations are consistent with recent
publications[32]. These suggest that heat induces conformational
changes in themicrotubules structures, or in another regulatory
protein, hence in-terfering with normal polymerization mechanisms
and blocking theime. Confocal microscopy was performed on samples
exposed to 43 C for 0, 15, 30, 45,to 43 C for 15 min compared to
the control. After 30 min of hyperthermia exposure,h continued
exposure. Differences in the actin laments could be observed
starting ate majority of samples. Nuclear deterioration in the form
of mitotic bodies (as shown
-
2248 C. Theriault et al. / Materials Science and Engineering C
32 (2012) 22422249cell at the mitotic spindle assembly checkpoint.
Recently, Michalakiset al. [63] showed very similar results
pertaining to the microtubuleorganization of HeLa cells after 1 h
exposure to a 39 C heat shock.Their research also showed similarity
between cells treated withheat shock and tubulin inhibitor
(paclitaxel, arsenide, etc), indicatingthat mitotic arrest is the
most likely the major apoptosis pathway.
Fig. 8. Growth as a function of heat shock cycles. (a)
MDA-MB-231 cell samples wereexposed to multiple cycles of a heat
shock duration of 45 min followed by a relaxationperiod of 12 h.
Cell growth arrest was observed for the rst 10 cycles with a small
non-statistically signicant decrease in the cell population after
10 cycles. After 20 cycles ofheat shock treatment, a large
statistically signicant drop (pb5105) in the cell pop-ulation was
observed. The underlying cause of cell death was hypothesized to be
theprolonged cell cycle arrest caused by the cyclic heat shock as
45 min of exposure to43 C was shown to be non-lethal on its own.
(b) PI staining of MDA-MB-231 cellsshowed a radial pattern of cell
death: (bA) shows the cell population and (bB) showsthe cells that
have died in the population. The hyperthermia device is located at
the ex-treme right of the gure. This result was similar to the
pattern observed during contin-uous hyperthermia exposure. Scale
bar in lower right corresponds to 500 m.
0
50
100
150
200
250
T0 T12 T0 T12 T0 T12 T0 T12
HS1 HS2 HS3 HS4
Hea
t Sho
ck P
rote
in (n
g per
106 c
ells
)
Heat Shock Trial (12hr Intervals)
37C
41C
43C
45C
Fig. 9. Heat shock protein assay results. The extent of HSP
expression fromMDA-MB-231cells increases with increasing
temperature.6.2. The biocompatible hyperthermia device and its
potential for medicalapplications
Although it is clear that this device could be used for treating
a widerange of cancers, thiswork suggests that initial attention
could be paid tobreast cancer, specically in patients with stage II
or III non-metastatictumors. In such scenarios, surgical removal of
the tumor is normally pos-sible, and radiotherapy as well as low
doses of chemotherapy are stillprescribed in an effort to quench
the body of any remaining cancercells and improve the patient's
prognosis. As an alternative, the hyper-thermia device developed in
this study could be implanted during thesurgical removal of a
tumor. The device could then not only serve as astand-alone
treatment modality, but could be used in combinationwith either
radiotherapy or chemotherapy. Moreover, multiple studies[34,64]
have shown additive and even synergistic results when hyper-thermia
is combined with radiotherapy.
Prior work on internal heat and pain receptors have also shown
thatsub-dermal temperatures of up to 43 C are very well tolerated
clinical-ly, with patients feeling only a slight discomfort [65].
Furthermore, theblood perfusion rate, convection due to surrounding
arteries, and uni-form thermal values could be taken into account
using the bio-heattransfer equations [66,67]. Under such
conditions, the heat ow canbe described by a three-dimensional (3D)
heat diffusion model ratherthan the two-dimensional version
presented here. A 3D model could,therefore, be used to predict
which regions of the surrounding tissuewill reach hyperthermic
temperatures by taking into account the actualpatient-specic data
retrieved from clinical imaging techniques, such asMRI and
ultrasound.
6.3. The potential use of an inductively coupled system for
powerautonomy
In order for a hyperthermia device to be clinically efcient,
there isa need for it to achieve a certain level of power autonomy.
One poten-tial design would be to create an implant consisting of a
closed coilembedded in PDMS. By introducing an alternating magnetic
eld(AMF) in its vicinity, a current could be induced, thereby
enabling re-sistive heating. The surrounding temperature would also
have to bemonitored and controlled using wireless technology to
maintain hy-perthermic temperatures. One potential pitfall with
this system in-cludes the fact that the implant coil would have to
be oriented in aspecic direction within the body for greatest power
efciency.
In any case, the resulting current ow could then be stored in a
re-chargeable battery to provide an autonomous power source.
Furtherwork is clearly needed to demonstrate the feasibility of
such systems.These should be combinedwith efforts to develop
autonomous powersupplies for sustained use in the human body. More
advanced tem-perature controls are also warranted. Finally, future
versions of thedevice should be capable of automatically delivering
the right amountof heat at the correct intervals without human
interaction.
6.4. Potential future directions
Future work is needed to determine the response of normal
breastcells, as well as other cancer cell lines, to the effects of
heat. A recentstudy by Rylander et al. [68] showed that the HSP
expression prole ofcancerous prostate cells was more sensitive to
hyperthermia than theirnon-cancerous counterparts. In future work,
uorescence-activated cellsorting (FACS), using propidium iodine or
Annexin V, could be used toexplore the portions of cells that are
in early or late stage apoptosis, ornecrosis. Furthermore, FACS
cell cycle analyses should reveal at whatpoint hyperthermia causes
mitotic arrest. Additionally, further quanti-cation of the cellular
HSP expression prole in relation to cellular prox-imity to the
hyperthermia-inducing device could provide additional
insights.
-
Furthermore, since the cellular responses of 2D models have
beenshown to be different from 3D in-vivo models, there is a need
to ex-plore the response to heat in more complex 3D in-vitro or
in-vivomodels that simulate tumor conditions. This could be done in
3Dmicro-environments or in-vivo models that could further
investigatethe use of regular short hyperthermia pulses in the
selective killingof cancer cells. Such experiments could be
performed with our with-out the proposed hyperthermia delivery
device.
Also, although the current results do indicate that, in a simple
2Dcell culture environment, a pulsed hyperthermia regimen was
efca-
[11] R.W. Habash, et al., Crit. Rev. Biomed. Eng. 34 (6) (2006)
459489.[12] U. Heine, et al., J. Ultrastruct. Res. 34 (34) (1971)
375396.[13] R.J. Palzer, C. Heidelberger, Cancer Res. 33 (2) (1973)
422427.[14] R. Strom, et al., Eur. J. Cancer 9 (2) (1965) 103112
1973.[15] R. Warocquier, K. Scherrer, Eur. J. Biochem. 10 (2)
(1969) 362370.[16] E.M. Levine, E.B. Robbins, J. Cell. Physiol. 76
(3) (1970) 373379.[17] B. Mondovi, et al., Eur. J. Cancer 5 (2)
(1965) 137146 1969.[18] J.A. Dickson, D.M. Shah, Eur. J. Cancer 8
(5) (1965) 561564 1972.[19] J. Overgaard, Cancer 39 (6) (1977)
26372646.[20] D.P. Higheld, W.C. Dewey, M.P. David, Methods in Cell
Biology, Academic Press,
1975, pp. 85101.[21] R.J. Martin, P.R. Schloerb, Cancer Res. 24
(11 Part 1) (1964) 19972000.[22] O.S. Selawry, M.N. Goldstein, T.
McCormick, Cancer Res. 17 (8) (1957) 785791.[23] A. Westra, W.C.
Dewey, Int. J. Radiat. Biol. 19 (5) (1971) 467477.[24] J.E. Sisken,
L. Morasca, S. Kibby, Exp. Cell Res. 39 (1) (1965) 103116.[25] H.B.
Kal, M. Hateld, G.M. Hahn, Radiology 117 (1) (1975) 215217.
2249C. Theriault et al. / Materials Science and Engineering C 32
(2012) 22422249micro-environments can be simulated using gels and
agars to mimictissue-like structures. The relationships between
cell viability and dis-tance from the device should also be modeled
quantitatively [69]after heat exposures in such 3D
micro-environments. These are clear-ly some of the challenges for
future work.
7. Summary and concluding remarks
This paper describes the results of an experimental and
computa-tional study of the effects of localized hyperthermia on
breast cancercells. Controlled hyperthermia was achieved using an
implantable de-vice. The results from the in-vitro studies show
clearly that low levelsof hyperthermia, when delivered in a short
pulse regimen, can causecell cycle arrest in tumor cells. This
leads to the inhibition of cell growthand increasing cell death
after more than ve days of cyclic hyperther-mia. The observed
increase in cell death is also associated with heatshock protein
expression. This study suggests that, after 5 days of cyclicheat
treatment, using the implantable device, MDA-MB-231 cells un-dergo
cell cycle arrest.
The current results clearly suggest a need for in-vivo testing.
Further-more, there is a need for additional studies that use
surface texture topromote the improved integration of the device
with biological tissue.
Acknowledgments
This workwas supported by theNational Science Foundation
(GrantNo. DMR 0231418), the Grand Challenges Program at Princeton
Univer-sity and the sponsor-id="gs3" id="gts0015">STEP-B Program
of theWorld Bank. The authors are grateful to these organizations
for theirsupport of the research.
References
[1] B.W. Stewart, P. Kleihues, in: C.I.d.R.s.l.C. (CIRC) (Ed.),
Le Cancer Dans le Monde,IARCPress, Lyon, 2005.
[2] A. Jemal, et al., CA Cancer J. Clin. 57 (1) (2007) 4366.[3]
B. Alberts, Molecular Biology of the Cell, 5th ed. Garland Science,
New York, 2008
1 v. (various pagings).[4] P.C. Prorok, J. Pediatr. Hematol.
Oncol. 14 (2) (1992) 117128.[5] N.E. Day, D.R. Williams, K.T. Khaw,
Br. J. Cancer 59 (6) (1989) 954958.[6] S.H. Taplin, et al., Cancer
Epidemiol. Biomarkers Prev. 6 (8) (1997) 625631.[7] R. Kurzrock, M.
Markman, Current Clinical Oncology, Humana Press, Totowa, N.J,
2008, xii, 445 pp.[8] M.J. Hassett, et al., J. Natl. Cancer
Inst. 98 (16) (2006) 11081117.[9] A.C. Society, American Cancer
Society, American Cancer Society, Atlanta, 2008.
[10] J. van der Zee, Ann. Oncol. 13 (8) (2002) 11731184.[26] R.
Cavaliere, et al., Cancer 20 (9) (1967) 13511381.[27] J.A. Dickson,
M. Suzangar, Cancer Res. 34 (6) (1974) 12631274.[28] J.G. Short,
P.G. Turner, Proc. IEEE 68 (1980) 133142.[29] S.B. Field, C.C.
Morris, Radiother. Oncol. 1 (2) (1983) 179186.[30] S.A. Sapareto,
W.C. Dewey, Int. J. Radiat. Oncol. Biol. Phys. 10 (6) (1984)
787800.[31] M.W. Dewhirst, et al., Int. J. Hyperthermia 21 (2005)
779790.[32] B. Hildebrandt, et al., Crit. Rev. Oncol. Hematol. 43
(1) (2002) 3356.[33] J.R. Lepock, International Journal of
Hyperthermia, Taylor & Francis Ltd., 2005,
pp. 681687.[34] H.H. Kampinga, Int. J. Hyperthermia 22 (3)
(2006) 191196.[35] R.W. Habash, et al., Crit. Rev. Biomed. Eng. 34
(6) (2006) 491542.[36] R.W. Habash, et al., Crit. Rev. Biomed. Eng.
35 (12) (2007) 37121.[37] R.W. Habash, et al., Crit. Rev. Biomed.
Eng. 35 (12) (2007) 123182.[38] R. Love, R.Z. Soriano, R.J. Walsh,
Cancer Res. 30 (5) (1970) 15251533.[39] K. Kase, G.M. Hahn, Nature
255 (5505) (1975) 228230.[40] U. Hfeli, Scientic and Clinical
Applications of Magnetic Carriers, Plenum Press,
New York, 1997 xiii, 628 pp.[41] A. Jordan, et al., J. Magn.
Magn. Mater. 225 (12) (2001) 118126.[42] D.C.F. Chan, D.B.
Kirpotin, P.A. Bunn Jr., J. Magn. Magn. Mater. 122 (1-3) (1993)
374378.[43] K. Shinohara, Int. J. Hyperthermia 20 (7) (2004)
679697.[44] S. Yukumi, et al., Anticancer Res. 28 (1A) (2008)
6974.[45] A.A. Sarnaik, et al., Recent Results Cancer Res. 169
(2007) 7582.[46] M.-C. Blanger, Y. Marois, J. Biomed. Mater. Res.
58 (5) (2001) 467477.[47] W. Strober, Trypan blue exclusion test of
cell viability, Current Protocols in Cell
Biology, 2002.[48] N.A.P. Franken, et al., Nat. Protoc. 1 (5)
(2006) 23152319.[49] J.L. Roti Roti, N. Turkel, Radiat. Res. 139
(1) (1994) 7381.[50] A. Mukhopadhaya, et al., Cancer Res. 67 (16)
(2007) 77987806.[51] F.A. Wiegant, et al., Cancer Res. 47 (6)
(1987) 16741680.[52] R.A. Coss, W.A. Linnemans, Int. J.
Hyperthermia 12 (2) (1996) 173196.[53] W.C. Dewey, Int. J.
Hyperthermia 10 (4) (1994) 457483.[54] H.R. Moyer, K.A. Delman,
Int. J. Hyperthermia 24 (3) (2008) 251261.[55] B.V. Harmon, et al.,
Int. J. Radiat. Biol. 58 (1990) 845858.[56] S.S. Mambula, S.K.
Calderwood, Int. J. Hyperthermia 22 (7) (2006) 575585.[57] S.K.
Calderwood, D.R. Ciocca, Int. J. Hyperthermia 24 (1) (2008)
3139.[58] J.G. Kiang, I.D. Gist, G.C. Tsokos, Mol. Cell. Biochem.
204 (12) (2000) 169178.[59] S.K. Calderwood, J.R. Theriault, J.
Gong, Eur. J. Immunol. 35 (9) (2005) 25182527.[60] S.K. Calderwood,
J.R. Theriault, J. Gong, Int. J. Hyperthermia 21 (8) (2005)
713716.[61] S.A. Rosenberg, N. Engl. J. Med. 350 (14) (2004)
14611463.[62] R.C. Davis, G.T. Bowden, A.E. Cress, Int. J. Radiat.
Biol. Relat. Stud. Phys. Chem.
Med. 43 (4) (1983) 379390.[63] J. Michalakis, et al., Cancer
Chemother. Pharmacol. 56 (6) (2005) 615622.[64] N.R. Datta, et al.,
Int. J. Hyperthermia 6 (3) (1990) 479486.[65] R.T. Pettigrew, et
al., Br. Med. J. 4 (5946) (1974) 679682.[66] H.H. Pennes, Analysis
of tissue and arterial blood temperatures in the resting
human forearm, 1948, pp. 93122.[67] W.L. Nyborg, Phys. Med.
Biol. 7 (1988) 785.[68] M.N. Rylander, et al., Int. J. Hyperthermia
26 (8) (2010) 748767.[69] P. Mulamba, Biomaterials modeling of
localized hyperthermia and drug delivery
for breast cancer, PhD Thesis, Ohio State University, 2008.cious
in halting cellular growth in MDA-MB-231 cells, in-vitro 3D
An in-vitro study of the effects of temperature on breast cancer
cells: Experiments and models1. Introduction2. Device design and
fabrication2.1. Device fabrication2.2. Design and testing of a
biocompatible implant for inducing hyperthermia using resistive
heating
3. Experimental procedures3.1. Cell culture experiments3.2. Cell
exposure to continuous hyperthermia3.3. Cyclic heat shock
procedure3.4. Propidium iodine assay for both cell death and
apoptotic body formation3.5. Cell cytoskeleton imaging3.6. Effects
of temperature (37C, 41C, 43C, and 45C) on heat shock protein
expression3.7. Statistical analysis
4. Modeling4.1. Numerical modeling of heat transfer4.2.
Two-dimensional analysis of the heat diffusion from the
hyperthermia biocompatible device into a fluid filled
environment
5. Results5.1. Isothermal hyperthermia5.2. Comparison of
measured and predicted temperatures profiles of device mediated
hyperthermia5.3. Staining of cell death and cell morphology under
isothermal hyperthermia5.4. Cyclic hyperthermia5.5. In-vitro cyclic
heating and heat shock protein expression
6. Discussion6.1. Morphological changes in the cell cytoskeleton
as a potential cause for both the disturbance of cellular function
and ...6.2. The biocompatible hyperthermia device and its potential
for medical applications6.3. The potential use of an inductively
coupled system for power autonomy6.4. Potential future
directions
7. Summary and concluding remarksAcknowledgmentsReferences