Mapping of hyperthermic tumor cell death in a microchannel under unidirectional heating.
ABSTRACT Hyperthermia can be used as an adjunctive method of chemotherapy, radiotherapy, and gene therapy to improve cancer treatment. In this study, we investigate the hyperthermic cell death of cervix cancer CaSki cells in a microchannel integrated with a directional heating scheme. Heat was applied from the inner end to the outer end of the channel and a temperature distribution from 60 °C to 30 °C was established. A three dimensional (3D) numerical model was conducted for the heat transfer simulation, based on which a simple fitting method was proposed to easily estimate the temperature distribution along the channel. Cell death along the channel was mapped 22 h after the heating treatment by dual fluorescent labeling and phase-contrast microscopy imaging. Upstream, where the temperature is higher than 42 °C, we observe necrotic death, late-stage and early stage apoptotic death in sequence along the channel. Downstream and in the middle of the channel, where the temperature is lower than 42 °C, significant cell detachment was noted. Vigorous detachment was observed even in the non-hyperthermic zone (temperature lower than 37 °C), which we believe is due to the direct effect of the hyperthermic zones (higher than 37 °C). The present work not only gives a vivid map of cell responses under a temperature gradient, but also reveals the potential interactions of the heated tumor cells and non-heated tumor cells, which are seldom investigated in conventional petri-dish experiments.
Article: Anoikis.Cell Death and Differentiation 12/2005; 12 Suppl 2:1473-7. · 8.37 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: To determine whether intracellular pH (pHi) is affected during hyperthermia in substrate-attached cells and whether acute extracellular acidification potentiates the cytotoxicity of hyperthermia via an effect on pHi. The pHi was determined in cells attached to extracellular matrix proteins loaded with the fluorescent indicator dye BCECF at 37 degrees C and during 42 degrees C hyperthermia at an extracellular pH (pHe) of 6.7 or 7.3 in cells. Effects on pHi during hyperthermia are compared to effects on clonogenic survival after hyperthermia at pHe 7.3 and 6.7 of cells grown at pHe 7.3, or of cells grown and monitored at pHe 6.7. The results show that pHi values are affected by substrate attachments. Cells attached to extracellular matrix proteins had better signal stability, low dye leakage and evidence of homeostatic regulation of pHi during heating. The net decrease in pHi in cells grown and assayed at pHe = 7.3 during 42 degrees C hyperthermia was 0.28 units and the decrease in low pH adapted cells heated at pHe = 6.7 was 0.14 units. Acute acidification from pHe = 7.3 to pHe = 6.7 at 37 degrees C caused an initial reduction of 0.5-0.8 unit in pHi, but a partial recovery followed during the next 60-90 min. Concurrent 42 degrees C hyperthermia caused the same initial reduction in pHi in acutely acidified cells, but inhibited the partial recovery that occurred during the next 60-90 min at 37 degrees C. After 4 h at 37 degrees C, the net change in pHi in acutely acidified cells was 0.30 pH unit, but at 42 degrees C is 0.63 pH units. The net change in pHi correlated inversely with clonogenic survival. Hyperthermia causes a pHi reduction in cells which was smaller in magnitude by 50% in low pH adapted cells. Hyperthermia inhibited the partial recovery from acute acidification that was observed at 37 degrees C in substrate attached cells, in parallel with a lower subsequent clonogenic survival.International Journal of Radiation OncologyBiologyPhysics 09/1997; 39(1):205-12. · 4.52 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: The sensitivity to hyperthermic damage (45Â°C +- 0.05Â°) and the interaction of hyperthermic damage with radiation damage was determined in Chinese hamster ovary (CHO) fibroblasts in monolayer culture. The survival curve for asynchronous cells heated at 45Â°C was exponential with a shoulder (Dâ = 3.6 min, D/sub q/ = 7.5 min, n = 7.3). Recovery of the capacity to accumulate sublethal damage (return of extrapolation number to 7.3) was complete and occurred within the first 4 hr. By 72 hr the Dâ returned to normal although the D/sub q/ persisted at 20 min at 45Â°C. Heating (17.5 min at 45Â°C) followed immediately by radiation (450 rad) or radiation followed immediately by heating did not produce significantly different survival values. Hyperthermia reduced the Dâ of the radiation--survival curve by a factor of 1.9 and increased the extrapolation number by a factor of 3. The D/sub q/ remained unchanged. For intervals up to 48 hr, prior hyperthermia increased the extrapolation number and D/sub q/ of the radiation--survival curve to approximately 1000 and 650 rad, respectively. However, the Dâ remained unchanged at approximately 83 rad over the entire interval of 0 to 72 hr. For the sequence of radiation followed by hyperthermia, the D/sub q/ of the hyperthermia--survival curve was reduced to 2.9 min at 45Â°C by simultaneous irradiation but returned to near-normal values after a fractionation interval of 2 hr. The Dâ of the hyperthermia--survival curve was unaffected by prior irradiation.Radiation Research 07/1976; 66(3):505-18. · 2.70 Impact Factor
Mapping of hyperthermic tumor cell death in a microchannel underMapping of hyperthermic tumor cell death in a microchannel under
unidirectional heating unidirectional heating
Fen Wang, Yuhui Li, Lei Chen, Dandan Chen, Xiaolei Wu et al.
Citation: Biomicrofluidics 6 6, 014120 (2012); doi: 10.1063/1.3694252
View online: http://dx.doi.org/10.1063/1.3694252
View Table of Contents: http://bmf.aip.org/resource/1/BIOMGB/v6/i1
Published by the American Institute of Physics.
Related ArticlesRelated Articles
Efficient manipulation of microparticles in bubble streaming flows
Biomicrofluidics 6, 012801 (2012)
Characterization and separation of Cryptosporidium and Giardia cells using on-chip dielectrophoresis
Biomicrofluidics 6, 012805 (2012)
Development and validation of a low cost blood filtration element separating plasma from undiluted whole blood
Biomicrofluidics 6, 012804 (2012)
Rapid and cost-effective fabrication of selectively permeable calcium-alginate microfluidic device using “modified”
embedded template method
Biomicrofluidics 6, 012821 (2012)
Magnetic energy coupling system based on micro-electro-mechanical system coils
J. Appl. Phys. 111, 07E734 (2012)
Additional information on BiomicrofluidicsAdditional information on Biomicrofluidics
Journal Homepage: http://bmf.aip.org/
Journal Information: http://bmf.aip.org/about/about_the_journal
Top downloads: http://bmf.aip.org/features/most_downloaded
Information for Authors: http://bmf.aip.org/authors
Mapping of hyperthermic tumor cell death in a
microchannel under unidirectional heating
Fen Wang, Yuhui Li, Lei Chen, Dandan Chen, Xiaolei Wu,a)and Hao Wanga)
College of Engineering, Peking University, Beijing, China
(Received 15 December 2011; accepted 26 February 2012; published online 20 March 2012)
Hyperthermia can be used as an adjunctive method of chemotherapy, radiotherapy,
and gene therapy to improve cancer treatment. In this study, we investigate the
hyperthermic cell death of cervix cancer CaSki cells in a microchannel integrated
with a directional heating scheme. Heat was applied from the inner end to the outer
end of the channel and a temperature distribution from 60?C to 30?C was
established. A three dimensional (3D) numerical model was conducted for the heat
transfer simulation, based on which a simple fitting method was proposed to easily
estimate the temperature distribution along the channel. Cell death along the
channel was mapped 22 h after the heating treatment by dual fluorescent labeling
and phase-contrast microscopy imaging. Upstream, where the temperature is higher
than 42?C, we observe necrotic death, late-stage and early stage apoptotic death in
sequence along the channel. Downstream and in the middle of the channel, where
the temperature is lower than 42?C, significant cell detachment was noted. Vigorous
detachment was observed even in the non-hyperthermic zone (temperature lower
than 37?C), which we believe is due to the direct effect of the hyperthermic zones
(higher than 37?C). The present work not only gives a vivid map of cell responses
under a temperature gradient, but also reveals the potential interactions of the
heated tumor cells and non-heated tumor cells, which are seldom investigated in
conventional petri-dish experiments. V
C 2012 American Institute of Physics.
In the past few years, hyperthermia has gained a lot of interest as an effective adjunctive method
of chemotherapy, radiotherapy, and gene therapy1–4in cancer treatments. Successful hyperthermic
treatments rely on the balance between killing the majority of cancerous cells, while protecting the
healthy cells when the tumorous tissues are exposed to high temperatures. It is of great importance
to identify the associated damage to both the cancer cells and the neighboring healthy tissues. There
is extensive research in this particular field, both in vivo5–7and in vitro,8–11but numerous challenges
remain due to the complex physical and physiological properties and variety of tissues.
Cell damage caused by hyperthermia has been studied for a few decades. Vidair et al.12pro-
posed two distinct modes of hyperthermic cell death based on time tracking. A “rapid” mode was
characterized by cell detachment and inhibited rates of macromolecular synthesis, which usually
dominates during the first few days post heating. A “slow” mode was defined to describe the proc-
esses after the cells had fully recovered from the heat-induced inhibition of macromolecular synthe-
sis, and cell detachment had ceased. Roti et al.13reviewed the mechanisms of hyperthermic induced
cell killing at the cellular level. Hyperthermia not only has effects on the plasma membrane protein
distribution which related with calcium spike, disrupts mitochondrial membrane potential resulting
in the redox status of the cells, but also has significant impacts on the nucleus such as double strand
break. Recent research by Shellman et al.14indicated that hyperthermia can also induce endoplas-
mic reticulum mediated apoptosis. Hyperthermia can induce necrotic and apoptotic cell death
a)Authors to whom correspondence should be addressed. Electronic addresses: firstname.lastname@example.org and hwang@coe.
C 2012 American Institute of Physics6, 014120-1
BIOMICROFLUIDICS 6, 014120 (2012)
depending on the temperature distribution and time spans. Harmon et al.15categorized the death
occurring after various heat loads and distinguished the necrosis and apoptosis by the extent of
membrane disruption or DNA damage. Moroi et al.7studied the regional dead-cell distribution by
examining the necrosis and apoptosis in the center and periphery of the solid nodules of gliosarcoma
(T9) in vivo by heating for 30min in a water bath at 43?C. They found necrosis was enhanced in
3?6 h post heating, while a regional difference in the rate of apoptosis was detected immediately af-
ter heating. Apoptosis can induce a group of proteins named HSPs (heat shock proteins) which will
trigger thermotolerance and protect cells from heat-induced apoptosis.16–18Even though necrosis
and apoptosis have been observed and differentiated in a variety of heating experiments, their sensi-
tivity to temperature in vivo or in vitro and their development might be better interpreted either
experimentally or theoretically.
The main drawbacks in the previous in vitro experiments is that the heating was performed
in either a dedicated water bath or a heating stage,15,19where the temperature in the test section
is approximately uniform. However, in practice, uniformly high temperature is hard to achieve
in living systems due to the diverse physical thermal conductivity of organs, tissues, and blood
flow, especially in cases when an external heat source, such as a heating probe, is applied.
Extensive research of temperature-gradient control has been conducted in PCR for DNA ampli-
fication in a microfluidic platform,20–22but very few studies have been conducted to investigate
temperature-gradient effects on cells. Das et al.23designed a microfluidic thermal gradient sys-
tem in an incubator environment and showed its potential in thermotaxis studies. Lucchetta
et al.24used microfluidic laminar flow to create a temperature difference around a Drosophila
embryo and observed the density difference of nuclei in the two halves of the embryo affected
by a T-step. However, few studies have been conducted to investigate the cell responses under
a unidirectional heating and the interactions of the heated cells and non-heated cells with a re-
gional or selected hyperthermic treatment of some of the cells in a microfluidic environment.
The present work aims to study the hyperthermic death of CaSki cells seeded in a single
PDMS (polydimethysiloxane) microfluidic channel with unidirectional heating along the chan-
nel length. The temperature system was easily controlled by a heating tungsten wire embedded
in the PDMS mold. A mathematical model was proposed to calculate the temperature distribu-
tion along the channel. The cell damage and the trend of the cell responses were mapped along
the channel by identifying the apoptosis/necrosis rates using a commercial apoptosis kit
(Annexin V-FITC&PI). Moreover, the interactions between hyperthermic and non-hyperthermic
zones were revealed.
II. MATERIALS AND METHODS
A. Microchip fabrication
Microfluidic devices were fabricated in PDMS using the standard soft lithography method
and replica molding.25The detailed fabrication procedure was presented in our previous
papers.26,27In short, microscale patterns of channels in the photomask were replicated photoli-
thographically using a negative photoresist (Microchem, Newton, MA) in a 100 class clean
room. The inlet and outlet reservoirs for cell loading or collecting were punched with a 3mm
aiguille. The holes for thermocouples were punched by a 20 gauge needle. The tungsten wire
(200lm diameter) was sandwiched between the PDMS slab and the cleaned 1mm thick,
25?75mm glass slides, which were sealed together by an oxygen plasma treatment (Plasma-
Therm Etcher, Diener electronic Co. LTD, Germany). The wire was positioned near the inlet
reservoir perpendicular to the channel. The PDMS chips and the heating wire were exposed to
UV light for around four hours in a safety cabinet (MSC-Advantage, Thermo Fisher Scientific
Inc., MA, USA) in order to sterilize them before each experiment.
B. Heating system and operation
Metal wires generate heat when an electric current is applied. In the present design, a piece
of tungsten wire was used as the heating wire. It was fixed near the inlet reservoir at a distance
014120-2 Wang et al. Biomicrofluidics 6, 014120 (2012)
of about 4.5mm, perpendicular to the longitudinal direction of the channel (Fig. 1). Three
K-type thermo couples, namely T1, T2, and T3, were embedded near the channel (2mm away
and parallel), 9200lm and 16 900lm away from the inlet, respectively. The temperatures were
recorded through a data acquisition system (UTL/D-08LS1V0N, UTOP Electronic Co. LTD,
Guangzhou, China). The power was supplied by a DC-stabilized source (WYK-5030, Huatai
Electronics Co. LTD, Yangzhou, China). After the microchip was mounted on the clean micros-
copy stage, a voltage was applied to the heating wire to induce a local high temperature near
the inlet and the temperature decreased along the microchannel. When the temperature distribu-
tion reached a steady state (which took around 15?20min), the temperature at the inlet was
?60?C and ?30?C at the outlet. The heating lasted for 105min (from the power being turned
on) and then the power supply was disconnected. It took 5?10min for the chip to cool down
to room temperature.
C. General cell culture
Cervical cancer CaSki cell line was chosen in our study which is known to be sensitive to
a combination of radiation and hyperthermia.28Cervical cancer CaSki cell line was cultured in
plastic culture petri-dish at 37?C, under 5% CO2 in Dulbecco’s modified Eagle’s medium
(DMEM, Hyclone, Logan, UT) supplemented with 10% (v/v) fetal bovine serum (FBS,
Hyclone, Logan, UT) and penicillin (100 units/ml, Hyclone, Logan, UT). CaSki cultures were
maintained in 60mm petri-dish (Corning) in a CO2 incubator (Forma Series II, Thermo Fisher
Scientific Inc., MA, USA) at 37?C and humidified atmosphere with 5% CO2. Cells were
diluted at a ratio of 1:5 every 3 days to maintain them in the exponential growth phase
(?1?106cells/ml). Cells were harvested by adding 0.25% Trypsin-EDTA (Hyclone, Logan,
UT) to the culture and centrifuged at 300g for 10min to remove the supernatant. The cells
were then resuspended in the culture medium at a higher concentration before seeding in the
D. Cell seeding and culture in the microchip
The channels were coated with fibronectin from bovine plasma (Sigma, St. Louis, MO) to
facilitate cell adhesion. The activity of fibronectin remained unchanged up to 55?C as observed
FIG. 1. Schematic diagram of the microchip integrated with a directional heating scheme: side view (a) and top view (b).
The inserts are the fluorescent images of the cells on the first day post-heating located at the positions corresponding to
each temperature measurement point.
014120-3Cell death under unidirectional heatingBiomicrofluidics 6, 014120 (2012)
previously by Vuento et al.29Therefore, cell detachment by the denaturation of fibronectin
could be eliminated in our experiment, if the temperature were lower than 55?C. Fibronectin
was prepared at a concentration of 50lg/ml in phosphate-buffered saline (PBS) (Beijing Chem-
icals Co. LTD., China) before each experiment and then was introduced into the microfluidic
device and incubated at 37?C for 30min. Prepared CaSki cells were suspended in the culture
media, injected into the sample reservoirs of the microfluidic device, and allowed to spread and
attach for 5–15min. After adding extra culture media in both reservoirs, the device with seeded
cells was incubated in a 37?C, 5% CO2 incubator (Forma Series II, Thermo Fisher Scientific
Inc., MA, USA) before hyperthermic experiments. Renewing the media in the reservoirs, every
few hours provided enough media to support cell growth inside the device.
E. Sample treatment
To test the apoptosis and necrosis, a dual staining using an Annexin V-FITC/PI apoptosis
detection kit (KeyGen Biotech. Co. LTD., Nanjing, China) was performed in a microfluidic cell
culture system. Apoptotic cells translocate the membrane phosphatidylserine (PS) from the
inner face of the plasma membrane to the outside which can be easily detected by staining with
a fluorescent conjugate of Annexin V, which has a high affinity for PS. Propidium iodide (PI)
is membrane-impenetrable and usually used to identify necrotic cells or late apoptotic cells
whose membrane is not intact. With the combination of these two chemicals, it was possible
for intact cells, initial apoptosis, middle or late apoptosis, and necrosis to be distinguished. The
reagent was diluted by binding buffer to a final concentration suggested by the supplier. The
culture medium was removed and the channel was washed twice with fresh PBS buffer. A drop
of prepared reagent (around 50ll) was added to the reservoir, and the culture channel was
filled. The chip was incubated in the CO2 incubator at 37?C for 10min before testing the cell
F. Fluorescence microscopy
Fluorescent and phase contrast images were taken on an inverted fluorescence microscopy
(TI-U INVERTED, Nikon, Japan) with a CCD camera (Monochrome Cooled Digital Camera
Head DS-Qi1Mc, Nikon, Japan). The setting of the CCD camera and the software were kept
identical from one experiment to another whenever comparison between experiments was
desired. In this experiment, Annexin V-FITC (465–495nm) was excited by blue light and emit-
ted green fluorescence, while propidium iodide (PI, 535/617nm) was excited by green light and
emitted red fluorescence.
III. RESULTS AND DISCUSSION
A. Directional heating
The growing cultured CaSki cells were exposed to the directional heating in the single
microfluidic channel. The temperature profiles T1, T2, and T3 at three positions (at a distance
of 0lm, 9200lm, and 16 900lm from the inlet reservoir, respectively) were recorded. Under
the heating power of about 0.47W, T1, T2, and T3 reached a steady state of around 6062?C,
4263?C, and 3362?C, respectively, which are represented by the three triangle points marked
in Fig. 2(a).
To have the temperature distribution along the channel, a 3D computational fluid dynamics
(CFD) model is established to simulate the heat transfer of the system. The conduction in the
chip, the heat dissipation on the chip surfaces, and the natural convection in the ambient air are
all considered in the model. The material properties employed in the simulation are shown in
Table I. Commercial software package FLUENT 6 is employed for the geometry setup and also
the numerical treatment. As a result, the temperature distribution on the chip is obtained as
shown in Fig. 2(b). The temperature distribution along the microchannel is extracted and shown
by the dashed curve in Fig. 2(a). A good agreement is seen with the three experimental points.
014120-4 Wang et al. Biomicrofluidics 6, 014120 (2012)
FIG. 2. Temperature along the chip: (a) temperatures at three locations measured by thermocouples (black triangle), the
simulation result (dashed curve), and the fitting result based on fin theory (solid curve); (b) the simulated temperature distri-
bution on the chip.
TABLE I. Material properties employed in simulation.
Properties PDMS Glass Air
Density (kg/m3) 1020 2500Incompressible ideal gas
Thermal conductivity (W/m K) 0.181.4 0.0242
Thermal capacity (J/kg) 1460840 1006.43
Viscosity (kg/m s)——
014120-5 Cell death under unidirectional heatingBiomicrofluidics 6, 014120 (2012)
The chip temperature distribution in Fig. 2(b) indicates that the temperature change is
mainly occurring along the x direction. Due to the simplicity of the present setup, the tempera-
tures along the channel at the steady states can be easily obtained by making a fitting based on
the heat transfer theories about a fin.30We justify this by noting that since the chip thickness
(?4mm) is much smaller than the chip length (?30mm), the heat is dissipating into the air
mainly on the top and bottom surfaces, and the chip can be treated like a fin. The temperature
distribution in the fin along x direction is30
h ¼ h0emxþ e2mHe?mx
1 þ e2mH
where h is the excess temperature (h=T?Tair), h0is the excess temperature at the fin base
and H is the length of the fin. Based on this equation, a temperature curve along the channel
can be fitted as shown in Fig. 2(b). This fitted curve will be employed in the following sections
as the temperature distribution along the channel.
B. Maps of cell death under the temperature gradients
The cells were cultured for 22 h following heating treatment. A map of cell damage along
the channel was then conducted by means of fluorescent staining and phase-contrast imaging.
The results are combined and shown in Fig. 3(a). A control experiment was also conducted and
no obvious change was detected except spontaneous apoptosis.
As introduced in Sec. III A, PI is membrane-impenetrable and usually used to identify ne-
crotic cells or late apoptotic cells whose membrane is not intact. In Fig. 3(a), the red fluores-
cence represents the nuclei that were stained with PI. The dense red dots in the upstream of the
channel from 0?7000lm (>45?C) indicate vast necrosis or late-stage apoptosis in this area.
Downstream along the channel, the proportion of the necrotic cells decreases and the proportion
of apoptotic cells increases. The apoptotic cells at the early stage were stained by FITC in the
membrane only (green fluorescence). Further downstream the cells were not stained by either
PI or FITC, but their morphology alteration, such as rounding up, might indicate that they were
The cell morphology along the channel can also be used to study the cell responses. The
magnified images of cell morphology at three different positions (I, II, and III) are illustrated at
the bottom of Fig. 3. The cell morphology before the heating treatment (0 h) is also given for
comparison. It is seen at the beginning of the channel (0?3800lm, 60?C to 51?C) that the
cytoplasm has collapsed and the nuclei appear to have swelled as shown in position I. The cells
have indistinct boundaries. During the heating period, membrane blebs were detected around
25min after the onset of the heating, then the cell cytoplasm suddenly collapsed (Fig. 4).
Downstream to 3800lm ?8000lm, 51?C ?43?C, the cells still showed a necrotic morphology
change but some cells shrank and became sequestered from each other (position II). The cell
boundary looked much clearer than the upstream cells. The mechanism behind this slight differ-
ence should be addressed in future studies by more frequent tracking after the heating treat-
ment. Beyond 8000lm, e.g., position III, the cells appear rounder which could be a sign of
early apoptosis.31,32Further downstream between from 9000lm and 18 500lm (about 42?C
down to 31.6?C), a great portion of the cells are estimated to detach after comparing with the
map before the heating (Fig. 3(b)). Worthy to note is that detachments were also detected in
the non-hyperthermic zone where the temperature was lower than 37?C. Finally, near the outlet
reservoir, where temperature was around room temperature during the heating treatment, there
was a very short area (from ?18 500lm to the outlet, below 32?C) where the cells were alive
and grew well. The distribution of cell amount along the channel is given in Fig. 5(a). The
amount of the detached cells is represented by the light gray bar and it is seen that the detach-
ment mostly occurred from 9000 to 16 000lm, corresponding to the temperatures ranging from
42 to 34?C. Figure 5(b) shows the percentage of cell detachment versus temperature along the
channel (statistical result over 5 experimental trials).
014120-6 Wang et al.Biomicrofluidics 6, 014120 (2012)
FIG. 3. (a) Map of cell damage stained by dual fluorescent dyes. Due to the confinement of the microscopy view field, a
seamless connection for images has been conducted. The CaSki cells were heat-treated from 60?C to 30?C for around
105min followed by incubation at 37?C for around 22 h. Necrotic damage, apoptosis at different stages, and detachment
and living cells were observed in sequence along the channel. Three positions I, II, and III were dotted circled in (a) and
the magnified images of cell morphology change from 0 h to 22 h were given. (b) Cell morphology map before heating in
the position where the temperature was below 37?C during heating.
014120-7 Cell death under unidirectional heating Biomicrofluidics 6, 014120 (2012)
Based on the above mapping of florescent staining and morphology imaging, it is expected
that high temperatures (>45?C) induce necrosis or late-stage apoptosis to the tumor cells, while
mild hyperthermia (from 45?C to 42?C) mainly causes apoptosis at different stages. Heat
shocks with the temperature below 42?C are likely to trigger cell detachment. It is interesting
to note that cell detachment also expanded into the non-hyperthermic zone (?37?C).
C. Cell detachment in hyperthermic zones
As shown in Fig. 3, cell detachment occurred in the hyperthermic zone from ?9000lm to ?12
750lm around 22 h post heating, corresponding to a temperature range between ?42?C and
?37?C. The adherent cells in this detachment zone were not stained using an apoptosis kit (Annexin
V-FITC&PI), but they were rounder and seemed to detach from the substrate, which indicates they
might undergo early apoptosis. Further tracking (42 h) at the location around 9100lm (?42?C)
revealed that some of the circular cells did detach from the bottom substrate (detached cells are
marked by dotted arrows in Fig. 6(d)). As shown by the photographs taken at 90 h post heating (Fig.
6(e)), the cells swelled and shrank in size, but still attached to the monolayer.
D. Cell detachment in non-hyperthermic zones
Cell detachment was also detected in the zone where the temperature was equal or below
37?C (around 12750lm from the inlet). Fig. 3(a), which was taken 22 h post heating, clearly
shows that the cell population downstream in the channel was significantly less than that before
the heating (Fig. 3(b)).
Vuento et al.29have studied the denaturation of fibronectin and found that this adhesion-
mediating protein remains unchanged up to 55?C, which is much higher than 37?C. Therefore,
the detachment in the non-hyperthermic zone cannot be due to the denaturation of the extracel-
lular matrix. Fig. 7 shows the cell detachment in the region near T3=33?C after culturing the
cells for 22, 42, and 90 h, respectively, following heating.
Cell detachment is one of the cellular responses in hyperthermic treatment33–37and the
mechanisms associated with it have been studied extensively.12,37–39Vidair and Dewey12found
cell detachment in Chinese hamster ovary cells (CHOs) several days after heating them for
periods of 20min or 30min at 45?C. They suggest that the acidification of the culture medium
FIG. 4. Real time images for the cell morphology change during the heating. The cells expanded in size (b) and the nucleus
swelled (c) and (d) in the first few minutes after applying heat. Then, necrotic blebs occurred near the plasma membrane
(arrowed in (e)) and bloomed around the cell ((f) and (g)). The blebs grew quickly in size and then disappeared and the cells
014120-8 Wang et al. Biomicrofluidics 6, 014120 (2012)
might be a possible reason of the observed detachment. Wahl et al.39detected that acutely
acidified cells (acute changes in the extracellular pH from 7.3 to 6.7 before onset of the heat-
ing) showed detachment from the substrate after hyperthermic treatment at 42?C. The unde-
tached cells were becoming rounder and more easily detached with a medium change. In
another study,38the cytoskeletal protein alteration or loss on the plasma membrane was found
to affect the cell adhesion to the substrate during hyperthermia-induced apoptosis, which results
FIG. 5. The distribution of the cell amount along the channel at the 22nd h post heating (gray bar) and the detached cells
within the 22 h (light gray bar) (a), and the cell detachment percentage versus the temperature along the channel (b).
014120-9 Cell death under unidirectional heatingBiomicrofluidics 6, 014120 (2012)
in cell rounding/blebbing or anoikis. However, few studies have investigated the mechanisms of
cell detachment below 37?C by heating nearby cells. A recently published study37found that
incubating the carcinoma cells (SK-OV-3) with the heat-shocked mesenchymal stem cells
(MSCs) could cause significant tumor cell death. It was observed that the secreted factors from
the heated MSCs induce cytoskeleton destruction and floating of the adherent tumor cells. The
contents released from the heated MSCs might also markedly weaken the tumor cell progres-
sion by reducing the expression of antiapoptotic protein Bfl-1.
Our study demonstrates that the heated tumor cells and non-heated tumor cells interact in a
microfluidic environment with a regional or selected hyperthermic treatment of some of the
cells. It is no doubt that the vigorous cell detachment in the non-hyperthermic zone is the result
of the upstream hyperthermic zone. The authors had conducted more than 10 experimental trials
and the results are similar, i.e., the distributions of the death along the channels are similar, and
the interaction between the heated and the non-heated cells are similar.
The current work did not include biological assays for analyzing the mechanisms of detach-
ment in non-hyperthermic zones. The authors tried to measure the pH value of the fluid in the
microchannel, finding that the volume of the fluid was too limited (about 0.2ll) to have accu-
rate pH measurement with traditional methods. Micro-sensor might be employed for accurate
measurement.40,41Future experiments will be designed and conducted soon to analyze in more
detail the effect of hyperthermic or non-hyperthermic zones on cellular damage rates.
E. Cell spreading downstream
Downstream and near the outlet, some cells are observed to spread on the substrate, as
shown in Fig. 7(d) (42 h). The enlarged part of one cell exhibits significant spreading, which
may indicate an early stage of adhesion. In Fig. 7(e) (90 h), some cells already have a polarized
shape on the substrate but the proliferation is still inhibited. By phase contrast mapping at 90 h
(data not shown), cell spreading occurs at a distance ?11 000lm from the inlet, where the cor-
responding temperature is lower than 39?C.
FIG. 6. Cell morphological changes at the position 9000 ?9400lm before (a), right after (b), 22 h (c) and (c’), 42 h (d),
and 90 h (e) after hyperthermic treatment. Rounding cells at 22 h were marked by arrows in (c) and (c’); detached cells at
42 h are marked by dotted arrows in (d) in order to make comparisons with (c). Cell detachment was observed at 22 h and
42 h post heating. The scale bar represents 100lm.
014120-10 Wang et al.Biomicrofluidics 6, 014120 (2012)
We have considered the hyperthermic cell death of CaSki cells in a microchannel in a
PDMS microchip integrated with a directional heating scheme. A three dimensional (3D)
numerical model was conducted for the heat transfer simulation, based on which a simple fitting
method was proposed to easily estimate the temperature distribution along the channel. Hyper-
thermic (>37?C) and non-hyperthermic (?37?C) zones were both achieved. Cell death along
the channel, from high (?60?C) to low (?30?C) temperatures, was mapped at 22 h post heat-
ing by fluorescent dyes and tracked by phase contrast microscope for several days.
Necrotic cells, apoptotic cells at different stages, detached cells, and living cells were dis-
tinguished by dual fluorescent labeling and recorded in sequence along the channel. Cell
detachment was observed in both the hyperthermic (>37?C) and non-hyperthermic (?37?C)
zones. By tracking the cells after incubating for 90 h, cell spreading was observed downstream
in the channel where the heating temperature was lower than 39?C. The present work not only
gives a vivid map of cell responses under a temperature gradient but also reveals the potential
interactions of heated tumor cells and non-heated tumor cells, which are seldom investigated in
conventional petri-dish experiments.
This work was supported by the National Natural Science Foundation of China (Grant No.
50876001). The authors thank the Lab of Phase Change and Interfacial Phenomena at Tsinghua
University for their help throughout the work. The authors also thank Dr. Adriana Setchi at Imperial
College London for the English editing.
FIG. 7. Phase contrast images of cell morphological changes at the position, 16 650?17 050lm before (a) and right after
(b) hyperthermic treatment; after 22 h (c), after 42 h (d), and after 90 h (e). The insert is an enlarged look of a radial spread-
ing cell at 42 h post heating. Cell detachment was observed. Spreading cells and polarized cells were recorded in (d) and
(e), respectively. The scale bar represents 100lm.
014120-11 Cell death under unidirectional heatingBiomicrofluidics 6, 014120 (2012)
1J. Bull, Cancer Res. 44(10 Supplement), 4853 (1984).
2T. S. Hauck, T. L. Jennings, T. Yatsenko, J. C. Kumaradas, and W. C. W. Chan, Adv. Mater. 20(20), 3832 (2008).
3R. S. Scott, R. J. R. Johnson, K. V. Story, and L. Clay, Int. J. Radiat. Oncol., Biol. Phys. 10(11), 2119 (1984).
4A. Ito, F. Matsuoka, H. Honda, and T. Kobayashi, Cancer Gene Ther. 10(12), 918 (2003).
5R. Gordon, J. Hines, and D. Gordon, Med.Hypotheses 5(1), 83 (1979).
6A. Wagner, H. Weber, L. Jonas, H. Nizze, M. Strowski, F. Fiedler, H. Printz, H. Steffen, and B. Goke, Gastroenterology
111(5), 1333 (1996).
7J. Moroi, S. Kashiwagi, S. Kim, M. Urakawa, H. Ito, and K. Yamaguchi, Int. J.Hyperthermia 12(3), 345 (1996).
8S. Bhowmick, J. Coad, D. Swanlund, and J. Bischof, Int. J.Hyperthermia 20(1), 73 (2004).
9K. Henle and D. Leeper, Radiat. Res. 66(3), 505 (1976).
10B. Hildebrandt, P. Wust, O. Ahlers, A. Dieing, G. Sreenivasa, T. Kerner, R. Felix, and H. Riess, Crit. Rev. Oncol. Hema-
tol. 43(1), 33 (2002).
11Y. Rabin, Int. J. Hyperthermia 18(3), 194 (2002).
12C. A. Vidair and W. C. Dewey, Radiat. Res. 116(1), 157 (1988).
13J. L. Roti Roti, Int. J. Hyperthermia 24(1), 3 (2008).
14Y. G. Shellman, W. R. Howe, L. A. Miller, N. B. Goldstein, T. R. Pacheco, R. L. Mahajan, S. M. LaRue, and D. A. Nor-
ris, J. Invest. Dermatol. 128(4), 949 (2007).
15B. Harmon, A. Corder, R. Collins, G. Gobe, J. Allen, D. Allan, and J. Kerr, Int. J. Radiat. Biol. 58(5), 845 (1990).
16H. M. Beere, J. Cell Sci. 117(13), 2641 (2004).
17A. Ito, H. Honda, and T. Kobayashi, Cancer Immunol. Immunother. 55(3), 320 (2006).
18A. Samali and T. G. Cotter, Exp. Cell Res. 223(1), 163 (1996).
19X. He and J. C. Bischof, Ann. Biomed. Eng. 33(4), 502 (2005).
20N. Crews, C. Wittwer, R. Palais, and B. Gale, Lab Chip 8(6), 919 (2008).
21H. Mao, T. Yang, and P. S. Cremer, J. Am. Chem. Soc. 124(16), 4432 (2002).
22C. Zhang and D. Xing, Nucl. Acids Res. 35(13), 4223 (2007).
23S. K. Das, S. Chung, I. Zervantonakis, J. Atnafu, and R. D. Kamm, Biomicrofluidics 2, 034106 (2008).
24E. M. Lucchetta, J. H. Lee, L. A. Fu, N. H. Patel, and R. F. Ismagilov, Nature 434(7037), 1134 (2005).
25D. C. Duffy, J. C. McDonald, O. J. A. Schueller, and G. M. Whitesides, Anal. Chem. 70(23), 4974 (1998).
26F. Wang, H. Wang, J. Wang, H. Y. Wang, P. L. Rummel, S. V. Garimella, and C. Lu, Biotechnol. Bioeng. 100(1), 150
27Y. Li, F. Wang, and H. Wang, Biomicrofluidics 4, 014111 (2010).
28M. Franckena, L. J. A. Stalpers, P. Koper, R. G. J. Wiggenraad, W. J. Hoogenraad, J. D. P. van Dijk, C. C. Warlam-
Rodenhuis, J. J. Jobsen, G. C. van Rhoon, and J. van der Zee, Int. J. Radiat. Oncol., Biol., Phys. 70(4), 1176 (2008).
29M. Vuento, E. Salonen, K. Salminen, M. Pasanen, and U. Stenman, Biochem. J. 191(3), 719 (1980).
30S. M. Yang and W. Q. Tao, Heat Transfer, 4th ed. (High Education Publisher, Beijing, China, 2006).
31M. B. Friis, C. R. Friborg, L. Schneider, M. B. Nielsen, I. H. Lambert, S. T. Christensen, and E. K. Hoffmann, J. Physiol.
567(2), 427 (2005).
32G. Nagy, G. Pinter, G. Kohut, A. L. Adam, G. Trencsenyi, L. Hornok, and G. Banfalvi, DNA Cell Biol. 29(5), 249
33A. Gilmore, Cell Death Differ 12, 1473 (2005).
34W. X. Li, C. Chen, C. Ling, and G. C. Li, Radiat. Res. 145(3), 324 (1996).
35P. Ross-Riveros and J. T. Leith, Radiat. Res. 78(2), 296 (1979).
36Y. J. Lee, J. H. Kim, S. Ryu, S. H. Kim, C. A. Vidair, E. P. Armour, and P. M. Corry, J. Thermal Biol. 19(5), 305 (1994).
37J. A. Cho, H. Park, H. K. Kim, E. H. Lim, S. W. Seo, J. S. Choi, and K. W. Lee, Cancer 115(2), 311 (2009).
38F. Luchetti, F. Mannello, B. Canonico, M. Battistelli, S. Burattini, E. Falcieri, and S. Papa, Apoptosis 9(5), 635 (2004).
39M. L. Wahl, S. B. Bobyock, D. B. Leeper, and C. S. Owen, Int. J. Radiat. Oncol., Biol., Phys. 39(1), 205 (1997).
40I. A. Ges, B. L. Ivanov, D. Schaffer, E. A. Lima, A. A. Werdich, and F. J. Baudenbacher, Biosens. Bioelectron. 21(2),
41S. E. Eklund, D. Taylor, E. Kozlov, A. Prokop, and D. E. Cliffel, Anal. Chem. 76, 519 (2004).
014120-12 Wang et al.Biomicrofluidics 6, 014120 (2012)