Interplay of chemical and thermal gradient on bacterial
migration in a diffusive microfluidic device
Nithya Murugesan,
1
Purbarun Dhar,
2
Tapobrata Panda,
1
and Sarit K. Das
3,a)
1
Department of Chemical Engineering, Indian Institute of Technology Madras,
Chennai 600 036, India
2
Department of Mechanical Engineering, Indian Institute of Technology Ropar,
Rupnagar 140001, India
3
Department of Mechanical Engineering, Indian Institute of Technology Madras,
Chennai 600 036, India
(Received 17 November 2016; accepted 13 March 2017; published online 24 March 2017)
Living systems are constantly under different combinations of competing gradients
of chemical, thermal, pH, and mechanical stresses allied. The present work is about
competing chemical and thermal gradients imposed on E. coli in a diffusive
stagnant microfluidic environment. The bacterial cells were exposed to opposing
and aligned gradients of an attractant (1 mM sorbitol) or a repellant (1 mM NiSO
4
)
and temperature. The effects of the repellant/attractant and temperature on
migration behavior, migration rate, and initiation time for migration have been
reported. It has been observed that under competing gradients of an attractant and
temperature, the nutrient gradient (gradient generated by cells itself) initiates
directed migration, which, in turn, is influenced by temperature through the
metabolic rate. Exposure to competing gradients of an inhibitor and temperature
leads to the imposed chemical gradient governing the directed cell migration. The
cells under opposing gradients of the repellant and temperature have experienced
the longest decision time (60 min). The conclusion is that in a competing
chemical and thermal gradient environment in the range of experimental condit ions
used in the present work, the migration of E. coli is always initiated and governed
by chemical gradients (either generated by the cells in situ or imposed upon
externally), but the migration rate and percentage of migration of cells are
influenced by temperature, shedding insights into the importance of such gradients
in deciding collective dynamics of such cells in physiological conditions.
Published by AIP Publishing. [
http://dx.doi.org/10.1063/1.4979103]
I. INTRODUCTION
Multiple gradients, viz., chemical, thermal, pH, gaseous concentration, and mechanical
stress co-exist naturally in a physiological system.
1
Interactive action of such gradients coordi-
nates various essential biological activities such as wound healing,
2
guidance of sperm cells
towards the egg,
3
cancer cell metastasis,
4
and host–microbe relationships (viz., migration, colo-
nization, secretion, and infection)
5
in living organisms.
To counter infected or diseased conditions, gradients are created within the system via
external sources by drugs (skin patches, injecting drug carrier molecules/particles, and chemo-
therapy) or radiation and laser therapy (hyperthermia and hypothermia).
6
Such gradients
imposed on cells may interact with the multiple gradients existed within the entity, and changes
in the activities of the diseased cells and/or the microbes inhabited the host. E. coli, a Gram
negative bacterium, which exists in the human gastrointestinal (GI) tract, is a highly motile
microbe whose migration is constantly guided by the interaction of multiple gradients (mainly
a)
Author to whom correspondence should be addressed. Electronic mail: skdas@iitrpr.ac.in.
1932-1058/2017/11(2)/024108/13/$30.00 Published by AIP Publishing.11, 024108-1
BIOMICROFLUIDICS 11, 024108 (2017)
chemical, arising due to enzymatic secretions from the cells in the GI tract and thermal, arising
due to the metabolism of cells in the GI tract) within the GI tract.
5
Even though most E. coli
strains are non-pathogens in the human system, colonization of harmful strains can lead to dis-
eases of the gastric and the intestinal systems.
7
Hence, it is of great importance to understand
the motility and mobility of E. coli population under the influence of such imposed gradients
for the effective design of preventive drugs and a priori estimation of internal therapeutic
processes.
As chemical and thermal gradients are among the most important agents affecting the
migration pattern of microbial cells within a physiological system, the effects of these individ-
ual gradients on E. coli have been reported by researchers.
5,8–15
Earlier, conventional methods
such as capillary assay
8
and Boyden chamber
9
have been used to generate chemical gradients
to study the chemotaxis of E. coli cells in vitro. Agar plates placed over aluminum blocks with
provision for the flow of hot and cold fluids
10
were used to generate thermal gradients to study
the thermotaxis of E. coli cells. Since the past decade, microfluidic devices has been the better
system than conventional assays for the creation of stable, predictable, reproducible, and mea-
sured gradients which are not possible in conventional assays.
5
This is due to the fact that
microfluidic environments can successfully mimic the requisite physiol ogical conditions, both
qualitatively and quantitatively. There have been reports on individual studies of E. coli cell
migration due to chemical
5,11–13
and thermal gradients
14,15
using microfluidic devices.
Although there have been reports on the effect of these gradients on E. coli cells in micro-
fluidic environments, there is no report on the navigation and behavior of such cells when
exposed to competing thermal and chemical gradients imposed on cells, which closely mimic
the physiological conditions and present during the defense mechanism by the body in conjunc-
tion with external drug molecules. A computational report
1
on migration of E. coli cells in the
presence of competing gradients of chemical, thermal, and pH in a microfluidic device is avail-
able, but a detailed survey of the literature suggests that the robust experimental reports of such
cellular migration in a fluidic microenvironment are really necessary. Understanding cellular
migration under such complex gradients is a need of the hour.
Previous reports on thermotaxis suggest that at the population level (in the absence of any
external chemical change in the environment), bacteria can themselves bring about a chemical
change in the environm ent.
16
This behavior is governed by the number density of the cells at a
particular location, and this can affect the thermotactic characteristics of the population.
16
It has
also been reported that below a threshold population density, E. coli cells behave as hot temper-
ature seeking, whereas above the threshold population density, they were cold temperature seek-
ing.
17
It has been explained that such reversal in behavior is caused by the formation of nutrient
gradients and a high Tar/Tsr ratio at high cell population density. Tar and Tsr are transmem-
brane chemoreceptors which play major roles in chemical and thermal sensing mechanisms of
E. coli.
17
Both Tar and Tsr are warm seeking receptors in the absence of ligands.
16,17
Reports
show that cells grown below 0.1 OD (optical density) are warm seeking since they possess an
abundance of Tsr receptors (Tsr being a warmth seeking receptor; however, in the presence of
ligands such as glycine or serine, it becomes insensitive to temperature). On the contrary, cells
of OD 0.3 are cold seeking since they have abundant Tar receptors (Tar being warmth seeking;
however, in the presence of ligands such as aspartate, it changes as cold seeking). An alterna-
tive transduction pathway through the change in intracellular or extracellular pH differences
other than the chemoreceptor has also been reported.
18
In all existent reports, the utilized attractant (aspartate/serine/glycine) was introduced into
the medium under observation itself before exposing to the temperature gradient (to make it a
nutrient rich medium).
17,18
In such procedures, the cells were not exposed to competing or
aligned chemical and thermal gradients, which are highly plausible during radiotherapy, chemo-
therapy, and induced photothermal hyperthermia.
6
Several pertinent questions may arise based
on the present knowledge. The behavior of bacterial cells and migration patterns of cells in a
situation wherein all the possible migratory options are unfavorable are unanswered questions.
Similarly, the factors that would lead to the bacterial population to make a collective deci-
sion and the time necessary for it to respond to imposed interacting chemical and thermal
024108-2 Murugesan et al. Biomicrofluidics 11, 024108 (2017)
gradients also require proper comprehension. E. coli migration patterns under the influence of
aligned and opposing gradients will shed insights into the behavior of the bacterial population
when the host location tends to be unfavorable or favorable compared to the neighboring micro-
environment which may be unfavorable or favorable. The present study attempts to understand
insights into such regions of knowledge gap by employing physiology, mimicking microenvir-
onments, and the observations are likely to replicate the migration dynamics of parasitic cells
in complex physiological environments.
II. EXPERIMENTAL SECTION
A. Bacterial culture preparation
An aliquot of 500 ll of untransformed Escherichia coli DH5a cells (M/S. Biogenei,
Bangalore, India) was used to inoculate 25 ml LB liquid broth. The culture was grown under
shaking conditions (180 rpm) at 37
C for 12 h. After incubation, the culture was centrifuged for
5 min at 6000g and at 4
C. Then, the sample was prepared by collecting the bacterial pellet
and further suspending in distilled water to an optical density of 0.3.
B. Device design, simulation, and fabrication
The device used for generating a combined gr adient under stagnant conditions of the
microbe chamber consists of 5 channels (3 parallel channels at the centre and 2 channels per-
pendicular to the centre channels, and one on each side). The channel, having a dimension of
10 mm 2mm 0.2 mm, in which the migration has been observed, contains no agarose. The
dimension of the channel in the device is described in detail in the
supplementary material of
our previous publication.
11
The channels parallel to (on either side) the channel under investiga-
tion carry the agarose matrix and are connected to the side channels (source and sink chan-
nels).
11
The above description of the device design can create steady, long range, and flow
free chemical gradients across a stagnant fluid. In addition, the device comprises of two parallel
tubes to supply hot and cold water for temperature gradient generation (cf. Fig.
1).
15
Computations have been performed to gauge the characteristics of the generated gradient since
the chemicals used in the experiments are colorless, and hence, the gradient cannot be deter-
mined quantitatively during experiments. The chemical and thermal gradient generation within
the device has been already characterized and validated with the “solute transport” and
“conjugate heat transfer” modules of COMSOL Multiphysics, 4.2.a., in previous reports by the
same authors.
11,15
Accordingly, the same simulation protocol has been used in the present study
to estimate a priori steady combined chemical (sorbitol/NiSO
4
) and thermal gradient generation
in the designed device. The boundary conditions are similar to those reported by the present
authors.
11,15
A PDMS (polydimethysiloxane) based device was fabricated using optical and soft
lithography techniques.
15
C. Experimental methodology
The fabricated device was autoclaved before the start of the experiments. The agarose
channels were filled with freshly prepared 1.5% agarose and kept undisturbed for 30 min for the
formation of the matrix. Further, E. coli DH5 a cells as obtained from the previous section
(Bacterial culture preparation) were injected in the central channel under investigation through
the cell inlet option provided in the device (cf. Fig. 1(b)) and kept undisturbed for 30 min for
the population to attain the equal distribution of cells throughout the channel. The microfluidic
device loaded with cells was then positioned over the stage of an inverted fluorescent micro-
scope (Leica DMI3000 B, Leica Microsystems, Wetzlar, Germany) along with the fittings for
chemical and thermal gradient generation (cf. Fig.
1).
Once the equal distribution of cells was achieved and confirmed qualitatively by viewing
through the 62 objective of the microscope, both chemical and thermal gradients were gener-
ated simultaneously within the device by flowing respective fluids in the respective channels
and tubes. A stead y chemical gradient was generated by pumping a chemo–effector (1 mM
024108-3 Murugesan et al. Biomicrofluidics 11, 024108 (2017)
sorbitol as an attractor or 1 mM NiSO
4
as an inhibitor) through one side channel and distilled
water through the other side channel, in a parallel flow mode and both at a flow rate of 100 ll/
h. A steady thermal gradient was simultaneously generated by pumpi ng hot (53
C) and cold
water (30
C) through the respective tubes inserted in the device in a counter flow mode at a
flow rate of 4000 ml/h. The surrounding temperature was maintained at 30
C. Fig.
2 gives vari-
ous combinations of thermal and chemical gradients generated. Given the fact that it is not fea-
sible to introduce the cells (at population level) in the main channel without disturbing the
existing gradients, the cells were introduced before generating the gradients.
Images and videos of the cell population were captured at 15 different axial locations of
the channel at a frequency of 30 min up to 2 h, and the images of the cells were analyzed for
the cell count using Image J software. The thermal image of the device was captured using an
infrared camera (FLIR T250, M/S. PCI middle east FZE, UAE), and the same was analyzed
using the associated FLIR Quick Report software to quantify the thermal gradient generated
across the channel under investigation. The cell migration behavior was studied in a combined
gradient microenvironment and further quantified using two population based matrices, viz., the
migration rate (MR) and the apparent migration coefficient (AMC).
11
The definitions of the
migration rate and the apparent migration coefficient have been reported by Murugesan et al.,
and they are given below.
11
The migration rate MR
ðÞ
¼
dN=N
avg
dt
;
where N
avg
¼ the number of cells irrespective of distance before the initiation of the gradient
and N ¼ the count at different times, after the initiation of the gradient, at specified distance in
the said channel. If the gradient of an unknown compound is considered from these values, it
will be easier to interpret whether the unknown compound is an attractant or a repellant. This
facilitates the detection of the unknown compound.
FIG. 1. (a) Sketch of the experimental setup for the cell migration study under imposed competing chemical and thermal
gradients. (b) Schematic illustration of the channels in the device. (c) Schematic view of the microfluidic device under
working conditions.
024108-4 Murugesan et al. Biomicrofluidics 11, 024108 (2017)
The apparent migration coefficient AMC
ðÞ
¼
X
5
i¼1
1 N=N
avg
a
;
where a ¼ the number of observation points and i ¼ the point or location at which the cell count
is made. If the microchannel is hypothetically divided into two halves, then AMC is the net
shift of the cells from one half to another. The clock has been started around 20 min from the
establishment of the gradients (i.e., 10 min) fo r the MR and AMC plots. The total number of
cells in the channel under investigation under the key conditions of the experiment during the
experimental duration (2 h) has been given in Figs. S1 to S7 (cf.
supplementary material)to
show that the total number of cells in the channel under investigation is the same under all the
conditions and that the migration results are not affected by this.
III. RESULTS AND DISCUSSION
A. Characteristics of the microfluidic environment
A steady combined chemical and thermal gradient has been generated in a stagnant micro-
fluidic environment. It took 10 min for the device to establish chemical and thermal gradients.
With an initial concentration of 1 mM of the chemo-effectors, the microfluidic device was able
to generate a concentration profile ranging from 0.2 mM at one end of the channel to 0.8 mM at
the other end of the channel (as obtained from simulation and illustrated in Figure
3).
FIG. 2. Combined gradient generation within the microfluidic device. (a) Opposing gradients of the attractant (sorbitol) and
temperature. (b) Aligned gradients of the attractant and temperature (c) Aligned gradients of the inhibitor (NiSO
4
) and tem-
perature. (d) Opposing gradients of the inhibitor and temperature.
024108-5 Murugesan et al. Biomicrofluidics 11, 024108 (2017)