Synthetic control of mammalian-cell motility by
engineering chemotaxis to an orthogonal
bioinert chemical signal
Jason S. Parka,b,c, Benjamin Rhaua,b, Aynur Hermanna,b, Krista A. McNallya,b, Carmen Zhoua,b, Delquin Gongb,d,
Orion D. Weinerb,d, Bruce R. Conklina,b,c,d,e, James Onuffera,b, and Wendell A. Lima,b,f,1
aDepartment of Cellular and Molecular Pharmacology,bThe Cell Propulsion Lab, a National Institutes of Health Nanomedicine Development Center,
University of California, San Francisco, CA 94158;
and Department of Biochemistry, University of California, San Francisco, CA 94143; andeDepartment of Medicine andfHoward Hughes Medical Institute,
University of California, San Francisco, CA 94158
cGladstone Institute of Cardiovascular Disease, San Francisco, CA 94158;dCardiovascular Research Institute
Edited by Peter N. Devreotes, The Johns Hopkins University School of Medicine, Baltimore, MD, and approved March 13, 2014 (received for review
February 6, 2014)
Directed migration of diverse cell types plays a critical role in
biological processes ranging from development and morphogenesis
to immune response, wound healing, and regeneration. However,
techniques to direct, manipulate, and study cell migration in vitro
and in vivo in a specific and facile manner are currently limited. We
conceived of a strategy to achieve direct control over cell migration
to arbitrary user-defined locations, independent of native chemo-
taxis receptors. Here, we show that genetic modification of cells
with an engineered G protein-coupled receptor allows us to redirect
their migration to a bioinert drug-like small molecule, clozapine-N-
oxide (CNO). The engineered receptor and small-molecule ligand
form an orthogonal pair: The receptor does not respond to native
ligands, and the inert drug does not bind to native cells. CNO-
responsive migration can be engineered into a variety of cell types,
including neutrophils, T lymphocytes, keratinocytes, and endothelial
cells. The engineered cells migrate up a gradient of the drug CNO
and transmigrate through endothelial monolayers. Finally, we dem-
onstrate that T lymphocytes modified with the engineered receptor
can specifically migrate in vivo to CNO-releasing beads implanted in
a live mouse. This technology provides a generalizable genetic tool
to systematically perturb and control cell migration both in vitro and
in vivo. In the future, this type of migration control could be a valu-
able module for engineering therapeutic cellular devices.
GPCR|cellular therapeutics|synthetic biology
is critical for their proper function. For example, immune cells
rapidly home to sites of infection, concentrating their powerful
cytotoxic and proinflammatory activities for maximum efficacy
while limiting damage to healthy tissue. In morphogenesis, cells
undergo a complex stereotyped process involving migration as well
as proliferation, differentiation, and programmed cell death to
produce fully developed multicellular structures. In wound healing
and regenerative processes, stem and progenitor cells home to
injured tissues from nearby sites—as well as from distant locations
including the bone marrow—to provide a stream of new cells to
replenish and provide trophic support to old and damaged cells.
Cell migration is also an important factor to consider in the
use of cells as therapeutic agents. The use of cells for the treat-
ment of a growing array of diseases including cancer, autoimmu-
nity, and chronic wounds is currently being explored (1–6). The
appropriate and efficient localization of therapeutic cells to sites
of disease has been identified as an important factor for successful
cell-based therapy (7–17). However, preclinical studies and clinical
trials to date have shown that the homing to sites of disease of
many cell types commonly used as therapeutics is frequently im-
paired or limited, especially after ex vivo expansion of cells in
culture (7, 12, 18, 19).
he ability of many cell types to migrate long distances within
the body and specifically localize to target sites of action
The ability to redirect the migration of cells to any user-
specified location in the body would be a powerful enabling
technology for basic research as well as for future applications, but
there are currently few easily generalizable strategies to accomplish
this goal. We conceived of an approach to direct cellular homing to
small molecules by expressing, in motile cells, engineered G protein-
coupled receptors (GPCRs) called receptors activated solely by a
synthetic ligand (RASSLs) (20, 21).
RASSLs are engineered to be unresponsive to endogenous
ligands but can be activated by pharmacologically inert orthog-
onal small molecules (Fig. 1A). Versions of these receptors exist
for the three major GPCR signaling pathways (Gαs-, Gαi-, and
Gαq-coupled receptors), and the design of a new arrestin-biased
variant has recently been reported (21, 22). Because GPCRs
control many important physiological functions, including cell
migration, we hypothesized that, by expressing these engineered
receptors in motile cells, we could develop a general strategy for
establishing user control over cell homing (Fig. 1B). Here, we
use a family of second-generation RASSLs, known as designer
receptors exclusively activated by a designer drug (DREADDs),
that are activated only by the small molecule clozapine-N-oxide
(CNO), an inert metabolite of the FDA-approved antipsychotic
drug clozapine (Fig. S1) (20). CNO is highly bioavailable in
rodents and humans, lacks affinity for any known receptors,
Directed migration of diverse cell types is critical in biological
processes ranging from development and morphogenesis to
immune response, wound healing, and regeneration. However,
techniques to specifically and easily direct, manipulate, and
study cell migration in vitro and in vivo are currently limited.
We conceived of a strategy to directly control cell migration to
arbitrary user-defined locations, independent of native che-
motaxis receptors. In this work, we demonstrate that genetic
modification of cells with an engineered G protein-coupled
receptor allows us to redirect their migration to a bioinert
drug-like small molecule, clozapine-N-oxide. This technology
provides a generalizable tool to systematically control cell mi-
gration in vitro and in vivo and could be a valuable module for
engineering future therapeutic cellular devices.
Author contributions: J.S.P., B.R., A.H., K.A.M., C.Z., D.G., O.D.W., B.R.C., J.O., and W.A.L.
designed research; J.S.P., B.R., A.H., K.A.M., C.Z., and D.G. performed research; J.S.P., B.R.,
A.H., K.A.M., C.Z., D.G., O.D.W., B.R.C., J.O., and W.A.L. analyzed data; and J.S.P. and W.A.L.
wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| April 22, 2014
| vol. 111
| no. 16www.pnas.org/cgi/doi/10.1073/pnas.1402087111
channels, and transporters, and does not cause any appreciable
physiological effects when systemically administered in normal
mice (20, 23, 24).
Identification of Orthogonal GPCR That Controls HL-60 Neutrophil
Motility. To rapidly test whether this family of engineered
orthogonal receptors could be used to control cell morphology
and motility, we first transiently expressed several variants of
these receptors (Dq, Di3, and Di) along with green fluorescent
protein (GFP) in HL-60 neutrophils. Transfection efficiencies
were routinely 40–45%, as measured by coelectroporation with
GFP and determination of % GFP-positive cells via flow cytom-
etry. We tested these engineered cells in a high-throughput im-
pedance-based adhesion/spreading assay in which cells are plated
on a fibronectin-coated electrode array and exposed to putative
chemoattractants (Fig. 1C). Cells that morphologically respond to
the chemoattractant adhere tightly to the surface and spread out,
and this cytoskeletal change is measured as an increase in electrical
impedance in real-time. We found that cells expressing the Gαi-
coupled receptors Di3 and Di responded to the drug CNO whereas
cells expressing the Gαq-coupled receptor Dq did not. This result
was consistent with the known fact that many natural Gαi-coupled
receptors are associated with chemotaxis (25). None of the cells
responded to vehicle treatment, and all of the cells maintained a
strong response to the positive control chemoattractant formyl-
Met-Leu-Phe (fMLP), which strongly attracts neutrophils (Fig. 1C
and Fig. S2). fMLP also induced a strong cell-spreading response
in Di receptor and vector control-transfected HL-60 neutrophils
We tested whether HL-60 neutrophils expressing the same
three engineered receptors would migrate directionally through
a porous membrane in response to a gradient of the drug CNO
in a Boyden-chamber transwell migration assay (Fig. 1D). The
number of migrating cells was quantitated by flow cytometry
using a fluorescent bead-counting standard. Consistent with the
results of the cell-spreading assay, cells expressing the Gαi-coupled
receptors Di3 and Di migrated in response to a gradient of CNO
whereas cells expressing the Gαq-coupled receptor Dq did not. All
of the cells maintained a strong migratory response to the positive
control chemoattractant, fMLP (Fig. 1D).
It is well known that polarization and cell migration in neu-
trophils involves highly conserved cellular signaling and positive
feedback loops that include the activation of the Rho-family
GTPase Rac and the generation of phosphatidylinositol-(3,4,5)-
Tris-phosphate by phosphotidylinositol 3-kinase (PI3K) at the
leading edge of the migrating cell. To confirm that these pathways
are activated in Di-expressing HL-60 neutrophils in response
to CNO stimulation, we stimulated cells in suspension and
performed immunoblotting for phosphorylated Akt and phos-
phorylated PAK as readouts for PI3K activity and Rac activity,
respectively (Fig. 1E and Fig. S4). We observed that, upon CNO
stimulation, levels of phosphorylated Akt and PAK significantly
increased in Di-expressing, but not control, cells. In contrast,
upon stimulation with the natural chemoattractant fMLP, levels
of phosphorylated Akt and PAK increased in both Di and con-
trol cells. Interestingly, the amplitude and duration of phospho-
Akt and phospho-PAK were slightly higher in Di-expressing
cells, both in response to CNO and fMLP (Fig. S4).
adhesion / spreading assay
boyden chamber assay
- = vehicle
+ = pos ctrl
- = vehicle
+ = pos ctrl
+ = pos ctrl
changes and chemotaxis of HL-60 neutrophils in response to CNO. (A) RASSLs
are engineered GPCRs that interact orthogonally with a bioinert small-
molecule drug. Natural ligands do not interact with the engineered recep-
tors, and the bioinert drug that activates the engineered receptors does
not interact with native receptors. (B) We tested whether certain second-
generation RASSLs known as DREADDs could mediate cell motility. (C)
Changes in electrical impedance that result from cell spreading in response
to drug or ligand are detected by an electrode array. HL-60 neutrophils
transiently transfected to express engineered GPCRs were plated on fibro-
nectin-coated impedance assay plates and stimulated with vehicle control,
100 nM fMLP (positive control chemoattractant) or 100 nM CNO. All cells
responded to fMLP whereas only Di3- or Di-expressing cells responded to
CNO. Mean ± SEM for n = 3 replicates is shown. (D) Cell migration of HL-60
neutrophils transiently transfected with engineered GPCRs was quantitated
in a porous transwell Boyden-chamber assay. All cells migrated in response
to fMLP whereas only Di3- or Di-expressing cells migrated in response to
CNO. Drug concentrations used: 100 nM CNO, 100 nM fMLP. Mean ± SEM for
n = 3 replicates is shown. (E) Polarization and cell migration in neutrophils
involves Rac and PI3K activation. Di-expressing HL-60 neutrophils were
treated with 100 nM fMLP or 100 nM CNO before immunoblotting for
phosphorylated Akt and phosphorylated PAK as readouts for PI3K and Rac
activity, respectively. Peak levels of phospho-Akt and phospho-PAK are
shown for each condition. Both were increased by CNO stimulation in Di cells
Engineered Gαi-coupled GPCRs Di3 and Di mediate cytoskeletal
but not in control cells (P < 0.01 by Student t test). Stimulation with fMLP
increased phospho-Akt and phospho-PAK levels in both Di and control cells
(P < 0.01 by Student t test), but Di cells showed higher peak levels of
phospho-Akt than did control cells (P < 0.01 by Student t test). Three (for
CNO) or four (for fMLP) independent experiments were performed and
mean ± SEM are shown.
Park et al.PNAS
| April 22, 2014
| vol. 111
| no. 16
Finally, we tested whether uniform stimulation with CNO is
sufficient to induce polarization, symmetry breaking, and random
motility in unpolarized Di-expressing HL-60 neutrophils, as is
known to be the case with natural chemoattractants, including
fibronectin-coated glass, and treated with CNO while being ob-
served via time-lapse microscopy. We observed that Di-expressing
changes and motile behaviors characteristic of neutrophils un-
dergoing chemokinesis upon treatment with CNO (Movie S1).
Directed Migration in a CNO Gradient. Next, we used a micropipet
migration assay with time-lapse microscopy to visualize the dy-
namic process of migration. This assay allows for visualization of
individual cell behavior and provides (i) a very steep local con-
centration gradient and (ii) the ability to rapidly move the source
of the gradient (Fig. S5A). Transiently transfected HL-60 neu-
trophils expressing Di and GFP (as a coelectroporation control)
migrated robustly and directionally to the micropipet point
source of CNO whereas cells transfected with an irrelevant plas-
mid control exhibited random migration (Fig. 2A and Movies S2–
S4). Further, cells migrating to CNO were able to reorient to
a changing gradient of the drug as can be appreciated when the
micropipet is moved in Movie S3.
To facilitate further quantitation of migration metrics of
engineered HL-60 neutrophil chemotaxis in vitro, we used
a microfluidic gradient generator developed and optimized in
collaboration with the CellASIC Corporation. The microfluidic
device is capable of generating a smooth, steady gradient over
a relatively large area, allowing the user to track and analyze
many cells within a field of view that are all experiencing a fairly
consistent chemical gradient environment (Fig. S5B). To im-
prove the homogeneity of receptor expression, we also generated
HL-60 cell lines stably expressing the Di receptor with a YFP
fluorescent protein fusion. Cells were loaded into the micro-
fluidic device and allowed to adhere to the fibronectin-coated
glass surface, and unbound cells were washed away, as can be
seen at the beginning of Movie S5. A diffusive CNO gradient was
applied (visualized by a fluorescent red tracer dye), and cells
were tracked by time-lapse microscopy. Image analysis was per-
formed, and cell tracks were generated, with initial cell positions
plotted at the origin (Fig. 2B). Di receptor-expressing cells mi-
grated directionally in response to the CNO gradient compared
with vehicle control, as determined by comparing the track ve-
locity, displacement rate, and directionality metrics between the
two treatment conditions (Fig. 2B). In a separate experiment, we
observed that Di receptor-expressing cells also migrated direc-
tionally toward the positive control chemoattractant fMLP, with
grossly comparable fold increases in track velocity, displacement
rate, and directionality in the presence versus the absence of
chemoattractant as in CNO experiments (Fig. S6 and Movie S6).
Orthogonal Control of Chemotaxis in Diverse Cell Types. Having
established that the Di receptor is a potent mediator of CNO
chemotaxis in HL-60 neutrophils, we asked whether this engi-
neered chemotaxis receptor is “portable” to other cell types. We
therefore generated a lentiviral vector to efficiently express an
mCherry fluorescent protein-tagged Di receptor construct in a
variety of cell types. Stable Di receptor-expressing cell lines were
then established from HL-60 cells, primary human T lymphocytes,
primary neonatal human epidermal keratinocytes, and primary
human umbilical vein endothelial cells (HUVECs) (Fig. 3A).
Expression of the Di receptor did not cause alterations in gross
cellular morphology, and cells expressing, and not expressing, Di
appeared indistinct on microscopic examination (Fig. S7).
We tested each of the above cell types in Boyden-chamber
transwell migration assays. In each case, Di receptor-expressing
cells migrated in response to a gradient of CNO. Control cells not
expressing the Di receptor did not migrate in response to CNO
(Fig. 3B). We also performed transwell checkerboard control
experiments, in which the putative chemoattractant is placed in
the top and/or bottom chamber of the transwell in all combinations
to distinguish between cellular chemotaxis (directed migration up
a gradient of chemoattractant) and chemokinesis (increased mo-
tility in the presence of chemoattractant). In these experiments,
HL-60 neutrophils and T lymphocytes exhibited directed mi-
gration toward CNO (chemotaxis) whereas keratinocytes and
endothelial cells showed only increased motility in the pres-
ence of CNO (chemokinesis) (Fig. S8).
Cellular migration in the body is complicated by mammalian
anatomy. A critical step in homing for cells that travel via the
bloodstream to reach target sites is exiting blood vessels to enter
surrounding tissues—a process known as diapedesis or trans-
endothelial migration (Fig. 4A). Therefore, in our next experi-
ment, we tested whether motile cells expressing the Di receptor
could migrate through an endothelial monolayer in vitro in re-
sponse to a gradient of CNO. We grew a tight monolayer of
HUVECs on a fibronectin-coated porous transwell membrane
gration in response to CNO. (A) HL-60 neutrophils coelectroporated with Di
and GFP were plated on a fibronectin-coated glass surface and observed by
time-lapse microscopy in the presence of a steep, micropipette-generated
gradient of CNO. Di- and GFP-expressing cells migrated directionally toward
the micropipet. Fluorescent dye Alexa 594 tracer is mixed with CNO solution
in micropipet to visualize the diffusive gradient. The micropipet gradient
source is marked by a magenta asterisk. Track start locations are marked by
black squares, and red triangles mark cell location and direction in each
frame. Traces (black and gray) connect track start locations (black squares)
and cell location (red triangles). Drug concentration used (at source): 1μM
CNO. See Movies S2–S4 for full movies. (B) HL-60 neutrophils stably expressing
Di were placed in the fibronectin-coated viewing area of a microfluidic che-
motaxis assay device capable of generating a smooth, stable gradient of CNO.
Time-lapse microscopy was used to track cell migration, and cell-tracking soft-
ware was used to quantitate various migration metrics. Cells migrated toward
the CNO gradient (trajectories plotted with cell start locations at origin) and
show increased track velocity, displacement rate, and directionality compared
with basal motility in the presence of vehicle control. Drug concentration used
(at source): 200 nM CNO. Mean ± SEM is shown for n = 61 cells tracked (**P <
0.0001 by Student t test). See Movie S5 for full movie.
Microscopic analysis of HL-60 neutrophil polarization and cell mi-
| www.pnas.org/cgi/doi/10.1073/pnas.1402087111Park et al.
for 4 d. Monolayer integrity was assessed by an observed increase
in transendothelial electrical resistance from a baseline of <7 Ω
to >60 Ω and barrier function in an FITC-dextran permeability
assay (Fig. S9). We then proceeded with a transwell migration
assay using HL-60 neutrophils and primary human T lympho-
cytes as the motile cell types. Both engineered HL-60 neutrophils
and primary human T lymphocytes exhibited a directed trans-
endothelial migration response to CNO as well as to a positive
control chemoattractant (fMLP for HL-60 neutrophils and SDF-
1a for T lymphocytes) (Fig. 4B).
Cells with Engineered Receptor Home to CNO Signal in Vivo. Finally,
we tested whether our approach of redirecting cellular homing
using a small-molecule drug could be feasible for use in vivo.
T lymphocytes are highly motile cells of the adaptive immune
system that play critical roles in cell-mediated immunity. Their
use is currently being heavily explored in cell-based therapeutic
applications in human clinical trials and in preclinical models,
especially in cancer and autoimmunity (1, 2, 26, 27). We there-
fore tested whether the homing of engineered T lymphocytes
could be redirected to the orthogonal CNO signal in a mouse.
Mouse T lymphocytes were retrovirally transduced with a bicis-
tronic construct encoding both an mCherry-tagged Di receptor
and an enhanced firefly luciferase to allow tracking of modified
cells (28). Biodegradable CNO-loaded poly-lactide-coglycolide
(PLGA) microspheres were formulated using standard techni-
ques to generate a slow-release source of CNO in the body (Fig.
S10). The encapsulated drug concentration was determined to be
4.1 μg/mg (encapsulation efficiency of 19.6%). Vehicle control
microspheres were generated in parallel by omission of CNO
in the protocol. CNO-loaded and vehicle control microspheres were
injected s.c. (suspension in PBS) into opposing flanks of albino B6
mice. Di receptor- and luciferase-transduced T lymphocytes were
injected i.v. via the lateral tail vein.
In this experiment, we observed that the Di receptor-expressing
T lymphocytes preferentially localized to sites of injection of
CNO-loaded beads versus vehicle control beads injected on the
contralateral flanks (Fig. 5A). This preferential localization was
also observed in mice where the injected flanks were switched
(CNO-left and vehicle-right versus CNO-right and vehicle-left
flank) (data points combined and analyzed together in Fig. 5B).
The luminescence of the T cells localized at each site was quan-
titated at 6 h, 4 d, and 7 d after T-cell injection (Fig. 5 B and C).
This study was also performed with T cells expressing luciferase
but not the Di receptor—in this case, these negative control cells
did not show preferential localization to CNO slow-release
microspheres (Fig. S11).
The technology we describe here represents a step forward in the
development of generalizable genetic tools with user-defined
orthogonal control for the study of cell migration in vitro and in
vivo. Of course, further work remains to optimize this technology.
For example, the small-molecule drug could be modified through
synthetic chemistry to optimize its properties as a gradient-generating
homing molecule. Alternative delivery formulations of the drug
(such as smart liposomes with antibody-based targeting and trig-
gered release characteristics) (29) could be used for delivery to
sites of disease in a targeted manner. In the longer term, it may be
possible to develop genetically encodable orthogonal receptor/li-
gand pairs to allow for biological expression of the homing signal
by cells. Protein engineering of the receptor could also be used to
develop variants with altered drug affinity, recycling properties, or
signaling capabilities. Such tools will allow researchers to uncouple
the control of motility from other signals and give them the ability
to systematically perturb motility and understand its role in diverse
processes such as development, immune response, wound healing,
An orthogonal tool to control cell migration like the one de-
scribed here could be of value not only as a research tool, but
types. (A) Gene construct with N-terminal signal sequence followed by
mCherry fluorescent protein fused to Di was inserted into a lentiviral plasmid
backbone for viral expression in various cell types. (B) HL-60 neutrophils,
primary human T lymphocytes, primary human epidermal keratinocytes, and
primary human umbilical vein endothelial cells were transduced to stably
express the Di receptor and tested for migration in the presence of a CNO
gradient in Transwell experiments. All of the above cell types exhibited in-
creased migration through the Transwell membrane in the presence of CNO
compared with vehicle control. Drug concentrations used: 25 nM CNO for T
lymphocytes, 100 nM CNO for all other cell types. Mean ± SEM is shown for
three repeats (***P < 1e−4, **P = 0.001, *P = 0.02 by Student t test).
The engineered chemotaxis receptor Di is portable to a range of cell
and transendothelial migration in immune cells. (A) Transendothelial mi-
gration is a critical step in the overall process of cellular homing that also
includes adhesion to endothelium and chemotaxis. (B) HL-60 neutrophils and
primary human T lymphocytes stably expressing Di were tested for their
ability to transmigrate through a tight endothelial monolayer grown on a
porous fibronectin-coated transwell membrane in response to both a CNO
gradient as well as a positive control chemoattractant (100 nM fMLP and
50 ng/mL SDF-1a, respectively). Both cell types exhibited migration in the
presence of CNO. Mean ± SEM for n = 3 (HL-60) or n = 4 (T lymphocytes)
replicates is shown (**P < 1e−4, *P = 0.02 by Student t test).
The engineered receptor Di is sufficient to mediate both chemotaxis
Park et al.PNAS
| April 22, 2014
| vol. 111
| no. 16
also in the future as applied in the emerging field of cell-based
therapeutics. For example, antitumor T-cell trafficking into tumors
is often quite inefficient, despite being critical for antitumor ac-
tivity: It has been observed that increased tumor infiltration by
T cells correlates with better prognoses in mouse studies and in
human clinical trials (10–15, 19, 30–34). There are currently limited
ways for physicians to steer cells to desired sites, however. Most
cells that are currently used in clinical trials, including immune cells
and stem cells, largely rely on the natural “tropism” of particular
cell types for certain tissues (e.g., hematopoietic stem cell homing
to the bone marrow niche) (35) or for disease-associated signals
[e.g., mesenchymal and neural stem cell homing to inflammation
(9, 18, 36, 37) or monocytes into tumors (38, 39)]. The use of a
simple system to guide cellular localization in the body to arbitrary
locations could in principle allow physicians to more effectively
harness powerful cellular therapeutic activities, including cell kill-
ing, repair/regeneration, sensing disease (40, 41), and delivering
therapeutic molecules (42–44) to treat disease, and potentially
broaden the range of uses for cells in medicine.
The use of a bioinert drug to orthogonally direct engineered
cell migration is conceptually distinct from (and complementary
to) past strategies reported in the literature for directing cell
migration. Other groups have described interesting approaches,
including the chemical or enzymatic modification of the cell
surface with specific adhesion molecules (45), materials engi-
neering of artificial scaffolds and tunable matrices to direct cell
adhesion and migration (46), expression or direct injection of
natural homing ligands such as chemokines into sites where in-
creased cell migration is desired (47, 48), and the expression in
therapeutic cells of natural receptors such as chemokine recep-
tors whose ligands are up-regulated in inflammation or cancer
(8, 7, 16, 17). These strategies rely on naturally existing homing
receptors and ligands, and they are powerful because they tap
into cells’ native migration axes. However, many homing ligands
are present in multiple locations throughout the body, the ex-
pression of these ligands may vary in time throughout the natural
course of disease or in response to therapy, many ligands (such
as chemokines) interact with multiple receptors and vice versa,
and native receptors for natural ligands can sometimes be found
not only on therapeutic cell types but also on cell types that are
detrimental for therapy (47, 49–52). In contrast, the work we
have demonstrated here benefits from the use of an orthogonal
receptor–drug pair. The drug has a low toxicity profile, which
decreases concerns of side effects in therapeutic settings. The
homing receptor for the drug is expressed uniquely on the cell
type of one’s choosing (and not on native cells). The user can
better control when, where, and how much drug is present at
a given site, and the drug cannot naturally be produced at off-
target sites. Cellular homing can be directed not only to sites of
disease where there are known chemotactic ligands or migration
signals, but also to any site where a drug can be delivered.
Another intriguing strategy to gain control over cell migration
is the use of light-sensitive proteins such as photoactivatable Rac
or opsin photoreceptors to tap into cell motility signaling path-
ways (53–55). These types of tools have already yielded valuable
insights into the basic biology and mechanism of cell migration in
vitro as well as in vivo in the optically transparent zebrafish
model. So far, however, the requirement for consistent in vivo
delivery of light remains an obstacle to the broader use of
so-called optogenetic tools in vivo and in therapeutic contexts.
We have demonstrated a simple approach to directing cell
migration in vitro and in vivo in a variety of cell types. A para-
digm of gaining synthetic control over complex cellular behaviors
using engineered proteins that respond to orthogonal chemical
signals is likely to be generally useful for basic research and in
future biotechnological and therapeutic applications.
Materials and Methods
A complete detailed description of materials and methods is provided in
SI Materials and Methods. Gene constructs were cloned using standard
molecular biology methods. The DREADD constructs were a generous gift
from Dr. Bryan Roth (University of North Carolina Medical School, Chapel
Hill, NC). The enhanced firefly luciferase gene (effLuc) was a generous gift
from Dr. Brian Rabinovich (M. D. Anderson Cancer Center, Houston). Stan-
dard sterile culture methods were used for cell culture and viral supernatant
production. The xCELLigence RTCA MP impedance array assay platform
(ACEA Biosciences/Roche) was used to monitor cytoskeletal changes (adhe-
sion/spreading) of HL-60 neutrophils on fibronectin-coated wells in response
to agonist. Boyden-chamber assays were used to assess the migration of cells
through porous membranes. Standard cell lysis and immunoblotting procedures
T lymphocyte imaging
expressing Di receptor &
time after T cell injection
fold T cell localization
(CNO vs vehicle)
6 hr 4 days 7 days
specifically localize to an s.c. implanted depot of CNO slow-release bio-
degradable microspheres. (A) Mouse T lymphocytes expressing Di and firefly
luciferase (to enable in vivo bioluminescent imaging) were systemically ad-
ministered (intravenously) to mice in which CNO-releasing biodegradable
PLGA microspheres were implanted s.c. Bioluminescent imaging was used to
track cell localization. T lymphocytes expressing Di specifically localized to
CNO-releasing microspheres compared with vehicle control microspheres
implanted on the contralateral flank. Location of spleen is denoted by the
letter “s.” (B) Quantitative analysis of bioluminescent imaging was per-
formed. Specific localization of T lymphocytes persists for at least 7 d.
Quantitation shown for 4 and 7 d postinjection of T lymphocytes and for two
different doses of implanted microspheres (analyzed for statistical signifi-
cance separately). Microsphere injection doses were 2 mg (triangles) and
6 mg (squares). Mean shown for n = 6 mice for each microsphere dose
(dashed line for 2-mg dose, solid line for 6-mg dose) (*P = 0.013 for day 4,
2 mg; P = 0.022 for day 4, 6 mg; P = 0.017 for day 7, 2 mg; P < 0.001 for day 7,
6 mg) (by Student t test). (C) Fold-differences in T lymphocyte luminescent
signal in CNO microsphere-injected flanks (black circles) versus vehicle mi-
crosphere-injected flanks (gray circles) at 6 h, 4 d, and 7 d after T lymphocyte
injection. Mean ± SEM shown for n = 6 mice (*P < 0.01 by Student t test).
Intravenously administered primary T lymphocytes expressing Di
| www.pnas.org/cgi/doi/10.1073/pnas.1402087111Park et al.
were used to assay protein phosphorylation in stimulated cells by Western blot. Download full-text
Micropipet gradients for chemotaxis assays were generated using the Narishige
MM-89 micromanipulator and glass capillaries pulled on a Sutter Model P-97.
The ONIX microfluidic platform with M04G gradient generator plate (CellASIC/
EMD Biosciences) was used to study HL-60 neutrophil migration. Biodegradable
microspheres loaded with CNO were generated in a sterile environment using a
standard oil-in-water emulsion method. Animal studies were conducted with the
University of California, San Francisco (UCSF) Preclinical Therapeutics Core under
a protocol approved by the UCSF Institutional Animal Care and Use Committee.
ACKNOWLEDGMENTS. We thank the University of California, San Francisco
(UCSF) Preclinical Therapeutics Core Facility (especially Byron Hann, Don
Hom, Donghui Wang, and Paul Phojanakong) for mouse experimental
support and helpful discussions. We also acknowledge the 2009 UCSF Inter-
national Genetically Engineered Machine (iGEM) competition team (espe-
cially Katja Kolar, Ryan Liang, Cathy Liu, Hansi Liu, Jackie Tam, and Eric
Wong) for their work on HL-60 neutrophil chemotaxis experiments and mo-
lecular cloning. This work was supported by National Institutes of Health (NIH)
Grant R01 HL60664-07 (to B.R.C.), pilot study funds from the Gladstone Insti-
tutes, NIH Nanomedicine Development Center Grant PN2EY016546 (The Cell
Propulsion Laboratory: Center for Synthetic Signaling and Motility Systems
Engineering) (to W.A.L.), NIH Grant P50 GM08187 (to W.A.L.), the National
Science Foundation Synthetic Biology Engineering Research Center, NIH
Grant R01 GM084040 (to O.D.W.), a California Institute for Regenerative
Medicine fellowship (Grant TG2-01153) (to J.S.P.), and the Howard Hughes
Medical Institute (W.A.L.).
1. Tang Q, Bluestone JA, Kang S-M (2012) CD4(+)Foxp3(+) regulatory T cell therapy in
transplantation. J Mol Cell Biol 4(1):11–21.
2. Burrell BE, Nakayama Y, Xu J, Brinkman CC, Bromberg JS (2012) Regulatory T cell
induction, migration, and function in transplantation. J Immunol 189(10):4705–4711.
3. Restifo NP, Dudley ME, Rosenberg SA (2012) Adoptive immunotherapy for cancer:
Harnessing the T cell response. Nat Rev Immunol 12(4):269–281.
4. Riddell SR, et al. (1992) Restoration of viral immunity in immunodeficient humans by
the adoptive transfer of T cell clones. Science 257(5067):238–241.
5. Lu P, et al. (2012) Long-distance growth and connectivity of neural stem cells after
severe spinal cord injury. Cell 150(6):1264–1273.
6. Bliss T, Guzman R, Daadi M, Steinberg GK (2007) Cell transplantation therapy for
stroke. Stroke 38(2, Suppl):817–826.
7. Kershaw MH, et al. (2002) Redirecting migration of T cells to chemokine secreted from
tumors by genetic modification with CXCR2. Hum Gene Ther 13(16):1971–1980.
8. Moon EK, et al. (2011) Expression of a functional CCR2 receptor enhances tumor
localization and tumor eradication by retargeted human T cells expressing a mesothelin-
specific chimeric antibody receptor. Clin Cancer Res 17(14):4719–4730.
9. Ankrum J, Karp JM (2010) Mesenchymal stem cell therapy: Two steps forward, one
step back. Trends Mol Med 16(5):203–209.
10. Cohen PJ, Lotze MT, Roberts JR, Rosenberg SA, Jaffe ES (1987) The immunopathology
of sequential tumor biopsies in patients treated with interleukin-2: Correlation of
response with T-cell infiltration and HLA-DR expression. Am J Pathol 129(2):208–216.
11. Cole DJ, et al. (1994) Histopathological analysis of metastatic melanoma deposits in
patients receiving adoptive immunotherapy with tumor-infiltrating lymphocytes.
Cancer Immunol Immunother 38(5):299–303.
12. Pockaj BA, et al. (1994) Localization of 111indium-labeled tumor infiltrating lym-
phocytes to tumor in patients receiving adoptive immunotherapy: Augmentation
with cyclophosphamide and correlation with response. Cancer 73(6):1731–1737.
13. Galon J, et al. (2006) Type, density, and location of immune cells within human co-
lorectal tumors predict clinical outcome. Science 313(5795):1960–1964.
14. Tosolini M, et al. (2011) Clinical impact of different classes of infiltrating T cytotoxic
and helper cells (Th1, th2, treg, th17) in patients with colorectal cancer. Cancer Res
15. Mitchell MS, et al. (2002) Phase I trial of adoptive immunotherapy with cytolytic T
lymphocytes immunized against a tyrosinase epitope. J Clin Oncol 20(4):1075–1086.
16. Craddock JA, et al. (2010) Enhanced tumor trafficking of GD2 chimeric antigen re-
ceptor T cells by expression of the chemokine receptor CCR2b. J Immunother 33(8):
17. Di Stasi A, et al. (2009) T lymphocytes coexpressing CCR4 and a chimeric antigen re-
ceptor targeting CD30 have improved homing and antitumor activity in a Hodgkin
tumor model. Blood 113(25):6392–6402.
18. Karp JM, Leng Teo GS (2009) Mesenchymal stem cell homing: The devil is in the de-
tails. Cell Stem Cell 4(3):206–216.
19. Rosenberg SA, et al. (1994) Treatment of patients with metastatic melanoma with
autologous tumor-infiltrating lymphocytes and interleukin 2. J Natl Cancer Inst
20. Armbruster BN, Li X, Pausch MH, Herlitze S, Roth BL (2007) Evolving the lock to fit the
key to create a family of G protein-coupled receptors potently activated by an inert
ligand. Proc Natl Acad Sci USA 104(12):5163–5168.
21. Conklin BR, et al. (2008) Engineering GPCR signaling pathways with RASSLs. Nat
22. Nakajima K, Wess J (2012) Design and functional characterization of a novel, arrestin-
biased designer G protein-coupled receptor. Mol Pharmacol 82(4):575–582.
23. Bender D, Holschbach M, Stöcklin G (1994) Synthesis of n.c.a. carbon-11 labelled
clozapine and its major metabolite clozapine-N-oxide and comparison of their bio-
distribution in mice. Nucl Med Biol 21(7):921–925.
24. Ray RS, et al. (2011) Impaired respiratory and body temperature control upon acute
serotonergic neuron inhibition. Science 333(6042):637–642.
25. Neptune ER, Bourne HR (1997) Receptors induce chemotaxis by releasing the beta-
gamma subunit of Gi, not by activating Gq or Gs. Proc Natl Acad Sci USA 94(26):
26. Fischbach MA, Bluestone JA, Lim WA (2013) Cell-based therapeutics: The next pillar of
medicine. Sci Transl Med 5(179):ps7.
27. Rosenberg SA (2012) Raising the bar: The curative potential of human cancer
immunotherapy. Sci Transl Med 4(127):ps8.
28. Rabinovich BA, et al. (2008) Visualizing fewer than 10 mouse T cells with an enhanced
firefly luciferase in immunocompetent mouse models of cancer. Proc Natl Acad Sci
29. Allen TM, Cullis PR (2013) Liposomal drug delivery systems: From concept to clinical
applications. Adv Drug Deliv Rev 65(1):36–48.
30. Rubin JT, Elwood LJ, Rosenberg SA, Lotze MT (1989) Immunohistochemical correlates
of response to recombinant interleukin-2-based immunotherapy in humans. Cancer
Res 49(24 Pt 1):7086–7092.
31. Ogawa M, et al. (1997) Enhanced induction of very late antigen 4/lymphocyte function-
associated antigen 1-dependent T-cell migration to tumor sites following administration
of interleukin 12. Cancer Res 57(11):2216–2222.
32. Pagès F, et al. (2005) Effector memory T cells, early metastasis, and survival in co-
lorectal cancer. N Engl J Med 353(25):2654–2666.
33. Musha H, et al. (2005) Selective infiltration of CCR5(+)CXCR3(+) T lymphocytes in
human colorectal carcinoma. Int J Cancer 116(6):949–956.
34. Ohta M, et al. (2005) The high expression of Fractalkine results in a better prognosis
for colorectal cancer patients. Int J Oncol 26(1):41–47.
35. Peled A, et al. (1999) Dependence of human stem cell engraftment and repopulation
of NOD/SCID mice on CXCR4. Science 283(5403):845–848.
36. Imitola J, et al. (2004) Directed migration of neural stem cells to sites of CNS injury by
the stromal cell-derived factor 1alpha/CXC chemokine receptor 4 pathway. Proc Natl
Acad Sci USA 101(52):18117–18122.
37. Wagner J, Kean T, Young R, Dennis JE, Caplan AI (2009) Optimizing mesenchymal
stem cell-based therapeutics. Curr Opin Biotechnol 20(5):531–536.
38. De Palma M, Murdoch C, Venneri MA, Naldini L, Lewis CE (2007) Tie2-expressing
monocytes: Regulation of tumor angiogenesis and therapeutic implications. Trends
39. Lewis CE, Pollard JW (2006) Distinct role of macrophages in different tumor micro-
environments. Cancer Res 66(2):605–612.
40. Folcher M, Fussenegger M (2012) Synthetic biology advancing clinical applications.
Curr Opin Chem Biol 16(3-4):345–354.
41. Khalil AS, Collins JJ (2010) Synthetic biology: Applications come of age. Nat Rev Genet
42. Ruoslahti E, Bhatia SN, Sailor MJ (2010) Targeting of drugs and nanoparticles to tu-
mors. J Cell Biol 188(6):759–768.
43. Stephan MT, Moon JJ, Um SH, Bershteyn A, Irvine DJ (2010) Therapeutic cell engi-
neering with surface-conjugated synthetic nanoparticles. Nat Med 16(9):1035–1041.
44. Yoo J-W, Irvine DJ, Discher DE, Mitragotri S (2011) Bio-inspired, bioengineered and
biomimetic drug delivery carriers. Nat Rev Drug Discov 10(7):521–535.
45. Sarkar D, et al. (2011) Engineered cell homing. Blood 118(25):e184–e191.
46. Lutolf MP, Hubbell JA (2005) Synthetic biomaterials as instructive extracellular
microenvironments for morphogenesis in tissue engineering. Nat Biotechnol 23(1):
47. Homey B, Müller A, Zlotnik A (2002) Chemokines: Agents for the immunotherapy of
cancer? Nat Rev Immunol 2(3):175–184.
48. Chada S, Ramesh R, Mhashilkar AM (2003) Cytokine- and chemokine-based gene
therapy for cancer. Curr Opin Mol Ther 5(5):463–474.
49. Tan MCB, et al. (2009) Disruption of CCR5-dependent homing of regulatory T cells
inhibits tumor growth in a murine model of pancreatic cancer. J Immunol 182(3):
50. Penn MS, et al. (2013) An open-label dose escalation study to evaluate the safety of
administration of nonviral stromal cell-derived factor-1 plasmid to treat symptomatic
ischemic heart failure. Circ Res 112(5):816–825.
51. Abastado J-P (2012) The next challenge in cancer immunotherapy: Controlling T-cell
traffic to the tumor. Cancer Res 72(9):2159–2161.
52. Hong M, et al. (2011) Chemotherapy induces intratumoral expression of chemokines
in cutaneous melanoma, favoring T-cell infiltration and tumor control. Cancer Res
53. Wu YI, et al. (2009) A genetically encoded photoactivatable Rac controls the motility
of living cells. Nature 461(7260):104–108.
54. Yoo SK, et al. (2010) Differential regulation of protrusion and polarity by PI3K during
neutrophil motility in live zebrafish. Dev Cell 18(2):226–236.
55. Levskaya A, Weiner OD, Lim WA, Voigt CA (2009) Spatiotemporal control of cell
signalling using a light-switchable protein interaction. Nature 461(7266):997–1001.
Park et al.PNAS
| April 22, 2014
| vol. 111
| no. 16