of 20
A microrobotic system guided by photoacoustic computed
tomography for targeted navigation in intestines
in vivo
Zhiguang Wu
1,*
,
Lei Li
2,*
,
Yiran Yang
1
,
Peng Hu
3
,
Yang Li
1,3
,
So-Yoon Yang
2
,
Lihong V.
Wang
1,2,†
,
Wei Gao
1,†
1
Andrew and Peggy Cherng Department of Medical Engineering, California Institute of
Technology, Pasadena, CA, USA
2
Department of Electrical Engineering, California Institute of Technology, Pasadena, CA, USA
3
Department of Biomedical Engineering, Washington University in St. Louis, St. Louis, MO, USA
Abstract
Recently, tremendous progress in synthetic micro/nanomotors in diverse environment has been
made for potential biomedical applications. However, existing micro/nanomotor platforms are
inefficient for deep tissue imaging and motion control
in vivo
. Here, we present a photoacoustic
computed tomography (PACT) guided investigation of micromotors in intestines
in vivo
. The
micromotors enveloped in microcapsules are stable in the stomach and exhibit efficient propulsion
in various biofluids once released. The migration of micromotor capsules toward the targeted
regions in intestines has been visualized by PACT in real time
in vivo
. Near-infrared light
irradiation induces disintegration of the capsules to release the cargo-loaded micromotors. The
intensive propulsion of the micromotors effectively prolongs the retention in intestines. The
integration of the newly developed microrobotic system and PACT enables deep imaging and
precise control of the micromotors
in vivo
and promises practical biomedical applications, such as
drug delivery.
Summary
An imaging-guided ingestible microrobotic system enables deep tissue navigation and enhances
targeted retention
in vivo
.
Introduction
Micro and nanorobots that can be navigated into hard-to-reach tissues have drawn extensive
attention for the promise to empower various biomedical applications, such as disease
diagnosis, targeted drug delivery, and minimally invasive surgery (
1
6
). Chemically powered
motors, in particular, show great potential toward
in vivo
application thanks to their
Correspondence: LVW@caltech.edu, weigao@caltech.edu.
*
These authors contributed equally.
Author contributions:
W. G. and L. V. W. conceived the project. W. G. and L. V. W. supervised the studies. Z. W., Y. Y., and S. Y.
prepared and characterized the micromotors and micromotor capsules. L. L. and Z. W. performed photoacoustic experiments. Y. L.
drew the schematic illustrations. L. L., P. H., and Y. L. analyzed the photoacoustic imaging data. Z. W., L. L., Y. Y., W. G., and L. V.
W. interpreted data and wrote the manuscript. All authors reviewed the manuscript.
HHS Public Access
Author manuscript
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Published in final edited form as:
Sci Robot
. 2019 July 31; 4(32): . doi:10.1126/scirobotics.aax0613.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
autonomous propulsion and versatile functions in bodily fluids (
7
11
). However, imaging
and control of micromotors
in vivo
remain major challenges for practical medical
investigations, despite the significantly advanced development of micromotors (
12
15
). The
ability to directly visualize the dynamics of micromotors with high spatiotemporal resolution
in vivo
at the whole-body scale is in urgent demand to provide real-time visualization and
guidance of micromotors (
14
). In addition to high spatiotemporal resolution, the ideal non-
invasive micromotor imaging technique should offer deep penetration and molecular
contrasts.
To date, optical imaging is widely used for biomedical applications owing to its high
spatiotemporal resolution and molecular contrasts. However, applying conventional optical
imaging to deep tissues is hampered by strong optical scattering, which inhibits high-
resolution imaging beyond the optical diffusion limit (~1–2 mm in depth) (
16
). Fortunately,
photoacoustic (PA) tomography (PAT), detecting photon-induced ultrasound, achieves high-
resolution imaging at depths that far exceed the optical diffusion limit (
17
). In PAT, the
energy of photons absorbed by chromophores inside the tissue is converted to acoustic
waves, which are subsequently detected to yield high-resolution tomographic images with
optical contrasts. Leveraging the negligible acoustic scattering in soft tissue, PAT has
achieved superb spatial resolution at depths, with a depth-to-resolution ratio of ~200 (
18
), at
high imaging rates. As a major incarnation of PAT, photoacoustic computed tomography
(PACT) has attained high spatiotemporal resolution (125-μm in-plane resolution and 50-μs
frame
−1
data acquisition), deep penetration (48-mm tissue penetration
in vivo
), and
anatomical and molecular contrasts (Text S1)(
19
21
). With all these benefits, PACT shows
great promise for real-time navigation of micromotors
in vivo
for broad applications,
particularly, drug delivery.
Drug delivery through the gastrointestinal (GI) tract serves as a convenient and versatile
therapeutic tool, owing to its cost-effectiveness, high patient compliance, lenient constraint
for sterility, and ease of production (
22
,
23
). Although oral administration of various micro/
nanoparticle-based drug delivery systems has been demonstrated to both survive the acidic
gastric environment and diffuse into the intestines, drug absorption is still inefficient due to
the limited intestinal retention time (
24
). A large number of passive diffusion-based targeting
strategies have been explored to improve the delivery efficiency, but they suffer from low
precision, size restraint and specific surface chemistry (
25
). With precise control,
microrobotic drug delivery systems can potentially achieve targeted delivery with long
retention times and sustainable release profiles, which are in pressing need (
26
). Due to the
lack of imaging-guided control, there is no report yet for precisely targeted delivery using
micromotors
in vivo
(
14
). Additionally, biodegradability and biocompatibility are required,
and an ideal microrobotic system is expected to be cleared safely by the body after
completion of the tasks (
5
,
27
,
28
).
In this paper, we describe a PACT-guided microrobotic system (PAMR), which has
accomplished controlled propulsion and prolonged cargo retention
in vivo
(Fig. 1A and
Movie S1). Owing to the high spatiotemporal resolution, non-invasiveness, molecular
contrast, and deep penetration, PACT provides an attractive tool to locate and navigate the
micromotors
in vivo
(Fig. 1B) (
18
20
). Ingestible Mg-based micromotors are encapsulated
Wu et al.
Page 2
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
in enteric protective capsules to prevent reactions in gastric acid and allow direct
visualization by PACT (Fig. 1A–C). PACT monitors the migration of micromotor capsules
(MCs) in intestines in real time; continuous-wave (CW) near-infrared (NIR) light irradiation
induces phase transition of microcapsules and triggers propulsion of the micromotors (Fig.
1D); the autonomous and efficient propulsion of the micromotors enhances the retention in
targeted areas of the GI tract (Fig. 1E). We believe that the proposed integrated microrobotic
system will substantially advance gastrointestinal therapies.
Results
Fabrication of the micromotor capsules (MCs)
The fabrication of MCs mainly consists of two steps: the fabrication of Mg-based
micromotors (fig. S1, Materials and Methods) and the formation of MCs (fig. S2, Materials
and Methods). In the first step, Mg microparticles with a diameter of ~20 μm were dispersed
onto glass slides, followed by the deposition of a gold layer, which facilitates the
autonomous chemical propulsion in GI fluids and enhances PA contrast of the micromotors.
An alginate hydrogel layer was coated onto the micromotors by dropping aqueous solution
containing alginate and drugs (e.g., doxorubicin) on the slides. A parylene layer, acting as a
shell scaffold that ensures the stability during propulsion, was then deposited onto the
micromotors. Figure 2A illustrates a fabricated spherical micromotor (~20 μm in diameter).
A small opening (~2 μm in diameter), attributed to the surface contact of the Mg
microparticles with the glass slides during various layer coating steps, acts as a catalytic
interface for gas propulsion in the intestinal environment. Next, the micromotors were
encapsulated into the enteric gelatin capsules by the emulsion method (fig. S2). The green
fluorescence from the fluorescein isothiocyanate-labeled bovine serum albumin (FITC-
albumin) and red fluorescence from doxorubicin (DOX) were observed from the
micromotors (fig. S3, Materials and Methods) and the MCs (fig. S4, Materials and
Methods), confirming a successful drug loading. The size of MCs can be varied by changing
the speed of magnetic stirring (fig. S5). The microscopic images in Fig. 2B show three MCs
with diameters of 68 μm, 136 μm, and 750 μm.
For deep tissue imaging
in vivo
, it is crucial that the MCs should have higher optical
absorption than the blood background. To evaluate the PA imaging performance of the MCs,
the PA amplitudes of the MCs, whole blood, and bare Mg particles were measured
(Materials and Methods). NIR light experiences the least attenuation in mammalian tissues,
permitting the deepest optical penetration. As shown in Fig. 2C, the MCs exhibit strong PA
contrast in the NIR wavelength region, ranging from 720 to 890 nm. In order to assess
quantitatively the optical absorption of the MCs, we extracted amplitude values from the
above PA images and subsequently calibrated them with optical absorption of hemoglobin
(
29
,
30
). At the wavelength of 750 nm, the MCs display the highest PA amplitude of 15.3
(Fig. 2D). The bare Mg particles display a similar PA spectrum, with a lower PA peak with
an amplitude of 10.0 at 750 nm. The difference due to the Au layer is expected to
significantly improve the imaging sensitivity in the NIR wavelength region (Fig. 2D) (
31
,
32
). In addition, the approximate 3-fold increase in PA amplitudes of the MCs compared to
that of the whole blood provides sufficient contrast for PACT to detect the MCs
in vivo
Wu et al.
Page 3
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
using 750-nm illumination. To evaluate the stability of the MCs under pulsed NIR PA
excitation, we measured the PA signal fluctuation of the MCs during PA imaging (fig. S6).
The negligible changes in the PA signal amplitude during the operation suggest a remarkably
high photostability of the MCs. Fig. 2E and F show the PA images and the corresponding PA
amplitudes of single MCs with different concentrations of micromotors. The dependence of
the PA amplitude on the NIR light fluence (i.e., energy per area) was also investigated
(Materials and Methods). As expected, the PA amplitude of the micromotors almost linearly
increases with the NIR light fluence (Fig. 2F,
inset
). We also studied the maximum
detectable depth of MCs using PACT (Fig. 2G, Materials and Methods). The micromotors
showed dramatically decreased fluorescence intensity when covered by thin tissues (0.7–2.4
mm in thickness) and became undetectable quickly (Fig. 2G,
inset
and fig. S7). By contrast,
PACT can image the micromotors inside tissue as deep as ~7 cm (Fig. 2G), which reveals
that the key advantage of PACT lies in the high spatial resolution and high molecular
contrast for deep imaging in tissues (
19
).
Characterization of the dynamics of the PAMR
in vitro
The high optical absorption of the MCs empowers the PAMR as a promising
in vivo
imaging
contrast agent. To evaluate the dynamics of the PAMR, we conducted the PA imaging
experiments initially
in vitro
, where silicone rubber tubes modeled intestines (Materials and
Methods). The tubular model intestine was sandwiched in chicken breast tissues (Fig. 3A).
The PA time-lapse images in Fig. 3B and Movie S2 illustrate real-time tracking of the
migration of an injected MC in the model intestine.
In addition to tracking and locating the MCs, propulsion of the micromotors upon
unwrapping from the microcapsules can be activated on demand upon high power CW NIR
irradiation (Fig. 3C, Materials and Methods). Due to the enteric coating and gelatin
encapsulation, the MCs show long-term stability in both gastric acid and intestinal fluid
(Fig. 3D and fig. S8). The Au layer of the micromotors can effectively convert NIR light to
heat, resulting in a gel-sol phase transition of the gelatin-based capsule followed by the
release of the micromotors. Such CW NIR-triggered disintegration of the MCs usually
occurs within 0.1 s. Therefore, CW NIR irradiation can activate autonomous propulsion of
the micromotors (Fig. 3E and Movie S3). Such a photothermal effect also significantly
accelerates the Mg-water chemical reactions and thus enhances the chemical propulsion of
the micromotors. As shown in fig. S9 and Movie S4, the micromotors exhibit efficient
bubble propulsion in various biofluids. Further quantitative analysis indicates that the
velocities of the micromotors are 45 μm s
−1
and 43 μm s
−1
in PBS solution and the model
intestinal fluid, respectively. Note that bare Mg particles have negligible propulsion in
neutral media (i.e., intestinal fluid) and disordered propulsion in acidic condition (fig. S10,
Materials and Methods). The highly efficient propulsion in the targeted areas in intestines
provides a mechanical driving force to enhance retention and delivery. The required NIR
power can be potentially adjusted by controlling the synthesis process and composition of
the MCs. It is worth mentioning that other triggering mechanisms in biomedicine, such as
magnetic or ultrasonic fields, can also be employed to activate propulsion of the
micromotors (
33
).
Wu et al.
Page 4
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Dynamic imaging of the PAMR
in vivo
The movement of a swarm of MCs was monitored
in vivo
by PACT (Materials and
Methods). The MCs were dispersed in pure water and then orally administered into 5–6-
week old nude mice. The mice were subsequently anesthetized, and the lower abdominal
cavity was aligned with the imaging plane of the ultrasonic transducer array for longitudinal
imaging (Fig. 1B). PACT images were captured at a frame rate of 2 Hz for ~8 hours. As
shown in Fig. 4A and Movie S5, where the blood vessels and background tissues are shown
in gray and MCs in intestines are highlighted in color. During the imaging period of the first
6 hours, the MCs migrated for ~1.2 cm, roughly 15% of the length of the entire small
intestine. After 5 hours, the PA signals of some MCs faded away as they moved downstream
in intestines that were outside of the imaging plane. The moving speed of the swarm MCs in
the intestines and the movements induced by respiratory motion were quantified (Fig. 4B–D,
fig. S11). As shown in Fig. 4B–D, the abrupt motion caused by respiration is much faster
than actual migration of the MCs. Despite the respiration-induced movement, PACT can
distinguish the signals from the slowly migrating MCs in the intestines (Materials and
Methods). These results indicate that PACT can precisely monitor and track the locations of
the MCs in deep tissues
in vivo
.
The evaluation of the PAMR toward targeted retention and delivery
Of particular biomedical significance is the retention of cargo carriers in the targeted region
in intestines. While most of the previous studies focused on improving the interactions
between particles and the mucoadhesives by engineering surface coatings on the passive
particles, the biofluid-driven propulsion of the active micromotors can dramatically prolong
their retention in intestine walls. When the MCs approach the targeted areas in intestines, we
can trigger the collapse of the capsules and activate the propulsion of micromotors on
demand (Fig. 5A, Materials and Methods). To investigate the use of the PAMR for targeted
delivery, we grew melanoma cells in mouse intestines and coated the intestines with tissues
as a model
ex vivo
colon tumor. Thanks to the high optical absorption of melanoma cells in
the NIR wavelength region, colon tumors can be clearly resolved by PACT. After injection
into the intestines, the MCs migrated toward the targeted colon tumor, as illustrated by the
time-lapse PACT images in Fig. 5B and Movie S6. Once the MCs approach the targeted
location, they are irradiated with CW NIR light to trigger a responsive release of the
micromotors. The PA signals from the MCs in the intestines were prolonged upon the CW
NIR irradiation, suggesting the release of the micromotors (Fig. 5C). The overlaid
microscopic images in Fig. 5D show the NIR-triggered release of the micromotors from an
MC in the intestines. The DOX-loaded micromotors, clearly observed with red fluorescence,
rapidly diffused into the surrounding area after the CW NIR irradiation (Fig. 5D, Methods).
To evaluate retention of the micromotors
in vivo
, the enteric polymer-coated micromotors
and the paraffin-coated Mg particles (as passive control particles) were orally administrated
into two mouse groups that underwent a fasting treatment for 8 hours. The mice were
euthanized 12 hours after the administration, and their GI tracts were collected to evaluate
the retention of the micromotors (Materials and Methods). The intestines from the mice
treated with micromotors retained a much higher number of micromotors than that with
passive particles (Fig. 5E,
left
and
middle panels
). The quantitative analysis displays a
Wu et al.
Page 5
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
nearly 4-fold increase in the density of the micromotors in the treated intestine segments
(Fig. 5E,
right panel
). Note that nearly all Mg has already degraded in the retained
micromotors 12 hours after administration, as illustrated by the hollow structures of the
micromotors in the intestine before and after acid treatment (fig. S12). These results confirm
the capability of PAMR for prolonged retention in targeted areas in intestines. Besides the
active propulsion, the enhanced retention
in vivo
may also be attributed to the elevated pH
and Mg
2+
concentration in the surrounding environment caused by Mg-water reactions (Fig.
5F, Methods) (
33
,
34
). It has been recently reported that high pH (~8.2–12.0) could trigger a
phase transition of the mucus and facilitate tissue penetration of the micro/nanoparticles
(
33
36
). To investigate the influence of the micromotors on the pH of the surrounding
environment, the micromotors were dispersed in water with phenolphthalein as a pH
indicator. The microscopic image in Fig. 5F shows red/orange color in the vicinity of a
micromotor, indicating an increased pH of the surrounding medium. In addition, an
increased concentration of divalent cation Mg
2+
can cause collapse of the mucus gel (
37
).
The enhanced diffusion of the micromotors in the mucus was further validated using a
previously reported method (
38
), as shown in Fig. 5G (Materials and Methods). Compared
with the negligible diffusion of the control silica particles in the mucus, diffusion of the
micromotors in the mucus shows a significantly enhanced profile within 40 minutes. To
investigate the cargo release kinetics of the micromotors, a fluorescent anticancer drug,
DOX, was encapsulated into the alginate layer of the micromotors (Materials and Methods).
The release of DOX from the micromotors was characterized utilizing an ultra-violet/visible
spectrophotometer. The cross-linking treatment of the hydrogel significantly improves the
efficiency of DOX loading (fig. S13A). By increasing the DOX loading amount from 0.5 to
4 mg, the dose of DOX per micromotor can be controlled from ~1 to 20 ng while the
encapsulation efficiency can be improved up to 75.9% (fig. S13B). A higher release rate was
observed in the DOX-loaded micromotors in comparison to the DOX-loaded MCs (fig. S14),
indicating the promise of using the micromotors for
in vivo
targeted therapy of GI diseases
such as colon cancer.
The biocompatibility and biodegradability of the PAMR are important for biomedical
applications. The materials of the MCs, such as Mg, Au, gelatin, alginate, and enteric
polymer are known to be biocompatible. To evaluate the toxicity profile of the PAMR
in
vivo
, healthy mice were orally administered with MCs or DI water once a day for two
consecutive days. Throughout the treatment, no signs of distress, such as squinting of eyes,
hunched posture, or lethargy, were observed in either group. Initially, the toxicity profile of
the MCs in mice was evaluated through changes in body weight. During the experimental
period, the body weights of the mice administered with MCs have no significant difference
from those of the control group (fig. S15, Materials and Methods). The histology analysis
was performed to evaluate further the toxicity of the PAMR
in vivo
. No lymphocytic
infiltration into the mucosa or submucosa was observed, indicating no signs of inflammation
(Fig. 5H, Materials and Methods).
Discussion
Two key challenges should be addressed for applying synthetic micromotors to practical
biomedical applications: 1) advanced imaging techniques to locate micromotors in deep
Wu et al.
Page 6
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
tissue at high spatiotemporal resolution with high contrast; 2) precise on-demand control of
micromotors
in vivo
. With high molecular sensitivity at depths, PACT allows real-time
monitoring of micromotors in intestines at high spatial resolution for subsequent control.
Here, micromotors with partial coating of functional multilayers are designed as both the
imaging contrast agents and the controllable drug carriers. An Au layer is employed to
significantly increase the optical absorption for PA imaging and the reaction rate for efficient
propulsion simultaneously. A gelatin hydrogel layer is used to enlarge the loading capacity
of different functional components, such as therapeutic drugs and imaging agents. A
parylene layer is applied to maintain the geometry of the micromotors during propulsion.
Our current platform, integrating real-time imaging and control of micromotors in intestines
in vivo
, leads to the next generation of intelligent microrobotic systems, and provides
opportunities for precise microsurgery and targeted drug delivery.
Although the current platform has been demonstrated in small animals, human clinical
translations may require tens of centimeters of tissue penetration. PACT can provide up to 7-
cm tissue penetration, which is limited by photon dissipation. By employing a more
penetrating excitation source—microwave and acoustic detection, thermoacoustic
tomography (TAT) promises tissue penetration for clinical translations (
39
,
40
). Moreover,
incorporation of a gold layer in the micromotor design provides an excellent microwave
absorption contrast for TAT owing to the high electrical conductivity, and thus greatly
enhances the deep tissue imaging capability of the microrobots for clinical applications.
Focused ultrasound heating can increase the depths of thermally triggered microrobot
release to the whole-body level of humans.
Currently, the passive diffusion-based delivery suffers from complex designs, particle size
constraints, low precision, and poor specificity. Our platform allows micromotors to reach
any targeted regions in intestines with high precision. It can be tailored to particles of any
sizes and can be applied to any biological media without additional design efforts. Our
platform can also be easily modified to carry various cargos, including therapeutic agents
and diagnostic sensors, with real-time feedback during delivery.
Biocompatibility and biodegradability of the micromotors are essential for practical
biomedical applications. The components of our micromotors, widely used as therapeutic
agents and in implantable devices, are studied to be safe for
in vivo
applications (
41
43
).
The micromotors have been eventually cleared by the digestive system via excrement,
without any adverse effects.
In summary, we report an ingestible microrobotic platform with high optical absorption for
imaging-assisted control in intestines. The encapsulated micromotors survive the erosion of
the stomach fluid and permit propulsion in intestines. PACT non-invasively monitors the
migration of the micromotors and visualizes their arrival at targeted areas
in vivo
. CW NIR
irradiation toward targeted areas induces a phase transition of the capsules and triggers the
propulsion of the micromotors. The mechanical propulsion provides a driving force for the
micromotors to bind to the intestine walls, resulting in an extended retention. The proposed
platform lays a foundation for targeted delivery in tissues and opens a new horizon for
precision medicine.
Wu et al.
Page 7
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Materials and Methods
Materials
Commercially available magnesium microparticles with a diameter of 20±5 μm were
purchased from TangShan WeiHao Magnesium Powder. Agarose, FITC-albumin, alginate,
gelatin, and doxorubicin hydrochloride were purchased from Sigma-Aldrich. Paraffin liquid,
wax, hexane, glutaraldehyde, phenolphthalein, hydrochloric acid, glass coverslip, and gene
frame were purchased from Thermo Fisher Scientific. Acrylic polymers (Eudragit L 100–55)
were purchased from Evonik Industries. Silicone rubber tubes (inner diameter: 0.5 mm)
were purchased from Dow Silicones.
Fabrication of the micromotors
The Mg-based Janus micromotors were constructed with an embedding method (fig. S1).
The Au, alginate, and parylene were deposited in a layer-by-layer manner. Mg particles were
first washed with acetone for three times and dried at room temperature prior to use. Mg
particles were dispersed in acetone with a particle concentration of ~0.1 g mL
−1
and then
spread on the glass slides at room temperature. After the acetone evaporated, Mg particles
were attached onto the surface of the glass slides through physical adsorption, exposing the
majority of the surface areas of the particles to air. Subsequently, the glass slides coated with
Mg particles were deposited with a Au layer (~100 nm in thickness) using an electron-beam
evaporator (Mark 40, CHA Industries). After the deposition, a mixture containing alginate
(2%, w/v) and doxorubicin was dropped on the glass slides and then dried with N
2
gas.
Aqueous CaCl
2
(0.2 mL of 5%, w/v) was then dropped onto the glass slides to cross-link
alginate. After 30 minutes, the glass slides were washed with pure water and dried with N
2
gas. The glass slides were coated with a parylene C layer (750 nm in thickness) using a
parylene coater (Labtop 3000, Curtiss-Wright). The resulted micromotors were collected by
scratching from the glass slides.
Preparation of the MCs
MCs were fabricated based on a controlled emulsion technique according to the previous
reports (fig. S2) (
44
,
45
). A mixture containing gelatin (5%, w/v) and micromotors (5%,
w/v) at 40–60 °C was extruded from a 30-gauge needle into 50 mL liquid paraffin at ~60 °C.
Pure water was used as the solvent here, in which micromotors remained stable due to the
formation of a compact hydroxide passivation layer on the Mg surfaces. Subsequently, an
enteric polymer solution consisting of 100 mg of Eudragit L-100 in 2 mL organic solvent
mixture (acetone:methanol = 1:1, v/v) as previously reported (
46
), was extruded into the
liquid paraffin. The extruded solution was kept at 60 °C for 4 hours to evaporate the acetone
and methanol, and then the temperature was lowered to 0 °C with an ice bath. In order to
harvest the MCs from the liquid paraffin, cold water (~4 °C) was added into the liquid
paraffin with magnetic stirring for more than 20 minutes, and most MCs were separated
from the liquid paraffin into the water. The water containing MCs was extracted and then
washed with hexane for three times. The size of the MCs can be controlled by varying the
rotational speed of magnetic stirring between 100 and 1000 rpm (fig. S5). The collected
MCs were rinsed with an aqueous hydrochloric acid solution (pH = 2) and then washed with
Wu et al.
Page 8
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
pure water to remove the hydrochloric acid. Subsequently, the MCs were cross-linked
through incubation with glutaraldehyde for 1 hour followed by water rinse.
Characterization of the structures of the micromotors and the MCs
SEM images of the Mg-based micromotors were acquired with a field emission scanning
electron microscope (FEI Sirion) at an operating voltage of 10 keV (Fig. 2A). The samples
were coated with a 5-nm carbon layer to improve the conductivity (Leica EM ACE600
Carbon Evaporator). The bright field and fluorescence microscopic images of the
micromotors and the MCs were taken with a Zeiss AXIO optical microscope (Fig. 2B, figs.
S3 and S4). To observe the structure of the DOX-loaded micromotors and MCs using
fluorescence imaging, the micromotors and the MCs were stained with FITC-albumin.
Labeling of FITC-albumin onto the micromotors was carried out by dip-coating the
micromotors-loaded glass slides in a 0.2 mL of FITC-albumin solution (0.2 mg mL
−1
),
followed by dip-coating in an alginate solution (2%, w/v). Labeling of the FITC-albumin
onto the MCs was conducted by adding FITC-albumin into the gelation solution.
Characterization of the PA performances of the MCs
Characterization of the PA performances of the MCs was conducted using a PACT system
(
19
). The MCs, bare Mg microparticles, and blood were separately injected into three
silicone tubes. Both ends of the tubes were sealed with agarose gel (2%, w/v). the PACT
system employed a 512-element full-ring ultrasonic transducer array (Imasonic SAS; 50 mm
ring radius; 5.5 MHz central frequency; more than 90% one-way bandwidth) for 2D
panoramic acoustic detection. Each element has a cylindrical focus (0.2 NA; 20 mm element
elevation size; 0.61 mm pitch; 0.1 mm inter-element spacing). A lab-made 512-channel
preamplifier (26 dB gain) was directly connected to the ultrasonic transducer array housing,
minimizing cable noise. The pre-amplified PA signals were digitized using a 512-channel
data acquisition system (four SonixDAQs, Ultrasonix Medical ULC; 128 channels each; 40
MHz sampling rate; 12-bit dynamic range) with programmable gain up to 51 dB. The
digitized radio frequency data were first stored in the onboard buffer, then transferred to a
computer and reconstructed using the dual-speed-of-sound half-time universal back-
projection algorithm (Fig. 2C–G, figs. S6 and S7) (
19
).
PACT of the migration of the MCs
PACT of the migration of the MCs in model intestines was carried out by injecting the MCs
into a silicone tube, which was covered by chicken breast tissues. Migration of the MCs in
the tube was driven by microfluidic pumping and was captured by PACT (Fig. 3B, Movie
S2).
For
in vivo
experiments, all experimental procedures were conducted under a laboratory
animal protocol approved by the Office of Laboratory Animal Resources at California
Institute of Technology. Three- to four-week-old nude mice (Hsd: Athymic Nude-FoxlNU,
Harlan Co.; 20–25-g body weight) were used for
in vivo
imaging. Prior to the imaging
experiments, the mice were fasted for ~8 hours followed by the oral administration with the
MCs. The mouse was then fixed to a lab-made imaging platform by taping the fore and hind
legs on the top and bottom parts of the holder in the PACT system. During imaging, the mice
Wu et al.
Page 9
Sci Robot
. Author manuscript; available in PMC 2020 July 06.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript