of 13
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
1 of 13
MICROROBOTS
Imaging-
guided bioresorbable acoustic
hydrogel microrobots
Hong Han
1
†, Xiaotian Ma
1
†, Weiting Deng
1,2
†, Junhang Zhang
3
†, Songsong Tang
1
, On Shun Pak
4
,
Lailai Zhu
5
, Ernesto Criado-
Hidalgo
1,6
, Chen Gong
3
, Emil Karshalev
1
, Jounghyun Yoo
1
, Ming You
1
,
Ann Liu
6
, Canran Wang
1
, Hao K. Shen
6
, Payal N. Patel
1
, Claire L. Hays
1
, Peter J. Gunnarson
7,8
,
Lei Li
1
, Yang Zhang
1
, John O. Dabiri
7,8
, Lihong V. Wang
1
, Mikhail G. Shapiro
1,6,9
, Di Wu
6
*,
Qifa Zhou
3
*, Julia R. Greer
1,2
*, Wei Gao
1
*
Micro-
and nanorobots excel in navigating the intricate and often inaccessible areas of the human body, offering
immense potential for applications such as disease diagnosis, precision drug delivery, detoxification, and mini-
mally invasive surgery. Despite their promise, practical deployment faces hurdles, including achieving stable pro
-
pulsion in complex in vivo biological environments, real-
time imaging and localization through deep tissue, and
precise remote control for targeted therapy and ensuring high therapeutic efficacy. To overcome these obstacles,
we introduce a hydrogel-
based, imaging-
guided, bioresorbable acoustic microrobot (BAM) designed to navigate
the human body with high stability. Constructed using two-
photon polymerization, a BAM comprises magnetic
nanoparticles and therapeutic agents integrated into its hydrogel matrix for precision control and drug delivery.
The microrobot features an optimized surface chemistry with a hydrophobic inner layer to substantially enhance
microbubble retention in biofluids with multiday functionality and a hydrophilic outer layer to minimize aggrega-
tion and promote timely degradation. The dual-
opening bubble-
trapping cavity design enables a BAM to maintain
consistent and efficient acoustic propulsion across a range of biological fluids. Under focused ultrasound stimula-
tion, the entrapped microbubbles oscillate and enhance the contrast for real-
time ultrasound imaging, facilitating
precise tracking and control of BAM movement through wireless magnetic navigation. Moreover, the hydrolysis-
driven biodegradability of BAMs ensures its safe dissolution after treatment, posing no risk of long-
term residual
harm. Thorough in vitro and in vivo experimental evidence demonstrates the promising capabilities of BAMs in
biomedical applications. This approach shows promise for advancing minimally invasive medical interventions
and targeted therapeutic delivery.
INTRODUCTION
The emergence of micro-
and nanorobotics heralds a pivotal shift in
biomedical engineering, introducing a new era filled with unparal-
leled prospects to refine a myriad of applications including disease
diagnosis, targeted drug delivery, detoxification, and minimally in-
vasive surgery (
1
12
). These miniature devices, adept at traversing
the complex and often unreachable recesses of the human body,
stand at the forefront of precision medicine. The diverse propulsion
mechanisms used by these microrobots, which include the deploy-
ment of local chemical fuels or external fields, such as a magnetic
field, an ultrasound field, or light, or the exploitation of natural
microorganisms’ motility, illustrate the approaches engineered to
navigate these devices across micro-
and nanoscale environments
(
13
18
). However, the practical application of these advancements
in vivo is impeded by challenges in the propulsion through complex
biofluids, real-
time imaging, precise control, and therapeutic assess-
ment of micromotors (
19
24
).
For an untethered microrobot to operate effectively in the com-
plex biological milieu of the human body, it must encompass a suite
of characteristics: reliable and steady propulsion through biological
media, enhanced imaging contrast for real-
time visualization under
deep tissue, meticulous control for accurate targeting of diseased
regions, substantial payload capacity for effective therapy, and high
biocompatibility with biodegradable characteristics to obviate the
need for surgical removal. Moreover, high-
resolution additive man-
ufacturing and high-
freedom design are critical for the widespread
clinical adoption of microrobots. Nonetheless, an intricate challenge
lies in integrating these optimal features within the confines of a
micro-
or nanoscale robotic platform, given the inherent limitations
posed by their diminutive size.
Among the emerging solutions, acoustically actuated microro-
bots show great promise owing to the distinct advantages of acous-
tic propulsion, including safety, noninvasive operation, deep-
tissue
penetration, robust propulsive forces, rapid response, and unteth-
ered control, which are well aligned with clinical demands (
25
38
).
In particular, microrobots that incorporate bubbles and leverage
microbubble oscillations induced by weak traveling acoustic waves
offer an attractive approach to in vivo manipulation (
39
42
). How-
ever, the practical deployment of these microrobots faces obstacles
because of the transient stability and short life span of the encapsu-
lated bubble within biological fluids and the lack of precisely con-
trolled propulsion in complex environments in vivo (
43
,
44
).
1
andrew and Peggy cherng Department of Medical engineering, Division of engineer
-
ing and a
pplied Science, c
alifornia i
nstitute of
technology, Pasadena, ca, USa.
2
Kavli
nanoscience i
nstitute, c
alifornia i
nstitute of
technology, Pasadena, ca, USa.
3
alfred
e. Mann Department of biomedical engineering, University of Southern c
alifornia,
los
angeles, ca, USa.
4
Department of Mechanical engineering, Santa clara University,
Santa clara, ca, USa.
5
Department of Mechanical engineering, national University
of Singapore, Singapore, Singapore.
6
Division of chemistry and chemical engineering,
california i
nstitute of
technology, Pasadena, ca, USa.
7
Graduate a
erospace labora-
tories, c
alifornia i
nstitute of
technology, Pasadena, ca, USa.
8
Department of Mechanical
and
civil
engineering,
california
institute of
technology, Pasadena,
ca, USa.
9
Howard
Hughes Medical i
nstitute, Pasadena, ca, USa.
*c
orresponding author. email: weigao@
caltech.
edu (W.G.); jrgreer@
caltech.
edu (J.R.G.);
qifazhou@ usc. edu (Q.Z.); dwwu@ caltech. edu (D.W.)
†t
hese authors contributed equally to this work.
copyright © 2024
the
authors, some rights
reserved; exclusive
licensee a
merican
association for the
advancement of
Science. no claim to
original U.S.
Government Works
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
2 of 13
In response to these multifaceted challenges, we introduce a
hydrogel-
based, imaging-
guided, bioresorbable acoustic microro-
bot (BAM) designed for navigation with high stability, precision,
and control within the human body (Fig. 1A and Movie 1). The
fabrication of BAMs involves two-
photon polymerization (TPP),
using a composite material consisting of poly(ethylene glycol) diacry
-
late (PEGDA); pentaerythritol tetraacrylate (PETA); and additional
functional components including Fe
3
O
4
nanoparticles (NPs) for
magnetic steering and therapeutic agents such as the anticancer
drug 5-
fluorouracil (5-
FU) (Fig. 1, A to C). A BAM’s refined design
incorporates an asymmetric dual-
opening bubble-
trapping cavity, sub
-
stantially elevating its propulsion capabilities beyond the single-
opening
configuration (Fig. 1, C and D, and fig. S1). A unique dual-
step surface
chemistry modification strategy endows BAMs with a hydrophobic
inner surface, substantially improving bubble retention within bio-
fluids, and a hydrophilic outer surface that prevents microrobot
aggregation and promotes hydrogel degradation (Fig. 1E). These
BAMs are capable of sustained, efficient acoustic propulsion in a
variety of biological fluids—including urine, gastrointestinal fluid,
and whole blood—over multiday durations (Fig. 1F).
In addition to their attractive propulsion abilities, the microbub-
bles trapped by BAMs serve as effective ultrasound imaging contrast
agents (
45
,
46
). Thus, the movement and localization of BAMs can
be monitored in real time with high spatiotemporal resolution using
A
D
G
Real-time imaging
Single-opening
Dual-opening
Dual-opening (=
90°)
Bubble oscillation
Hydrolysis
within biofluids
∆t
∆t
ii
Drug release
i
Tu
mor targeting
Tw
o-photon polymerization
US imaging probe
BAMs
T
umor
Bladder
US transducer
US transducer
iii
Biodegradation
E
(ii) Plasma etching
(i) SAM
C
H
OF
Si
F
Ultrasound imagin
g
Acoustic propulsion
iv
v
DETC
5-FU
PET
A
PEGDA
Fe
3
O
4
Urine
Whole blood
Gastrointestinal
fluid
US imaging probe
US transducer
N
S
N
S
N
S
N
S
N
S
Outer
surface
Inner
surface
c
c
c
c
c
c
Real-time guidance
BAMs
B
C
O
Fe
F
S
N
N
S
Fig. 1.
Imaging-
guided BAMs.
(
A
) Schematic illustration of the preparation and in vivo applications of ba
Ms, covering the fabrication of ba
Ms by
tPP with a customized
polymer resin (i), the efficient propulsion of the dual-
opening ba
M configuration (ii), ultrasound imaging-
guided ba
Ms for targeted drug delivery (iii), sustainable drug
release at the diseased area from ba
Ms (iv), and biodegradation of the ba
M backbone in vivo over time (v). US, ultrasound. (
B
) SeM images of a 3D-
printed ba
M array.
Scale bar, 50
μ
m. (
C
) SeM image of a single ba
M and the corresponding energy-
dispersive x-
ray spectroscopy mapping of o
, F, and Fe elements, respectively. Scale bar,
10
μ
m. (
D
) Schematic illustration of the substantially enhanced propulsion behavior of dual-
opening ba
Ms compared with single-
opening ba
Ms.
θ
, the angle of the two
openings. (
E
) Schematic illustration of the dual-
step surface chemistry strategy used by the ba
Ms to preserve microbubbles for extended periods of operation in biologi-
cally relevant environments.
θ
c
, surface contact angle. (
F
) Schematic illustration of the efficient acoustic propulsion, ultrasound imaging, and magnetic control of ba
Ms in
different biofluids in vivo. (
G
) Schematic of the hydrolysis-
based biodegradation of ba
Ms in biofluids.
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
3 of 13
an ultrasound probe under deep tissue and directed with precision
via wireless magnetic navigation (Fig. 1F). The harmonious combi-
nation of advanced and stable ultrasound-
driven propulsion with
concurrent imaging capabilities emphasizes the considerable potential
of BAMs for in vivo biomedical applications. Last, the hydrolysis-
mediated biodegradability of the BAM design ensures safe bio
-
logical compatibility, devoid of harmful residue risks after operation
(Fig. 1G). BAMs’ performance, functionality, and efficacy for cancer
therapy have been evaluated through extensive in vitro and in vivo
experimental validations.
RESULTS
The design and motion characterization of BAMs
The hydrogel-
based BAMs, fabricated via high-
resolution TPP, fea-
ture a spherical design with an outer diameter of 30
μ
m and a spher
-
ical 18-
μ
m-
diameter internal cavity designed to trap microbubbles
in aqueous environments (Fig. 1, B and C). Under an ultrasound
field, the encapsulated gas bubbles within BAMs pulsate, oscillating
at their resonant frequency, which could be tuned by the size of the
internal cavity (table S1). This oscillation of the gas-
liquid interface
produces powerful microstreaming vortices around the openings
(Fig. 1D), generating a propulsive force (
F
p
) in the opposite direc-
tion of the flow.
A notable aspect of BAM movement is the influence of secondary
Bjerknes force (
F
SB
), an attractive force that acts perpendicular to the
boundary (
47
,
48
). For microrobots with a single opening, the com-
bined forces orient the opening toward the boundary and lead to un-
stable and inefficient translational movement (Fig. 2A). In contrast,
BAMs designed with dual openings benefit from an additional
propulsive force that runs parallel to the boundary, notably improv-
ing both speed and stability of movement (Fig. 2, B to D; fig. S2; and
movie S1). An asymmetric design, in which the geometric center of
the entrapped bubble deviates from the geometric center of BAM,
was observed to exhibit enhanced propulsion performance (fig. S3A).
This effect may be attributed to the closer proximity of the liquid-
gas
interfaces to the openings in the off-
centered design, which generates
more pronounced streaming flows (movie S2) and, thus, improves
propulsion. As a result, asymmetric BAMs achieve propulsion speeds
exceeding twice those of the symmetric design (fig. S3B).
Our experiments revealed the crucial role of the angle between
the BAMs’ openings (
θ
) in determining their performance. Specifi-
cally, the average speed of BAMs in phosphate-
buffered saline (PBS)
varied with
θ
, reaching a peak at 90° before decreasing. Speed mea-
surements at 0°, 60°, 90°, 120°, and 150° were 413
±
69, 1251
±
188,
2043
±
280, 754
±
166, and 500
±
47
μ
m s
1
, respectively. Given that
one of the openings will face the substrate, the optimal speed at
θ
=
90° is attributed to the propulsive force generated from the sec-
ond opening being absent of a
y
-
direction component; thus,
F
p2
for
θ
=
90° contributed all to motion when comparing that with the
other angles (
θ
=
60°, 90°, 120°, and 150°; fig. S4), as corroborated by
simulations indicating the maximum streaming force at this angle
(Fig. 2D and Supplementary Methods). When moving close to the
wall, BAMs could be attracted to the wall because of the secondary
Bjerknes force and move along the wall (fig. S5 and movie S3).
The microstreaming patterns generated by single-
opening and
dual-
opening BAMs under the same ultrasound field were experi-
mentally characterized with the passive particle tracers around the
BAMs anchored on the substrate (Fig. 2, E and F; fig. S6; and movie
S4). Particle image velocimetry (PIV) analysis revealed pronounced
microstreaming patterns from both openings in the dual-
opening
configuration (Fig. 2, E and F). This further indicates the capability
of generating propulsive thrust in additional directions compared
with the single-
opening design, as validated by both experimental
data and simulation outcomes (Fig. 2, E to H; fig. S6; and Supple-
mentary Methods). Several triple-
opening BAMs were further in-
vestigated but displayed slower speeds than the dual-
opening ones.
This might be due to the partial cancellation of propulsion forces in
the triple-
opening structures (fig. S7). These findings underline the
effectiveness of the dual-
opening architecture in enhancing the pro-
pulsion and maneuverability of BAMs in fluidic environments.
Surface modification of BAMs for enhanced operation
longevity in biofluids
The performance of BAMs in vivo relies on their ability to sustain
microbubbles, which is crucial for both propulsion and imaging. The
short life span of microbubbles in existing bubble-
based microrobots
has limited their practical use in biomedical applications. Previous
reports have used surface engineering and structural design to pro-
long bubble retention (
39
,
41
,
49
). However, the use of bioresorbable
materials in BAMs presents additional hurdles, particularly because
of the hydrophilic nature of hydrogel surfaces, which complicates
prolonged bubble retention.
Movie 1. Overview of the imaging-
guided bioresorbable acoustic hydrogel
microrobots.
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
4 of 13
To overcome these obstacles and ensure stable in vivo operation,
we adopted a unique surface chemistry approach aimed at preserv-
ing microbubbles for extended durations while facilitating the easy
transfer and immersion of microrobots (Fig. 2I). Our strategy in-
volved a two-
step surface modification process. Initially, we treated
BAMs with fluorosilane to create a highly hydrophobic self-
assembled
monolayer (SAM) on both the inner and outer surfaces. This hydro-
phobicity is essential for trapping bubbles; however, it also predisposes
BAMs to self-
aggregation and adherence to surfaces, complicating
their deployment in fluids. To mitigate this, we subsequently applied
a hydrophilic treatment via O
2
plasma etching to the SAM-
coated
BAMs, achieving a hydrophilic exterior while preserving the hydro-
phobic interior.
The experimental results indicate that this dual-
step surface
modification substantially enhanced microbubble stability. For ex-
ample, unmodified hydrogel microrobots lost their bubbles within
1 min of immersion in biofluids such as urine, attributed to the
hydrophilic PEGDA film’s 0° contact angle (Fig. 2J). In contrast, the
G
10
5
10
4
10
3
10
2
10
1
10
0
10
-1
ms
-1
]
10
5
10
4
10
3
10
2
10
1
10
0
10
-1
ms
-1
]
F
I
M
SAM
E
O
2
pl
as
ma
L
In
ur
in
ef
or
1
mi
n
θ
c
c
= 80.1
°
c
= 133.4°
θ
c
J
In
uri
ne
for
14 day
s
A
Si
ngl
e-open
in
g
2
s
F
SB
F
P
H
Dual
-opening
(
=9
)
2
s
F
P1
F
P2
F
SB
D
K
c
= 0°
Outer
surfac
e
Inner
surfac
e
Outer
surfac
e
Inner
surfac
e
2
s
B
Dual
-openi
ng
(
=
60°
)
F
P2
F
P1
F
SB
x
y
Days
of
over
80%
mi
crorobot
sr
etai
ni
ng
bubble
s
0
2
4
6
8
10
12
14
PBS
Ur
in
e
Se
ru
m
150
500
0
1000
1500
2000
2500
0.
5
1.
0
1.
5
Si
mu
la
ted force
(a
.u.)
120
90
06
0
Speed
m
s
-1
)
16
IP-D
ip
Robo
ts
BAMs
Untr
eate
dB
AMs
C
Se
ru
m
Afte
r
7
day
s
Afte
r
4
day
s
PBS
2
s
O
P
Q
N
3
s
Ur
in
e
Wh
ol
e
bl
ood
12
s
Wo
und
fl
ui
d
3
s
Gastro
in
testi
nal
fl
ui
d
3
s
c
= 0°
2
s
Fig. 2.
Characterization of the propulsion of BAMs.
(
A
to
C
)
the propulsion trajectories of a single-
opening ba
M (
a), a dual-
opening (
θ
=
60°) ba
M (b), and a dual-
opening (
θ
=
90°) ba
M (
c). Scale bars, 50
μ
m. (
D
)
the dependence of the experimentally measured velocity of ba
Ms and the simulated combined force on
θ
. error bars
represent the SD from 10 ba
Ms. a.u., arbitrary units. (
E
to
H
)
the time-
averaged streamline patterns of the vortex profiles of a single-
opening ba
M [(e) and (G)] and a dual-
opening
ba
M [(F) and (H)] at the resonant frequency, obtained via PiV analysis of the passive particle tracers [(e) and (F)] and numerical simulation [(G) and (H)].
the ba
M’s
body is fixed on the substrate with the opening parallel to the substrate. Scale bars, 15
μ
m. (
I
) Schematic illustration of the two-
step surface treatment strategy to preserve
microbubbles for extended periods in biofluids and maintain easy transferability. (
J
and
K
) optical images of ba
Ms without (J) and with (K) the surface treatment after
incubation in urine, and the corresponding
θ
c
values of the internal and external surfaces of ba
Ms. Scale bars, 50
μ
m. (
L
) bubble-
retaining capabilities in the biologically
relevant fluids (i.e., PbS, urine, and serum) of 3D-
printed microrobots based on iP-
Dip, untreated ba
Ms, and surface-
treated ba
Ms. error bars represent the SD from three
batches. (
M
) optical images of the bubble-
retaining ba
Ms and the motion trajectories of ba
Ms after being immersed in PbS for 7 days and in serum for 4 days. Scale bars,
50
μ
m. (
N
to
Q
) Schematics and motion trajectories of the microrobots in human urine (n), porcine gastrointestinal fluid (
o), murine wound fluid (P), and whole blood (Q).
Scale bars, 100
μ
m.
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
5 of 13
modified BAMs, featuring a hydrophobic interior (with a contact
angle of 133.4°) and a hydrophilic exterior (with a contact angle of
80.1°), successfully maintained microbubbles in human urine for
14 days (Fig. 2K).
Our comprehensive study on the durability of bubble retention
in various microrobots fabricated through TPP further revealed
BAMs’ substantial improvements in bubble longevity. The microro-
bots based on traditional TPP resin IP-
Dip lost their bubbles within
1 hour, and untreated hydrogel-
based BAMs lost bubbles in 1 min
in all tested biologically relevant fluids. In contrast, surface-
treated
BAMs demonstrated robust bubble retention, with over 80% of the
BAMs sustaining bubbles for 3 days in serum, 12 days in PBS, and
13 days in urine (Fig. 2L and fig. S8). This extended bubble reten-
tion did not compromise BAMs’ mobility, which remained efficient
even after several days of immersion in various biofluids (Fig. 2M
and fig. S9).
The unique asymmetric dual-
opening design, combined with
asymmetric surface hydrophobicity, was crucial for stabilizing mi-
crobubbles and ensuring robust operation within complex biologi-
cal media. The BAMs exhibited the ability to navigate through
untreated body fluids with varying viscosities—such as urine,
gastrointestinal fluid, wound fluid, and whole blood—with high
efficiency, achieving speeds of 896
±
262, 812
±
195, 621
±
228, and
121
±
19
μ
m s
1
, respectively (Fig. 2, N to Q; figs. S10 and S11; and
movie S5). The trajectory of BAMs could be substantially influ-
enced by the heterogeneity of biofluids. Moreover, the incorpora-
tion of the Fe
3
O
4
NPs granted BAMs a superparamagnetic property
(fig. S12), enabling precise control over the acoustic propulsion of a
single BAM or a swarm of BAMs through an external magnetic
field (fig. S13 and movie S6). These capabilities underscore BAMs’
promising potential for real-
world biomedical applications, dem-
onstrating their effectiveness and adaptability in navigating the
challenging environments of the human body.
Characterization of the material properties of BAMs
The main composition of BAMs is PEGDA, a US Food and Drug
Administration–approved hydrogel material known for its nontox-
icity and biodegradability. In aqueous environments, the hydrolysis
of ester bonds within the cross-
linked PEGDA framework triggers
the breakdown of polymer chains, leading to the gradual degrada-
tion of BAMs (Fig. 3A) (
50
). This degradation process is notably
faster under alkaline conditions because of enhanced hydrolysis
via saponification. The hydrolytic degradation of BAMs was as-
sessed in 50 mM NaOH, chosen for its minimal influence on
human health (
51
). Routine optical inspections charted the degra-
dation kinetics of BAMs, initially observing an expansion because
of swelling followed by collapse within a 3-
day period (Fig. 3B). The
biocompatibility of Fe
3
O
4
NPs and Fe
3
O
4
NP–loaded BAMs was
evaluated through incubating murine macrophage cells (RAW
264.7) and human bladder cancer (T24) cells with Fe
3
O
4
NPs and
passive BAMs, respectively. After a 72-
hour incubation, both Fe
3
O
4
NPs and BAMs exhibited negligible effects on the metabolic activity
of the cells (Fig. 3, C and D, and fig. S14, A and B). The scant pres-
ence of red-
stained dead cells further indicated BAMs’ high com-
patibility with cellular environments (Fig. 3E and fig. S14C).
For therapeutic applications, the anticancer drug 5-
FU was effi-
ciently loaded into BAMs during the TPP fabrication process, facili-
tated by its minimal absorption at the TPP laser wavelength of 780 nm
(
52
) and its low reactivity with photoinitiators (
53
). Given that the
terminal half-
life of 5-
FU, when administered parenterally, is under
20 min because of rapid hepatic breakdown (
54
), encapsulation
within BAMs substantially extends its half-
life, thereby increasing its
therapeutic efficacy. The drug-
loading efficiency of BAMs was deter
-
mined through ultraviolet-
visible (UV-
vis) spectroscopy, revealing
that 1000 BAMs contained a total 5-
FU dose of 183.1 ng. This amount
represents 62.0% of the theoretical drug-
loading capacity, on the ba-
sis of the maximum solubility of 5-
FU in PEGDA resin (24 mg ml
1
),
indicating effective drug incorporation within BAMs (fig. S15). The
release profile of 5-
FU from BAMs was monitored by observing the
decrease in fluorescence intensity, with
~
60% of the drug being re-
leased into urine within the initial 12 hours because of diffusion and
polymer degradation (Fig. 3, F and G). Comparative analysis re-
vealed that larger model drug molecules, such as rhodamine B, dis-
played a slower release rate, taking approximately 19 hours to achieve
60% release in human urine (fig. S16).
The in vitro therapeutic efficacy of 5-
FU–loaded BAMs was
assessed using a three-
dimensional (3D) tumor spheroid model
composed of T24 cells. This model closely simulates the mass trans-
port dynamics and structural complexity of avascular tumor tissues
(
55
57
). In the experiment, T24 spheroids,
~
300
μ
m in diameter,
were treated with PBS, 5
μ
M free drug, passive BAMs, or acousti-
cally propelled BAMs carrying an equivalent amount of drug. After
treatment, propidium iodide (PI) staining was applied to identify
dead cells within the spheroids (Fig. 3H). The results revealed that
acoustically propelled BAMs exhibited the strongest PI fluores-
cence intensities compared with the spheroids treated with PBS,
free 5-
FU, and passive BAMs (Fig. 3I). This can be attributed to the
enhanced interaction between BAMs and tumor spheroids, a direct
result of efficient propulsion of BAMs (fig. S17). This confirms the
effectiveness of BAMs’ propulsion and their polymer matrix in con-
trolling drug release, substantially enhancing drug delivery effi-
ciency and anticancer activity.
Ultrasound imaging of acoustically propelled BAMs in vitro
and in vivo
The microbubbles trapped inside BAMs can increase the ultrasound
imaging contrast, especially in deep tissue visualization. This is at-
tributed to the acoustic impedance mismatch between air and water
and the volumetric oscillations induced by microbubbles when ex-
posed to ultrasound waves (Fig. 4A) (
45
,
46
). BAMs, whether indi-
vidually or in arrays and across varying densities, can be efficiently
detected and located within biological tissues or tissue-
simulating
phantoms using ultrasound imaging techniques (Fig. 4, B and C; fig.
S18; and movie S7).
To achieve concurrent ultrasound imaging and acoustic propul-
sion of BAMs within soft tissue environments, a dual-
probe ap-
proach was used. This approach comprises an ultrasound imaging
probe for real-
time imaging of the bubbles and a focused ultra-
sound (FUS) probe designed for effective propulsion. The choice of
FUS over piezoelectric disks here is informed by its strong field at a
low input voltage (fig. S19), which enhances its effectiveness and
durability for the acoustic pressure amplitude attenuation through
soft tissues (
58
). Adjusting the inner diameter of BAMs allows for
the tuning of the resonant frequency of the trapped microbubbles
to align with the FUS’s center frequency, thereby maximizing op-
erational efficiency (table S1 and fig. S20).
In vitro real-
time imaging of the propulsion of BAMs was con-
ducted within an agarose chamber, simulating a bladder environment
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
6 of 13
and using a 50-
MHz linear array ultrasound probe in the B-
mode
imaging configuration (Fig. 4, D to G). The propulsion of BAMs,
facilitated by the ultrasound field generated by the FUS probe at an
excitation frequency of 480 kHz and an applied acoustic pressure of
626 kPa (peak-
to-
peak pressure), was observed. BAMs with an inner
diameter of 14
μ
m, which have a resonant frequency close to the ex-
citation frequency (table S1), demonstrated highly efficient motion,
as captured in ultrasound imaging scans (Fig. 4E and movie S8).
Guided by the ultrasound imaging, the propulsion direction of BAMs
can be precisely and wirelessly controlled by an external magnetic
field, leveraging their superparamagnetic properties (Fig. 4, F and G,
and movie S8).
The evaluation of in vivo propulsion and imaging of BAMs was
conducted within the mouse bladder using the capabilities of a
setup that aligns ultrasound propulsion and imaging (fig. S21).
This setup ensured that the imaging probe’s central axis intersected
the focal point of the propulsion transducer, hence aligning the
ultrasound imaging zone with the propulsion field. BAMs were
administered into the mouse bladder through intravesical instilla-
tion, demonstrating vortex-
like motion upon activation by ul-
trasound (movie S9). Ultrasound imaging revealed that, under
magnetic control, BAMs moved toward a specific bladder region
(Fig. 4, H to K, and movie S9). The potential of BAMs for in vivo
tumor targeting was assessed in a mouse with orthotopic bladder
cancer. Ultrasound imaging enabled the visualization of the blad-
der tumor, aiding in the precise delivery of BAMs to the tumor site
via acoustic propulsion coupled with magnetic steering (Fig. 4, L
and M, and movie S10). Once attached to the tumor tissue or
Contr
ol
250 BAMs
50 BAMs
Cumulative drug release (%)
01
0203040506
0
0
20
40
60
80
100
Time (h)
Fast release
by diffusion
Slow rele ase
by degradat
ion
Contro
l
RAW 264.7
Live/Dead
T24
RAW 264.7
0
50
100
24
48
72
Time (h)
Cell viability
(%)
C
G
0day
1day
2day
s3
days
0
h4
h1
2 h
B
HI
5-FU
PBSP
assive BA
Ms
Actuated BAMs
1 h
Tumor
Spheroi
d
BAMs
Microplate culture
US
propul
sion
50 BA
Ms
+
Hydrol
ysis
A
PEGD
A
O
O
O
H
n
O
O
O
O
n
HO
O
H
O
H
250 BA
Ms
D
Dead ce
lls
staining
T24
Tumor
Control
250 BAMs
50 BAMs
+
E
F
0
50
100
24
48
72
Time (h)
Cell viability
(%)
Fig. 3.
In vitro evaluation of biodegradability, drug release, and biocompatibility of BAMs.
(
A
) Schematic of the degradation mechanism of PeGD
a, the backbone
material of ba
Ms. (
B
) bright-
field microscopic images demonstrating the degradation of ba
Ms to debris within 3 days in 50 mM naoH. Scale bar, 50
μ
m. (
C
and
D
) c
ell
viability of R
aW 264.7 (
c) and
t24 (D) cells cocultured with ba
Ms. error bars represent the SD of four measurements. (
E
) Representative live (green)/dead (red) fluorescence
images of R
aW 264.7 (top) and
t24 cells (bottom) seeded with PbS (control) and ba
Ms after 3 days of culture. Scale bar, 100
μ
m. (
F
)
the time-
dependent cumulative drug
release of 5-
FU from ba
Ms in urine. (
G
) Fluorescence images of the time-
dependent release of 5-
FU from ba
Ms in urine. Scale bar, 50
μ
m. (
H
) Schematic of the in vitro
tumor spheroid treatment process using ba
Ms. (
I
) Representative fluorescence images of dead cells (red color) stained with Pi dye in
t24 spheroids after a 15-
min expo
-
sure to PbS, 5
μ
M 5-
FU, nonactuated ba
Ms, and acoustically actuated ba
Ms, respectively, followed by washing and a 48-
hour incubation.
the outline depicts the shape of
the spheroid at its largest cross section. Scale bar, 100
μ
m.
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025
Han
et al.
,
Sci. Robot.
9
, eadp3593 (2024) 11 December 2024
Science RoboticS
|
ReSeaR
cH
aR
ticle
7 of 13
embedded within the bladder wall, BAMs remained firmly in
place and ultrasound responsive despite changes in magnetic di-
rection. The enduring attachment and responsiveness were veri-
fied through Doppler mode ultrasound imaging (Fig. 4N).
Evaluation of the anticancer therapeutic efficacy of
BAMs in vivo
The therapeutic efficacy of BAMs was assessed in vivo through a
murine orthotopic bladder tumor model established by the inoculation
B
BAMs
H
I
0 s
20 s
30 s
10 s
F
G
1 s3
s
5 s
0 s
BAMs
Magnet
L
M
US ON
Tumor
US OF
F
BAMs
D
E
BAMs
FUS
0 s
2 s
3 s
1 s
B
J
K
BAMs
0 s
5 s
15 s
10 s
A
C
US Im
aging probe
Z
water
≠ Z
air
B
BAMs
US imaging probe
v
v
4 s
0 s
N
v
v
Emission
Reflection
Trans
mission
v
4 s
0 s
Tumor
BAMs
Fig. 4.
Characterization of the propulsion, control, and imaging of BAMs in vitro and in vivo.
(
A
) Schematic of deep-
tissue ultrasound imaging of a ba
M with
the trapped microbubble serving as the contrast agent. (
B
and
C
) SeM and ultrasound images of ba
M arrays with varied densities: one ba
M per 50
μ
m
2
by 50
μ
m
2
(b) and
200
μ
m
2
by 200
μ
m
2
(c). Scale bars, 50 and 500
μ
m for SeM and ultrasound images, respectively. (
D
to
G
) Schematics and the corresponding ultrasound images of
in vitro acoustic propulsion of ba
Ms without [(D) and (e)] and with [(F) and (G)] magnetic guidance in an agarose chamber–based artificial bladder. Scale bars, 1 mm.
(
H
to
K
) Schematics and the corresponding ultrasound images of propulsion of ba
Ms with magnetic guidance to the left side [(H) and (i)] and right side [(J) and (K)] in vivo
in the bladder of a mouse. Scale bars, 2 mm. (
L
and
M
) Schematic and the corresponding ultrasound images of acoustic propulsion-
based, magnetically navigated tumor
targeting of
ba
Ms in vivo in a mouse bladder with orthotopic bladder tumor. Scale bar, 2 mm. (
N
) Doppler mode ultrasound images of
ba
Ms remaining attached to the
tumor despite alterations in magnetic orientation while retaining responsiveness to ultrasound stimuli. Scale bar, 2 mm.
Downloaded from https://www.science.org at California Institute of Technology on January 28, 2025