Abstract
Clinical approaches for cancer therapy face several interrelated
challenges involving inefficient drug delivery, potential adverse side
effects, and inconvenience. Here, we present an integrated wearable
flexible ultrasound microneedle patch (wf-UMP) that serves as a
portable platform for convenient, efficient, and minimally invasive
cancer therapy. The wf-UMP adopts an all-in-one bioelectronic concept,
which integrates a stretchable lead-free ultrasound transducer array
for acoustic emission, a bioadhesive hydrogel elastomer for robust
adhesion and acoustic coupling, and a dissolvable microneedle patch
loaded with biocompatible piezoelectric nanoparticles for painless drug
delivery and reactive oxygen species generation. With soft mechanical
properties and enhanced electromechanical performance, wf-UMP can be
robustly worn on curved and dynamic tissue surfaces for easy and
effective manipulation. In preclinical studies involving mice, wf-UMP
demonstrated notable anticancer effects by inducing tumor cell
apoptosis, amplifying oxidative stress, and modulating immune cell
proliferation. Furthermore, the synergistic immunotherapy induced by
wf-UMP and Anti-PD1 further improved anticancer immunity by activating
immunogenic cell death and regulating macrophages polarization,
inhibiting distant tumor growth and tumor recurrence.
Subject terms: Biomedical engineering, Cancer immunotherapy, Electronic
devices
__________________________________________________________________
Clinical approaches for cancer therapy face several interrelated
challenges. Here, Xue et al. presented an integrated wearable flexible
ultrasound microneedle patch that serves as a portable platform for
convenient and minimally invasive cancer therapy.
Introduction
Cancer, or malignant tumor, remains a critical global health challenge
and a leading contributor to the global disease burden, according to
the World Health Organization^[48]1. Clinical approaches to cancer
treatment primarily include surgical removal, radio/chemotherapy,
targeted therapy, and immunotherapy. Despite significant advancements,
cancer remains difficult to cure^[49]1. Nevertheless, several
interrelated challenges pertaining to current clinical protocols lead
to restricted use and limited therapeutic efficacy, including
inconvenience, potential adverse side effects, and inefficient drug
delivery^[50]2,[51]3. Piezocatalytic therapy (PCT) is an emerging
reactive oxygen species (ROS)-driven dynamic therapeutic strategy that
relies on ultrasound (US)-activated anisotropic piezocatalysts with
built-in electric fields, resulting in the generation of ROS and
catalytic behavior^[52]4. In contrast to other dynamic therapies, PCT
overcomes the penetration limits of conventional photodynamic therapy
in the treatment of some deep tumors due to the deep tissue penetration
nature of US. Additionally, PCT effectively generates ROS without
relying on external oxygen, superior to chemodynamic therapy in hypoxic
tumors^[53]5,[54]6. PCT is a promising emerging technology with
additional advantages that are still being researched. With the
advances in nanomedicine, PCT based on piezoelectric nanoparticles
(PNPs) offers potential advantages in precision over conventional
radio/chemotherapy, making it a favorable platform for cancer
immunotherapy^[55]6–[56]8. For therapeutic purposes, these PNPs are
commonly injected into the body, but this procedure might cause
discomfort, pain, and potential tissue damage. An emerging transdermal
drug delivery approach, dissolvable microneedle (MN) patch, offers a
minimally invasive manner to bypass the skin barrier by creating
microchannels that allow loaded drugs to act locally on superficial
tumors or enter the systemic circulation for further distribution.
Notably, the diameter, height, and density of MNs are adjustable for
painless insertion and for improved and expanded drug
delivery^[57]9,[58]10. However, challenges persist, including limited
penetration depth, uneven drug distribution, and the difficulty in
achieving adequate therapeutic doses for deeper or larger tumors.
Additionally, the clinical translation of most piezocatalyst-loaded MNs
is hindered by the cumulative toxicity of heavy metals and the risk of
foreign body rejection.
For PCT, US, known for its high tissue penetration and biosafety, is
also essential for diffusing and activating PNPs. Commonly used US
transducers are rigid probes that pose challenges in effectively
coupling with skin for US propagation, thus requiring the use of
abundant coupling agents and needing to apply pressure to achieve
effective acoustic contact on curved body parts^[59]11,[60]12.
Meanwhile, this procedure relies on hand-held manipulation or
mechanical fixation devices (for example, straps and tapes) to hold
bulky probes to the skin^[61]13. These hamper the ease, comfort, and
continuity of the treatment. Moreover, most commercial US probes are
developed using toxic lead zirconate titanate (PZT) ceramics, raising
concerns about the harm to the body, especially being used as the
wearable/implantable electronics^[62]11–[63]19. Wearable, flexible, and
lead-free US devices possess extraordinary mechanical and acoustic
properties for stable and effective treatment or monitoring over time
without the need of coupling agents and external fixation, especially
with their biocompatibility. We hypothesize that an all-in-one wearable
US-MN device incorporating flexible, lead-free US transducers and
dissolvable MN patches can address the comprehensive challenges posed
by current PCT. However, it remains a noteworthy challenge to realize
such high-level integrated and functional wearable electronics, yet
with an applicable paradigm for cancer immunotherapy.
In this work, we present a fundamental methodology for integrating
wearable flexible ultrasound microneedle patches (wf-UMP) from
materials engineering, device architecture and system integration
(Fig. [64]1a), which can serve as a portable platform for cancer
therapy. Distinguished from the reported studies, the wf-UMP does not
rely on traditional PCT programs but adopts an US-MN all-in-one
concept, which comprises three main components: (1) a stretchable
conformal lead-free US transducer array for US emission, (2) a
bioadhesive hydrogel elastomer containing N-hydroxysuccinimide (NHS)
for robust adhesion (interfacial toughness >500 J m^−2) and acoustic
coupling (acoustic impedance ~1.54 MRayl), and (3) a biocompatible
PNPs-loaded dissolvable hyaluronic acid (HA) MN patch for minimally
invasive and painless drug delivery. The wf-UMP with soft mechanical
properties and enhanced electromechanical performance can be robustly
worn on curved and dynamic tissue surfaces, exhibiting efficient
acoustic emission, enhanced drug delivery, and improved generation of
ROS. In preclinical experiments, wf-UMP demonstrated notable anticancer
effects by inducing tumor cell apoptosis, amplifying oxidative stress,
and modulating immune cell proliferation. RNA sequencing analysis
revealed that the cytokine-cytokine receptor interaction and T cell
receptor signaling pathway were upregulated with wf-UMP treatment.
Moreover, the synergistic immunotherapy induced by wf-UMP and Anti-PD1
further improved anticancer immunity and effectively targeted distal
cancers in dual tumor models by further activating dendritic cells
(DCs) maturation through immunogenic cell death (ICD) and up-regulating
macrophage M1 phenotype polarization, while augmenting antitumor immune
memory to prevent tumor recurrence.
Fig. 1. Overview of wf-UMP.
[65]Fig. 1
[66]Open in a new tab
a Schematic illustration of wf-UMP application for cancer therapy.
Created by figdraw.com. b Schematic of the integrated system-level
wf-UMP electronics, consisting of a flexible US transducer array for
effective US emission, a bioadhesive hydrogel elastomer for robust
adhesion and acoustic coupling layer, and a mKNN PNPs-loaded MN patch
for drug delivery. c, d Schematic illustration of the operational
procedure of wf-UMP for tumor immunotherapy. Step 1: Insertion and
dissolution of MNs (b). Step 2: Diffusion of mKNN PNPs and ROS
generation under US stimulation, along with activation of systemic
immune responses (c). e SEM image of mKNN MN. Scale bar, 100 μm. f TEM
image of mKNN PNPs. Scale bar, 50 nm. g Schematic illustration of the
generation of ROS by mKNN PNPs under US stimulation. h Schematic
illustration of the robust adhesion of the bioadhesive hydrogel via
amide bond formation at the interface. i, j Mechanical properties of
the US transducer array under tensile stress, demonstrated by
experimentally stretching the transducer array (i) and corresponding
FEA simulated strains (j). Scale bar, 2.5 mm. k Photograph of the US
transducer array adhering to curved surfaces, showing its mechanical
flexibility. Scale bar, 4 mm. l Photograph of wf-UMP adhered seamlessly
on skin. Scale bar, 5 mm. m Photograph demonstrating the robust
adhesion of wf-UMP with skin. Scale bar, 5 mm. n H&E staining of mouse
skin after the insertion of MNs. Scale bar, 100 μm.
Results
Design of wf-UMP
The idea of the integrated wf-UMP for cancer therapy is illustrated in
Fig. [67]1a, which adapts to the tumor site and enables portable
treatment with minimal activity restrictions. The device orderly
integrates a stretchable conformal transducer array for US emission, a
bioadhesive hydrogel elastomer for robust adhesion and acoustic
coupling, and a dissolving MN patch for drug delivery (Fig. [68]1b and
Supplementary Fig. [69]1). Upon wf-UMP apposition, the soluble MN patch
punctures the skin and dissolves, subsequently releasing loaded PNPs,
which accelerates diffusion and generates ROS for tumor therapy under
US stimulation, further inducing improved anticancer immunity
(Fig. [70]1c, d).
The cone shaped MNs with a height of 600 μm, a base diameter of 250 μm,
and an inter-needle spacing of 550 μm were evenly arranged in a 20 × 20
array over an area of 10 mm × 10 mm, enabling effective penetration of
the skin and access to the dermis (Fig. [71]1e and Supplementary
Figs. [72]2–[73]5)^[74]20,[75]21. The relatively thin backing film
(~50 μm) of the MN patch ensures sufficient structural support for the
MNs without exerting significant mechanical resistance during insertion
(Supplementary Fig. [76]6). The loaded chemically modified
0.95K[0.48]Na[0.52]Nb[0.97]Sb[0.03]O[3]-0.05Bi[0.5]Na[0.5]ZrO[3]-0.2%
Fe[2]O[3] (abbreviated as mKNN, Supplementary Figs. [77]7–[78]12,
Supplementary Note [79]1, and Supplementary Table [80]1) PNPs acting as
piezocatalysts with a size of <100 nm are sensitive to structural
deformation induced by ultrasonic pressure, which facilitates charge
separation and thus ensures high-efficient ROS generation (Fig. [81]1f,
g)^[82]22. An NHS ester-based bioadhesive hydrogel, with a thickness of
~1 mm, was applied between the US array and the MN patch for optimal
tissue-device integration. This hydrogel exhibits notable
stretchability, and importantly, can form covalent amide bonds with
interfaces (both transducer and skin), resulting in strong adhesion for
continuous wearable US therapy (Fig. [83]1h and Supplementary
Fig. [84]13)^[85]13,[86]23,[87]24. The US transducer incorporates a
4 × 6 matrix array of high-performance mKNN 1–3 piezoelectric composite
units (2.5 mm × 2.5 mm × 1.5 mm) that are interconnected via
island-bridge electrodes and encapsulated in biocompatible silicone
elastomer, thereby achieving overall stretchability (Supplementary
Figs. [88]14–[89]16)^[90]11,[91]15. We further quantified the strain on
the transducer when stretched by 60% along the x-axis (Fig. [92]1i).
Finite element analysis (FEA) results indicate a maximum tensile strain
of ~1.5% in the copper interconnects (Fig. [93]1j). Overall, the MN
patch, with its small size and thin backing film, can be fully covered
and securely adhered to the skin by the larger bioadhesive hydrogel
layer, ensuring reliable attachment of wf-UMP throughout the
application (Supplementary Figs. [94]17 and [95]18).
The final transducer array is displayed in Fig. [96]1k, highlighting
its flexibility and conformal contact on curved surfaces. The robust
adhesion of the wf-UMP is shown in Fig. [97]1l, m and Supplementary
Fig. [98]19, where it withstands high pulling forces (~5 N) and
maintains seamless adhesion to skin, demonstrating its potential for
application to non-planar surfaces. It also adheres well to skin under
dynamic and sweaty exposures, enabling real-world applications
(Supplementary Figs. [99]20 and [100]21 and Supplementary
Movie [101]1). Furthermore, the effective penetration of the MNs into
the mouse skin was validated by the hematoxylin and eosin (H&E)
staining (Fig. [102]1n). In conclusion, wf-UMP demonstrates stable and
robust adhesion to soft tissues, ensuring seamless contact even when
coupled to complex surfaces, thus facilitating effective, painless drug
delivery and sustained therapy (Supplementary Fig. [103]22).
Characterizations of wf-UMP
In the context of biosafety for wearable electronics, high-performance
mKNN ceramics were selected as the functional material for
electromechanical conversion of the transducer array (Supplementary
Note [104]2), which were further processed into 1–3 piezo-units with a
kerf of 200 μm and a pitch of 700 μm (Fig. [105]2a and Supplementary
Fig. [106]23) to enhance the thickness electromechanical coupling
coefficient (k[t]) and reduce acoustic impedance
(Z[a])^[107]11,[108]25. The efficiency of electromechanical conversion
can be quantified by the resonance frequency (f[r]) and antiresonance
frequency (f[a]) of the impedance spectra (Fig. [109]2b and
Supplementary Fig. [110]24). Compared to the isotropic bulk ceramic,
the anisotropic 1–3 piezo-unit exhibited higher k[t] (~0.53) due to the
suppression of the shear vibrating modes (Fig. [111]2c). Moreover, the
lower Z[a] (~10 MRayl) of 1–3 piezo-units further improved the acoustic
coupling and thus acoustic transmission efficiency. Since the filled
epoxy polymer is without ferroelectric polarization, the polarization
and strain values of the 1–3 piezo-units were degraded (Supplementary
Fig. [112]25); however, the well-preserved piezo/ferroelectricity still
ensures stable electrical performance and US emission. Ultimately, the
1–3 piezoelectric transducer exhibits a center frequency of 1.2 MHz and
a −6 dB bandwidth of 58% (Fig. [113]2d).
Fig. 2. Electrical, mechanical, and acoustic performance validation of
wf-UMP.
[114]Fig. 2
[115]Open in a new tab
a Schematic illustration of the active 1–3 piezo-unit in wf-UMP for US
emission. b Impedance and phase angle spectra of 1–3 piezo-unit. The
two peaks in the impedance spectrum are defined as the resonance
frequency (f[r]) and the antiresonance frequency (f[a]), respectively.
Inset presents the equivalent RLC circuit for 1–3 piezo-unit. c
Comparison of thickness-mode electromechanical coupling coefficient and
acoustic impedance between mKNN ceramic and its 1–3 piezo-unit. d
Pulse-echo response and frequency spectrum of the 1–3 piezo-unit. e
Impedance spectra of wf-UMP under different bending curvatures. f
Optical image of a 2 × 2 array released from the 80% x-axis tensile
strain. Scale bar, 2.5 mm. g 180° peeling test of the bioadhesive
hydrogel on porcine skin. Scale bar, 1 cm. h FEA simulation of the
acoustic field generated by a 1–3 piezo-unit. i FEA simulation of the
acoustic field generated by wf-UMP. j Acoustic pressure of wf-UMP with
increasing input voltages. k I[SPTA] and MI values of wf-UMP with
increasing input voltages. l Thermal images of wf-UMP on an agarose
hydrogel operated continuously for 30 min. Top, 0 min. Bottom: 30 min.
Scale bars, 1 cm. m Comparison of drug delivery performance of wf-UMP
on agarose hydrogels with/without US stimulation. Scale bars, 2 mm. n
Confocal microscopy images of porcine skin treated with Rho B labeled
wf-UMP with/without US stimulation at a depth of 328 μm. Scale bars,
250 μm.
Furthermore, we assessed the electrical impedance of the device under
deformation. The stable impedances of wf-UMP with increasing bending
curvatures indicate that mechanical deformation has minimal influence
on device performance (Fig. [116]2e). Tensile properties were evaluated
by stretching experiments on wf-UMP along the x-axis from 0 to 80%, as
visualized by piezo-units in a 2 × 2 array region (Supplementary
Fig. [117]26). The serpentine interconnects unraveled, rotated, and
twisted when loaded with tensile stresses, thereby relieving the strain
on the electrode islands^[118]11,[119]26,[120]27. The wf-UMP can be
reversibly stretched up to 60% in one direction. With further
stretching up to 80%, plastic deformation of the serpentine
interconnect as well as partial delamination from the silicone
elastomer can be observed upon release (Fig. [121]2f and Supplementary
Fig. [122]26)^[123]11,[124]15. These special geometrical and electrical
designs enable wf-UMP to exhibit an elastic strain level of ~60% and an
optimal operating frequency of 1.2 MHz. To assess the adhesion
capability of wf-UMP, we performed the standard 180°-peeling test to
quantify the interfacial toughness of the bioadhesive hydrogel on
porcine skin (Fig. [125]2g). The interfacial toughness between the
hydrogel and porcine skin is >500 J m^−2, demonstrating a robust
adhesion of wf-UMP on soft tissues^[126]23,[127]28. Under mechanical
stresses such as bending and stretching, wf-UMP maintains stable
adhesion without delamination or loss of contact (Supplementary
Fig. [128]27). Moreover, the acoustic impedance of the bioadhesive
hydrogel was experimentally measured to be ~1.54 MRayl, which is
comparable to commercial US gels and skins (~1.5 MRayl) and ensures the
efficiency of US transmission (Supplementary Fig. [129]28)^[130]25.
We employed FEA to simulate the US beam pattern emitted by wf-UMP.
Compared to the bulk ceramic unit exhibiting a spreading US field, US
beams emitted by 1–3 piezo-unit show improved longitudinal directivity,
magnitude, and penetration depth of >15 mm (Fig. [131]2h and
Supplementary Figs. [132]29 and [133]30). In terms of wf-UMP, it
demonstrates more converged beams with penetration depths up to
~100 mm, enabling efficient diffusion and activation of mKNN PNPs
injected directly into the skin via the MN patch or into the body via
subsequent systemic circulation (Fig. [134]2i). The larger the
piezoelectric array, the greater the ultrasonic power and the deeper
the penetration. The acoustic pressure generated by wf-UMP over the
input voltage ranging from 10 Vpp (peak-to-peak voltage) to 120 Vpp was
measured, which increased with input voltages and was higher than that
of the device fabricated by mKNN ceramics (Fig. [135]2j and
Supplementary Fig. [136]31). Regarding the safety assessment of
acoustic exposures on human tissue, we calculated safety parameters of
wf-UMP, including the spatial-peak temporal average (I[SPTA]), the
spatial-peak pulse average (I[SPPA]), the mechanical index (MI), and
the thermal index (TI) (Fig. [137]2k and Supplementary Fig. [138]32).
All of these parameters are positively correlated with the input
voltage, indicating that the acoustic output of wf-UMP can be flexibly
adjusted to ensure safe operation (input voltage <70 Vpp, based on U.S.
Food and Drug Administration (FDA) limits)^[139]12,[140]29,[141]30.
Furthermore, the thermal effects of wf-UMP were evaluated by direct
imaging with an infrared camera. The temperature of wf-UMP increased by
3.2 °C when the device was placed on an agarose hydrogel and operated
at an input voltage of 20 Vpp for 30 min, indicating a negligible
device thermal effect and thermal-induced drug diffusion (Fig. [142]2l
and Supplementary Figs. [143]33 and [144]34)^[145]31.
To reveal the drug delivery performance of wf-UMP, the device was
adhered on the agarose hydrogel to visualize the release and diffusion
behavior of the drug-loaded MN patch^[146]32. Notably, the MN patch
demonstrated excellent solubility, ensuring rapid dissolution and drug
release upon wf-UMP application (Supplementary Figs. [147]35 and
[148]36 and Supplementary movie [149]2). To further investigate the
effect of US stimulation, Rho B-labeled wf-UMP devices were pressed
into agarose hydrogels and held in place for 2 min, during which one
group of agarose hydrogels received US stimulation (US+) while the
other served as a control without US stimulation (US-) (Supplementary
Fig. [150]37). Cross-sectional images of hydrogels were recorded at
different time points to track drug diffusion (Fig. [151]2m). The
results demonstrated enhanced diffusion in the US+ samples, with a drug
penetration depth of ~4.6 mm at 30 min, which was twice that observed
in the control group (~2.3 mm). While these results highlight a trend
of improved diffusion under US stimulation, it is important to note
that agarose hydrogel serves only as a preliminary model for diffusion
visualization and does not replicate the complex structural and
mechanical properties of skin tissues. To further validate these
findings, confocal microscopy was used to analyze the permeation depth
of Rho B into porcine skin. wf-UMP devices were applied to porcine skin
under identical conditions (pressed for 2 min), and the diffusion depth
of Rho B was examined. At a depth of 328 μm, the US+ sample exhibited
significantly higher fluorescence intensity compared to the control
group (Fig. [152]2n). The total diffusion depth in the US+ samples
reached ~500 μm, compared to <400 μm in the US- group (Supplementary
Fig. [153]38). These results further confirm that US stimulation
enhances the diffusion of Rho B into tissue, indicating the potential
of wf-UMP for efficient transdermal drug delivery.
US-activated catalysis of mKNN PNPs
The distribution of mKNN PNPs within the MNs was observed to be
relatively homogeneous, with slight clustering localized near the tip
region (Fig. [154]3a). Upon application, the dissolvable mKNN MN
portion in wf-UMP will rapidly dissolve in aqueous environments or
tissues, releasing the mKNN PNPs, enabling them to be activated by US
stimulation and exert their catalytic functionality (Fig. [155]3b). The
optimized mKNN PNPs are irregularly shaped with a size of <100 nm and
the contained elements are homogeneously distributed (Fig. [156]3c and
Supplementary Fig. [157]39). High-resolution transmission electron
microscope (HR-TEM) images showed clear and well-ordered lattice
fringes extending throughout the mKNN crystals (Fig. [158]3d and
Supplementary Fig. [159]40). The calculated interplanar spacings (d) of
0.2799 nm and 0.2827 nm belong to (101) lattice plane of tetragonal (T)
phase and (111) of orthorhombic (O) phase, respectively. The
selected-area electron diffraction (SAED) patterns also displayed
distinct rings (Supplementary Fig. [160]41). Subsequently, the crystal
structure of mKNN PNPs was evaluated by X-ray powder diffraction (XRD)
pattern, which resembles the multiphase coexisting mKNN ceramic, as the
lowered energy barrier will promote polarization rotation and thus
enhance piezoelectricity (Fig. [161]3e, f). Notably, the diffraction
peaks of mKNN PNPs were significantly weakened and widened due to the
decrease in crystallinity and particle size^[162]33,[163]34. The
chemical nature of mKNN PNPs was investigated by XPS analysis, which
revealed that oxygen vacancies could further enhance the piezocatalytic
activity (Fig. [164]3g, Supplementary Fig. [165]42, and Supplementary
note [166]3). The piezoelectricity of PNPs on the nanoscale was
confirmed by switching spectroscopy PFM (SS-PFM). Both mKNN and KNN
(abbreviation for unmodified K[0.48]Na[0.52]NbO[3]-0.2% Fe[2]O[3]) PNPs
demonstrated characteristic butterfly-shaped amplitude curves and
rectangular phase loops (Fig. [167]3h and Supplementary Fig. [168]43).
However, mKNN PNPs displayed higher amplitude (>900 pm) and phase
contrast (~180°) compared to KNN PNPs (amplitude <300 pm, phase
contrast ~120°) under an applied voltage of ±20 V, indicating favorable
intrinsic piezorespone and complete domain switching in mKNN PNPs.
Additionally, these PNPs were developed into piezoelectric energy
harvesters (PEHs), and their electrical outputs were measured under
mechanic force. Under identical instantaneous force impacts, the
average open-circuit output voltage (~16 V) of the mKNN-based PEH was
four folds higher than that of the KNN-based PEH (~4 V), further
validating the enhanced piezoelectric properties of mKNN PNPs
(Supplementary Fig. [169]44).
Fig. 3. Mechanism, structural characterizations, and US-activated
piezocatalysis of mKNN PNPs.
[170]Fig. 3
[171]Open in a new tab
a Fluorescence images of blank and mKNN MNs. mKNN MNs showed reduced
fluorescence particularly at the tips. Scale bars, 600 μm. b Schematic
illustration of US-activated piezocatalysis of mKNN PNPs loaded in
dissolvable MNs. c TEM images and corresponding EDS mapping results of
mKNN NPs. Scale bars, 100 nm. d HR-TEM image of mKNN PNPs. Scale bar,
5 nm. e XRD patterns of mKNN ceramic and PNPs. f The spontaneous
polarization vectors of mKNN materials with R-O-T phase coexistence. g
XPS spectra of O 1s for mKNN PNPs. h SS-PFM amplitude curve and phase
loop of mKNN PNPs. i, j Electronic band structure (i) and density of
states (j) of mKNN PNPs determined by DFT calculations. k UV-vis
diffuse reflectance spectra of mKNN PNPs. l, Schematic illustration of
electron and hole generation, transfer, and utilization in mKNN PNPs
for piezocatalysis. m Degradation of Rho B by mKNN PNPs under US
stimulation. n Piezocurrents of KNN, mKNN, and non- piezoelectric
Nb[2]O[5] PNPs under periodic US stimulation. o FEA simulation of
surface piezo-potential distribution of a mKNN PNP under the acoustic
pressure generated by wf-UMP. p EPR spectra of DMPO-^•OH and DMPO-^•
[MATH:
O2
− :MATH]
over mKNN PNPs under US stimulation.
In view of the importance of energy band structure in piezocatalytic
activities, we first employed density function theory (DFT)
calculations to theoretically investigate the energy band structure of
mKNN. The mKNN exhibits an indirect electronic band gap of 2.91 eV,
where the valence band maximum (E[VB]) originates from the O 2p
orbital, while the conductive band minimum (E[CB]) arises from the
coupling of the O 2p and Nb 3 d orbitals (Fig. [172]3i, j). The
experimentally determined energy gap (E[g]) of mKNN PNPs via UV-vis
spectra was measured to be 3.91 eV (Fig. [173]3k, Supplementary
Fig. [174]45, and Supplementary note [175]4), exceeding the calculated
value. Generally, a decrease in particle size (especially at the
nanometer size) would enhance the energy barrier among charge carriers,
resulting in wider energy gaps^[176]33,[177]35.The mechanism of
piezocatalysis is schematically illustrated in Fig. [178]3l. Periodic
US stimulation induces the generation of a built-in piezo-potential
field and energy band bending in mKNN PNPs, which promotes continuous
separation and transmission of charge carriers (for example, electrons
and holes) and thus facilitates the surface redox reactions, ultimately
leading to the production of ROS •OH and •
[MATH: O2− :MATH]
^[179]22,[180]33.
Next, we assessed the piezocatalytic activity of mKNN PNPs through the
degradation of Rho B. After 25 min of US stimulation, notable decreases
in absorption peak intensity were observed, with a high degradation
rate of 110 × 10^−3 min^−1 (Fig. [181]3m, Supplementary Fig. [182]46,
Supplementary note [183]5, and Supplementary table [184]2). Generation,
separation and migration of electron-hole pairs are crucial factors
determining the efficiency of piezocatalysis^[185]4,[186]36,[187]37.
Therefore, we measured the transient piezoelectric currents of mKNN
PNPs under periodic US stimulation (Fig. [188]3n). The results show
prominent current peaks upon exposure to US stimulation, indicating
effective separation and migration of charges. In contrast,
non-piezoelectric Nb[2]O[5] (raw material of mKNN and KNN) PNPs showed
no response to US stimulation, further demonstrating that the charge
separation and migration are induced by the piezoelectric effect.
Moreover, the piezocurrents of mKNN PNPs were proportional to the US
intensity (Supplementary Fig. [189]47). FEA findings reveal the surface
piezo-potential distribution of the mKNN PNP under the acoustic
pressure generated by wf-UMP, which indicates the sufficient energy for
charge carrier migration (Fig. [190]3o)^[191]22. Subsequently, electron
paramagnetic resonance (EPR) spin-trapping experiments were carried out
to detect the generated free radicals using 5, 5-dimethyl-1-pyrroline
N-oxide (DMPO). Typical peaks corresponding to DMPO-•OH and DMPO-
[MATH: ⋅O2− :MATH]
were observed under US stimulation, confirming the existence of ROS •OH
and
[MATH: ⋅O2− :MATH]
in the piezocatalytic process (Fig. [192]3p). Note that the mKNN PNPs
remained structurally stable after US stimulation, highlights the
negligible effect of US on the piezoelectric properties of mKNN PNPs
and their continuous piezocatalytic applications (Supplementary
Fig. [193]48).
In vitro anticancer activity of mKNN PNPs
We investigated the formation of cellular ROS to elucidate the
mechanism of nontoxicity and biocompatibility mKNN PNPs as a
piezocatalyst for anticancer treatment (Supplementary Fig. [194]49 and
Supplementary note [195]6). The results indicated that the
non-piezoelectric HA and Nb[2]O[5] did not affect ROS production in
cells. However, 4T1 cells treated with mKNN+US exhibited a substantial
increase in intracellular ROS levels, as evidenced by higher green
fluorescence and significant average fluorescence intensity
(Fig. [196]4a, b and Supplementary Fig. [197]50). Therefore, the mKNN
PNPs, possessing superior piezoelectric performance, enhance PCT to
tumors primarily by increasing ROS production while minimize the impact
on surrounding healthy cells. Flow cytometry (FCM) analysis of
apoptosis confirmed a significantly higher apoptosis level in 4T1 cells
treated with mKNN+US compared to those treated with mKNN only, whereas
no significant increase in apoptosis levels was observed in the HA and
Nb[2]O[5] groups, regardless of US stimulation (Fig. [198]4c and
Supplementary Fig. [199]51). Furthermore, JC-1 dyes were employed to
investigate changes in mitochondrial membrane potential (MMP)^[200]38,
as excessive ROS production is associated with mitochondrial
dysfunction. The mKNN+US group exhibited the most intense green
fluorescence and weakened red fluorescence compared to the control
groups, suggesting that mKNN+US significantly heightened oxidative
stress and reduced MMP in 4T1 cells, which may account for the induced
apoptosis (Fig. [201]4d, e and Supplementary Fig. [202]52). Similarly,
treatment with mKNN+US in 4T1 cells significantly increases the ratio
of dead to live cells compared to mKNN alone (Supplementary
Fig. [203]53). Moreover, the number of migratory 4T1 cells treated with
mKNN+US was markedly reduced (Fig. [204]4f and Supplementary
Fig. [205]54). Consequently, intracellular ROS generated by mKNN PNPs
under US stimulation effectively exert cytotoxic effects on 4T1 cells,
inducing apoptosis and inhibiting cell migration.
Fig. 4. In vitro and in vivo antitumor effects of mKNN PNPs and wf-UMP.
[206]Fig. 4
[207]Open in a new tab
a, b Intracellular fluorescence images of ROS stained with DCFH-DA in
4T1 cells post various treatments (a) and corresponding quantification
(b) (n = 3 independent samples). Scale bar, 100 μm. c Quantitative
analysis of early and late apoptotic cell populations in each group
(n = 3 independent samples). d, e Fluorescence images of 4T1 cells
stained with JC−1 post-treatment (d) and the corresponding green/red
fluorescence ratios (e) (n = 3 independent samples). Scale bar, 100 μm.
f, Quantification of 4T1 tumor cell invasion capacity by Transwell
assay following different treatments (n = 3 independent samples). g,
Schematic diagram illustrating the application of wf-UMP to mouse and
the treatment protocol. Created in BioRender. Jin, J. (2025)
[208]http://BioRender.com/x39n446. h, i Average (h) and individual (i)
tumor growth curves for five groups of mice across distinct therapeutic
regimens (n = 5 mice). j Comparative tumor weights of mice in each
group on day 14 (n = 5 mice). Box plots show the median (center line),
first and third quartiles (edges), and data range (whiskers). k
Survival rate curves for tumor-bearing mice in each group (n = 5 mice).
Statistical significance was estimated by the log-rank (Mantel‒Cox)
test with the mKNN MN group. l Body weights of mice in each group after
14 days of treatment (n = 5 mice). m, Histological analysis of tumors
via H&E, Ki−67, TUNEL, and CD31 staining post-treatment (n = 3
independent samples). Scale bar, 100 μm. n–p Quantification of relative
percentages of mean fluorescence intensity of Ki-67 positive cells (n),
TUNEL (o), and CD31 (p) in each treatment group (n = 3 independent
samples). Group annotations: G1, Control; G2, US; G3, MN + US; G4, mKNN
MN; G5, wf-UMP. Results are presented as mean ± s.e.m. Statistical
difference was calculated using a two-tailed unpaired student’s t-test.
Significance thresholds: *p < 0.05, **p < 0.01, ***p < 0.001,
****p < 0.0001. Source data are provided as a Source Data file.
In vivo anticancer activity of wf-UMP
Next, we carried out in vivo experiments using a 4T1 subcutaneous tumor
model to investigate the capability of wf-UMP in inhibiting tumor
growth (Supplementary Fig. [209]55). Figure [210]4g illustrates the
treatment design of wf-UMP for early-stage tumors. It is evident that
tumors were effectively suppressed in the wf-UMP group within 2 weeks,
with negligible effects observed upon US stimulation and MN insertion
(Fig. [211]4h and Supplementary Fig. [212]56). Specifically, the
efficacy of wf-UMP in delaying and inhibiting tumor growth was superior
to the other groups, which was further validated by the reduced weights
and volumes of the resected tumors (Fig. [213]4i, h). The survival
monitoring results of mice after treatment in each group showed that
wf-UMP could effectively prolong the survival of mice (Fig. [214]4k).
Notably, wf-UMP demonstrated effective antitumor therapeutic effects in
a wearable and minimally invasive manner, with exceptional biosafety
evidenced by minimal weight fluctuations and no significant toxic side
effects observed in major organs (heart, liver, spleen, lung, and
kidney) (Fig. [215]4l and Supplementary Fig. [216]57). Tumor sections
stained with H&E demonstrated the therapeutic efficacy of wf-UMP,
characterized by the largest areas of tumor necrosis (Fig. [217]4m).
Immunofluorescence staining of Ki-67 revealed a significantly reduction
in highly proliferative tumor cells following wf-UMP treatment
(Fig. [218]4m, n), and TUNEL staining histological analysis further
confirmed its enhanced therapeutic effect of inducing tumor cell
apoptosis (Fig. [219]4m, o). Anti-CD31 antibody labeling of tumor
vasculature showed the smallest red area in the wf-UMP group,
indicating the suppression of tumor blood vessel proliferation
(Fig. [220]4m, p). Collectively, wf-UMP effectively suppresses tumor
cell growth and proliferation by mediating oxidative stress
amplification, and offers a secure and efficient PCT approach for
early-stage cancers.
In vivo biosafety of wf-UMP and metabolic distribution of mKNN PNPs
We further evaluated the biosafety of repeated use of the wf-UMP
through in vivo experiments. Blank MN patches, mKNN MN patches, and
wf-UMP were applied daily to mice for 5 consecutive days. Mild red
traces of inflammation were observed at the application sites but
resolved by day 10, with no significant skin damage detected during
subsequent weekly applications (Fig. [221]5a). By day 10 and day 28,
H&E staining of skin tissue revealed no significant differences in
epidermal thickness between the wf-UMP group and the other treatment
groups, confirming that no skin damage occurred during these treatments
(Fig. [222]5b, c and Supplementary Fig. [223]58). Systemic biosafety
was further assessed by examining primary organs, including the heart,
liver, spleen, lungs, and kidneys, which showed no toxic side effects
in any group (Fig. [224]5d). Hematological evaluations on day 28
indicated normal levels of white blood cells (WBC), red blood cells
(RBC), and hemoglobin (HGB) across all groups (Fig. [225]5e–g).
Additionally, blood biochemical analysis revealed no abnormalities in
liver and kidney function markers, such as alanine aminotransferase
(ALT), aspartate aminotransferase (AST), and creatinine (CREA), in the
wf-UMP group compared to other groups (Fig. [226]5h–j). These results
collectively confirm that wf-UMP exhibits excellent biosafety in in
vivo cancer therapy.
Fig. 5. In vivo biosafety of wf-UMP and metabolic distribution of mKNN PNPs.
[227]Fig. 5
[228]Open in a new tab
a Representative images of skin condition at the treated site (blue
dotted box) in healthy Balb/c mice at days 0, 3, 6, 10, 14, and 28
after treatment with blank MN patch (MN), mKNN MN patch (mKNN MN), and
wf-UMP (n = 4 mice). Scale bars, 1 cm. b, c H&E staining (b) and
epidermal thickness of the skin (c) in healthy Balb/c mice following
various treatments on day 28 (n = 4 independent samples), while ns
denotes no significance. Scale bar, 200 μm. d, H&E staining of major
organs (heart, liver, spleen, lung, and kidney) from mice treated with
control and wf-UMP at the experimental endpoint. Scale bars, 100 μm.
e–g Blood routine test results for WBC (e), RBC (f), and HGB (g) in
mice after treatments on day 28 (n = 4 biologically independent mice).
h-j, Liver aminotransferase levels for ALF (h), AST (i), and creatinine
(j) levels in the blood after different treatments detected on day 28
(n = 4 biologically independent mice). k Quantification of Nb
concentration in tumors over time following wf-UMP treatment, measured
by ICP-OES, showing the retention of mKNN PNPs at the tumor site (n = 3
independent samples). l Representative biological TEM images showing
distribution of mKNN PNPs within tumors post-treatment with wf-UMP. Red
arrows indicate clusters of nanoparticles. Scale bars, 2 μm (top) and
500 nm (bottom). All data are represented as mean ± .e.m. Statistical
difference was calculated using a two-tailed unpaired student’s t-test.
Source data are provided as a Source Data file.
The metabolism of mKNN PNPs at the treatment site was evaluated by
measuring the concentration of Nb. Following wf-UMP application, Nb was
detected in the subcutaneous tumor of mice, with its concentration
gradually decreasing over time. Nevertheless, the mKNN PNPs remained
localized at the treatment site throughout the treatment period,
enabling sustained anti-tumor effect (Fig. [229]5k). To further
investigate the distribution of mKNN PNPs after wf-UMP application,
transmission electron microscopy (TEM) was employed to assess their
localization in mouse subcutaneous tumors 4 h post-application. TEM
images revealed that the mKNN PNPs were accumulated and randomly
distributed throughout the tumor tissue and maintained their morphology
(Fig. [230]5l).
Immunomodulation of wf-UMP on tumor progression
To elucidate the underlying molecular mechanism between wf-UMP and its
inhibition on breast cancer tumor growth, we conducted RNA sequencing
analysis. A distinct pattern of gene expression was observed for the
mKNN MN group without US activation (Supplementary Fig. [231]59). Upon
comparison between the wf-UMP and control groups, we identified 467
differentially expressed genes (DEGs), comprising 248 down-regulated
genes and 219 upregulated genes (Fig. [232]6a). Gene Ontology (GO)
chord diagram analysis showed that the related genes regulated by
wf-UMP were mainly involved in different gene groups related to immune
processes (Fig. [233]6b). Further analysis using the Kyoto Encyclopedia
of Genes and Genomes (KEGG) pathway enrichment analysis revealed the
significant impact of wf-UMP on relevant signaling pathways, such as
natural killer cell mediated cytotoxicity, T cell receptor signaling
pathway, cytokine-cytokine receptor interaction, and NF-κB signaling
pathway (Fig. [234]6c). Note that both cytokine-cytokine receptor
interaction and T cell receptor signaling pathway exhibited a low
p-value and a high enrichment index in the enrichment analysis,
indicating their crucial role in wf-UMP-induced tumor regression
(Fig. [235]6d, e).
Fig. 6. Transcriptome and anti-tumor immune mechanism analysis of wf-UMP.
[236]Fig. 6
[237]Open in a new tab
a Volcano plots of DEGs between wf-UMP and control cohorts. Significant
DEGs (red and blue dots): adjusted p-value < 0.05, fold change >2. b
Gene ontology (GO) chord plot revealed that the regulation of
immune-related genes in different clusters. c Enrichment study
highlighting KEGG pathways relevant to immune function. d, e Gene set
enrichment analysis-enriched pathways of cytokine-cytokine receptor
interactions (d) and T cell receptor signaling pathways (e) in wf-UMP
compared to control. f FCM results of CD3^+CD8^+ T cells and CD3^+CD4^+
T cells infiltrated in tumors. g, h Quantitative analysis of
tumor-infiltrating CD3^+CD8^+ T cells (g) and CD3^+CD4^+ T cells (h)
(n = 3 independent samples). i Representative flow cytometric analysis
of MDSCs. j-l, Relative quantification of MDSCs (j), NKT cells (k), and
TAM-M1 (l) (n = 3 independent samples). m–p Quantification of IFN-γ
(m), IFN-β (n), IL-6 (o), and TNF-α (p) across treatment groups (n = 3
independent samples). All data are represented as mean ± s.e.m.
Statistical difference was calculated using a two-tailed unpaired
student’s t-test. Significance thresholds: *p < 0.05, **p < 0.01,
***p < 0.001, ****p < 0.0001. Source data are provided as a Source Data
file.
Adaptive immunity, renowned for its persistence and specificity,
represents a pivotal immune response capable of targeting perinecrotic
tumor cells. To elucidate the mechanism by which wf-UMP modulates the
tumor immune microenvironment, we collected 4T1 tumors 14 days after
wf-UMP treatment and analyzed immunophenotypic changes using FCM
analysis. The results showed that wf-UMP induced 8.50% CD8^+ cytotoxic
T lymphocyte infiltration (gated by CD3^+ cells), surpassing that in
other groups (Fig. [238]6f, g). Moreover, the FCM analysis confirmed
elevated expression of CD4^+ T cells in tumors of the wf-UMP group
(Fig. [239]6h). Compared to the other three groups, tumors treated with
wf-UMP showed a reduction in myeloid-derived suppressor cells (MDSCs),
along with a substantial increase in NKT cells (Fig. [240]6i–k). It is
known that tumor-associated macrophages-M1 (TAM-M1) exhibit
pro-inflammatory properties and potent effector activities against
infections and cancer cells, whereas tumor-associated macrophages-M2
(TAM-M2) promote tumor cell development and metastasis^[241]39. The
proportion of TAM-M1 and M1/M2 ratio in the tumors of 4T1 tumor-bearing
mice was significantly higher in the wf-UMP group compared to the
control group (Fig. [242]6l and Supplementary Fig. [243]60).
Subsequently, serum samples from each group of mice were collected for
enzyme-linked immunosorbent assay (ELISA) analysis to detect cytokine
secretion levels, including interferon-γ (IFN-γ), interferon-β (IFN-β),
interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α), aiming to
further assess the antitumor immune response of wf-UMP. The results
show that the secretion of IFN-γ, IFN-β, IL-6, and TNF-α induced by
wf-UMP was 1.1, 1.18, 1.17, and 1.16-fold higher, respectively,
compared to the mKNN MN group (Fig. [244]6m–p). Collectively, through
the reactivation of immune cells and the enhancement of cytokine
release, wf-UMP sustains and stimulates immune cell proliferation,
contributing to antitumor therapy.
Abscopal effect of wf-UMP combined with PD1 blockade
Based on the immunological response induced by wf-UMP in the
subcutaneous tumor model of 4T1 breast cancer mice, we further
investigated the synergistic effect of wf-UMP and Anti-PD1 in inducing
systemic antitumor immunity using a dual tumor model. Five days prior
to treatment initiation, 1 × 10^6 4T1 cells were injected into the
right side of the mice to construct primary tumor, and 5 × 10^5 4T1
cells were subcutaneously inoculated into the left side three days
prior to treatment to establish a distal tumor model. Anti-PD1 was
injected intraperitoneally on days 0, 2 and 4 to monitor the abscopal
effect. The primary tumor was treated with wf-UMP for five consecutive
days, while the contralateral tumor remained untreated (Fig. [245]7a).
The results indicated a limited effect of Anti-PD1 on primary and
distant tumors (Fig. [246]7b, c and Supplementary Fig. [247]61).
Notably, while the Anti-PD1+mKNN MN treatment effectively inhibited
primary tumor growth, the Anti-PD1+wf-UMP treatment exhibited notable
antitumor efficacy, with complete eradication of primary tumors in
three out of five cases (Supplementary Figs. [248]62 and [249]63).
Furthermore, the Anti-PD1+wf-UMP group demonstrated significant
efficacy in reducing untreated contralateral tumors (Fig. [250]7c, d).
H&E-stained primary and distant tumors reveled substantial tumor
damage, further indicating that synergistic immunotherapy can improve
breast cancer outcomes (Supplementary Fig. [251]64). Moreover, there
were no notable changes in body weight among the treated mice, and H&E
staining of major organs revealed no significant toxic side effects
across all treatment groups, indicating the safety and tolerability of
synergistic immunotherapy (Supplementary Figs. [252]65 and [253]66). In
view of the promising efficacy of Anti-PD1 combined wf-UMP
immunotherapy in tumor suppression, we further conducted fluorescent
staining of Ki-67, TUNEL, and CD31 for bilateral tumors. While Anti-PD1
treatment led to a modest decrease in Ki-67 staining and a limited
number of TUNEL-positive cells due to antibody-mediated immunity, the
incorporation of Anti-PD1 with wf-UMP resulted in a substantial
reduction in Ki-67 staining and a higher percentage of TUNEL-positive
cells, indicating significantly inhibited tumor cell proliferation and
enhanced apoptosis in both primary and distant tumors. CD31
fluorescence staining also confirmed a notable reduction in
angiogenesis in tumors subjected to Anti-PD1+wf-UMP treatment, further
demonstrating the efficacy of synergistic immunotherapy with
Anti-PD1+wf-UMP (Supplementary Fig. [254]64).
Fig. 7. Distal effects of wf-UMP induced synergistic immunotherapy and T
cell-mediated anti-tumor immunity.
[255]Fig. 7
[256]Open in a new tab
a Schematic illustration of bilateral tumor treatment strategies.
Created in BioRender. Jin, J. (2025) [257]http://BioRender.com/x39n446.
b Average tumor growth curves for distant tumors (n = 5 independent
samples). c Tumor weights of distant tumors after various treatments
(n = 5 independent samples). Box plots show the median (center line),
first and third quartiles (edges), and data range (whiskers). d Gross
images of distal tumors in each group (n = 5 independent samples). e, f
Representative FCM analysis (e) and quantification of matured DCs (f)
in primary tumors after different treatments (n = 3 independent
samples). g Quantification of CD8^+ T cells in primary tumors (n = 3
independent samples). h, i Representative FCM analysis (h) and
quantification (i) of CD8^+ T cells in distant tumor (n = 3 independent
samples). j Relative quantification of CD8^+CD69^+ T cells in distant
tumors following different treatments (n = 3 independent samples). k, l
Representative FCM plots (k) and relative quantification (l) of
macrophage-M1 in distant tumors (n = 3 independent samples). m
Quantification of NKT cells in distant tumors (n = 3 independent
samples). n, o Representative FCM plots (n) and quantification (o) of
MDSCs in distant tumors (n = 3 independent samples). p, Quantification
of tumor-infiltrating CD8^+PD−1^+ T cells in distant tumors (n = 3
independent samples). q–t Quantification of IFN-γ (q), IL-6 (r),
IL-12P70 (s), and TNF-α (t) levels in different groups (n = 3
independent samples). u, v Relative quantification of CD8^+ T cells (u)
and MDSCs (v) in the spleen (n = 3 independent samples). w Multicolor
immunostaining images of distant tumors showing CD4^+, CD8^+, Foxp3^+,
and DAPI cells in the control and Anti-PD1+wf-UMP groups. Scale bars,
50 μm. Group annotations: G1, Control; G2, US; G3, Anti-PD1; G4,
Anti-PD1-PD1+ mKNN MN; G5, Anti-PD1+wf-UMP. All data are presented as
mean ± s.e.m. Statistical difference was calculated using a two-tailed
unpaired student’s t-test. Significance thresholds: *p < 0.05,
**p < 0.01, ***p < 0.001, ****p < 0.0001. Source data are provided as a
Source Data file.
Synergistic immunotherapy-induced T cell-mediated antitumor immunity
To elucidate the therapeutic mechanism of wf-UMP combined with PD1
blockade, immune cells in primary and distant tumors were analyzed
using FCM on day 14 post-treatment. The results showed that
Anti-PD1+wf-UMP group exhibited strong immune activation and
significantly improved maturity of DCs in the primary tumors, which was
much higher than that of other treatment groups (Fig. [258]7e, f). As
professional antigen-presenting cells (APCs), DCs play a critical role
in activating T cells by presenting tumor antigens, and their
maturation correlates with antigen presentation capacity^[259]40. As a
result, the Anti-PD1+wf-UMP group exhibited a significant increase in
CD3^+ T cells, CD8^+ T cells, and CD4^+ T cells in the primary tumors
(Fig. [260]7g and Supplementary Figs. [261]67 and [262]68). For
adaptive immunity, the percentage of intratumoral CD8^+ T cells in
Anti-PD1+wf-UMP group was significantly higher than that of the other
treatment groups in the distant tumors (Fig. [263]7h, i). In addition,
CD69 was identified as an early marker of T cell activation, and the
results showed that the proportion of CD8^+CD69^+ T cells in the distal
tumors of the Anti-PD1+wf-UMP treatment group was much higher than
other groups (Fig. [264]7j)^[265]41. This group also showed dramatic
increases in TAMs-M1, M1/M2 ratio and NKT cells in distant tumors,
while MDSCs and regulatory T cells (Tregs) were significantly
decreased, compared to both PD-1 blockade alone and Anti-PD1+mKNN MN
treatment groups (Fig. [266]7k-o and Supplementary Figs. [267]69 and
[268]70). CD8^+ T cell infiltration in the tumor microenvironment (TME)
is important for the efficacy of immune checkpoint inhibitors (ICIs).
Previous studies have established a correlation between the frequency
of CD8^+ T PD-1^+ cells in the TME and the clinical efficacy of PD-1
blockade therapy^[269]42. Furthermore, the depletion of CD8^+ T cells
due to PD-1 overexpression is a key mechanism of immune evasion in
tumors^[270]43. The results showed that the proportion of CD8^+ T
PD-1^+ cells in the Anti-PD1+wf-UMP treatment group was significantly
reduced compared to other control groups (Fig. [271]7p). Moreover,
serum cytokine assays (including IFN-γ, IL-6, IL-12P70, and TNF-α)
revealed that Anti-PD1+wf-UMP treatment induced markedly higher
cytokine secretion compared to all other groups (Fig. [272]7q–t).
To facilitate immunosuppression, monocytes derived from the enlarged
spleen migrate to the TME^[273]44. The spleen weights in the control
and US groups were significantly higher compared to the other groups,
whereas the Anti-PD1 and Anti-PD1+mKNN MN groups exhibited a slight
reduction in spleen weights. Notably, spleens in the Anti-PD1+wf-UMP
group showed normal morphology alongside significantly reduced weight
compared to the other groups (Supplementary Fig. [274]71). These
results suggest that synergistic immunotherapy attenuates splenomegaly
and diminishes immunosuppression. Further, the synergistic
immunotherapy was examined through FCM analysis to determine its
capability to promote the expansion and proliferation of cytotoxic T
lymphocytes in peripheral immune organs, aiming to elucidate the
mechanism of interaction between Anti-PD1 and wf-UMP in enhancing
antitumor efficacy. Evaluation of immune cells in the spleen reveled a
significant increase in tumor-infiltrating T cells (CD3^+CD8^+) in the
Anti-PD1+wf-UMP group with a suppression of MDSCs (Fig. [275]7u, v and
Supplementary Fig. [276]72). These findings indicate that this
synergistic immunotherapy poses a great potential in activating immune
responses, thereby effectively promoting the activation of cytotoxic T
lymphocyte and downregulating immunosuppressive cells.
Furthermore, polychromatic immunofluorescent staining of primary and
distant tumors showed that Anti-PD1+wf-UMP treatment resulted in
increased CD8^+ and CD4^+ staining in both tumors, which is essential
for inducing cytotoxic T cell-mediated immunotherapy. Conversely,
staining for Forkhead box P3-positive (Foxp3), a marker indicative of
Tregs, notably decreased (Fig. [277]7w and Supplementary Fig. [278]73).
Moreover, there was a significant increase in F4/80^+ and CD86^+
staining following treatment with Anti-PD1+wf-UMP, indicating an
increase in M1 macrophages in both primary and distant tumors
(Supplementary Fig. [279]74).
Long-term immune memory function
Immune memory response is critical for tumor prevention and recurrence.
To assess the immune memory effect of Anti-PD1+wf-UMP treatment, we
performed a tumor rechallenge study. Tumors were reinoculated on the
left side of the mice at day 30 after the end of combination therapy,
and controls were sex- and age-matched mice that had not been
previously inoculated with tumors (Fig. [280]8a). The results showed
that the Anti-PD1+wf-UMP group effectively inhibited the growth of
reinoculated distant tumors, whereas the tumors of the control mice
progressed rapidly, which was also confirmed by tumor weights
(Fig. [281]8b–f and Supplementary Fig. [282]75). To further investigate
the long-term antitumor immune memory induced by the Anti-PD1+wf-UMP
treatment, we analyzed the expression of antigen-specific memory T
cells in the spleen, including effector memory T (Tem) cells (CD44^+
CD62L^−) and central memory T cells (Tcm) cells (CD44^+ CD62L^+). Tem
subsets, which are localized in peripheral tissues, provide rapid
responses to antigen exposure and immediate protection through the
production of multiple cytokines^[283]39. FCM results showed that the
Anti-PD1+wf-UMP treatment elicited a sustained antitumor response
compared to other groups, characterized by a significant transition
from the Tcm to the Tem phenotype. Specifically, this treatment
upregulated Tem (CD44^+ CD62L^−) and down-regulated Tcm (CD44^+
CD62L^+) in both CD8^+ and CD4^+ T cells (Fig. [284]8g–i). These
findings indicate that Anti-PD1 combined with wf-UMP not only induces
specific long-term protection but also effectively prevents tumor
recurrence. Therefore, this combination immunotherapy inhibits systemic
tumor formation and recurrence over an extended period. Collectively,
Anti-PD1+wf-UMP presents a promising PCT strategy that improves
antitumor immunity by further activating DCs maturation through ICD and
the up-regulation of macrophage M1 phenotype polarization, thereby
eliciting T cell-mediated anti-tumor immunity and effectively
inhibiting tumor recurrence (Fig. [285]8j).
Fig. 8. Long-term immuno-memory effects of wf-UMP in combination with PD1
blocker.
[286]Fig. 8
[287]Open in a new tab
a Schematic of the experiment timeline for the tumor rechallenge study.
b Representative tumor images captured 15 days after tumor rechallenge
(n = 5 independent samples). c–e Tumor growth curves during the
rechallenge study: overall growth curves (c), individual tumor growth
curves in the control group (d), and the Anti PD1+wf-UMP group (e).
Data are presented as mean ± SD (n = 5 independent samples). f Tumor
weights measured 15 days post-rechallenge (n = 5 independent samples).
g, h Representative flow cytometric analysis (g) and quantification of
effector memory T cell (Tem, CD44^+ CD62L^−) and central memory T cell
(Tcm, CD62L^+ CD44^+) from CD8^+ T cells (h) in the spleen after
different treatments (n = 5 independent samples). i Quantitative
analysis of Tem and Tcm subsets from CD4^+ T cells in the spleen after
different treatments (n = 5 independent samples). j Proposed mechanism
of wf-UMP synergistic checkpoint blockade immunotherapy for distant
tumor suppression. TAM-M1, tumor-associated macrophages-M1; TAM-M2,
tumor-associated macrophages-M2; iDC, immature dendritic cells; mDC,
mature dendritic cells. Created in BioRender. Jin, J. (2025)
[288]http://BioRender.com/r08a561. All data are presented as
mean ± s.e.m. Two-sided Student’s t-test was used to calculate the
statistical difference between two groups. Significance thresholds:
*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Source data are
provided as a Source Data file.
Discussion
We presented a principled design approach and implementation of an
integrated wf-UMP for convenient, efficient, and minimally invasive
cancer therapy based on a ROS-driven dynamic therapeutic strategy. By
systematically optimizing materials and device architecture, the wf-UMP
ensures to be robustly worn on curved and dynamic tissue surfaces for
easy and effective manipulation in view of the soft mechanical
properties and superior acoustic-electrical transduction
characteristics, thereby efficiently generate ROS with optimized PNPs.
Preclinical studies in mice have shown that wf-UMP is effective not
only in treating localized cancers and modulating immunity, but also in
inducing a synergistic immune response through the DCs maturation
activated by ICD and the modulation of macrophage polarization in
combination with Anti-PD1, thereby capable of targeting distantly
located cancers and inhibiting tumor recurrence. Our study offers an
approach to cancer immunotherapy and underscores the promise of
clinical translation of such a wearable, portable platform.
Furthermore, wf-UMP promotes the transformation towards preventive and
proactive health care, enhancing accessibility and affordability of
cancer therapy.
Considering the rapid advancements in wearable medical systems and the
urgent need for personalized healthcare, the future integration of
modules such as modulus, temperature, pH, and tissue oxygenation
detections into wf-UMP could enable close-loop real-time monitoring and
targeted therapy via artificial intelligence networks, contributing to
the progression beyond current one-size-fits-all healthcare
models^[289]45. For future iterative optimization, it is possible to
implement a 2D phased array as the wearable ultrasonic probe to take
full advantage of the deep penetration and directionality of US, which
not only allows for zoned stimulation of tumors by beamforming
technology without disturbing other healthy tissues, but also allows
for continuous tracking of tumor status as well as physiological
signals from deeper tissues before and after therapy by through
clinical-quality US imaging^[290]13,[291]19,[292]24,[293]46. The
bioadhesive hydrogel can be further integrated with electrical,
thermal, optical, and biological functions to facilitate the
human-machine interaction of the device^[294]47,[295]48. Biologically
and genetically engineered PNPs are advised to interface with in vivo
microbial system and enhance the bioavailability of loaded
nanomedicines^[296]49,[297]50. Moreover, leveraging the benefits of
wf-UMP, the loaded nanomedicines can be functionalized and customized
to target and treat various diseases, whether they present
superficially on the skin or infiltrate deeper into tissues.
Given the complex physiological environment of living organisms, wf-UMP
may demonstrate significant differences in therapeutic behavior and
efficacy. Therefore, more in-depth studies into the immune reaction and
immunosuppressive microenvironment are necessary to evaluate these
therapeutic formulations and develop optimized strategies for cancer
immunotherapy. For clinical translation, the realization of scalable
and reproducible wf-UMP-based therapies that meet high standards may be
challenging and require additional efforts. Further studies involving
large animal models, clinical trials, and patient feedback are
essential for advancements towards clinical applications. Meanwhile,
establishing a clear regulatory pathway for wearable US electronics and
setting standards for PCT will accelerate their safe and effective
integration into clinical validation and potential FDA regulatory
approval, ultimately transforming wf-UMP from a highly promising
prototype to an indispensable instrument for personalized
medicine^[298]46.
Methods
Preparation of piezoceramics and corresponding PNPs
The piezoceramics of
0.95K[0.48]Na[0.52]Nb[0.97]Sb[0.03]O[3]-0.05Bi[0.5]Na[0.5]ZrO[3]-0.2%
Fe[2]O[3] (abbreviated as mKNN) and K[0.48]Na[0.52]NbO[3]-0.2%
Fe[2]O[3](abbreviated as KNN) were synthesized using the conventional
solid-state reaction method. High-purity raw materials including
K[2]CO[3] (99.0%), Na[2]CO[3] (99.8%), Nb[2]O[5] (99.5%), Sb[2]O[3]
(99.99%), Bi[2]O[3] (99.99%), ZrO[2] (99.0%), and Fe[2]O[3] (99.0%)
purchased from Sinopharm Chemical Reagent Co., Ltd were accurately
weighted according to their stoichiometric ratios, respectively. Then,
the raw materials were ball-milled with alcohol and ZrO[2] balls for
24 h, followed by calcination at 850 °C for 6 h. Subsequently, the
calcined ceramic powders were blended with 8 wt.% polyvinyl (PVA,
Sinopharm Chemical Reagent Co., Ltd) and compressed into ceramic
pellets under a pressure of 10 MPa. After the PVA was burned off, the
ceramics were obtained by sintering the pellets at 1085 °C (mKNN) and
1090 °C (KNN) for 3 h. The poling process was carried out at room
temperature with a direct current electric field of 2 kV mm^−1. For the
piezocatalytic PNPs, the as-calcined powders were secondary ball-milled
and then sand-grounded (VB-0.3Q, Suzhou Vgreen Nano-chem Technology
Co., Ltd, China) at 2000 rpm for 8 h. The control material of Nb[2]O[5]
NPs was fabricated by directly sand-grounding the raw material
Nb[2]O[5].
Preparation of MN patches
For the MN patches, 300 mg of hyaluronic acid (HA, 10 kDa, Meilunbio,
Dalian, China) were dissolved into 4 mL deionized water. 500 μL of the
homogeneous solution were then dropped onto the polydimethylsiloxane
(PDMS, Sylgard 184) MN mold. After keeping in a vacuum oven of 0.08 MPa
for 10 min to remove bubbles, the MN mold fulfilled with HA solution
was dried at 37 °C for 12 h. The MN patch was obtained by gently
peeling it off from the mold (Supplementary Fig. [299]1). HA MN patches
with chemical loadings were prepared by the same process, with 20 mg
sufficient chemicals (mKNN PNPs) added at the beginning. To prepare the
Rho B MN patches, 300 mg of HA were dissolved into 4 mL 1 mg mL^−1 Rho
B solution. The subsequent preparation process was similar to that of
the HA MN patch mentioned above.
Preparation of polyacrylamide bioadhesive hydrogel and agarose hydrogel
First, 2 g of gelatin (Sinopharm Chemical Reagent Co., Ltd) and 6 mL of
acrylic acid (AR, ChengDu Chron Chemicals Co., Ltd) were dissolved in
14 mL of deionized water. Subsequently, the photoinitiator
α-ketoglutaric acid (0.04 g, 98%, Macklin), the cross-linker
N,N’-methylenebisacrylamide (0.01 g mL^−1, 98%, Adamas-beta) and
N-hydroxysuccinimide (NHS, 1% w/w, 98%, Tianjin Kemiou Chemical Reagent
Co., Ltd) were added to the solution. After stirring for 10 min, the
mixture was subjected to US to remove air bubbles. The bioadhesive
hydrogel was then obtained by curing in silicon molds with an
ultraviolet light (UV) chamber (284 nm, 10 W). Prior to use, the
bioadhesive hydrogel was sealed in plastic bags and stored at −20 °C.
The agarose hydrogel was fabricated by adding 2 g of agarose into 40 ml
of deionized water and stirring for 30 min in a 100 °C oil bath. The
solution was promptly poured into silicon molds and maintained at 4 °C
until the agarose hydrogel was obtained.
Fabrication of wf-UMP
The detailed process of wf-UMP fabrication was shown in Supplementary
Fig. [300]1. To optimize the performance of piezo-element, mKNN
ceramics were initially processed into ceramic/epoxy (ELINOPTO E106-7
A/B) 1–3 composites via the dice-and-fill method (DS9260).
Specifically, the ceramic pillars were designed with a width of 500 μm
and a spacing of 200 μm. After being polished to the desired thickness,
the 1–3 piezo-units were sputtered with Au electrodes on both sides
(TPR450). Then the composites were cut into 2.5 mm × 2.5 mm × 1.5 mm
cubes and poled. The 24 poled piezo-units were then neatly arranged and
connected to the laser-cut island-bridge Cu/PI electrodes through
silver paste (E-Solder 3022). The entire transducer array was
subsequently encapsulated in silicon (Ecoflex 00–30) to ensure the
flexibility and stretchability. Finally, the bioadhesive hydrogel and
mKNN MN patch were implanted onto the surface of the encapsulated
transducer array to form the device, wf-UMP.
Fabrication of PNPs-based PEHs
The PEH was manufactured by mixing PDMS, curing agent and mKNN/KNN PNPs
in a weight ratio of 10:1:4. The resulting homogenous slurry was then
spin-coated onto quartz glass. After drying at 80 °C for 4 h, the
PDMS-mKNN/KNN films were peeled from the glass substrate. Subsequently,
the films were cut into 2 cm × 2 cm squares and packaged with Cu/PI
electrodes and Kapton films to create the PEHs.
Structure and performance characterization of mKNN and MN patch
The microstructure of MN patches was characterized by scanning electron
microscope (SEM, agellan400, FEI Company) and optical microscope
(SZM7045, SOPTOP). The distribution of mKNN PNPs within the MNs was
further observed by inverted fluorescence microscope using Rho B as a
marker. The permeation of the loaded drugs at different skin depths was
characterized by confocal laser microscopy (Nikon A1si). The
morphology, lattice fringes, and corresponding energy dispersive
spectrometer were obtained by the high-resolution transmission electron
microscopy (HR-TEM, Thermo Scientific Talos F200X). To investigate the
phase structure, both bulk ceramics and PNPs were assessed by XRD with
Cu Kα radiation (Bruker D8 Advanced XRD, Bruker AXS Inc., Madison, WI,
USA). In addition, the temperature-dependent dielectric constants of
bulk ceramics were investigated by the Dielectric Spectroscopy test
system (TZDM-200-300, Harbin Julang Technology Co. Ltd., Harbin China)
at elevating frequencies from 1 kHz to 100 kHz. PFM measurements were
recorded by the piezoelectric force microscopy (PFM, MFP-3D, Asylum
Research, Goleta, CA) with Dual AC Resonance-Tracking and SS-PFM modes.
The P-E loops and S-E curves were characterized using a ferroelectric
equipment (TF Analyzer 2000: aixACCT, Aachen, Germany) at 1 Hz. The
surface chemical nature and valence band of PNPs were detected by the
X-ray photoelectron spectroscopy (XPS, AXIS Ultra DLD). The bandwidth
of PNPs was examined using the UV-vis spectrophotometer (UV-3600,
Hitachi, Japan).
Performance of mKNN PNPs under US stimulation
To evaluate the piezocatalytic activity of mKNN PNPs, degradation
experiments of Rho B under US stimulation (40 kHz, 120 W) were carried
out. Specifically, 100 mg mKNN PNPs were added into 50 ml Rho B
solution (5 mg L^−1). Note that the degradation experiments were
carried out in darkness and the ambient temperature was preciously
controlled within ± 1 °C to avoid photocatalysis and pyroelectric
catalysis. The mixture was stirred in darkness for 30 min to reach
adsorption-desorption equilibrium. Then, 3 mL of the mixture was
pipetted and centrifuged every 5 min, and the supernatant from each
centrifuged sample was further examined by a UV-vis spectrophotometer
to determine the dye concentration.
The output voltages of the PNPs-based PEHs under instantaneous
compression were recorded using an electrometer (6517B, Keithley).
Transient piezo-currents of PNPs were tested by an electrochemical
workstation (CHI660E, Shanghai Chenhua Instrument Corporation, China)
equipped with a three-electrode system. Platinum served as the counter
electrode, Ag/AgCl as the reference electrode, and FTO glass coated
with PNPs functioned as the working electrode. The experiments were
conducted in a 0.1 M Na[2]SO[4] electrolyte solution. Then, the
generation of ROS was detected by EPR (Bruker EMXplus X-Band EPR
spectrometer) experiments after 5 min of US stimulation.
Characterization of wf-UMP
The impedances and phase angles of bulk ceramic, 1–3 piezo-unit, and
wf-UMP were evaluated by an impedance analyzer (E4991B, Keysight, USA).
The k[t] and k[eff] values can be calculated through the following
Eqs. [301]1 and [302]2^[303]11:
[MATH: kt=
π2frf
atanπ2fa−frf
r :MATH]
1
[MATH: keff=1
−fr2fa2 :MATH]
2
The acoustic pressure of wf-UMP was measured by a hydrophone (3646,
Precision Acoustics Ltd, UK). Thermal imaging pictures of wf-UMP in
operation were recorded by a thermal imaging camera (H21Pro, Hikmicro).
Mechanical tests
The mechanical properties of MN patches were measured by the electronic
universal testing machine (TSC504B), operating at a compression speed
of 1 mm min^−1. For bioadhesive hydrogels, mechanical characterizations
were conducted using an electronic universal testing machine (Instron
5967). Rectangular hydrogel samples, measuring 60 mm in length, 20 mm
in width, and 1.6 mm in thickness, were evaluated at a constant tensile
speed of 50 mm min^−1. To assess the adhesion performance of the
bioadhesive hydrogel on porcine skin, 180° peeling tests were
performed. Adhesion samples with an adhesion area measuring 2 cm in
width and 6 cm in length were prepared and tested at a constant speed
of 25 mm min^−1.
FEA simulations
FEA simulations on the mechanical performances, the acoustic outputs of
wf-UMP, and the piezo-potential distribution of mKNN PNPs were
conducted using COMSOL Multiphysics 6.1, ensuring strict adherence to
the experiment settings. The physical parameters involved have been
carefully considered and appropriately selected.
Theoretical calculations
First-principles calculations of the electronic properties of mKNN were
conducted using the Vienna Ab initio Simulation Package (VASP) within
the framework of density functional theory (DFT). To describe the
exchange-correlation functional, the PBEsol parametrization of the
generalized gradient approximation was employed. The wave functions
were expanded in a plane wave basis set with a cut-off energy of
520 eV. For Brillouin zone sampling, Γ-centered k-point meshes with a
grid of spacing 0.04 × 2π Å^−1 were utilized. The self-consistency
cycle was iterated until achieving an energy convergence of 10^−5 eV,
the Kohn-Sham orbitals were updated.
Cells and animals
The 4T1 mouse breast cancer cell line (Cat CRL-2539) was sourced from
the American Type Culture Collection (ATCC, Manassas, VA), while Hacat
cells were obtained from the Laboratory of Radiation Medicine, School
of Basic Medical Sciences, Sichuan University. 4T1 cells were cultured
in RPMI Medium 1640 basic, and Hacat cells were maintained in
Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal
bovine serum (FBS, Gibcom, USA) and 1% penicillin/streptomycin (w/v).
Cells were cultured in a humidified incubator at 37 °C with 5% CO[2].
6–8 weeks old Balb/c mice (female) were purchased from GemPharmatech
LLC (Jiangsu, China) and housed in an environment with a 12 h
light-dark cycle, an ambient temperature of 18~22 °C, and 50 ~ 60
humidity. All animal experiments were approved by the Animal Care and
Use Committee of Sichuan University and conformed to the Guide for the
Care and Use of Laboratory Animals published by the National Institutes
of Health.
Cytotoxicity measurement
The cytotoxicity of mKNN against Hacat cells was evaluated using a CCK8
assay. Hacat cells were cultured at a density of 5 × 10^3 cells per
well in 96-well plates and incubated overnight. Next, various
concentrations of mKNN (0, 0.1, 0.2, 0.3, 0.4, and 0.5 mg mL^−1) were
added to cell-containing medium for an additional 24 h. Finally, the
standard CCK8 assay protocol was performed, and the optical density
(OD) at 450 nm per null was detected using a microplate reader. The in
vitro antitumor efficacy of mKNN against 4T1 cells was evaluated by
CCK8 assay. In 96-well plates, 4T1 cells were cultured at a density of
5 × 10^3 cells per well. After 24 h of culture, mKNN at varying
concentrations (0, 0.025, 0.05, 0.1,0.2, and 0.3 mg mL^−1) was added to
4T1 cells and co-cultivated for 4 h, followed by US stimulation
(1.2 MHz, 30 s). Cells were stained with CCK8 solution and assessed
using a microplate reader after an additional 24 h of incubation. Each
experiment was performed independently three times.
ROS measurement
To assess intracellular ROS levels, 4T1 cells were inoculated into
6-well plates at a density of 2 × 10^5 cells per well and cultured for
24 h in fresh medium containing PBS, HA, Nb[2]O[5], and mKNN at a
concentration of 0.2 mg mL^−1. After 4 h of incubation, US stimulation
(1.2 MHz, 30 s) was applied. The cells were then labeled with DCFH-DA
and observed using a fluorescence photographic recording.
Apoptosis assay
4T1 apoptosis was assessed as follows: 4T1 cells were inoculated into
6-well plates at 1 × 10^5 cells per well and cultured overnight. Cells
were divided into 8 treatment groups: (1) PBS (control), (2) HA, (3)
Nb[2]O[5], (4) mKNN, (5) US, (6) HA + US, (7) Nb[2]O[5] + US, and (8)
mKNN+US. Cells in each group were treated for 4 h with the same
concentrations of HA, Nb[2]O[5] and mKNN, then stimulated with US
(1.2 MHz, 30 s). After a further 24 h of incubation, the cells were
stained with FITC-Annexin V/propidium iodide (PI) apoptosis detection
kit and analyzed by FCM.
Live/dead cell staining and JC-1 staining
4T1 cells were seeded overnight in 6-well plates and divided into 4
groups: (1) PBS (control), (2) US, (3) mKNN, and (4) mKNN + US. After
exposing the cells to the respective treatments for 4 h, the cells were
stimulated with US for 30 s (1.2 MHz). Following cell staining with
calcein-AM/PI and JC-1, the cells were incubated at 37 °C with 5% CO[2]
for 30 min in a humidified environment. Subsequently, the cells
underwent three PBS washes and were analyzed with a fluorescent
microscope.
Transwell assay
To assess the impact of mKNN on cell migration, a transwell assay was
conducted using the same treatment grouping and workflow as in the
live/dead experiments. After US stimulation (1.2 MHz, 30 s), cells were
collected, and 8 × 10^4 cells per well were seeded into the upper
chamber with FBS-free 1640 medium, and the lower chamber, which
contained medium supplemented with 10% FBS. Following a 24 incubation,
cells that had migrated to the bottom of the polycarbonate membrane
were fixed with 4% paraformaldehyde and stained with 0.1% crystal
violet. The migration of cells was then recorded using a fluorescence
microscope.
In vivo tumor therapy
Female Balb/c mice were acclimatized to the environment for 1 week
before tumor model induction. Unilateral tumor model was established by
subcutaneous injection of ~1 × 10^6 4T1 cells suspended in 100 μL of
PBS into the right posterior ventral side of each mouse. Upon reaching
a tumor volume of ~60 mm^3, the mice bearing 4T1 tumors were randomly
assigned to five distinct groups (n = 5 mice): (1) control (no
treatment), (2) US, (3) MN + US (blank MN + US), (4) mKNN MN (mKNN
PNPs-loaded MN), (5) wf-UMP (mKNN MN + US). Under isoflurane
anesthesia, MN patches were administered to the mice. The wf-UMP was
pressed onto the tumor surface with the thumb. 2 min later, the MNs
were completely dissolved, and the wf-UMP was operated with a trigger
voltage of 10 Vpp for 5 min on days 0, 1, 2, 3, and 4. Subcutaneous
tumor volume and mouse weight were recorded every 2 days. Tumor volumes
(mm^3) were calculated using the formula:
V = (length) × (width)^2 × 2^−1. At the end of 14 days of treatment,
tumors were harvested and weighted after euthanasia of mice.
Subcutaneous tumors and major organs of mice were fixed with 4%
formaldehyde and analyzed by H&E staining. Immunofluorescence staining
(Ki-67, CD31, TUNEL) was performed on tumor samples from each group.
Meanwhile, single-cell suspensions were prepared from mouse
subcutaneous tumors for immune FCM. In order to detect MN-induced
mircoholes, mice were executed after tumor insertion of MN, and tumors
were fixed with 4% paraformaldehyde and then stained with H&E.
In vivo biosafety assessments
To evaluate the biosafety of the wf-UMP, healthy Balb/c mice were
randomly divided into 4 groups (n = 4 mice): (1) control group; (2)
blank MN patch group (MN); (3) mKNN MN patch group (mKNN MN); (4)
wf-UMP group (mKNN MN + US). Patches or wf-UMP were applied
continuously for the first 5 days of treatment, followed by once-weekly
applications thereafter. Skin recovery was recorded on days 0, 3, 6,
10, 14, and 28 post-treatments, and on day 10 and day 28, treated skin
samples were collected for H&E staining to assess epidermal thickness.
To further assess the in vivo biocompatibility, blood samples were
collected from the orbital veins of mice in each group to analyze
routine blood and blood biochemical parameters. Major organs (heart,
liver, spleen, lung, and kidney) of mice were fixed with 4%
formaldehyde and analyzed by H&E staining to assess potential
histopathological changes. Transmission electron microscopy (TEM)
(JEM-1400FLASH, JEOL) was used to confirm the distribution of mKNN
PNPs. Inductively coupled plasma optical emission spectroscopy
(ICP-OES) (Thermo ICAP PRO) was used to determine the concentrations of
niobium ions.
RNA sequencing
In the above-mentioned subcutaneous tumor experiments in mice, on day
14, tumor samples were collected from the control, mKNN MN, and
mKNN+wf-UMP groups using sterile techniques for RNA sequencing. RNA
libraries were constructed using Bioanalyzer 2100 and RNA 6000 Nano
LabChip Kit (Agilent, CA, USA) following manufacturer’ s instructions,
ensuring the creation of high-quality RNA sequencing libraries. These
libraries were then sequenced using the Illumina Novaseq 6000^TM
platform (LC-Bio Technology CO., Ltd., Hangzhou, China), generating
double-ended reads for subsequent bioinformatics analysis. In order to
regulate the false-positive ratio, the p-value modified by the multiple
hypothesis testing (Padj) was examined.
FCM analysis
Single-cell suspensions were prepared by digesting isolated tumor
tissue or spleen. The immune cells were construed following the
standard protocol. To prevent non-specific binding, the cell suspension
was incubated with BD Fc Block^TM purified anti-mouse CD16/CD32
monoclonal antibody (BD Pharmingen^TM, Catalog: 553141) for 15 min at
4 °C. Then, cells were stained with surface antibodies: CD45-APC-CY79
(BD Pharmingen^TM, Catalog: 557659), CD3-BUV395 (BD Pharmingen^TM,
Catalog: 563565), CD4-BUV661 (BD Pharmingen^TM, Catalog: 741461),
CD8-BV711 (BD Pharmingen^TM, Catalog: 563046), CD49b(BD Pharmingen^TM,
Catalog: 746974), CD44-FITC (BD Pharmingen^TM, Catalog: 561859),
CD62L-PerCP-Cy5.5 (BD Pharmingen^TM, Catalog: 570282), CD69-PE-CY7 (BD
Pharmingen^TM, Catalog: 552879), CD25-BV786 (BD Pharmingen^TM, Catalog:
564368), PD-1-BV650 (BD Pharmingen^TM, Catalog: 744546), CD11B-BV421
(BD Pharmingen^TM, Catalog: 562605), F4/80-BV605 (BD Pharmingen^TM,
Catalog: 743281), CD86-BUV737 (BD Pharmingen^TM, Catalog: 741757),
Gr-1-BUV496 (BD Pharmingen^TM, Catalog: 750615), CD11C-BUV563 (BD
Pharmingen^TM, Catalog: 749091), MHC II-R718 (BD Pharmingen^TM,
Catalog: 752163), CD80-BV510 (BD Pharmingen^TM, Catalog: 740130).
Following fixation and permeabilization, cells were stained with
intracellular markers: FOXP3-PE (BD Pharmingen^TM, Catalog: 563101) and
CD206-AF647 (BD Pharmingen^TM, Catalog: 565250). Finally, using BD
FACSymphony^TM A5 for filtering and detection of stained cells and
FlowJo 10.4 for data analysis. Flow cytometry gating strategy involved
selecting single cells based on FSC-A/SSC-A, removing dead cells using
a viability dye, and identifying specific cell populations through
marker expression (e.g., CD45^+ for immune cells) (Supplementary
Figs. [304]76 and [305]77).
Serum collection and enzyme-linked immunosorbent assay
For each group, blood samples were obtained by ocular phlebotomy in
dried EP tubes on day 14 after treatments. The upper serum was then
centrifuged at 4 °C according to standard steps. The expression levels
of cytokines (including IFN-γ, INF-β, IL-6, IL-12P70, and TNF-α) in the
sera of mice were detected by a commercial ELISA kit (ZCIBIO
Technology, Shanghai, China).
In vivo synergistic immunotherapy
To establish a bilateral tumor model, 1 × 10^6 4T1 cells were injected
into the right abdominal skin of female Balb/C mice. Following a 2 day
interval, 5 × 10^5 4T1 cells were then injected into the left side.
After 3 days, the mice were divided into 5 groups (n = 5 mice): (1)
control (no treatment), (2) US, (3) Anti-PD1, (4) Anti-PD1+mKNN MN
(Anti-PD1 + mKNN PNPs-loaded MN), and (5) Anti-PD1+wf-UMP
(Anti-PD1+mKNN PNPs-loaded MN + US). The right tumor served as the
primary site for synergistic immunotherapy, while the left tumor
represented a distant tumor and remained untreated. Following the
treatment schematic, Anti-PD1 was injected intraperitoneally to the
relevant groups on days 0, 2, and 4 (10 mg kg^−1 per mouse initially,
followed by 5 mg kg^−1 per mouse for the second and third times).
Anti-PD1 antibody was kindly provided by BeiGene, Ltd. Subsequently,
both the Anti-PD1+mKNN MN and Anti-PD1+wf-UMP groups underwent
treatment with mKNN MN patches consecutively for the first 4 days.
Following this, the Anti-PD1+wf-UMP group received an additional US
stimulation with wf-UMP (1.2 MHz, 10 Vpp, 5 min). Bilateral tumor
volumes and mouse body weights were recorded every 2 days. On the 14th
day, spleens and tumors were collected, weighed, and subjected to
immune FCM analysis. Similar to the unilateral tumor model, major
organs and bilateral tumors were stained for H&E, while bilateral
tumors further underwent immunofluorescence (Ki-67, CD31, TUNEL)
staining and polychromatic immunofluorescent staining. Simultaneously,
serum samples from each group were collected to assess cytokine levels.
Polychromatic immunofluorescent staining
Tumor tissue sections were collected for four-color multiplex
fluorescence immunohistochemical staining. Sections were incubated
overnight with CD4 (servicebio, Catalog: GB15064, diluted at 1:1000),
CD8 (servicebio, Catalog: GB15068, diluted at 1:1000), FOXP3
(servicebio, Catalog: [306]GB112325, diluted at 1:1000) antibodies or
F4/80 (servicebio, Catalog: [307]GB113373, diluted at 1:5000), CD86
(servicebio, Catalog: [308]GB115630, diluted at 1:1000), CD163
(servicebio, Catalog: [309]GB113751, diluted at 1:3000) antibodies at
4 °C. Following three washes with PBS, the sections were counterstained
with DAPI for nuclei visualization before being sealed. Subsequently,
all sections were scanned with a fluorescence scanning camera (Akoya
Vectra Polaris).
Tumor rechallenge study
To establish a tumor rechallenge mouse model, 1 × 10^6 4T1 cells were
inoculated into the right abdomen of mice. Five days after inoculation,
the mice were treated with the Anti-PD1 antibody in combination with
wf-UMP. On day 30 after the end of treatment, 1 × 10^6 4T1 cells were
subcutaneously inoculated into the left abdomen of mice in the combined
treatment group and age-matched healthy mice (not inoculated with
tumors). After monitoring tumor growth for 15 days, the mice were
euthanized and tumors were excised and weighed, and spleens were
obtained for FCM analysis.
Inclusion and ethics
Every experiment involving animals have been carried out following a
protocol (approve number: 20240514002) approved by an ethical
commission of the Animal Care and Use Committee of Sichuan University.
Statistics and reproducibility
Experiments were repeated three times with similar results.
Statistical analysis
Fluorescence images were quantitatively analyzed using ImageJ software
(version 1.52a). Data were analyzed using the unpaired, two-tailed
Student’s t-test in GraphPad Prism 9.0 statistical analysis software to
determine p-values. Survival curves of mice were analyzed using the
log-rank (Mantel-Cox) test. Significant differences were defined as
p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).
Reporting summary
Further information on research design is available in the [310]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[311]Supplementary Information^ (7MB, pdf)
[312]41467_2025_58075_MOESM2_ESM.docx^ (12.6KB, docx)
Description of Additional Supplementary Information
[313]Supplementary Movie 1^ (4.5MB, mp4)
[314]Supplementary Movie 2^ (108MB, mp4)
[315]Reporting Summary^ (142.7KB, pdf)
[316]Peer review file^ (4.3MB, pdf)
Source data
[317]Source Data^ (8.2MB, xlsx)
Acknowledgements