Imaging-guided deep tissue in vivo sound printing

Abstract

Three-dimensional printing offers promise for patient-specific implants and therapies, but is often limited by the need for invasive surgical procedures. To address this, we developed the imaging-guided deep tissue in vivo sound printing (DISP) platform. By incorporating crosslinking agent-loaded low temperature sensitive liposomes into bioinks, DISP enables precise, rapid, on-demand crosslinking of diverse functional biomaterials using focused ultrasound. Gas vesicle-based ultrasound imaging provides real-time monitoring and allows for customized pattern creation in live animals. We validated DISP by successfully printing near diseased areas in the mouse bladder and deep within rabbit leg muscles in vivo, demonstrating its potential for localized drug delivery and tissue replacement. DISP’s ability to print conductive, drug-loaded, cell-laden, and bioadhesive biomaterials demonstrates its versatility for diverse biomedical applications.

Three-dimensional (3D) bioprinting has emerged as a transformative tool in medicine, enabling the creation of patient-specific implants (1, 2), intricate medical devices (3, 4), and tissue replacements (5–7). Advancing bioink formulations and extrusion- or light-based printing systems (8, 9) have driven this progress, unlocking applications in diverse fields such as tissue regeneration (10, 11), bioelectronics (12, 13), drug delivery (14, 15), and wound sealing (16, 17). However, the implantation of these constructs often requires invasive surgeries, limiting their utility for minimally invasive treatments (18, 19). in vivo printing technologies would enable direct fabrication of bioconstructs at defect sites within the body, eliminating the need for traditional implantation and facilitating rapid, on-site tissue repair. While near-infrared (NIR) light has been explored as a biosafe energy source for in vivo printing, its applications remain limited to subcutaneous tissues due to the limited light penetration (20, 21).

Ultrasound technology, known for its deep tissue penetration and non-invasive nature, offers a promising platform for in vivo printing. Its real-time imaging capabilities enable precise targeting and control during in situ fabrication of biomaterials (22). Focused ultrasound (FUS), in particular, allows for targeted energy delivery, facilitating processes like acoustic cavitation induced radical polymerization of materials such as polydimethylsiloxane (PDMS) (23, 24) or sonothermally induced polymerization of poly(ethylene glycol) diacrylate (PEGDA) (25) (table S1).

A major challenge to in vivo printing lies in developing versatile bioink formulations capable of accommodating diverse biomaterials for wide-ranging applicability across various medical scenarios while ensuring high biocompatibility and minimal toxicity from residual prepolymers. Comprehensive in vitro and in vivo studies are essential to address these issues. Additionally, advancing these technologies requires systems capable of large-scale and high-resolution printing, as well as seamless integration with real-time imaging to ensure precise focal point positioning, minimize off-target tissue effects, and accelerate the clinical translation.

We developed an imaging-guided deep tissue in vivo sound printing (DISP) platform, which utilizes low temperature sensitive liposomes (LTSLs) as carriers for crosslinking agents to enable precise and controlled in situ fabrication of biomaterials within deep tissues (Fig. 1, A to C and table S1). In DISP, LTSLs enhance biocompatibility by encapsulating crosslinking agents, preventing premature interaction with surrounding tissues, and enabling on-demand release via FUS. Designed to respond to mild temperature changes slightly above body temperature, LTSLs enable activation of various crosslinking mechanisms, including ionic, oxidative, and free radical polymerization (Fig. 1, D and E).

Fig. 1. Imaging-guided deep tissue in vivo sound printing (DISP).

(A) Schematic of the DISP platform. The DISP system utilizes an ultrasound-responsive bioink (US-ink) composed of non-crosslinked prepolymer, crosslinking agent-loaded low temperature sensitive liposomes (LTSLs), and gas vesicles (GVs). The US-ink is injected into the body to noninvasively fabricate a precise functional biostructure in vivo. Integrated GV-based ultrasound imaging is employed to monitor the target organ, detect the presence of the prepolymer, and ensure accurate targeting and successful ultrasound-induced gel (US-gel) formation. (B) In vivo printing setup for focused ultrasound (FUS) generation and monitoring. RF, radio frequency; T/R, transmitter/receiver. (C) Transmission electron microscopy (TEM) image of the crosslinking agent-loaded LTSLs. Scale bar, 100 nm. (D) Scanning electron microscopy (SEM) image of freeze-dried 3D printed alginate US-gel. Scale bar, 20 μm. (E) Functional hydrogel structures printed with sound in vivo printing. Scale bars, 5 mm. (F to H) DISP-based in vivo printing of bioelectronic devices for sensing and recording (F), biocarriers for drug delivery and tissue regeneration (G), and bioadhesives for wound sealing and device/tissue interfaces (H).

DISP achieves high-resolution printing (~150 μm) and fast printing speeds (up to 40 mm s⁻¹). A wide range of functional biomaterials, including conductive, drug-loaded, cell-laden, and bioadhesive hydrogels, were successfully printed (Fig. 1, F to H). Gas vesicle (GV)-based ultrasound imaging (26, 27) is integrated into the printing platform, allowing for real-time monitoring of the printing process, precise focal point positioning, and in situ crosslinking verification. As a proof-of-concept, we demonstrated in vivo printing within the bladders and muscles of live animals, with post-procedure analyses confirming the high biocompatibility of both the prepolymers and printed hydrogels.

The design and mechanism of deep tissue in vivo sound printing

DISP utilizes ultrasound-responsive bioinks, termed US-inks, specifically designed for precise and controlled in situ fabrication. US-inks are composed of biopolymers, crosslinking agent-encapsulated LTSLs, and GVs that act as ultrasound imaging contrast agents. These bioinks are delivered to the target sites through injection or catheters, and are located using an ultrasound imaging setup integrated into 3D printing platform. This system accurately targets the FUS focal point onto the US-ink and continuously monitors the printing process.

The FUS transducer, controlled by an automatic positioning system, scans over the US-ink following a predefined G-code. Localized heating induced by FUS triggers the release of crosslinking agent from the LTSLs, enabling immediate in situ crosslinking of the US-ink (Fig. 1, A and B). Transmission electron microscopy (TEM) confirmed the integrity of the LTSLs (Fig. 1C), while scanning electron microscopy (SEM) revealed uniform crosslinking and the formation of US-gels (Fig. 1D). DISP enables the synthesis and precise patterning of functional US-gels with diverse properties, including conductivity, biocarrier capacity, and bioadhesion (Fig. 1E), expanding potential applications in bioelectronics, drug delivery, tissue regeneration, wound sealing, and other medical interventions (Fig. 1, F to H).

Low temperature sensitive liposomes for controlled crosslinking

Low temperature sensitive liposomes are widely used in drug delivery due to the ability to precisely control release temperatures by engineering lipid bilayer properties. While certain liposomes can be activated using low-frequency ultrasound from probe sonicators to compromise their membranes for crosslinking applications (28), LTSLs enable remote and precise activation via FUS, allowing superior spatial control. In this study, the LTSLs were designed to remain stable at 37 °C and rapidly release encapsulated materials at ~41.7 °C (29). Upon FUS exposure, temperature increases locally within the US-ink, which induce a phase transition in the LTSL lipid bilayer from a solid to a liquid state, creating nanopores in the bilayer structure (Fig. 2A). Grain boundary defects in the solid phase were found to be crucial for the formation of stable nanopores in the lipid bilayers. To enhance lipid mobility and facilitate pore formation, pore-forming lysolipids and a small percentage of PEGylated lipids were incorporated into the lipid bilayer (fig. S1). During heating, these defects expand rapidly (29, 30), enabling controlled payload release while minimizing premature leakage at physiological temperatures, ensuring precise and reliable crosslinking for sound printing.

Fig. 2. Synthesis and characterization of low temperature sensitive liposomes for controlled release of crosslinking agents.

(A) Schematic illustrating the formation of nanopores in lipid bilayers of LTSLs due to a phase transition from solid to liquid induced by mildly elevated temperatures. (B) Mass production of crosslinking agent (e.g., Ca²⁺)-loaded LTSLs through the extrusion process. (C) Dynamic light scattering (DLS) analysis of LTSLs before and after extrusion. (D) Fluorescent imaging of Ca²⁺-loaded LTSLs using fura-2-acetoxymethyl ester as an intracellular calcium indicator. Scale bar, 3 μm. (E) Stability study showing Ca²⁺ release from LTSLs stored at 4 °C and 25 °C after one month and six months. (F) Ultraviolet-visible analysis of LTSLs subjected to 43 °C for various durations. (G) Temperature-dependent Ca²⁺ release from LTSLs at 43 °C and 37 °C. (H) Crosslinking time for alginate US-inks under various heating temperatures when LTSLs concentration is fixed at 50 wt.%. (I) Ionic crosslinking of alginate US-inks with varying LTSL concentrations, evaluated by storage (Gʹ) and loss (Gʹʹ) modulus. Inset: images of the gelation status of alginate US-inks containing 0%, 15%, and 50% LTSLs after 30 s of exposure to 43 °C. Scale bars, 5 mm. (J) Crosslinking time measurements for alginate US-inks with different LTSL concentrations. (K) Live/Dead staining images of human dermal fibroblast cells cultured for 7 days with the alginate, alginate US-ink containing 50% LTSLs, and alginate US-gel. Scale bar, 100 μm. The error bars in the figures indicate the standard deviation from the mean (n=3).

LTSLs were synthesized using a dry lipid film hydration process followed by extrusion to achieve uniform size distribution (Fig. 2B, figs. S2 and S3). Dynamic light scattering (DLS) analysis confirmed the successful transition from multilamellar to unilamellar vesicles with smaller and uniform sizes (Fig. 2C) (29). A zeta potential of −17.31 mV indicated high dispersibility and stability (fig. S3D). Additionally, two-photon microscopy verified the successful encapsulation of CaCl₂ crosslinking agent within the LTSLs for alginate-based US-inks (Fig. 2D), while unencapsulated CaCl₂ and other residues were effectively removed through ultracentrifugation (fig. S4).

The LTSLs demonstrated long-term stability, maintaining consistent crosslinking performance even after 6-month storage (Fig. 2E). Ultraviolet-visible (UV-vis) spectroscopy revealed a substantial Ca²⁺ release peak at 650 nm after 30 s of heating to 43 °C (Fig. 2F). The LTSLs showed over 77% release at 43 °C with negligible release occurred at 37 °C, ensuring controlled, on-demand crosslinking for US-inks (Fig. 2G). Optimizing lipid composition, lysolipid content, and liposome size can enhance release profiles while balancing permeability and stability to prevent premature leakage, enabling effective crosslinking during sound printing.

Alginate US-ink was prepared by redistributing centrifuged LTSLs into an alginate solution. Energy dispersive X-ray spectroscopy (EDS) confirmed the presence of calcium in the alginate US-ink (fig. S5). Elevated temperatures accelerated calcium release and reduced gelation times at fixed LTSL concentrations (Fig. 2H). Rheological studies demonstrated rapid gelation at LTSL concentrations of above 32 wt.% (Fig. 2, I and J and fig. S6A). However, concentrations exceeding 50 wt.% led to undesirable self-association within 1 h at 37 °C (fig. S6, B and C). Therefore, 50 wt.% was selected as the optimal concentration for alginate US-inks, enabling on-demand in situ alginate crosslinking (fig. S6D) and addressing the limitations of conventional ion-diffusion-based crosslinking, which is impractical for in vivo applications.

Biocompatibility is critical for both prepolymer and printed structures in biomedical applications. Live/dead viability assays performed on human dermal fibroblasts confirmed high cell viability and normal morphology after 7 days of culture (Fig. 2K, fig. S7).

The US-ink design strategy is versatile, supporting not only ionic crosslinking but also free radical polymerization. As a proof of concept, PEGDA-based US-inks were developed using tetramethylethylenediamine (TEMED)-loaded LTSLs, synthesized and purified in a similar manner (fig. S8A). TEMED release from LTSLs occurred rapidly, within 30 s at 43 °C (fig. S8, B and C). These LTSLs were incorporated into a prepolymer mixture containing PEGDA monomers and ammonium persulfate initiator (fig. S9A). At 12 wt.% TEMED LTSL concentration, the shortest crosslinking times were achieved at 43 °C, with no gelation observed after 1 h at 37 °C (fig. S9, B to E). Importantly, FUS exposure during in vivo printing further reduced crosslinking times due to rapid localized heating, outperforming conventional rheological setups.

Focused ultrasound-induced high-resolution 3D printing

FUS waves precisely converge on a focal point (Fig. 3A) and effectively penetrate deep into tissues, reaching depths of several centimeters or more, far surpassing other energy delivery methods such as UV, visible, or NIR light, which are limited to a few millimeters due to absorption and scattering (Fig. 3B). In the DISP platform, FUS enables high-resolution in vivo printing with sub-millimeter focal zone.

Fig. 3. Characterization of focused ultrasound-induced 3D printing.

(A) Schematic of FUS wave propagation, illustrating precise targeting of US-ink. (B) Comparison of tissue penetration depths for ultrasound waves versus various light sources, highlighting ultrasound’s superior penetration. UVA, ultraviolet A; UVB, ultraviolet B; NIR, near-infrared. Note the inverse relationship between the ultrasound frequency with penetration depth. (C) Thermal simulations showing the temperature distribution at the focal point under different frequencies and exposure times. Scale bar, 2 mm. (D) Temperature profile at the focal point of FUS at 8.75 MHz during and after 10 s of ultrasound exposure. (E to H) Normalized pressure maps at the focal point using a 2.65 MHz transducer: experimental measurements using a hydrophone in the X-Z plane (E) and in the X-Y plane (F), and simulation results in the X-Z plane (G) and in the X-Y plane (H). (I) DISP-printed US-gel patterns. Scale bar for inset, 400 μm. Scale bars for patterns on the right, 4 mm. (J) Printability of the alginate US-ink with an 8.75 MHz transducer at various power levels and printing speeds. (K) Printing resolution in terms of line width for alginate US-ink using an 8.75 MHz transducer at different power levels and printing speeds. Scale bar, 5 mm. (L) Printing resolution of alginate US-ink, measured as line width, when printed under 15 mm thick pork loin tissue at 18 W, with varying frequencies and printing speeds. Inset, a deep tissue printed pattern on pork tissue. Scale bar, 5 mm. (M) Dissociation of alginate US-gels patterned on tissue using DISP, achieved by 5 min treatment with 0.025 M EDTA solution. The error bars in the figures indicate the standard deviation from the mean (n=3).

FUS exposure generates localized temperature increases at the focal point through acoustic energy absorption (31). Thermal characterization of the focal point reveals that the heating zone size depends on exposure time and frequency (Fig. 3C), with increased transducer power enhancing both pressure and temperature at the focal point (Fig. 3D, fig. S10, A and B) (32). Pressure zone measurements (Fig. 3, E and F) closely align with simulation results (Fig. 3, G and H and fig. S10, C and D), confirming that focal zone dimensions correlate with transducer frequencies. Simulations further indicate that enlarging transducer aperture reduces the focal zone size, thereby improving precision (fig. S11).

Within US-ink, localized heating and temperature gradients induced by ultrasound absorption trigger LTSL phase changes and initiate crosslinking (fig. S12). The temperature changes remained consistent across multiple FUS exposures (fig. S13). DISP-printed US-gel patterns were optimized by adjusting printing parameters, such as power and speed, at 8.75 MHz (Fig. 3, I–K, figs. S14–S16, and Movie S1), achieving resolutions as fine as 150 μm. Higher frequency FUS with a smaller focal zone and increased printing speed can further enhance resolution. Line width and height characterizations were performed for free-standing printing in a deeper 7 mm US-ink tank (fig. S17). Gradient-sized patterns were produced by varying printing power and speed during the process (fig. S14E).

DISP’s capabilities were further demonstrated by printing alginate US-gels beneath ~15 and ~40 mm-thick pork and chicken tissues at various resolutions (Fig. 3L and fig. S18). Additionally, the printed alginate US-gels can be selectively de-crosslinked and removed from tissue using chelate ethylenediaminetetraacetic acid (EDTA) (33) (Fig. 3M and fig. S19), providing flexibility for dynamic tissue applications.

Printing of functional biomaterials

The DISP technology offers a versatile platform for printing a wide range of functional biomaterials, unlocking applications in bioelectronics, drug delivery, tissue engineering, wound sealing, and beyond (tables S2 and S3). By enabling precise control over material properties and spatial resolution, DISP is ideal for creating functional structures and patterns directly within living tissues.

Using DISP, we successfully printed hydrogel bioelectronics using alginate US-inks containing conductive additives such as carbon nanotubes (CNTs), Ag nanowires (AgNWs), and MXene flakes (Fig. 4A). DISP’s nozzle-free approach avoids clogging issues commonly encountered in traditional printing methods when using inks with high additive concentrations. By optimizing printing parameters—such as speed, power, and CNT concentration—tailored hydrogel circuits with adjustable resistance and patterns were achieved (fig. S20). A calibration curve demonstrated an inverse relationship between resistance and line width, consistent with the theoretical predictions (Fig. S20C). Conductive US-inks with CNTs substantially improved both ohmic and ionic conductivity of alginate by nearly an order of magnitude through enhanced electron transfer within CNT networks (fig. S21). Despite increased viscosity due to CNT additives, the ink’s shear-thinning properties allowed for effective injection (fig. S22A). CNT-based inks also exhibited shorter gelation times compared to US-inks with other conductive additives (fig. S22, B to D) and slightly wider printed lines than their non-conductive counterparts due to higher acoustic absorption (fig. S22E). These conductive US-gels maintained stable conductivity during tensile and compressive loads, making them ideal for wearable and implantable bioelectronics (Fig. 4B and fig. S20, D to F). Additionally, a hybrid ink combining CNTs and AgNWs achieved a 10-fold conductivity increase. Scanning electron microscopy confirmed the structural integrity and entanglement of CNT fibers within the US-gels (fig. S23). Biocompatibility studies showed no apparent cytotoxicity for conductive US-ink and US-gels (fig. S24). DISP-printed hydrogel bioelectronics demonstrated promising applications for personalized health monitoring, including resistive skin temperature sensors with high sensitivity across physiologically relevant temperature ranges (Fig. 4C) and biopotential electrodes for reliable monitoring of physiological vital signs such as electrocardiogram (ECG) and electromyography (EMG) (Fig. 4D).

Fig. 4. Deep tissue in vivo sound printing-based 3D printing of functional biomaterials for various medical applications.

(A) Schematic of conductive US-ink composed of carbon nanotube (CNT) additives entangled within alginate US-ink, crosslinked using FUS. (B) Conductive US-gel patterns maintain stable electrical properties under cyclic bending deformations. (C) Temperature sensing using printed conductive US-gels. Inset, consistent and reversible temperature sensor response upon contact with human skin. RT, room temperature. (D) DISP-printed conductive US-gel sensors for electrocardiogram (ECG) and electromyography (EMG) recording in a human participant. (E) Integration of therapeutic biomolecules within US-ink, forming biocarrier US-gels for potential drug delivery applications. (F) Continuous and sustainable release of a model drug rhodamine B from US-gels. (G) Cell-encapsulated US-gels prepared by integrating cells within biocompatible US-inks followed by printing using an 8.75 MHz transducer at 7 W and a printing speed of 10 mm min⁻¹. (H) Live/Dead staining images of C2C12 mouse myoblast cells encapsulated within alginate US-gels on Day 1 and Day 3 post-printing. Scale bars, 100 μm. (I) Metabolic activity of cells assessed from Day 1 to Day 7 post-printing. Insets, images of the cell-laden US-gel pattern, printed with an 8.75 MHz transducer at 7 W and 10 mm min⁻¹, showing live cells 3 days post-print. Scale bar, 200 μm. (J) Catechol-modified gelatin-caffeic acid conjugates (GelCA) US-inks mixed with NaIO₄ liposomes for bioadhesive applications. (K) Adhesion strength of GelCA US-ink before and after crosslinking. Inset, images of the GelCA US-ink before and after mild heating and crosslinking. (L) Ex vivo adhesion testing of GelCA US-gel for sealing the punctured heart tissue. Scale bar, 5 mm. (M) In vivo US-induced adhesion, where FUS facilitates prepolymer jetting towards tissue, followed by in situ crosslinking of alginate US-ink to achieve mechanical interlock. (N) Alginate US-gels printed in vivo following intradermal injection of US-ink and crosslinking using a 2.65 MHz transducer at 7 W, with a printing speed of 20 mm min⁻¹ on live animals. The strong interfacial adhesion is observed between the alginate US-gel and tissue, with blue dye applied for visibility. Scale bars, 6 mm. The error bars in the figures indicate the standard deviation from the mean (n=3).

DISP-printed US-gels can also function as biocarriers, capable of encapsulating a wide range of small and large molecules, thus enabling targeted in vivo printing of therapeutics at the desired location for sustained and localized drug delivery (Fig. 4, E and F, figs. S25–S27). In vitro studies with 3D tumor microspheroids revealed a substantially higher cell death after exposure to doxorubicin (Dox)-loaded US-gels compared to the free drug administration. This enhanced therapeutic effect persisted for three days, even after multiple media changes, simulating more realistic in vivo conditions (fig. S28).

DISP enables minimally invasive tissue regeneration through in vivo printing of cell-laden hydrogels directly into affected areas. Patterns of cell-laden alginate US-gels were printed at low ultrasound amplitudes, ensuring safe temperature conditions for the cells (fig. S29). The printed cells exhibited high viability for seven days post-printing (Fig. 4, G to I and fig. S30A), with no substantial cytotoxic effect on morphology or function. Additionally, incorporating cells into the US-ink did not notably alter the prepolymer’s rheological properties (fig. S30B). To evaluate the stability of the Us-gel, its swelling and degradation in cell media was assessed (fig. S30C). This advance opens new possibilities for minimally invasive tissue regeneration deep within the body.

In vivo printing of bioadhesives offers a non-invasive approach to seal tissue ruptures and lacerations, enabling on-demand formation of bioadhesive patterns over wounds of various shapes. To achieve this, we developed a prepolymer solution of catechol-modified gelatin-caffeic acid conjugates (GelCA) (34), with sodium periodate (NaIO₄) encapsulated in the LTSLs to form GelCA US-ink. Upon FUS exposure, NaIO₄ was released, triggering crosslinking and forming a tissue adhesive hydrogel (Fig. 4J and fig. S31A). NaIO₄ was released from LTSLs within 30 s of mild heating at 43 °C (fig. S32), producing GelCA US-gel with increased bioadhesion strength (Fig. 4, K and L). The optimal NaIO₄ LTSL concentration of 20 wt.% ensured effective crosslinking without premature gelation after 1 h at 37°C (fig. S31).

Additionally, the typically non-adhesive alginate US-inks developed adhesive properties when printed using higher US amplitudes and extended exposure times, aligning with literature reports on US-induced adhesion (35). The ultrasound facilitates the integration of the prepolymer into the tissue, followed by in situ crosslinking, which enhances the mechanical interlocking of alginate US-gels within the tissue (Fig. 4, M and N).

In vivo evaluation of DISP printing in live animals

in vivo experiments were conducted using a mouse model (Fig. 5A and fig. S33). The shear force experienced during the in vivo injection of the prepolymer did not substantially impact the rheological properties of the US-ink (fig. S34A). Additionally, the alginate US-ink stored at 4 °C exhibited a long shelf-life of at least 450 days, reinforcing its suitability for DISP applications (fig. S34B). As a proof of concept, the DISP technique was used to print US-gels in the mouse bladder, with the goal of developing a potential platform for treating severe bladder cancer.

Fig. 5. Imaging-guided deep tissue sound printing in vivo.

(A) Setup for sound printing in vivo in live animals, employing ultrasound imaging for precise targeting. Inset: a linear pattern printed in vivo in a mouse. Scale bar, 4 mm. (B and C) Schematic of AM-mode ultrasound imaging with a GV contrast agent used to monitor US-ink distribution in vivo (B) and to ensure precise targeting (C). Ultrasound image inset in (C): A line of GV-integrated alginate US-ink printed and imaged in cross-section. GVs in areas not exposed to FUS remained intact, while those exposed to FUS collapsed. (D) In vivo printing of US-gels on a tumor site in the bladder of an anesthetized mouse. Successful targeting confirmed by GV collapse. After printing, the mouse bladder was extracted to verify successful printing. Scale bar, 4 mm. (E) In situ Ca²⁺ sensing using GV Ca²⁺ sensors integrated into alginate US-inks, designed to activate upon exposure to Ca²⁺. A line was printed and imaged in cross-section using AM-mode ultrasound imaging. Higher pressures in the center of the printed line led to partial collapse of GV Ca²⁺ sensors, while GV Ca²⁺ sensors at the boundary of the printed US-gel were activated, confirming the shape. (F) US-gel line printed using a 2.65MHz FUS at 11 Watt and 15 mm min⁻¹ on the abdominal muscle in a rabbit model. (G) The US-gel line printed deep into the adductor muscle and below the biceps femoris muscle using 2.65 MHz FUS at 20 Watt and 10 mm min⁻¹. (H) In vivo biocompatibility study of US-ink injected intradermally and ultrasound-printed US-gel in mice, assessed through hematoxylin and eosin (H&E) staining of skin tissues at one week and four weeks post-printing. Scale bars, 200 μm.

Ultrasound imaging plays a critical role in precisely positioning the prepolymer within deep tissue, which is vital for effective in vivo applications such as tissue regeneration and drug delivery. Accurate targeting, achieved through meticulous calibration of the FUS transducer relative to the imaging transducer, minimizes the risk of off-target effects and prevents unintended exposure of healthy tissues. Moreover, integrating stimuli-responsive GV contrast agents and GV Ca²⁺ sensors with US-ink provides real-time feedback during the printing process (26, 36). This integration confirms US-gel formation, monitors print shape, and tracks crosslinking dynamics. Compared to magnetic resonance imaging-guided techniques used for temperature evaluation at the focal point, GVs offers direct visualization of the printed US-gel (37, 38). Integrating stimuli-responsive GV contrast agent with prepolymer led to a high contrast allowed for easy visualization of the US-ink using amplitude-modulation (AM) mode ultrasound imaging (27). The inclusion of GVs not only facilitated accurate bioink delivery to the bladder but also enabled confirmation of gel formation upon FUS exposure, as the collapse of GVs indicated successful targeting (Fig. 5, B and C). This capability enables precise printing directly onto the diseased area of the organ, and the presence of GVs in the US-ink had a negligible effect on printing resolution (fig. S33E).

DISP was further explored for localized drug delivery, where drug-loaded US-gels were printed near the tumor site. Ultrasound imaging was employed to monitor the printing process while the animal was under anesthesia. The catheter’s position was monitored using B-mode ultrasound imaging to ensure successful instillation before injecting the US-ink into the bladder (fig. S35A). FUS was then applied to target the US-ink near the tumor, leading to the collapse of some GVs, confirming effective targeting. Clear US-gel was observed from the abdomen of the sacrificed animal, confirming successful in vivo printing (Fig. 5D and fig. S35).

To further assess the Ca²⁺ release zone and determine the size and shape of the printed US-gel in situ, GV Ca²⁺ sensors were incorporated into the US-ink (39). These sensors remained inactive during AM mode imaging but became activated upon FUS-induced Ca²⁺ release, producing noticeable contrast (Fig. 5E and fig. S36). At higher ultrasound amplitudes, some GV Ca²⁺ sensors collapsed at the center of the printed structure, where maximum pressure was applied, while the boundary of the printed US-gel remained clearly distinguishable.

To demonstrate in larger animal models, both ex vivo and in vivo experiments were conducted on rabbits (Fig 5F and G, figs. S37 and S38, and Movie S2). Successful in vivo printing was achieved on the exposed abdominal muscle and deep within the adductor muscle and beneath the biceps femoris muscle, highlighting the ability to target deeper tissue layers for applications like tissue replacement.

Practical in vivo printing faces challenges such as tissue heterogeneity and dynamic environments that affect ultrasound absorption and crosslinking. Minimal disruption from tissue movement was observed during experiments under anesthesia, but printing on dynamic organs like the lungs or heart remains challenging. GV Ca²⁺ sensors, coupled with AM ultrasound imaging, enable real-time monitoring and adjustment of printing parameters to address these challenges. Future advancements, such as machine learning-based adaptive transducer positioning, could further enhance precision in complex scenarios like cardiac printing.

In vivo biocompatibility analyses, conducted through hematoxylin and eosin (H&E) and immunofluorescent staining, confirmed biocompatibility of both the US-ink and resulting US-gel (Fig. 5H and fig. S39). H&E staining showed no tissue damage or abnormal immune cell infiltration, while immunofluorescent staining revealed minimal pro-inflammatory activity and macrophage infiltration. CD80 markers showed no significant increase in pro-inflammatory activity, and F4/80 markers indicated macrophage distribution consistent with normal physiological remodeling, without excessive or clustered immune responses. No signs of toxicity were observed, and tissues from other organs showed no notable changes (fig. S40). In the group that received only US-ink injection, the US-ink was completely cleared by the body in seven days. However, in the group exposed to FUS, the US-gel persisted (fig. S41).

Conclusions

Ultrasound-guided in vivo sound printing using crosslinking agent-loaded LTSLs enables precise, high-speed, high-resolution fabrication of functional biostructures deep within the body. A wide range of bioinks—including conductive, drug-loaded, cell-laden, and bioadhesives formulations—were designed utilizing diverse crosslinking chemistries. Real-time ultrasound imaging ensures precise targeting and controlled in situ crosslinking. Both in vitro and in vivo studies confirmed high biocompatibility of both prepolymers and printed hydrogels. As a proof of concept, in vivo printing was successfully demonstrated in the mouse bladder and rabbit leg muscles, showcasing its potential for targeted therapeutic interventions and tissue replacement.

Data and materials availability:

The customized finite element analysis code used for thermal simulations of the ultrasound printing process is available upon request. All other data are available in the main text or the supplementary materials.