Sperm cell empowerment: X-ray-guided magnetic fields for enhanced actuation and localization of cytocompatible biohybrid microrobots.

Magnetic iron oxide nanoparticles were synthesized using the precipitation method (Supplementary Fig. 2 ). 28 Briefly, 10 g FeCl 3 and 4.05 g FeCl 2 were dissolved in 550 mL of deionized water under magnetic stirring for 90 °C. Then, 30 mL of NH 4 OH solution (30 wt%) was gradually added (10 mL/min) to the stirring iron salt solution to anticipate the redox reaction. The reaction vessel was maintained at 90 °C under magnetic stirring for 1 h. The resulting precipitate was collected, followed by rinsing with deionized water and ethanol, and dried in the air. Vibrating sample magnetometer (VSM): magnetization measurements were obtained at room temperature using a vibrating sample magnetometer (VSM, 8600 Series Magnetometer) with a maximum magnetic field of 16 kOe and a step size of 1 kOe. A small amount of magnetic nanoparticle powder was mixed with Krazy glue and allowed to dry. Care was taken to avoid introducing any damage or defects to the samples during this process, ensuring that they were prepared appropriately for further analysis. Dynamic light scattering (DLS): an aqueous solution of iron oxide nanoparticles was prepared with a concentration of 0.1 wt%. To achieve homogeneity, the solution underwent sonication and vortexing to prevent the agglomeration of the nanoparticles and promote uniform dispersion. The solution was then transferred to the cuvette and loaded into the Malvern Zetasizer Nano-series. Zeta Potential: The zeta potential of the synthesized iron oxide nanoparticles was measured using a Malvern Zetasizer (Westborough, MA). The aqueous solution of the nanoparticles (0.1 wt%) was sonicated for 5 min before being transferred into the folded capillary cell (Zetasizer nanoseries, DTS 1061). Transmission electron microscopy (TEM): The morphology of the synthesized magnetic nanoparticles was investigated by transmission electron microscopy. After removing ethanol and drying the nanoparticles, the powder sample was dispersed in water (1 mg/mL), and about 10 μL of the dilute dispersion was poured onto the surface of a 300 mesh grid, which was then dried at an ambient temperature for 24 hs before taking the images on a Philips CM10 transmission electron microscope. The biohybrid microrobots were fabricated from bull sperm obtained from Semex Inc. (Guelph, Ontario, Canada) and iron oxide nanoparticles. The cryopreserved sperm were stored in 0.2 mL straws in liquid nitrogen until use. In a centrifugation tube, 2 mL of Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco) was warmed to 37 °C. Four straws of bull semen were placed in a water bath at 37 °C for 2 min to thaw. Afterwards, the 4 straws were opened and poured into the tube containing 2 mL DMEM. The tube was carefully inverted to mix the semen with the DMEM. The sperm suspension was washed following a washing protocol. The tube was placed in a centrifuge at 300 g-force, acceleration and deceleration of 5 for 5 min. The supernatant was removed from the tube and replaced with 2 mL of distilled water, carefully not disturbing the pellet. The pellet was mixed by pipetting and the washing procedure was repeated 3 times in total. Sperm cell concentration was measured and adjusted by counting sperm in a hemocytometer. Sperm concentration was adjusted to be 2.5 × 10 7  cells/mL. For all sperm-based biohybrid microrobot procedures, this sperm cell concentration remained consistent. Three concentrations of nanoparticles were formulated for the fabrication of sperm-based biohybrid microrobots. The nanoparticle solution was created by adding 5 mg of Fe 3 O 4 to 500 μL in a glass vial. The solution underwent vortex mixing for a duration of 5–10 min or until complete homogenization was achieved. Subsequently, 350 μL of sperm suspension was placed in four microcentrifugation tubes each. In the second tube, 50 μL nanoparticle solution was added; 100 μL of nanoparticle solution was added to the third tube; 150 μL of nanoparticle solution was added to the fourth tube. All tubes were labeled accordingly, filled with distilled water to a total volume of 350 μL (no nanoparticles were added to the first tube), and placed into a microtube rotator for overnight incubation. The size of the fabricated biohybrid clusters was on average 0.92 ± 0.63 mm across all concentrations, with a minimum size of 0.1 mm (for clusters with a few sperm cells) and a maximum size of 2.7 mm. The histogram in Fig. 5 A shows the distribution of cluster sizes depending on the nanoparticle concentration. The lower concentration of 1 mg/mL resulted in a lower average cluster size of 0.36 ± 0.18 mm, whereas 2 and 3 mg/mL resulted in larger cluster sizes, on average 1.23 ± 0.55 mm and 0.94 ± 0.65 mm, respectively. MRI data, obtained from The Cancer Imaging Archive, served as the reference dataset for model development. The outer muscular layer of the uterus, known as the perimetrium, was segmented using 3D Slicer. To accurately represent the uterine cavity, a 2D image was superimposed to capture its overall shape. Subsequently, mean measurements were researched, including parameters such as thickness, width, length, and circumference, to refine the digital phantom. The model’s optimization and hollowing were achieved using Meshmixer, ensuring its suitability for 3D printing on the Form3 resin printer (Formlabs, WatIMake lab, Waterloo) with Elastic 50A resin (Formlabs, Canada). Following the printing process, a series of post-processing steps was performed. These steps included a 20-min ethanol wash, the application of a thin layer of Elastic 50A resin to enhance transparency, ethanol flushing to remove any residual resin, a 20-min curing period, and the removal of support structures. To visualize the female reproductive tract model with the biohybrid magnetic microrobot clusters, a Siemens Healthineers Artis Pheno system (Siemens, Germany) was used. First, the reproductive tract model was filled with 0.9% saline and a biohybrid microrobot cluster. Subsequently, the microrobot cluster was manually guided toward its desired locations. Once in position, the model was imaged using a cone beam computed tomography (CBCT)-scan. Following the scanning process, the reproductive tract model was reconstructed with a specific window for the Hounsfield values, such that both the reproductive tract model and the biohybrid magnetic microrobot cluster are easily distinguishable (Fig. 1 A, B). The integration of the rotating permanent magnet longitudinally magnetized, as shown in Fig. 2 G-iii utilized a Maxon 18 V brushless DC motor with a gear ratio of 3.7:1. This motor was mounted onto a KUKA 6-DOF manipulator (KUKA KR-10 1100-2, KUKA, Augsburg, Germany) using a 3D printed mount. The operation of the robot was facilitated through the RoboDK program, enabling input of desired trajectories for the robotic arm. Trajectories were generated using a CBCT reconstruction of the phantom, with the centerline of the desired path imported into RoboDK. During traversal of the trajectories, the rotating permanent magnet maintained a pitch and roll angle of 0 degrees while rotating at a frequency of 1.5 Hz, which was likely due to reaching the step-out or optimum actuation frequency. Although robotic arm paths were automated, teleoperation was employed for the rotating permanent magnet to reverse motion when necessary. X-ray Fluoroscopy images were captured using an Artis Pheno C-arm system (Siemens, Germany) under a left oblique angle of 20°. The clinical radiation settings used were 45.5 kV for the peak voltage (kVp) and a tube current of 239 mA. The Fe 3 O 4 nanoparticle coverage was measured through the use of a brightfield microscope (Zeiss Axio Observer, Zeiss, Germany) and ImageJ software. Utilizing a micropipette, 10 μL of each individual sample was carefully dispensed onto four distinct glass slides (cleaned with 70% ethanol). The glass slides were placed under the brightfield microscope at 400× magnification, and at least 50 images of individual sperm-based biohybrid microrobots from each nanoparticle concentration were taken for thorough image analysis. In ImageJ, a global scale was established across all images using the scale function. Subsequently, the nanoparticles were delineated using the freehand tool, followed by the computation of their respective areas through the measurement tool. Fixation with 2.5% glutaraldehyde was performed on the 4 sperm-based biohybrid microrobot samples. Five microlitres of each of the 4 different sperm-based biohybrid microrobot samples were placed on clean silicon wafer pieces, followed by applying 20 μL of the glutaraldehyde solution. Following the 2 h fixation, the samples were dehydrated sequentially for 3 min in 50%, 70%, 80%, 95% and 100% ethanol. This process was repeated a total of 3 times for each ethanol concentration. After dehydration, the samples were air-dried for 1 h at room temperature. The silicon wafers were sputter-coated with gold and observed using a scanning electron microscope (Fig. 2 B). Initial cytotoxicity assays were performed on fibroblasts, and then on uterine carcinoma MES-SA cells (CRL-1976, ATCC, Gaithersburg) at 24 h and 72 h. Fibroblasts and MES-SA cells were cultured in T25 flasks to confluence in complete DMEM or McCoy medium, respectively (containing 10% fetal bovine serum and 1% penicillin) and released from the cell culture flask with Trypsin-EDTA (T4174, Millipore Sigma, USA). To assess cell viability and metabolic activity of the cells when in contact with sperm-based biohybrid microrobots, XTT assays (4891-025-K, RnD Systems, Minneapolis, USA) were performed based on the following protocol. Three replicates were performed for each assay. For each experimental run, the volumes of medium, XTT reagent, and activation solution were calculated as 8.68 mL, 4.34 mL, and 86.8 μL, respectively, following the manufacturer’s recommendations. The 24-well plate of cultured fibroblast cells was labeled accordingly. All four samples contained 400 μL of sperm-based biohybrid microrobots solution. A volume of 100 μL of sample 1 was distributed across three individual wells each. This was replicated for the remaining 3 samples and their respective wells. A volume of 100 μL of nanoparticles solution was poured into 3 empty wells each. The remaining 4 wells were left as controls. After filling the well plate, it was incubated at 37 °C for 24 and 72 h. Activated XTT solution (0.5 mL) was added to each tube well. The well plate was incubated for 2.5 h. Sample aliquots were transferred to a 96-well plate and measured in duplicate. Absorbance was measured at 490 nm with a reference wavelength of 630 nm on a SpectraMax Plus 384 Microplate Reader. Results were normalized to the control sample (no sperm cells/no nanoparticles). Results are presented in Fig. 4 D). Following the XTT assay, a Live/Dead staining was conducted on the fibroblasts and uterine cells to further evaluate cell viability. Calcein-AM (C1430, Thermofisher Scientific Chemicals, USA) and PI (P4170, Millipore Sigma, USA) were added to each well 1 μL of PI and 1 μL of Calcein-AM stock solution was added. The tray was incubated in the dark at 37 °C for 5 min. Brightfield and fluorescence images were taken with a Zeiss Axio Observer to record live (green) and dead (red) stained cells at 100× magnification (MES-SA cells displayed in Fig. 4 C, D, fibroblasts displayed in Supporting Fig. S5 ). Statistical analysis is summarized in Tables 1 and 2 in Supporting Information. Cell viability was calculated as a percentage by dividing the number of live cells by the sum of live and dead cells, and the results were presented in a bar chart (Fig. 4 D). A rotating magnetic field was generated using a disk magnet (Neodymium DX08B-N52) with dimensions of 25.4 mm in diameter and 12.7 mm in height, positioned 50 mm above each sample. The setup was powered by a 24 V DC motor (Maxon 2332.968-51.236-200) and supported using a robotic arm (Igus 4-DoF). The permanent magnet’s position remained fixed during each trial, with its rotational direction periodically reversed to record both clockwise and counterclockwise rolling speeds of the clusters. During testing, the clusters were placed in a glass tube that contained deionized water and had a diameter of 15 mm and a length of 120 mm. The clusters of magnetically coated biohybrid microrobots were fabricated using 1, 2, and 3 mg/mL concentrations of Fe 3 O 4 nanoparticles and bovine sperm. Tracking each concentration entailed measuring 10 clusters, ranging from 0.1 to 2.7 mm, across frequencies of 2–10 Hz. The lower limit of the frequency range was set by the motor’s capabilities, while the upper limit reflected the maximum rotational force the clusters could tolerate before disintegration. The cluster actuation was recorded using a Basler camera recording at 50 frames per second. The videos were analyzed using the Nano-micromotor Tracking Tool (NMTT) ( https://github.com/rafamestre/NMTT-nanomicromotor-tracking-tool ). To assess the CNR of biohybrid microrobots using ultrasound imaging, we utilized a Siemens Acuson 14L5 probe. The biohybrid microrobots were prompted to initiate motion, and ultrasound imaging was performed during this actuation. Afterward, we identified regions of interest (ROI) in various frames of the grayscale video generated, isolating specific areas within the rolling sample contained within a tube and distinguishing them from the background noise within the tube. The CNR was then calculated using the following equation: 1 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\rm{CNR}}=\frac{| {S}_{S}-{S}_{B}| }{{\sigma }_{B}},$$\end{document} CNR = ∣ S S − S B ∣ σ B , where S S represents the mean grayscale value of the sample’s ROI, S B represents the mean grayscale value of the background’s ROI, and σ B stands for the standard deviation of the background’s ROI. For the X-ray CNR assessment, fluoroscopy images were captured using a Siemens Healthineers Artis Pheno system, with clinical radiation settings of a tube current of 219 mA and a peak voltage of 43.1 kV. Similar to the ultrasound assessment, multiple frames were analyzed for each sample. The CNR was calculated by comparing the X-ray absorbance values of both ROIs. The data are presented in bar graphs and plots (Fig. 4 D–F) as mean ± standard error of the mean. The sample size ( n ) for each statistical analysis was n  = 80–100 clusters for each nanoparticle concentration for the frequency response data displayed in Fig. 4 E, n  = 50 for the particle coverage analysis (Table 1 ), and n  = 3 for the cytotoxicity assays on fibroblast cultures, each obtained from triplicate wells, and counting at least 100 cells per sample for the viability tests, respectively. Statistical analysis on cell viability and XTT assays was performed using OriginLab software with an ANOVA test with Bonferroni condition, which showed only small deviations for the XTT assay and live/dead stains (see Supporting Information Tables 1 and 2 ). Table 1 The influence of the ratio of nanoparticles (NP) per sperm cell area within biohybrid clusters is assessed using optical imaging ( n  = 50 images per concentration), motion analysis, noninvasive imaging, and cytotoxicity studies on uterine and fibroblast cells Structure Magnetic response CNR Cytotoxicity NP (mg/mL) Cov. [μm 2 ] Prop. [%] ω ft [Hz] U [mm/s] US X-ray XTT [%] Viability [%] 1 25.7 26.9 9.0 12.3 ± 1.1 3.41 ± 1.56 0.89 ± 0.29 100 ± 4 75 ± 4 2 66.3 69.4 10.0 11.5 ± 1.2 2.50 ± 0.59 0.53 ± 0.37 100 ± 3 75 ± 2 3 101.6 106.4 9.0 9.7 ± 1.1 4.22 ± 1.06 1.10 ± 0.46 89 ± 1 89 ± 4 The acronyms Cov and Prop indicate nanoparticle area coverage on sperm cells, and proportion of coverage related to the total sperm cell area as measured by phase contrast microscopy, respectively. The average maximum speed, U , is measured at the cluster’s fragmentation threshold frequency, ω ft , before they disintegrate. The influence of the ratio of nanoparticles (NP) per sperm cell area within biohybrid clusters is assessed using optical imaging ( n  = 50 images per concentration), motion analysis, noninvasive imaging, and cytotoxicity studies on uterine and fibroblast cells The acronyms Cov and Prop indicate nanoparticle area coverage on sperm cells, and proportion of coverage related to the total sperm cell area as measured by phase contrast microscopy, respectively. The average maximum speed, U , is measured at the cluster’s fragmentation threshold frequency, ω ft , before they disintegrate.

Here, we tackle two significant challenges encountered by biohybrid microrobots, directly influencing the localization and wireless actuation of sperm cells in vivo. First, for noninvasive localization using X-ray Fluoroscopy images or ultrasound images, the microrobots must attenuate radiation or reflect the ultrasound waves, respectively. We employ an electrostatic-based self-assembly process to coat sperm cells with nanoparticles (Fig. 2 B, C). The role of the nanoparticles for imaging purposes is twofold: to absorb radiation during X-ray-guided wireless manipulation and to reflect mechanical waves during preliminary imaging using pulse-echo techniques. These two imaging methods are implemented using clusters of biohybrid microrobots, as shown in Fig. 2 E, because the spatial resolution of modern imaging systems does not allow for the localization of a single cell (~70 μm-long and ~5 μm in width). Second, we demonstrate wireless X-ray-guided motion control inside a fluid-filled one-dimensional lumen (Fig. 2 G-i) and phantom of the reproductive tract (Fig. 2 G-ii) using controlled rotating magnetic fields (Fig. 2 G-iii). The advantage of this actuation method, in contrast to magnetic-based techniques such as MRI systems, lies in our ability to precisely project any desired magnetic torque. This enables the control of ferromagnetic torque-driven microrobots while effectively managing the magnetic force to assist in propulsion, as shown in Fig. 2 H. In these cases, both the single nanoparticle-coated sperm cell (see Supporting Information Fig. 1 ) and the cluster are torque-driven. A rotating magnetic field with a strength of about 7mT applied about the long axis of the sperm cell induces a traveling wave along the flagellum (Supporting Information Fig. 1 ). The same field also enables the cluster to rotate, providing rolling locomotion. Recent studies have explored the use of X-ray fluoroscopy imaging for localizing soft artificial medical microrobots, particularly in the context of vessel embolization 27 . This imaging technique becomes especially relevant when considering the wireless actuation of microrobots, such as coating a single sperm cell with superparamagnetic iron oxide nanoparticles, enabling their manipulation using a time-periodic magnetic field, as illustrated in Fig. 2 H. The iron oxide nanoparticles are characterized via transmission electron microscopy (TEM). The TEM image reveals the morphology of a cluster of iron oxide nanoparticles with a diameter of 121.8 ± 0.8 nm (Fig. 2 B). These clusters are composed of spherical nanoparticles, each measuring approximately 15 nm in diameter. The iron oxide nanoparticles exhibit superparamagnetic behavior with a small hysteresis loop and coercivity with magnetization of 0.06 emu per gram of dry weight, demonstrating their robust magnetic properties, as shown in Fig. 2 F. Additionally, the zeta potential indicates the average surface charge of the synthesized nanoparticles as −38.1 ± 4.3 mV (Fig. 2 D). A cluster of nanoparticle-coated cells provides a different locomotion mechanisms compared to single cells (Supporting Information Fig. 1 ), as it depends on a nearby solid boundary for rolling, as shown in Figs. 2 H and 3 A. A cluster of cells with magnetic particles is more practical for medical imaging and can also transport a larger cargo load 9 . As the X-ray beam passes through the entangled cells, its intensity diminishes. This attenuation occurs as the X-rays travel from the source to the detector array, enabling us to visualize the cluster, its surroundings, and its actuating permanent magnet (see Methods). Figure 3 A shows the rolling motion of the cluster along the y -axis and three representative X-ray Fluoroscopy images captured at different time points with a Fluoroscopy dose rate of 30.8 mGy cm 2  s −1 , which is within the range typically used in clinical settings for comparison. In clinical settings, X-ray fluoroscopy dose rates typically range from 10 to 100 mGy cm 2  s −1 , depending on the procedure and imaging requirements. Therefore, the dose rate of 30.8 mGy cm 2  s −1 used in our study falls within this clinical range. In this experiment, the rotating permanent magnet is controlled robotically to translate above the fluid-filled lumen at a translational speed comparable to that of the rolling cluster. As the permanent magnet translates and rotates to keep the cluster coupled with the rotating magnetic field, we measure the contrast-to-noise ratio (CNR) of clusters with nanoparticle concentrations of 1–3 mg/mL (see Methods). Each motion control trial involves the cluster rolling forward using clockwise rotation of the permanent magnet about the x -axis and then rolling backward to the starting point using counter-clockwise rotation. Fig. 3 Biohybrid microrobots (nanoparticles-coated sperm cells) are actuated using a rotating magnetic field and localized using X-ray fluoroscopy images. A Two clusters are actuated and localized simultaneously using a rotating magnetic field about the x -axis, yielding rolling along the y -axis. B – D Clusters with nanoparticle concentrations ranging from 1 to 3 mg/mL are detectable using X-ray fluoroscopy images (Supplementary Movies 2–4 ). A Two clusters are actuated and localized simultaneously using a rotating magnetic field about the x -axis, yielding rolling along the y -axis. B – D Clusters with nanoparticle concentrations ranging from 1 to 3 mg/mL are detectable using X-ray fluoroscopy images (Supplementary Movies 2–4 ). Note that the source–detector axis of the C-arm is rotated by 20° with respect to the z -axis in the frame of reference depicted in Fig. 3 A. This rotation enables unobstructed motion of the actuator magnet above the fluid-filled lumen during X-ray-guided actuation of the sperm cell clusters. The forward and backward rolling motion of a cluster with a nanoparticle concentration of 1 mg/mL is illustrated in Fig. 3 B. This trajectory forms an isosceles triangle when the contrast is measured over time. The contrast shifts with time as the cluster rolls along the + y -axis. When the direction of rotation of the actuator magnet is reversed, the slope of the line becomes negative, indicating motion reversal along the − y -axis. A similar response is observed for clusters with nanoparticle concentrations of 2 and 3 mg/mL, as shown in Fig. 3 C, D, respectively. This result demonstrates that the cluster’s structure is adequately dense to absorb sufficient radiation at any of the 3 tested concentrations, facilitating effective localization. The response of biohybrid microrobots, functionalized with varying nanoparticle concentrations, to rotating magnetic actuation is analyzed. These concentrations range from low (1 mg/mL) to medium (2 mg/mL) to high (3 mg/mL) magnetic content (see Fig. 4 A, B). As the concentration of magnetic nanoparticles increases, the size of the resulting clusters also increases. At 1 mg/mL nanoparticle content, the average cluster size measures approximately 0.36 ± 0.18 mm. However, at higher concentrations of 2 mg/mL and 3 mg/mL, the cluster sizes increase, averaging about 1.23 ± 0.55 mm and 0.94 ± 0.65 mm, respectively. These measurements, drawn from more than 200 clusters, are depicted in Fig. 5 . Figure 5 B displays no significant relation between cluster size and resulting forward velocity, although there is a slight trend of larger clusters moving faster, but no direct correlation to the magnetic content. Fig. 4 The influence of the nanoparticle concentration on magnetic actuation and cytotoxicity is examined. A Scanning electron microscopy (SEM) image depicting ascending concentrations of nanoparticles on bull sperm. (i) Bull sperm cells. (ii) cells covered with 1 mg/mL of nanoparticles. (iii) Covered with 2 mg/mL of nanoparticles. (iv) Covered with 3 mg/mL of nanoparticles. B Phase contrast microscopic images depicting increasing concentrations of nanoparticles adhering to bull sperm. (i) Bull sperm cells. (ii) Cells covered with 1 mg/mL of nanoparticles. (iii) Covered with 2 mg/mL of nanoparticles. (iv) Covered with 3 mg/mL of nanoparticles. C Representative bright field images of MES-SA uterine epithelial cells incubated for 24 h with increasing concentrations of nanoparticles on bull sperm (from left to right). D Fluorescent overlayed images displaying Calcein-AM stained live cells(green) and Propidium Iodide stained dead MES-SA uterine cells (red). E Cytotoxicity studies performed on uterine epithelial MES-SA cells across the various samples, measured by XTT assay with average relative corrected absorbance normalized to the control (red), and uterine epithelial cell Live/Dead viability staining (blue) evaluated after 72 h incubation. Data were obtained from 3 independent experiments with 3 replicate wells each, and were counted at least 10 3 cells for each condition in the viability assay. Error bars are the standard error of the mean. Statistical annotations: n.s. no significant difference to control, * indicates p -value < 0.05, ** indicates p -value < 0.01, *** indicates p -value < 0.001. Fig. 5 Sperm-nanoparticle cluster sizes and resulting forward rolling velocity. A Histogram of the size distribution of sperm-nanoparticle clusters depending on nanoparticle concentration. B Cluster translational velocity depending on cluster size. A Scanning electron microscopy (SEM) image depicting ascending concentrations of nanoparticles on bull sperm. (i) Bull sperm cells. (ii) cells covered with 1 mg/mL of nanoparticles. (iii) Covered with 2 mg/mL of nanoparticles. (iv) Covered with 3 mg/mL of nanoparticles. B Phase contrast microscopic images depicting increasing concentrations of nanoparticles adhering to bull sperm. (i) Bull sperm cells. (ii) Cells covered with 1 mg/mL of nanoparticles. (iii) Covered with 2 mg/mL of nanoparticles. (iv) Covered with 3 mg/mL of nanoparticles. C Representative bright field images of MES-SA uterine epithelial cells incubated for 24 h with increasing concentrations of nanoparticles on bull sperm (from left to right). D Fluorescent overlayed images displaying Calcein-AM stained live cells(green) and Propidium Iodide stained dead MES-SA uterine cells (red). E Cytotoxicity studies performed on uterine epithelial MES-SA cells across the various samples, measured by XTT assay with average relative corrected absorbance normalized to the control (red), and uterine epithelial cell Live/Dead viability staining (blue) evaluated after 72 h incubation. Data were obtained from 3 independent experiments with 3 replicate wells each, and were counted at least 10 3 cells for each condition in the viability assay. Error bars are the standard error of the mean. Statistical annotations: n.s. no significant difference to control, * indicates p -value < 0.05, ** indicates p -value < 0.01, *** indicates p -value < 0.001. A Histogram of the size distribution of sperm-nanoparticle clusters depending on nanoparticle concentration. B Cluster translational velocity depending on cluster size. The frequency response is illustrated in Fig. 6 A, demonstrating an enhanced speed of biohybrid microrobots as the frequency increases from 2 to 10 Hz, to an average speed of 8–12 mm/s across the different magnetic particle content of the microrobots. Higher frequency data points were excluded from the analysis due to the breakdown of clusters into smaller pieces, attributed to the magneto-elasto-hydrodynamic balance between the cluster and the surrounding fluid. An example of this disassembly effect is demonstrated in Supplementary Movie 5 . Below the fragmentation threshold frequency, ω ft , increasing the actuation frequency would elevate the angular velocity of the cluster, consequently intensifying the drag torque on the cluster for the same magnetic torque, ultimately leading to the breakdown of the cluster. As the clusters perform a rolling motion over a surface, their size directly correlates with the resulting forward motion in response to the rotating field. This means that larger clusters tend to exhibit larger translational motion over the same time period, as depicted in Fig. 5 B. Overall, there appears to be no significant correlation between the magnetic nanoparticle concentration and the resulting speed, indicating that actuation performance is not significantly improved with increasing magnetic content. Clusters at varying particle content levels all respond to the magnetic rotating field, indicating a consistent behavior across different concentrations. This can be attributed to the motion mechanism, which relies on aligning the magnetic microrobot axes along the rotating field, rather than utilizing field-gradient pulling, to which the microrobots’ response will be more influenced by the volume of the magnetic nanoparticle. This, in turn, results in the translational motion of the clusters, which is observed as a rolling motion along the contacting surface. The frequency response also shows that the individual cluster behavior depends on the density of sperm cells and particles, as well as the specific three-dimensional arrangement and shape of the cluster. Consequently, a wide variety of frequency response results is evident (shaded regions in Fig. 6 A). Fig. 6 Analysis of magnetic motion and detectability of sperm-nanoparticle clusters. A The rolling speed of the cluster increases up until reaching the fragmentation threshold frequency, ω ft . Beyond this frequency, the cluster disintegrates into smaller fragments. Error bars display the standard error of the mean. The sample size ( n ) for each statistical analysis was n  = 80–100 clusters for each nanoparticle concentration for the frequency response. B Contrast-to-noise ratio is quantified through ultrasound and X-ray fluoroscopy images, showing no significant effect of nanoparticle concentration. See also Supporting Videos 2–4, 6–9 and 11. A The rolling speed of the cluster increases up until reaching the fragmentation threshold frequency, ω ft . Beyond this frequency, the cluster disintegrates into smaller fragments. Error bars display the standard error of the mean. The sample size ( n ) for each statistical analysis was n  = 80–100 clusters for each nanoparticle concentration for the frequency response. B Contrast-to-noise ratio is quantified through ultrasound and X-ray fluoroscopy images, showing no significant effect of nanoparticle concentration. See also Supporting Videos 2–4, 6–9 and 11. The detectability of sperm-based biohybrid microrobot clusters in both X-ray fluoroscopy and ultrasound imaging is examined across varying levels of magnetic content. The results indicate that magnetic nanoparticle concentrations ranging from 1 to 3 mg/mL provide a sufficient signal for ultrasound and X-ray fluoroscopy imaging (see Fig. 6 B and Supporting Videos). Within the tested concentration range, there is no correlation between the particle concentration and the contrast-to-noise ratio (CNR). However, ultrasound images generally exhibit a higher CNR compared to X-ray fluoroscopy images (see cluster moving in Supporting Video 11). Despite the low CNR in X-ray fluoroscopy images, it remains sufficient for operators to control the biohybrid microrobots within the one-dimensional fluid-filled lumen (Figs. 7 and 8 ) and the female reproductive tract phantom. While higher CNR values would offer clearer imaging, the current CNR level provides enough contrast for operators to visualize and track the clusters effectively, as shown in Figs. 7 and 8 . Operators rely on real-time feedback from X-ray fluoroscopy images to guide the clusters’ movement. This visual feedback allows operators to make precise adjustments and corrective actions as needed, ensuring accurate navigation and control of the microrobots in the phantom. Therefore, although improvements in CNR would enhance imaging quality, the existing level still enables operators to maintain control and execute necessary maneuvers during the experimental trials. Fig. 7 Trajectory evolution over time visualized through camera footage (left) and X-ray fluoroscopy imaging (right). Ellipses illustrate the positions, within specific time frames, of all components linked to the primary cluster when separated. A The top images depict trajectories within the right fallopian tube. Three representative X-ray fluoroscopy images demonstrate the detectability of the cluster at three different time points: in the cervix, and at the proximal and distal ends of the right fallopian tube. B The bottom images represent the left fallopian tube. Three representative X-ray fluoroscopy images demonstrate the detectability of the cluster at three different time points: in the cervix, and at the proximal and distal ends of the left fallopian tube. (Supplementary Movies 6 and 7 ). Fig. 8 Biohybrid microrobots are precisely navigated within the female reproductive tract phantom using X-ray-guided magnetic fields. A Iterative motion control trials demonstrate nanoparticle-coated sperm cells starting from the cervix and rolling toward the right fallopian tube. Five representative X-ray fluoroscopy images show the detectability of the cluster at five different time points: the proximal and distal ends of the cervix, halfway along the right fallopian tube, and at its distal end. B To move sperm cells from the cervix toward the left fallopian tube, modifications are made to the magnetic fields, enabling rolling motion along their path. Five representative X-ray fluoroscopy images show the detectability of the cluster at five different time points: the proximal and distal ends of the cervix, halfway along the left fallopian tube, and at its distal end. (Supplementary Movies 8 and 9 ). The images show iterative trials that indicate the reproducibility of the navigation. Ellipses illustrate the positions, within specific time frames, of all components linked to the primary cluster when separated. A The top images depict trajectories within the right fallopian tube. Three representative X-ray fluoroscopy images demonstrate the detectability of the cluster at three different time points: in the cervix, and at the proximal and distal ends of the right fallopian tube. B The bottom images represent the left fallopian tube. Three representative X-ray fluoroscopy images demonstrate the detectability of the cluster at three different time points: in the cervix, and at the proximal and distal ends of the left fallopian tube. (Supplementary Movies 6 and 7 ). A Iterative motion control trials demonstrate nanoparticle-coated sperm cells starting from the cervix and rolling toward the right fallopian tube. Five representative X-ray fluoroscopy images show the detectability of the cluster at five different time points: the proximal and distal ends of the cervix, halfway along the right fallopian tube, and at its distal end. B To move sperm cells from the cervix toward the left fallopian tube, modifications are made to the magnetic fields, enabling rolling motion along their path. Five representative X-ray fluoroscopy images show the detectability of the cluster at five different time points: the proximal and distal ends of the cervix, halfway along the left fallopian tube, and at its distal end. (Supplementary Movies 8 and 9 ). The images show iterative trials that indicate the reproducibility of the navigation. The clarity difference between CBCT scans and X-ray fluoroscopy images is primarily due to the imaging techniques’ inherent differences. CBCT scans provide a clearer image because they capture volumetric data from multiple angles, allowing for reconstruction of a 3D image with higher resolution, as shown in Fig.1. On the other hand, X-ray fluoroscopy images provide real-time imaging but with lower resolution and contrast. While CNR may vary between the two techniques, CBCT typically yields higher CNR due to its superior spatial resolution and image quality compared to X-ray fluoroscopy images. However, the latter is more suitable for use during motion control through direct teleoperation. Finally, the cytotoxicity of biohybrid microrobot clusters with varying magnetic content is investigated by subjecting them to a 72 h exposure to human uterine epithelial cells. Two types of toxicity assays were conducted: cell viability assessed through Live/Dead staining (Fig. 4 D, E) and XTT assays, which analyze the metabolic activity of the uterine cells and are a good measure for overall cell health (see Fig. 4 E). Both toxicity studies confirm that the magnetic content has little impact on uterine epithelial cell viability across the various particle concentrations. Exposure to sperm-based biohybrid microrobots at concentrations of 1, 2, and 3 mg/mL resulted in no detectable toxicity as confirmed by XTT assays. Live/Dead staining of uterine cells revealed little toxicity after 72 h, with only 25%, 25%, and 11% toxicity observed at concentrations of 1, 2, and 3 mg/mL of magnetic nanoparticle, respectively. For statistical analysis, see Supporting Information Table 2 . The sample containing 10 mg/mL pure magnetic nanoparticles demonstrated no increased toxicity compared to the control, as evidenced by the two cytotoxicity assays (Fig. 4 E). This demonstrates the good cytocompatibility of the superparamagnetic particles, even at elevated concentration. Interestingly, pure sperm samples resulted in a reduction in uterine cell viability, with a decrease of 13% detected in the live/dead staining. This phenomenon could be attributed to residues of cryopreservant or pathogens present in the semen, which may not be entirely removed during sperm preparation. This is most likely also the reason for decreased viability in the 1 and 2 mg/mL samples at the 72 h mark in the live/dead cell staining. It is also noteworthy that the cytotoxicity of the sperm-based magnetic microrobots (74–88% viability and 89–100% in XTT assays) is comparable to that of pure sperm samples (87% and 99%, respectively), suggesting that significant toxicity does not originate from the magnetic nanoparticles. Biohybrid microrobots traverse the 3D printed female reproductive tract model using controlled rotating magnetic fields, initiated from the cervix and directed through the uterine cavity toward the right or left fallopian tube (Fig. 7 and Fig. 8 ). In this experiment, biohybrid microrobots, with a nanoparticle concentration of 3 mg/mL magnetizing the sperm cells, were navigated through the 3D printed organ model while the reproductive tract phantom was filled with a 0.9% saline solution. Fig. 7 A illustrates the rolling behavior from the cervix toward the distal end of the right fallopian tube, using the centerline of the pathway to determine the magnetic field rotation axis as the cluster rolls inside the phantom. The motion of the cluster is simultaneously tracked using camera footage and X-ray Fluoroscopy images (Supplementary Movies 6 and 7 ). At an actuation frequency of 1.5 Hz, the cluster rolls from the cervix toward the bifurcation points of the fallopian tubes under an inclination. At the bifurcation point ( t  = 15 s in Fig. 7 A and t  = 13 s in Fig. 8 B), the rotating permanent magnet is robotically controlled to rotate the field rotation axis along the centerline of the right fallopian tube, guiding the microrobots along this path while rolling. As the biohybrid microrobots roll past the bifurcation point ( t  > 15 s), the configuration becomes more planar, and the average rolling speed increases from 3 mm/s to 6 mm/s. Finally, the field gradient pulling is reduced, enabling controlled vertical rolling along the negative z -axis. The rotating magnetic field still exerts torque on the cluster during this maneuver, enabling the microrobots to reach the distal end of the right fallopian tube in 44 s (see Fig. 8 A). Similar to the motion control in the right fallopian tube, the biohybrid microrobots are precisely guided from the cervix toward the left fallopian tube, as depicted in Fig. 8 B, completing the journey in 37 s at an average rolling speed of 3.5 mm/s up until the descending point of the fallopian tube. Notably, as the cluster rolls toward the left or right fallopian tubes, it breaks into smaller clusters. Supplementary Video S5 shows the clusters moving at high frequency and starting to separate into smaller clusters. Its fragmentation into smaller clusters can be attributed to several factors. Firstly, the structural integrity of the cluster, reliant on the entanglement between the nanoparticle-coated sperm cells, can be compromised due to mechanical stresses encountered during movement. Additionally, interactions with the surrounding fluid environment, such as shear forces and hydrodynamic interactions, exert forces on the cluster, potentially leading to its fragmentation. In addition, variations in surface interactions as the cluster navigates different regions of the reproductive tract may also contribute to its destabilization and fragmentation. Despite these challenges, the rotating magnetic fields maintain cohesion among the smaller clusters, facilitating their synchronized movement toward the target site. While the fragmentation of the cluster might impact its detectability in X-ray fluoroscopy images, the achieved contrast allows for reliable automatic detection throughout the entire trajectory. Motion control within the reproductive tract model was repeated for a total of five successful trials in each fallopian tube (see Fig. 7 ), demonstrating the high controllability of these microrobots. Localization using X-ray fluoroscopy imaging showed difficulty at the distal end due to reduced contrast, with most trials showing no trace of cluster locomotion (Fig. 8 ). As a result, the run times in the processed X-ray data did not correspond accurately with the actual trajectory times due to contrast discrepancies between trials. Nevertheless, the clusters were able to reach the distal end in each trial in less than 50 s (Supplementary Movies 8 and 9 ). Supplementary Video 10 shows the actuation along the z -axis inside the reproductive tract model, demonstrating the cluster’s ability to overcome gravity and to move against it.

This study marks a significant milestone in the development and evaluation of biohybrid microrobots customized for applications within the female reproductive tract. Our method of coating sperm cells with magnetic nanoparticles has not only led to improved mobility, as evidenced by increased translational speed with higher actuation frequency, but it has also unveiled a transformative capability: the potential for real-time visualization of sperm cells using X-ray fluoroscopy images in organ models. While examining the detectability of these microrobot clusters in medical imaging, we observed that higher nanoparticle concentrations slightly enhance detection in both ultrasound and X-ray fluoroscopy images. However, all tested concentrations ranging from 1 to 3 mg/mL remain detectable. Additionally, our cytotoxicity evaluations suggest that the inclusion of magnetic nanoparticles does not cause significant harm to uterine or fibroblast cells. This finding, especially concerning X-ray fluoroscopy imaging, highlights the potential of these biohybrid microrobots for precise medical interventions, including drug delivery and diagnostic applications in the female reproductive tract. Possible target diseases include cervical, uterine, fallopian tube or ovarian cancer, but also endometriosis and uterine fibroids, for which currently no targeted and effective drug administration strategies exist. Our previous study on drug loading of sperm clusters demonstrated a simple loading strategy with a capacity of loading up to 500 pg doxorubicin-hydrochloride, a model anti-cancer drug, into a cluster containing 120 sperm cells and an overall cluster size of 0.1 mm 2 9 . The cargo load is directly correlated to the cluster size and can be tuned by the number of cells in the cluster. The drug loading amount is specific to each pathology and treatment, and it would need to be optimized, together with the release kinetics, for the best therapeutic effect. It is subject to future studies. With the capability to visualize sperm cells using X-ray fluoroscopy imaging in a life-sized human anatomical organ model, this research lays the groundwork for potentially new diagnostic tools to study sperm migration and reasons for infertility, e.g., tubal obstructions, contributing to reproductive science and medicine. Investigating potential triggering methods for cluster disaggregation will further enhance the versatility and efficacy of these biohybrid microrobots, ensuring their applicability in a wide range of clinical scenarios. Additionally, the navigation in high viscosity reproductive tract fluids and in tight spaces with folds and channels inside a soft tissue model should be the subject of future research. Possible robot obstruction or adhesion to the surrounding tissues will need to be studied in ex vivo or in vivo experiments. While the current focus is on optimizing the design and functionality of the biohybrid microrobot as a whole, exploring triggering methods opens up new avenues for enhancing the versatility and adaptability of these microrobots. Such methods could enable selective activation of individual components within the cluster, allowing for targeted functionality in specific environments or applications. Additionally, triggering methods may facilitate controlled release of payloads carried by the microrobots, further expanding their utility in drug delivery and therapeutic interventions. Overall, investigating triggering methods represents an important step toward unlocking the full potential of biohybrid microrobots for a wide range of biomedical and nanomedicine applications. Further, the actuation of the sperm clusters in fluids mimicking uterine fluid rheological properties, as well as actuation against fluid flow should be investigated in future studies. In summary, our study revealed that increasing the concentration of iron oxide nanoparticles led to improved coverage and clustering, resulting in enhanced actuation and imaging capabilities, all while maintaining uterine cell viability within acceptable limits. These results are highly promising and represent a pivotal step toward advancing sperm cell-based biohybrid microrobots for clinical applications, particularly in the realm of drug delivery within the reproductive tract. Potential target locations include precise, targeted drug delivery to the uterine wall or the fallopian tubes (Fig. 1 ), enabling the treatment of conditions such as endometriosis, hormonal imbalances, or cancer. This research demonstrates a significant contribution to the field of reproductive medicine, offering novel possibilities for minimally invasive and highly targeted therapeutic interventions. We anticipate future advancements in the development of live magnetic sperm clusters, which will unlock additional applications for in vivo diagnostics of sperm migration, targeted delivery, and assisted fertilization.

The lack of real-time imaging of sperm within the human reproductive tract has hindered our understanding of reproductive health and the advancement of diagnostics and treatments. Many fundamental processes, such as sperm migration, storage and interactions with the different parts of the female reproductive tract, are still poorly understood and contribute to a high rate of unexplained infertility, lack of diagnostic tools and treatment options 1 , 2 . The ability to image spermatozoa inside the female reproductive tract would contribute to advances in the diagnostics of reproductive disorders and fundamental studies of gamete transport and fertilization. It would also enhance assisted reproductive technologies and drug delivery approaches to target cancer and other diseases of the reproductive tract. Current methods for sperm imaging in vivo include optical approaches such as fluorescent or ink-based imaging established with the help of fluorescent reporter mice and two-photon microscopy 3 . Through the help of fibered confocal microscopy, insights can be given into in situ sperm motility in the genital tract, but it requires surgical intervention and anesthesia of the animal 4 . Optical coherence tomography is another approach that has been explored to provide label-free in vivo imaging of sperm cells, but is limited to a few millimeter imaging depth 5 . However, none of these methods provides a label-free, noninvasive method to track sperm motion in vivo. In recent years, soft biohybrid microrobots have emerged as a promising technology with significant potential in various biomedical applications, including drug delivery 6 – 11 , cell manipulation 12 and assisted fertilization 13 – 15 . These microrobots represent a fusion of biological cells and artificial components, capitalizing on the advantages of naturally evolved organisms in conjunction with cutting-edge technologies to address a wide range of biomedical challenges. Notable characteristics of these microrobots include their capability for wireless actuation in response to external stimuli 8 , 13 , 16 , cargo loading ability 6 , 9 , and their nontoxicity to surrounding cells 9 , 13 . Given their potential for in vivo applications, it is crucial that these microrobots can be both remotely controlled and detected using various imaging modalities. A variety of imaging techniques have been utilized for microrobot localization, including optical 16 , magnetic 17 , computed tomography 18 , photoacoustic 19 , electrical impedance tomography 20 , and pulse-echo 21 techniques, each presenting unique advantages and limitations 22 . Among these methods, magnetic-based techniques and computed tomography are advantageous because the former can be used for simultaneous actuation and localization for ferromagnetic field-driven microrobots, while the latter provides an unobstructed workspace between its emitter and detector, allowing for the incorporation of a wireless manipulation system 23 . So far, there has not been any demonstration of simultaneous wireless actuation and localization of torque-driven biohybrid microrobots using an imaging system scalable to the size of in vivo applications, for two key reasons. First, this limitation arises from the fact that magnetic-based localization methods, such as magnetic resonance imaging (MRI), solely permit control over the magnetic field gradient. This constraint limits the use of this imaging system for actuating and localizing ferromagnetic torque-driven microrobots 24 . Hence, it is crucial to generate the actuating magnetic field without interfering with the imaging signal. Second, the amount of artificial components that can be incorporated into an organism is fundamentally limited, either by its size or its surface charge. This, in turn, restricts the acoustic impedance and the absorption of radiation by the radiolucent component of the biohybrid microrobot, as observed in pulse-echo and computed tomography techniques, respectively. The ratio of inorganic to organic matter within the biohybrid microrobot is likely to enhance its magnetic response and detectability while potentially compromising its cytocompatibility. To ensure cytocompatibility, magnetic actuation, and imaging capability, it is essential to synergistically fuse biological cells, which can be either alive or non-living, and artificial components, yielding a distinct response unattainable through natural or synthetic means alone 25 . In a recent approach, sperm-templated biohybrid microrobots were fabricated by electrostatic self-assembly of bovine spermatozoa and magnetic nanoparticles, resulting in flexible swimmers that can be actuated by rotating magnetic fields and successfully loaded with cancer drugs for cargo delivery inside the female reproductive tract 8 . This approach offers a simple fabrication method to obtain flexible magnetic micro-scale swimmers with loading ability. In a follow-up study, the impact of segmented magnetization on the flagellar propulsion of sperm-templated microrobots was investigated, elucidating that the location of the magnetic component governs the resulting magnetic swimming performance of the flexible sperm-based microrobots 26 . A previous study looked at the interactions of a range of particle sizes and surface charges and their effect on sperm cell binding 12 . This study emphasized that the sperm cell has a distinct morphology with different membrane charges across its surface, which also change over the lifetime of sperm due to reorganization of membrane components (proteins, glycoproteins, etc.). Due to these dynamic changes, the study pointed out that differently charged particles tend to bind to different areas of the cell and that this technique can be used for charge mapping of sperm cells, which is linked to sperm quality 12 . In the next step, the approach of using sperm-nanoparticle clusters rather than single cells, or single biohybrid robots, was motivated by the demand to increase the cargo load and imaging contrast by clinical imaging modalities. The first sperm cluster study demonstrated that the drug cargo load correlates directly with the number of sperm cells, as the cells serve as drug carriers, and that clusters can move faster than single cells in a rolling motion 9 . Further, the clusters on a large micrometer scale can be imaged in medical ultrasound, which cannot be achieved for single sperm cells with magnetic nanoparticles. This motivated us to investigate further the effect of magnetic particle content of the clusters on contrast-to-noise ratio in medical imaging, actuation performance and cytotoxicity. Considering the potential of using a cluster of sperm cells for cargo delivery inside the female reproductive tract, precise motion control and simultaneous imaging are needed. Non-living cells can serve as natural loading units due to their cell membrane surrounding the cytosol 9 , 26 . Additionally, sperm-templated fabrication stands out for its simplicity in fabrication, offering loading capacity without the need for further functionalization or microfabrication processes, and taking advantage of the entanglement of sperm cells with each other to form flexible and magnetic clusters. In vivo localization of these cells can only be achieved by incorporating nanoparticles that provide a detectable signal. The primary groundbreaking contribution of this study lies in the capacity to image sperm cells using X-ray technology in a female reproductive tract anatomical model. We demonstrate that sperm particle clusters can be formed with satisfying uterine cell viability, magnetic response and detectability in X-ray fluoroscopy. Simultaneous actuation and imaging with a C-arm can be achieved, demonstrating the first example of X-ray guided magnetic actuation of sperm clusters. In Fig. 1 A, cone beam computed tomography (CBCT)-scans of a female reproductive tract phantom demonstrate an untethered cluster of nanoparticles-coated sperm cells, allowing for controllable movement and localization by attenuating radiation. The cluster is navigated inside the reproductive tract phantom and can be moved toward either the right fallopian tube (Fig. 1 B) or the distal end of the left fallopian tube (Fig. 1 C). Leveraging the inherent flexibility of sperm cells and incorporating nanoparticle coatings enables magnetic manipulation and enhances acoustic impedance 8 , 9 , 26 . We demonstrate that nanoparticles with both positive and negative surface charge (Fig. 2 A–E) can electrostatically self-assemble around sperm cells based on their local surface charge along their length, thereby imparting a magnetic moment to them (Fig. 2 F). According to the vibrating sample magnetometry data (Fig. 2 F), the magnetization saturation of the magnetic particles used in the nanoparticle-sperm clusters is approximately 17 emu/g, indicating typical superparamagnetic behavior, showing very small hysteresis loops and coercivity. However, the incorporation of magnetic nanoparticles introduces biocompatibility concerns that need to be investigated and are suspected to result in a trade-off between actuation, localization, and biocompatibility. While actuation and localization are expected to improve, biocompatibility may be compromised with increasing amounts of magnetic nanoparticles. Furthermore, solely improving the acoustic impedance of soft biohybrid microrobots falls short because pulse-echo techniques have a limited field of view, restricting the procedure’s applicability to smaller animals or for preliminary testing purposes only. Fig. 1 Sperm-based biohybrid microrobots targeting minimally invasive treatment in the female reproductive tract with weak, low-frequency, controlled external magnetic fields and X-ray fluoroscopy imaging for localization of the magnetizable sperm cell clusters. A Left: Schematic illustrating the concept of the simultaneous actuation and detection of sperm clusters in the reproductive tract (created with BioRender.com). Right: A cluster of nanoparticle-coated sperm cells is clearly visible in the cone beam computed tomography (CBCT) scans of the female reproductive tract phantom and located in the right ( B ) and left ( C ) fallopian tube. The phantom was reconstructed from MRI images of the reproductive tract and represents the three-dimensional anatomical features of the human organ in real size. The sperm cells cluster can be precisely controlled, responding to the influence of the external rotating magnetic fields within the female reproductive tract, rendering them detectable and controllable within the reproductive tract for potential applications in diagnostics and treatment in the reproductive tract (Supplementary Movie 1 ). Fig. 2 Fabrication and actuation of biohybrid microrobots consisting of magnetic nanoparticle-coated sperm cells. A Workflow of magnetic biohybrid microrobot fabrication, including nanoparticle fabrication, washing of sperm cells, incubation of sperm cells with nanoparticles and cluster formation (Supporting Information Fig. 2 ). B Transmission electron microscopic image of nanoparticles with an average size of 15 nm, forming clusters of about 121 nm in size (Supporting Information Fig. 3 ). C Scanning electron microscopic image of a single biohybrid microrobot consisting of magnetic nanoparticles and bull sperm, here displaying the paddle-shaped sperm head with attached 121 nm-sized nanoparticles. D Zeta potential measurements indicate a large population of negatively charged magnetic particles, and a smaller population with positive charges (Supporting Information Fig. 4 ). E Scanning electron microscopic image of a biohybrid cluster. F Vibrating sample magnetometer measurements of the superparamagnetic nanoparticles indicate strong magnetization. G A X-ray-guided robotic platform enables simultaneous actuation and localization. G -i Frequency response and detectability experiments are conducted within a one-dimensional saline-filled lumen. G -ii Camera footage and X-ray Fluoroscopy images are captured simultaneously during locomotion inside the female reproductive tract. G -iii The RPM-cluster gap can be increased to 15 cm. H A cluster of sperm cells coated with nanoparticles moves by rolling in response to a rotating external magnetic field. A Left: Schematic illustrating the concept of the simultaneous actuation and detection of sperm clusters in the reproductive tract (created with BioRender.com). Right: A cluster of nanoparticle-coated sperm cells is clearly visible in the cone beam computed tomography (CBCT) scans of the female reproductive tract phantom and located in the right ( B ) and left ( C ) fallopian tube. The phantom was reconstructed from MRI images of the reproductive tract and represents the three-dimensional anatomical features of the human organ in real size. The sperm cells cluster can be precisely controlled, responding to the influence of the external rotating magnetic fields within the female reproductive tract, rendering them detectable and controllable within the reproductive tract for potential applications in diagnostics and treatment in the reproductive tract (Supplementary Movie 1 ). A Workflow of magnetic biohybrid microrobot fabrication, including nanoparticle fabrication, washing of sperm cells, incubation of sperm cells with nanoparticles and cluster formation (Supporting Information Fig. 2 ). B Transmission electron microscopic image of nanoparticles with an average size of 15 nm, forming clusters of about 121 nm in size (Supporting Information Fig. 3 ). C Scanning electron microscopic image of a single biohybrid microrobot consisting of magnetic nanoparticles and bull sperm, here displaying the paddle-shaped sperm head with attached 121 nm-sized nanoparticles. D Zeta potential measurements indicate a large population of negatively charged magnetic particles, and a smaller population with positive charges (Supporting Information Fig. 4 ). E Scanning electron microscopic image of a biohybrid cluster. F Vibrating sample magnetometer measurements of the superparamagnetic nanoparticles indicate strong magnetization. G A X-ray-guided robotic platform enables simultaneous actuation and localization. G -i Frequency response and detectability experiments are conducted within a one-dimensional saline-filled lumen. G -ii Camera footage and X-ray Fluoroscopy images are captured simultaneously during locomotion inside the female reproductive tract. G -iii The RPM-cluster gap can be increased to 15 cm. H A cluster of sperm cells coated with nanoparticles moves by rolling in response to a rotating external magnetic field. This study represents a significant advancement in the field of biohybrid microrobots, specifically those tailored for applications within the female reproductive tract. The primary groundbreaking contribution lies in the capacity to image sperm cells using X-ray technology in a female reproductive tract anatomical model. We demonstrate that satisfying cell viability, magnetic response, and simultaneous imaging with a C-arm can be achieved, demonstrating the first example of X-ray guided magnetic actuation of sperm clusters. This innovative approach offers novel means to visualize and study sperm behavior with unprecedented detail, thanks to the magnetic nanoparticles attached to the sperm clusters. The expected potential impact of this research includes revolutionizing diagnostics and treatment strategies in reproductive health, as well as advancing the field of microrobotics for biomedical applications.

Supplementary Information Supplementary Information