1. Introduction
The
in vitro primary cultivation of patient-derived tissues has significant value for accurate diagnosis, precision medication, individualized therapy, and tissue engineering [
1], [
2], [
3]. Cultivation of patient-derived primary tumor tissues is regarded as one of the most convenient, instructive, and reliable means to reflect the real situation of patients because their characteristics are similar to those of patient tissues, and the culture conditions are not overly demanding [
4], [
5], [
6]. Thus, for tumor therapy and prognosis, the establishment of a tumor-on-a-chip from primary tumor tissues for high-throughput drug screening has attracted great attention [
7], [
8], [
9], [
10]. To extract and cultivate patient-derived primary tissues, commonly used procedures include obtaining clinical tissue samples, manually cutting them into small random pieces with sizes less than cubic millimeters, and culturing them in media [
11]. Despite their widespread application, some shortcomings still remain, particularly in terms of tissue slicing. Such fine slicing is not easy for manual operations and relies on technical experience; the sliced tissue pieces are stochastic and irregular, making standardization difficult. In addition, owing to the small size of the sliced tissue pieces, it is difficult to separate and manipulate them for subsequent implementation. Therefore, novel strategies for clinical tissue slicing and cultivation are highly desirable.
Here, we present a controllable histotomy technique based on hierarchical magnetic microneedle array robots for tailoring primary tissues and constructing a high-throughput tissue-on-a-chip, as schemed in
Fig. 1. Fabricated by three-dimensional (3D) printing of mortise-tenon structures, the slicing device for such histotomy featured a magnetic particle-loaded, pagoda-shaped microneedle scaffold. This microneedle scaffold can entrap tumor tissues, tightly fix them via mechanical interlocking, and avoid tissue slipping. After the microneedle scaffold and the penetrated tissues were assembled into the device, they were cut into small pieces meeting the size requirements for
in vitro cultivation. Owing to the encapsulated magnetic microneedle fragments, these tissue pieces can be manipulated to move and rotate in response to magnetic fields and act as biohybrid microrobots to realize separation, transportation, and dynamic culture. Based on these features, we applied this technique to operate tiny pieces of mouse-derived pancreatic cancer tissue, developed multilayered microfluidic chips, and established a tumor-on-a-chip platform for efficient high-throughput drug screening. These results indicated the promising clinical prospects and practical value of our novel biomedical technique.
2. Materials and methods
2.1. Materials
Ecoflex® 00-30 was bought from Smooth-On, Inc. (USA). Polyethylene glycol diacrylate (PEGDA; average molecular weight (Mn) ≈ 700), polyethylene glycol (PEG; Mn = 200), 2-hydroxy-2-methylpropiophenone (HMPP; purity 97%), dimethyl sulfoxide (DMSO), and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT) were purchased from Sigma-Aldrich (USA). Neodymium-iron-boron (NdFeB) magnetic particles (LW-BA(16-7A)-2000) were bought from Guangzhou New Nord Transmission Parts Co., Ltd. (China). Polydimethylsiloxane (PDMS; Sylgard 184) was obtained from Dow Inc. (USA). Roswell Park Memorial Institute (RPMI)-1640 culture medium, Dulbecco’s modified eagle medium (DMEM), fetal bovine serum (FBS), and penicillin-streptomycin (PS) were acquired from Gibco (USA). Matrigel was obtained from Corning Incorporated (USA). Calcein-AM and propidium iodide (PI) were provided by Invitrogen (USA). Complete culture medium for human pancreatic cancer-derived cells was obtained from Wuhan Procell Life Science & Technology Co., Ltd. (China). Anti-tumor drugs, 5-fluorocrail (5-FU), oxaliplatin (Oxa), gemcitabine (Gem), and paclitaxel (Pac), were provided by Nanjing Drum Tower Hospital (China). Phosphate buffered saline (PBS) solution, histological staining kits, and immunohistochemical antibodies (MUC1, EGFR, and Ki67) were supplied by Wuhan Servicebio Technology Co., Ltd. (China). Water in all experiments was deionized water (dH2O) purified by a water purification system (Milli-Q Plus 185; Millipore, USA) with the resistivity over 18.2 MΩ·cm. All reagents were of the best grade available and used as received.
2.2. Cells and animals
NCI-H1975 (human non-small cell lung cancer) cell lines were provided by Shanghai Cell Bank, Chinese Academy of Sciences, China. Cells were cultured at 37 °C with 5% CO2 in the medium of 89% (v/v) RPMI-1640, 10% (v/v) FBS, and 1% (v/v) PS. FG pancreatic carcinoma cell lines were offered by Nanjing Drum Tower Hospital and were cultured at 37 °C with 5% CO2 in the medium of 89% (v/v) DMEM, 10% (v/v) FBS, and 1% (v/v) PS. Twelve male Balb/c nude mice (6-8 weeks, 20-25 g) were obtained from the Model Animal Research Center of Nanjing University, raised in specific pathogen free (SPF) environments and were treated in strict accordance with the Regulations of Beijing Municipality on the administration of laboratory animals under the approval from the Animal Investigation Ethics Committee of Nanjing Drum Tower Hospital (No. 2020AE01054).
2.3. Fabrication of the magnetic microneedle scaffold
A 3D-printed mold with orderly microneedle array was fabricated by a 3D-printer (nanoArch S140; BMF Precision Tech Inc., China). A part and B part of the Ecoflex silicone were mixed, degassed, poured on the 3D-printed mold, and completely cured at room temperature for 6 h. Then by stretching the cured Ecoflex silicone rubber and using ethanol as lubricant, the 3D-printed mold could be integrally detached and the negative template was acquired. The magnetic microneedle scaffold was composed of a magnetic microneedle array and a nonmagnetic substrate. Mixed prepolymer of 50% (v/v) PEGDA, 20% (v/v) PEG, 1% (v/v) HMPP, and 100 mg∙mL−1 RdFeB particles was first added to the surface of the negative template and treated by vacuum for 5 min to fill the cavities. After removing the redundant solution, prepolymer with the same formula except that RdFeB particles was supplemented and filled other parts of the negative template. After that, ultraviolet (UV) irradiation was applied for 2 min to concurrently solidify both the prepolymers. By finally stretching the negative template and using dH2O as lubricant, the negative template was removed and the microneedle scaffold was prepared.
2.4. Procedures of tissue slicing via the microtomy device
To prepare the microtomy device, a cover, a shell with parallel grooves, and a holder were 3D printed at first. The microtomy device was assembled from these 3D-printed modules together with a surgical blade and the magnetic microneedle scaffold. During the slicing process, the microneedle scaffold was pressed on the tissues to pick them up and fix them on the microneedle tips. Then the scaffold was placed on the holder and inserted inside the shell. After installing the cover in the uppermost grooves, the scaffold was lifted up by the holder to contact the cover. The surgical blade was slid in through the bottom grooves to slice the tissues together with the microneedles. Since the distance between the nearby grooves was set to be 0.5 mm, thin tissue slices containing magnetic microneedle fragments could be cut off for every up move. After that, the remnant scaffold was dragged down by the holder and the tissue slices were gathered with the help of an external magnet. This up-cutting-down cycle was repeated until the slicing was completed. Eventually, the tissue slices were trimmed by a scalpel and tissue pieces with sizes smaller than 1.0 mm × 1.0 mm × 0.5 mm (length × width × thickness) were obtained.
2.5. Magnetic control of the biohybrid microrobots
The magnetic microneedle scaffold was imparted with polarity and strengthened magnetism by an impulse magnetizer (J302; Nanjing Jinlang Electric Co., Ltd., China) with the output voltage of 600 V in advance. A magnetic stirrer (MS-H280-Pro; DLAB Scientific Co., Ltd., China) and a cylindrical permanent magnet (diameter: 6 cm, height: 1 cm; 4000 GS) were used to manipulate the biohybrid microrobots. The experiments were conducted in PBS or culture medium.
2.6. Construction of the microfluidic chip
The multi-layer microfluidic chip was constructed via step-by-step template replication and sealing. For every layer, PDMS containing 10% (v/v) curing agent was poured on the polymethyl methacrylate (PMMA) mask template and solidified at 70 °C for 4 h. Then the PDMS layer was detached from the mask template and each side was treated by oxygen plasma for 10 min. Subsequently, the layers were stacked together with a binder clip in vacuum for 1 h to form PDMS/PDMS bonding. After removing the binder clip, the microfluidic chip was fabricated. This microfluidic chip contained four inputs, sixty incubation chambers (four as a group), and fifteen outputs, which enabled fifteen different combinations as described below:
$Each input: C_{4}^{1}=4 $
$Pairs: C_{4}^{2}=6 $
$Three by three : C_{4}^{3}=4 $
$All : C_{4}^{4}=1$
2.7. Microfluidic tissue culture
The other six layers except the up cover were bonded together using the above method at first. After the freshly-acquired mice-derived pancreatic tumor was sliced into desired tissue pieces, they would be positioned inside the incubation chambers of the microfluidic chip with the assistance of a magnet and a squeegee. During this process, the tumor tissue pieces were first added to one end of the chamber layer and the filling began. Then the squeegee moved forward on the chamber layer to push the tissue pieces to the other end, and the magnet followed to attract the tissue pieces to the bottom of the chambers. Finally, the misplaced tissue pieces would be rearranged by tweezers after removing the squeegee and the magnet. After filling all the chambers, we pressed the plasma-treated up cover onto the microfluidic chip with a heavy iron block. The iron block remained on top of the chip during the following tissue culture experiments. Complete culture medium for human pancreatic cancer-derived cells was injected from the inputs of the chip to culture these tissue pieces. After culturing for three days, the tissue pieces were treated by Calcein-AM to stain the live cells and PI to stain the dead cells. Scanning electron microscopy (SEM) characterization, Hematoxylin and eosin (H&E) staining, and immumohistochemical staining were also applied to the tissue pieces cultured for three days.
2.8. Drug screening via organ-on-chips
The four drugs were 5-FU, Oxa, Gem, and Pac, respectively. Each drug was mixed with the complete culture medium for human pancreatic cancer-derived cells with the final concentration of 0.5 µmol∙L−1 to prepare the work solutions. The four drug working solutions were injected into the four inputs of the microfluidic chip by a peristaltic pump (Longer, USA) at the speed of 1 mL∙h−1. In this case, the groups 1-15 were 5-FU, Oxa, Gem, Pac, 5-FU+Oxa, 5-FU+Gem, 5-FU+Pac, Oxa+Gem, Oxa+Pac, Gem+Pac, 5-FU+Oxa+Gem, 5-FU+Oxa+Pac, 5-FU+Gem+Pac, Oxa+Gem+Pac, and 5-FU+Oxa+Gem+Pac, respectively. The control group without any drug treatment was termed as group 0. After culturing for three days, the tissue pieces were treated by Calcein-AM to stain the live cells and PI to stain the dead cells. The software ImageJ was applied to measure the relative areas of live and dead cells.
3. Results and discussion
In a typical experiment, a microneedle scaffold that replicated the topological structure of a 3D-printed mold was fabricated. A 3D-printed mold consisting of pagoda-shaped microneedles was immersed in Ecoflex silicone rubber and cured at room temperature (Fig. S1 in Appendix A). It should be mentioned that the cured Ecoflex silicone rubber is highly elastic and flexible [
12], [
13], [
14], which could be stretched to over 275% of its original length (Fig. S2 in Appendix A). By stretching the cured Ecoflex silicone rubber, the 3D-printed mold was integrally detached, and a negative template was fabricated. By filling the negative template with prepolymers, solidifying it under UV irradiation, and finally removing the negative template through stretching, a microneedle scaffold was obtained, which was further magnetized to achieve magnetic polarity (
Fig. 2(a)). The microneedle scaffold perfectly inherited the pagoda shape of the original 3D-printed mold, indicating good structure preservation during the entire replication process (
Figs. 2(b)-(d)). The microneedles had six layers, and the overall length was approximately 1400 μm, ensuring that the microneedles were long enough to completely penetrate the tissue. It was also recorded that the height and base diameter of each layer of the microneedle were centered at 230 and 188 µm, respectively (Fig. S3 in Appendix A). Additionally, the sharp microneedle tip was characterized using SEM, as shown in
Fig. 2(e).
The magnetic responsiveness of the microneedle tips was attributed to the embedded magnetic microparticles (size: approximately 5 µm), as shown by the energy dispersive spectrum elemental mapping in
Figs. 2(f) and
(g). These magnetic particles can be imparted with a fixed magnetic polarity after magnetization [
15], [
16], [
17], [
18], [
19]. It is worth mentioning that the concentration of the magnetic particles influenced both the solidification time and magnetic responsiveness of the microneedles. In particular, the curing time of the microneedle prepolymer increased with the particle concentration, indicating that solidification was more difficult (Fig. S4 (a) in Appendix A). By contrast, the maximum interaction distance between the microneedles and the external magnet increased with more particles, indicating stronger magnetic responsiveness, although this effect was no longer obvious beyond 100 mg∙mL
−1 (Fig. S4(b) in Appendix A). Considering both parameters, a final magnetic particle concentration of 100 mg∙mL
−1 was used for the follow-up experiments. In addition, the penetrated tissue affected the magnetic response. To determine this, agarose blocks were used to mimic tissues and penetrated with microneedles. With the agarose block, the maximum reaction distance decreased, and the overall magnetic attraction of the microneedles became weaker, mainly because the impact of the increased mass outweighed that of the magnetism (Fig. S5 in Appendix A). Based on this premise, microneedles with and without tissue can be separated by simply reducing the strength of the external magnetic field.
The tumor tissue penetration ability of the microneedle scaffold was evaluated. Tumor tissues were collected from the mice and manually pressed onto a microneedle scaffold for 10 min. The overhead view of the SEM image showed that the microneedles left holes with a width of approximately 200 μm in the tumor tissue (Fig. S6(a) in Appendix A). In addition, the sectional view of the H&E staining image further proved that the microneedle could reach deep tissue and form a conical passage (Fig. S6(b) in Appendix A). Furthermore, the pagoda-like multilayered microneedle shape enhanced tissue fixation and adhesion [
20], [
21], [
22]. To quantify this, the force required to pull the microneedle scaffold from the tumor tissue was measured in real time. The maximum pulling force of the multilayered microneedle scaffold was found to be 4.66 times higher than that of its smoothly surfaced counterpart (Fig. S7 in Appendix A). In addition, because of mechanical interlocking, the connection time between the multilayered microneedle scaffold and the tissue was elongated, as reflected by the prolonged force-sensing time.
To construct the microtomy device, the microneedle scaffold was assembled with a 3D-printed cover, a surgical blade, a 3D-printed holder for carrying the microneedle scaffold, and a 3D-printed shell with each groove 0.5 mm apart (
Fig. 3(a)). The characterization of the different parts and the assembled microtomy device were shown in
Fig. 3(b) and Fig. S8 in Appendix A. To cut the tissues, the magnetic microneedle scaffold was first pressed onto the tissue to fix it, then fitted into the holder and placed inside the shell. After the cover slid through the uppermost groove, the microneedle scaffold was lifted by the holder until it touched the cover (
Figs. 3(c) and
(d)). The blade was inserted through the groove below to slice the tissue together with the microneedle scaffold. Eventually, the holder carrying the remnant microneedle scaffold was moved down, and sliced tissue pieces containing the microneedle fragments were transferred and collected with the aid of an external magnet. This up-cutting-down cycle was repeated until all the tissues were sliced. Notably, by staining the sliced tissue pieces with H&E, the microneedle fragment was embedded and completely encircled by the tissue (
Fig. 3(e)). The fabricated tissue pieces were uniform and smaller than 1.0 mm × 1.0 mm × 0.5 mm (length × width × thickness), which fulfilled the clinical requirements for
in vitro primary tissue culture (
Figs. 3(f)-(h)).
Benefiting from the embedded magnetic microneedle fragments, the tissue pieces were imparted with magnetic responsiveness and thus behaved as magnetic biohybrid microrobots. These biohybrid microrobots could be manipulated under an external magnetic field, as shown in Movie S1 in Appendix A. Specifically, attracted by a magnetic field with fixed direction and constant intensity, the biohybrid microrobot could move from one end to the other (
Fig. 4 (a)). Additionally, under a rotating magnetic field, the biohybrid microrobot rotated along a circular trajectory (
Fig. 4(b)). Moreover, multiple biohybrid microrobots can be simultaneously controlled by a rotating magnetic field and can move independently along their own routes, akin to satellites (
Fig. 4(c)). In addition, the microneedle scaffold materials and magnetic particles showed good compatibility with tumor cells, as demonstrated by co-culturing tumor cells with the magnetic microneedle scaffolds. The results showed that the cells grew and proliferated well in the presence of magnetic microneedle scaffolds, indicating cytocompatibility and laying the foundation for primary tissue cultivation (Fig. S9 in Appendix A).
High-throughput microfluidics have been utilized in many biomedical areas [
23], [
24], [
25], [
26], [
27], [
28], [
29], [
30]. One remarkable example is the use of organ-on-chip for drug screening. The aforementioned biohybrid microrobots can be easily filled into the chambers of the microfluidic chip with the help of a magnet and a squeegee, as shown in
Fig. 4(d). Specifically, the squeegee moved across the surface of the microfluidic chip to brush the tissue pieces from one end to the other (
Fig. 4(e)). This step was necessary because squeegees removed misplaced tissue pieces and ensured that each chamber contained only one tissue piece. Simultaneously, a magnet below the chip was used to attract the tissue pieces to the bottom of the chamber. A small number of misplaced tissue pieces were rearranged using tweezers after withdrawing the squeegee and magnet. Using this approach, these tissue pieces can be neatly placed into all the chambers of the chip.
Here, for efficient antitumor drug screening, we designed a microfluidic chip that contained four entrances for inputting four types of drugs, allowing the simultaneous monitoring of fifteen different types of drug combinations (Fig. S10 in Appendix A). For each drug combination, there were four parallel incubation chambers and a common output to collect the metabolites. To achieve complex fluid transport and mixing, the fluid channels of the chip were divided into five layers, which, from top to bottom, were the outlets of the incubation chambers, chambers, inlets of the chambers, main fluid channels, and a space for bypassing to avoid channel crisscrossing (Fig. S11 in Appendix A). It is worth noting that this down-in and up-out design helped the fluid to flow fully into the incubation chamber, and this equal-length path design mixed the fluids from different inputs well. To demonstrate this theory, a numerical simulation was conducted, as shown in Fig. S12 in Appendix A. The results showed that after entering the incubation chamber from below, the fluid gently filled the entire space at a significantly decreased flow rate. However, the flow rate increased near the upper outlet to discharge the fluid quickly. In addition, at all confluences, including the two-way parallel crossing, two-way rectangular crossing, and three-way rectangular crossing, fluids that traveled through equal-length paths could be evenly mixed.
5-FU, Oxa, Gem, and Pac are antitumor drugs commonly used clinically. In recent years, combined chemotherapy, such as the FOLFIRINOX regimen, has been proven effective in treating malignant tumors, such as pancreatic cancer [
31], [
32], [
33]. Thus, in this proof-of-concept experiment, we used sliced mouse-derived pancreatic tumor pieces and a microfluidic chip to evaluate the therapeutic effects of these four drugs and the effects of their different combinations. The microfluidic chip was covered with a lid, and the four drug solutions were injected into the chip from the four inputs, as shown in
Figs. 5(a) and
(b). In this way, the fifteen combinations were 5-FU (group 1), Oxa (group 2), Gem (group 3), Pac (group 4), 5-FU+Oxa (group 5), 5-FU+Gem (group 6), 5-FU+Pac (group 7), Oxa+Gem (group 8), Oxa+Pac (group 9), Gem+Pac (group 10), 5-FU+Oxa+Gem (group 11), 5-FU+Oxa+Pac (group 12), 5-FU+Gem+Pac (group 13), Oxa+Gem+Pac (group 14), and 5-FU+Oxa+Gem+Pac (group 15). Tumor pieces without any drug treatment were set as group 0.
Notably, to prove that cell death was mainly caused by drugs rather than the slicing process, and that the on-chip culture did not alter the physiological state of the cells, only complete culture medium was injected into the chip, and the tumor tissue pieces were cultivated for three days. Calcein-AM and PI were used to stain the tissue pieces simultaneously, where live cells were stained green and dead cells were stained red. The large region of green cells and the very small area of red cells indicated that the cells grew well without drugs, as shown by the live/dead staining results (
Fig. 5(c)). We also conducted SEM imaging and H&E staining of the tumor tissue pieces. The SEM image shows that the cancer cell clusters from the tumor tissue pieces retained their disordered morphology, with cells squeezing each other and stacking (
Fig. 5(d)). H&E images further revealed deeply stained and enlarged cell nuclei, as well as an imbalanced nucleocytoplasmic ratio, indicating that cells from the tumor tissue pieces retained cancerous features (
Fig. 5(e)). In addition, the massive expression of MUC1 (a mucin marker for pancreatic cancer diagnosis), EGFR, and Ki67 (proliferation markers of tumor cells) in these tumor tissue pieces demonstrated that their specific function and high activity remained during on-chip culture (
Fig. 5(f), Fig. S13 in Appendix A).
By contrast, when the tissue pieces were treated with any of the chemotherapeutic drugs, widespread cell death occurred, as indicated by the dramatically increased red area in the fluorescence images (
Figs. 5(g)-(l)). In particular, by simultaneously applying the four drugs, the killing effects on tumor cells were significantly enhanced. For quantification, we recorded the live/dead cell ratios of the control group (group 0) and all 15 drug combinations (groups 1-15) by calculating the area ratios of the green and red regions (
Fig. 5(m)). Compared to the live/dead cell ratio of the control, which was 9.77 ± 2.19, the ratios of all the drug-treated groups were below 1.50. The four-drug-combination group had the lowest live/dead cell ratio of 0.78 ± 0.13. These data further demonstrated the improved antitumor efficiency of the drug combination therapy. Previous clinical research has reported the benefits of multidrug combination therapy for pancreatic cancer. For instance, Gem plus Pac treatment [
34] or Gem with Oxa [
35] were more efficient than using only one drug. Different combinations of Oxa, 5-FU, Gem, and Pac have shown promising therapeutic effects [
36], [
37]. Our results were consistent with these findings. All results indicated that the presented tissue slicing and cultivation techniques could be a reliable strategy for clinical diagnosis.
4. Conclusions
In summary, we developed a tissue slicing technique to produce magnetic biohybrid microrobots for culturing primary tumor tissues and establishing tumor-on-chips. This technique was based on a 3D-printed, mortise-tenon-structured microtomy device that consisted of a cover, surgical blade, holder, shell, and magnetic microneedle scaffold. As the pagoda-shaped microneedles were loaded with polarized magnetic particles, the scaffold could not only fix the tissues via mechanical interlocking to avoid relative movements during slicing but also impart the sliced tissue pieces with magnetic responsiveness to construct the biohybrid microrobots. During the slicing process, the tumor tissue was first pressed onto the microneedle scaffold. After assembling into the microtomy device, the microneedle scaffold was moved up to touch the cover and cut using a sliding blade. Because the distance between the cover and blade was controlled at approximately 0.5 mm, every upward movement led to the generation of a very thin tissue slice. Tissue pieces smaller than 1.0 mm × 1.0 mm × 0.5 mm could be acquired by further trimming the sliced tissue with a scalpel. The thickness limit of the tissue pieces was determined by the thickness of the back of the blade. Here, the back was approximately 0.2 mm thick, and the limit was thus approximately 0.2 mm. The resulting biohybrid microrobots could be easily manipulated using magnetic fields, facilitating follow-up tissue separation and cultivation. By employing mouse-derived pancreatic cancer tissues to construct high-throughput tumor-on-chips and evaluate drug efficacy, we demonstrated the clinical application of this technique in precise tumor medicine.
Considerable effort is required to further improve this technique. Therefore, it is important to automate the horizontal sectioning and production of cubes. This can be achieved by replacing the remnant microneedle scaffold with a grid blade array after each up-cutting-down cycle and moving it toward the sliced tissue to cut it vertically. The length and width of the tissue cubes can be adjusted by changing the grid size. The overall size of the current microtomy device is only a few centimeters, and it is expected to be scaled up to simultaneously slice more tissues and achieve a higher throughput. In this proof-of-concept study, tissue positioning inside the chip was performed manually. This would be a good direction for improvement in automating this process. Some of these methods are worth exploring in the future. For example, an array of tiny magnetic coils can be placed under the chambers, and a squeegee can be assembled with a toy crane. By using a computational program to change the on-off and strength of the magnetic coils at different positions, it is possible to automatically move and fill the tissue pieces. By adjusting the moving speed of the crane to make it consistent with the magnetic field change, it is possible to automatically remove the misplaced tissue pieces. Additionally, we utilized this tumor-on-a-chip to test clinically promoted pancreatic cancer drugs and therapies. It is believed that such tumor-on-chips are more valuable for screening newly developed drugs, such as targeted and gene-based drugs. In addition, the application of our technique can be extended beyond cancer to other types of patient-derived primary tissues, including skin, testis, liver, and kidneys. Moreover, because this technology causes little damage to tissues, it can be applied to long-term primary tissue cultivation and observation.
Acknowledgments
This work was supported by the National Key Research and Development Program of China (2020YFA0908200), the National Natural Science Foundation of China (T2225003, 52073060, 61927805), the Nanjing Medical Science and Technique Development Foundation (ZKX21019), the Clinical Trials from Nanjing Drum Tower Hospital (2022-LCYJ-ZD-01), and the Guangdong Basic and Applied Basic Research Foundation (2021B1515120054).
Authors’ contribution
Yanjin Zhao conceived the idea, revised the manuscript, and supervised the project. Xiaoxuan Zhang conducted the experiments, arranged the figures, and wrote the manuscript. Hanxu Chen assisted in the numerical simulation and microfluidic chip fabrication, and checked the writing. Taiyu Song assisted in the animal experiments and revised the relevant contents. Jinglin Wang provided the technical and clinical support.
Compliance with ethics guidelines
Xiaoxuan Zhang, Hanxu Chen, Taiyu Song, Jinglin Wang, and Yuanjin Zhao declare no competing interests.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2024.05.004.
PDF
(2572KB)
