High-throughput solutions in tumor organoids: from culture to drug screening

Abstract

Tumor organoids have emerged as an ideal in vitro model for patient-derived tissues, as they recapitulate the characteristics of the source tumor tissue to a certain extent, offering the potential for personalized tumor therapy and demonstrating significant promise in pharmaceutical research and development. However, establishing and applying this model involves multiple labor-intensive and time-consuming experimental steps and lacks standardized protocols and uniform identification criteria. Thus, high-throughput solutions are essential for the widespread adoption of tumor organoid models. This review provides a comprehensive overview of current high-throughput solutions across the entire workflow of tumor organoids, from sampling and culture to drug screening. Furthermore, we explore various technologies that can control and optimize single-cell preparation, organoid culture, and drug screening with the ultimate goal of ensuring the automation and high efficiency of the culture system and identifying more effective tumor therapeutics.

Graphical Abstract

Graphical Abstract.

Introduction

An appropriate in vitro disease model is crucial for future research.¹ The tumor organoid culture system, based on three-dimensional (3D) cell culture technology, has emerged as a promising tumor model capable of simulating the growth and development of tumors in the human body.² This system utilizes 3D scaffolds formed by in vivo matrix analogs to recapitulate the in vivo tumor microenvironment (TME); this is achieved by adding various growth factors and small-molecule inhibitors to tumor cells obtained from patient-derived surgical, puncture tissue, or bodily fluid samples.³ This approach enables the induction and culture of organoids with tumor properties. The tumor organoid culture system exhibits high replicability of source tissues and stability,⁴ making it the best in vitro 3D model to date. Patient-derived organoids form heterogeneous microstructures.⁵ Heterogeneity helps researchers to study the mechanism of tumor development and perform drug screening and efficacy assessments.^6,7 However, the establishment and application of tumor organoid culture systems are hindered by complex experimental operations, expensive time and labor costs, and a lack of standardized processes and uniform identification rules. These limitations underscore the need for high-throughput solutions to facilitate the widespread adoption of tumor organoid models. Compared with other tumor models, organoids offer the advantages of rapid culture and effective simulation of organogenesis and its physiological and pathological states.⁸ To maximize these advantages, a series of high-throughput solutions are needed during the culture process to increase accuracy and efficiency. Tumor organoid manipulation involves several key steps that can be addressed in a high-throughput manner, including biological sample processing, image acquisition, drug screening, and cell viability evaluation.⁹ This review aims to provide a comprehensive overview of high-throughput solutions used in tumor organoid culture and discuss their applications, limitations, and future prospects. High-throughput solutions for these critical steps can enable more precise control and optimization of key aspects of the organoid culture process, such as single-cell preparation, cell cultivation, and drug screening. This approach improves the efficiency of the culture system and has significant implications for the discovery of more effective tumor therapies.

Tumor organoid culture technology and applications

Organoid culture technology is an extension of in vitro 3D culture technology, with a key distinction being the provision of a self-renewal condition that supports the self-maintenance and differentiation of stem cells, thereby better simulating the physiological environment of the human body. Notably, organoids can maintain the physiological functions of cells, including proliferation, apoptosis, and differentiation. In contrast, traditional 2D culture is limited to forming cell monolayers, which fail to replicate the 3D structure of human tissues, and the cells often lose their original physiological functions during passaging. Animal models of different species and strains exhibit distinct physiological characteristics and responses but are hindered by high heterogeneity⁵ and high costs. Conversely, tumor organoids are derived from patient-derived tumor tissues. Culturing these organoids involves adding growth factors specific to the corresponding cancer type, thereby establishing an in vitro tumor model that recapitulates the in vitro microenvironment. This approach not only directly reproduces the growth process of tumors but also individually reproduces the characteristics of the source tumor tissues, enabling the tumor organoid model to more accurately capture interindividual differences and providing a foundation for personalized tumor therapy.

The tumor organoid culture system comprises a matrix gel and a cocktail of growth factors that maintain organoid growth and viability. The matrix gel, primarily comprising collagen, nestin, laminin, and other components, serves as a 3D scaffold for tumor cells, recapitulating the key properties of the natural extracellular matrix (ECM), such as fiber composition, permeability, void size, and mechanical stability. This biomimetic environment enables the replication of complex cell–cell and cell–ECM interactions in vitro, providing ideal conditions for cell aggregation, proliferation, and migration on the scaffold, thereby accurately simulating the human body’s internal environment.

In addition to using matrix gel, the successful construction of tumor organoid models relies on supplementing specific growth factors and small molecule inhibitors (Figure 1). For example, lung cancer organoids require the addition of epidermal growth factor (EGF) to the culture medium to promote organoid proliferation, whereas R-Spondin-1, Noggin, and other factors are necessary to maintain stemness. Furthermore, fibroblast growth factor 7 (FGF7) and FGF10 are required to promote the differentiation of stem cells toward the distal end of the cell lineage. Another essential component of the culture system is small molecule inhibitors, such as A83-01,¹⁰ which inhibits epithelial–mesenchymal transition, and SB202190,¹¹ which inhibits cell differentiation by targeting the mitogen-activated protein kinase signaling pathway. Notably, adding these factors is more labor intensive than using serum, and the specific factors required for organoids constructed from different tissue sources are not identical, necessitating adjustments according to the characteristics of each tissue.

Figure 1.

Organoid culture system. Depending on the source of the tumor cells, different combinations of cytokines need to be added to the tumor organoid medium (eg, R-Spondin-1, Noggin, FGF7, FGF10, EGF, Wnt-3a, Y-27632, nicotinamide, N-acetyl-L-cysteine, A83-01, and SB202190) to promote or inhibit signaling pathways involved in organoid formation to obtain the desired organoids.

Establishing organoids involves a series of steps, including sampling, culturing, and subsequent evaluations. Tumor tissues are typically obtained from surgical specimens, puncture biopsies, and endoscopic biopsies, highlighting the advantage of requiring only a minimal number of cells for constructing organoid models.¹² Notably, there is a risk of contamination during tissue sampling and transportation, necessitating the addition of antibiotics, such as penicillin-streptomycin, to the sample tubes. The initial step in processing involves enzymatic digestion of tumor tissue into single cells, typically via a protocol that includes collagenase, DNAase, and hyaluronidase. The digested suspension is subsequently centrifuged to obtain a cellular pellet, which is then treated with erythrocyte lysates to generate a single-cell suspension of the tumor. Next, the single-cell suspensions are mixed with matrix gel and seeded into well plates, followed by the addition of a tumor organoid culture medium. The cultured organoids are then identified through multiple approaches. First, the cellular morphology, arrangement, and structure are evaluated at the cellular level to determine the consistency between the tumor organoids and the source tissue. Second, immunohistochemical staining with specific antibodies is employed to identify the expression of tumor markers at the protein level. Third, high-throughput sequencing technology is utilized to analyze gene expression profiles and gene mutations at the gene level. Finally, the utility of tumor organoids in drug screening and therapeutic strategy development is assessed by comparing the response of tumor tissues and organoids to chemotherapeutic drugs (Figure 2).

Figure 2.

Organoid culture process. Tumor tissue (surgery, tissue biopsy) or liquid biopsy samples of patient origin are made into single-cell suspensions after a series of treatments, and seed plates are made according to a certain number of cells. Drug-sensitive tests can be performed to observe cell viability when organoids are formed, and data collection and analysis can be performed.

The standardization of organoid culture protocols is crucial for ensuring the reliability of experimental results. In the context of tumor organoid culture systems, every step, from single-cell acquisition to organoid culture, drug screening, and validation, must adhere to standardized operating procedures (SOPs).¹³ This approach encompasses prelaboratory preparation, laboratory protocols, and postlaboratory processing.¹⁴ The construction of tumor organoid models can be broadly divided into the following stages: tumor tissue digestion and dissociation, mixing of isolated tumor stem cells with matrix gel, plating them in a well plate, and adding cytokine-rich culture medium to promote organoid growth. This process requires significant labor and time costs, thereby limiting the applicability of organoids.

The initial step in tumor organoid culture involves identifying a suitable cell source. Tumor cells can be derived from patient samples, cell lines, or stem cells. Purified tumor cells are obtained through digestion, isolation, and culture.¹⁵ However, this process is fraught with challenges. Digestion times vary substantially across different cancer types, with breast tumor tissue, for example, requiring 4-6 hours because of its high fibrous content, whereas gastrointestinal tumors can be digested within 1-2 hours. Moreover, distinct cancer types necessitate different digestive enzyme systems. These factors pose obstacles for researchers, including prolonged waiting times for digested tissues, low cell viability due to overdigestion, and insufficient digestion, which can hinder the generation of single-cell suspensions and subsequently affect downstream experiments. The emergence of these challenges underscores the need for a simplified, efficient, and reliable digestion method capable of yielding single-cell suspensions.

Cell cultivation requires precise control over cell density and distribution to ensure the propagation of tumor cells.¹⁶ Cell counting and viability assessments are necessary to determine the number of live cells following the generation of single-cell suspensions. The corresponding culture system matrix gel is then mixed with the single-cell suspension and plated in a well plate. Although this process is relatively straightforward, generating large numbers of organoids within a short timeframe through manual labor alone while maintaining control over the quality of each organoid sphere is challenging. This limitation is particularly pronounced in high-throughput drug sensitivity assays, where plating organoids in 384-well plates or higher-throughput formats necessitates automated machinery. Subsequently, an organoid culture system was established in which cell suspensions are inoculated in a 3D matrix to mimic the in vivo growth environment of tumors and induce organoid formation.¹⁷

Evaluating the effects of drugs on tumor organoids requires a multifaceted approach, encompassing drug preparation, drug screening, and data analysis.¹⁸ The drug preparation process involves selecting, dissolving, and diluting the drug to ensure that its quality and concentration meet the experimental requirements. Drug screening involves treating tumor cells with a drug and observing its efficacy and toxicity, among other parameters. Data analysis involves organizing and analyzing the experimental data to elucidate the potential mechanism of the drug and its therapeutic effects.

Difficulties in culture and standardization beset conventional organoid culture.¹⁹ Furthermore, organoid production, with variable sizes and shapes, is often time-consuming and costly, making large-scale production a significant challenge. Therefore, high-throughput and automated manipulation are crucial solutions for implementing organoid culture progress in terms of standardization, labor alleviation, and cost reduction. To ensure that the overall process of organoid culture is conducted with high precision and efficiency, high-throughput solutions must be designed, considering factors such as automation, high throughput, and reproducibility.²⁰ Examples of such solutions include using microfluidics for organoid implantation,¹⁵ automated robots for drug treatment and detection,²¹ high-resolution imaging for cell state analysis, and data analysis software for result processing. By implementing high-throughput SOPs, experimental errors and uncertainties can be minimized, and the credibility and reproducibility of experiments can be improved.

Substrate technologies involved in current ex vivo organoid culture and high-throughput solutions for various applications

High-throughput solutions are being investigated to alleviate the labor-intensive and critical steps in the organoid culture process. These solutions can be broadly categorized into isolation and culture, drug screening, and activity evaluation. The formats of these solutions encompass equipment, methods, and software that integrate high-throughput capabilities. High-throughput devices and methods include automated sample processing equipment, microfluidics, organ-on-a-chip, 3D printing technology, and high-content imaging systems. High-throughput software can rapidly and accurately acquire and analyze large amounts of experimental data, thereby improving the efficiency and accuracy of data processing.²² Imaging systems, in particular, can simultaneously acquire various types of biological information, such as cell morphology, fluorescence intensity, and cell activity, providing a more comprehensive understanding of the mechanism of action and therapeutic efficacy of drugs on tumor cells.

Microfluidics refers to the technology of systems that process minute quantities of fluid (10^–9-10^–18 l) in channels of tens to hundreds of micrometers. This technology has four main applications: molecular analysis, biosensing, molecular biology, and microelectronics.²³ In biology and medicine, microfluidics plays an increasingly important role in cell culture, single-cell analysis, and clinical diagnosis.²⁴ The application of microfluidics in biological research can significantly reduce sample sizes, thereby decreasing experimental costs, enabling fast and high-throughput detection of biological samples, maximizing the collection of more information from a small number of samples, creating better simulations of the physiopathological environment of the human body, and facilitating more in-depth and comprehensive research.²⁵ For example, the number of circulating tumor cells (CTCs) can indicate the degree of malignancy of a tumor and the risk of metastasis; however, the scarcity of CTCs in peripheral blood has hindered their application. Nevertheless, with the continuous development of microfluidic technology, several microfluidic-based CTC capture technologies have emerged.²⁶ Notably, Jinho Kim et al.²⁷ constructed a microfluidic device (SIM-Chip) that can enrich CTCs from whole blood and identify and separate them.

Organoid microfluidic technology integrates organoids and microfluidics to enable in vitro culture and simulation of organoids by designing microchannels and reaction units on microfluidic chips.²⁸ This technology leverages microfluidic devices and computational software to precisely control the microenvironment, mimicking the TME²⁹ to study tumor mesenchymal interactions.³⁰ Microfluidics facilitates the precise and prompt delivery of cells, drugs,³¹ and nutrients to cancer cells. Moreover, microfluidics allows for automated fluid transport speed, concentration, and duration control, enabling dynamic screening with drug mixtures and nutrients.³² Organoid microarray, a microfluidic-based organoid culture platform, simulates the in vivo TME, enabling the study of tumor growth and metastasis.³³ This technique requires the construction of microfluidic channels and cell culture chambers before subsequent experiments and analyses.³⁴ Furthermore, 3D printing technology can be combined with microfluidics to create innovative solutions. 3D printing technology is an additive manufacturing method that fabricates 3D physical objects from computer model data by stacking layers.³⁵ The emergence of 3D printing technology has provided new possibilities for microfluidics. First, 3D printing technology provides broad prospects for the standardization and mass production of microfluidic chip preparations. Second, 3D bioprinting technology enables precise control over cell and biomaterial distributions, better simulating the complex microstructures of biological tissues. 3D bioprinting technology can create high-resolution microstructures to accurately reproduce the complex characteristics of the TME.³⁶ In addition, these 3D-printed models can be used as preclinical models for versatile research and applications in tumor biology. By combining 3D printing technology and microfluidics, many stable tumor organoid models can be constructed quickly, maximizing the advantages of organoid models.

High-content screening (HCS) is an imaging-based assay that leverages automated imaging and data analysis to provide unbiased, multiparameter data at the single-cell level with high-resolution, high-throughput, and high-information content.³⁷ HCSs enable the capture of images and extraction of information at various levels, including cellular and molecular levels, allowing for the detection of changes in gene expression, protein localization, and morphology.³⁸

High-content imaging techniques can be employed following organoid culture completion, involving the acquisition of organoid data through automated high-throughput photography, such as high-speed microscope imaging.³⁸ The primary function of HCSs is to rapidly acquire a large quantity of data with fewer cells, shorter time, and higher throughput while reducing labor consumption. This technology can collect multichannel data, including cell morphology, status, quantity, and function information. Furthermore, HCSs can be integrated with drug delivery devices to observe the immediate response of cells to drugs in real time.

In recent years, there has been a substantial surge in artificial intelligence (AI) technology. In the medical field, AI can facilitate the design of drugs³⁹ and aid in cancer detection, diagnosis, and treatment optimization, among other applications. The integration of high-throughput experiments and AI technology in medicine has garnered increasing attention. AI can be leveraged throughout the entire process of generating tumor organoids, from construction to application. AI-based automation can assist in the formation of stable organoid spheres. Intelligent algorithms incorporating image recognition and data analysis can rapidly analyze experimental results and draw conclusions.⁴⁰ Various AI-based image recognition algorithms can also enable high-throughput applications for organoids. Intelligent analysis systems equipped with AI modules can select appropriate data analysis methods according to the type of data collected and the purpose of the analysis.⁴¹ The features, advantages, and disadvantages of high-throughput substrate technologies are summarized in Table 1.

Table 1.

High-throughput substrate technologies during tumor organoid culture systems.

High-throughput substrate technology	Microfluidics			High-content imaging technology	Image recognition algorithms
	Organ-chip	3D printing	Automated drug delivery devices
Characteristics (construction, materials, and processes)	Lithography	Light curing systems	Automated micropumps, multilayer microfluidic chips	Z-stack multiple color channels	Deep learning algorithm
Capacity	Simultaneous cultivation of hundreds of organoids	Six-384-well plates	Six-1536-well plates	Ninety-six-384-well plates	Fast image processing in minutes
Advantages	Can provide a complex microenvironment that mimics the real in vivo environment	Prints multiple cells, autocalibration, adjustable droplet size	Dynamic drug delivery, real-time monitoring, low drug consumption	Enables rapid and accurate detection of cell morphology, number, and viability	Fast, high-quality processing and reconstruction of highly complex images
Limitations	Continuous, stable vascular structures need to be constructed and ensured to be functional to act as selective barriers for the transport of drugs and active molecules. And also the interactions between organs need to be modeled in order to completely mimic the elements required for drug studies	Due to the complex structure of certain organs, 3D printing technology cannot directly replicate them. In addition, the physical properties of certain materials used to create the models are not sufficient for 3D printing.	Lack of relevant industrial support	May have poor focus and complex interface	Large amounts of data need to be preimplanted to process images more accurately
Reference	36,37,42-45	20,46-48	35,36	31,32,49	25,50

The features, advantages, and disadvantages of high-throughput substrate technologies are summarized in Table 1.

High-throughput solutions for organoid culture processes

Application of high-throughput techniques in tissue dissociation

Dissociating patient-derived tumor tissue masses into single cells is a critical initial step in tumor organoid culture. This step ensures the number and viability of harvested single cells and provides a stable data source for subsequent experiments. Instruments such as fully automated tissue processors and single-cell suspension preparation instruments have been developed to facilitate high-throughput automated harvesting of single cells. These instruments can prepare human tumor tissues as single cells through machine grinding and enzymatic reactions. Tissue dissociators, which combine mechanical milling and enzymatic digestion, can isolate single-cell suspensions from tissues and prepare tissue blocks as homogenates. These instruments utilize an automated system to customize applications, such as temperature, time, and speed, for standardized process operations and the easier handling of samples. The advantages of these instruments include increased precision and efficiency and decreased cell damage. The instruments can process up to 8 different tissues and samples simultaneously, with sample sizes ranging from 20 to 4000 mg, and are easy to operate. For example, simply placing the tissue sample into the separator tube, adding enzymes and other solutions through the septum-enclosed holes, and setting up the desired program allow the machine to automatically process the tissue sample through a grinding process, resulting in a high-quality single-cell suspension or subcellular material. For example, Jungblut et al.⁵¹ successfully processed mouse lung tissue into a high-quality single-cell suspension using an instrument. Automated tissue sample collection is another crucial aspect of tumor organoid culture. Using robot-assisted technology or automated sampling devices enables the efficient collection and processing of large numbers of samples. Implementing automated tissue collection systems and robotic arm-assisted operations can increase the speed and accuracy of sample collection.

Instruments for tissue processing at this stage may fulfill most of these needs. With the rapid development of AI technology, the automated processing of tissues will also become increasingly intelligent and personalized. In the future, the process may involve not only simple digestion but also automatic sensing devices during digestion to automatically regulate the temperature, enzyme concentration, and other conditions in real time, which is a worthwhile future research direction.

Application of high-throughput technology in organoid implantation

Organoid culture is a crucial step in simulating the process of tumor development. However, manual operation during organoid cultivation can lead to significant variations in individual organoid droplets’ composition, morphology, and function, which may impact the final experimental results. Moreover, the efficiency of manual cultivation is limited, making it challenging to prepare large numbers of organoids within a short timeframe. In contrast, introducing automation and standardization into high-throughput culture systems, such as automated organoid culture platforms based on microfluidics technology, can standardize and homogenize each organoid, enabling a monodisperse and high-throughput culture of organoids. This approach can better mimic the malignant phenotype and metastatic ability of tumors. Automated cell culture and monitoring can significantly improve the efficiency and accuracy of tumor organoid culture. For example, using microtiter plates and automated equipment for organoid cultivation and detection can streamline the process, ensuring consistency and reproducibility.

Jiang et al.¹⁵ developed an automated organoid platform integrating microfluidically linked 3D printing technology, enabling rapid fabrication and culture of normal tissues and tumor organoids with automated manipulation. In this study, primary tissue-derived cells were mixed with matrix gel, and the resulting cell suspension was injected into a microfluidic system to generate uniform cell-loaded matrix gel droplets via a microfluidic droplet printer. The droplets were gelatinized within the microfluidic pipeline, yielding homogeneous organoid precursors at high throughput. These precursors were then printed into 96-well plates, a design facilitating high-throughput imaging characterization of organoids during culture. Notably, these organoid precursors achieved successful culture within 1 week, at a lower cost and with reduced time, and were ready for drug testing. The results demonstrated that tumor organoids prepared via this method exhibited 81% predictive accuracy, indicating a high degree of reproducibility in mimicking the response of patients’ tumors to anticancer drugs.

Du et al.⁹ leveraged genetically engineered human colonoid organoids to establish a miniaturized 3D organoid culture platform, enabling high-throughput screening. Researchers have employed a multipoint combination dispenser, also referred to as a liquid dispenser, to automatically dispense cell–gel mixtures or drug reagents into 384/1536-well plates and ensure accurate and consistent dispensing. This approach yielded rapid and high-quality assay data. The 3D organoid culture platform has been demonstrated viable for large-scale primary compound screening by facilitating high-throughput screening of compound activity in cells and miniaturizing organoids in 384-well plates, achieving ultrahigh-throughput sequencing in 1536-well plates. Furthermore, incorporating a liquid handling robot equipped with a cooling table and pipetting arm enabled the testing of multiple concentrations of matrix gels in routine assay development, thereby rapidly optimizing organoid growth conditions.²¹ Fully automated workstations with liquid-handling robots can be set up for automated seeding of organoids.⁵²

A microfluidic device was also developed to pattern hydrogels, utilizing laminar flow to create hydrophobic channels on both sides while generating a hydrophilic channel in the middle. When a liquid hydrogel is injected into the hydrophilic region, it is confined to this path by the surrounding hydrophobic region. This device provides a controlled 3D microenvironment for cultured cells and tissues, overcoming the costly and complex handling challenges associated with microcolumns and small pressure barriers, thereby improving the efficiency of microfluidic cell culture analysis. The feasibility of this protocol was demonstrated by culturing human umbilical vein endothelial cells within the device.⁵³

Photolithography is a widely employed technique in the fabrication of organ chips, where a photoresist is utilized to transfer the pattern of a photomask onto the chip under the influence of light, thereby enabling the production of organ chips with arbitrary patterns.²⁹ Notably, Hu et al.⁵⁴ developed an integrated superhydrophobic microwell array chip (InSMAR-chip) for the culture of lung cancer organoids, enabling the simultaneous culture of hundreds of organoids and facilitating high-throughput culture and analysis.

With the continuous progress and improvement of high-throughput technology, increasingly complex and refined organoid models will be developed in the future, and it will likely also be possible to simulate the complete human body system and consider the treatment of tumors from this perspective. Thus, the refinement of the work will increase. As human error is unacceptable in this process, it is necessary to develop machines to conduct this work. However, at the same time, the replacement of manual labor by machines has its own problems, for example, clumps of cells or tiny tissues that may be present in the organoids, leading to uneven implantation. In the future, it may be necessary to develop more techniques such as fluid dynamics to make his distribution more uniform.

Application of high-throughput technology in drug screening

Pharmacovigilance screening and validation are essential components of high-throughput solutions. Cultured organoids offer a promising tool for drug screening, as they can recapitulate the in vivo TME, thereby enabling more accurate predictions of drug efficacy. An automated drug screening platform integrated with machine learning algorithms can rapidly and accurately detect the effects of drugs on tumor organoids. By comparing the physiological parameters of tumor organoids under different drug treatments, drugs with potential therapeutic efficacy can be initially identified. The existing applications and prospects of organoid modeling for peptide drug screening have been systematically described in previous reviews.⁵⁵ We have not discussed the traditional approach to organoid drug screening here, focusing only on applying high-throughput organoid drug screening.

Schuster et al.³¹ developed a microfluidic 3D organoid culture platform that can be automated, leveraging microfluidic technology and programming software to achieve dynamic, time-dependent drug delivery and combination therapy to organoids. Furthermore, the platform is coupled with an image analyzer to monitor organoid morphology and viability in real time. The feasibility of the microfluidic platform was validated through drug screening of patient-derived pancreatic tumor organoids, which demonstrated that combination and dynamic treatments were more effective in treating tumor organoids than were constant single-drug regimens.

Existing research has focused primarily on drug gradients or combinations of two drugs. However, Li et al.⁵⁶ developed a high-throughput, open-space, reusable multilayer microfluidic chip for organoid-based combination drug screening to identify increasingly complex drug combinations. This microfluidic chip features a programmable flow control system connected to a data collection tool. The chip consists of an inlet layer and multiple dispersion layers, enabling the simultaneous loading of different samples from the inlet layer, followed by the sample solution flow into the dispersion layer below. Individual micropumps control fluid flow in each channel, minimizing cell dilution and facilitating histological analysis via microscopy techniques.

Komen et al.⁵⁷ designed a microfluidic platform comprising a drug layer and a cell layer. The two layers are separated by a membrane, with two channels intersecting at a cross shape, allowing cells to be inoculated from the intersection. Upon completion, the bottom channel is closed, and two independent syringe pumps modulate the drug concentration in the drug layer, which then diffuses through the membrane into the lower space, thereby providing dynamic drug dosing. Zhai et al.⁵⁸ reported a digital microfluidic system featuring an innovative control structure and chip design. Their drug dispenser preloads the desired drug dose onto the chip, mixes it with a droplet of cell suspension to form a drug sieve model with varying drug concentrations, and finally delivers the mixture to the location of cultured cells. This technique was validated using two chemotherapeutic drugs. Compared with conventional 96-well plates, the system offers several advantages, including reduced cell and drug consumption and minimal spatial requirements. Within the organ-on-a-chip approach, Ryuji Morizane’s team has significantly contributed to renal disease research by developing an organoid model to improve organoid quality and disease model fidelity. Ryuji Morizane’s team reported a renal organoid that can be flow-cultivated on a microfluidic chip,⁵⁹ and their use of endogenous endothelial progenitors provides new ideas for research related to the construction of tissue-engineered blood vessels. Ryuji Morizane’s team added fluid flow to the organoid chip and used the model to test two new drugs, highlighting the great potential of organoid models in elucidating complex disease mechanisms for therapeutic testing and discovery.⁶⁰

Notably, drug screening is an area of substantial interest, and one key issue is the evaluation of the cell viability status after drug addition. The final evaluation of drug screening is the cellular response to the drug, and we often evaluate the state of the cells. The classical methods are MTT and CCK-8 assays based on cytotoxicity. As the technology has developed, assay sensitivity has also increased, and methods based on ATPase activity may be used. Commercialized kits are available for these methods; however, one disadvantage is that these methods are not performed in real time and do not reflect the state of the cells at the early stage of dosing; thus, new solutions, such as the C3 protein and ATP ratio, are needed. As research techniques advance, so do the tools available to researchers. It is now possible to transfect CFP-Asp-Glu-Val-Asp (DEVD)-YFP fusion proteins^61,62 into cells to determine caspase-3 expression levels and then use multiphoton microimaging⁶³ for data acquisition to obtain timed or delayed data on tumor sensitivity to drugs. In addition, Ryuji Morizane’s team utilized a genetically encoded fluorescent biosensor, PercevalHR, which senses the ATP:ADP ratio and thus reflects the state of cellular energy metabolism, for high-throughput drug screening in organoids. The application of this technology to kidney organoids, where drugs are evaluated for nephrotoxicity, suggests that it may be applicable to high-throughput drug screening in the future.⁶⁴ Notably, Raman spectroscopy can also be used to study drug metabolism in cancer cells.⁶⁵

Because of the vast number of compounds that exist in nature or are synthesized by humans, it is necessary to conduct experiments on a large scale to select the best drugs. Thus, the high-throughput application of drug screening has developed rapidly in recent years, and many pharmaceutical companies already have a complete set of equipment and methods for conducting high-throughput drug screening. Future high-throughput technology development may focus on cost reduction and refinement of the model to address the complex structure of drug molecules and enable the data to be combined with AI, big data analysis, and other technologies, comprehensively investigating drug mechanisms, efficacy, and safety. For targeted drugs, a more sophisticated method of assessing the activity of the drug is needed to evaluate its efficacy while also observing toxic side effects in humans. There are also drugs that need to be combined with other drugs or other therapies in order to be effective and that may require the development of a more comprehensive system for observing efficacy during the course of treatment or at a particular stage. This is likely to be a trend for the future, but at the same time, it will require a higher level of automation and economy.

Data acquisition and data analysis in organoid culture

Data acquisition

The organoid imaging and data analysis process is more complex than that of 2D cell culture, increasing the requirements for the imaging tools and analysis techniques used. To objectively present the phenotype and structure of organoids and realize a high-throughput organoid drug screening process, we need to combine automation with 3D imaging and machine learning approaches. Ryuji Morizane’s team⁶⁶ can now map the structure of kidney organoids via automated 3D imaging, thus overcoming the difficulties that the complex structure of organoids poses for high-throughput drug screening.

Cellular imaging is a crucial tool for understanding the culture status of tumor organoids and assessing drug efficacy. High-resolution microscopy and image analysis techniques enable rapid and accurate cell morphology, number, and viability detection during culture and drug screening. In our experimental protocol, live-cell fluorescence microscopy was employed to acquire organoid data. An inverted microscope equipped with an organoid platform on a panning stage was used for observation, with an enclosure set up to control the cellular environment of the organoids and maintain the necessary temperature, humidity, and CO2 gas composition for organoid cultivation.³¹ ImageXpress software was used to acquire images.⁵¹ The microscope-connected software enabled automatic image acquisition at different positions, z-stacks, and multiple color channels.

In addition, light-sheet microscopy has been employed to assess drug efficacy in tumor organoids by fusing multiview images, allowing for complete 3D tomography. This approach compensates for the low quality of images obtained at the deepest part of the 3D structural sample and enables fluorescent molecules to reach the core of the organoid, which is as large as 1 mm in diameter, thereby enabling the collection of relevant data from different regions.⁶⁷

Data analysis

A substantial quantity of data is generated during the culture process, including data on cell viability and physiological indicators. Data visualization techniques transform complex data into intuitive charts, facilitating easy understanding and analysis. These data provide valuable references for subsequent clinical trials and drug development. A microscope connected to a computer equipped with a MATLAB script has been used to analyze the fluorescence images to quantify apoptosis, providing the average fluorescence intensity of the desired fluorescent marker for each organoid and the corresponding image.³¹

AI analysis is a crucial tool for data analysis in organoid culture. By leveraging machine learning algorithms, AI analytics can analyze and predict experimental data for drug screening, enabling us to understand the mechanism of action of drugs on tumor organoids, predict drug efficacy and side effects, and optimize drug screening strategies. AI image analytics software (Aivia) offers a comprehensive platform for image visualization, analysis, and data combination from 2D to 5D and is capable of rapidly processing and reconstructing highly complex images in a few minutes with high quality. AI image analysis simultaneously integrates traditional image analysis and deep learning algorithms, opening new avenues for image analysis techniques by extracting specific data from digital microscope images through dedicated software. The following illustration (Figure 3) illustrates how the aforementioned high-throughput techniques can be utilized for tissue dissociation, organoid growth, drug screening, data analysis, and data acquisition.

Figure 3.

Overall process of performing tissue dissociation, organoid implantation, drug screening, data analysis, and data collection via high-throughput technology. (A) In the initial sample preparation step, patient-derived tumor tissues are processed via a tissue dissociator to obtain a single-cell suspension. (B) 3D printing is performed with microfluidic technology, and the resulting organoids are planted into 96-well plates or microfluidic chips. (C) During drug screening, drugs can be automatically administered via a drug printer or microfluidically delivered. (D) Image data are then acquired via a high-content imaging system, and data analysis is performed via AI analysis.

In the process of organoid culture, we found that the heterogeneity of cells is significant, including the determination of live and dead may show interference by many factors. So algorithms that exclude complex factors and differences in organoids formation between tissues and individual patients may need to be developed. With the development of imaging technology, there are difficulties in image processing algorithms for assessing cellular activity, and future image processing may require higher computation, so large servers are required to provide more powerful arithmetic. In addition, a computational model based on data optimization can be used to solve the problem of data analysis with relatively small energy consumption.

Summary and outlook

The tumor organoid culture system has a broad range of applications for recapitulating the TME in vitro. By mimicking the TME in vivo, this technology enables more accurate prediction of drug effects on tumors. Introducing high-throughput solutions facilitates the automation and scale-up of processing from culture to drug screening, improving culture efficiency and screening accuracy. Currently, organoid technology cannot fully recapitulate the TME. In addition, the lack of support from related industries hinders the widespread application of microfluidic technology, which cannot replace traditional organoid culture platforms.³¹ The chemical properties of naturally derived matrices, such as temperature sensitivity and polymerization at room temperature, which lead to the clogging of microfluidic channels, are also a challenge for current microfluidic and high-throughput technologies.⁹ Furthermore, the lack of cryogenic control of Matrigel loaded in microfluidics is a significant challenge in 3D printing technology. The limited availability of patient-derived tumor samples also hinders the evaluation of concentration-dependent assessments and combination therapies, highlighting the need for improved cell extraction efficiency from patient samples and standardized culture techniques.¹⁵ Organoids are currently in rapid development, and an increasing number of complex models, such as the organoid chip model made by Ryuji Morizane’s team, will be developed in the future to address the real-world situation of disease development.⁶⁶ Moreover, high-throughput solutions should be developed relatively rapidly. High-throughput experiments will also become more rapid, accurate, automated, and capable of handling large-scale experimental data. In the future, strengthening the integration of the tumor organoid culture system with high-throughput experiments, such as drug screening, is essential to identify more effective tumor treatments and drugs.

Author contributions

J.Z.: Investigation, Methodology, Writing—original draft. Y.F.: Project administration, Writing—review & editing, Visualization. R.W.: Funding acquisition, Supervision, Formal Analysis. S.L.: Conceptualization, Supervision, Project administration.

Funding

The authors would like to express their gratitude to the High-level Talent Innovation Support Program of Dalian Science and Technology Bureau (grant number 2021RD02) and the National Natural Science Foundation of China (grant number 82172822) for their financial support under grant number.

Conflicts of interest

There are no conflicts to declare.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.