As our understanding of cancer behavior and drug resistance improves, it has become increasingly clear that there is a need for better diagnosis, characterization, and therapeutic targeting of cancer subtypes. During the last twenty years, publications in the field of personalized medicine (also known as precision medicine) have been on the rise1,2. Indeed, precision medicine is defined as the practice of tailoring medical treatment based on the individual patient's disease characteristics3. This follows our current understanding that there is no one-drug-fits-all approach, which has become particularly clear and pertinent in cancer therapy. Currently, however, the lack of clinically relevant precision medicine approaches for most cancer patients means that the process of selecting the most suitable treatment strategy is by trial and error. This results in significant delays and unnecessary discomfort with diminished quality of life and shortened survival. Importantly, a patient with diminished quality of life after several ineffective chemotherapy regimens may opt to switch to palliative treatment4. With a suitable, personalized screening method, drug selection can be performed with higher precision and without delay. While it is possible to perform genetic and proteomic sequencing on cancer cells5, such molecular data is overall poorly predictive of patient outcome and often has little clinical value6. Within drug development, having a precision medicine approach as early as possible greatly improves success rates at the various milestones of the drug development pipeline. Indeed, understanding early on which drug candidates have the broadest inter-individual efficacy profile, or what cancer subtypes are particularly sensitive to each drug candidate, de-risks and improves the efficiency of late preclinical tests and clinical trials7. With rising pharmaceutical R&D costs8, growing political pressure to lower drug prices9,10, and advances in AI-driven target validation11, the need for a large-scale, efficient, and precision-medicine-centered screening method in drug development has never been greater.
Here, zebrafish enter the picture, specifically zebrafish larvae at 48 hour post-fertilization (hpf). Tumor xenograft models based on zebrafish larvae as hosts constitute a rapid in vivo screening platform to forecast the clinical responses to chemotherapy6,12,13,14. Zebrafish tumor xenograft models have emerged as a powerful alternative to both molecular precision medicine (i.e., genomics or molecular pathology approaches to guide choice of treatment for cancer patients), as well as organoid and mouse xenograft studies for preclinical drug development15,16. With zebrafish tumor xenografts, fluorescent tumor cells (genetically or chemically labelled) are engrafted into zebrafish larvae with high efficiency13. Drugs or drug candidates with a relatively fast mode of action (e.g., such that target tumor cell viability directly or indirectly via, for example, activation of T-cell cytotoxicity) can then be screened in just three days, leading to fast, functional readouts of anti-cancer and anti-metastatic drug efficacy. Following this, an informed selection of the therapy that shows a good efficacy can be made, with studies showing that this should translate to a strong treatment outcome for the patient or a chance for success in a drug development project respectively12,17,18,19. Moreover, this screening process is highly relevant for selecting compounds to be moved forward in the drug development pipeline.
However, to become a viable diagnostic tool in clinical settings and to enable large-scale drug screens within cancer research and drug development, the zebrafish tumor xenograft (ZTX) model must be fast, reliable, reproducible, and cost-effective. Automation is essential for achieving these goals, as it minimizes operator variability while increasing throughput and enhancing precision in tumor cell injection and analysis. By standardizing these procedures, we can improve data reproducibility and eliminate human error, while the automated workflows present opportunities for high-throughput screening and making large-scale studies feasible20,21. Recently, we have developed an automated injector that injects liquids, particles, or cells of various origins at three different anatomical sites with equal or higher speed, precision, and reproducibility than trained zebrafish researchers22. Here, we present a protocol for the automated injection of tumor cells into zebrafish larvae in a precise and controlled manner. Using advanced image recognition, zebrafish larvae are located on the injecting plate, orientation is determined, and the target site for the injection is identified. The needle is then guided automatically, and a successful puncture is determined by the software. With minimal training, operators can use this system to perform complex injections, for example, creating microtumors in live zebrafish larvae, thus enabling efficient screening of treatment strategies.