在抗PD1治疗的患者来源类器官型肿瘤球体（PDOTS）中对免疫激活与肿瘤反应进行纵向图谱绘制

Nivolumab is an anti-PD-1 immune checkpoint inhibitor approved for colorectal cancer (CRC) with an Microsatellite Instability-High (MSI-Hi) profile, yet patient-to-patient variability and the short-term dynamics of immune and tumor signaling is not well characterized. Patient-derived organotypic tumor spheroids (PDOTS) preserve the multicellular architecture and immune microenvironment of individual tumors, providing an ex vivo platform to monitor these dynamics under controlled conditions. In this study, we used CRC PDOTS from eight patients to characterize day-scale effects of nivolumab on tumor burden and transcriptional programs over a 72 hour peroid. Freshly resected tumors were processed into PDOTS, loaded into microfluidic devices, and treated with nivolumab (100 μg/mL) or a media control. The tumor response to nivolumab was quantified on days 1, 2, and 3 using Hoechst/propidium iodide imaging and automated analysis of live tumor area and the percent tumor killing relative to matched controls. In parallel, RNA was isolated from treated and control PDOTS at each timepoint and profiled using the NanoString IO360 panel to capture coordinated changes in tumor-associated transcripts and immune- and signaling-related gene sets. Across the cohort, nivolumab induced distinct temporal patterns of transcriptional change rather than a uniform on/off signature. A subset of PDOTS showed rapid day-1 increases in immune- and inflammation-related transcriptional programs that remained elevated through 72 hours. A second subset appeared transcriptionally cold at day 1 but developed coordinated shifts in immune and signaling pathways on days 2-3, with preferential upregulation of genes associated with interferon signaling, T-cell engagement, and leukocyte trafficking. A third group displayed minimal changes in these programs over the full 72-hour window. Notably, many PDOTS with strong day-3 pathway modulation showed little or no change from day 1, underscoring the risk of underestimating drug-associated biology from single early timepoints. These temporal patterns of immune activation loosely preceded tumor response. These findings demonstrate that longitudinal profiling in CRC PDOTS reveals discrete patterns of nivolumab-induced immune and tumor signaling dynamics that are not apparent from single terminal measurements, and support the use of time-resolved ex vivo platforms to refine mechanistic studies and biomarker development for immune checkpoint blockade.