用于研究TME并评估抗癌药物的肺癌原位移植模型的开发及其临床前研究应用

Heterotopic subcutaneous (SC) tumor transplantation models are widely used in cancer research owing to their technical simplicity and ease of tumor monitoring. However, these models fail to recapitulate the native tumor microenvironment (TME). To overcome this limitation, we developed a bronchoscopy-based orthotopic (ORT) lung cancer model in mice and compared its TME and drug response with those of the SC model.Three luciferase-expressing lung adenocarcinoma models were used: A549-Luc, RERF-LC-KJ-Luc, and LC-58-Luc (a cell line derived from a patient-derived xenograft, CIEM-established). In the ORT model, 3% lauryl-10 solution was administered via bronchoscopy to the left lung lobule three days prior to transplantation, followed by instillation of 1×10⁵ cells in 20 µL medium containing 10% Matrigel. SC tumors were established by subcutaneous injection of the same cells (1×10⁵ cells in 100 µL with 50% Matrigel). Tumor growth was monitored via bioluminescence (ORT) or caliper measurements (SC). Mice received carboplatin (CBDCA, 50 mg/kg, i.p., 3 doses at 5-day intervals) and paclitaxel (PTX, 15 mg/kg, i.p., 5 doses at 3-day intervals). VEGF inhibitor combination chemotherapy was also evaluated. RNA sequencing was performed using 150 bp paired-end NovaSeq, with the reads processed using TrimGalore, Xenogsort, STAR, and RSEM. The human-mouse interactome was analyzed using MultiNicheNet. CBDCA and PTX showed significant antitumor effects in both SC and ORT models. In the SC tumors, combination chemotherapy with the VEGF inhibitor significantly reduced the number of mCD31-positive endothelial cells, which is consistent with the anti-angiogenic response. By contrast, no such effect was observed in the ORT tumors, highlighting the TME-dependent differences in vascular responses. RNA sequencing showed ORT-specific upregulation of surfactant genes ( SFTPA1 , SFTPB , and SFTPC ), indicating lung-specific adaptation. MultiNicheNet analysis revealed ORT-enriched ligand-receptor pairs linked to perineural invasion and adhesion, underscoring enhanced host-tumor crosstalk. This ORT model better mimics clinical TME dynamics and enables a more translationally relevant evaluation of anticancer agents. Bronchoscopy-based ORT lung cancer models recapitulate physiologically relevant tumor-host interactions and TME-dependent drug responses. By incorporating VEGF inhibitor response and transcriptomic profiling, this model provides a better preclinical platform for evaluating antitumor agents and improving translational predictability compared with SC models.