Dongdong Zhan

Background: Molecular profiling is increasingly important for identifying candidate therapeutic targets in precision oncology. For lung adenocarcinoma brain metastases, individualized treatment selection remains limited by lesion-level heterogeneity, primary-metastatic molecular discordance, and the absence of actionable genomic alterations in some patients. Lesion-level tumor proteome profiling may uncover patient-specific therapeutic vulnerabilities, but these candidates require functional validation before they can inform prioritization of candidate therapies for preclinical evaluation. Methods: In this proof-of-concept study, we established a proteomics-guided functional precision oncology workflow using surgically resected brain metastases from two patients with lung adenocarcinoma. Lesion-level tumor proteomes were quantified by liquid chromatography–tandem mass spectrometry (LC–MS/MS) and benchmarked against a pan-cancer non-malignant reference proteome to identify patient-specific overexpressed proteins. Candidate targets were prioritized based on protein abundance, druggability, and available clinical evidence, and matched patient-derived xenograft (PDX) models were established for in vivo functional validation. Results: Reference-based tumor proteome analysis identified patient-specific actionable protein outliers and generated target-drug hypotheses for functional testing. In Patient LUAD-BM-01, whose tumor lacked actionable genomic alterations, MAP2K1 was markedly overexpressed, and trametinib suppressed tumor growth in matched PDX models across two host systems but was associated with lower body weight. In Patient LUAD-BM-02, ERBB2 protein overexpression was observed together with ERBB2 alterations, and trastuzumab deruxtecan (T-DXd; DS-8201) induced marked tumor inhibition in the matched PDX model without a statistically significant body-weight difference. Trametinib achieved tumor growth inhibition rates of 70.45% and 81.98% in BALB/c nude and NOD-SCID PDX models, respectively, whereas DS-8201 achieved a tumor growth inhibition rate of 93.56% in the matched BALB/c nude PDX model. Conclusions: In two illustrative cases, lesion-level tumor proteome profiling generated candidate target–drug hypotheses, and the corresponding agents showed antitumor activity in matched PDX models. This omics-informed functional precision oncology workflow may support prioritization of candidate therapies for preclinical evaluation for selected patients with lung adenocarcinoma brain metastases.

Chromosomal instability (CIN) covers approximately 65 to 70% of colorectal cancer patients and plays an essential role in cancer progression. However, the molecular features and therapeutic strategies related to those patients are still controversial. R-loop binding proteins (RLBPs) exert significant roles in transcription and replication. Here, integrative colorectal cancer proteogenomic analysis identified two RLBPs subtypes correlated with distinct prognoses. Cluster I (CI), represented by high expression of RLBPs, was associated with the CIN phenotype. While Cluster II (CII) with the worst prognosis and low expression of RLBPs was composed of a high percentage of patients with mucinous adenocarcinoma or right-sided colon cancer. The molecular feature analysis revealed that the active RNA processing, ribosome synthesis, and aberrant DNA damage repair were shown in CI, a high inflammatory signaling pathway, and lymphocyte infiltration was enriched in CII. In addition, we revealed 42 tumor-associated RLBPs proteins. The CI with high expression of tumor-associated proteins was sensitive to drugs targeting genome integrity and EGFR in both cell and organoid models. Thus, our study unveils a significant molecular association of the CIN phenotype with RLBPs, and also provides a powerful resource for further functional exploration of RLBPs in cancer progression and therapeutic application.

Gastric cancer (GC) is one of the most common malignant tumors with a high mortality rate worldwide and lacks effective methods for prognosis prediction. Postoperative adjuvant chemotherapy is the first-line treatment for advanced gastric cancer, but only a subgroup of patients benefits from it. Here, we used 833 formalin-fixed, paraffin-embedded resected tumor samples from patients with TNM stage II/III GC and established a proteomic subtyping workflow using 100 deep-learned features. Two proteomic subtypes (S-I and S-II) with overall survival differences were identified. S-I has a better survival rate and is sensitive to chemotherapy. Patients in the S-I who received adjuvant chemotherapy had a significant improvement in the 5-year overall survival rate compared with patients who received surgery alone (65.3% vs 52.6%; log-rank P = 0.014), but no improvement was observed in the S-II (54% vs 51%; log-rank P = 0.96). These results were verified in an independent validation set. Furthermore, we also evaluated the superiority and scalability of the deep learning-based workflow in cancer molecular subtyping, exhibiting its great utility and potential in prognosis prediction and therapeutic decision-making.