BackgroundImmunotherapy for malignant tumors has made great progress, but many patients do not benefit from it. The complex intratumoral heterogeneity (ITH) hindered the in-depth exploration of immunotherapy. Conventional bulk sequencing has masked intratumor complexity, preventing a more detailed discovery of the impact of ITH on treatment efficacy. Hence, we initiated this study to explore ITH at the multi-omics spatial level and to seek prognostic biomarkers of immunotherapy efficacy considering the presence of ITH.MethodsUsing the segmentation strategy of digital spatial profiling (DSP), we obtained differential information on tumor and stromal regions at the proteomic and transcriptomic levels. Based on the consideration of ITH, signatures constructed by candidate proteins in different regions were used to predict the efficacy of immunotherapy.ResultsEighteen patients treated with a bispecific antibody (bsAb)-KN046 were enrolled in this study. The tumor and stromal areas of the same samples exhibited distinct features. Signatures consisting of 11 and 18 differentially expressed DSP markers from the tumor and stromal areas, respectively, were associated with treatment response. Furthermore, the spatially resolved signature identified from the stromal areas showed greater predictive power for bsAb immunotherapy response (area under the curve=0.838). Subsequently, our stromal signature was validated in an independent cohort of patients with non-small cell lung cancer undergoing immunotherapy.ConclusionWe deciphered ITH at the spatial level and demonstrated for the first time that genetic information in the stromal region can better predict the efficacy of bsAb treatment.Trial registration numberNCT03838848.

Epithelial-mesenchymal transition (EMT)-related long non-coding RNAs (lncRNAs) may be exploited as potential therapeutic targets in gliomas. However, the prognostic value of EMT-related lncRNAs in gliomas is unclear. We obtained lncRNAs from The Cancer Genome Atlas and constructed EMT-related lncRNA co-expression networks to identify EMT-related lncRNAs. The Chinese Glioma Genome Atlas (CGGA) was used for validation. Gene set enrichment and principal component analyses were used for functional annotation. The EMT-lncRNA co-expression networks were constructed. A real-time quantitative polymerase chain reaction assay was performed to validate the bioinformatics results. A nine-EMTrelated lncRNAs (HAR1A, LINC00641, LINC00900, MIR210HG, MIR22HG, PVT1, SLC25A21-AS1, SNAI3-AS1, and SNHG18) signature was identified in patients with glioma. Patients in the low-risk group had a longer overall survival (OS) than those in the high-risk group (P < 0.0001). Additionally, patients in the high-risk group showed no deletion of chromosomal arms 1p and/or 19q, isocitrate dehydrogenase wild type, and higher World Health Organization grade. Moreover, the signature was identified as an independent factor and was significantly associated with OS (P = 0.041, hazard ratio = 1.806). These findings were further validated using the CGGA dataset. The low-and high-risk groups showed different EMT statuses based on principal component analysis. To study the regulatory function of lncRNAs, a lncRNA-mediated ceRNA network was constructed, which showed that complex interactions of lncRNA-miRNA-mRNA may be a potential cause of EMT progression in gliomas. This study showed that the nine-EMT-related lncRNA signature has a prognostic value in gliomas.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.