Transcriptome dysregulation in hyper-progressive disease samples with immune checkpoint blockade

Dan Xue, Tengteng Zhu, Hongguang Lin, Peilin Guo, Mengling Li, Mei'e Yu, Fan Yang, Sheng Yang, Xiangqi ChenChinese 2023 · DOI · OpenAlex

Abstract

To the Editor: Tumor immunotherapy has made rapid progress in recent years. However, immune checkpoint blockade (ICB) therapy may cause clinical symptoms of hyper-progressive disease (HPD), especially for patients with alterations or amplifications in driver genes, such as mouse double minute 2 (MDM2), epidermal growth factor receptor(EGFR), and fibroblast growth factor 4 (FGF4). Recently, Kamada et al[1] investigated the molecular mechanism of HPD to explore whether immune checkpoint inhibition caused HPD in patients in clinical trials. Tumor invasive hyperplasia (Ki67+) effector regulatory T cells (eTregs) in HPD patients were found to be increased after PD-1 monoclonal antibody treatment. Other studies have also found that immune cells in the tumor microenvironment of patients with HPD show obvious changes before and after treatment.[2] Tumor microenvironment changes in HPD imply the potential presence of unknown gene interactions that promote rapid growth of tumor cells after cancer immunotherapy. Although the effects of immunotherapy drugs on changes in tumor mutations were reviewed previously, transcriptional changes in HPD samples have not yet been elucidated. In this study, we collected clinical samples of HPD from the Gene Expression Omnibus database (GSE126044)[3] and analyzed the changes in transcriptome expression and transcript structure corresponding to the pathological characteristics of HPD. All statistical analyses were conducted using R 3.6.2 (www.r-project.org). One-way analysis of variance (ANOVA) was used to test differences of continuous variables between groups and Chi-square test for categorical variables. The Kaplan–Meier method was used for survival analysis. Detailed bioinformatics steps were listed in supplementary methods file, https://links.lww.com/CM9/B650. Functional validation of splice factor over-expression was conducted via transwell invasion assay of lung cancer cell lines [Supplementary File, https://links.lww.com/CM9/B650]. According to the survival data of each patient and the definition of HPD,[2] four non-drug-responding samples were assigned to the HPD group [Supplementary Table 1, https://links.lww.com/CM9/B651]. The median progression-free survival time of HPD patients was 0.6 months, and the median overall survival time was 1 month. No sex or age difference was found among the HPD group, non-responding and non-HPD (NR) group and drug-responding (DR) group [Supplementary Table 2, https://links.lww.com/CM9/B651]. We estimated immune cell abundance in tumor tissues according to the immune-marker expression pattern. The median immune score of HPD samples was found to be lower than that of DR samples but higher than that of NR samples [Figure 1A]. The cell content in HPD samples had a more similar pattern to that in the DR group than the NR group. Differential expression analysis showed that HPD samples had unique expression characteristics [Supplementary Figure 1, https://links.lww.com/CM9/B651]. We compared the differences in gene expression between the HPD group and the DR group. A total of 2210 genes were found to be significantly altered [Supplementary Figure 1A, https://links.lww.com/CM9/B651]. Genes highly expressed in the major histocompatibility complex-II (MHC-II) pathway, such as legumain (LGMN) and cathepsin B (CTSB), were significantly enriched in the HPD group [Figure 1B]. The intestinal immune network for immunoglobulin A production was also enriched. Notably, the gene expression of the MHC-I pathway did not change significantly [Figure 1B]. MDM2, EGFR, and FGF4 showed no expression difference in the HPD group compared with the DR group.Figure 1: (A) Comparison of immune cell score in HPD and ICB drug responding (DR) and non-responding non-HPD samples (NR); (B) HPD upregulated genes (red-colored) and unchanged genes (green-colored) in the antigen processing and presentation pathways, hsa 04612; (C) gene pathway enrichment of differentially spliced genes for HPD compared with ICB response genes; (D) overexpression of SRSF2 in A549 and H1650 cell lines. Left: The fold change of the relative invasion. * P <0.05, † P <0.01 vs. control. Right: Representative images of invasion ability test examined by transwell invasion assay. Scale bar was 50 μm in HE staining. CIIV: Class II vesicles; E.R.: Endoplasmic reticulum; HE: Hematoxylin-eosin; HPD: Hyper-progressive disease; ICB: Immune checkpoint blockade; MHC: Major histocompatibility complex; MIIC: Minnesota Immunization Information Connection; NC: Control cell line; NK: Natural killer.Changes in immune cell infiltration and antigen delivery pathways imply that the endogenous tumor mutation-related antigen may have been shifted in HPD samples. We evaluated sequence changes in which base changes, indels and alternative splicing events (ASE) presentable by MHC pathways. The single-nucleotide variant and insertion–deletion mutations did not show significant changes in HPD [Supplementary Table 3, https://links.lww.com/CM9/B651]. ASE, especially intron retention (IR) changes, were clustered together in HPD samples [Supplementary Figure 2, https://links.lww.com/CM9/B651]. The functional enrichment of ASE-genes showed that the pathways related to tumor immune cells changed significantly [Figure 1C and Supplementary Figure 3, https://links.lww.com/CM9/B651]. The most significant gene of IR is Nod-like receptor (NLR) family caspase recruitment domain (CARD) containing 5 (NLRC5), a regulator of cancer immune escape. Exon usage of NLRC5 in HPD samples was significantly lower than that in DR samples [Supplementary Figure 4, https://links.lww.com/CM9/B651]. Additionally, karyopherin subunit alpha 3 (KPNA3) showed a significant increase in exon usage. The predicted enrichment of binding proteins for alternative splicing sites at the transcriptome level showed that serine and arginine rich splicing factor 2 (SRSF2) was the most significant binding gene in HPD samples [Supplementary Figure 5, https://links.lww.com/CM9/B651]. This was consistent with the increased expression level of SRSF2 in HPD samples. We validated the effect of SRSF2 on transcriptome selective splicing using the lung cancer cell lines A549 and H1650. After the overexpression of the SRSF2 gene, the transcriptome level of splicing events increased significantly. We also conducted SRSF2 cell invasion experiments and found that tumor cells overexpressing SRSF2 had a significantly increased invasion ability [Figure 1D]. The validation of clinical data in lung cancer samples in TCGA showed that samples with high SRSF2 expression had significantly reduced overall survival [Supplementary Figure 6, https://links.lww.com/CM9/B651]. In addition, systematic screening in knockdown experiments from the DepMap database showed SRSF2 was likely to be essential in diverse human cancer cell lines [Supplementary Figure 7, https://links.lww.com/CM9/B651]. These results suggest that SRSF2 might be an important factor in promoting the hyper-progression of HPD tumor cells. In this study, we analyzed the transcriptome characteristics of HPD samples from patients with non-small cell lung cancer. Compared with the results of a previous analysis of the sample dataset, the MHC-II pathway was specifically highly expressed in the HPD group. Tumor-specific MHC-II expression (tsMHC-II) can increase the immune system's recognition of tumors and thus play an important role in immunotherapy.[4] Some studies in multiple cancer types have found an association between tsMHC-II and a good prognosis.[4] MHC class II-restricted neoantigens were reported to have a key function in the anti-tumor response independent of the function of MHC class I-restricted neoantigens.[5] Consistent with the upregulation of the MHC-II pathway, immune cells show a unique composition profile in HPD samples. Our results show that alternative splicing is an important factor in HPD occurrence. For instance, we found lower exon usage of NLRC5, a transcriptional coactivator of genes involved in the MHC class I antigen-presentation pathway. Genetic and epigenetic alterations in NLRC5, such as copy number loss, DNA methylation, or somatic mutations, reduce MHC class I expression or activity, leading to impaired CD8+ T cell activation. We identified splicing factors at splicing sites in the whole transcriptome, and our results showed the key role of SRSF2. Recent studies have reported the mechanisms of other members in the SRSF family during cancer development. Our functional verification revealed that dysregulated SRSF2-related splicing might be an important feature in the occurrence of HPD in patients undergoing ICB therapy. We recognize that major shortcoming of this study is the limited number of HPD samples. We will continue to collect clinical samples to validate the HPD transcriptome mechanism in future studies. To conclude, our results showed that the tumor neoantigen delivery MHC-II pathway was upregulated, and the spliceosome factor SRSF2 was involved in changes in HPD-specific splicing. These results may be helpful for future clinical practice of combination therapy in patients with ICB-induced HPD. Acknowledgments The authors acknowledge and appreciate two anonymous reviewers for their constructive feedback that improved this paper. We also thank Dr. Melissa Crawford, Dr. Yunjing Wang, and Dr. Jingxiang Xu for their comments and editing the English text of a draft of this manuscript. Funding This work was supported by the Joint Funds for the Innovation of Science and Technology, Fujian Province, China (Nos. 2016Y9028, 2019Y9055 and 2020Y9090), Fujian provincial health technology project (No. 2020CXB016) and the Natural Science Foundation of Fujian Province (No. 2021J01772). Conflicts of interest None.
BibTeX
@inproceedings{xue2023transcriptome,
  author = {Dan Xue and Tengteng Zhu and Hongguang Lin and Peilin Guo and Mengling Li and Mei'e Yu and Fan Yang and Sheng Yang and Xiangqi Chen},
  title = {{Transcriptome dysregulation in hyper-progressive disease samples with immune checkpoint blockade}},
  booktitle = {Chinese},
  year = {2023},
  doi = {10.1097/cm9.0000000000002550},
}

Loading related content...

Figure