Over the years, various studies have shown an increased exosome concentration during cancer, and most exosomes in this condition come from cancer cells to support its function. Dissecting the cargo of cancer-derived clinically relevant exosomes can reveal key elements responsible for establishing tissue-tissue communication, tumor development, factors responsible for priming the secondary sites to receive metastasized cells and identifying biomarkers for early diagnosis and prognosis or predictive markers for precision medicine. Thus, unbiased analysis of the exosomal whole transcriptome, without any inherent prejudice of methodologies such as size selection, is necessary to know the overall landscape of RNA species packaged in such cancer-derived exosomes and their involvement in cancer. In this study, a prespecified population of lung cancer cell-derived exosomes from clinical plasma samples was enriched utilizing a novel immunomagnetic capture approach that allowed us to analyze the abundance of different RNA species packaged into the lung cancer cell-derived exosomes and their differential expression in cancer vs. healthy controls. Both protein-coding and noncoding RNA transferred through the exosomes can affect cancer development and progression directly or indirectly by establishing interaction with the cancer microenvironment. Noncoding RNA such as lncRNA comprised the most abundant RNA species detected in these samples, followed by protein-coding RNA and other noncoding RNA. It could be due to the absence of a size selection step in the whole transcriptome library preparation. For each sample in the study, an average of 54.4 million reads were generated. Extracellular vesicles are enriched with rRNA fragments, which also form a larger cellular RNA composition (80–90%) where the cargo of these vesicles is packaged. The rRNA percentage composition in EVs could reach up to 97%, as described in Supplementary Table 11. Contrary to cell RNA, many studies involving whole transcriptome sequencing of EVs did not perform rRNA depletion (as described in Supplementary Table 11), essentially because the current rRNA depletion kits require a high starting RNA quantity (> 100ng) which is far more than the RNA content available in typical clinical patient-derived samples (< 1ng). Even performing the rRNA depletion could further reduce the RNA quantity for downstream processing. Sometimes, even the kit chemistry or design of probes is not efficient enough to remove fragmented rRNA present in EVs; for example, Jenjaroenpun et al. reported that the short probes size of RiboMinus Eukaryote Kit could not efficiently target the exosomal rRNA fragments to deplete it (Jenjaroenpun et al. 2013). As listed in Supplementary Table 11, EV RNA sequencing studies suggest that the type of library preparation also has a role in circumventing rRNA detection. The size selection step incorporated in small RNA library preparation enriches the small RNA, avoiding rRNA due to the difference in transcript length. However, it may provide a biased view of EV transcriptome by excluding a substantial protein-coding and long noncoding RNA population. Therefore, more than 50 million reads per sample were generated for this study to avoid the bias that the high rRNA reads could bring in the comparative analysis of other RNA biotypes in the samples.
The most abundant protein-coding RNA and lncRNA packaged in the lung cancer cell-derived exosomes are listed in Fig. 4B and Fig. 5B. Many of the abundantly packaged lncRNA found here have been previously identified as biomarkers in cancer diagnosis, prognosis, or predictive markers, as described in Supplementary Table 4. Several reports have shown that exosomes transmit specific lncRNAs that lead to the proliferation of NSCLC and other cancers (Li et al. 2019; Zang et al. 2020; Zhang et al. 2021). The expression of tumor-related exosomal lncRNAs has been correlated with the disease severity in patients. Thus, exosomal lncRNAs are an emerging class of cancer biomarkers. Some lncRNAs detected here have not been described previously in lung cancer, but their involvement was described in other cancers. For example, the role of lncRNA LRRC75A-AS1 has been identified as a ceRNA that facilitates cell proliferation and invasion in triple-negative breast cancer (Li et al. 2020); in contrast, in colorectal cancer, it acts as an anti-oncogene that inhibits cell proliferation and migration (Chen, Lan, et al. 2019). This lncRNA also targets the miR-199b-5p/PDC4 axis to repress multiple myeloma (Pang et al. 2020). C6orf48 or SNHG32 is a relatively new noncoding; this primate-specific gene has been identified in the glioblastoma and prostate cancer network. SNHG32 has been associated with several ribosomal proteins that interact with lncRNA GAS5 (oncogene), suggesting its role in coordinating translation in response to GAS5-induced growth arrest in colon cancer (Maertens et al. 2020). Exosomal GAS5 has a diagnostic value in NSCLC, and it can distinguish between early-stage NSCLC patients and healthy controls with an AUC of 0.822 (Li et al. 2019). However, in this study, we did not notice a difference in the expression pattern of GAS5 between patient and control. LINC01506 is another abundant lncRNA associated with immune system regulation and responses, and it has been linked to better overall survival in lung adenocarcinoma (LUAD) (Salavaty, Rezvani, and Najafi 2019).
Protein-coding RNA is the next abundant RNA species packaged into the exosomes, followed by lncRNA. These lung-derived exosomes were enriched with protein-coding genes involved in ribosomal subunit assembly and protein synthesis, ribosome biogenesis, translational elongation, alpha, and beta-globin synthesis, consistent with some previous reports. For example, the presence of elongation and translation factors has been demonstrated in brain tumor-derived exosomes (Graner et al. 2009), and mRNA associated with many ribosomal proteins has been found in mouse mast cell line-derived exosomes (Valadi et al. 2007). It has been hypothesized that these exosomal cargo mRNAs are translated into proteins to support the ribosomal functions for ensuring efficient translation of other mRNAs packaged inside exosomes when the exosomal content is released into the cellular compartment of recipient cells (Ratajczak et al. 2006; Valadi et al. 2007; Jenjaroenpun et al. 2013).
The functional analysis of the DEGs identified between NSCLC patients and healthy controls revealed the enriched biological processes, cellular components, molecular functions, and signalling pathways associated with the overall DEGs enriched in the lung-derived exosomes of the patient vs. healthy controls, as shown in Fig. 6. The enriched terms were involved in the biological processes such as I) regulation of transmembrane ion transport- ion channels form a major class of membrane proteins that play critical roles in signalling with adjacent cells and extracellular events. Dysregulated ion channels can play a causative role in many diseases, including cancer; for example, an in vivo experiment on mice showed that the ion channel inhibitors diminish tumor growth. II) Wnt signalling is reported to be among the main signalling pathways that maintain lung homeostasis, and any aberration in this pathway can cause several lung diseases. Wnt signalling plays an important role in lung carcinogenesis. III) cAMP catabolic processes are also highly enriched in these exosomes, which is in line with the previous reports that claim a significant upregulation of CREB (cAMP-responsive element-binding protein) expression and phosphorylation in NSCLC tumor tissues compared to the normal adjacent tissues that have been associated with poor survival of patients. cAMP/PKA signalling pathways have also been associated with cellular adaptation to hypoxia by hypoxia-induced upregulation of adenylyl cyclases, enzymes involved in cAMP production. The enriched genes of the exosomes were associated with the following cellular components cell adhesion, plasma membrane, and cell surface, which are the cellular components involved in exosome biogenesis.