LncRNAs are increasingly considered critical regulators of many cellular functions. In intestinal tissue, they modulate several signaling pathways that are crucial for maintaining its homeostasis [21]. Conversely, their dysregulation in cancer can alter these signaling cascades and allow malignant cells to proliferate and spread [12, 22]. Expression profiling of lncRNAs can identify potential targets that can be used for early disease detection. Despite the increased interest in lncRNA identification, their use as disease biomarkers remains largely unexplored. Importantly, RNA sequencing of EV content offers the possibility of developing biomarkers derived non-invasively from blood serum or plasma.

Our comprehensive study presents significant insights into the roles of lncRNAs from sEVs in CRC. This research was conducted in two phases: an exploratory phase that focused on RNA sequencing and a subsequent validation phase involving a larger cohort for further analysis of identified lncRNAs. However, the process of purifying and characterizing sEVs was an important aspect of our research. For sEV characterization, we used DLS and TEM to determine the size and concentration of EV fractions obtained by SEC from blood serum. Our analyses confirmed the presence of sEVs in samples from both CRC patients and healthy controls. A slightly elevated concentration of sEVs was observed in the samples from CRC patients, which could support the hypothesis that sEVs play a significant role in the pathogenesis of tumor development. Additionally, we identified a marginal subset of particles exhibiting larger sizes, suggesting slight heterogeneity in the vesicle population. Further confirmation of sEVs in our samples was performed by Western blot, which successfully detected key protein markers characteristic of these vesicles. Interestingly, in samples from CRC patients, there was an enhanced signal for CD81, indicative of a potentially higher abundance of sEVs in CRC patients compared to healthy controls. This observation aligns with findings from previous research [23,24,25,26]. In the study by Ricklefs et al., the authors demonstrated that in cancer tissues and cell lines, CD81 was significantly upregulated and associated with decreased overall survival. This pattern of increased protein marker expression extends beyond CD81, as shown by Tian et al., who reported higher abundance levels of CD63 in EVs from metastatic breast cancer patients compared to non-metastatic breast cancer patients and healthy donors. Additionally, Odaka and colleagues found that serum levels of CD63-positive EVs were significantly higher in pancreatic ductal adenocarcinoma patients compared to healthy controls.

The detection of a weakly positive calnexin signal in our SEC-isolated samples suggested the possible co-isolation of a different EV subtype, with a different size or distinct biogenesis pathway. To obtain a pure fraction of sEVs, a combination of different approaches is recommended; however, it is not feasible without high losses of vesicles.

Our exploratory cohort aimed to identify the lncRNA profiles in sEVs from blood serum of 76 CRC patients and 29 healthy controls. First, to perform the downstream analysis, we isolated RNA from the obtained sEV samples. However, RNA concentrations were below the detection limit of conventional techniques, so we employed vacuum evaporation to concentrate the samples. Despite lower RNA input, we were able to enhance RNA concentration and successfully prepare sequencing libraries. Using a high-throughput RNA sequencing approach, we detected differences in the sEV levels of 460 genes, which included mainly protein coding genes, lncRNAs, and pseudogenes. This differential analysis was statistically significant (P < 0.01), with a majority of these genes showing higher abundance in sEVs from CRC patients. Notably, the analysis revealed that about 20% of genes were lncRNAs, specifically differentiating between patients and healthy controls. Further statistical analysis highlighted the most significantly dysregulated lncRNAs, revealing tumor-specific lncRNAs not yet described in the context of CRC. We have also performed additional bioinformatic analyses related to stage and grade of CRC patients, the results are included in the supplementary data section (Tables S4–S12).

Next, RT-qPCR was used for validation of lncRNAs in larger study cohort of 159 CRC patients and 138 healthy controls. Of the top 20 lncRNAs from exploratory phase, 11 were selected for the validation phase of the study. However, quantifying these lncRNAs through RT-qPCR proved challenging due to their low concentration in the sEV samples. To overcome this, we prepared sample pools based on similar clinicopathological data and concentrated the RNA before cDNA preamplification and qPCR validation. These steps including preamplification enabled the measurement of previously undetectable molecules. However, it is important to note that while preamplification increases the detectability of low-abundance transcripts, it may also introduce artifacts in the amplification process. The RT-qPCR analysis confirmed the upregulation of three lncRNAs (NALT1, AL096828, and LINC01637) in CRC patients, which was in agreement with our sequencing data. Additionally, the RT-qPCR results also revealed elevated levels of AC055788 and AC016933, which were not identified as upregulated in the RNA sequencing analysis. Certain methodological factors, particularly the limited volume of blood serum used for RNA isolation from sEVs and the absence of RNA concentration measurements, might have contributed to the discrepancy observed between our sequencing results and RT-qPCR validation.

The dysregulation of NALT expression was explored in the study by Wang et al. [27] that demonstrated a significant upregulation of NALT in association with NOTCH1 in human samples in pediatric T cell acute lymphoblastic leukemia. High expression of NALT correlated with increased levels of NOTCH1, and their interaction promoted cell proliferation both in vitro and in vivo. A similar observation was described by Ye and colleagues [28], who showed upregulated levels of NALT1 in patients with advanced CRC stage and in CRC cell lines. In their study, NALT1 contributed to cancer progression by acting as a molecular sponge for microRNA-574-5p. This interaction led to increased expression of the PEG10 gene, promoting CRC cell proliferation, migration, and invasion. In another study [29], NALT1 was significantly overexpressed in gastric cancer tissues and cells, and this overexpression was closely associated with tumor invasion, metastasis, and poor prognosis in gastric cancer patients. In our study, high-throughput RNA sequencing results supported the findings of the referenced studies, showing higher levels of NALT1 in cancer patients, specifically in sEVs isolated from peripheral blood of individuals with CRC. Additionally, our validation testing confirmed significantly increased levels of NALT1 in CRC patients compared to healthy controls. Although not statistically significant, we also observed a higher abundance of NALT1 in more advanced stages of the disease.

Similar to the reported roles of NALT1 in various cancers, apart from our CRC findings, dysregulation of LINC02499 was detected in a hepatocellular cancer (HCC). The study by Ma et al. [30] revealed that LINC02499 was significantly downregulated in HCC and its lower expression was associated with poorer patient survival. Furthermore, the overexpression of LINC02499 in vitro had an inhibitory effect on the proliferation, migration, and invasion of HCC cell lines. A similar observation was reported by Zhang et al. [31] who showed LINC02499 to be downregulated in HCC tissues compared to adjacent normal tissues. The authors identified LINC02499 as the lncRNA most significantly correlated with a range of clinicopathological factors in HCC and demonstrated its significance in predicting overall survival in HCC patients. LINC02499 was recognized as a protective factor against the progression of the disease. While the function of LINC02499 has been described in relation to HCC, its role in CRC, particularly in sEVs, remains unexplored. In CRC, we observed a similar downregulation of LINC02499 in the sequencing analysis of patient-derived sEVs, reflecting its expression pattern in HCC. This could suggest a potentially universal role of LINC02499 as a tumor suppressor across different cancer types. Despite the lack of confirmation in the validation phase for differences between CRC patients and healthy controls, we observed a noticeable trend indicating LINC02499's differential abundance between early (I + II) and late (III + IV) stages of CRC. This trend was close to reaching statistical significance.

Chung and colleagues [32] found that the lncRNA LINC01013 was prominently overexpressed in tumor tissue specimens of anaplastic large-cell lymphoma (ALCL), as well as being significantly upregulated in invasive ALCL cell lines. This lncRNA influenced tumor behavior and promoted cell proliferation, suggesting its use as a prognostic marker in ALCL. Similarly, Wang et al. [33] showed that LINC01013 was significantly overexpressed in HCC tumors, and its upregulation was associated with a worse prognosis of HCC patients. Moreover, loss- and gain-of-function experiments revealed that LINC01013 could promote HCC cell proliferation and tumor progression by enhancing stemness of cells both in vitro and in vivo. In contrast, our sequencing data interestingly revealed that LINC01013 was significantly downregulated in sEVs isolated from CRC patients compared to healthy controls, suggesting a distinct role of LINC01013 in CRC. However, this observation was not significant in our validation cohort, highlighting a potential complexity in the behavior of LINC01013 across different biological matrices and cancer types.

In pancreatic adenocarcinoma (PAAD), LINC01637, also known as XXbac-B135H6.15, was identified as significant in the study by Deng et al. [34]. In this study, the high expression of LINC01637 was associated with better overall survival in PAAD patients, indicating its potential as a protective factor against disease progression. Additionally, its expression inversely correlated with the increasing risk score in PAAD, suggesting its importance as a potential prognostic biomarker for this type of cancer. Huang et al. [35] identified LINC01637 as being overexpressed in bladder cancer cell lines T24 and J82 compared to a less aggressive cell line of bladder cancer. However, overexpression of LINC01637 in the cell lines did not translate to enhanced levels in the exosomes derived from these cells. In contrast to its roles in PAAD and bladder cancer, our study investigates LINC01637 in the context of CRC, specifically examining its abundance in sEVs. The analysis of RNA sequencing data revealed a significant elevation of LINC01637 in patient samples relative to healthy controls, indicating its distinct role in CRC compared to documented functions in other cancers. Importantly, we validated these findings by a larger study cohort, which confirmed the high abundance of LINC01637 in sEVs from CRC patient blood serum, suggesting its potential as a non-invasive biomarker in CRC diagnostics.

While our study provides substantial insights into the relative abundance of lncRNAs in sEVs from CRC patients, we have encountered some limitations. Firstly, the pooling of samples, while necessary due to low RNA concentrations, could mask individual variability and relevant differences between patients. This approach, combined with the challenges of quantifying low amounts of RNA, may limit the direct clinical applicability of our findings.

Secondly, while preamplification enables the detection and quantification of RNA molecules that would otherwise be below the threshold of detection, it is not without its drawbacks. This process can introduce amplification biases and non-specific artifacts that can lead to disproportionate representation of certain RNA sequences, which may not accurately reflect their true abundance in the original sample. Despite these challenges, the use of preamplification was a necessary compromise given the current technological constraints and the low RNA yield from sEVs.

Thirdly, although our RNA sequencing approach identified a significant number of lncRNAs with different levels between CRC patients and healthy controls, the analytical power of specific lncRNA for clinical use might be limited. This could be partially due to the technical challenges associated with the isolation of sEV by SEC, which can introduce variability by co-isolation of other EV subtypes. Nevertheless, our findings highlight the biological significance of lncRNAs isolated from sEVs, revealing their potential as non-invasive biomarkers of CRC.