Abstract
Colorectal cancer (CRC) is the second highest cause of cancer-related mortalities in Aotearoa, New Zealand, corresponding to 1,265 deaths annually, with an incidence of 40.27 per 100,000 individuals. Rates of early-onset CRC in individuals under 50, has increased in incidence from 4.41 to 8.03 per 100,000 individuals within 20 years. The gold standard for CRC diagnosis is a colonoscopy, the capacity of which is a major limitation for diagnosis in Aotearoa. Therefore, the need for a minimally invasive and accessible diagnostic tool, such as blood-based biomarkers, would aid in improving diagnostic pathways. Extracellular vesicles (EVs) are membrane-bound nanoscale particles that are released by a wide range of healthy and diseased cells, including CRC tumour cells, into the circulation. EVs carry bioactive cargo, such as proteins and nucleic acids, which can mediate paracrine communication to different cell types and promote tumorigenesis, making them useful biomarkers of disease. MicroRNA (miRNA) cargo of EVs isolated from the circulation have the potential to differentiate CRC from healthy individuals. However, previous studies have identified diverse and inconsistent miRNA panels for CRC stratification, which indicate the need for pre-clinical models to identify EV cargo that are tumour-specific and representative of physiological conditions. Two-dimensional (2D) cell cultures are traditionally used for EV studies; however, multi-spheroid three-dimensional (3D) cultures embedded in extracellular matrix materials better replicate the cell-cell and cell-matrix interactions in solid tumour microenvironments. This project aimed to characterise EVs from 3D cultures, with the hypothesis that their characteristics of size and concentration and miRNA cargo would differ from EVs derived from 2D cultures.
Optimisation of 3D spheroid formation of two CRC cell lines, DLD-1 (colon) and SW837 (rectal) identified 3.8×105 cell/mL and 1.3×106 cell/mL, respectively, to provide the most reliable spheroid formation. Conditioned cell culture medium (CCM) and cell lysate (CL) were collected from 2D and 3D cultures of SW837 (n = 3, paired) and DLD-1 (n = 1, unpaired) cell lines. EVs were isolated from CCM using ultrafiltration and size exclusion chromatography. CRC-EVs were characterised using tuneable resistive pulse sensing and a bicinchoninic acid assay. Total RNA isolation followed by miRNA-specific cDNA synthesis was performed for both EV and paired CL samples. A small subset of miRNAs that have been suggested as biomarkers in CRC with links to tumour properties were selected for assessment by RT-qPCR. The panel included transcripts with reported oncogenic roles, miR-181a-5p, miR-221-3p, and miR-23a-3p and tumour-suppressive roles, miR-193a-5p, miR-101-3p, and miR-98-5p. EV-miRNA expression was normalised to the endogenous reference and the seeded cell number to estimate EV-derived transcripts per cell.
In the SW837 cell line, we found statistically significant increase of miR-221-3p and miR-23a-3p in 3D culture-derived EVs in comparison to 2D-derived EVs (4.0- and 3.9-fold; p = 0.028 and p = 0.039, respectively) and their 3D cell origins (2.8- and 5.6-fold; p = 0.005 and p = 0.012, respectively). Although miR-181a-5p did not reach significance, it showed an increased trend in 3D-EVs compared to 2D-EVs (4.6-fold; p = 0.09). The DLD-1 cell line aligned with these trends, however it cannot be determined whether these descriptive observations are biologically robust as the sample size did not capture variation within this cell line. To explore the biological effects of the enriched 3D EV-miRNA, enriched pathway analysis was performed for pooled targets of miR-221-3p and miR-23a-3p and indicated that 3D culturing promoted secretion of EVs with greater oncogenic potential.
While further investigation with larger sample sizes is required to validate these biological trends, our findings demonstrate differential EV-miRNA expression between 2D and 3D CRC cultures. This contributes to the knowledge that 3D cultures show different EV-secretion compared to 2D cultures and provides a more promising pre-clinical model for future EV-based studies in CRC diagnostics.