CLL Exosomes Modulate the Transcriptome and Behaviour of Recipient Stromal Cells and Are Selectively Enriched in miR-202-3p

Abstract

Bi-directional communication with the microenvironment is essential for homing and survival of cancer cells with implications for disease biology and behaviour. In chronic lymphocytic leukemia (CLL), the role of the microenvironment on malignant cell behaviour is well described. However, how CLL cells engage and recruit nurturing cells is poorly characterised. Here we demonstrate that CLL cells secrete exosomes that are nanovesicles originating from the fusion of multivesicular bodies with the plasma membrane, to shuttle proteins, lipids, microRNAs (miR) and mRNAs to recipient cells. We characterise and confirm the size (50-100 nm) and identity of the CLL-derived exosomes by Electron microscopy (EM), Atomic force microscopy (AFM), flow cytometry and western blotting using both exosome- and CLL-specific markers. Incubation of CLL-exosomes, derived either from cell culture supernatants or from patient plasma, with human stromal cells shows that they are readily taken up into endosomes, and induce expression of genes such as c-fos and ATM as well as enhance proliferation of recipient HS-5 cells. Furthermore, we show that CLL exosomes encapsulate abundant small RNAs and are enriched in certain miRs and specifically hsa-miR-202-3p. We suggest that such specific packaging of miR-202-3p into exosomes results in enhanced expression of 'suppressor of fused' (Sufu), a Hedgehog (Hh) signalling intermediate, in the parental CLL cells. Thus, our data show that CLL cells secrete exosomes that alter the transcriptome and behaviour of recipient cells. Such communication with microenvironment is likely to have an important role in CLL disease biology.

Fig 1. Characterization of CLL derived exosomes.

A) Transmission electron microscopy (TEM) image of exosomes enriched from CLL cell culture medium after 48 hrs. (a) Exosomes are visible as small 50–100 nm vesicles with bi-layered membranes (scale bar: 100 nm). (b) Exosomes were immune-isolated using magnetic beads coated with anti-HLA antibody and ultrathin sections of resultant beads processed for image analysis. The representative TEM image shows a bead coated with exosomes (scale bar: 0.2μm). B) AFM images of exosomes from a representative CLL case immobilized on a mica surface using Peakforce Tapping mode (Multimode 8, Bruker). A) Topographic image of exosomes. B) DMT Modulus image C) Adhesion image: The exosomes appear as circular biconvex vesicular structures. The DMT Modulus and adhesion images of exosomes show explicit sub-structures at the centre of the vesicles (arrows). D) Schematic cross-section following the line in the indicated exosome in C). C) FACS analysis of B-cell antigens on CLL exosomes prepared from primary CLL cases by density ultracentrifugation and adsorbed onto 4μm aldehyde-sulphate latex beads and incubated with isotype control, anti-HLA-A, B, C or CD19 antibodies followed by FITC-conjugated secondary antibody. Mean fluorescence intensities are plotted (mean ± standard error of mean (S.E.M.) (n = 9)). D) Representative analysis showing surface expression of CD81, HLA-DR, CD37, and integrin α4 (ITGA4) in CLL derived exosomes coupled to aldehyde-sulphate latex beads. Binding of FITC-conjugated isotype controls are included for comparison. E) Immunoblot analysis of CLL derived exosomes: Lysates from CLL exosome were probed for abundance of HLA-DR, MHC, IgM, Gp96, Lyn kinase, TSG 101 and Calnexin. Images are representative of analyses of 3 cases. WCE-Whole cell extract.

Fig 2. CLL-derived exosomes are internalised by stromal cells and localise to late endosomes.

Exosomes purified from CLL cell culture supernatants were labelled with the lipophilic dye PKH67. HS-5 stromal cells were cultured in the absence (control) or presence of fluorescent-labelled exosomes for 24hrs. Paraformaldehyde fixed HS-5 cells were permeabilised and stained for Lamp-1 or TSG-101, followed by FITC-conjugated secondary antibody and visualised by fluorescence microscopy (original magnification, ×100). Nuclei were stained using DAPI. Internalised exosomes co-localise with Lamp-1, a marker for late endosomes but not with TSG-101. A) Exosomes were internalised by HS-5 Cells and visualized for PKH67 (green), Lamp-1 (red) and DAPI (blue). Co-localisation of Lamp-1 and PKH67 appear in yellow. B) HS-5 Cells were visualized for PKH67 (green), TSG-101 (red) and DAPI (blue) and do not show co-localisation C) As controls. HS-5 cells were stained with free PKH67 dye without exosomes or antibody isotype control to exclude non-specific binding. (Scale bars: 10μm).

Fig 3. CLL derived exosomes modulate the gene expression profiles of HS-5 cells and alter the rate of their proliferation.

(A) HS-5 cells were treated with MEC-1 exosomes for 24hrs and a pathway array analysis of cellular RNA was performed (S2 File). Untreated HS-5 cells ‘spiked’ with equivalent doses of MEC-1 exosomes served as control. Validation of c-fos and ATM expression level changes determined by additional RT-qPCR is shown. Error bars represent ± S.E.M. of three independent experiments. p values indicating level of significance are shown. (B) Exosomes influence c-fos activation in recipient cells: c-fos heterodimer complex DNA binding activity was detected using an ELISA transcription factor assay kit (Active Motif). HS-5 cells were cultured alone or with exosomes as indicated. Nuclear extracts from MEC-1 and CLL exosome treated HS-5 cells show increased c-fos activation (p = 0.009 and p = 0.06 respectively). (C) HS-5 cells plated at a density of 3000 cells/well were incubated in the absence or presence of exosomes derived from primary CLL cell culture medium (CLL CM), CLL patient plasma or normal healthy donor plasma, (150ug/ml) and proliferation measured after 48 hrs using the CyQuant assay. Paired t-test was used to calculate p values. All data represent mean ± S.E.M of triplicate experiments.

Fig 4. MicroRNA profiles and RT-qPCR of purified exosomes and parental CLL cells.

A) A volcano plot of LNA array miRNA profiles of CLL cells vs exosomes samples is depicted. The x-axis shows the Log2 fold-change in miRNA expression between cellular cases and exosomes. The y-axis shows the -Log10 of the p value for each miRNA. Expressed miRs that are statistically significant between the two groups appear above the line (p < 0.05). B) The miRs let-7g, miR-21, miR-29a, miR-29b, and miR-26a were selected for validation of LNA array by RT-qPCR analysis (n = 5). Data represent mean ± S.E.M of triplicate experiments. C) miR-202-3p is enriched in exosomes compared to parental cells in the LNA array data (mean ± S.E.M). D) Fold change of miR expression in exosomes vs cellular CLL samples showing the enrichment of miR-202 by RT-qPCR in two representative cases.

Fig 5. Expression of miR-202-3p and its target Sufu mRNA in CLL cells.

(A) HS-5 cells incubated with MEC-1 exosomes for 24 hrs show increased expression of miR-202-3p by RT-qPCR analysis. The results represent the findings of three independent experiments. (B) HS-5 cells incubated with MEC-1 exosomes for 24 hrs show decreased expression of Sufu, one of the miR-202-3p targets, by RT-qPCR analysis. The bar graphs are derived from three independent experiments. (C) RNA extracted from HS-5 cells transfected overnight with a synthetic miR-202-3p mimic or scrambled control, was subjected to RT-qPCR to assess Sufu mRNA expression. (D) RT-qPCR of miR-202-3p, using RNA prepared from CLL cells (n = 19) or normal B cells (n = 4). Box plots show the expression level of miR-202-3p relative to a miR-U6B control for normalisation. (E) Primary CLL cell lysates (n = 12) were subjected to RT-qPCR analysis for miR-202-3p and Sufu mRNA expression levels, normalised to RNU6 and GAPDH expression respectively, and show statistically significant inverse correlation. Plotted data represent mean of experiments performed in triplicate. (F) Sufu mRNA expression, using RT-qPCR and GAPDH as reference, in primary CLL cells of 12 unmutated and 10 mutated CLL cases is shown. Data demonstrate relatively higher (p = 0.05) levels of Sufu expression in the IgVH unmutated cohort. The expression in individual cases represents the mean of triplicate measurements.

Fig 6. Expression of miRs 202-3p, 29a, and 21 in exosomes derived from plasma of CLL patients versus normal donors.

(A) Exosomes sourced from plasma obtained from CLL patients or healthy donors (n = 3 each) were subjected to RT-qPCR analysis for absolute levels of miR-202-3p using a standard curve method. The levels of miR-202-3p are significantly higher in exosomes purified from CLL plasma (p = 0.03). (B) Absolute levels of miR-29a and miR-21 in exosomes harvested from CLL patient and healthy donor plasma (n = 3) were determined by RT-qPCR analysis. Standard curves for miR-29a and miR-21 were generated using the miRVana^TM miRNA reference panel v9.1. Statistical analysis was performed using an unpaired t-test and yielded p values of 0.01 and 0.2 for miR-29a and miR-21 respectively.