Single-cell-derived clones display different phenotypic features

To investigate the cellular and molecular mechanisms involved in inter-clonal communication in PDHGG, we generated and characterized the phenotypic features of single-cell-derived clones from patient-derived primary cell lines. To this purpose, we have generated the optical barcoded (OB) clones from multifluorescent PDHGG-WT (OPBG-GBM002) and a DMG-H3K27-altered (OPBG-DIPG002) patient-derived cell lines, using the multifluorescent marking technology as previously described [11]. The OB clones and their bulk cell lines were characterised for their phenotypic features related to cell morphology, growth, migration, invasion and adhesion.

Individual clones displayed different cell morphologies (Fig. 1A), as exemplified by the clone 1C5 which had a larger and round cell shape when compared to the more elongated features of 2B4, both derived from OPBG-DIPG002. Moreover, the clone 5E2, derived from OPBG-GBM002, presented a smaller cell size compared to clones 1D3 and 2G7 which were fusiform with well-pronounced cell protrusions (Fig. 1A). Interestingly, the 2 bulk cell lines and each clone differed also in terms of invasion (Fig. 1B, C) and migration (Fig. 1D, E) ability. The cells from the OPBG-DIPG002-derived 1C5 clone presented an ameboid-like invasion and migration pattern in contrast to the more mesenchymal-like pattern observed with the cells of the 2B4 (Fig. 1B). These clones also displayed a significantly different degree of invasion and migration when compared to their bulk population (Fig. 1B, D). The OPBG-GBM002 bulk cell line and its derived clones had a very low cell invasive capability even if, the 5E2 and 2G6 showed a significantly higher degree of invasion than the bulk and the other cellular clones (Fig. 1C). On the contrary, the bulk cell population and its derived clones displayed high but heterogeneous migration capacity (Fig. 1E). The clone 5E2 demonstrated a significantly higher degree of migration when compared to the OPBG-GBM002 bulk cell line and the other 3 derived clones, with the 1D3, 2G6 and 2G7 being the least migratory.

Fig. 1

Single-cell-derived clones display different phenotypic features. A Representative brightfield images of the OPBG-DIPG002 and OPBG-GBM002 patient-derived and the clone cell morphologies. Scale bar = 100 μm. B, C Invasion assay was performed for 96 h with the patient-derived cells OPBG-DIPG002 (B) and OPBG-GBM002 (C) and with the respective derived clones. Brightfield images were segmented as indicated to quantify the invaded area as reported on the graph bar. Scale bar = 500 μm and 200 μm. D, E Representative images of the migration assay performed onto Matrigel for 96 h with cells of OPBG-DIPG002 (D), OPBG-GBM002 (E) and the respective derived clones. Brightfield images were segmented as indicated to quantify the migration area as reported on the graph bar. Scale bar = 500 μm and 200 μm. F, G Graph bar representing the results of the cell proliferation assay performed with OPBG-DIPG002 (F), OPBG-GBM002 (G) and the respective derived clones. After seeding, brightfield images were acquired at the indicated times and cell confluency was determined as described in the material and methods. The percentage of cell confluence corresponds to the cell confluency at the indicated time point normalised to the confluency at the time zero (t0). H, I Graph bar representing the results of the OPBG-DIPG002 (H), OPBG-GBM002 (I) and the derived clone cell adhesion assays using the 7 indicated extracellular matrices. All brightfield images were acquired and analysed with the Celigo imaging cytometer and are representative of 3 independent biological repeats. Data are mean ± SD, n = 3. (****) p < 0.0001; (***) p < 0.001; (**) p < 0.01; (*) p < 0.05

Furthermore, the clones exhibited overall different growth rates, also when compared to the bulk cell line from which they were derived (Fig. 1F, G). The single-cell-derived clones and the bulk cell lines exhibited a heterogeneous adhesion capacity (Fig. 1H, I). Overall, the cell lines OPBG-DIPG002 and OPBG-GBM002 displayed a higher adhesion capability on all seven matrices tested compared to their respective clones. For the OPBG-DIPG002, 1C5 showed stronger adhesion property on Laminin (mouse and human), Fibronectin and Tenascin-C, while 2B4 adhered more on Vitronectin, Collagen I and IV (Fig. 1H). Furthermore, the OPBG-GBM002-derived 2G6 and 2F4 clones, displayed the highest and the lowest adhesion capacity respectively, on all the matrices tested (Fig. 1I). Therefore, the clones and the bulk cell lines exhibited a heterogeneous adhesion phenotype.

Then, we looked at the genomic and transcriptomic profiles of both OPBG-DIPG002 and OPBG-GBM002 bulk cell lines and their respective derived clones. Next Generation Sequencing (NGS) analysis was performed at high depth using a custom-designed targeted panel. In addition to the common mutations (e.g. H3F3A, TP53, ATRX), the clones showed several shared mutations, not identified in the bulk and in the patient tumour tissue (e.g. MAX, KMT2D, SETD1B), and clone-specific mutations (e.g. AMER1, WINT11) (Fig. 2A, B). RNAseq was performed to obtain the transcription profile of OPBG-DIPG002 and OPBG-GBM002 and their respective clones. Both bulk cell lines and the clones showed heterogeneous transcriptomic profiles. While the transcriptomic profile of the clone 1C5 was closely related to the one of the OPBG-DIPG002 cell line, a diverse profile characterised all the clones derived from the OPBG-GBM002 bulk cell line (Fig. 2C, D). Based on the interesting and distinctive phenotypic features evidenced for the clones (e.g. motility, proliferation), we performed gene set enrichment analysis (GSEA) for the 1C5 and 2B4 clones derived from OPBG-DIPG002 and for 5E2 and 1D3, the two most phenotypically different clones derived from OPBG-GBM002. Interestingly, for the OPBG-DIPG002, 2B4 showed enrichment in genes associated with positive regulation of cell migration and extracellular matrix organization when compared to 1C5 (Fig. 2C). For the OPBG-GBM002, 5E2 displayed enrichment in genes associated with the positive regulation of cell proliferation and cell migration when compared to 1D3 clone (Fig. 2D). Interestingly, the clones with a particular aggressive phenotype, 2B4 for OPBG-DIPG002 and 5E2 for OPBG-GBM002, are associated with unique gene signatures (Additional file 2: Table S1), which correlate with a worse overall survival in patients (Additional file 1: Fig S1).

Fig. 2

Single-cell-derived clones display different genomic and transcriptomic features. A, B Next-generation sequencing was carried out on the OPBG-DIPG002 (A), OPBG-GBM002 (B) cell lines and the respective tumour samples and corresponding derived clones. Blood samples were used as a control. The heatmaps show the gene variants identified in at least one specimen. C, D Heatmaps showing the differential gene expression obtained from RNA sequencing analysis performed with cells from OPBG-DIPG002 (C), OPBG-GBM002 (D) and from the corresponding derived clones. Gene set enrichment analysis was performed for the top 50 most highly differentially expressed genes between the OPBG-DIPG002 derived clones 2B4 and 1C5 (C) and the OPBG-GBM002 derived clones 5E2 and 1D3 (D). All cell preparations were sequenced n = 1 and statistical comparisons were made by Gene Set Enrichment Analysis

These results confirm an intrinsic heterogeneity in PDHGG that is phenotypically displayed and retained by the distinct clones.

Inter-clonal interactions during migration and invasion

Next, we wanted to investigate how the inter-clonal interaction affects PDHGG tumour cell motility. To this end, we used two phenotypically and transcriptionally distinct clones, 1C5 and 2B4 for the OPBG-DIPG002, and 5E2 and 1D3 for OPBG-GBM002, and compared their invasion and migration phenotypes when grown individually and in co-culture (Additional file 1: Fig S2A, D, G). When in co-culture, 1C5 and 2B4 clones displayed significantly higher migration and invasion phenotypes compared to their mono-culture condition (Additional file 1: Fig S2B, E). Similarly, 5E2 and 1D3 clones also displayed an enhanced migration phenotype in co-culture compared to their mono-culture condition (Additional file 1: Fig S2H).

Taking advantage of having optical barcoded clones, we were able to image the cells with the Operetta CLS and look at the behaviour of each clone when they were in co-culture. Looking first at the percentage of cells in the invasion and migration area, we observed that the less motile clone had a higher percentage of invading and migrating cells compared to the other clone in co-culture (Additional file 1: Fig S2C, F, I). Moreover, by using the single-cell tracking on 3D migration and invasion assays, we analysed in-depth additional features of cell motility and looked at how these features were affected by the inter-clonal interaction when the clones were in co-culture (Fig. 3).

Fig. 3

Single-cell tracking of 3D migration and invasion. Representative fluorescent images acquired with the Operetta CLS 48 h after the initiation of the 3D migration (A) and invasion (E) assays for OPBG-DIPG002 single-cell-derived clones 1C5 (Venus) and 2B4 (m-Orange2) either in mono or in co-culture (overlay of m-Orange2 and Venus) Scale Bar = 200 μm. For cell tracking experiments, images were acquired every 30 min with the Operetta CLS over 48 h. Single-cell tracking was performed using the Harmony software as represented with the lines and arrows overlayed on the fluorescent images and the mean cell displacement (B–F), speed (C–G) and accumulated distance (D–H) were determined as reported on the Graph bar. Data are mean ± SD, n = 3. (****) p < 0.0001; (***) p < 0.001; (**) p < 0.01; (*) p < 0.05

Significant differences were observed in terms of cell speed, displacement, and accumulated distance between the clones in mono- and co-culture, in migration (Fig. 3A–D) and invasion (Fig. 3E–H) assays, with the clone 2B4 (m-Orange2) being generally more “motile” than the 1C5 (Venus). Interestingly, 2B4 also seemed to advantage of the co-culture condition with the 1C5 clone, as it significantly showed higher speed and accumulated distance when compared to its mono-culture (Fig. 3C, D).

These results, together with our previous observation [10, 11], support our hypothesis on the role of inter-clonal interactions in contributing to more invasive and migratory phenotypes of PDHGG tumour cells.

Primary characterization of exosomes derived from PDHGG bulk cell lines and single-cell-derived clones

We isolated and characterised extracellular vesicles (EVs) from conditioned medium (CM) of the OPBG-DIPG002 and OPBG-GBM002 bulk cell lines and their corresponding derived clones. The total EV protein content was quantified, showing a relatively variable protein amount between the bulk and the clones (Fig. 4A). Western Blot (WB) analysis demonstrated an enrichment of exosome-specific proteins, such as the tetraspanin CD63 and the tumour susceptibility gene 101 protein (TSG101) and, as expected, a decrease in Golgin subfamily A member 2 protein (GolgA2), which, instead was found in the total cell lysate (Fig. 4B). The SEM analysis, used to analyse the morphology and size of individual vesicles, with a detection limit of 0.5 nm [22], revealed single and aggregated round-shaped EVs, the majority of which ranged from 30 to 100 nm for both multifluorescent bulk cell lines and clones (Fig. 4C). The NanoSight tracking system analysis, which is used for distribution and particle concentration, with a detection limit of 60 nm [22], showed a relatively consistent EV size distribution with peaks between 100 and 150 nm (Fig. 4D).

Fig. 4

Primary characterization of exosomes and their cellular uptake. A Determination of the exosomal protein concentration. The quantification of the total exosomal protein obtained from 106 cells of OPBG-DIPG002 and OPBG-GBM002 and of the derived clones is shown. B The characterization of the exosomes from the different cell lines and derived clones as in (A) was carried out by Western Blot for the exosomal markers, CD63 and TSG101, and non-exosomal and cell membrane marker GolgA2. HSP90 was used as the loading control. C Images of the Scanning Electron Microscopy show a population of heterogeneously sized exosomes isolated from the OPBG-DIPG002 and OPBG-GBM002 cell lines and the from the respective clones. Scale bar = 200 nm. The table shows the minimum and maximum size of isolated exosomes. D Graphics representing the size distribution of the nanoparticles resulting from the NanoSight particle-tracking analysis performed with the exosomes obtained from OPBG-DIPG002 and OPBG-GBM002 and the derived clones culture medium. E–F Representative images of the exosome uptake experiments carried out with the clones derived from OPBG-DIPG002 (E) and OPBG-GBM002 (F). The recipient clone was cultured for 24 h in the presence of 10 µg/mL PKH67-labelled exosomes isolated from the donor clone. Graph-bar represent the quantification of the PKH67 fluorescent spots by cells, corresponding to the number of PKH67-labelled exosomes internalized by cells, as determined using the Harmony software. Scale bar = 20 μm and 10 μm. Data are mean ± SD, n = 3. (****) p < 0.0001; (***) p < 0.001; (**) p < 0.01; (*) p < 0.05

Altogether, these results confirmed that we successfully isolated, from our cell cultures, EVs corresponding to exosomes.

From donor to recipient clone: exosome uptake

To explore whether the exosomes secreted by the clones could exert a role in the inter-clonal interaction, we first performed exosome uptake experiments. Exosomes were isolated from CM of distinct clones: 1C5 and 2B4 derived from OPBG-DIPG002 and 5E2 and 1D3 from OPBG-GBM002. PKH67-labelled exosomes isolated from one “donor” clone, were added to the culture of the “recipient” clone, and vice versa. After 24 h, internalized fluorescent green signals were visualized in the cytoplasm of the cells (Fig. 4E, F). The quantification of the number of spots per cell showed a differential uptake between the clones. In particular, from the OPBG-DIPG002 bulk cell line, the 2B4 clone significantly internalized more 1C5-derived exosomes than the reverse (Fig. 4E). For the OPBG-GBM002 cell line, the 1D3 clone showed a higher exosome uptake compared to 5E2 (Fig. 4F).

These results demonstrate an active internalization of exosomes between “donor” and “recipient” PDHGG-derived clones.

The inhibition of exosome biogenesis affects the single-cell-derived clone motility in mono- and co-culture conditions

Next, we wanted to explore the possibility that exosomes could play a role in regulating the motility of these cells. To address this issue, we first evaluated the effect of exosome education on cell migration and invasion. Exosomes isolated from the “donor” clone were used to stimulate the “recipient” clone, while cells were undergoing migration or invasion. At the end of the education period, we observed that neither the migration nor the invasion of the recipient clones was affected by the exosomes of the donor clones when compared to liposomes-treated control cells (Additional file 1: Fig S3A, B).

To further explore the possibility that exosomes affect the migratory/invasive capability of the clones, we used the phospholipase inhibitor GW4869 [23], a known inhibitor of exosome biogenesis.

We used the compound at doses that do not affect cell viability (10 μM for OPBG-DIPG002 and 20 μM for OPBG-GBM002 clones, respectively, Additional file 1: Fig S4) and showed that GW4869 inhibits exosome biogenesis/secretion in our clones at a variable rate, up to 65% of inhibition, with the clone 2B4 been more consistently affected (Additional file 1: Fig S5).

Based on this, we then tested the effect of GW4869 on cell motility. We performed 3D migration assays with the clones grown in mono- and co-culture condition in presence or absence of the GW4869 compound (Fig. 5). GW4869 inhibited, in a dose-dependent manner, the cell migration of the clones either in mono- or co-culture.

Fig. 5

GW4869 treatment affects the migration capability of the single-cell-derived clones. A, B Representative fluorescent images of the cell migration assays performed with the OPBG-DIPG002 derived clones (A) 1C5 (Venus) and 2B4 (m-Orange2), and the OPBG-GBM002 clones (B) 5E2 (Venus) and 1D3 (m-Orange2), in mono-culture and in co-culture (overlay of m-Orange2 and Venus) and in presence or absence of 10 μM and 20 μM of GW4869 or DMSO as vehicle control for 48 h. The graph bar shows the quantification of the cell migration areas after segmentation and analysis of the images using the Harmony software. Images were acquired with the Operetta CLS every 30 min over 48 h. At the 48 h time point, Scale bar = 200 μm. Data are mean ± SD, n = 2 with 6 different technical repeats. (****) p < 0.0001; (***) p < 0.001; (**) p < 0.01; (*) p < 0.05

Furthermore, we performed single-cell tracking analysis to evaluate the effect of the inhibition of the exosome biogenesis on several parameters linked to cell motility, including accumulated distance, cell displacement and speed (Additional file 1: Fig S6A, B). For the OPBG-DIPG002 cell line, both clones, 2B4 and 1C5, appeared to be affected by the treatment with GW4869 (Additional file 1: Fig S6A). Upon treatment, cells of the clone 1C5 showed a significantly reduced displacement and speed either in mono- and co-culture conditions compared to the untreated control, while the cells of the clone 2B4 displayed a significant reduction in all the measured parameters. For the OPBG-GBM002 clones, the 5E2 and the 1D3 showed a strong reduction of the measured parameters (Additional file 1: Fig S6B).

Finally, we explored the possibility that exosomes could re-stimulate cell migration following the GW4869 treatment. To this end, exosomes from the “donor” clone were used to stimulate the migration in the “recipient” clone, in the presence or absence of GW4869. Interestingly, in the presence of GW4869, the exosomes obtained from the “donor” clone were able to rescue, even if partially, the GW4869-mediated inhibition of cell migration of the recipient clone, when compared to the relative control (Additional file 1: Fig S7).

These results indicate that exosome biogenesis plays a role in the motility of PDHGG cells and suggest that the inter-clonal communication mediated by the exosomes could contribute, at least in part, to the aggressive PDHGG cell phenotype.

miRNA analysis of exosome cargo from single-cell-derived clones

To determine how the inhibition of the exosome biogenesis affects PDHGG cell motility, we first analysed the cargo of the exosomes secreted by the single-cell-derived clones in terms of miRNAs.

For the identification of exosomal miRNAs (exo-miRNAs), we used a 384 miRNome PCR panel. Following the analysis, we did not identify differentially expressed exo-miRNAs between the paired clones, 1C5 and 2B4 for OPBG-DIPG002 and 5E2 and 1D3 for OPBG-GBM002. However, we found that the expression of some of the exo-miRNAs was clone specific. The miR-200c-3p was found exclusively expressed in the exosomes isolated from the clone 1C5, while miR-887-3p, miR-885-5p, and miR-582-5p were found exclusively expressed in the exosomes secreted by the clone 2B4, both OPBG-DIPG002 derived clones (Table 1). For the clones derived from OPBG-GBM002, miR-203a and miR-877-5p were found exclusively expressed in the exosomes of 5E2, while miR-572, miR-376a-3p, and miR-22-3p were only expressed in the exosome isolated from 1D3 (Table 1).

Table 1 Exosome miRNAs in OPBG-DIPG002 and OPBG-GBM002 single-cell-derived clonesIdentification of exosome miRNA target genes associated with migration and invasion

Given our interest in the potential role of the exosomes in migration and invasion processes, we performed further analysis to identify the target genes of the clone-specific exo-miRNAs involved in these processes. For the OPBG-DIPG002-derived clones, we identified 9 genes targeted by miR-200c-3p exclusively identified in the 1C5-derived exosomes (Table 1) and 6 genes predicted to be targets of miR-885-5p and miR-582-5p, that were specifically identified in the exosomes of the 2B4 clone (Table 1), with MNX1 and RAC12 genes being common targets of the two miRNAs. For the OPBG-GBM002 derived clones, we identified 5 genes as predicted targets of miR-877-5p, found in 5E2-derived exosomes (Table 1), and 7 targets for miR-376a-3p and miR-22-3p exclusively expressed in the 1D3 exosomes (Table 1).

We identified 3 genes, NTRK2, DDIT4 and NR2F2, as common targets of the exo-miRNAs identified in the clones derived from both OPBG-DIPG002 and OPBG-GBM002. Interestingly, these 3 genes are involved in the regulation of cell invasion and/or migration phenotype.

The inhibition of exosome biogenesis modulates the expression of genes regulating migration/invasion in single-cell-derived clones

We hypothesized that the inhibition of exosome biogenesis in PDHGG clones, and hence the consequent disruption of inter-clonal communication via exosomes, could affect their motility capability (Fig. 5A, B).

To test this hypothesis, we first studied the expression levels of the exo-miRNA target genes regulating migration and invasion discussed above, for the individual clones in mono and in co-culture (Table 1 and Fig. 6A a, B and C). When in co-culture, and in comparison to the mono-culture, we observed an overall modulation of the gene expression for both 1C5 and 2B4 clones derived from OPBG-DIPG002 (Fig. 6B), as well as for the clones 5E2 and 1D3 derived from OPBG-GBM002 (Fig. 6C). In the co-culture condition, out of the 15 genes identified as targets of the exo-miRNAs of OPBG-DIPG002 derived clones (Table 1), 4 of them, NR2F2, TUBB2A, VEGFA and FN1 were upregulated in 1C5, and 8 were found either downregulated, RAC1, NR2F2, NTRK2, DDIT4, FN1 and FLNA, or upregulated, PTPRZ1, in 2B4 (Fig. 6B). Two of the 15 target genes, MNX1 and TBX20, were not detected in any of the conditions analysed. For the OPBG-GBM002 derived clones, out of the 11 genes targeted by the identified exo-miRNA (Table 1), 9, including NTRK2, FUBP1, DDIT4, PDS5A, AMOTL2, MAPK8, NRF2, NDE1 and TUBB2 were found down-regulated and one FOXG1 upregulated in 1D3, while 4, NTRK2, DDIT4, NR2F2 and TUBB2B, were found upregulated in 5E2 (Fig. 6C). Interestingly, the mRNA expression levels of NTRK2, DDIT4 and NR2F2, identified as common targets of the exo-miRNAs, were found modulated in the clones 2B4, 1D3 and 5E2, except for the clone 1C5, for which only the NR2F2 mRNA expression level was found modulated. Furthermore, proteomic analysis performed on the clones from mono and co-culture conditions, demonstrated the expression and the modulation at protein level, of several miRNA predicted target genes (Additional file 1: Fig S8 and Additional file 3: Table S2).

Fig. 6

The inhibition of exosome biogenesis modulates the expression of migratory/invasive related genes in single-cell-derived clones. A Scheme representing the hypothetical mode of action of GW4869. Single-cell-derived clones in co-cultures were treated with the vehicle control (a) or with the GW4869 (b) for 96 h. B, C The graph-bar shows the fold change in mRNA expression between a clone grown in mono and co-culture conditions. The mRNA levels were analysed by Q-PCR for the indicated and selected set of genes involved in migration and invasion phenotypes and for the clones 1C5 and 2B4 (B) derived from OPBG-DIPG002 and the clones 5E2 and 1D3 derived from OPBG-GBM002 (C). D, E Graph-bars represent the fold change in mRNA expression of the indicated genes, analysed by Q-PCR between cells treated and not treated with 10 μM of GW4869 for the OPBG-DIPG002 derived clones 1C5 (D) and 2B4 (E) grown for 96 h in mono- (dash-bars) or in co-culture (plain coloured-bar) conditions. F, G Same as D and E but with the clones 5E2 (F) and 1D3 (G) derived from OPBG-GBM002 and treated with 20 μM of GW4869. For the gene see Table 1. Data are mean ± SD, n = 3. (****) p < 0.0001; (***) p < 0.001; (**) p < 0.01; (*) p < 0.05

These data suggest that, when the clones are in co-culture, the modulation of the specific exo-miRNAs targeted genes is the result of an inter-clonal cross-talk.

We have demonstrated that the inhibition of the exosome biogenesis affects the cell motility of PDHGG clones (Fig. 5). Moreover, we have identified exo-miRNA target genes involved in processes such as cell migration and invasion (Table 1). Therefore, to confirm the role of the exosomes in inter-clonal communication and, in particular, in the modulation of genes relevant to PDHGG inter-clonal motility, we repeated the mono- and co-culture experiments in the presence or absence of the exosome biogenesis inhibitor, GW4869 (Fig. 6A a and b). After 96 h of GW4869 treatment, we analysed, in each clone, the mRNA expression levels of the specific genes targeted by the exo-miRNAs produced by the “brother” clone (Table 1). For the OPBG-DIPG002 derived clones, a decrease in RAC1, NR2F2 and CTBNN1 mRNA expression levels was observed in the cells of the clone 1C5 in the co-culture condition (Fig. 6D) and, on the other hand, a significant increase in the mRNA levels of NTRK2, PTPRZ1, DDIT4 and GLI3 in the cells of 2B4, was detected when two clones were in co-cultures (Fig. 6E) and in comparison to their respective untreated control. Regarding the OPBG-GBM002 cell line, 5E2 was characterised by a significant modulation of the mRNA expression levels of NTRK2, PEX5, DDIT4 and FOXG1 upon GW4869 treatment either in mono- and co-culture conditions (Fig. 6F). The 1D3 clone also showed a modulation of gene expression. In particular, we observed a significant increase in MAPK8 expression when 1D3 was in co-culture with 5E2, while there was an overall significant decrease of the NR2F2, AMOTL2, NDE1 and TUBB2B expression in both mono and co-culture conditions (Fig. 6G).

Altogether, these results strongly support the involvement of the exosomes in the inter-clonal regulation of the mRNA expression of genes implicated in PDHGG cell migration and invasion through the transport of clone-specific miRNAs.