Programmed vesicle transfer of green fluorescent protein from a stably transduced cell line to primary hematopoietic cells

To the editor:

Release and trafficking of cell membrane–derived vesicles is a constitutive cellular function important to cell-cell communication.¹  Montecalvo and colleagues recently described the transfer of microRNA (miRNA) between dendritic cells.^2,3  The authors report the selective incorporation of specific miRNA in vesicle compartments as a function of cellular-differentiation stage and demonstrate endogenous miRNA transfer to bystander cells via nano-sized vesicles with translational suppression of cognate mRNA targets. Even as the molecular mechanisms for protein or RNA incorporation into vesicles remain to be clarified, these elegant studies confirm earlier work demonstrating cell-cell transfer of coding and noncoding small RNA via vesicles with modulation of the target cell phenotype.^1,4,5  At the same time they leave open to what extent vesicle trafficking occurs between nonimmune cells and whether it can be exploited for the programmed transfer of nonendogenous protein or RNA.

To test the hypothesis that protein or coding RNA can be deliberately transferred between nonimmune cells via vesicles, we genetically modified Jurkat T cells with a vesicular stomatitis virus G protein–pseudotyped lentivector bearing a green fluorescent protein expression cassette (Jurkat-GFP). Vesicle release by Jurkat cells has recently been studied by others and our transmission electron microcopy studies (Figure 1A) confirm that these were predominantly in the exosome size range (30-110 nm; Figure 1B). Jurkat-GFP cells were propagated in bulk, enriched by flow-cytometric sorting for > 90% purity, with stable proviral integration indicated by GFP expression over serial (> 20) passages (Figure 1C). Twenty-four-hour culture supernatant from Jurkat-GFP and parental Jurkat cells was collected and processed to generate cell-free vesicle preparations by standard sequential gradient centrifugation.⁶  To demonstrate successful transfer we exposed murine bone marrow cells (mWBM, C57B/6 strain) to Jurkat-GFP–derived purified vesicles (Jurkat-GFP-V) for 48 hours and detected GFP expression in mWBM by flow cytometry (Figure 1D). We tested vesicle transfer in 2 additional cell-cell contact-independent settings: as vesicle rich media (VRM) after removing debris from the harvest culture supernatant (2000g × 15 minutes) and in 0.4-μm transwell coculture (TW). After 48-hour coculture we observed successful amplification of GFP transcripts in DNAse-treated RNA from mWBM cells under all 3 conditions (Figure 1E). To highlight the potential relevance of our findings, these studies were repeated using human CD34⁺ (hCD34; StemCell Technologies) target cells, with identical results, showing cell contact–independent vesicle transfer of GFP transcript from Jurkat-GFP cells (Figure 1F).

Taken together, our experiments demonstrate the cell-cell contact-independent transfer of nonendogenous coding RNA from stably transduced Jurkat cells to murine and human hematopoietic targets after purified vesicle, vesicle-rich media, or transwell exposure. As a caveat, these observations do not yet resolve to what extent GFP protein transfer, versus mRNA translation, accounts for the observed expression in hematopoietic target cells. Notwithstanding, our report extends the observations by Montecalvo and colleagues by revealing the feasibility of vesicle trafficking for the programmed transfer of stably introduced transgenes between nonimmune cells. We believe the deliberate cell-cell vesicle transfer of nonendogenous RNA may be useful for the experimental, and potentially therapeutic, manipulation of hematopoietic and other stem cells.