Brain:

For brain targeting, we chose to use RVG, which is a short peptide from Rabies Virus, as our targeting protein. RVG can specifically recognize acetylcholine receptor in the central nervous system [2], thus we engineer the RVG peptide into the lamp 2b and we use pcDNA 3.1(+) as our vector.
To produce the brain-specific RVG exosomes, we first transfected the exosome-producing cells, HEK 293T cells, with the plasmids encoding the fusion protein of lamp 2b and RVG peptide.
Also, we use a siRNA as an indicator to assess the targeting effect of RVG-exosomes and loaded the siRNA into the RVG-modified exosomes by direct transfection of the HEK 293T cells with siRNA probe.

Experimental procedure and Results:

1. Quantification of siRNA contained in the exosomes
We first quantify the amount of siRNA encapsulated into the exosomes. siRNA was transfected to the HEK 293T cells(which has already been transfected with RVG-Lamp2b plasmids) before exosome purification. Control group was not transfected with such siRNA.
By quantitative PCR analysis of a series of siRNA with known concentration, we drew a standard curve and then use it to calculate the absolute amount of siRNA in the exosomes. As shown in Fig.1, the amount of siRNA in the negative control is quite low (background) while the siRNA contained in the RVG exosomes transfected with siRNA probe reached as high as 0.8 fmol (RNA)/μg (exosome).
The result demonstrates that siRNAs were successfully encapsulated into RVG exosomes.

Figure.1 Empty exosome is collected from HEK 293 T cells without transfection of siRNA probe while the RVG exosome +siRNA is collected from the HEK 293T cells after transfection of siRNA probe. The amount of siRNA contained in the exosome is measured by qPCR.

2.In vitro evidence for the entry of RVG exosomes into the primary cortical neuron
To confirm the targeting capability of modified exosomes, we co-culture the siRNA-containing RVG exosomes with primary cortical neurons, and then measure the amounts of siRNA probe in the neurons to see if our RVG exosome can successfully get into the neurons.
From Fig.2, we can see that siRNA probe labeled with Alexaflour 555（red ）successfully get into the neurons.

Figure.2 Confocal microscopy image of the internalization of Alexaflour 555（red ）labeled siRNA into neurons.

In qPCR analysis of the relative siRNA probe level in the neurons co-cultured with both empty RVG-exosomes and siRNA containing RVG-exosomes, respectively (Fig.3), almost no siRNA was detected from the neurons co-cultured with 40 μg empty RVG-exosomes, while the siRNA detected in neurons co-cultured with siRNA probe containing RVG-exosomes show both significant and dose-dependent increase.

Figure.3 The RNA was extracted from the primary cortical neurons co-cultured 24h with 40 μg empty exosome, 20μg siRNA containing RVG-exosomes,40μg siRNA containing RVG-exosomes, respectively. And the RNA extracted was measured by qPCR using probe for the siRNA encapsulated in the exosomes.

3.In vivo evidence for the entry of RVG-exosomes entry into the brain
To further investigate whether the RVG-exosome can get into brain, we intravenously injected the empty RVG-exosomes, siRNA containing RVG-exosomes, respectively, into the mice. Then we extracted brain tissue and measured the siRNA level in the cortex and medulla.

As shown in Fig.4, no siRNA was detected in both cortex and medulla after the injection of empty RVG-exosomes and siRNA containing exosomes, while for the siRNA containing RVG-exosomes, the siRNA detected in the cortex and medulla are significant higher than the others.

Figure.4 The mice were intravenously injected with 200 μg empty exosomes,siRNA containing exosomes and siRNA containing RVG-exosomes, respectively once a day, and continued for four days. On the fifth day, the mice were killed and their brains were taken out. The RNA from their cortex and medulla were measured using siRNA probe for the siRNA encapsulated in the exosome.

4.In vivo evidence for the targeting effect of RVG-exosome
The siRNA contained in the exosomes were designed to silence OPRM gene, which encodesμ- opioid receptor in neuron cells. We intravenously injected three groups of mice: control group is injected with PBS (Mock), second group is injected with empty exosome, third group is injected with RVG-exosome which has siRNA in it. Next, by qPCR analysis we determined the relative OPRM mRNA level in mice brain. The result (Fig.5) shows that RVG exosome successfully target to neuron cells and the siRNA brought by the exosome down-regulate the gene significantly.

Figure.5 in vivo evidence for targeting effect of RVG-exosome. First group (Mock) was injected with PBS and second group was injected with Empty exosome. The last group was injected with RVG-exosome which contained siRNA that could silence OPRM gene. The result shows that modified exosome successfully crossed blood-brain-barrier, entered neuron cells and thus significantly silenced OPRM gene.

Finally, with all these data, we demonstrated that our modified exosome can not only cross the blood-brain-barrier but also specifically deliver siRNA to target tissue and thus silence specific gene.