Figure 5. Characterization of secreted exosomes after overexpression of nSMase2 in HEK293 cells. (A and B) Representative screen shots of the NTA videos for EV from HEK293 cells under normal condition (left) or after transfection with nSMase2 plasmid (right). (C and D) Size and intensity of EV from HEK293 cells under normal condition (left) or after transfection with nSMase2 plasmid (right). (E) Concentration of different particle sizes of exosomes with (red line) or without (blue line) transfection with nSMase2 plasmid. The peak of size distribution of EV shifted from 170 nm to 120 nm after transfection with nsMase2 plasmid, indicating the increase in quantity of secreted exosomes.
2 Targeting capability validation (in vitro and in vivo)
2.1 RVG exosomes specifically deliver fluorescent-labeled oligonucleotide into neuronal cells
To determine whether RVG exosomes can deliver siRNAs into neuronal cells, Neuro2A cells were selected as the recipient cells to incubate with RVG exosomes loaded with Alexa Fluor 555-tagged oligonucleotide (red fluorescence). First, untreated Neuro2A cells or cells treated with RVG exosomes but without loading the fluorescent-labeled oligonucleotide, which served as the controls, were not fluorescently labeled under fluorescence confocal microscopy. In contrast, significant fluorescence signals were observed in Neuro2A cells treated with RVG exosomes loaded with Alexa Fluor 555-tagged oligonucleotide, whereas the fluorescence signals were dramatically lower in cells treated with unmodified exosomes loaded with Alexa Fluor 555-tagged oligonucleotide. The results suggest that RVG exosomes can specifically deliver siRNA to cells of neuronal origin, while unmodified exosomes are generally rejected by neuronal cells. Interestingly, a greater number of Alexa Fluor 555-tagged oligonucleotides accumulated in non-neuronal cells including C2C12 (skeletal muscle origin), A549 (lung origin) and MCF-7 (breast origin) when these cells were incubated with unmodified exosomes compared with those with RVG exosomes, suggesting that RVG exosomes are, in contrast, rejected by non-neuronal cells. In summary, the results indicate that RVG peptide on the exosomal membrane efficiently guides exosomes to enter neuronal cells bearing the acetylcholine receptor on their membranes but prevents exosomes from entering non-neuronal cells lacking the surface acetylcholine receptor.
Figure 6. Confocal microscopy images of fluorescent-labeled oligonucleotide (Alexa Fluor 555, red) in untreated control cells or in cells incubated with RVG exosomes (RVG exosome), unmodified exosomes loaded with fluorescent-labeled oligonucleotide (oligonucleotide-exosome) or RVG exosomes loaded with fluorescent-labeled oligonucleotide (oligonucleotide-RVG exosome). Images of four cell lines (Neuro2A, C2C12, A549 and MCF-7) were acquired.
2.2 RVG exosomes specifically deliver MOR siRNA into neuronal cells
Subsequently, MOR siRNA levels were assayed in recipient Neuro2A cells incubated with RVG exosomes loaded with MOR siRNA. The siRNA concentrations were barely detected in untreated control cells or in cells treated with RVG exosomes or unmodified exosomes loaded with MOR siRNA. In contrast, a significant amount of siRNAs were detected in Neuro2A cells after treatment with RVG exosomes loaded with MOR siRNA. As a control, C2C12 cells were treated with RVG exosomes loaded with MOR siRNA, and MOR siRNA was barely detected. Taken together, these results clearly demonstrate that the RVG peptide modification on the exosome membrane specifically guides exosomes to target neuronal cells bearing the surface acetylcholine receptor, allowing for the efficient delivery of MOR siRNA into the recipient cells.
Figure 7. Quantitative RT-PCR analysis of MOR siRNA concentrations in Neuro2A and C2C12 cells treated with RVG exosomes (RVG exosome), unmodified exosomes loaded with MOR siRNA (siRNA-exosome) or RVG exosomes loaded with MOR siRNA (siRNA-RVG exosome).
2.3 GFP levels in the brains of GFP transgenic mice decrease after tail vein injection of a plasmid that expresses GFP siRNA
To determine whether siRNA delivered via RVG exosomes can pass through the BBB and regulate endogenous gene expression, we packaged siRNA against green fluorescent protein (GFP) into RVG exosomes and injected them into GFP-transgenic mice through the tail vein. Then, the GFP levels in various tissues were determined by measuring fluorescence emission using a fluorescence microscope. Compared with control mice, injection of the RVG exosomes loaded with GFP siRNA dramatically reduced GFP levels in different parts of the brain of GFP-transgenic mice. In contrast, unmodified exosomes loaded with GFP siRNA did not induce obvious GFP silencing in mouse brain. However, while unmodified exosomes loaded with GFP siRNA had a significant effect on GFP levels in the lungs, livers and spleens of GFP-transgenic mice, RVG exosomes loaded with GFP siRNA only induced slight but non-significant GFP silencing in these tissues. The results successfully demonstrate that exosome-packaged siRNA can be delivered to various tissues, thus silencing endogenous gene expression. The results also indicate that RVG peptide on the surface of exosomes has some selectivity for neuronal tissues, which may simultaneously prevent siRNA from spreading to non-neuronal tissues.
Figure 8. Fluorescence confocal microscopy images showing sections from different tissues of GFP-transgenic mice. GFP-transgenic mice were intravenously injected with saline (control) or with GFP siRNA loaded into normal exosomes (siRNA-exosome) or RVG exosomes (siRNA-RVG exosome).
3 Silencing capability validation (in vitro and in vivo)
3.1 RVG exosomes loaded with MOR siRNA specifically reduce MOR expression in neuronal cells
We next evaluated the effect of RVG exosome-delivered siRNA on MOR expression in vitro. MOR expression levels were assayed in Neuro2A cells after treatment with RVG exosomes loaded with MOR siRNA. Compared with control cells, MOR protein and mRNA levels were dramatically reduced in cells treated with RVG exosome-delivered siRNA, while no reduction in the MOR protein and mRNA levels were observed in cells treated with exosomes without the RVG peptide on their surface. These results suggest that the RVG peptide modification on the exosome membrane can specifically guide exosomes to target neuronal cells, allowing for the delivery of MOR siRNA into neuronal cells to reduce MOR expression levels.
 |
 |
Figure 9. RVG exosome-delivered siRNA specifically enters Neuro2A cells and reduce MOR expression. Left panel: Western blot analysis of MOR protein levels in untreated control Neuro2A cells or cells treated with MOR siRNA loaded in normal exosomes or RVG exosomes. Right panel: qRT-PCR analysis of MOR mRNA levels in untreated control Neuro2A cells or cells treated with MOR siRNA loaded in normal exosomes or RVG exosomes.
3.2 The effects of siRNA delivered by RVG exosomes on morphine-induced CPP
MOR and its signaling pathway are known to be involved in the dependence and relapse of opioids such as morphine and heroin. Importantly, relapse always disrupts the process of opioid withdrawal. Subsequently, we focused on investigating the effect of exosomal siRNA of MOR on opioid relapse. We evaluated the consequences of MOR knockdown by exosomal siRNA in animals by conducting the morphine-induced conditioned place preference (CPP) test, a mouse model of morphine wanting/liking behaviors. In CPP test, researchers place an animal in a distinctively designed chamber and inject or infuse it with the substance being studied (e.g., morphine in this study). Once trained to associate the chamber with the test agent, the animal is placed in an anteroom connecting a neutral chamber and the drug-associated chamber. The animal indicates its preferred chamber by spending more time there; its choice reveals the animal’s preference for or aversion to the test substance. In our CPP test, mice learn to associate the rewarding effect of morphine with a drug-paired environment. The CPP test was designed to mimic the process of relapse of morphine. Before conditioning, the mice showed a preference for visiting the black chamber. Then, morphine dependence was developed when the mice were place-conditioned by intraperitoneal injection with morphine in the white chamber on even-numbered days (on days 2, 4, 6, 8 and 10) and with saline in the black chamber on odd-numbered days (on days 3, 5, 7, 9 and 11). On day 12, CPP test 1 was conducted by allowing the mice to freely visit the morphine-paired white chamber or saline-paired black chambers. As expected, the mice showed a significant preference for visiting the morphine-paired white chamber, suggesting the development of morphine dependence. Then, the morphine treatment was removed for several days. On day 26, CPP test 2 was conducted, and the mice spent less time in the morphine-paired white chamber than in the saline-paired black chamber, suggesting the disappearance of morphine dependence. Then, the mice were intravenously injected with saline or with siRNAs loaded into normal exosomes or RVG exosomes once every two days for a total of four times, and CPP test 3 was performed on day 32. The mice maintained their natural preference for the black chamber, suggesting that MOR siRNA had no effect on the behavior of the mice. Finally, the mice were relapsed on morphine on day 33, and CPP test 4 was performed the next day. Interestingly, the mice treated with RVG exosome-delivered siRNAs maintained their natural preference for the black chamber, while the mice treated with saline or with siRNAs loaded into normal exosomes showed a preference for the morphine-paired white chamber, suggesting that the MOR siRNAs delivered by RVG exosomes restrained drug addiction when the mice were re-exposed to morphine.
Figure 10. The effects of siRNA delivered by RVG exosomes on morphine-induced CPP. A flow chart depicting the experimental design is shown. The upper panel is represented by the value of the time the mice stayed in morphine-paired white chamber minus the time the mice stayed in the saline-paired black chamber (from left to right: pre-test, CPP test 1, CPP test 2, CPP test 3, CPP test 4).The lower panel is the representives of the heatmap of the mouse mobility.
3.3 The effects of siRNA delivered by RVG exosomes on MOR expression in vivo
After the CPP tests, the mice were sacrificed, and total RNA and protein were extracted from the mouse brains to evaluate the expression levels of MOR in vivo. Both MOR protein and mRNA levels were reduced in the mice treated with RVG exosome-delivered siRNA. In contrast, siRNAs delivered by unmodified exosomes could not reduce MOR mRNA and protein levels in mouse brains. Thus, these results clearly demonstrate that exosomes with RVG modification passed through the BBB and delivered MOR siRNA into the central nervous system to regulate MOR expression, while natural exosomes without the RVG modification were not capable of delivering siRNA into the central nervous system or of regulating target gene expression.
 |
 |
Figure 11. RVG exosomes can transfer MOR siRNA through the BBB and reduce MOR expression levels in vivo. (A) Western blot analysis of MOR protein levels in the brains of mice following injection with saline or with MOR siRNA loaded into normal exosomes or RVG exosomes. (B) qRT-PCR analysis of MOR mRNA levels in the brains of mice following injection with saline or with MOR siRNA loaded into normal exosomes or RVG exosomes.
4 Safety validation
MOR plays a critical role in mediating the rewarding effects of brain. A major concern of our project is whether transiently block dopamine-mediated reward system by targeting at neurons and downregulating MOR may cause any physiological disorder and related safety issues, namely depression, reinforcing addict’s tendency towards negative mood and behaviors. In order to assess the side effect of our medicine on the brain, we performed a classic experiment called forced swimming test.
Forced swimming test was developed by Porsolt and is based on the assumption that a rodent will try to escape an aversive (stressful) stimulus and centered on a rodent’s response to the threat of drowning, whose result has been commonly interpreted as measuring susceptibility to negative mood and the effectiveness of antidepressants. In the forced swimming test, the animal is placed in a container of water from which it cannot escape. Most animals will attempt to escape by actively swimming. When the animal stops swimming and floats on the surface of the water it is considered to have “given up”. An animal that gives up relatively quickly is thought to be displaying characteristics similar to human depression. The validity of this test stems from the finding that physical or psychological stress (which can induce depression in humans) administered prior to the test causes animals to give up sooner and treatment with an antidepressant drug will increase the time an animal spends in escape attempts.
In our forced swimming test, mice were randomly divided into control group and test group. Mice in the control group received saline injection and mice in the test group received exosome injection (200 μg RVG exosomes loaded with MOR siRNA per mouse). After two days, mice were subjected to two trials during which they were plunged individually and forced to swim in a tank (height 30 cm, length 50 cm, width 40 cm) filled with 15 cm of water maintained at 25℃. The first training trial lasted 15 min and was performed one day before testing. Then, after 24 h, a second trial was performed and lasted for 10 min. The time that the test animal spent in the second trial without making any movements beyond those required to keep its head above water was measured and further analyzed. Both of two trials were videoed.
Classically, immobility in the second test has been interpreted as a behavioral correlate of negative mood, representing a kind of hopelessness in the animal. In our test, mice receiving exosome (RVG exosomes loaded with MOR siRNA) injection swam as hard as control mice. The total duration of active time showed no significant difference between mice in two the groups (P > 0.05). The test indicated that knocking down MOR in the neuronal cells by RVG exosomes-delivered MOR siRNA did not bring explicit negative moods and depression on the physiological status of mice.
Figure 12. Mobile time of mice injected with saline or RVG exosomes loaded with MOR siRNA.
Figure 1. Relative levels of MOR mRNA in Neuro2A cells after transfection of MOR siRNA plasmids.
Figure 2. The concentration of MOR siRNA in unmodified or RVG-modified exosomes.
Figure 3. TEM images of the exosomes with outside RVG modification and inside siRNA loading.
