OAS are differentially expressed in the cells of the NVU
While an association was demonstrated between the OAS gene expression and the progression of several viral infections and autoimmune diseases [44-47], no studies have characterized the expression of the members of the OAS family in human brain or brain microvasculature. Here, we describe the protein expression levels of OAS1, OAS2, OAS3, and OASL in cells forming the NVU, such as primary human brain pericytes, astrocytes, EC, immortalized human microglial cells, and SH-SY5Y neuroblastoma cell line.
Among the cells of NVU, pericytes expressed the highest levels of OAS1, followed by microglia, EC, and SH-SY5Y. The lowest expression of OAS1 was found in astrocytes (Fig. 1A). In the case of OAS2, SH-SY5Y cells expressed significantly higher levels than other studied cells, followed by EC, and pericytes. Microglia and astrocytes expressed the lowest levels of OAS2 (Fig. 1B). For OAS3, the highest expression was found in SH-SY5Y cells, followed by significantly lower expression in astrocytes and microglia (Fig. 1C). The lowest levels of OAS3 were found in EC and pericytes. Lastly, the highest levels of OASL proteins were detected in microglia and SH-SY5Y cells, with significant lower expression in astrocytes, EC, and pericytes (Fig. 1D). Overall, these results indicate a differential expression pattern of the OAS family members in cells composing the NVU, suggesting that these cells are involved in antiviral protection.
Ocln regulates IFN genes and alters the STAT signaling pathway
Ocln has been traditionally consider as a tissue barrier regulating protein; however, recent evidence indicated multifunctional role of this protein in controlling cellular metabolism and HIV-1 infection [14, 16, 22]. Therefore, we evaluated the impact of ocln on mRNA and protein expression of the IFN genes as the main component of innate immunity. These experiments focused on pericytes as the NVU cells, which can harbor HIV-1 infection [6, 13, 16, 18]. Pericytes were transfected with the PCMV3-OCLN expression vector, and the expression levels of several IFN genes were evaluated by real time q-PCR. The results indicated that ocln overexpression led to significantly increased IFNα5 (Fig. 2A), and IFNβ (Fig. 2C). In contrast, no significant changes were found in IFNα2 (Fig. 2B) and the expression of IFNγ where not detectable.
We next analyzed the STAT signaling pathway that initiate the transcription of IFN-stimulated genes (ISGs) [48, 49]. One of the key elements in the STAT signaling pathway are the signal transducer and activator of transcription (STAT)1 and STAT2 proteins, which upon phosphorylation, translocate into the nucleus where they initiate the transcription of ISGs. Therefore, we examined if modulations of ocln levels can alter the expression of STAT proteins. The results indicate that ocln overexpression markedly induced the expression of STAT1 at the gene and protein levels but not STAT2 (Fig. 2F-G). Moreover, ocln overexpression led to an increase in phosphorylated STAT1 (pSTAT1) at Tyr701 (Fig. 2F). To analyze the functional consequences of ocln upregulation on the activity of STAT1, cells were co-transfected with firefly luciferase constructs under the control of the STAT1 promoter. As shown in Fig. 2E, ocln upregulation led to an increase in STAT1 binding activity. Moreover, interferon regulatory factors (IRFs) are molecules that execute positive feedback with type I IFN. For example, activated STAT1 recruits IRF9 forming a complex that translocate to the nucleus. Consistent with this mechanism, ocln overexpression resulted in upregulation of IRF9 gene expression (Fig. 2D).
Ocln regulates OAS expression levels
We next focused on the OAS genes and protein family as a prominent component of native immunity regulated by IFN. As in Figure 2, pericytes were transfected with the PCMV3-OCLN vector for ocln overexpression or with the PCMV3 vector as a negative control, and the expression of ocln, OAS1, OAS2, OAS3 and OASL was analyzed by qPCR and immunoblotting. Ocln overexpression led to remarkably significant increase in OAS1, OAS2, OAS3, and OASL mRNA and protein (Fig. 3A-E, respectively) levels. Interestingly, this increase was notably higher for OASL when compared to OAS1, OAS2 or OAS3.
In the next series of experiments, we measured the expression of the OAS family in pericytes with silenced ocln gene. The controlled experiments were performed with silenced ZO-1, another tight junction protein to determine specificity of ocln-mediated responses. Briefly, pericytes were transfected with control siRNA, ocln siRNA, or ZO-1 siRNA, and the expression of the OAS genes was analyzed by q-PCR. Ocln silencing (Fig. 4A), but not ZO-1 silencing (Fig. 4B), resulted in a significant decrease in the expression of OAS1, OAS2, OAS3, and OASL mRNA (Fig. 4C-F). Along with Figure 3, these results indicate a regulatory influence of ocln on the expression of the members of the OAS family. They also suggest a novel mechanism by which ocln can influence innate immunity and protect against viral infection.
OASL, but no other members of the OAS family, alters ocln expression levels
Because ocln can modify the expression of the OAS genes, we next investigated if the reverse modulation can also occur. Pericytes were transfected with control siRNA, OAS1 siRNA, OAS2 siRNA, OAS3 siRNA, or OASL siRNA, and ocln mRNA and protein expression levels were measured by qPCR and immunoblotting, respectively. The efficiency of silencing of the individual OAS genes was confirmed by qPCR (Fig. 5A). We then measured ocln expression in these samples. Among studied OAS genes, only OASL silencing led to a relatively small but a significant decrease in ocln levels at mRNA (Fig. 5B) and protein levels (Fig. 5C). In contrast, no changes in ocln expression were found after silencing the remaining members of the OAS family.
Cross-regulation of the expression among the OAS family members
The OAS family members may have overlapping functions in regulation of innate immunity; therefore, we examined whether they could interact among themselves and regulate each other´s expression. Such study has never been performed in the literature. Pericytes were transfected with control siRNA, OAS1 siRNA, OAS2 siRNA, OAS3 siRNA, or OASL siRNA, and mRNA and protein expression levels were measured for individual members of the OAS family. OAS1, OAS2 and OAS3 silencing significantly reduced OASL mRNA levels but not by other members of the OAS family (Fig. 6A-C). Furthermore, OASL silencing significantly decreased OAS1 mRNA levels; the effect, which did not apply to OAS2, or OAS3 (Fig. 6D).
At the protein level, downregulation of OAS1 decreased OAS2 expression, but no changes were found in OAS3 or OASL expression (Fig. 6A). OAS3 silencing increased the expression of OAS2 protein but it did not alter OAS1 or OASL protein levels (Fig. 6C). Furthermore, no changes were detected in the expression of any OAS members after OAS2 or OASL downregulation (Fig. 6B, D). These results indicate that individual members of the OAS family can influence each other expression; however, this input appears to be highly specific.
HIV-1 infection alters the expression levels of the OAS genes and proteins
The OAS family has been studied in viral infections and autoimmune disorders; however, only limited information is available on the role of these protein in HIV-infection [23, 50]. Moreover, no studies have defined the impact of the OAS gene family on HIV-1 infection in human brain pericytes. To investigate this relationship, pericytes were mock‐infected or infected with HIV‐1 for 24, 48 or 72 hours, and the mRNA and protein levels of individual members of the OAS family were evaluated. Infection with 60 ng/ml of HIV-1 for 24 or 48 hours resulted in a significantly increase in the expression of OAS1, OAS2, OAS3, and OASL at the mRNA level (Fig. 7A-D). Interestingly, this increase was gradually reduced over time, and only OAS2 and OAS3 showed a slight increase in mRNA expression after 72 hours of infection as compared to mock infection (Fig. 7B, C).
At the protein levels, there were notable differences in the response of individual OAS members to HIV-1 infection. The expression of OAS1 protein was increased only 24 hours after infection and returned to control levels after a longer infection period (Fig. 7A). There was a significant increase in the expression of OAS2 and OAS3 proteins 24, 48 or 72 hours post HIV-1 infection; however, these changes were more prominent for OAS 2 than those for OAS3 (Fig. 7B, C). Finally, no changes were detected in OASL protein levels after HIV-1 infection (Fig. 7D).
Ocln regulates HIV-1 infection through an OAS-mediated mechanism
Studies from our laboratory have shown that human brain pericytes can regulate the extent of HIV-1 infection in various cell types, including brain pericytes [14, 16].
Pericytes were transfected with ocln siRNA, OAS1 siRNA, OAS2 siRNA, OAS3 siRNA, or OASL siRNA and cultures were either mock-infected or infected with HIV-1 for 12 hours, followed by extensive washing to remove the unbound virus before addition of fresh medium. The levels of p24 antigen, the major structural component of HIV-1, were analyzed 48h after infection in the supernatants of cell cultures as the indicator of active HIV-1 replication. Downregulation of ocln resulted in increased p24 levels (Fig. 8A). Most interestingly, silencing of OAS1, OAS2, OAS3, or OASL markedly increased HIV-1 replication in human brain pericytes (Fig. 8B), providing the first evidence that the OAS family can regulate HIV-1 infection in human pericytes.
HIV-1 infection of human brain pericytes does not affect RNaseL expression
Activated OAS can catalyze the oligomerization of ATP into 2´, 5- linked oligoadenylates (2-5A), which then activates RNaseL, one of the key elements of the OAS/RNaseL pathway by catalyzing ssRNA or rRNA [51]. Given the lack of information about RNaseL in human brain pericytes, we first aimed to characterize the expression of this endoribonuclease in individual cell types forming the NVU. Interestingly, pericytes along with microglia cells, express RnaseL to the highest extension, with a significantly lower expression in astrocytes, SH-SY5Y, and EC (Fig. 9A).
We next investigated whether RNaseL levels could be influenced by HIV-1 infection or by changes in ocln expression levels. No changes were found in the expression of RNaseL at mRNA or protein levels (Fig. 9C) after HIV-1 infection. Moreover, no changes were detected in RNaseL mRNA or protein levels after ocln overexpression (Fig. 9B).