Browsing by Author "Cullen, Bryan R"

Epstein-Barr virus (EBV) is a ubiquitous human gamma-herpesvirus which chronically infects >95% of the global population, and can give rise to a number of malignancies in B cells and epithelial cells. In EBV latently infected epithelial cells, such as nasopharyngeal carcinoma (NPC) and gastric carcinoma (GaCa) cells, viral protein expression is low. In contrast, a cluster of viral microRNAs (miRNAs) called miR-BARTs is highly expressed. MiRNAs are small non-coding RNAs which regulate gene expression by binding to complementary sequences in mRNAs. It is likely that miR-BARTs play a crucial role in EBV-infected epithelial cells, however a comprehensive understanding of miR-BARTs is currently lacking. Here, I present two studies utilizing the phenotypic and the target approaches, respectively, to demonstrate that miR-BARTs can inhibit apoptosis and activate the Wnt signaling pathway. To discover miR-BARTs that can inhibit apoptosis, I individually expressed miR-BARTs in the EBV- GaCa cell line AGS, and identified five miR-BARTs that conferred this phenotype. To identify pro-apoptotic genes targeted by the five anti-apoptotic miRNAs, I validated one previously published target and identified nine novel targets by performing photoactivatable ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP) in the EBV+ NPC cell line C666. Next, I thoroughly demonstrated that the 10 candidate target genes were substantially suppressed by expression of the relevant miR-BARTs, as measured by 3’UTR-containing firefly luciferase (FLuc) expression, mRNA and protein levels, and knockdown of seven of the 10 candidate genes could suppress apoptosis, mimicking the effects of relevant miR-BARTs. On the other hand, in order to identify miR-BARTs that can activate the Wnt signaling pathway, I analyzed the PAR-CLIP data set of C666 cells and discovered nine anti-Wnt signaling targets of miR-BARTs, including seven novel genes and two pro-apoptotic genes identified above. Using FLuc 3’UTR indicator assays, I proved that the 3’UTRs of all seven newly identified anti-Wnt signaling genes were indeed targeted by the relevant miR-BARTs identified by PAR-CLIP. Utilizing a Wnt signaling FLuc reporter TOPflash which measures the Wnt signaling activation, I confirmed that expression of many miR-BARTs that target Wnt signaling inhibitors can indeed upregulate the Wnt signaling pathway. Together, my results identified and validated a substantial number of novel targets of miR-BARTs involved in apoptosis and the Wnt signaling pathway, indicating that EBV may employ miR-BARTs to heavily target these two pathways to facilitate chronic infection.

The CRISPR/Cas system has emerged as a powerful tool for the precise modification of DNA in mammalian systems. In addition to the robust utility of Cas9 loaded with its cognate single guide RNA (sgRNA) to modify cellular DNA targets, there is great interest in using this system to target the DNA of viral targets. Human immunodeficiency virus type 1 (HIV-1) remains a major human pathogen with currently no curative therapy available and serves as an ideal target for disruption through CRISPR/Cas. Indeed, multiple reports have demonstrated the effectiveness of the CRISPR/Cas system to impair the ability of HIV-1 to successfully replicate. However, there is now evidence that this repression is not permanent, with HIV-1 being to escape Cas9-mediated inhibition through mutation of the target site. Here, we demonstrate through in vitro cell culture assays that not only is CRISPR/Cas an effective tool to inhibit de novo HIV-1 infection and that cleavage of the proviral DNA occurs both before and after integration, but that targeting of the highly conserved TAR stem-loop structure can prevent HIV-1 breakthrough replication.

In addition to targeting HIV-1, multiple reports have demonstrated the effectiveness of targeting other DNA viruses in vitro. There is now significant enthusiasm for translating these studies into an in vivo setting. Thus, the development of vector technology that can effectively deliver the Cas9 and sgRNA components to target cells in vivo is of critical importance. The most likely vectors for these studies would be adeno-associated virus (AAV) vectors due to their potent ability to transduce various tissue types in vivo. Unfortunately, AAV vectors are hampered by a strict packaging limit and, given the large size of the Cas9 protein, technology that can aid in overcoming this barrier is necessary. Here, we characterize a novel and compact method of sgRNA expression using tRNAs as promoters. With this method, the sgRNA is initially transcribed as a long precursor transcript that is subsequently cleaved by tRNAse Z, liberating the mature sgRNA where it is loaded into Cas9. Importantly, these sgRNAs are able to guide Cas9 to a target sequence to an equivalent degree compared to sgRNAs expressed from a canonical, and substantially larger, U6 promoter. These findings should greatly facilitate the development of AAV vector technology that can efficiently deliver CRISPR/Cas components to target cells for in vivo studies.

The cellular microRNA (miRNA) pathway has emerged as an important regulator of host-virus interactions. While miRNAs of viral and cellular origin have been demonstrated to modulate viral gene expression and host immune responses, reports detailing these activities in the context of mammalian retroviruses have been controversial. Using modern, high-throughput small RNA sequencing we provide evidence that the spumaretrovirus bovine foamy virus expresses high levels of viral miRNAs via noncanonical biogenesis mechanisms. In contrast, the lentivirus human immunodeficiency virus type 1 (HIV-1) does not express any viral miRNAs in a number of cellular contexts. Comprehensive analysis of miRNA binding sites in HIV-1 infected cells yielded several viral sequences that can be targeted by cellular miRNAs. However, this analysis indicated that HIV-1 transcripts are largely refractory to binding and inhibition by cellular miRNAs. In addition, we demonstrate that HIV-1 exerts minimal perturbations on cellular miRNA profiles and that viral replication is not affected by the ablation of mature cellular miRNAs. Together, these data demonstrate that the ability of retroviruses to encode miRNAs is not broadly conserved and that lentiviruses, particularly HIV-1, have evolved to avoid targeting by cellular miRNAs.

Glycogen storage disease type Ia (GSD Ia) is an autosomal recessive metabolicdisorder caused deficiency of glucose-6-phosphatase (G6Pase) resulting from pathogenic variants in the G6PC gene. G6Pase catalyzes the hydrolysis of glucose-6-phosphate to release glucose which can then enter the bloodstream. GSD Ia patients have excess glycogen accumulation mainly in the liver and kidneys and suffer from life threatening hypoglycemia. The current treatment for GSD Ia is dietary therapy that requires patients to frequently consume uncooked cornstarch on a strict schedule. Cornstarch provides a complex carbohydrate that slowly releases glucose to prevent hypoglycemia. This treatment fails to prevent long-term complications associated with GSD Ia including renal failure and the development of hepatocellular adenomas and carcinomas. This lab and others have developed adeno-associated virus (AAV) vector based gene therapies to deliver and therapeutic G6PC transgene to affected tissues in GSD Ia animal models. However, the therapeutic effect is limited as AAV vector genomes are rapidly lost and the biochemical correction declines. Currently no treatment for GSD Ia exists that provides stable, robust expression of G6Pase that can clear glycogen and prevent hypoglycemia. This study employed a novel genome editing approach designed to insert the therapeutic G6PC into the endogenous locus in canine and murine models of GSD Ia. Integration of the transgene into the genome will promote stable expression of G6Pase and prevent the decline of vector genomes and the therapeutic benefit. This genome editing approach utilizes the CRISPR/Cas9 system to generated targeted double stranded DNA breaks at a targeted site in the genome. The G6PC transgene is present in a Donor template with homology to the DNA break to drive homology directed repair (HDR) resulting in the integration of the transgene into the genome. In a canine model of GSD Ia, editing and incorporation of the transgene was achieved in both adult dogs and puppies. Up to 1.0% of alleles were edited in the dog livers and contained the transgene. G6Pase production from the integrated transgene was detected, which correlated with prevention of hypoglycemia during fasting. This demonstrated genome editing in the liver of a large animal model for an inherited metabolic disorder using HDR to insert a therapeutic transgene. A subsequent study in GSD Ia mice also showed incorporation of a G6PC transgene in the mouse genome. Mice were treated with either the Donor transgene vector alone or with both the Donor and a CRISPR/Cas9 vector to assess to role of nuclease activity on integration. Mice treated with both vectors demonstrated improved blood glucose concentrations during fasting, decreased liver glycogen, and increased vector genome copies. Treatment with the pan PPAR agonist bezafibrate increased the efficiency of genome editing. Mice treated with bezafibrate that received both editing vectors had 5.9% of alleles that contained the integrated transgene, whereas only 3.1% of alleles contained the transgene in mice not treated with the drug. This work showed that integration of a therapeutic transgene using CRISPR/Cas9 based genome editing is possible in murine and canine models of GSD Ia. Editing resulted in biochemical correction and sustained transgene expression. These data support the further development of genome editing technologies for GSD Ia and other inherited metabolic disorders.

Integration of HIV-1 DNA is an essential step in the viral life cycle, and how the unintegrated proviral DNA intermediate is transcriptionally silenced has been an ongoing question in the field. Here we show that this transcriptional silencing is epigenetic and demonstrate that the Tax transcription factor encoded by Human T-cell Leukemia Virus 1 (HTLV-1) can reverse the repressive epigenetic modifications on unintegrated, episomal HIV-1 circular DNA. Tax expression in HIV-1 mutants lacking functional integrase, and are normally transcriptionally silenced, causes the recruitment of NF-κB to unintegrated viral DNA promoter, reverses the repressive epigenetic modifications on the chromatinzed viral DNA, and leads to a robust, spreading infection. In addition, using an unbiased screen, we identify the host SMC5/6 complex as essential for epigenetically silencing unintegrated/ pre-integrated HIV-1 DNA. The SMC5/6 complex binds chromatinized HIV-1 DNA and triggers epigenetic silencing by inducing its SUMOylation through NSMCE2, an E3 SUMO ligase. Inhibiting this SUMOylation, by knocking out members of the SMC5/6 complex, mutationally disrupting its E3 ligase function, or by using a SUMOylation inhibitors like TAK-981 prevents this epigenetic silencing and rescues both viral gene transcription from, and replication of, integrase-deficient HIV-1. Finally, we show that inhibiting this initial SUMO-mediated silencing of HIV-1 DNA in integration-competent viruses with a functional integrase by the SMC5/6 complex drastically interferes with the establishment of latent HIV-1 infections in both CD4+ T cell lines and primary human T cells. We thus propose a new model to explain the establishment of the HIV-1 latent reservoir.

Epstein Barr virus (EBV), a γ-herpesvirus that latently infects over 90% of the adult population, is known to play a role in the development of both B cell and epithelial cell tumors. In a laboratory setting, transformation of primary B cells into lymphoblastoid cell lines (LCLs) is used to model EBV infection and latency. LCLs have a latency III gene expression profile akin to the EBV-associated tumors diffuse large B cell lymphoma (DLBCL) and post-transplant lymphoproliferative disorder (PTLD). During latency III, EBV expresses the full repertoire of latency genes including two microRNA (miRNA) clusters, the BHRF1 cluster and the BART cluster. The BHRF1 miRNA cluster is composed of three virally encoded primary miRNAs (pri-miRNAs) which are processed into four highly expressed mature ~22nt miRNAs that target mRNA for translational inhibition or degradation: miR-BHRF1-1, 1-2, 1-2*, and 1-3. Virally encoded miRNAs have been shown to enhance pathogenesis by inhibition of cellular genes. Thus, characterizing the role the BHRF1 miRNAs during latency is important for understanding EBV pathogenesis.

Collectively, the BHRF1 miRNAs are crucial for EBV-induced transformation of primary B cells into LCLs. They also play a role in LCL proliferation by enhancing cell cycle progression of LCLs. However, the individual roles of the miRNAs in the BHRF1 cluster were previously not known. We sought to address this gap in knowledge by ectopically expressing the BHRF1 miRNA cluster in LCLs transformed with an EBV bacmid that lacks the BHRF1 miRNAs (∆123 LCLs). We hypothesized that ectopic expression of the BHRF1 miRNAs would rescue the growth deficiency of ∆123 LCLs. Because of technical difficulties inherent to previously established lentiviral vectors we developed a new lentiviral vector bearing a reverse intron (pTREX) to express the BHRF1 miRNAs at physiological levels. We demonstrated by flow cytometry, Western blot, and TaqMan qPCR that pTREX was better at expressing transgenes and miRNAs than a lentiviral vector without an intron (pTRIPZ). We additionally showed by infection assays that pTREX generated higher viral titers than pTRIPZ when encoding a cluster of miRNAs. Thus, we were able to use the pTREX vector to show that ectopic expression of the BHRF1 miRNAs at WT levels in ∆123 LCLs does not rescue the ∆123 LCL growth deficiency.

Since the BHRF1 miRNAs did not regulate growth through their canonical role in trans, we investigated other processes that were dysregulated in ∆123 LCLs. In doing so, we demonstrated that the BHR1-2 and 1-3 pri-miRNAs play a novel role in viral gene expression. Processing, by the RNAse III protein Drosha, of the BHRF1 miRNAs out of EBNA-LP transcripts, as part of the miRNA biogenesis pathway, allows the BHRF1 miRNAs to regulate EBNA-LP expression in cis. We used a meta-analysis of RNA-Seq data and performed PacBio sequencing to show that splicing from EBNA-LP to the HF exon, which contains the BHRF1-2 and 1-3 pri-miRNAs, occurs at a high frequency. This had not been previously established due to the limitations of short reads used in RNA-Seq analysis. Our PacBio sequencing also revealed several novel splice junctions that prevent expression of EBNA-LP from the Cp and Wp promoters. We also show, using a reporter assay, that the BHRF1-2 and 1-3 pri-mRNAs are able to regulate gene expression in cis in a Drosha-dependent manner. Furthermore, targeting of the HF exon with amiRNAs was capable of reducing EBNA-LP expression. We propose that during EBV infection the BHRF1 miRNAs are encoded in the 3’UTR of EBNA-LP, and that their processing by Drosha regulates EBNA-LP protein expression. We show, by 4SU labeling of nascent transcripts and Western blot, that the increase in EBNA-LP protein expression observed in ∆123 LCLs is correlated with an increase in EBNA-LP regulated transcripts and proteins. Taken together, we developed a new lentiviral vector, pTREX, which allows for high protein and miRNA expression, and revealed that the growth deficit of ∆123 LCLs is not due to miRNA regulation of cellular mRNAs in trans. By contrast, we demonstrate a novel role of the BHRF1 miRNAs in regulating the expression of the virally-encoded EBNA-LP protein through Drosha processing in cis. This is an underappreciated role for miRNAs as there are very few examples of Drosha processing regulating gene expression in cis. These results suggest that regulation of EBNA-LP expression as important for LCL growth and transformation. Further studies examining the role of EBNA-LP expression are necessary for understanding EBV-induced transformation.