Our Research Focus Areas
Delving into the Mechanisms and Therapeutic Potential of Our Research
The Kay lab has a long history of work in the area of Gene Therapy/Genome editing and Non-coding RNA biology. A historical summary of selected accomplishments is outlined in the following sections.
Pioneering Gene Therapy with rAAV vectors
Our early work pursued retroviral, lentiviral, and adenoviral vectors for gene therapies. In the mid 1990s we shifted our attention to recombinant AAV vectors. We were the first to demonstrate successful rAAV mediated liver gene transfer in small and large animals, which set the stage for the first-in-man systemic delivery of rAAV vectors. Dr. Mark Kay was the original Sponsor (IND holder) of this first trial. Although the AAV2 vector capsid induced an immune response in humans unlike non-human primates, our lab has since developed more robust AAV vectors that have entered two clinical trials.
Key Accomplishments & Discoveries:
- Pioneering rAAV-mediated liver gene transfer: First to demonstrate successful rAAV gene transfer in small and large animal models.
- First-in-man- systemic delivery: Set the stage for the first human clinical trial using systemically derived rAAV vectors.
- Immune Response Insights: Identified host immune responses to AAV2 vector capsids in humans, unlike animal models.
- Developed Robust AAV Vectors: Created improved AAV vectors that have advanced to clinical trials.
Selected Publications:
Advancing AAV Vector Design and Nuclease-Free Genome Editing
Over the years we have made important contributions that involve unraveling the mechanism of AAV transduction in vivo. We specifically focused on how the vector goes from a single-stranded DNA into a double-stranded episomal DNA and how different capsid variants transduce the same cells/tissues differently. We were the first to establish how differential vector uncoating kinetics can affect transduction parameters.
As a participant in the more recent AAV-8 hemophilia B clinical trial, it became clear that there was a difference in predicted dose response based on animal studies. While the AAV-2 dose response was accurately reflected in the human trial, the AAV-8 dose response was more than 10-times less effective in humans. We used a chimeric humanized liver mouse to demonstrate that the dose response result was the result of an inherent difference in species transduction. More importantly, we propose that this xenotransplant model may more accurately predict the dose response even when compared to non-human primates. Recognizing early on how small sequence differences can affect species, cell and tissue transduction parameters, we constructed the first reported multi-species capsid shuffled AAV library and used mutliple selection schemes. We used this library to isolate an AAV chimeric humanized mouse liver model. This serotype is currently being evaluated for use in humans. We had also previously selected an AAV capsid (DJ) that has proven to be especially robust for use in ex vivo and neurobiology applications.
Classical rAAV vectors have two major limitations: 1) episomal genomes are lost during cell division; 2) rAAV-delivery into young mice results in an elevated risk of hepatocellular carcinoma because of the selected growth of cells that have vector promoter insertion near an oncogenic locus. To overcome these limitations, we have developed a promoterless genome targeting vector without the need for a nuclease. In this approach, the vector DNA is designed to facilitate homologous recombination into a desired locus in such a manner that the genomic locus produces a new single mRNA that not only continues to produce the protein from the endogenous locus, but the desired protein encoded in the vector sequence. This approach was used to cure hemophilia B mice and is currently being used in a clinical trial for the treatment of methylmalonic acidemia.
Our work continues to select improved capsids for delivery into the human CNS and pancreas for the treatment of diabetes. We are using advanced chemical modification paradigms for targeting AAV vectors. In addition, we are now pursuing studies showing how the capsid proteins can influence the epigenomic state of the vector and this can explain at least some of the discordance between vector transduction between various tissues and species.
Selected Publications:
Optimizing Non-Viral Gene Delivery: Minicircles and MIPs
Non-viral vectors
We were the first to develop a DNA transposon for gene therapy applications in mammals. While we continued to develop these for additional years, during our study of rAAV-vector transduction, we found that episomal DNA plasmids have the potential to last indefinitely in quiescent tissues. However, canonical plasmids are transcriptionally silenced in the liver. Over the years, we established the mechanisms responsible for silencing and designed robust non-canonical plasmid vectors that were persistently transcribed. We learned that the length (>1kb) rather than the specific sequence of DNA contained outside of the recombinant expression cassette (classically occupied by the bacterial origin of replication and selectable marker e.g., amp or kan) is responsible for silencing. To overcome the silencing effect, we have generated several new plasmid variants that are becoming more popular in the gene therapy community. Two such vectors are named minicircle and mini-intronic plasmid vectors (MIP). Both provide 10-1000 times more persistent expression when delivered into quiescent tissues compared to their canonical plasmid counterparts. Additional mechanistic findings in the epigenomic state of the vector DNAs in whole tissues are providing new general insights into general eukaryotic transcriptional paradigms.
Selected Publications:
Decoding RNA Interference (RNAi) and MicroRNA (miRNA) Biogenesis
RNAi based therapies and miRNA biogenesis/function
We published the first study establishing the use of siRNA and transcriptional RNAi in whole mammals and subsequently worked towards developing a platform for delivering AAV-shRNA vectors. Our first targets were hepatitis viruses B and C. During our studies we found that overexpression of shRNAs can induce liver toxicity and even fatality. From this point, we continued studies to unravel the mechanism of shRNA overexpression and hepatic toxicity. Along the way we explored various mechanisms involved in miRNA biogenesis. As a result, we started to investigate the mechanisms involved in miRNA-mediated gene regulation, processing, and RISC loading. Over the years, we have provided new insights into Dicer processing of both exogenously expressed miRNAs/shRNAs and endogenous miRNAs and discovered a Dicer loop counting rule. This rule when applied to shRNA design can drastically improve the homogenous products derived from transcriptional shRNAs increasing efficiency and decreasing the off targeting. We recently discovered a new function for some pri/pre-miRNAs. In this example, the precursor can bind to some target mRNAs and protect the target from the action of the mature miRNA. Thus, in such examples, the ratio of the primary/precursor and mature miRNA is what dictates the degree of mRNA down regulation. Furthermore, this provides an example where a single miRNA locus can regulate two mRNAs differently in the same cell. During our studies we have discovered a new class of non-coding RNAs derived from tRNAs. This is discussed in the next section.
Importantly, these studies have resulted in a novel finding that a long-non-coding RNA (Inc122), who's only known function is as a precursor to the hepatic specific miR122 present in 50,000 copies per hepatocyte has a separate and sometimes opposing function in regulating liver homeostasis and cancer.
Selected Publications:
Uncovering the Regulatory Power of tRNA-Derived Small RNAs (tsRNAs)
Gene regulation of tRNA derived small RNAs
We recently found that a 22 nucleotide (nt) 3' end of the LeuCAG transfer-RNA-derived small RNA (LeuCAG3'tsRNA) binds to the human RPS28 mRNA, unwinds the double-stranded secondary structure, which enhances RPS28 mRNA translation. Small changes in RPS28 protein production were also shown to regulate rRNA processing and ribosome biogensis. Inhibition of this specific tsRNA inducted apoptosis in rapidly dividing cells in culture and suppressed the growth of human hepatocellular carcinomas in vivo, making it a bona-fide target for cancer therapeutics. A decrease in translation of RPS28 mRNA blocks pre-18S ribosomal RNA processing, resulting in a reduction in the number of 40S ribosomal subunits.
These data establish a post-transcriptional mechanism that can fine-tune gene expression during different physiological states and provide a potential new target for treating cancer. We also found that RPS28 mRNA and ribosome biogenesis were similarly regulated by the same 3' ts RNA in the mouse. In addition, using various inhibitors of protein synthesis and polysome gradient analysis in both mouse and human cells we establish that the 3'tsRNA regulates translation at the elongation step. We also found that the 3'tsRNAs are generated from charged tRNAs providing more insight into their regulation. Our results suggest a conserved functional role for 3'tsRNAs to fine tune translation of mRNAs. We propose that this may represent a feedback loop to regulate the components of protein translation and may represent new targets for treating cancer. We are currently exploring the expression patters of the 161 3'tsRNAs in tissues including cancer. Using various screening approaches, we have identified several 3'tsRNAs that may also regulate proteins involved in translational regulation.