2024 Symposium Poster Session Abstracts - Center for RNA Biomedicine

8th Annual RNA Symposium Poster Session Abstracts

8th Annual RNA Symposium

March 8, 2024 || Biomedical Science Research Building || Kahn Auditorium

Abstracts

Max Baymiller,¹ Stephanie L. Moon¹

1 Department of Human Genetics, University of Michigan, Ann Arbor, MI

Cells suppress translation initiation when stressed, leading to the formation of biomolecular condensates of non-translating mRNAs known as stress granules. A requirement for stress granule assembly is that mRNAs are not associated with polysomes. However, stress can cause ribosomes to stall during elongation, and these stalled ribosomes are emerging as key signaling platforms for stress pathways. We hypothesized that such stressors break an RNA cycle by preventing mRNAs from exiting translation and condensing into stress granules. Here we tested this by targeting aminoacyl-tRNA synthetases to disrupt translation elongation. We found that inhibition of prolyl-tRNA charging with the chemical halofuginone caused phosphorylation of the translation initiation factor eIF2. Typically, P-eIF2 drives stress granule formation, and blocking P-eIF2 function inhibits stress granules. Yet we observed only infrequent and atypically small stress granules in cells treated with halofuginone. Predicting that this defect in mRNA condensation was due to stalled ribosomes, we used the translation inhibitor puromycin to eject mRNAs from polysomes. Excitingly, co-treatment with puromycin and halofuginone led to robust stress granule assembly. Prolonged treatment with halofuginone also inhibited formation of stress granules by arsenite, which was rescued by puromycin. Treatment with halofuginone is predicted to affect mRNAs in a proline content-dependent manner. Accordingly, we found that after a short halofuginone treatment, arsenite-induced stress granules do form, but are depleted in a candidate proline-rich mRNA. Overall, our results demonstrate that disrupting translation elongation can lead to altered assembly and mRNA content of stress granules. This has implications for how stress granules can regulate translation, and how disease-associated mutations in translation elongation factors may activate stress pathways but prevent mRNA condensation.

Grace McIntyre,^1-3 Jose Colina,^1-3Aaron Robida,⁴ Peter Toogood,⁵ Andrew Alt,⁴ Analisa DiFeo^1-3

1 Department of Pathology, Rackham Graduate School, University of Michigan, Ann Arbor, MI
2 Rogel Cancer Center, University of Michigan, Ann Arbor, MI
3 Michigan Ovarian Science and Innovation Consortium
4 Center for Chemical Genomics, University of Michigan, Ann Arbor, MI
5 Michigan Drug Discovery, University of Michigan, Ann Arbor, MI

High grade serous cancer (HGSC) remains the deadliest gynecological cancer and subtype of ovarian cancer. Despite high genomic instability being a key phenotype of HGSC, there are very few genetic drivers contributing to its progression prompting the exploration of new genetic drivers and therapeutic targets. Our group has discovered that microRNA-181a (miR-181a) is highly expressed in HGSC and promotes oncogenesis through the activation of the WNT and TGF-β signaling pathways thus mediating epithelial to mesenchymal transition and transcription of pro-survival signals. Additionally, miR-181a inhibits innate immune function by suppressing STING, making it an attractive therapeutic target. Therefore, in this study we aim to identify a series of small molecule inhibitors specifically targeting miR-181a. To achieve this, we functionally screened >54,000 drug-like compounds using a previously characterized high-throughput miR-181a biosensor drug screen that can simultaneously assess changes in miR-181a function and cell viability. From this collection, 32 compounds reduced miR-181a activity by >20% while also decreasing cellular viability. Interestingly, of the 32 compounds 5 of the compounds clustered as structurally similar RNA binding molecules. At least 4 of these have been shown to target miR-181a post-transcriptionally and at least 1 reduces TGF-β mediated epithelial to mesenchymal transition. Future cell-based profiling of these compounds will confirm and identify our lead hits to be optimized using structure activity relationship analysis. Identification of a miR-181a small molecule inhibitor will be the first of its kind to be used in HGSC and will establish a novel and foundational therapeutic avenue to be exploited in gynecological cancers.

Guoming Gao,^1-2 Emily Sumrall,^1-2Nils G. Walter^2-3

1 Biophysics Graduate Program, University of Michigan, Ann Arbor, MI
2 Center for RNA Biomedicine, University of Michigan, Ann Arbor, MI
3 Department of Chemistry, University of Michigan, Ann Arbor, MI

Ribonucleoprotein (RNP) granules are phase separated biomolecular condensates that regulate many aspects of cellular RNA metabolism, including transcription, translation, decay, and processing. RNP granules selectively recruit RNAs and thus impart distinct functional fates to them, yet little is known about the spatiotemporal organization of the composite RNAs and proteins within condensates. Here, we used dual-color single-particle tracking (SPT) to measure RNA and protein diffusion within single reconstituted model condensates. We purified a full-length native form of RNA-binding protein Fused in sarcoma (FUS) – known for its important roles in cellular physiology and pathology – and co-assembled it with low concentrations of fluorophore-labeled messenger RNAs (mRNAs), microRNAs (miRNAs) and FUS into RNP granules. SPT showed that long mRNA molecules, but not short miRNAs and FUS, become predominantly corralled within nano-domains with a low, but observable, probability of escape. Immobilized and corralled diffusion states extend the dwell time of an RNA within a condensate, potentially enhancing the regulatory function of an RNP granule on its selectively recruited RNAs. We further measured biophysical characteristics of the granules that lead to a model wherein the nano-domains form as local percolation networks that capture single molecules upon their partial unfolding. Taken together, the observed spatiotemporal heterogeneity within single RNP granules is likely to play a critical role in regulating the intra-condensate diffusion and thus functionality of RNAs captured within.

Emily Ellinger,¹ Yichen Liu,¹ Adrien Chauvier,¹ Jason C. Porta,²Meliane Ohi,² Nils G. Walter¹

1 Single Molecule Analysis Group and Center for RNA Biomedicine, Department of Chemistry, University of Michigan, Ann Arbor, MI
2 Life Sciences Institute, University of Michigan, Ann Arbor, MI

Riboswitches are structural RNA elements mostly found within the untranslated regions of bacterial messenger RNAs (mRNAs) and play an important role in regulating gene expression at the level of transcription or translation. Here we use techniques of cryo-electron and single molecule microscopy in order to investigate the regulation of transcription termination mediated by the fluoride-sensing riboswitch. We have observed global conformational changes of the RNAP upon the binding of fluoride to the riboswitch which provides insight into understanding the mechanism of the termination regulation by the fluoride riboswitch. We have identified RNAP-RNA interactions further elucidating this mechanism. We have applied methods such as transcriptional gels and single-molecule experiments, specifically Single-Molecule Kinetic Analysis of RNA Transient Structure (SiM-KARTS) assays to expand on this mechanism by analyzing mutants of the identified key interactions on both RNAP and RNA and observing the effect on riboswitch structure. Our single molecule experimental setup allows for the investigation of riboswitch mechanisms in the context of the transcriptional machinery which is known to modulate RNA structure within a cellular context. These results further explain the regulatory process by which the fluoride riboswitch modulates transcription efficiency and improve our understanding of riboswitch dynamics, structure, and regulatory mechanism. Identifying the pivotal interactions between riboswitches and the transcription machinery has the potential to assist in designing new classes of anti-bacterial drugs.

Audrey Hoelscher,¹ Charlotte Clark-Slakey,¹ Rachel Niederer¹

1 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI

MRNA translation plays an important role in gene expression, which influences cellular diversity, stress response, and disease. While the characteristics of mRNA which control protein output are not well understood evidence suggests that mRNA 5’ UTR sequences contain potent features which regulate protein output and control gene expression. Here, we were interested in whether expression of 5′-UTR isoforms known to differentially recruit ribosomes can drive phenotypic differences in yeast. Plasmids containing the long or short 5′-UTR isoform of the gene YHR0202W sequence were generated and transformed into yeast cells with the endogenous gene knocked out. The effects of each isoform on protein output were then measured with a luciferase reporter. Next, we will determine if preferential expression of one isoform mediates differential tolerance to stress. Finally, to determine the exact location of the functionally relevant features we are generating scanning mutations in the unshared sequence of the long isoform. Luciferase output of the mutants will be measured and compared to the original isoform protein output. The insights found on 5’UTR regulatory features will provide a better understanding of the role of translational control elements in human diseases.

Dalia M. Soueid,¹ Amanda L. Garner¹

1 Department of Medicinal Chemistry, College of Pharmacy, University of Michigan, Ann Arbor, MI

RNA-binding proteins (RBPs) make up a class of over 2,000 proteins that bind to and regulate the diverse functions of various types of RNAs, and accordingly, are involved in controlling many cellular processes. Disruption of RNA-Protein Interactions (RPIs), consequently, has been implicated in human diseases ranging from neurodegenerative and autoimmune diseases to several human cancers. Hence, targeting RBPs and RPIs has surfaced as a new frontier in RNA-targeted drug discovery which takes advantage of the endogenous regulation of messenger RNA (mRNA). The aim of this work is to characterize the high-affinity interactions of RBPs with mRNA motifs through live-cell detection using an assay previously developed for the detection of pre-miRNAs and their RBP partners, RNA-interaction with Protein- mediated Complementation Assay (RiPCA). Our goal is to be able to expand RiPCA to allow us to study other more complex RPIs in cells composed of RNAs that are larger and more structurally diverse than pre-miRNAs and that bind to proteins which perform a variety of cellular functions. Initial studies to accomplish this included three representative examples motifs found in 3’ untranslated regions (UTRs) of mRNA such as (1) expanded repeat RNA, (2) UAGUAG target sequence, and (3) AU-rich elements (AREs). My efforts towards developing these assays will be discussed, as well as future directions aimed at further improvement of this assay technology. Through RiPCA optimization, we hope to generate a platform for detecting and validating various RPIs in live cells to enable screening and drug discovery efforts.

Minli Ruan,¹ Kristin S. Koutmou^1-2

1 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
2 Department of Chemistry, University of Michigan, Ann Arbor, MI

Pseudouridine (Ψ) is among the most abundant post-transcriptional RNA modifications. Consistent with the broad influence that Ψ exerts on RNA structure and function, the misregulation of Ψ-incorporating enzymes, pseudouridine synthases (PUSs), is associated with a variety of deleterious health outcomes including human-inherited intellectual disorders. PUS enzymes have been studied for decades because of their ability to modify essential non-coding RNAs such as rRNA and tRNA. However, over the last decade it has become apparent that a sub-set of eukaryotic PUS enzymes also target protein coding mRNAs. How these enzymes select their mRNA targets, as well as the ultimate influence of mRNA pseudouridylation on gene expression remains to be determined. There is mounting evidence that factors beyond inherent enzyme preference for particular RNA sequences or substrates strongly influence PUS target selection in cells. Here, we tested how enzyme localization influences which mRNAs are modified by one of the primary mRNA pseudouridinylating enzymes, PUS7. Using Nanopore sequencing we find that the addition of a cytoplasmic localization tag to PUS7 greatly increases the number and variety of pseudouridylated mRNAs. Furthermore, we identify multiple stress conditions (heat shock and oxidative stress) under which PUS7 naturally relocalizes and demonstrates an altered substrate scope. Notably, the localization of PUS7 to the cytoplasm enhances the fitness of cells subject to these same stresses. These data raise the possibility that shifts in PUS7 localization provide cells an avenue to directly alter translation in response to stress.

Megan Dykstra,^1-2 Kaitlin Weskamp,^1-2 Emile Pinarbasi,² Nico Gomez,² John Moran,² Corey Stewart,^1-2 Sami Barmada,^1-2 Kristin S. Koutmou^1-2

1 Neuroscience Graduate Program, University of Michigan, Ann Arbor, MI
2 Department of Neurology, University of Michigan, Ann Arbor, MI

TDP43 is a critical RNA binding protein involved in several aspects of RNA metabolism. In nearly all individuals with ALS and FTD, TDP43 is mislocalized to form cytoplasmic aggregates. Even so, we understand little about the underlying reasons for TDP43 mislocalization, or the impact of TDP43 cytoplasmic deposition and/or loss of nuclear TDP43 on neuronal survival.

Our previous studies suggested that TDP43 mislocalization may be due to the production of alternatively spliced and truncated TDP43 isoforms that are prone to aggregation and actively exported from the nucleus. These ‘short’ (s)TDP43 isoforms are evolutionarily conserved, but their regulation and function remain fundamentally unclear.

Here, we show that sTDP43 is produced by the same negative feedback loop that regulates TDP43 levels in all cells, wherein TDP43 binds to its own RNA, resulting in alternatively spliced transcripts that are destabilized by nonsense mediated RNA decay (NMD). We found that the transcripts generated in the process of TDP43 autoregulation are those that encode sTDP43. Using OPL, a non-invasive method for measuring protein turnover in situ, we noted that sTDP43 is rapidly cleared by macroautophagy, further contributing to low steady-state levels in healthy cells.

Bypassing these regulatory mechanisms through overexpression of sTDP43 recapitulates TDP43 mislocalization together with neurodegeneration. We found that sTDP43-dependent toxicity requires both RNA binding and dimerization with flTDP43, suggesting both gain- and loss-of-function mechanisms contributing to toxicity. Together, these investigations may prove essential for elucidating the function of sTDP43, along with the consequences of sTDP43 accumulation in ALS and FTD.

Wenxi Yu,¹ Sophie F. Hall,^1-2 Yumei Huangi,³ Julie Ziobro,⁴ Faith Reger,¹ Xiaoyan Jai,³ Miriam H. Meisler^1-2

1 Department of Human Genetics, University of Michigan, Ann Arbor, MI
2 Neuroscience Graduate Program, University of Michigan, Ann Arbor, MI
3 Center for Genomic Technologies, Greater Bay Institute of Precision Medicine, Guangzhou, Guangdong, China
4 Fudan University, Guangzhou, Guangdong, China

Rationale: Gain-of -function mutations of SCN8A cause Developmental and Epileptic Encephalopathy, a rare genetic disorders characterized by refractory seizures, developmental delay, and impaired movement. Affected individuals are heterozygous for wildtype and mutant alleles. Current AEDs and Scn8a ASOs do not distinguish between wildtype and mutant proteins/transcripts. Specific reduction of the pathogenic transcript alone would be preferable in order to limit reduction to 50% of total activity.

Methods: We targeted the mutant allele in the D/+ mouse line carrying the patient SCN8A mutation p.Asn1768Asp (N1768D). In addition to N1768D, this mutant allele contains 8 nearby silent base substitutions. Allele-specific sgRNAs were designed and validated in in vitro assay and cultured cells. The most effective sgRNA was cloned into an AAV.PHP.eB vector and administered to D/+ mice carrying a ubiquitously expressed Cas9 gene (JAX 026179).

Results and conclusions: Significant editing of the mutant allele was detected in mouse brain. Frame-shifting indels were only detected in mutant transcripts but not in wildtype transcripts. Allele-specific inactivation of 25% of mutant Scn8a alleles was obtained in brain of heterozygous mutant mice and was sufficient to rescue seizures and survival.

Kira Holton,^1-2 Brittany Bowman,^1-3 Kristin Koutmou,^2,4-5 Chase Weiddman^1-3,5

1 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
2 Chemistry-Biology Interface Training Program, NIH, GM132046
3 Rogel Cancer Center, University of Michigan, Ann Arbor, MI
4 Department of Chemistry, University of Michigan, Ann Arbor, MI
5 Center for RNA Biomedicine, University of Michigan, Ann Arbor, MI

Human pseudouridine synthase (PUS) enzymes isomerize uridine bases to pseudouridine (Ψ) in coding and noncoding RNAs. The PUS enzyme PUS7 is responsible for a large fraction of pseudouridylation events, and altered PUS7 activity is implicated in several diseases. The mechanisms of PUS7-depedent substrate modification and how Ψ contributes to disease pathogenesis are unclear. Without understanding the cellular contexts that direct PUS7 target selection, we cannot predict the functional consequences of PUS7-dependent Ψ in RNA. We hypothesize that distinct RNA structural contexts and protein-RNA interactions drive selection of Ψ sites by PUS7 in cells. We are employing live-cell chemical probing and sequencing technologies to identify these cellular contexts. Preliminary data from interaction network probing in human cells suggest a trend for limited protein-binding at Ψ sites compared to unmodified sites. We will probe these networks in cells that do not express PUS7 to understand the underlying molecular features of PUS7-modified sites in the absence of Ψ. We will similarly chemically probe whether RNA structural motifs are conserved at PUS7-modified sites. We anticipate that the knowledge generated from this research will allow us to predict novel PUS7-dependent Ψ sites and may be suggestive of a novel gene regulatory mechanism based on RNA modifications.

Xufei Zhou,¹ Rucheng Diao,² Xin Li,¹ Christine A. Ziegler,¹ Max J Gramelspacher,¹ P. Lydia Freddolino,^1-2 Zhonggang Hou,¹ Yan Zhang¹

1 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
2 Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI

Cas9 is an RNA-guided effector enzyme in prokaryotic adaptive immunity and widely used for eukaryotic genome editing. RNA-free apoCas9 is perceived as non-functional. Here we report a new function of apoCas9 in CRISPR biology. Bacteria create immune memories by acquiring viral DNA pieces into CRISPR locus as spacers using Cas1-Cas2 integrase. Type II CRISPR-Cas acquisition also requires Cas9 to select memories from protospacer adjacent motif (PAM)-flanked DNA. Here we found an acquisition-stimulatory activity of apoCas9 in meningococci, which is usually repressed by crRNA-tracrRNA pair to mitigate autoimmunity caused by forming host genome-derived spacers. We further demonstrate the physiological role of apoCas9 in sensing shallow immune memory depth and in augmenting acquisition to rapidly expand the memory repertoire. In the neogenesis and post-contraction replenishment of type II-C CRISPR arrays, apoCas9 is both a sensor of insufficient crRNAs produced from shorter arrays and an actuator to dynamically upregulate acquisition. Lastly, these features of Cas9 in spacer acquisition are evolutionarily conserved across multiple II-C orthologs.

Olivia Swaim,^1-3 Anbarasu Kumaraswamy,^1-2 Hannah Beck,^1-2 Eva Rodansky,^1-2 Thomas Westbrook,^1-2 Yehyun Choi, ^1,4 Amy Kasputis,^2,5 Visweswaran Ravikumar,^2,6 Lisa McMurry,^2,7 Christine Caldwell-Smith,^2,7 Janice Bell,^3,7 Erica Rabban,^2,7 Xuhong Cao,^2,7Dan Robinson,^2,7 Yi-Mi Wu,^2,7 Kening Zhao,^2,7 Chandan Kumar-Sinha,^2,7 Zachery Reichert,^1-2 Arvind Rao,^2,6 Todd Morgan,^3,5 Arul Chinnaiyan,^2,5,7 Euisik Yoon,^3-4 Joshi J. Alumkal^1-2

1 Department of Internal Medicine, University of Michigan, Ann Arbor, MI
2 Rogel Cancer Center, University of Michigan, Ann Arbor, MI
3 College of Literature, Science, and the Arts, University of Michigan, Ann Arbor, MI
4 Department of Electrical Engineering, University of Michigan, Ann Arbor, MI
5 Department of Urology, University of Michigan, Ann Arbor, MI
6 Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI
7 Department of Pathology, University of Michigan, Ann Arbor, MI

Androgen deprivation therapy (ADT) is the principal treatment for metastatic prostate cancer. In recent years, more potent inhibitors of the androgen receptor (AR), a luminal lineage-specifying transcription factor, have been approved for use in men whose tumors become resistant to ADT alone. These AR signaling inhibitors (ARSI) improve overall survival. Unfortunately, progression after ARSI treatment is nearly universal. ARSIs are now given upfront with ADT in patients with metastatic prostate cancer, but mechanisms of resistance to upfront combination therapy are largely unknown. Our goal is to understand the clonal architecture of metastatic prostate cancer patient tumors and how upfront ADT+ARSI treatment changes the architecture of these tumors and promotes resistance. Therefore, we initiated a prospective cohort study to serially profile newly-diagnosed metastatic prostate cancer patient tumors called Michigan Oncology Multi-omic Assessment of Tumor Change and Heterogeneity (Mi-ONCOMATCH).

MiONCOMATCH enrolls patients with metastatic prostate cancer who are beginning treatment with ADT+ARSIs (e.g., abiraterone, enzalutamide). Patients undergo metastatic tissue biopsy profiling through our institutional next generation DNA and RNA-sequencing platform called MiONCOSEQ. At progression, patients undergo a repeat metastatic biopsy for profiling. Since our study began in January 2022, we have enrolled 33 patients , 8 of whom have progressed. Comparison of the baseline vs. progression biopsies has clarified mechanisms that may have contributed to resistance in specific patients. We anticipate completing accrual of MiONCOMATCH this year, which will hopefully enable us to clarify baseline determinants of resistance to upfront ADT+ARSI treatment and further clarify resistance mechanisms that emerge with treatment.

Zoe C. Yeoh,^1-2 Jennifer L. Meagher,² Chia-yu Kang,³ Jennifer A. Bohn,⁴ Paul D. Bieniasz,⁴ Melanie D. Ohi,^{1-2, 5} Janet L. Smith^1-3

1 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
2 Life Sciences Institute, University of Michigan, Ann Arbor, MI
3 Biophysics at Michigan, University of Michigan, Ann Arbor, MI
4 Laboratory of Retrovirology, Rockefeller University, New York, NY
5 Department of Cell and Developmental Biology, University of Michigan, Ann Arbor, MI

The human genome is CpG-suppressed due to the methylation and subsequent deamination of cytosines to uracil in open chromatin, which leads to the conversion of the complementary nucleotide to an adenine by DNA repair pathways. This lack of CpGs allows host cells to use the CpGs in viral RNA as a marker for targeting their destruction. A protein called Zinc finger Antiviral Protein (ZAP) binds CpG dinucleotides, and even though it has no intrinsic RNase activity, leads to degradation of viral RNA1–3. It is thought that the nuclease activity is provided by KHNYN, a ZAP co-factor4. KHNYN has three domains – a KH (K-Homology) domain with predicted RNA binding activity; an NYN (N4BP1, YacP-like Nuclease) domain with predicted RNase activity; and a C-terminal domain (CTD), which is predicted to recognize ZAP. How KHNYN and ZAP act together to promote RNA degradation is not understood and is the focus of our biochemical and structural studies using recombinant proteins.

We solved crystal structures of the KHNYN KH domain, and showed by EMSA that KH does not bind RNA. We also determined that the KHNYN NYN domain is a single-stranded ribonuclease that lacks specificity for CpG dinucleotides. Using purified domains, we found that the KHNYN C-terminal domain (CTD) is required for binding to the ZAP RNA binding domain (ZAP-RBD). A minimal ZAP-nuclease complex, ZAP-RBD:NYN-CTD, hydrolyzed RNA more rapidly than the NYN-CTD by itself.

Together, these results suggest a model where ZAP and KYNYN form a complex that facilitates the degradation of ZAP-bound RNA.

Samantha J. Grudzien,^1-3 Amy Krans,^1,4 Sinem S. Ovunc,¹ Peter K. Todd^1,4

1 Department of Neurology, University of Michigan, Ann Arbor, MI
2 Neuroscience Graduate Program, University of Michigan, Ann Arbor, MI
3 Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI
4 VA Medical Center, Ann Arbor Neurology Clinic, University of Michigan, Ann Arbor, MI

Fragile X-associated tremor/ataxia syndrome (FXTAS) is a neurodegenerative disorder characterized by ataxia, action tremors, and dementia that stems from a trinucleotide CGG repeat expansion ranging from 55 to 200 repeats in the 5’ UTR of FMR1. CGG repeats are thought to drive neurodegeneration through either CGG RNA mediated sequestration of RNA binding proteins or through repeat-associated non-AUG (RAN) translation of toxic homopolymeric peptides, the most abundant being a polyglycine protein (FMRpolyG). To assess the relative contributions of CGG-repeat RNA gain of function toxicity and FMRpolyG generation to neurodegeneration in vivo, I generated adeno-associated viruses (AAVs) that permit assessment of the mechanistic contributions separately or in combination through intracerebroventricular (ICV) injection in neonate mice. Preliminary data demonstrates that expression of CGG-repeat RNAs that support RAN translation of FMRpolyG exhibit robust P62+ and ubiquitin positive inclusions and Purkinje cell dropout that correlate with impaired motor behavioral deficits at 6 months of age, recapitulating many key pathological features observed in FXTAS patients. However, the expression of AAVs that either make FMRpolyG in the absence of the CGG-repeat or that express the CGG repeat in a fashion that markedly diminishes RAN translation do not elicit robust disease pathology or behavioral abnormalities. This AAV model of FXTAS recapitulates key pathological findings which include gait ataxia, purkinje cell loss, and neuronal inclusion formation. Neither FMRpolyG nor CGG-repeat RNA alone are sufficient to elicit maximal toxicity. Future work will investigate the mechanism that underlies the RAN-component toxicity – either RAN translation itself, or pathological RAN translation evoked pathways.

Kayla Lenshoek,¹ Brittany Bowman,² Chase Weidmann^2,3

1 Department of Cellular and Molecular Biology, University of Michigan, Ann Arbor, MI
2 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
3 Center for RNA Biomedicine and Rogel Cancer Center, University of Michigan, Ann Arbor, MI

Long noncoding RNA Metastasis Associated Lung Adenocarcinoma Transcript 1 (MALAT1) is linked to multiple cancer types. True to its names, MALAT1 upregulation often correlates with aggressive metastasis. The mechanism by which MALAT1 enacts altered gene regulatory programs is not well understood. Here, we create and validate A549 lung adenocarcinoma cell lines where CRISPR-based gene deletions and insertions alter MALAT1 expression or destabilize the MALAT1 transcript. We find that MALAT1 accumulates at high levels in nuclear speckles, but loss of MALAT1 does not strongly alter nuclear speckle formation. Increased MALAT1 expression promotes a mesenchymal phenotype, whereas reduction in MALAT1 expression promotes a more epithelial cell state. Further, we find that MALAT1 expression is inversely correlated with expression of EZH2, the catalytic subunit of the Polycomb Repressive Complex 2 (PRC2). Consequently, levels of the PRC2 epigenetic mark, Histone H3 Lysine 27 tri- methylation (H3K27me3) are inversely correlated with MALAT1 levels. These results are consistent with a model wherein MALAT1 reprograms metastatic cancer cells through reorganization of epigenetic marks.

Genesis Rodriguez,^1-2 Elizabeth Tank,² Roberta Fuller,^1,3 Andrew Tidball,² Tracy Qiao,²Jack Parent,^1-2Sami Barama^1-2

1 Neuroscience Graduate Program, University of Michigan, Ann Arbor, MI
2 Department of Neurology, University of Michigan, Ann Arbor, MI
3 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI

Chorea acanthocytosis (ChAc) is an autosomal-recessive neurodegenerative disorder caused by mutations in the gene encoding vacuolar protein sorting factor 13A (VPS13A). Studies show that VPS13A aids in the transport of lipids between organelles and that some domains are homologous to evolutionarily conserved autophagy proteins, suggesting a function for VPS13A in specialized branches of autophagy. Using HEK293T cells in which the native VPS13A has been knocked out (VPS13A KO), we found that VPS13A loss only moderately autophagic flux. Furthermore, mass spectroscopy failed to demonstrate an upregulation of autophagy substrates in VPS13A KOs – arguing against an essential role for VPS13A in these processes. Therefore, we broadened our approach and found that candidate proteins that that were consistently affected at the protein level also demonstrated changes at the RNA level, implying transcriptional regulation upon VPS13A KO. To investigate whether VPS13A may regulate transcription, we conducted next generation RNA-sequencing in VPS13A KO HEK293T cells and controls. These findings imply that VPS13A may be regulating RNA expression, through secondary effects on mRNA transport or transcription factors. Concurrent with these studies, we observed an elevated risk of death in iPSC-derived neurons from ChAc patients through longitudinal analyses. Future studies will examine cell type-specific functions of VPS13A in human neurons compared to HEK293T cells, with a focus on pathways that mediate neuronal survival. Collectively, these investigations will help define a function for VPS13A in neurons, outline disease mechanisms, and highlight pathways that may be targeted to prevent neuron loss in ChAc.

Yaping Liu,¹ Cade T. Harkner,² Sarah C. Keane^1-2

1 Biophysics Program, University of Michigan, Ann Arbor, MI
2 Department of Chemistry, University of Michigan, Ann Arbor, MI

MicroRNAs are small non-coding RNAs that post-transcriptionally regulate gene expression. To maintain proper microRNA expression levels, the enzymatic processing of primary and precursor microRNA elements must be strictly controlled. However, the molecular determinants underlying this strict regulation of microRNA biogenesis are not fully understood. Due to the essential regulatory roles of miRNAs in gene expression, these enzymatic processing steps are often post-transcriptionally regulated by RNA-binding proteins (RBPs). Knockdown of heterogenous nuclear ribonucleoprotein (hnRNP) A2B1 has been shown to reduce the expression of mature miR-20a, an oncogenic miRNA, in esophageal cancer cells via an unknown mechanism. We were able to validate these findings and determine that hnRNPA2B1 functions to promote Drosha processing of the primary miRNA, as we observe an increase in pre-miR-20a levels upon addition of purified hnRNPA2B1. We therefore sought to elucidate the mechanism by which hnRNPA2B1 promotes processing of pri-miR-20a. Here, we determined the solution structure of pri-miR-20a using NMR spectroscopy. Additionally, we found that hnRNPA2B1, which contains two tandem recognition motif (RRM) domains, forms a 1:1 complex with the apical loop of pri-miR-20a. Our ITC data indicates that each isolated RRM can interact with the apical loop of pri-miR-20a with similar binding affinity. We found that an hnRNPA2B1 construct containing two tandem RRMs results in the unfolding of the pri-miR-20a apical loop. We hypothesize that remodeled apical loop structure of pri-miR-20a, induced by hnRNPA2B1 binding, is a better substrate for Drosha processing. Our results highlight that the apical loop of pri-miR-20a is critical binding site for RBPs which promote miR-20a maturation by modulating the apical loop structure.

Chunyi Hu,¹ Mason Myers,² Xufei Zhou,²Zhonggang Hou,² Macy Lozen,²Hi Kyun Nam,³ Yan Zhang,² Ailong Ke²

1 Department of Molecular Biology and Genetics, Cornell University, Ithaca NY
2 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI
3 College of General Education at Kookmin University, Seoul, South Korea

Type I CRISPR-Cas systems utilize the RNA-guided Cascade complex to identify matching DNA targets, and the nuclease-helicase Cas3 to degrade them. Among seven subtypes, Type I-C is compact in size and highly active in creating large-sized genome deletions in human cells. Here we use four cryo-electron microscopy snapshots to define its RNA-guided DNA binding and cleavage mechanisms in high resolution. The non-target DNA strand (NTS) is accommodated by I-C Cascade in a continuous binding groove along the juxtaposed Cas11 subunits. Binding of Cas3 further traps a flexible bulge in NTS, enabling efficient NTS nicking. We identified two anti-CRISPR proteins AcrIC8 and AcrIC9, that strongly inhibit N. lactamica I-C function. Structural analysis showed that AcrIC8 inhibits PAM recognition through direct competition, whereas AcrIC9 achieves so through allosteric inhibition. Both Acrs potently inhibit I-C-mediated genome editing and transcriptional modulation in human cells, providing the first off-switches for controllable Type I CRISPR genome engineering.

Yihua Li,¹ Samuel Carlson,² Gabrielle Law,² Victoria D’Souza,² Janet L. Smith,¹ Melanie D. Ohi¹

1 Life Sciences Institute, University of Michigan, Ann Arbor, MI
2 Department of Molecular and Cellular Biology, Harvard University, Cambridge, MA

Acquired Immunodeficiency Syndrome (AIDS) caused by Human Immunodeficiency Virus (HIV) is a public health threat worldwide. Currently, there is no effective cure for HIV-AIDS. The long-term survival of HIV-infected individuals requires lifelong treatment with drugs targeted to steps in the retrovirus life cycle. Development of a long-acting and curative treatment for HIV-AIDS requires detailed knowledge of HIV structural biology. After HIV infects a host T cell, the viral RNA is reverse transcribed to DNA for integration into the host genome. The reverse transcriptase (RT) enzyme is an important drug target, and its elongation activity has been broadly studied. However, the initiation step is less well understood. RT initiates the synthesis of DNA by binding to an RNA:RNA duplex formed by a viral RNA template and a host lysine tRNA primer. A structural understanding of the mechanism of formation of the initiation complex could significantly guide drug design.

Our study aims to uncover the mechanism of reverse transcription initiation in HIV-1. Our previous studies identified that a critical RT interaction with a specific viral RNA stem-loop induces a primer-ready conformation of the viral template. We utilized cryo-electron microscopy (cryo-EM) to obtain a three-dimensional view of RT engaged with this RNA stem-loop. We then used Nuclear Magnetic Resonance (NMR) to map the interactions between RT and this RNA. Our structural models provide a new view of RT initiation. Overall, our study provides a valuable contribution to the field of HIV research by providing new structural evidence for how RT initiates reverse transcription of the HIV-1 RNA.

Jingxuan Tang,¹ Ben Dodd,² Andreas Schimidt,¹ Stephanie Moon,^2-3 Nils Walter^1,3

1 Department of Chemistry, University of Michigan, Ann Arbor, MI
2 Department of Human Genetics, University of Michigan, Ann Arbor, MI
3 Center for RNA Biomedicine, University of Michigan, Ann Arbor, MI

The ATPase valosin-containing protein (VCP), with its hexametric structure, is crucial for unfolding ubiquitinated protein substrates to facilitate proteasomal degradation. A single amino acid mutation within VCP can cause amyotrophic lateral sclerosis/frontotemporal dementia (ALS/FTD). The mechanism of pathogenesis remains undefined. Studies have revealed persistent ubiquitin-positive aggregates in patient tissues with the VCP mutation, indicating that the ubiquitinated misfolded proteins may not be properly unfolded and degraded. This may be due to a loss of function caused by the mutation. Surprisingly, in vitro studies have demonstrated enhanced ATPase activity and unfoldase rate in mutated VCP homomers when compared to the wildtype. Given that these studies were conducted only with homomeric mutant VCP, but most patients have heterozygous VCP mutations, our hypothesis is that mutant VCP monomers can co-assemble with wildtype, leading to compromised motor protein function. To test whether wildtype and mutant VCP co-assemble, we co-expressed SNAP- and Halo-tagged, distinctly fluorophore-labeled wildtype and mutant VCP, respectively, in HeLa cell lysate and visualized the assembly stoichiometry using single molecule pull-down (SiMPull) combined with total internal reflection fluorescence (TIRF) microscopy-based single molecule photobleaching (SMPB). Unfolding assays will further be employed to ask whether co-assembled heteromeric VCP shows loss of function.

Nicholas Rossiter,¹ Stephen DeAngelo,² Ingrid Kilde,³ Brandon Chen,¹ Anna Anders,⁴ Catherine Wilhelm,³ Kyle Flickinger,^5-6Caela Fedraw,⁴ Jason Cantor,^5-6Brandon Ruotolo,⁴Markos Koutmos,^2-3 Costas Lyssiotis,^1,2,7 Yatrik Shah^1,2,7

1 Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, MI
2 Doctoral Program in Cancer Biology, University of Michigan, Ann Arbor, MI
3 Program in Chemical Biology, University of Michigan, Ann Arbor, MI
4 Department of Chemistry, University of Michigan, Ann Arbor, MI
5 Morgridge Institute for Research, Madison, WI
6 Department of Biochemistry, University of Wisconsin-Madison, Madison, WI
7 Department of Molecular and Integrative Physiology, University of Michigan, Ann Arbor, MI

Human mitochondria retain a compact genome encoding 13 integral subunits of the oxidative phosphorylation (OXPHOS) system. The expression of these genes is dependent on a suite of mitochondria-specific RNA synthesis, processing, and degradation machineries. However, the mechanisms that regulate mitochondrial RNA (mtRNA) dynamics remain incompletely understood. REXO2 is a mitochondrial exonuclease that catalyzes the final step of mtRNA degradation. Loss of REXO2 leads to accumulation of short oligoribonucleotides and noncoding mtRNAs, decreased mitochondrial transcription, and impaired respiration. To our knowledge, no work has investigated whether REXO2 activity is altered under varying cellular contexts.

To address this gap, we analyzed multiple functional genomics datasets to identify mitochondrial genes associated with REXO2. We discovered a robust association between REXO2 and the mitochondrial zinc exporter SLC30A9. Depletion of SLC30A9 causes mitochondrial Zn2+ accumulation, impaired respiration, and decreased abundance of mitochondrially-encoded proteins. Despite having a significant consequence for cellular fitness, the mechanism of zinc-induced mitochondrial dysfunction following loss of SLC30A9 remains unresolved.

To determine if zinc alters REXO2 activity, we purified recombinant human REXO2 and measured its ability to degrade short RNAs. Our data indicate that low levels of Zn2+ have a potent inhibitory effect on REXO2 activity. Additionally, we found that SLC30A9 and REXO2 knockdown cells exhibit impaired respiration, destabilization of OXPHOS complexes, and accumulation of POLRMT-inhibiting 7S RNA. Based on these results, we propose a novel axis of zinc toxicity in human cells, wherein elevated mitochondrial zinc following SLC30A9 depletion inhibits REXO2, disrupting mtRNA turnover and transcription, ultimately leading to impaired respiration.

Noah Helton¹, Stephanie Moon^1-2

1 Department of Human Genetics, University of Michigan, Ann Arbor, MI
2 MI Faculty Scholar of the Center for RNA Biomedicine, University of Michigan, Ann Arbor, MI

Stress-induced gene expression and RNA-protein condensates called stress granules (SGs) are crucial for cellular resilience to stress through poorly understood mechanisms. The prevailing model in the field is that SGs sequester non-translating mRNAs and suppress mRNA translation. However, stress-induced genes, such as ATF4 and GADD34 must be translationally upregulated during stress. Contradictory evidence in the field suggests a fraction of stress-induced gene mRNAs can localize to stress granules and reporter RNAs that are translationally upregulated during stress can translate within SGs. Thus there is a gap in knowledge as to what role (if any) SGs play in translation regulation, stress-induced gene expression, and how this impacts cellular survival following stress. By imaging single mRNA molecules in U-2 OS cells and human iPSC-derived spinal motor neurons, we investigated abundance and localization of a subset of stress-induced gene mRNAs. Surprisingly, these experiments revealed variable changes in abundance and localization of stress-induced gene mRNAs to SGs that were generally similar in U-2 OS and motor neurons. Polysome profiling coupled with RT-qPCR and the use of translation inhibitors in U-2 OS cells suggested that translation of stress-induced gene mRNAs prevented their localization to SGs. Importantly, cells deficient in forming SGs had reduced stress-induced gene expression and decreased viability during chronic stress. These results suggest SGs play a key role in mediating stress-induced gene expression and cell survival. This work has shed light on fundamental roles of SGs in biology and translation regulation, thus giving insight into potential therapeutics for diseases associated with SG dysregulation.

Daniel Wilinski,¹ Monica Dus¹

1 Molecular, Cellular, and Developmental Biology Department, University of Michigan, Ann Arbor, MI

Regulation of the insulin system is critical for glucose homeostasis. RNA control plays an important yet understudied aspect of the dynamic regulation of the insulin system. N6-methyladenosine RNA (m6A) is a pervasive messenger RNA modification that modulates the fate of RNAs. As the generalized roles of m6A in RNA control are beginning to be unveiled, the specific roles for m6A over the insulin system remain unclear. Here we show that m6A is essential for proper glucose homeostasis. Insulin-producing cells monitor glucose levels and release the hormone insulin to stimulate the uptake of glucose and thus reduce circulating glucose. When m6A methylation is inhibited, specifically in insulin-producing cells, animals become hyperglycemic and develop diabetes. We found by m6A individual-nucleotide-resolution cross-linking and immunoprecipitation (mi-CLIP) that messenger RNAs important for the insulin system harbor m6A including insulin messenger RNA. Furthermore, using direct RNA-sequencing combined with machine learning we identified the precise nucleotide that is methylated in the insulin transcript. Polysome profiling and in vivo mutational analysis demonstrated that m6A enhances the translation of insulin messenger RNA but does not affect the stability of the message. Finally, our data reveal that methylation of the insulin RNA is necessary to maintain adequate levels of insulin in insulin-producing cells to respond to high levels of glucose. These insights uncover a new regulatory layer of the insulin system and a novel context in which m6A increases translation. The results underscore how the reduction of m6A could lead to functional impairment of insulin-producing cells and ultimately diabetes.

William K. Storck,^1-2 Joshua A. Kuleape,^1-2 Diana Flores,^1-2 Chao Zhang,^1-2 Canping Chen,^3-4 Eva Rodansky,^1-2 Anbarasu Kumaraswamy,^1-2 Raymond Cavalcante,⁵ Zhi Duan,^1-2 Zheng Xia,^3-4 Joel A. Yates,^1-2 Joshi J. Alumkal^1-2

1 Department of Internal Medicine, University of Michigan, Ann Arbor, MI
2 Rogel Cancer Center, University of Michigan, Ann Arbor, MI
3 Knight Cancer Institute, Oregon Health & Science University, Portland, OR
4 Biomedical Engineering Department, Oregon Health & Science University, Portland, OR
5 BRCF Epigenomics Core, University of Michigan, Ann Arbor, MI

Prostate adenocarcinomas rely on the androgen receptor (AR) for proliferation and survival. AR signaling inhibition is the principal treatment for this disease, but resistance is nearly universal. An increasingly prominent resistance mechanism is lineage plasticity, exemplified by loss of AR signaling and epithelial differentiation. Neuroendocrine prostate cancer (NEPC) is the most virulent example, for which there are limited treatment options. Loss of the tumor suppressors TP53/RB1 is ubiquitous in NEPC, but molecular mechanisms that explain how TP53/RB1 loss promotes NEPC are largely unknown. Our prior work demonstrated that BET bromodomain inhibition (BETi) is a promising approach to block a neuronal survival program in NEPC tumors. Importantly, a subset of NEPC patients experienced prolonged tumor control in our prior BETi clinical trial. However, disease progression was universal, demonstrating the need to address other critical mechanisms that contribute to NEPC in order to improve patient outcomes.

Prior work demonstrated NEPC patient tumors harbor differential DNA methylation vs. adenocarcinomas. Using TP53/RB1-knockout vs. intact mouse models, we determined that TP53/RB1 loss leads to global DNA methylation changes, including in pathways associated with epithelial differentiation. DNA methyltransferase inhibition (DNMTi) suppressed growth of TP53/RB1-knockout NEPC cell lines but combination therapy with BETi was even more effective in vitro. We confirmed these results using a TP53/RB1-loss NEPC patient-derived xenograft model in vivo. Currently, we are clarifying the mechanisms that explain the superior anti-tumor activity of this combination. Altogether, our work suggests combination DNMTi + BETi is a promising new treatment approach for NEPC.

Anbarasu Kumaraswamy,^1-2 Olivia A. Swaim,^1-3 Rahul Mannan,⁴ Xiao-Ming Wang,^4-5 Aaron Udager,⁵ Colm Morrissey,⁶ Arul M. Chinnaiyan,^2,4-5,7-8Joel A. Yates,^1-2 Joshi J. Alumkal^1-2,5

1 Department of Internal Medicine, University of Michigan, Ann Arbor, MI
2 Rogel Cancer Center, University of Michigan, Ann Arbor, MI
3 College of Literature, Science, and the Arts, University of Michigan, Ann Arbor, MI
4 Department of Pathology, University of Michigan Medical School, Ann Arbor, MI
5 Michigan Center for Translational Pathology, Ann Arbor, MI
6 Department of Urology, University of Washington, Seattle, WA
7 Department of Urology, University of Michigan Medical School, Ann Arbor, MI
8 Howard Hughes Medical Institute, Ann Arbor, MI

Lysine-specific demethylase 1 (LSD1) is a histone demethylase and regulator of differentiation, including in cancer. A neuronal-specific isoform of LSD1—LSD1+8a—was previously shown to play a key role in promoting neuronal differentiation in the developing brain. We previously determined that LSD1+8a transcripts were detectable in an aggressive subtype of prostate cancer harboring a neuronal program—neuroendocrine prostate cancer (NEPC)—while LSD1+8a was not expressed in the more common prostate adenocarcinomas harboring a glandular program. However, it was unclear if LSD1+8a was expressed at the protein level in NEPC and the role of LSD1+8a in NEPC was unknown because suitable antibodies to measure LSD1+8a were not available. We generated a rabbit monoclonal antibody to specifically detect LSD1+8a protein expression. Using stable LSD1+8a-overexpressing cells and normal neuronal tissues, we confirmed that our antibody detected LSD1+8a protein by Western blotting and immunohistochemistry. However, we failed to detect LSD1+8a protein by immunohistochemistry in prostate cancer patient-derived xenografts and patient tissue samples—including those that we determine to express low but detectable levels of LSD1+8a transcript. In summary, LSD1+8a protein is not detectable in NEPC tissues. However, measuring LSD1+8a transcript levels may assist in the diagnosis of NEPC, which can sometimes be challenging.

Abigail S. Mauger,¹ Sarah C. Hanks,¹ Alexandria J. Shumway,² Arushi Varshney,³ Dan L. Ciotlos,¹ Nandini Manickam,³ Narisu Narisu,⁴ Peter Orchard,³ Michael R. Erdos,⁴ Mike Sweeney,¹ Jeffrey Okamoto,¹ Heather M. Stringham,¹ Lori L. Bonnycastle,⁴ Timo A. Lakka,^5-7 Karen L. Mohlke,⁸ Jaakko Tuomilehto,^9-11 Markku Laakso,¹² Heikki A. Koistinen,^13-15 Michael Boehnke,¹ Stephen C. J. Parker,³ Francis S. Collins,⁴ Praveen Sethupathy,² Laura J. Scott¹

1 Department of Biostatistics and Center for Statistical Genetics, School of Public Health, University of Michigan, Ann Arbor, MI
2 Department of Biomedical Sciences, Cornell University, Ithaca, NY
3 Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI
4 Center for Precision Health Research, National Human Genome Research Institute, NIH, Bethesda, MD
5 Institute of Biomedicine, School of Medicine, University of Eastern Finland, Kuopio, Finland
6 Department of Clinical Physiology and Nuclear Medicine, Kuopio University Hospital, Kuopio, Finland
7 Foundation for Research in Health Exercise and Nutrition, Kuopio Research Institute of Exercise Medicine, Kuopio, Finland
8 Department of Genetics, University of North Carolina, Chapel Hill, NC
9 Public Health Promotion Unit, Finnish Institute for Health and Welfare, Helsinki, Finland
10 Department of Public Health, University of Helsinki, Helsinki, Finland
11 Diabetes Research Group, King Abdulaziz University, Jeddah, Saudi Arabia
12 Institute of Clinical Medicine, Internal Medicine, University of Eastern Finland, Kuopio, Finland
13 Department of Public Health and Welfare, Finnish Institute for Health and Welfare, Helsinki, Finland
14 Department of Medicine, University of Helsinki and Helsinki University Hospital, Helsinki, Finland
15 Minerva Foundation Institute for Medical Research, Helsinki, Finland

There are extensive sex differences in gene expression in muscle tissue, but the mechanisms underlying these differences are unclear. MicroRNAs (miRNAs) are small, non-coding RNAs that post-transcriptionally regulate gene expression primarily by binding to 3’ untranslated regions and promoting degradation. Sex differences in miRNA expression may contribute to sex differences in mRNA regulation and expression. To analyze differential expression of miRNAs, we used small RNA-seq to measure bulk miRNA expression in vastus lateralis tissue biopsied from 256 living Finnish donors (FUSION Tissue Biopsy Study). We also generated bulk mRNA-seq and single-nucleus RNA-seq (snRNA-seq) data from these samples. We tested 755 miRNAs for sex differences with DESeq2, correcting for the estimated cell-type composition from snRNA-seq. We found 156 (20.7%) miRNAs differentially expressed by sex (sex-biased miRNAs). miRNAs processed from the same primary transcript, as well as intragenic miRNAs and the corresponding host mRNAs, had concordant directions of effect by sex. To investigate the relationship between sex-biased miRNAs and the mRNAs they regulate, we utilized target predictions from TargetScan. We did not find evidence that mRNAs targeted by sex-biased miRNAs were more likely to be sex-biased than non-targeted mRNAs. We also tested for enrichment of sex-biased genes among the predicted targets of each of 698 miRNA families and did not find evidence for enrichment, though this analysis was limited by non-tissue-specific predictions. We found extensive sex differences in miRNA expression in muscle tissue. The concordance of sex-biased expression between co-transcribed RNAs suggests miRNAs are primarily regulated at the level of transcription.

Joshua A. Kuleape,^1-2 William K. Storck,^1-2 Diana Flores,^1-2 Chao Zhang,^1-2 Canping Chen,^3-4 Eva Rodansky,^1-2 Anbarasu Kumaraswamy,^1-2, Raymond Cavalcante,⁵ Karan Bedi,^2,6 Zhi Duan,^1-2 Joel A. Yates,^3-4 Zheng Xia,^3-4 Joshi J. Alumkal^1-2

At diagnosis, nearly all prostate cancers express a luminal differentiation program regulated by the androgen receptor (AR). However, certain tumors lose AR-dependence and undergo lineage plasticity, exhibiting features of epithelial-mesenchymal transition (EMT) and differentiation change. Combined RB1 and TP53 loss in patient tumors is strongly linked to prostate cancer lineage plasticity and poor outcomes. Moreover, experimental prostate cancer mouse models show that combined loss of RB1/TP53 promotes lineage plasticity and tumor aggressiveness. However, mechanisms by which RB1/TP53 loss reprograms cells are unclear. Further, there are no effective therapies to block prostate cancer lineage plasticity induced by RB1/TP53 loss.

To understand how RB1/TP53 loss induces lineage plasticity, we performed an integrative analysis of RNA-seq and ATAC-seq using RB1/TP53-intact and knockout cell lines from mouse models. Gene set enrichment analysis implicated activation of pathways linked to neuronal differentiation, EMT, and stemness. Our prior work in AR-activity low lineage plasticity tumors demonstrated that BET bromodomain protein inhibition (BETi) is a promising treatment direction in models with similarly upregulated pathways. Treatment of RB1/TP53 knockout cells with BETi suppressed many of the key pathways induced by RB1/TP53 loss and suppressed tumor growth. Our results shed new light on mechanisms that drive cellular reprogramming by RB1/TP53 loss and suggest BETi may be a promising treatment approach for these tumors.

Dhruv Khokhani,^1-3 Zhi Duan,^1-2 Anbarasu Kumaraswamy,^1-2 Olivia A. Swaim,^1-3Chao Zhang,^1-2 Diana Flores,^1-2 Joel A. Yates,^1-2 Joshi J. Alumkal^1-2

The principal treatment for prostate cancer is lowering levels of androgens (male hormones) that activate the androgen receptor (AR), a nuclear hormone receptor that promotes a luminal, or glandular program. Lineage plasticity (LP), or differentiation change, is increasingly recognized as an emergent resistance mechanism after therapeutic targeting of the AR, in which tumors lose reliance on the AR. We now know that LP is a continuum ranging from AR activity-low tumors with persistent AR expression but low AR target gene expression, AR-null tumors without neuroendocrine differentiation (double-negative prostate cancer [DNPC]) or AR-null tumors with neuroendocrine differentiation (neuroendocrine prostate cancer [NEPC]). Importantly, there are no effective treatments for prostate cancers that have undergone LP.
Our previous study revealed that the transcription factor PROX1 was the most significantly upregulated gene in matched biopsies that underwent LP to DNPC after AR inhibitor treatment. Examination of additional patient datasets demonstrates PROX1 is upregulated earlier in the LP continuum and gradually increases in both DNPC and NEPC tumors, indicating the potential significance of PROX1 for LP’s emergence. Since the PROX1 gene is not amplified, we focused on epigenetic mechanisms. We found that the PROX1 promoter becomes hypomethylated as tumors undergo LP and upregulate PROX1 mRNA and protein expression. Importantly, our functional studies demonstrate that ectopic PROX1 overexpression in AR-driven prostate cancer cells suppresses AR expression and signaling while also promoting alternate differentiation programs. We are currently focused on how PROX1 contributes to LP and targeting those mechanisms.

Joel A. Yates,^1-2 Ya-Mei Hu,³ Chao Zhang,^1-2 Joshua Kuleape,^1-2 Zheng Xia,³ Joshi J. Alumkal^1-2

The androgen receptor antagonist enzalutamide (enza) is one of the principal treatments for men with metastatic castration-resistant prostate cancer (mCRPC) whose tumors progress after androgen deprivation therapy, the principal treatment for metastatic prostate cancer. Most mCRPC patients respond to enza treatment. However, tumors from a subset of patients exhibit extreme enza non-response, progressing within three months of treatment. These patients have a significantly shorter overall survival vs. patients who have more durable response to enza.

To clarify mechanisms of extreme enza non-response, we analyzed RNA-seq data from pre-treatment metastatic biopsies obtained in our prior enza clinical trial in men with mCRPC. We focused on men with extreme enza non-response (progression within 3 months) vs. extreme response (progression only after 24 months). Transcriptional profiling of the patient tumors identified a gene signature associated with extreme enza non-response. High expression of that signature in independent clinical datasets was linked to poor patient outcomes, highlighting its clinical relevance.

To understand candidate regulators of the extreme enza non-response program, we used an integrative systems biology approach. Master regulator analysis implicated specific factors, including Cyclin Dependent Kinase 1 (CDK1). We identified prostate cancer cell lines that harbor high expression of our extreme enza non-response signature and confirmed that CDK1 targeting with inhibitors or siRNA reduced cell viability. Our results demonstrate the importance of an integrative systems biology approach to understand extreme enza non-response and suggest CDK1 may be a target of interest in these tumors.

Huibin Yang,¹ Radhika Suhas Hulbatte,¹ Natalie Gratsch,¹ Mario Ashaka,¹ Alan Kelleher,^2-3 John Raupp,² Yin Wang,² Kristen Hong Dorsey,⁴ Troy Halseth,⁴ Anna Schwendeman,⁴ Philip L. Palmbos,^2,5 Mats Ljungman^1,5-7

1 Department of Radiation Oncology, University of Michigan, Ann Arbor, MI
2 Department of Internal Medicine, University of Michigan, Ann Arbor, MI
3 Doctoral Program in Cancer Biology, University of Michigan; Ann Arbor MI
4 Department of Pharmaceutical Sciences and Medicinal Chemistry, University of Michigan; Ann Arbor MI
5 Rogel Cancer Center, University of Michigan; Ann Arbor MI
6 Department of Environmental Health Sciences, University of Michigan; Ann Arbor MI
7 Center for RNA Biomedicine, University of Michigan; Ann Arbor MI

Current cancer therapies typically give rise to dose-limiting normal tissue toxicity. We have developed KLIPP, a precision cancer approach that specifically kills cancer cells using CRISPR/Cas9 technology. The approach consists of guide RNAs that target cancer-specific structural variant junctions to nucleate two parts of a dCas9-conjugated endonuclease, Fok1, leading to its activation. We show that KLIPP causes induction of DNA double strand breaks (DSBs) at the targeted junctions and cell death of cancer cells both in cell culture and in vivo. When cancer cells were grown orthotopically in mice, local LNP delivery of KLIPP reagents resulted in tumor cell death. This therapeutic approach has high specificity for tumor cells and is independent of tumor-specific drivers. Individualized translation of KLIPP to patients would be transformative and lead to consistent and simplified cancer treatment decisions.

Clare M. Wieland,^1-3 Shannon E. Wright,^1-2,4 Melissa Asher,¹ Peter K Todd^1,4

1 Department of Neurology, University of Michigan, Ann Arbor, MI
2 Neuroscience Graduate Program,, University of Michigan, Ann Arbor, MI
3 Medical Scientist Training Program, University of Michigan; Ann Arbor MI
4 Department of Neuroscience, Picower Institute, Cambridge, MA
5 Ann Arbor Veterans Administration Healthcare, Ann Arbor, MI

Expansion of short tandem repeats (STRs) cause over 60 neurological diseases including C9orf72 associated frontotemporal dementia/amyotrophic lateral sclerosis (C9 FTD/ALS) and Fragile X Associated Tremor/Ataxia Syndrome (FXTAS). These expanded repeats form GC-rich regions which create stable secondary structures such as G-quadruplexes, alter RNA dynamics, and support a non-canonical initation process known as repeat-associated non-AUG (RAN) translation. While canonical translation initiation requires various factors to bind to the 5’ cap of mRNA and then scan the mRNA until the start codon (AUG) is reached, RAN translation initiation can occur in two distinct ways: cap-dependent, in which initiation factors bind to the cap but initiate translation at a near cognate codon due to stalling at the repeat, and cap-independent, in which translation does not require a cap and utilizes poorly characterized internal ribosome entry sites (IRES). Here, we examined the influences of cell-type on cap-independent translation. While RAN translation is largely a cap-dependent process in cell lines and in vitro, cap-independent translation of RAN translation specific reporters is upregulated in primary rat hippocampal neurons and in iPSC-derived neurons. In primary rat hippocampal neurons, cap-independence is not influenced by G-quadruplex destabilization and responds differently to ER stress compared to findings in HEK293 cell lines. Taken together, these data indicate an effect of cell type on the cap-dependence of RAN translation, with implications for how repeat expansion disorders repeats might elicit toxicity primarily within the nervous system.

Ariel McShane,^1-2 Ishwarya Venkata Narayanan,² Michelle Paulsen,² Mario Ashaka,² Brian Magnuson,³ Karan Bedi,^4-5 Thomas E. Wilson,^3,6 Mats Ljungman,^2,7

1 Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, MI
2 Departments of Radiation Oncology, University of Michigan, Ann Arbor, MI
3 Department of Pathology, University of Michigan; Ann Arbor MI
4 Department of Biostatistics, School of Public Health, Ann Arbor, MI
5 Rogel Cancer Center, University of Michigan; Ann Arbor MI
6 Department of Human Genetics, University of Michigan; Ann Arbor MI
7 Center for RNA Biomedicine, University of Michigan; Ann Arbor MI

Current estimates suggest that 75-85% of the genome is pervasively transcribed by RNA polymerase, although only ~30% has been annotated. While the relevance of intergenic transcription has been questioned, thousands of unannotated putative long non-coding RNAs (lncRNAs) have been identified, and a large proportion of these lncRNAs may be a consequence of non-productive transcription of annotated genes. While our interest and understanding of transcription-associated RNAs have grown with the advent of functional genomics techniques to identify them, information regarding their genomic locations, patterns of transcription, and functions is still lacking. We used the ENCODE4 deeply profiled cell lines (DPCL), a collection of 13 sequencing assays performed across 16 human cell lines, to explore the phenomenon of pervasive transcription and determine the prevalence of three transcription-associated RNA species–PROMPTs, eRNAs, and readthrough transcripts–genome-wide. We evaluated these RNAs relative to their associated gene and found that, although they are expected to be dependent on the production of the primary transcript, they do not always follow the anticipated pattern of transcription. In addition, we identified features of the local and 3D chromatin environment that correlate with the patterns of transcription we see for each RNA species, however, these trends are not absolute. Our results support the idea that intergenic, transcription-associated RNAs are not simply byproducts of genic transcription and may have some inherent regulation of their own, and we provide a useful resource that can be used to identify these regions of intergenic transcription to orient future studies of these RNAs.

Dana Beseiso,¹ Rachel O. Niederer,¹

1 Department of Biological Chemistry, University of Michigan, Ann Arbor, MI

Breast adenocarcinoma arises in the epithelial cells of breast glands. Cancerous epithelial cells can undergo Epithelial to Mesenchymal Transition (EMT) where they lose their adhesive properties and gain a mesenchymal-like morphology with increased migration and invasion abilities. Tumors that have undergone EMT are often classified as highly aggressive with a poor prognosis. EMT in breast cancer is associated with genome-wide changes in the transcriptome, including alternative splicing and polyadenylation. However, alternative Transcription Start Site (TSS) usage, which defines the 5′ Untranslated Region (5′ UTR) of an mRNA, remains comparatively understudied. Given that 5′ UTRs often harbor translational control elements that alter ribosome recruitment to mRNAs, switches in TSS usage can alter gene expression and thus facilitate the cellular reprogramming necessary to drive EMT and cancer. Using luciferase reporters, we have identified novel putative translational control elements in the 5′ UTR isoforms of genes relevant to EMT in breast cancer. To characterize 5′ UTR-embedded translational control elements in breast cancer in high throughput, we will utilize the recently developed Direct Analysis of Ribosome Targeting (DART). We anticipate that widespread TSS switching throughout metastatic progression will produce distinct 5′ UTR isoforms that alter translational output via enhancer or repressor motifs. Identifying novel translational control elements in 5′ UTRs will elucidate the 5′ UTR regulatory code and inform novel anticancer therapeutic strategies.

Smith, J.S.,¹ Dolinoy, D.C.,¹ Jones, T.J.,¹ Colwell, M.,¹ Perera, B.P.U.¹

1 Department of Environmental Health Sciences, University of Michigan, Ann Arbor, MI

The study explores the expression of Piwil mRNA in mouse embryonic head and placenta during 13-14 days post conception (dpc), aiming to characterize the PIWI-piRNA system in somatic in utero tissues. Prenatal exposures can impact offspring health outcomes through epigenetic mechanisms, including small non-coding RNA (ncRNA) like PIWI-interacting RNAs (piRNAs). While traditionally associated with the germline, recent research from our lab indicates piRNAs and their associated PIWIL proteins are expressed in somatic tissues.

Adult C57BL/6 mice were time mated and 10 embryos and placentas (5 M, 5 F) were harvested 13-14 dpc from a total of 8 different litters. No more than 1 male and 1 female were used per litter. Tissues were homogenized and RNA was extracted and verified for quality. RT-qPCR was conducted using Taqman probes to analyze Piwil1, 2, and 4 expressions in the fetal brain and placenta. Results revealed low expression of Piwil1, 2, and 4 in the fetal brain, while in the placenta, Piwil1 was absent, and Piwil2 showed significantly higher expression than Piwil4. All results were compared with adult testis tissue, our positive control.

The findings suggest the presence of Piwil machinery in mouse somatic tissues during embryonic development. Future research will involve characterizing the piRNA-ome in these tissues, particularly exploring the impact of environmental toxicants, such as lead. The study aims to uncover epigenetic mechanisms influencing early life exposures and persistent adverse outcomes, with potential implications for intervention and therapeutic strategies, including epigenetic editing approaches.

Jeffrey N Dudley,^1-2 Mats Ljungman,³ Sethu Pitchiaya^1,4-5

1 Program in Cell and Molecular Biology, University of Michigan, Ann Arbor, MI
2 Medical Scientist Training Program, University of Michigan, Ann Arbor, MI
3 Department of Radiation Oncology, University of Michigan, Ann Arbor, MI
4 Department of Pathology, University of Michigan, Ann Arbor, MI
5 Department of Urology, University of Michigan, Ann Arbor, MI

In response to environmental stress, mammalian cells globally downregulate transcription, but selectively induce genes that facilitate cell fate decisions. One emerging feature of such selectivity is the extensive transcription (~30 kb) of genes beyond their canonical end sites, i.e. readthrough-transcription (RTxn), which possibly results from disruptions of transcription termination (DOTT). We and others have shown RTxn in diverse contexts, including renal cell carcinoma, influenza virus infection, cellular senescence, and osmotic stress. However, the molecular mechanisms by which RTxn occurs and the role of these transcripts in stress response is unknown. To this end, we used a combination of cutting-edge omics and single-molecule imaging methods. Nascent transcriptome sequencing (via Bru-Seq) of cells exposed to stresses that mimic the tumor microenvironment and physiological perturbations (e.g. hyperosmotic, oxidative, and proteotoxic stress) revealed the widespread persistence of DOTT with up to 400 genes exhibiting RTxn. Notably, transcript lengthening had a negative relationship with gene expression. Using epigenomic interrogation by assay for transposase accessible chromatin via sequencing (ATAC-Seq), we found that genomic loci exhibiting readthrough gain chromatin accessibility upon stress, suggesting stress-responsive selective remodeling of the epigenome. Finally, using single-molecule RNA fluorescence in situ hybridization (smFISH) we find that lengthened transcripts are largely nuclear with most transcripts remaining at the site of transcription, paving the way toward extending this imaging toolkit for identifying DOTT regulators at high-throughput. Overall, our work strongly indicates dynamic alteration of the epigenome and chromatin organization for transcriptional regulation during stress, which together potentially impact cell fitness.

D. L. Ciotlos,^1-2 S. C. Hanks,² A. Varshney,², M. R. Erdos³, N. Manickam², H. M. Stringham¹, P. Orchard², E. Hill-Burns¹, N. Narisu³, L. Bonnycastle³, M. D. Sweeney¹, M. Laakso^4,5, J. Tuomilehto⁶, T. A. Lakka⁴, K. L. Mohlke⁷, M. Boehnke¹, H. A. Koistinen⁶, F. S. Collins³, S. C. J. Parker^2,8, L. J. Scott¹

1 Biostatistics and Center for Statistical Genetics, University of Michigan, Ann Arbor, MI
2 Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI
3 NHGRI, NIH, Bethesda, MD
4 Inst Clinical Medicine, Internal Medicine, Univ of Eastern Finland, Kuopio, Finland
5 Dept Medicine, Kuopio Univ Hospital, Kuopio, Finland
6 Dept of Public Health and Welfare, Finnish Inst for Health and Welfare, Helsinki, Finland
7 Genetics, Univ of North Carolina, Chapel Hill, NC
8 Human Genetics, Univ of Michigan, Ann Arbor, MI

Skeletal muscle plays a role in regulation of glucose metabolism, and physical activity is a preventive factor for type 2 diabetes (T2D). To understand the relationships of T2D status and vigorous physical activity (VPA) with gene expression and chromatin regulation in skeletal muscle cell types, we analyzed skeletal muscle single nucleus RNA-sequencing and ATAC-sequencing data from 279 Finnish adults from the FUSION Tissue Biopsy Study. Participants were 60.2 (SD 7.4) years-of-age, 41.5% were female, and had varying glucose tolerance, ranging from normal glucose tolerance (NGT, n=102) to newly-diagnosed T2D (n=72). Nucleus-level data were clustered using LIGER. We used negative binomial models to test for associations between glucose tolerance or VPA and cell-type composition, gene expression, or chromatin accessibility (genes: 23,841; peaks: 924,519). Across muscle fiber types, associated genes generally had the same direction of gene expression when having higher VPA or NGT (as compared to T2D). We observed a similar pattern for chromatin accessibility. For example, in Type 1 muscle fiber, NGT (as compared to T2D) or higher VPA were each associated with higher expression of cellular respiration genes (VPA: OR=1.77, p=1.4e-9, T2D vs. NGT: OR=0.80, p=8.9e-5) and lower expression of protein K48-linked polyubiquitination genes (VPA: OR=0.64, p=5.2e-3, T2D vs. NGT: OR=2.03, p=6.2e-5). These findings suggest that both normal glucose tolerance and higher vigorous physical activity have similar regulatory effects on gene expression.

Ngai So,¹ Stacy Hovde,² Bruno Caetano Trindade,¹ Jiachen Chen,¹ Sergey Seregin,¹ Adam H. Courtney,³ Ronald W. Henry,² Grace Y. Chen¹

1 Department of Internal Medicine, University of Michigan, Ann Arbor, MI
2 Department of Biochemistry, Michigan State University, East Lansing, MI
3 Department of Pharmacology, University of Michigan Medical School, Ann Arbor, MI

Non-coding RNAs play critical roles in cellular homeostasis, and their altered expression has been associated with cancer. It has been shown that patients with colorectal cancer have heightened circulating levels of a fragment of the U2 small nuclear RNA (snRNA). However, little is known about the target and function of these small U2 RNA fragments. Using a mouse model of colon tumorigenesis, we observed increased expression of U2 snRNA in colon tumors, which was also associated with elevated levels of circulating U2 fragments (U2-f). We also determined that U2-f is capable of modulating immune responses. Specifically, U2-f inhibited the production of Th1 cytokines, such as IFN, after activation of T cells. RNA-seq analysis also revealed suppression of gene expression related to DNA replication and cell cycle control. Consistently, we demonstrated that U2-f induced cell cycle arrest of activated T cells and reduced T cell proliferation. Importantly, intratumoral injection of U2-f in congenic MC38 and MC38-OVA colorectal cancers in mice resulted in increased growth and reduced tumor regression, respectively. Our studies suggest that: 1) the U2 snRNA is upregulated in colorectal cancers, which leads to increased abundance of U2-f within the circulation and within the tumor microenvironment, and 2) increased levels of U2-f affect T cell responses that promote the growth of tumors by inhibiting anti-tumor immunity. These studies point to a role for U2-f in the regulation of T cells and suggest that U2-f may be a novel immunotherapeutic target in the treatment of colorectal cancer.

Stephen DeAngelo,¹ Sofi Dziechciarz,¹ Catherine Wilhelm,¹Markos Koutmos,¹ Yatrik Shah¹

1 Departments of Cancer Biology, Chemistry, and Biophysics, University of Michigan, Ann Arbor, MI

Colorectal cancer (CRC), an increasingly prevalent disease in the US, is particularly sensitive to disruptions of redox homeostasis and oxidative cell death. Despite this, many inducers of oxidative cells death have failed to transition to the clinic. One emerging strategy to induce targeted cell death in advanced CRC is inhibition of the selenoproteome – 25 selenocysteine containing enyzmes that maintain cellular redox homeostasis through ROS detoxification. Here we take a novel approach targeting the selenoproteome in cancer through investigation of essential RNA modifications on tRNA-selenocysteine. The tRNA methyltransferase AlkBH8 methylates U34 in the anticodon loop of tRNA-selenocysteine and knockout of AlkBH8 leads to an inability to translate selenoproteins. Our preliminary data demonstrates that AlkBH8 is overexpressed in CRC, and that overexpression of AlkBH8 is correlated with a worse prognostic outcome. Therefore, we hypothesized that knockdown or inhibition of AlkBH8 in CRC would lead to global disruption of the selenoproteome and might present a novel therapeutic vulnerability in this disease.

To understand the functional role of AlkBH8 in CRC, we generated inducible knockdown (KD) lines of AlkBH8. The results revealed a reduction in selenoproteins and diminished growth across all CRC cell lines as assessed by cell counts and colony forming assays. Notably, this effect was cancer selective. However, contrary to our current understanding of the selenoproteome, a subset of cell lines exhibited growth arrest without oxidative cell death following AlkBH8 KD. These data suggest that there may be multiple biological mechanisms through which AlkBH8 influences CRC proliferation and tumor-forming capabilities.

Brandon J.C. Klein,¹ Jose Reyes-Franceschi,² Emilio Cardenas,¹ Chase Weidmann,³ Amanda L. Garner^1-2

1 Department of Medicinal Chemistry, College of Pharmacy, University of Michigan, Ann Arbor, MI
2 Program in Chemical Biology, University of Michigan, Ann Arbor, MI
3 Department of Biological Chemistry, Center for RNA Biomedicine, University of Michigan, Ann Arbor, MI

Affinity-based chemical probes for RNAs offer an exciting look at putatively ligandable RNA structures within the transcriptome. These techniques couple bespoke small-molecule probes to an NGS read-out to identify chemical matter that binds to RNAs. Current approaches to analyze these datasets use either differential expression analysis or quantification of reverse-transcriptase stops as the key read-out for feature detection. This work compares these approaches to a novel ChIP-seq-inspired analysis pipeline. This novel containerized workflow aims to identify and rank-order the binding loci of these small molecule probes through a combination of measures. These measures combine traditional enrichment-based feature identification with alternative measures of RNA structure and function to rank features of biological importance. These analyses demonstrate the limitations of photo-affinity-based chemical probing when experimental designs fail to include adequate controls. Informatic analysis and experimental replication reveals previously reported features in these published datasets that are attributable to input expression levels and not solely affinity-based enrichment.

Shurong Zhou,¹ Suling Yang,¹ Guizhi Zhu¹

1 Department of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI

Cytosolic dsRNA activates cytosolic RNA sensors such as MDA5 and of MAVS‐mediated type-I interferon (IFN) signaling, which can benefit tumor immunotherapy[1]. Adenosine deaminase acting on RNA type I (ADAR1) is a ubiquitous endogenous RNA base editor in vertebrates that edit adenosine (A) to inosine (I) in double-stranded RNA (dsRNA). ADAR editing reduces the ability of dsRNA to elicit type-I IFN responses, making ADAR1 a potential therapeutic target for the immunotherapy of interferon-stimulated gene (ISG)-positive cancers. Specifically, IFN-stimulated positive tumor cells, such as triple negative breast cancer (TNBC)[2], can be sensitized via ADAR1 loss via ISG activation and tumor cell apoptosis[3]. CRISPR-Cas13d is an emerging technology with great RNA knockdown efficiency and low off-target effects[4]. CRISPR-Cas13d system is composed of a RNA cleaving enzyme, Cas13d, and a guide RNA (gRNA), the latter of which is comprised of a Cas13d-binding hairpin-like RNA and an RNA region complementary to target RNA. However, the development of CRISPR-Cas13d RNA editing therapeutics can be limited by 1) pre-existing host immunity against exogenous Cas13d protein as reported in humans[5], and 2) safety concerns associated with Cas13d-expressing plasmids. Here, we present the development of Cas13d mRNA delivered together with ADAR1-specific circular gRNA to sensitize tumor cells for TNBC immunotherapy. In contrast to linear gRNA, circular gRNA improves its biostability and ADAR1 knockdown efficiency in vitro and in vivo in a 4T1 triple negative breast cancer tumor model. Overall, these results provide preliminary support for the development of ADAR1-targeting Cas13d mRNA and circular gRNA for TNBC combination immunotherapy.

X. You,¹Z. Yu,¹ Z. Guizhi¹

1 Department of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI

Despite mRNA vaccines in the combat the emergent threat of SARS-Cov2, an inhalation-based mRNA therapy for lung diseases remains challenging. Current modified mRNA vaccines have low stability, suboptimal prophylactic/therapeutic potency and duration, and limited safety, leading to ineffective therapy. We report here an inhalable small circular mRNA (circRNA) vaccine with highly stable, limit potential toxicities, and sustained production of antigens, resulting in robust and long-lasting adaptive immunity against lung diseases. Artificial intelligence was used to optimize circRNA sequences for high stability and gene expression capability. Relative to state-of-the-art modified mRNAs, circRNA showed stable physicochemical properties for at least 12 h in 10% FBS, and high loading copies per LNP. Moreover, inhalable circRNA vaccines achieved high levels of both circulating T cell memory and lung resident memory T cells, with superior safety and low protein kinase R (PKR) activation. Thus, the development of a stable and efficacious small circRNA pulmonary delivery system would enable high therapeutic concentrations locally in the lungs to improve efficacy and limit potential toxicities. Overall, small circRNA vaccines are promising for versatile applications.

Xiang Liu,¹ Yu Zhang,¹ Guizhi Zhu¹

1 Department of Pharmaceutical Sciences, University of Michigan, Ann Arbor, MI

Messenger RNA (mRNA) has emerged as a novel vaccine platform and has proven highly effective during pandemic. This platform had also found application in the development of cancer vaccines.

Despite the encouraging outcomes observed in early-phase clinical trials, there are still certain concerns associated with current mRNA-based cancer vaccines such as limited stability, suboptimal tumor therapy potency and durability, and limited safety. Here, we report novel highly stable antigen-encoding small circular mRNA (circRNA) vaccines that elicit potent and durable T cell responses for robust tumour immunotherapy. Compared to modified mRNA and current long circRNA, small circRNA shows superior stability, minimal protein kinase R activation, and low cytotoxicity. This allows small circRNA vaccines to sustain efficient antigen translation over an extended period, while activating pattern recognition receptors for innate immunostimulation. Relative to several modified mRNAs, nanocarrier-delivered small circRNA vaccines show superior safety and elicit up to 10-fold antigen-specific T cells, accounting for ~25%-75% total peripheral CD8+ T cells over 6 months in mice. circRNA vaccines are applicable for various major histocompatibility complex I/II-restricted tumour and viral (neo)antigens to elicit CD8+ and CD4+ T cell responses, in young adult and immunosenescent aged mice. In mice, mono-/multi-valent circRNA vaccines plus immune checkpoint blockade reduced tumour immunosuppression and eradicated multiple types of tumours, including immunotherapy-resistant BrafV600E melanoma. Overall, small circRNA vaccines represent a promising advancement in the field of tumor immunotherapy, offering optimal stability, safety, and the potential to generate potent and enduring immune responses against a variety of tumors.

Yi Xu,¹ Kate Kuntz,¹ Jeffrey Pina,¹ Zhenfeng Liu¹

1 Zymo Research Corporation, Irvine, CA

In a total RNA sample, ribosomal RNA (rRNA) usually makes up as high as 90% of the content. Hence, rRNA depletion is often implemented in total RNA library prep for a more efficient and informative gene expression analysis and transcript quantification. Most existing rRNA depletion strategies are probe-based and species dependent, well-established for commonly studied mouse and rat models besides human samples. Research utilizing other models and non-model organisms often lacks a simple and reliable total RNA library prep method with rRNA depletion integrated. To address this need, Zymo Research developed a novel RNA library prep procedure that integrates a probe-free and species-independent rRNA depletion. Total RNA from eight species (mouse, rat, human, cow, tomato, wheat, yeast, and green algae) were utilized in RiboFree library preparation. The resulting libraries achieved an average unique alignment rate of 75.7% across all samples with > 92% uniquely aligned in samples from human. rRNA was depleted to ≤ 9% across the tested species. Exceptional numbers of genes were also detected: over 30,000 genes were detected in the human samples where > 8,000 were lncRNA; in the C. albican samples, over 6,100 genes were detected among the ~ 6,260 annotated genes in the reference assembly. Furthermore, Zymo Research deposited an RNA-Seq pipeline to a no-code platform called Aladdin to provide a ready-to-use data analysis tool for the scientific community. Within several “point-and-click”, researchers can execute essential steps including trimming, alignment, and quality control. The pipeline integrates differential gene expression analysis, the most common type of RNA-Seq data analysis, and supports 14 reference assemblies across 12 different species so far. The powerful Zymo-Seq RiboFree total RNA library prep protocol and the streamlined, autolaunchable Aladdin platform will greatly empower researchers from diverse backgrounds to make impactful contributions for a broader wealth of knowledge and innovation.

Contact: yxu@zymoresearch.com