DNA is the backbone of the genetic code so, logically, decoding and understanding the genetic sequence along with epigenetic modifications came first. Yet, RNA and the emerging field of epitranscriptomics are playing increasingly crucial roles in both health and disease. This class of nucleic acids controls the proteins that keep cellular biology chugging along, or, in the case of disease, misfiring.
Take these molecules together with all of the cellular signals that regulate RNA and protein production, and the degree of complexity is rapidly apparent, if not overwhelming, to decryption by the human brain. Companies, particularly AI-first TechBio companies, are harnessing AI to rapidly process data and the endless biological possibilities to accelerate the field of genetics-based medicines.
Although RNA therapeutics may target similar diseases to one-and-done DNA therapeutics, they offer several unique advantages. The underlying oligonucleotide technology has already been proven and, thus, the regulatory pathway is not as arduous. Any deleterious reactions are temporary because of the transience of RNA therapies. The effects of RNA therapeutics are transient, so repeated dosing is necessitated for the therapeutic to remain effective, but that also means any deleterious reactions are temporary. Dosage is expected to become less frequent as therapeutics and delivery mechanisms move to next-generation.
As mRNA therapeutics advance, especially in the proof-of-concept disease AATD (alpha-1 antitrypsin deficiency), clinical data are expected later this year. In addition, new takes on therapeutics are emerging as companies such as AIRNA explore emerging options to impact human health through the introduction of healthy variants.
Optimizing an endogenous system
The OPERA® (oligonucleotide-promoted editing of RNA) platform is a precise, programmable approach for selective editing of adenosine to inosine using Adenosine Deaminase Acting on RNA (ADAR) enzymes. “We optimize the structure and stability of our fully synthetic, chemically-modified oligonucleotides for high-precision, efficient editing across a broad range of diseases, while minimizing off-target and bystander edits,” summarized Loïc Vincent, PhD, CSO at Korro Bio.
Four synergistic foundational pillars make the OPERA platform an advanced and scalable RNA editing system. First, the Korro team has a deep understanding of ADAR, which empowers OPERA to specifically and precisely harness endogenous editing machinery. Second, expertise in oligonucleotide chemistry to develop CHORD (Customized High-fidelity Oligonucleotides for RNA Deamination) allows fine-tuning of stability, potency, and safety.
Third, an in-house AI/ML framework continually learns and facilitates target identification and in silico design and optimization. “This translates to a rapid design and test iteration of oligonucleotides in under four weeks,” said Vincent. “Lastly, our delivery strategy is to go beyond the liver to tissues that have historically been difficult to reach with oligonucleotides.”
The company is currently executing a 3-2-1 pipeline strategy with the goal of establishing three clinical programs to two different tissue types using the OPERA platform.
KRRO-110, a wholly-owned therapeutic candidate, is a potential best-in-class, disease-modifying treatment for AATD. The REWRITE clinical trial is a two-part study to evaluate KRRO-110’s safety and durability. Up to 64 healthy and PiZZ genotype participants will help establish the dosage for later-stage studies. In addition, the FDA has granted Korro Bio orphan drug status for use of KRRO-110 for treatment of AATD.
Other early-stage programs target both rare and highly prevalent diseases, as well as transcriptome rewiring in complex diseases. “Our goal is to show versatility across different disorders and tissues,” Vincent emphasized. Another notable program focuses on pain, where a significant unmet need for safe and effective, non-opioid management remains. The company is also collaborating with Novo Nordisk with an initial focus on cardiometabolic disease.
Vincent believes that incorporating more AI capability would act as an accelerant. “With AI, we can continue to refine OPERA’s power, expand the design space, and turn empirical drug discovery into a faster and more robust, predictive and specific process,” said Vincent.
Healthy variant therapeutics
The 2019 Nature publication describing the use of an oligonucleotide to recruit an endogenous ADAR enzyme to make an RNA base edit kicked off the commercial interest in the molecule.1
“There are an infinite number of possibilities on how to design the oligonucleotides. If you compound all of the different parameters, the optimization of adenosine to inosine editing is quite complex. This is why the field has struggled with an in vitro to in vivo translation,” said Kris Elverum, president and CEO of AIRNA. “Our optimization of the potency, safety, and delivery of these oligonucleotides to achieve a greater editing efficiency is driven by our backbone base modification, structure-based design as well as optimized GalNAc delivery to the liver.”
Overcoming technical challenges is ongoing. ADAR itself is complex: For example, the sequence for the human ADAR1 cDNA predicts an open reading frame of 1226 amino acids, and the enzyme has different isoforms, some of which are cytosolic, nuclear, expressed in response to stimuli, or continually present.2
“We are now at an inflection point with the earliest clinical data coming out,” said Elverum. Other parameters, such as manufacturing, safety, and preclinical models, are well-established for oligonucleotides.
AIRNA plans to submit their clinical trial application in the second half of this year for their lead candidate, AIR-001, to treat AATD. A recent round of funding provides the resources for this effort.
RNA editing is positioned to be a breakthrough modality for multiple medicines. “We are incorporating proven components from other modalities for mechanism of action, administration, distribution, etc. With our RESTORE+ platform, we believe our medicine will be preferred by patients because of the key elements of our technology, which positively impact potency, safety, and delivery,” said Elverum.
AIRNA is also focusing on how individuals with superior health evolve. “We can identify cohorts who have superior health against a given characteristic and identify the genetic alterations driving these differences,” said Elverum. “Although point mutations will vary among individuals, long term, we believe the introduction of healthy variants to address disease will open up a whole new field for medicines and improve health.”
AI driving new discovery
Deep Genomics’ mission is to pioneer AI foundation models to transform R&D of novel genetic medicines. The AI-first TechBio company’s flagship foundation model, BigRNA, has 1.8 billion tunable parameters trained on over one trillion genomic signals. The more data fed in by “lab-in-the-loop,” the more intelligent BigRNA becomes.
For example, the initial training data did not involve oligonucleotides, but “once BigRNA was trained, we were able to provide the foundation model with a gene and ask it to find oligonucleotides to increase its expression. The oligonucleotides it found were then experimentally validated,” said Tehmina Masud, PhD, MBSS, vice president of systems and target biology at Deep Genomics.
BigRNA was recently piloted in one of the largest explorations into complex genetic conditions. “By leveraging a massive amount of computing power and Google’s specialized TPU architecture, we increased our understanding of the genetic and biological determinants of complex and rare diseases to create a database that will drive innovation,” Masud highlighted.
The modality-agnostic target discovery platform is uniquely equipped to identify RNA editing, knockdown, or upregulation targets. Recently added models—REPRESS, DeepRNAi, and DeepADAR—support siRNA, RNA editing, and mRNA therapeutic design. DeepADAR enables the design of high-performing gRNAs by accurately predicting the effects of ADAR-recruiting guides on the target sites.
Going beyond BigRNA, Deep Genomics’ foundation model platform contains a variety of foundation models for nucleic acids, proteins, and systems biology to address multifaceted challenges in therapeutic development, and analyze vast biological datasets to tackle multiple problems simultaneously. The models demonstrate emergent intelligence, solving problems beyond their original training.
“We believe the model where TechBio companies deliver early-stage breakthroughs and pharma partners lead later-stage development is not only scalable but the key to unlocking AI’s potential in drug discovery,” said Masud. “We want to raise the tide of the entire industry, reshaping how genetic medicines are discovered and developed.”
Masud is particularly excited about the new patient-centric and multimodal data integration approach to further advance the model. Accessing large patient cohorts and biologically relevant multimodal data, including detailed clinical records, has long been a major challenge for conducting well-powered analyses in study designs.
The emerging role of epitranscriptomics
RNA modifications play a critical role in modulating most aspects of RNA biology, including protein translation. Intricately regulated by complex cellular pathways, RNA modifications are frequently dysregulated in various diseases, including cancer, cardiometabolic conditions, and neurological disorders.
But the lack of accurate, robust, and user-friendly methods for their detection and analysis limits the use of epitranscriptomics in drug development. Indeed, only a small number of highly specialized laboratories perform this type of analysis. Currently, the most common synthetic application of RNA modifications is in antisense oligonucleotides (ASOs) and RNA vaccines to enhance their stability, reduce immunogenicity, and improve pharmacokinetics.
“A steady year-over-year increase in the number of scientific publications and the first emerging translational studies illustrate the field’s rapid evolution,” said Gudrun Stengel, PhD, CEO and co-founder at Alida Biosciences.
Alida Biosciences intends to embed epitranscriptomic analysis seamlessly into diagnostics, drug discovery, and RNA therapeutic development, where it hopes to unlock novel insights and add measurable value. Their EpiPlex™ platform offers a comprehensive, end-to-end workflow for the detection and analysis of RNA modifications, enabling the interrogation of multiple RNA modifications in a single reaction.
The streamlined approach significantly reduces the need for input material—allowing for biopsy use—and minimizes time and labor. Most crucially, it provides actionable insights into the interrelationships and correlative changes between different RNA modifications, including the most abundant modifications m6A and inosine, and—soon—pseudouridine.
By harnessing the power of proximity barcoding, the EpiPlex assay, which includes the detection of m6A and inosine, and ongoing efforts to incorporate pseudouridine, effectively translates the presence of RNA modifications into unique barcodes that are read using NGS.
The accompanying cloud-based Epi-Scout™ analysis software employs an ML-based algorithm to interpret the data, delivering the transcript location and relative abundance of each RNA modification, along with gene expression data.
“We aspire to capture the full epitranscriptomic landscape of mRNA to facilitate a more comprehensive understanding of disease mechanisms and the development of multifactorial diagnostic tools,” Stengel said. “To remove barriers to adoption, we prioritize robustness, ease-of-use, high throughput, cost-efficient sequencing, and the integration of powerful bioinformatics tools.”
References
1. Merkle T, et al. Nat Biotechnol. 2019 Feb;37(2):133-138. doi: 10.1038/s41587-019-0013-6.
2. Samuel CE. J Biol Chem. 2019 Feb;29 4(5):1710-1720. doi: 10.1074/jbc.TM118.004166.
