The Evolution of the New RNA World
Updated: Mar 26
By Rabia Khan, CEO Serna Bio
We are Serna Bio: our mission to design novel small molecule modulators of RNA function.
We believe in the vast, untapped power of using the most successful drug modality (small molecules) to target the largest constituent of the human genome (RNA).
Over the next few blog posts, I hope to shed light on why we believe this combination is transformational and present an overview of the current competitive landscape and approaches in this burgeoning field.
Our expanding appreciation of RNA
Proteins have long been thought of as the “functional units of biology”, and that the remainder of the human genome (the non-protein coding genome) was “junk DNA”. This idea stemmed from the belief that a gene would encode mRNA which would make protein, called the Central Dogma of Biology. Prior to sequencing the human genome, it was naively believed that the number of protein coding genes would increase with the complexity of the organism, and each gene would be associated with a disease. Thereby, sequencing the human genome would rapidly open up our understanding of disease - gene associations.
Yet, as expected, biology is messy. With the first full sequence of the human genome in 2000 [Human Genome Project Timeline] came the realisation that the central dogma of biology (DNA makes mRNA makes protein) wasn’t that simple and RNA carries a multitude of functions beyond coding for proteins.
The Central Dogma of Biology
As former NIH Director Francis Collins put it:
“Only a decade ago, most scientists thought humans had about 100,000 (protein coding) genes. When we analyzed the working draft of the human genome sequence three years ago, we estimated there were about 30,000 to 35,000 genes, which surprised many”
Over 20 years on, we still do not understand the function and role of the majority of the human genome, and genetics has largely focused on studying the proteome. By current estimates, ~80% of the human genome encodes RNA, while only 2% encodes proteins. Of the whole genome, only a scant 0.4% has been explored by classical, protein-based drug discovery methodologies [Principles for targeting RNA with drug-like small molecules | Nature Reviews Drug Discovery]. The remaining 99.6% of the genome, and in particular the non-coding transcriptome, represents an unexplored universe of novel potential therapeutic targets [Beyond the RNA-dependent function of LncRNA genes] .
This so called “junk DNA” is anything but, and we know it is encoding a vast library of RNAs [The roles of structural dynamics in the cellular functions of RNAs., Non-coding RNA: what is functional and what is junk? , Junk DNA and the long non-coding RNA twist in cancer genetics ].Recent projects like the Gene Regulation Observatory [Gene Regulation Observatory (GRO)] and ENCODE [Encode Project ] continue to expand our understanding of the function of this “junk DNA” . Interestingly, recent work questions the most basic classification of RNAs as “non-coding” or “coding”, with initial evidence in this recent article stating that many lncRNA’s can express proteins, opening up the idea that more of our genome may be “coding” than initially appreciated.
We believe that RNA, the transcriptome, represents the next frontier in drug discovery. An untapped area of human biology that has yet to be understood.
Over the past decade, our understanding of RNA, structure and function has rapidly accelerated. This is driven by a combination of biochemical and computational advances that are enabling the dissection of cellular transcriptomes at unprecedented resolution. The increasingly complex RNASeq datasets combined with our ability to study splicing isoforms at cellular level and in transcriptome-scale mapping of RNA structures, we are just beginning to unlock a unique understanding of RNA’s role in disease biology which will enable precise targeting of disease- and even tissue-specific RNA.
The COVID-19 pandemic accelerated the path of mRNA as a modality, but RNA as a target for small molecules is an area where a systemic, repeatable solution still does not exist. We believe targeting RNA represents a unique opportunity to unleash pharma’s workhorse, the small molecule, on an entirely new target class. This approach does have its own challenges, of course, and the most common question we receive is:
Does RNA Form Druggable Pockets for Small Molecules?
Although many antibiotics target RNA [Antibiotic drugs targeting bacterial RNAs], it is frequently quoted that RNA does not form druggable pockets like those found in proteins. This idea is due to the relative simplicity of RNA, which has 4 nucleotide building blocks, compared to the 20 amino acid building blocks in proteins. This leads to the assumption that RNA cannot fold into structures that are sufficiently “diverse” to enable specific and selective interactions with small molecules.
However, we (and many others in the field) believe that this view does not take into account that an n-nucleotide long RNA can potentially form nearly 2n different structural conformations [RNA secondary structures and their prediction]. Additionally, each RNA nucleotide has 8 degrees of freedom (six backbone torsions, one torsion between the ribose ring and the base, and two possible configurations for the ribose ring's pucker), ensuring a tremendous structural diversity across the transcriptome [RNA Folding: Conformational Statistics, Folding Kinetics, and Ion Electrostatics]. Critically, these different structural conformations adopted by the RNA are not merely cosmetic and can regulate multiple aspects of the RNA function, including protein binding, translation and splicing.
The tide is turning on this misconception, with ever-growing evidence that RNAs do indeed harbor “druggable pockets'' comparable to those found within proteins [Evidence for ligandable sites in structured RNA throughout the Protein Data Bank]. For more on the druggability of RNA, we recommend these comprehensive reviews [Amanda Garner Review, Matt Disney Review and Kevin Weeks Review,]
The rapid increase in our understanding of the biology of RNA, its druggability, and how it can be manipulated to impact disease-related pathways has inspired the formation of a number of biotechnology companies.
In targeting RNA, we are referring to a target space that is 35x larger than what is being explored by classical drug discovery. I believe that a decade from today, there will be more RNA targeting companies than protein targets. With this blog, I wanted to share our perspective on how we segment the market. A perspective that we find useful is here [The emerging landscape of RNA-targeted small molecules].
Small Molecule targeting of RNA: An Competitive Landscape
I am frequently asked about our perspective on the competitive landscape for small-molecule RNA targeting companies. Although there isn’t a clear differentiation, we have classified companies as “functional-phenotypic screening” or “target based drug discovery” [Promises and Challenges of Target-Based Drug Discovery]
Phenotypic screening-based companies
Prior to the Human Genome Project, all drug discovery efforts focused on identifying molecules that led to a change in a disease-related phenotype, a technique known as phenotypic screening. This approach is undoubtedly successful, with the
“Majority of the first-in-class drugs approved by the US Food and Drug Administration (FDA) between 1999 and 2008 were discovered empirically without a drug target hypothesis” [Modern Phenotypic Drug Discovery ].
Phenotypic screening is a natural fit for identifying RNA-targeting drugs, and since it is agnostic to the target and can be used to assay a concrete functional read out, such as changes in splicing pattern of a gene, using reporter systems. A recent pre-print even suggests that a large number of FDA approved drugs also interact with RNA [Pervasive Transcriptome Interactions of Protein-Targeted Drugs]. Unlike classical phenotypic drug discovery, where the target is unknown, here we refer to functional-phenotypic drug discovery, where a single target-function combination (eg: SMN2-splicing) is known and phenotypic screening is used to assess the functional impact of the small molecule. Here, the mechanism of action of the small molecule can be via an RBP (RNA binding protein)-RNA interface rather than modulation of the RNA structure itself.
This approach works, as Risdiplam, the first RNA-targeted drug modulating the splicing of a human gene, was discovered via a phenotypic screen by PTC Therapeutics [FDA approves RNA-targeting small molecule]. Since then, a number of other splicing-focused, phenotypic screening companies (Remix and SkyHawk Therapeutics, ReviR) have launched. The advantages of a phenotypic screening approach, particularly for RNA, is it allows you to develop molecules that affect RNA function, without having to understand the RNA structure. Furthermore, one can use well established methods (RNASeq) to study the off-target profiles of the small molecule.
Phenotypic screening drug discovery is not without its challenges, however, and many of these are directly tied to the lack of knowledge of the target, including unexpected toxicity and off-target profiles. A review of the AstraZeneca pipeline [Lessons learned from the fate of AstraZeneca's drug pipeline: a five-dimensional framework] showcases how the cause of drug failure has shifted from toxicity to lack of efficacy overtime, and it is hypothesized that this mirrors the shift in the industry away from phenotypic screening-based drug discovery and towards target-based drug discovery. As applied to RNA-small molecule drug discovery, the same challenges persist and phenotypic screening can result in downstream challenges in tox, compounded by the lack of tools to study RNA-off-target profiles.
After the sequencing of the human genome, and understanding of the genetic drivers of disease, we entered an era of target based drug discovery, and the “druggable genome”, which is actually the “druggable proteome” [The druggable genome and support for target identification and validation in drug development].
RNA-targeting drug discovery has followed a similar trend, starting with phenotypic screening and then incorporating more target-based approaches. The introduction of low cost sequencing and high throughput structure probing methods has meant for the first time, we can study RNA structures to begin to understand RNA 2D structures. [RNA Structure Analysis at Single Nucleotide Resolution by SHAPE]
The ability to study RNA structure and specifically modulate a function is yet to be proven in the clinic, but the opportunity is large. This approach enables us to target classically “undruggable” proteins, like KRAS and MYC but also opens up an entirely novel space of targets - microRNA, lncRNA, tRNA to name a few. A number of companies operate in this space, using 2D/3D methods (Arrakis, Ribometrix, Nymirum, Ranar, as key examples).
Target-based approaches rely on the 2D and 3D structure of a particular RNA to identify a druggable pocket with sufficient information content to hit with a small molecule. This approach presents the opportunity to systematically design a molecule at the atomic level to interact with the desired RNA at specific locations in order to affect a biological function. Furthermore, it is worth emphasizing that while any RNA will adopt some sort of 2D and 3D structure, not all these structures will be functional. Therefore, functional characterisation of RNA structures is a key prerequisite of targeted screenings.
This wave of target-based companies benefits from the incredible pace of innovation in our ability to study RNA at a 2D, and increasingly at a 3D level, along with ever-improving computational methods (both bioinformatics and machine learning) to study the confirmations of RNA. Research performed at both companies, such as the newly launched Atomic AI, and within academia, such as that from the lab of Rhiju Das, [CASP15 RNA 3D Prediction]is pushing the whole space forward. We believe that more companies will continue to grow in this space, using now well-established structure probing methods in combination with machine learning and cryo-EM, XRay, and NMR-determined RNA structures.
For ease of exploration, we have included a list of companies in this space in a separate blog post.
While target-based approaches are on the rise and improving rapidly, the number of rationally designed small molecules targeting RNA is low, and PTC Therapeutics and phenotypic screening-based drug discovery still leads the way as the major method to bring RNA targeting molecules into the clinic.
It is important to note that despite their differences, both of the approaches described here face the same limitation: they are inherently bottom-up, focusing either on a narrow-range of phenotypes or targets. With >80% of the human genome transcribed into RNA, tapping into the full therapeutic potential of RNA will require evolving from case-by-case investigations to transcriptome-scale exploration. Additionally, this space is ripe for integration with new technologies developed in the protein-targeting space, most notably synthetic biology-enabled drug discovery methodologies, gene circuit science, high throughput data generation, and further integration of machine learning.
This field is at its infancy, with new methods, new companies and new modalities yet to be proven. But given the sheer size and scope of the druggable transcriptome, the pace of innovation, and the need for new therapeutic strategies, we believe that a decade from now there will be more RNA-targeting companies than protein-targeting companies. At Serna Bio, we’re excited to be on the cusp of this paradigm shift.
A thank you to Aaron Lazarus, Danny Incarnato and Lauren Richardson for their input.