The vast non-coding portion of the human genome has had many names, including ‘junk’ and ‘dark matter.’ But research, including the international ENCODE project, has shown that these regions are dynamically active and generate thousands of therapeutically unexplored long non-coding RNAs (lncRNAs). Since the initial inklings from tiling microarrays in 2002, there has been remarkable progress in understanding these lncRNAs, with a clearer picture of these molecules’ features and functional versatility. Current knowledge indicates that lncRNAs are key drivers of chronic disease processes and can be therapeutically targeted with high specificity.
HAYA Therapeutics is a Swiss precision therapeutics company that has developed tools and methods to interrogate lncRNA biology to design tissue- and cell-type-specific genomic medicines. Headquartered at the life sciences park Biopôle in Lausanne, Switzerland, with laboratory facilities at JLABS in San Diego, the company’s proprietary drug discovery engine targets lncRNAs that are drivers for fibrotic diseases and other severe health conditions associated with aging, including cancer. HAYA is utilizing emerging RNA-targeting modalities, such as modified antisense oligonucleotides (ASOs), to target and inhibit proprietary lncRNAs from preventing and reversing fibrosis.
HAYA—derived from the Arabic and Hebrew word ‘Hayat,’ which means life—is developing a pipeline of lncRNA-targeting candidates for the tissue-specific treatment of fibrotic diseases in other tissues, including lungs, kidneys, liver, and the micro-environment of solid tumor cancers. HAYA’s lead development candidate (HTX-001) is a modified ASO targeting the lncRNA Wisper, a cardiac myofibroblast enriched driver of fibrosis.
GEN Edge talked to Samir Ounzain, PhD, co-founder and CEO of HAYA Therapeutics, at BIO 2022 in San Diego to catch up on how the company is carving the path to commercializing lncRNA-targeting medicines.
GEN Edge: How and when was HAYA launched?
Ounzain: HAYA was founded in 2019, based on my academic research over the previous ten years. I’ve always been fascinated by the genetic switches that control gene programs linked to development and disease. My research focused on how the heart responds to the environment to drive remodeling disease processes. We realized that most of the information processing in cells when they react to the environment is not through proteins or the protein-coding portion of your genome—it’s the 98% of your genome that we used to call junk. That inspired me in this whole field when I was a student in 2001 and was told that 98% of our genetic material is junk! We have the same number and complement of genes as worms or flies. That spark drove my career—discovering what 98% of our DNA does.
Over the past 10-15 years, we learned that it’s the computational engine of the cell. It’s how your cells respond to the environment to change their state. The most common, chronic, and underserved diseases are driven by the environment and how the cell responds to the environment. Over the last decade, we realized most of the RNA is associated with the dark genome. This RNA is critical for this interface with the environment to control the cell state.
We discovered hundreds of lncRNAs when we created the first-ever atlas of these molecules in the heart. One of the critical features was that lncRNAs are incredibly disease-, cell- and tissue-specific. When you look at the clinical side and the clinical development of drugs for different processes like fibrosis, one of the things that became apparent was that most antifibrotics fail because of on-target toxicity in off-target tissues. If you’re drugging a pleiotropic protein signaling molecule—for example, TGF-beta—it can block fibrosis, but the same modality will block TGF-beta everywhere else. We have an atlas and molecules that control fibrosis just in the heart. So, if you can systemically administer therapy to drug them, you can have an antifibrotic effect only in the tissue where these molecules are present.
Our lead program targets a lncRNA we discovered called ‘Wisper.’ This is the first example of a tissue-specific—in this case, the heart—regulator of fibroblast activity. We use antisense oligonucleotides (ASOs) that can be systemically administered. Most ASOs tend to go in the kidney and liver, but you have enough exposure to other tissues, including the heart, to have an on-target effect. That’s what we observe. We can block fibrosis specifically in the heart. In the last two years, HAYA has applied these insights to many cell states in many tissues, including the lung and other types of fibroblast populations.
GEN Edge: Is HAYA targeting other indications?
Ounzain: We’re applying our discovery engine to additional fibrotic indications. The kidney and liver are areas of interest. We’ve also been working on the lung and have some exciting programs in pulmonary fibrosis. In pulmonary fibrosis, there is a fantastic consortium of single-cell data sets in patient samples in idiopathic pulmonary fibrosis (IPF). We take the IPF cell atlas and other tremendous public datasets, reanalyze them using our computational approaches, and couple that with in vitro models where we essentially profile every aspect you could imagine of the regulatory genome—chromatin conformation, genome topology, chromatin states, and nascent transcription—to build our atlas.
The reference genome annotations from transcriptome data currently are terrible. Most of the interesting transcripts are not annotated in reference genomes. We showed that in the first paper in 20141,2, where we laid the framework and discovered hundreds of novel RNAs in the heart. The reason we could find them was because of the sequencing depth. We sequenced to a depth of 250 million paired reads per sample.
Even today, most people cap depth at 50 million paired-end reads, but then you’re missing lowly expressed RNAs that are very tissue and cell-specific. One of the questions that always arises is that they’re so lowly expressed, so how can they have a function? We now know you can have super stoichiometric activity of these RNAs because of their ability to induce phase-separated domains, where many of these remarkable biochemical processes happen. That’s how we think about selecting tissues, and cell states: What are the public datasets available, and what can we generate rapidly internally?
GEN Edge: What is the current state of HAYA’s pipeline?
Ounzain: The process runs both in vivo and in vitro in two species, humans and mice. We’ll either generate or obtain various functional genomic transcriptomic data sets from the public domain from the disease patients. We’ll then work on a cell state with relatively translatable in vitro models.
For example, in lung fibrosis, you can obtain primary human lung fibroblasts. There are suitable protocols to differentiate these into myofibroblasts, and then we execute single-cell analysis in vivo and in vitro. We do single-cell RNA- and ATAC-seq so we can also map the chromatin landscape or accessibility landscape. The first step before we go deep is to ask how the cell states we have in the dish compare to the cell states you see in vivo in the disease setting, both in terms of transcriptional signatures at the single-cell level and in the chromatin landscape.
Suppose there is an overlap, which we see nicely in the lung. Human lung myofibroblasts in a dish are very comparable to the human lung myofibroblasts you see in IPF. We take our in vitro model and map and build the atlas over 12 different orthogonal genomic approaches, which we now use to map this landscape. Then it goes into what we call DiscoverHAYATM, leading to the creation of our HAYAtlasTM atlas, which we call our search engine, because once you have the atlas, you can start searching for features you care about. You can look for lncRNAs associated with super-enhancers or mega-trans-enhancers—pick your favorite. We focus on those associated that are highly tissue cell and disease-specific, like lung myofibroblasts.
Then we start integrating human genetics. Most genetic variations are associated with common and chronic diseases in these regulatory loci. We impute GWAS studies relevant to the disease or the cell we care about and ask, do any of these GWAS variants in linkage disequilibrium (LD) overlap the lncRNA loci. We can go from 5,000-10,000 lncRNA loci of interest, usually depending on the tissue and the heterogeneity in the cells. We can end up with a high-priority list of around five to twenty targets, which is manageable. Then we screen them in vitro. We use our antisense design platform to design ASOs against lncRNAs.
There are many features in lncRNAs that people don’t realize that influence their drugability, which you don’t have to consider when you are designing ASOs against mRNA molecules necessarily. We use the Nanopore platform to do direct RNA sequencing to look at RNA modifications, structure, and editing. We don’t know why, but lncRNAs are incredibly dynamic structures that influences your antisense design approach. We take all those features, start designing ASOs, and then screen them.
GEN Edge: Is HAYA monetizing the proprietary computation approaches?
Ounzain: It’s something that is taken into consideration. How can we maximize the value of the platform? How can we accelerate the value of the platform? I think the atlas that we built and are building is very much our proprietary information, but how we can leverage those atlases to accelerate specific programs is something we’re always open to discussing. I always like to think about the biology that drives these cell state transitions. What is the real biology underpinning that? We have insights, a toolbox, and modalities to start drugging that biology.
GEN Edge: Where is HAYA in its journey to bring a lead candidate to the clinic?
Ounzain: For the lead program targeting this RNA Wisper, we now have preclinical proof of concept in several rodent models. We are in the middle and almost finalizing our large animal translational study. The initial exploratory analysis looks very encouraging. We validated these ASOs across multiple species for our lead program. By the end of this year, we should have our development candidate for our lead target.
Next year we will go into IND enabling talks, looking for hopefully entering the clinic in early 2024 for our lead program. For the new targets we’ve identified in the lung, we are currently performing in vivo target proof of concept studies and, very soon, in vivo therapeutic proof of concept for these pulmonary assets.
The discovery is now being applied to other cell states and tissues. We recently were awarded a grant from the Swiss government to apply our approach to cancer-associated fibroblasts and the tumor microenvironment. We are identifying cancer-associated fibroblast targets that could be interesting to be combined with immunotherapies. One of the major issues in many solid tumors is the matrix of the cancer-associated fibroblast population. We are trying to ablate that population by targeting a lncRNA.
GEN Edge: Where do you want to be in a year, but also further out, say the next 5-10 years?
Ounzain: In five years, we’ll hopefully be a clinical-stage company. In 10 years, I’d like to see us having a clinically approved therapeutic in one of our indication areas. We’re starting with a small indication: non-obstructive hypertrophic cardiomyopathy, which has a strong fibrotic mechanistic underpinning. You can go quite fast in that patient population through clinical development.
Ultimately, I see HAYA changing how we think about drugs and common and chronic diseases because we know quite a lot now about where and how the genome is being programmed in response to the environment. We know most of these highly underserved, common, and chronic diseases where we have very limited, effective, safe, and accessible medicines are being hardwired there. If we find a way to map and understand the regulatory genome and how to drug it, that opens up many possibilities for different diseases. The ultimate vision of HAYA is to create truly effective, safe, accessible, and relatively simple medicines.
GEN Edge: How is HAYA approaching tissue-specific delivery of ASOs?
Ounzain: It’s less of an issue for us because, for most of the tissues we select, we know you can get uptake into those tissues. You still get exposure, even if it’s minimal compared to the liver and kidney. Many of the toxicity issues are typically around on-target liver and kidney engagement. If that’s where most of the ASO goes, and the target is also expressed in the liver and kidney, you’ll get most of the on-target effect there rather than in the heart. We don’t have those concerns because the targets are highly cell-type and tissue-specific. There are some tissues and cell types where you need to improve delivery, even if your target is very tissue specific.
You also need to consider the subcellular target when you start thinking about drugging the regulatory genome by targeting lncRNA. Most of the lncRNA targets are in the nucleus, and most are involved in regulating dynamic chromatin processes.
GEN Edge: What are your company’s major challenges?
Ounzain: The funding landscape right now is a little bit challenging! However, we can go forward at the moment. The talent pool is an issue. One of the reasons we came to San Diego was that there’s a lot of talent here in regulatory genome and RNA biology. It’s a competitive area, and many people are in the RNA space now. Talent is somewhat of a roadblock.
On the technical side, being the first to examine this biology makes it a challenge. When presenting our value proposition to investors or potential partners, we’re constantly asked where the benchmark is on drugging lncRNAs. These molecules have not been targeted before, so there may be many unknowns.
- Ounzain S, Pezzuto I, Micheletti R, et al. Functional Importance of Cardiac Enhancer-Associated Noncoding RNAs in Heart Development and Disease. J Mol Cell Cardiol 2014;76:55–70; doi: 10.1016/j.yjmcc.2014.08.009.
- Ounzain S, Micheletti R, Beckmann T, et al. Genome-Wide Profiling of the Cardiac Transcriptome after Myocardial Infarction Identifies Novel Heart-Specific Long Non-Coding RNAs. Eur Heart J 2015;36(6):353–368; doi: 10.1093/eurheartj/ehu180.