Size Matters: Using oligonucleotide siRNAs for Targeted Therapeutics

Coming to you LIVE from the 3rd annual RNA Symposium: Advancing RNA Bioscience into Medicine. Follow us on Twitter or the tag #umichrna!

Live blogger: Sarah Kearns. Editor: Whit Froehlich.


Neurodegenerative diseases and genetic conditions lack effective treatments. Patients with disorders like Huntington’s disease (HD) and congenital amyotrophic lateral sclerosis (ALS) thus have unmet medical needs. To begin to get to the heart of these disorders, researchers like Dr. Anastasia Khvorova, a professor at UMass Medical School, are looking for strategies to target RNA in order to develop treatments.

These strategies involve oligonucleotides – small DNA molecules that bind to mRNAs – that can prevent the mRNA from producing that encoded protein. Oligonucleotides by themselves have little clinical relevance because they target their specific mRNA sequences wherever they arise in the body. For diseases that mainly affect certain areas, localization is required for effective treatments. Appending chemical additions or modifications to the backbone of the oligonucleotides allows them to have appropriate distribution and specificity for their target.

Therapeutic oligonucleotides could be the sequence-specific drugs certain genetically determines disorders like HD or congenital ALS need for treatment, and some are in clinical trials. Dr. Khvorova is developing novel oligonucleotide treatments that are safe, durable, and can be widely delivered to the brain and spinal cord.

She is uniquely capable of doing this research and making these oligonucleotides as the founder of the UMass Nucleic Acid Chemistry Core. This is the only nonprofit facility in North America capable of synthesis of modified oligonucleotide in the quantity needed to perform these experiments. She’s also the named inventor on over 150 patents and 200 patent applications, and is defining the field of RNA drug design and development.

Returning to smaller forms of RNA, its unique chemical structure allows it to be a modifiable chemical backbone for drug development. “There is enormous excitement when something is realized” she claims, showing a slide of the Technology Hype Curve showing the market size (and optimism) of using siRNA in clinical contexts. The hype comes from the concept that drugs, instead of being a daily pill, could be administered only once a year as an injectable RNA. What Dr. Khvorova wants to understand is how to effectively use RNA as a treatment and direct its localization to have the desired minimally invasive therapeutics.

We’re now in the plateau of productivity, but many advances continue to be made! Image source.

“Chemistry matters,” she insists, showing the many different types of modifications that are necessary for the stabilization of siRNA. The ribose portion of siRNA in particular needs to be modified so it does not become degraded within the cell. Terminal phosphates, too, will be cleaved instantaneously without chemical protection, where the phosphate is necessary for proper recognition. Modified siRNAs, compared to those that are unmodified, have significantly more stability and long-term tissue retention.

The ribose and phosphate of RNA need to be protected for an effective biologic. Image source.

Trying to address pharmacokinetic properties of the siRNA therapeutic, they turn to co-administering with lipids. Constructing a library of lipophilic conjugates and administering these conjugates with the modified siRNA therapeutics, they found there is better distribution across the body in a mouse model. Looking at two different targets, Huntingtin and cyclophilin B, there is targeted, functional, and selective delivery. This is not contradictory because, while the biologic becomes widely distributed in the body, it is only functional in a certain tissue type where there is aberrant pathophysiology.  

One focus is preeclampsia (PE), the leading cause of premature births, which is caused by oversecretion by the placenta factor sFlt-1; because it has a significant unmet medical need. Levels of sFLT1 mRNA are significantly higher in the placenta than in any other tissue type allowing for an siRNA therapeutic to be a viable option. There are two isoforms of sFlt-1, a short and long construct, where the short form is only found in the placenta and the long form in the mother’s liver and kidney.

A schematic of the role of sFlt-1 in preeclampsia. Image source.

Using the facility of the UMass Nucleic Acid Core, they are able to synthesize and validate the siRNAs for their studies. They find that there is selective sFtl-1 downregulation that does alleviate the hypertension associated with preeclampsia.

Testing siRNA delivery on CNS-related disorders yielded a surprising result. She recounts excitedly alerting her collaborator after the first experiment: “Look, a pink brain!” Dr. Khvorova shows in this pink-colored brain that the siRNA was able to cross the blood-brain barrier, a significant challenge in drug development. Specifically targeting Huntington’s Disease in sheep, she showed that there was sustained modulation of Huntingtin protein expression. With only 2ng of drug per mg of tissue, there can be sustained action and effectiveness of the therapeutic.

Doing the same test in non-human primates, she finishes her talk by showing a very pink brain, with the hope that the siRNA therapy could soon be used in human clinical trials to treat Huntington’s Disease.

Stressed out about RNA Granules; Roy Parker

Coming to you LIVE from the 3rd annual RNA Symposium: Advancing RNA Bioscience into Medicine. Follow us on Twitter or the tag #umichrna!

Live blogger: Whit Froehlich. Editor: Sarah Kearns.

Roy Parker, Ph.D., is a Professor at the University of Colorado and Investigator at the Howard Hughes Medical Institute who studies the regulation of translation and degradation of messenger RNA (mRNA) using yeast as a model organism. Degradation of mRNA is accomplished in coordination with the removal of the “poly-A tail” at its end, which precedes degradation from either end of the molecule. His other work includes investigation of the processes around mRNA decapping and storage in P-bodies. He is speaking today about RNA granules, and generally about aggregated RNA structures, as well as some of their roles in disease.

Firstly, RNA granules, aggregations of protein and RNA, come in a number of different forms including Cajal bodies in the nucleus, and P-bodies and stress granules in the cytoplasm. Stress granules are assemblies of non-translating mRNAs with a number of the enzymes normally involved in translation and processing of proteins. These granules have a role in transitions between different states of mRNAs, assembly of organelles, and disease processes.

RNA granules aggregate into P-bodies (pink) and stress granules (green) in response to cellular stress and in diseased states. Image source.

One example is dendrites of neurons, where mRNA is first stored in granules and then later translated in certain circumstances. This process is important for synaptic plasticity and habit formation. An experiment around this function involved observing first that flies become habituated to an unpleasant scent, and then removing the function of Atx2, an important regulator protein. This removed their ability to habituate, and so they no longer would eventually ignore the unpleasant odor. Furthermore, they no longer accumulated the RNA granules seen in the normal flies suggesting the importance of the granules in habit formation. The genetic domains were identified which coded for the granule proteins, where the deletion of just these domains had the same effect on fly habituation, by inhibiting it.

Granules also play a role in development, regeneration, stress responses, and neurodegenerative disease. It has been observed that mutations causing neurodegenerative disease often alter the presence of RNA stress granules. It turns out that they code for RNA binding proteins. Likewise, genes coding for complexes that disassemble stress granules, when mutated, can correlate with the development of neurodegenerative disease.

Stress granules have been purified, and the proteins in them identified. Less is known about their RNA content, however. In analyzing the RNAs that appear in stress granules, there is a continuum in the range of genes that are expressed at higher, lower, or similar levels to elsewhere in the cell. This invited the suggestion that perhaps the sequencing was inaccurate; but single molecule validation performed on stress granules confirmed correct localization. So, back to the stress granules: it turns out that only about 10-15% of mRNAs appear in stress granules, but none particularly predominate, although there are a few genes for which most of their transcribed mRNA ends up in stress granules.


Looking at the different types of RNA, longer mRNA sequences only 10% of mRNA ends up in stress granules with longer, poorly translated sequences forming the bulk of the granule. Image Source.


The genes that are related to the mRNA within granules have two identified properties: poor translation and length (~7.5kb). As to the latter, it was considered that perhaps length itself, and the correspondingly large amount of protein-binding site, was the predisposing factor. However, enrichment with binding protein found that this was not actually important. On the other hand, a longer mRNA also has a greater opportunity to bind with itself. An experiment performed with roughly cell-like quantities of RNA in the absence of granule proteins found that they still become entangled nonetheless. Analysis of the RNAs that tended to self-bind discovered that these were the same RNAs that assemble in stress granules. The suggestion from these various findings is that protein-protein interactions and RNA-RNA interactions, rather than protein-RNA interactions, drive stress granule formation.

Going back to proteins, it is found that while stress granules can form with either mostly RNA or mostly protein, the formation is still limited by inhibition of either. Notably, RNA tend to be more or less likely to self-aggregate depending on what type of RNA they are: tRNAs are not observed to clump together, while longer strands encoding genetic sequences with many repeats will. In general, the more structured an RNA molecule is in its functional state, the less likely it is to self-aggregate.

Other aspects of the environment affect the tendency of RNA to form granules: crowding, length, and the absence of RNA binding proteins all make granule formation more likely. Dr. Parker even suggests that, without protein complexes, the lowest-energy state of RNA might still be in forms that promote aggregation. Based off of that, another factor that affects granule formation is the presence of polyamines (amino acid clusters of arginine and glycine or proline), which drive RNA self-assembly. These polyamines, because of their positive charge, interact with the negative charges on RNA and prevent RNA granule formation. The effect of polyamines is paralleled by the deletion of the domains found to encode granule proteins playing a role in habituation; deletion of the same domains also inhibited necrosis of fly eyes caused by a similar process.

Through Space and Time: Monitoring RNA Translation; Jonathan Weissman

Coming to you LIVE from the 3rd annual RNA Symposium: Advancing RNA Bioscience into Medicine. Follow us on Twitter or the tag #umichrna!

Live blogger: Sarah Kearns. Editor: Whit Froehlich.

Suppose you have some extremely important information in the form of a blueprint and it’s your job to protect it. It’s not just a blueprint for a top-secret location – it’s the blueprint to life; it specifies how every cell in the body should function.

Needless to say, you have to keep this information in a secure spot. This is exactly what eukaryotic cells do with their DNA by storing it in the nucleus, a membrane-enclosed compartment. The genetic material itself never leaves the nucleus; instead it’s transcribed (essentially making a copy through the genetic pairing process) as messenger RNA (mRNA) molecules that leave the nucleus. From there, to be translated into a protein product, mRNA has to go to a molecular machine called the ribosome. This molecular sandwich, also made of RNA, “reads” the information on the mRNA and translates it into the amino acid sequence that it spells out, making proteins.

The ribosome sandwich, made of large (orange) and small (white) subunits, feeds through the mRNA sequence (purple) matching amino acids with the mRNA sequence (using tRNA, green). This results in a growing protein chain. Image source.

An understanding of how the ribosome works is important for a scientific understanding of life due to its role in protein production. As the RNA instructs the ribosome to assemble amino acids in the proper order, the new polypeptide – or protein – starts to fold into a compact blob. This final blob shape is important because a protein’s function is significantly determined by its structure. As such, if the protein does not fold into the correct shape, then it cannot do what it’s supposed to do in the cell. Many disease states, including sickle cell anemia, Parkinson’s Disease, and Alzheimer’s Disease, are related to protein misfolding.

Dr. Jonathan Weissman, a professor of cellular and molecular pharmacology at the University of California, San Francisco, has helped develop a key technique called ribosome profiling that monitors protein translation. Pairing together this large-scale approach with smaller mechanistic investigations, Dr. Weissman looks at the role of the ribosome in aiding correct protein folding.

The ribosome’s work of translation is itself a very dynamic process. So far, structures obtained from x-ray crystallography and cryo-EM allow for a reconstruction of a simulated video, but the dream would be to actually visualize this process in the cell. A step closer is getting the full time course of translation, which Dr. Weissman has been working on through ribosome profiling.

Ribosome profiling globally profiles protein translation. Unlike typical analysis of a lifetime of behaviors, this technique takes a snapshot of all of the ribosomes within a cell at a particular time. The mRNA sequences that are protected by the ribosome from the processing process are counted, yielding a measurement of the mRNA sequences being translated when the analysis is performed. This technique can further monitor when other proteins are interacting with the ribosome or mRNA.

Image source.

The precision achieved by combining billions of these footprinting experiments can tell us how much RNA and protein are being synthesized and where along the known gene sequence that transcription is occurring. The most abundant proteins, he finds, have yet to be marked on the genetic code, suggesting a significant number of unidentified open reading frames within the genome.

Ribosome profiling as well more precisely reveals the rate of protein expression, which is currently measured by counting mRNAs. Taking F-ATPase as an example, because there are different subunits in different amounts that create the protein, he finds that the ribosome density along the mRNA correlates with the ratios of the different domains in the final protein. This makes sense, as it results in efficient production of the protein.

The next question is: where is translation occurring? Because the cell is relatively large, at least compared to a single ribosome, to make sure the protein synthesis is occuring in the right place, the mRNA might be brought to the correct spot within the cell before it’s translated.

By adding a tag specifically to the endoplasmic reticulum (ER), they can look at the specific ribosome profile occuring there. Proteins that need to travel to the membrane need to be bundled and sent there through signalling done by the ER. Comparing the ER-ribosome footprints with the overall ribosome footprints, they find that proteins destined for secretion through the membrane, or installation in the membrane, are translated at the ER. So this tag works to identify known and localized mRNA sequences.

Image source.

To test the same technique with other organelles, they add the tag to mitochondria instead of the ER, but run into some complications (what would science be without its annoying barriers?). This is because there is a protein complex (ERMES) that tethers the mitochondria to the ER. Nevertheless, they were able to tease apart the ribosome profiles by enriching the cell for either mitochondria or ER. Adding this tag to other organelles, proteins, and other factors will allow discovery of the spatiotemporal map of translation across the whole cell!

But the ER has a tough job! Not only is it responsible for trafficking membrane proteins, it has roles in protein synthesis and stabilization as well. With proteins having variable lengths, amino acid compositions, and electrochemical properties, there are many different routes in protein trafficking and protein folding, each of which can malfunction. The resulting misfolded or misplaced proteins can causes diseases including cystic fibrosis, hypercholesterolemia, and many others.

Chaperone proteins help with both of these problems (localization and folding). By interacting with misfolded proteins, they pull on tangled polypeptide strands, not unlike how we untangle pocketed headphones, to get the protein into a correct conformation. Because there are so many different types of proteins, chaperones are localized at the final destination of the protein. As such, membrane protein chaperones, like EMC, are at the membrane surface.

By using the tagging technology previously discussed, Dr. Weissman and his team added a tag onto the EMC to find out what ribosomal substrates were being synthesised. It turns out there’s a sequence-specificity for transmembrane domains and hydrophobic patches. These interactions of the EMC and the ribosome occur during translation to ensure that there are chaperones at the site of protein synthesis, avoiding any downtime prior to its proper folding. This suggests that the ribosome, too, has to be localized.

By putting together the results from ribosome profiling and tagging, a map of where transcription occurs is made, raising questions of how these multisubunit complexes are being transported in the crowded cell.

Genetics of Colorectal Cancer; Eric Fearon

Coming to you LIVE from the 3rd annual RNA Symposium: Advancing RNA Bioscience into Medicine. Follow us on Twitter or the tag #umichrna!

Live blogger: Sarah Kearns. Editor: Whit Froehlich.

Most of the work done in your cells is done by complex molecules called proteins. Made up of long chains of amino acids, they are required for the structure, function, and regulation of your body’s tissues and organs. The design of these proteins comes from DNA, the genetic code of life. But the pathway from double helix DNA to protein, as it turns out, is a complicated one.

The Central Dogma of molecular biology describes the process as two steps — DNA is transcribed into RNA, and RNA is translated into proteins. While accurate, this model glosses over the regulation done by the various different types of RNA molecules. One such regulator is microRNAs (miRNAs) which are small RNA molecules that regulate gene expression by interacting with messenger RNA (mRNA), the product of transcription.

Particularly in cancers, miRNAs have been found to bind to mRNA before the mRNA has a chance to be translated into key gene regulator proteins. But as miRNAs play a significant role in the progression of cancer, they can also serve as a tool for prognosis.  Dr. Eric Fearon, the director of the UM Comprehensive Cancer Center (which recently received an impressive $150M donation), is looking at how miRNA can be used for early detection and potential treatment of cancer, focusing on colorectal cancer.

The Keynote

Diving into the genetics of cancer, Dr. Fearson introduces Apc which is involved in the development of many cancers by activating a kinase signaling pathway. Mutations in Apc leading to a loss of its N-terminus lead to aberrant activation within cancerous cells. Where the cell’s defense system usually catches incorrect protein, mutant proteins within tumorigenesis can evade the immune system, which further aids in resistance to medical cancer therapies. As Dr. Fearon notes, “if you’ve seen one cancer, you probably haven’t seen one cancer,” because there are so many aberrant mutations that lead to subtle differences in genetics even within one tumor.

Within colorectal cancers (CRCs) in particular, there are specific recurrent somatically mutated genes that typically occur within nucleotide repeats, a group to which the Apc gene belongs. Unlike other cancers, there are many mutations that occur throughout the genome that further drive the gain-of-function of the Ras pathway. These collectively lead to the “big bang” model of tumor growth.

The Apc protein. Image source

The Apc protein, responsible for tumor suppression, huge for a protein — just under 300 kD. Towards the end of the Apc gene reading frame, the portion of the gene that codes for the protein, there’s a cluster of methionine amino acid (Met) codons towards the C-terminus of the protein. Met also functions as the “start” codon for protein translation so the cluster of them can “confuse” translation to have multiple truncations. The loss of the tail of the protein is a problem because it includes key domains for its function and regulation.

Within the early development of colon cancers, the normal epithelial cells become crypts. Targeting the Apc genes within this cell-type population results in promotion of cell growth and proliferation. But what mechanisms control this?

Looking at a particular signaling pathway (Wnt), β-catenin is an important component of gene regulation of oncogenic transcripts. In the presence of the Wnt signaling molecule, β-catenin gets into the nucleus where it can act as a transcription factor. Without Wnt, Apc prevents β-catenin from becoming a transcription factor by recruiting a proteasome, a complex that breaks down the protein. This is important because Dr. Fearson shows that the knockdown of β-catenin prevents tumorigenesis in mice colons. By turning off and on patterns of expression in Apc-mutant crypts, he further demonstrates the destruction in crypt formation due to proper tumor suppressor function.

Frequency of gene mutations in colorectal cancer. Image source

Just like other cancers, colorectal cancers are impacted by alterations in p53 tumor suppressor protein signaling. Dominant negative mutation of the p53 gene is the component that leads to tumorigenesis, and not just the addition of Apc and KRAS. Using CRISPR-Cas9, Dr. Fearon wants to make an accurate mouse model that incorporates all of the mutations that are known to be important in colorectal cancers. By making the transgenic mice with this gene editing tool, his research team can tease out the different interactions of independent proteins.  

He concludes his very protein-signaling-based keynote with the hope that better understanding of RNA will aid in discovering the interactions and pathology of the complex progression of cancers. Alluding to his RNA tool development, he suggests that they could also be used for early cancer prognosis.

Modern(a) developments in mRNA theraputics; Melissa Moore

Coming to you LIVE from the 3rd annual RNA Symposium: Advancing RNA Bioscience into Medicine. Follow us on Twitter or the tag #umichrna!

Live blogger: Whit Froehlich. Editor: Sarah Kearns.

Melissa Moore, Ph.D., is currently the Chief Science Officer at Moderna Therapeutics, having previously been on the faculty at the University of Massachusetts as Professor of Biochemistry & Molecular Pharmacology and Eleanor Eustis Farrington Chair in Cancer Research with a concurrent appointment as Investigator at the Howard Hughes Medical Institute. Her work ranges widely in RNA, currently focusing on pre-mRNA processing for medicinal applications.

She starts with a few brief examples of drugs developed over the past 80 years, noting that biologics, therapeutics that have a biological structure, have grown to account for $180 of $600 billion spending on medications. These contrast with traditional (“small-molecule”) drugs, which contain within a single small molecule both a dianophore (localizing portion) and pharmacophore (active portion), by including more distinct and larger such portions. This increases the complexity and challenge of development. And drug development currently takes 12 years (and $2.5 billion)!

Differences between small-molecule and biologics as drug therapies. Image Source.

Each portion of the central dogma has been represented by biologic therapies: 400 protein therapies, 40 RNA therapies, and 5 DNA therapies. But that’s not very much – there’s a lot more to come!

The use of mRNA as a therapy was described back in the early 1990s and is used to target diseases before protein production. As such they are considered less dangerous than editing at a gene level. Further, they might be tolerated better by the human body than attempts to use recombinant protein therapies. However, there are some issues in administering RNA therapies.

Typically therapeutics are most effective if they are delivered directly, but mRNA is really big! At 1500-3000 bp, it has 10x the weight of the protein it encodes. This makes it really hard to get it where we’d want it in the cell. Additionally, the body has defense mechanisms as a part of its immune system. Key elements of the innate immune system are the receptors TLRs and RLRs that have evolved to work against exogenous mRNAs, usually encountered by the body in the form of viruses. Finally, you can’t just get the mRNA in – you’ve also got to get the ribosome, the complex that takes mRNA and transcribes it into protein, on board with translating the mRNA into the protein with the final intended effect.

A lot of design has to go into the mRNA therapies to avoid the immune system and making sure there’s enough protein product. Image Source.

Moderna and other companies have been working on these challenges because the potential is large. One part of this work is in siRNA with modified bases, which has been demonstrated to evade TLRs, but (with one exception) it hasn’t been able to engage the ribosome effectively. Now, sometimes it works to just inject “naked” mRNA into muscle with resulting protein translation – an upcoming drug in the works between AstraZeneca and Moderna is a formulation of VEG-F mRNA for intracardiac application in myocardial infarction patients (data from pigs are shown; currently in phase 2 clinical trials). In many cases, however, this technique does not sufficiently address the size problem.

For other cell types, and for systemic delivery, mRNA needs a delivery mechanism. The mechanism currently employed is a lipid delivery system composed of lipid nanoparticles forming an amino-lipid envelope. One such example of this system has been developed at Alnylam using MC3 to deliver other types of RNA treatments. Moderna is building on these efforts because mRNA has a much greater need for these encapsulations than siRNA. Dr. Moore shows that using Moderna’s lipid nanoparticles, mRNA tagged with a fluorescent tag is localized to the liver in a mouse model, as demonstrated by a bright glow in their livers. This shows promise for the application of specific and localized therapeutics.

Vaccine development, in particular, is a focus of her research because viruses are known to hijack the transcription system. But vaccine development takes quite a long time because we’re fighting the pace of evolution with slower tools, especially with influenza. If we were able to base a vaccine directly on the genetic code of the target organism, we could more rapidly produce an appropriate vaccine. Hence mRNA is a promising approach for a more responsive flu vaccine.

Another disease presenting an even more daunting challenge is CMV, a disease that currently has no vaccine or treatment. In particular, the CMV virus has a pentamer that hides behind a similar sequence, and so has thus far resisted efforts to develop a protein-based vaccine. One more disease currently being targeted is Zika virus, which demonstrates a rare case of acceleration of the drug timeline – the first construct was ordered in December 2015, and the first vaccine was in humans in December 2016.

Oh, and Moderna is big! Now over 600 employees. Big advances and one of the top medical research companies, now hiring!