I’m “at” a virtual meeting on the novel Coronavirus, SARS-Cov-2, and the disease it causes COVID-19. A meeting like this would normally be held on campus here at CSHL, and with months and months of prep. But they were able to bring together leading scientists from around the world to meet virtually and it’s a great privilege and tremendous opportunity to be able to listen in. As I learn the up-to-date-est details, I thought it would be a good time to review the molecular biology and biochemistry of this virus, why it’s trickier to treat than bacterial infections, and some of its Achilles’ heels scientists and doctors are working on exploiting.
When you have an infection that’s caused by bacteria, doctors can prescribe drugs called antibiotics that target bacterial Achilles’ heels. But when an infection is caused by a virus, there’s less doctors can do to specifically treat the infection – instead they normally rely on providing supportive measures to test the symptoms – things like fever reducers if you have a high fever or oxygen if you’re having trouble breathing.
It’s harder to target the actual viruses in part because there’s really not much to them – just their genetic blueprint (viral genome) with some packaging. Different viruses have different compositions, but SARS-Cov-2 is a single-stranded RNA virus of the Coronaviridae family. Instead of double-stranded DNA like we have, it holds its genetic info in a single strand of RNA, wrapped up around nucleocapsid (N) proteins to form a compact, protected “nucleocapsid” and enveloped by a membrane coat embedded with proteins, with Spike proteins jutting out to help it bind to receptors on cells to sneak inside where it can hijack the cell to make lots of copies of itself (replicate), break out, and then infect new cells. Speaking of sneaky, despite the havoc they can wreak, viruses typically aren’t even considered “living” because they can’t really do anything on their own, instead relying on host cells for basic functioning.
Bacteria, on the other hand, may be small, but they are whole, living, organisms. Single-celled ones, yes – but that only ups the pressure on them to make sure that each and every cell makes everything it needs to grow, thrive, and divide – and each of those daughter cells has everything it needs. So it has to make, and hold genetic instructions for, a lot more things, which offers a lot more potential drug targets.
To put things into perspective, the SARS-Cov-2 genome is only about 30,000 letters long, the bacteria E. Coli has a genome that is about 4.6 million letters long and our own genome is about 3 billion letters long. In order to keep its genome so small, the virus uses its host cells’ machinery whenever possible, and, for the machinery it needs that we can’t provide, it uses clever space-saving tricks to write the instructions in as few letters as possible. This allows it to more easily carry out its non-life’s purpose: making lots more copies of its genome quickly, packaging them up in small packages, and shipping them out to infect new cells. Since all that’s taking place inside a cell, the virus, including its genome, has to be really small.
Bacterial genomes don’t have this size restriction, so they can have more genetic instructions and make more things. This includes making some structures we don’t have. For example, since it’s not like it has skin to protect itself, each bacterial cell needs to be hardy – so they have thicker cell coatings than our cells. In addition to the fatty membranes like we have, bacteria have “cell walls” made up of sugars and peptides (short strands of protein letters). Since our cells don’t have such cell walls, they and the enzymes responsible for constructing them, are great targets for antibiotics. This, for example, is how penicillin, ampicillin, and related β-glycan antibiotics work.
But viruses don’t have these cell walls – and, from a cell’s perspective they look a lot like a neighboring cell because when the virus buds out of its host cell, it gets coated in the host cell’s membrane coat – told you they were sneaky!
So let’s look elsewhere… other antibiotics target other pieces of bacterial machinery like bacterial ribosomes – ribosomes are protein-making complexes made up of proteins and ribosomal RNAs. Before a protein gets made, a messenger RNA (mRNA) copy of its gene recipe gets made. And the ribosomes travel along this mRNA, following the mRNA’s instructions on which amino acid (protein letter) to add in which order (3 RNA letter “words” called codons “spell” one amino acid and a helper molecule called tRNA brings that letter to the ribosome to link up, using a complementary “anticodon” to provide specificity).
We have ribosomes too, but our ribosomes are different enough from bacterial ones that we can use drugs to target the bacterial ones while not hurting our own. An example of this type of antibiotic is kanamycin.
Viruses, too, rely on ribosomes to make their viral proteins, but they use their host cells’ ribosomes, so there aren’t “viral ribosomes” to target – instead, if we targeted viral protein writing, we’d target our own protein writing.
I said “writing” there and not “making” because, although the peptide “writing” machinery is the same, there’s an important distinction between viral protein making and our protein making. The “writing” process of ribosomes using mRNA to link up amino acids into a long “polypeptide” is called translation. And that’s the same for us and for viruses. For most of our proteins, the polypeptide that results from translation is a single protein (it gets folded and might get a few post-translational modifications, but, it’s ONE protein.) But viruses, with that constraint to keep their genome super small and packable, often write multiple proteins in one long “polypeptide” that they then cut (using enzymes called endoproteases) into several proteins.
For example, SARS-Cov-2’s biggest gene is a “replicase” gene which contains instructions for making proteins the virus needs to replicate. And, in another of its space-saving tricks, the virus can make 2 different polyproteins from this one gene – pp1a and pp1ab. It’s able to do this because it uses a “slippery sequence” and a pseudo knot (a stretch of the RNA folds into a “roadblock”) to stall the ribosome on its translation journey and get it to slip up, change reading frame (e.g read “aredcat” as “are dca t” instead of “a red cat”), and keep going past the rep1a stop sign to make pp1ab.
As a result, pp1ab has the same stuff pp1a but it also has extra protein instructions – pp1a has nsps 1-11 and pp1ab has nsps 1-16. Nsps stands for Nonnstructural Proteins and they’re called that to distinguish them from the virus’s structural proteins (including the spike protein, S; envelope protein, E; membrane protein, M; and nucleocapsid protein, N) – which are important for packaging the viral genome and getting it to new cells. These structural proteins (and a few others) have their own individual genes downstream of the replicase gene.
Nsps include things like enzymes (reaction mediators/speed-uppers) for helping copy the RNA and processing the polypeptides, which need to be chopped up into their individual proteins. Within the polyproteins are 2 proteases, with different sequence specificities, which cut the polyproteins into their individual nsps. A Papain-like protease (PLpro) is encoded within nsp3 and a serine type “main protease” (Mpro) is encoded by nsp5. PLpro is responsible for the first few cuts (separating nsp1/2, 2/3, and 3/4) and Mpro handles the rest.
Since viruses need proteases to make functional proteins from polyproteins, some antivirals are protease inhibitors – including the HIV treatment combo Lopinavir/ritonavir (trademark Kaletra). Early on, scientists had hoped that Kaletra also might help treat covid19 but the results of a small trial found it ineffective and side-effect-causing.
So that was disappointing news. But not that surprising since viral proteases are specific and different from one another. So scientists are working to custom design inhibitors for SARS-CoV-2’s main proteases, literally! As part of a “COVID Moonshot” project, Diamond Light Source in the UK collected thousands of x-ray crystallography “3D pictures” of Mpro bound to different drug fragments from a compound library. These fragments are like pieces of potential drugs that can be modified and/or joined together to make “full-size” drugs using information about how and where they bind. That fragment-based design takes a lot of work, so they crowdsourced it and scientists from around the world answered the call, submitting lots and lots of potential drugs. The project chose the most promising to synthesize and test and you can learn more about this ongoing project here: https://bit.ly/mproinhibitors
There’s been some talk about trying to inhibit the human proteases that aid SARS-CoV-2 in its cellular entry – mainly TMPRSS2, but also cathepsin & furin. These human proteases do NOT cut the virus’ polyprotein – the virus is on its own for that. But these human proteases DO cut the SARS-CoV-2 Spike protein, allowing it to undergo a dramatic shape-shift (conformational change) that allows it to shoot out its inner protein part, latch onto the host membrane, and pull it so close to the viral membrane that they merge. https://bit.ly/coronavirusspike
If the Spike protein doesn’t get cut, it can’t do that gymnastics. So inhibiting those proteases could inhibit the virus. But we need those proteases too! TMPRSS2 is less widely-needed than the other ones (which is likely part of the reason certain cell types are more susceptible to invasion) so it would be more potentially possible to target it than to target cathepsin or furin. But it still would be a much better idea to target something that the virus needs but we do not. So that we only hurt the virus and not the human.
One such class of “I-only-attack-viruses” treatments are RNA-Dependent RNA Polymerase (RdRP) inhibitors, such as remdesivir, which acts as an adenosine analog (it mimics the RNA letter A) to mess up viral replication (genome copying).
SARS-Cov-2’s small size makes copying less time-consuming, but the fact that it’s written in RNA, not DNA, means the virus needs to provide its own copier, since our cells don’t have an RNA to RNA copier (an RNA dependent RNA polymerase (RdRP)). So one of the nsps (nsp12) is an RdRP. Since we don’t have one of these it offers an attractive target.
When you go to target an enzyme, and you really want to put a stake in its heart, you go to that heart – the “active site,” where reacting molecules come together and get changed. For an RNA polymerase, this active site is where the RNA letters bind to get linked. So if we want to mess things up we can use something that looks like an RNA letter (a nucleotide analog) but is modified in some way that either cripples the enzyme or screws up the product.
But there’s a problem. The copier may be different, but the letters it’s using are the same – it’s using the same RNA letters our cells use, just linking them in the order that it wants based on its RNA template. Instead of what our cells do when writing RNA – linking letters in the order we want based on a DNA template (a process called transcription). So, when we target RdRP we have to be sure we’re only targeting RdRP and not our own DNA-dependent RNA polymerase. Thankfully, our own polymerase isn’t as easily fooled so the drug remdesivir doesn’t interfere with them.
But how exactly does it interfere with the viral polymerase? Laurel Oldach wrote a really great article for ASBMB about remdesivir – and it’s a really cool format – you can click on various parts of the molecule to see how the different “functional groups” give the drug different chemical “superpowers”. http://bit.ly/2Wttg76
I encourage you to check out her article to learn more, and I also wrote a post on it: https://bit.ly/remdesivirbiochemistry
but here’s an overview. RNA letters (ribonucleotides) have 3 main parts – a ribose sugar linked up to phosphate groups (at the “5’ position”) and 1 of 4 unique nitrogenous bases at the 1’ position (A, U, C, or G). At the 3’ position on the ribose is a hydroxyl (OH) group. Letters link up 5’ to 3’ using the phosphate and hydroxyl groups.
Unlike some nucleotide analogs, which act as obligate chain terminators (add one and you’re done) because they don’t have the 3’ hydroxyl (OH) group needed to latch onto another letter, remdesivir *does* have a 3’OH, so multiple messed up letters can get added on before it causes the chain to terminate, potentially because the slight differences in the fake letter’s shape make the RNA less comfortably held by the polymerase. And having a 3’ OH might allow it to sneak past the viral proofreaders, so the virus doesn’t realize it’s making messed up RNA and try to fix it.
Remdesivir is a prodrug, meaning that our bodies have to tweak it a bit to get the active form. The changes that need to be made for remdesivir involve the 5’ phosphate – nucleotides are added as triphosphates, so three phosphate groups in a row (2 of which get kicked off to pay the energy cost of linking) but remdesivir just has one – and it’s “hidden” by a bulky protective group. Inside our bodies, that group gets cleaved off and the other two phosphates added on by enzymes our cells already have – so that activation is “like no big deal”. But could remdesivir be a big deal?
Early clinical trial data show that it’s likely not a huge deal, but it does provide some modest benefit for some patients. In the trial of 1,063 patients (one of those gold-standard double-blind, randomized, placebo-controlled ones), remdesivir shortened the average hospitalization time from 15 to 11 days. And it lowered the death rate a little, from 11.9% for the placebo to 7.1% in the treated group, but it wasn’t enough to reach statistical significance (i.e. the less death could just be a fluke). https://bit.ly/3c7v6i7
Further trials are still underway, but results from this NIH-funded study, published in the New England Journal of Medicine, https://bit.ly/2X6Axd4 , led the FDA to issue Emergency Use Authorization (EUA), meaning that doctors can give it to hospitalized patients outside of clinical trials https://bit.ly/2M567BB
The FDA also recently reversed an EUA it had issued for another drug that got a lotta lotta hype with not a lotta (basically no) data – chloroquine, or its more stable form, hydroxychloroquine. It’s an antimalarial drug that’s been around for over half a century.
Malaria is a parasite, not a virus, but chloroquine also has some anti-viral activity. It is thought to prevent viruses from getting into and/or out of cells by messing with their modes of transportation – when the virus enters the cell, sometimes it does so by the cell kinda pinching in the part of the membrane that the virus had bound to (using that spike protein to latch onto a receptor on the cell surface). So when the virus gets inside the cell, it’s still “quarantined” inside of a membrane-bound pouch called an endosome. Chloroquine messes with the pH in endosomes, which was theorized could prevent the virus from getting out (and it does some other stuff too).
Early on in the pandemic, results were published from a VERY small “open-label” study (patients and doctors knew which patients were being treated with what) which show initial promising results, including when chloroquine was combined with the antibiotic azithromycin (which might help combat secondary bacterial infections that can develop once the patient’s lungs are made vulnerable by the virus) BUT the study was really badly set up, multiple patients dropped out, and they used a really low threshold for success. A great thread about it by Jason Pogue can be found here: https://twitter.com/jpogue1/status/1241138975802359813?s=20
Further trials (including well-designed ones) failed to find any benefit – and these drugs can have SERIOUS side effects so it is NOT a “nothing to lose” situation. I encourage you to check out this Science article by Derek Lowe that lays out the results of studies to date on all these drugs http://bit.ly/3ddpyVg
So the hydroxychloroquine hype seems to finally be abating.
You can’t blame people for hoping. And for wanting something quickly. But the hard reality is that developing new therapeutics is a long process. The reason these drugs were able to go to trial so soon is that they’d already been approved or at least were close to approval for other purposes. So they were found to be “safe” (at least safe enough to be used compared to their benefit for a certain condition) they just needed to get the okay to be reused – sometimes through “compassionate use” exceptions and sometimes through official clinical trials. These official trials are important in order to know if the drug has unpredicted side effects specific to this new condition it’s being used for and thus which weren’t an issue when they were going through testing for their other uses. Additionally, the controlled trials are crucial to figure out if a drug really is more effective than the current standard of care or if patients on it are just getting better because they would have gotten better anyway.
Speaking of patients that have gotten better – the “getting better” involves an adaptive immune response in which the patient’s immune system learns to recognize parts of the virus as foreign (usually parts of the spike proteins in the case of these types of viruses) by developing little proteins called antibodies that bind specifically to that spike protein and trigger an immune response.
Even after the infection is cleared, a low level of those antibodies stick around to “keep watch” – this is why antibody tests can detect if someone had the virus in the past and recovered. And it’s why researchers are giving blood plasma (blood without the blood cells) donated by recovered patients to currently-sick patients. This treatment is called “convalescent plasma” (CP) therapy. It’s showing promise, but it also isn’t very scalable, and it’s VERY batch-to-batch different-y. In addition to antibodies against SARS-CoV-2, the plasma contains antibodies against a bunch of other stuff the donator had encountered. And even the SARS-CoV-2 antibodies differ from donor to donor because the development of antibodies in response to an infection is a random trial-and-error approach of mixing and matching different antibody parts to see what specifically binds a virus, and different peoples’ immune systems find different solutions.
Scientists are hoping that they can isolate some of the best antibodies from the blood of patients who have recovered from covid19, mass produce them, and use them to treat and/or vaccinate patients against covid19 by putting their immune system on watch and/or blocking its access into cells. Antibodies that can do this viral-entry-blocking are called Neutralizing Antibodies (NAbs) and they typically bind the Spike protein, often in a way that prevents the Spike protein from binding to the cellular ACE2 receptor to prevent viral docking.
Antibody-based therapies have been found to be effective against other viruses – Ebola for example – remdesivir was actually somewhat “abandoned” as an Ebola treatment before getting FDA approval after scientists found that antibody-based treatments were more effective. There are currently a number of companies working to produce and test such treatments. This includes Eli Lilly/AbCellera who are testing a monoclonal antibody (single antibody (but lots of copies of it)) and Regeneron is testing an “antibody cocktail” which contains 2 different antibodies – the hope being that even if the virus evolves a mutation that lets it escape one, it will still be “caught” by the second.
Speaking of catching, in addition to treating patients who are already sick, antibody treatments might even be able to prevent people from catching the disease in the first place if taken as a prophylactic. But these antibody-based treatments are a lot more time-consuming and expensive than making small molecule drugs like remdesivir. You can synthesize remdesivir by combining different chemicals under the right conditions (o-chem really is useful!). But “synthesizing” an antibody means making – and purifying – a whole protein – and a lot of it. So don’t expect everybody to just start taking preventative antibodies – but people at high risk might be able to benefit (e.g. if someone’s partner gets COVID-19, they could take antibodies to try to keep them from catching it).
But that’s still a ways off. Any new drug will have to go through studies on safety and efficacy, which can be a lengthy process but it’s really important it be done correctly. In the meantime, many companies are hedging their bets and making lots of the drugs anyway so that, if approval does come, they’ll be ready to provide. Whether they will be provided at a reasonable cost – and to developing nations – that’s a whole ‘other story.
So – bottom line – drugs *are* in the process of being developed and tested – including some I haven’t mentioned, such as famotidine (pepcid), which a couple CSHL doctors are actually part of and which looks like it might help based on some observational data but no one really knows why… There’s currently a real clinical trial underway in New York so stay tuned https://bit.ly/37wQJYy
One theory is that famotidine have is modulating the patients’ immune response to prevent the kind of “overreaction” called a cytokine storm that can lead to some of the most serious problems. Other known immune modulators are also in trials, including steroids like Dexamethasone which was found to reduce mortality in really sick patients (but NOT in mildly infected people – so you shouldn’t go try to buy it). https://bbc.in/3fHCmUr
But immune system modulators are these are tricky because 1) the immune system itself if tricky and 2) there’s a fine line between under-reacting and over-reacting. So doctors need to know when they need to ramp up the response and when they need to tamp it down. And then prescribe things accordingly. And to know this we need more research. This research includes clinical trials as well as trying to see if there are certain “biomarkers,” such as elevated levels of various signaling molecules in the blood in patients that are more likely to respond to various treatments.
One of the hopes is that, if people continue to practice social distancing, in addition to spreading cases out over time so that we don’t overburden the health system, we can buy some time to find effective therapeutics. So, if you’re finally getting out of shutdown like I am, enjoy it – yes, but responsively!
more on topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0