There are a lot of living things that are quite good at killing humans. Tigers, anthrax, lions, cows, bears, and other people do away with thousands of us each year.
There are a few non-living things that are also quite good at offing us. Good old water manages to take quite a few. In good years, we don’t lose anyone to the nerve gasses sarin or VX (Unfortunately, the last few years haven’t been good ones in that regard).
What about those liminal critters though? Viruses and prions aren’t really alive in the traditional sense. They can replicate, they can even evolve, but they lack the hallmarks of life, foremost among them the ability to reproduce. Both of them find ways to hijack the machinery of living organisms and use them for their own ends.
These self-replicating patterns and their potential to wipe us out are the subject of this blog post.
This blog post grew out of a conversation with my friend Malcolm Ocean. We were discussing all the ways the world could end and our conversation drifted towards the biological. We started with proteins (where my background is) and moved on from there towards viruses (where I make a whole bunch of assertions; if you’re a virologist, please correct me).
Can you engineer a protein to wipe out humanity?
When Malcolm posed this question, my first thought went to Ricin, the incredibly deadly protein poison. A favourite of communist assassins, less than 2mg of Ricin can kill an adult human. But while Ricin is active when inhaled, it’s not the easiest to disperse. Nerve gasses like VX are far more deadly and much easier to deliver and disperse to boot.
For a protein to have any advantage over these tailor made weapons, it would need the ability to self-replicate and jump from person to person as it kills them. Anything short of that and you may as well use something else. When it comes to death, we live in an age that’s a grim parody of a cell phone advertisement – whatever you desire, there is almost certainly already a weapon for that.
Unfortunately, there exist self-replicating protein patterns. They are called prions and they’re still poorly understood.
The first prions were probably caused by genetic mutations. These mutations still exist – they’re the cause of diseases like fatal familial insomnia (so named because it is passed down in families, it causes its sufferers to lose the ability to sleep, and because it is invariably fatal).
Prions are mis-folded proteins that somehow catalyze other proteins mis-folding in the same way. This leads to aggregates of proteins. Many neurodegenerative diseases have an element of protein aggregation in them – it’s been implicated in both ALS and Alzheimer’s, for example.
But what makes prion disease unique is their transmissibility. If someone else’s prions get into your brain, they’ll cause the same aggregation process to occur. Suddenly, you have a prion.
We hear about these cases on occasion. Mad Cow (or Creutzfeldt–Jakob Disease) is a prion disease. In cows, it’s known as Bovine Spongiform Encephalopathy (BSE), while it takes on the Creutzfeldt–Jakob Disease (CJD) moniker in humans. Like FFI, it’s invariably fatal. Outbreaks of BSE/CJD occurred because factory farming can involve feeding dead cows to living cows. When cows eat BSE tainted beef, they too contract BSE. One cow with BSE can end up infecting many other cows if it enters the feed supply.
If humans consume any part of the brain of an infected cow, the prion can end up in us. Evidentially, there is enough similarity between the underlying protein for the cow prion to catalyze the formation of aggregates in human hosts.
We’ve gotten better at not feeding cows with symptoms of BSE to other cows, which has cut down on the incidences of CJD. In the interim, it killed about 200 people.
It is maybe within the realm of possibility that you could make a prion more transmissible than BSI. Currently the only method of human-human prion transmission is cannibalism (more on that in a moment). This is ill suited to wiping us all out, because cannibalism is uncommon .
Perhaps some villainous biochemist could design a prion that gets out of the brain, enters systemic circulation, irritates the lungs, and gets coughed into other people’s faces, where it goes up their nose and (maybe) right past the blood-brain barrier. This is the best I can come up with and I’m pretty skeptical (the Wikipedia article points out we’ve yet to successfully transmit anything therapeutically relevant across the blood brain barrier).
But begging the question of plausibility and assuming this could happen, has our biochemist created a civilization destroying plague?
Enter the Fore people. They live in Papua New Guinea and they practiced ritual cannibalism – they ate their dead relatives. Like out BSE infected cows, some of these relatives had prions. In this case, the culprit prion is known as Kuru.
Eating prion laced brains is not a recommended survival strategy. It tends to lead to contracting prions yourself. And indeed, many of the Fore people contracted Kuru from their cannibalism.
Remember how I said that prions are invariably fatal? So is being alive (at least, for now). Sometimes people with prions don’t die of prions because something else kills them first.
Evolution optimizes for reproductive success. In the funeral practices of the Fore, children and the elderly ate the brains. Prions tend to kill people in a few years. A few years really isn’t enough time for young children to beget the next generation.
Imagine the spread of Kuru. The first sufferer dies from it and is consumed. Years later, those who ate her begin to show the same symptoms. They in turn die and are consumed. The infection spreads exponentially, but slowly, first in bursts every few years, but eventually continuously as the differing survival times lead to staggered deaths.
But humans are wonderfully diverse. In each group of infected, there would be those who, by some quirk of biology, survive longer. Invariable or inevitable doesn’t mean quick. Those who survived would reproduce. And whatever quirk of their physiology allowed them to last longer against Kuru would be passed down to their offspring.
If you aren’t a member of the Fore people, Kuru would kill you in a few years. But if you’re a member, you might survive thirty years after contracting it. Thirty years is plenty of time to die of something else.
Even in a small, isolated ethnic group in Papua New Guinea, there was enough genetic variation for the genes that protect against Kuru to be “found” and amplified.
Imagine our hypothetical prion plague spreading across the globe. It would kill many, many people, but it would kill most of them slowly. They would have time to pass down their skills, to shut down any reactors they were in charge, etc.; essential to prepare for the end of their lives in an orderly fashion.
And among these people, there would be some who are mostly unaffected and some who are completely immune. It would be a crap-shot – we’d need immense amounts of coordination and altruism to pull through this sort of prion pandemic. Or rather, our current society would require desperate measures if it were to survive. Humanity would come out – if not fine, then alive. Living without running water and antibiotics isn’t pleasant, but we pulled it off for half a million years and we can (hopefully) do it for another half a million if we have to.
What about a virus?
Ebola killed roughly 8,000 people in 2014. Influenza killed about 4.5 times that many people. In the USA alone.
Ebola infected about 30,000 people in its worst outbreak ever. Influenza infects 200,000 people in the USA each year. The common cold infects an incalculable number of people.
Ebola grabs all the headlines. The flu is the target of a new vaccine every year. And yet the common cold is more reproductively successful, when we look at total number of people it infects.
The common cold is successful because it is so mild. You get the cold and feel like shit, but you’re still mostly okay to go out and work. Your partner will still kiss you. People might still sit next to you on the train. And so the virus gets passed on.
When you get influenza (the real deal, not the 24 hours of stomach pain sometimes called the flu but in reality caused by food poisoning), it tends to floor you. Even partners do their best to limit their exposure to you and your boss wants you well away from her, thank you very much. And so you pass it on to fewer people.
When Ebola is in a city, people become scared to even leave their homes. All non-essential social contact stops. Transmission is rare.
I know I’m comparing apples to oranges here – influenza and the common cold are both airborne and Ebola is not.
But stop and think about what these case numbers mean. We often portray viruses as an inimical threat, as organisms hell-bent on our destruction. This is false. This is false for reasons beyond the inherent mistake of ascribing agency to non-sentient machines.
Viruses are replicating patterns that need the machinery of another organism to replicate. They are the simplest possible parasite. Viruses without hosts are just inert RNA or DNA in a thin protein coating. Viruses need their hosts, insomuch as they need anything.
Ebola isn’t particularly dangerous to human civilization because outbreaks of Ebola tend to burn themselves out. It kills more quickly than it can be transmitted. And so transmission stops and the viral particles again become inert.
Viruses that kill their hosts too quickly face selective pressure to slow down. This can occur over the course of an outbreak too. Is it any wonder that the strain of Ebola that caused a particularly large outbreak ended up being less likely to kill than most other strains?
Maybe this is actually just the result of better care. We’d need to run a lot of sequencing of a lot of Ebola samples to be sure. But we know how selective pressure on viruses work and those models would predict that this strain of Ebola evolved to be less virulent over the course of the epidemic or was more successful because it was less virulent from the start.
There’s an opposite pressure that bears mentioning here as well. Viruses can’t go too far in the other direction. They need to be fairly virulent if they want to spread successfully. Each virus particle is another chance to infect another host. They want their current host to make a lot more of themselves. They’re just incentivized by evolution to do it at a reasonable pace.
I came at this point the long way around, but here it is:
Any “super” virus created in a lab will face this selective pressure.
In parallel to this pressure, humans will face an immense selective pressure for immunity or resistance.
The end result of the release of any virus engineered to kill all of humanity would probably be a new equilibrium. The virus would tend to become less virulent with each case and many survivors (and everything we know suggests there would be survivors; even HIV and rabies don’t get everyone) would probably have an innate resistance that they could pass on to their children.
Perhaps you’ve come up with a super-virus and are aware of this problem. You know your virus will eventually betray you and become less virulent, so you decide to trap it in amber, so to speak.
Viruses tend to have poor DNA or RNA replication machinery. This results in lots of errors every time they reproduce. With many viruses, you can tell the number of them inside a cell by the number of variations in the genome. Their copying machinery is so error-prone than every single virus has a mutation.
But you don’t want that, so you give your virus the best possible DNA/RNA replication and repair machinery. Its DNA/RNA will be exactly as you intended it in every single copy. It won’t become any less virulent now!
You’ve done two things. First, we’ve already established that some people will be immune. If your virus doesn’t evolve at all (and evolution requires variation for selection, which requires mutation), then these people (and all their descendants) will always be immune.
You’ve also just given whoever has to come up with the vaccine and other treatments for it an orgasm.
You know the flu shot that you don’t get? The one that you’re supposed to get every year because the surface of the influenza virus changes every year and you lose any immunity to it? This is because of influenza’s bad RNA replication mechanisms. It’s the same with the common cold. This isn’t true for many viruses. Once you get chicken pox, you’ll probably never get it again. Your body has learned its surface markers and knows to kill them on sight. These markers change very rarely, so your immunity doesn’t expire.
These changing markers are the best (and only) defense a virus has against vaccination. Take that away from them and vaccination becomes much easier.
Vaccination isn’t the only thing that becomes easier when viruses don’t evolve. We’ve had designer drugs for fighting viruses for a while now. Spurred on by the crystal structure of HIV integrase, scientists have begun to figure out how to take the structure of a virus and develop molecules likely to stop it.
It is true that HIV can evolve around some drugs. This is why we’ve switched to regimes of drugs, which combine several medicines with different routes of action. This strategy is much more effective at preventing resistance. Instead of needing one mutation to develop resistance, the virus must develop several, all at once. This almost never happens, even in fairly fast changing viruses. In one that was deliberately fixed it would be almost impossible.
Here’s what the response to a large scale bioengineered pandemic would look like:
- People would start showing signs of infection. Doctors, confused by the mortality rate and the rapidity of the spread would eventually send it off for sequencing (I’m eliding a host of difficulties here, but they’re solved difficulties; simple doesn’t mean easy).
- Sequencing results would come back and show that this virus is something new.
- There’d be a quarantine. In addition, people in unaffected areas would reduce the time they spend outside their houses as a precaution.
- As the death toll mounts, more and more scientists would begin working on the virus. Several avenues would be exploited in parallel.
- A vaccine would be developed using conventional techniques
- Large scale, massively parallelized and automated efforts would be undertaken to get the virus crystal structure
- Antiretrovirals, antivirals, and the kitchen sink would be tried on affected patients in the hopes that something would serendipitously work.
- Something would work. We’ve done this too many times now not to have the process figured out. At this stage:
- Either the pandemic had good DNA/RNA repair machinery, in which case scientists would be slightly surprised by the lack of resistance to treatment. Not having to work around resistance would allow doctors to give patients only the cheapest or best tolerated treatments, reducing the economic or human toll of the disease.
- Or the pandemic is free to evolve. In this case, we’ll see a steadily decreasing lethality rate. Part of this would be natural evolution and part would be the effort of scientists. Fully annihilating it would be hard and will probably take decades, but increasingly effective combinations of therapies would be developed, killing off even highly resistant strains.
If your next work of fiction involves a virus or prion killing all the world’s population, I’m sorry. May I suggest writing about nukes instead?
Epistemic Status: Falsifiable
4 thoughts on “Skepticism About X-Risk: Viruses and Prions Edition”
Kudos for writing this up!
> ” Once you get chicken pox, you’ll probably never get it again. Your body has learned its surface markers and knows to kill them on sight. These markers change very rarely, so your immunity doesn’t expire.
> These changing markers are the best (and only) defense a virus has against vaccination. Take that away from them and vaccination becomes much easier.”
Could you somehow allow the markers to evolve rapidly while the functional parts remain consistently hyper-virulent?
The normal mechanism by which markers change is just the DNA/RNA polymerase enzyme. It’s responsible for transcribing the whole genome, so you can’t be selective with it.
There are certain sequences that are prone to mutation, but these are either:
a) non-coding sequences (so they can’t really be used for things like surface markers) that have a lot of repeats
b) In Eukaryotes, non-methylated stretches of DNA may be more prone to mutation, but viruses don’t have the same sort of epigenetic methylation system that humans have; it’s the sort of thing that’s much more complicated than their genomes.
I created a map about biological risks, I hope it could be interesting to you
Hey there! I disagree, somewhat, for reasons I’m not keen to over-publicise. The FHI has recently hired its first exclusive biotech guy, so I’ll ask him when he gets in today.
Feel free to email email@example.com if you want to discuss this further.
Comments are closed.