The Genesis of the 1918 «Spanish Influenza» Pandemic

Lecturer
Michael Worobey

I want to start tonight with some numbers. So influenza doesn’t just cause pandemics, worldwide epidemics. It’s also a seasonal virus. It gets established in the human population. There are two lineages, H1N1 and H3N2, that circulate every year. And here are some numbers. So every year 5% to 30% of the population gets Infected. Most of us were probably infected by the time we were five years old. And that first infection kind of sets the stage for the rest of your life. Your immune system remembers that first virus, and for the rest of your life kind of responds to it every time you’re reinfected. You get it for about a week and then you clear the virus. And it’s gone until the next time you encounter it. So just regular old flu every year accounts for, on average, 31 million outpatient visits in the US. 200,000 people end up in the hospital, for a total of 3.1 million hospitalized days. Total economic burden $87 billion— that’s medical cost, work lost, work time lost, and so forth. And the bottom line is it kills something like 30,000 people in an average year. So this is a serious disease, more than half a million worldwide usually each year. And if you tally those up, every year in the US, people’s lives are cut short by a total of more than 600,000 years. Let’s go back to 1918 now. So we’re missing the audio here but we’re getting— It was the worst epidemic this country There it is. — Has ever known.

It killed more Americans than all the wars this century combined.

But you never knew from day to day who was going to be next on the death list. There were so many people dying that you ran out of things that you’d never considered running out before— caskets.

So 1918 was different. There’s never been an epidemic as lethal as what happened with flu in 1918. You can see it here on this graph of mortality rates by year, from about 1900 up to the 1990s. It’s not too hard to see 1918. Absolutely unprecedented, not just for flu, but for any other pathogen to kill so many people in such a short time. In the US, at a time when the population was maybe a third of what it is now, 675,000 people died, most of them within about three months— October, November, December.

In that year, the culling by the Spanish flu actually dropped the average lifespan in the US by 12 years. Around the world, something like 50 million people died. Estimates range even as high as 100 million, and possibly for instance, 20 million just in India alone. So this was just unlike anything we’ve ever seen before or since. And still about 100 years after it happened, we don’t really know why.

So what I do is evolutionary genomics. So I take the genomes sequenced from things like influenza virus, and when possible I line them up against other genomes from the same germs, and then try to read the history of those germs from the DNA, essentially.

This is a figure from the paper that Joaquin alluded to. In 2009 when the swine flu jumped into humans, we tried to figure out where it came from, what species that jumped from, and when it made the move. And I’ll come back to that. But basically at this stage, I just want to communicate how rich the information is, now that we have full genomes. And for flu we have tens of thousands of complete genomes that are available. I used to read National Geographic when I was young. And I can remember looking at stories where you have an archaeological dig, and little hints about things that happened in the deep past. And thinking, you know, that’s really great, but this kind of wistful feeling of, there must be so much that we’ll never know about. And genomics, and evolutionary genomics in particular, is taking us back in time to places where 15, 20 years ago, we thought we couldn’t go. But when you have a collection of genomes, that collected information can tell really important stories.

Every one of these virus particles presents two main things to the outside world, i.e. our immune system— an H and an N. The H you can think of like a key. That’s the key for the virus to get into the cell. And the N you can think of like a little scalpel. When those new virions, those new virus particles, are made in the cell, after the genome, those eight segments takes over the machinery inside the cell, and turns it into a factory for making new flu viruses. The N protein on the outside of the virus cuts a little hole, essentially, for those near virus particles to escape. Those antibodies that you saw, this is promotional material from a drug company. They make antibodies essentially as drugs. Our own body makes antibodies. And every time we’re exposed to a new germ like flu, it’s like a little R&D project. Our immune system comes up with essentially a drug, a molecule that binds either to HA, the H, or to N, the scalpel. And it’s just physical binding. It just gets in the way of those viral proteins from doing what they need to do for the virus to complete its life cycle. It’s as simple as that. And those eight genome segments contain the information not just for the H and the N, but for all the other dozen or so proteins that allow the flu to do its thing. And so here is a flu genome. This is the one and only complete genome that you’re likely to see during this series of talks, because RNA viruses are very sensible and thrifty. And they can do the whole thing in about 10,000 nucleotides, OK? So that’s one segment. And you can read the A’s, C’s, G’s and T’s. That’s the segment that codes for the key, H. And that’s the scalpel, N. And that’s it. That’s all that the flu needs to take over our giant bloated genome and start turning the knobs and making it do what it wants our genome to do, which is to create more copies of the virus. And so you can think of the battle between flu virus and the immune system as a boxing match. And in one corner you have the virus that brings to the table this H protein, the key to get in, and the N protein, the scalpel to cut out. But we come to the table with something as well. After that initial exposure when we’re children, we make antibodies, sorry, to both of those proteins. So that the next time we see an influenza virus in our body, we bind those antibodies to it, and if not kill it, slow it down. And that can make the difference between pneumonia and death after you get flu, or a kind of cough and a mild fever.

So this kind of dynamic, where you have the immune system kind of putting up this barrier, and the flu attempting to get around it, sets up what evolutionary biologists call Red Queen dynamics. And it’s named after the Red Queen from Through the Looking Glass, who said «it takes all the running you can do to keep in the same place.» And you see this a lot in nature. It’s kind of an arms race. Our antibodies bind to the flu, so once you get this year’s flu strain, next year’s flu strain had better be a little bit different in the H and the N, otherwise it’s not going to be able to pierce your defenses. And that Red Queen arms race kind of dynamic is why you need to get a new flu vaccine every year. Unlike things like measles, the flu evolves, so that your initial immunity doesn’t necessarily do the trick next time you encounter the virus.

So there’s another small element of the story that I want to introduce you to here. It was discovered by this guy, Tom Francis Jr., who actually trained Jonas Salk, and developed the first flu vaccine that was used during World War II, Pretty much the same technology that we use today, actually. And he noticed something really weird in the late 1940s. So in the late 1940s there was a complete vaccine failure. The flu that was out there had changed so much that the previous vaccine just didn’t do any good whatsoever. And so he was doing experiments, and noticed that when he gave a vaccine based on the new flu variant that had replaced the old one to teenagers, instead of making antibodies to the new variant as you would hope your immune system would do, it actually boosted up the old antibodies that it remembered from— that each person’s immune system remembered from their childhood infection. And he named that process the original antigenic sin. So it’s kind of like the other original sin, that it kind of taints you for life, potentially. And it’s part of the story. We’ll come back to it. But what that’s important for is the year to year dynamics of the things like H1N1 and H3N2 that are the seasonal strains now.

It’s less important for this process which is the other big way that influenza comes up with evolutionary novelty. Instead of just making minor tweaks to the H or the N, every once in a while, a new flu virus is created that mixes and matches segments from two different viruses. So if you have a person infected with H1N1, and as happens from time to time, that person is simultaneously infected with a bird virus that normally doesn’t infect humans, the virus that comes out that person can be an H1N7, if you have one segment or more from that bird virus jumping in. We call that re-assortment. And that’s, as you might imagine, a kind of game changer. Whatever immunity you had to N1 from your childhood infection or lifetime exposure, it’s useless. N7 is so different antigenically, that your antibodies to the previous version don’t really help you. And that process, this jumping in of complete genome segments from nature, from non-human reservoirs, is what ushers in what we call pandemics, which are these worldwide big epidemics that are occasioned by completely new genetic diversity from the pool of various subtypes of virus that exist in these animal hosts, jumping into humans and starting to spread. And so when that happens, let’s say in 1957, after H1N1 circulated, an H2N2 virus jumped into humans in 1957.

At that point on an individual level and a population level there was almost no protection from age H2N2. And so you have a big year, a couple million people dying, instead of half a million. In 1918 you had the same process. You had some new stuff coming in, but 50 million people died. So this is the kind of layered history of the virus from now, when we have H1N1 and H3N2 circulating, back in time. If you were born before ’57, your childhood infection, your first exposure was to H1N1. That set the stage for the rest of your life. No one was safe. In Washington, Victor Vaughn was working late trying to make sense of the hellish chaos. He uncovered an unnerving fact. Usually influenza kills only the weak, the very young, and very old. But this time it had a different target. People in the very prime of life from 21 to 29 were the most vulnerable of all. This infection, like war, kills the young, vigorous, robust adults. The husky male either made a speedy and rather abrupt recovery, or was likely to die. The situation was upside down and backwards. A disease that’s supposed to be a mild disease is killing people. The people it’s killing are the strongest members, the most robust members of our society. [END PLAYBACK] OK so the plot thickens. Not only was 1918 unusual in the large number of people dying, but the wrong people were dying. Flu’s supposed to kill very old people and very young infants. But in 1918, you had this extra peak in mortality in 20 to 40 year olds, and really focused in on people about 25 to 29, who normally almost never die from flu compared to the rest of the age groups.

And so there are several kind of mysteries, but there are also clues that we need to address here if we want to figure out what happened in 1918. So first is that young people, young adults, were dying. And that’s unusual. The second is while young adults were dying, the very age that normally does most of the dying in a flu epidemic, the most elderly, they actually died in fewer numbers than they did during seasonal flu. So if you just compare the mortality among, say, 85 year olds in 1918, in the midst of the worst pandemic, the worst epidemic of disease that humanity has ever faced, the most vulnerable people actually died at lower levels than they had in the five years before from whatever flu strain existed then. Very strange. Another key thing, when you watch documentaries about the flu, the 1918 flu, or hear about it, you often hear scary stories of people turning blue, feeling fine, and dropping dead a couple of hours later. Those cases happened occasionally. But the vast majority of people who died from flu, including those young people in 1918, died of secondary bacterial pneumonia. Which is the way flu has always killed, and it’s the way it kills now. So I have a picture up here of a garbage can. And it is a garbage can, not a trash can, because it’s Canadian. And this garbage can is one like the one that killed my 93-year-old grandfather a few years ago. So a few years ago he fell in his kitchen. He busted a rib on the garbage can. And several weeks later, after contracting pneumonia and getting congestive heart failure, he died. And you can think of flu, in 1918 and otherwise, like that garbage can, or like a shot to the ribs. It lands you on your back for a few weeks. It creates inflammation, fluid buildup in your lungs where you need them clear for oxygen, for gas exchange. And pneumonia is really, really deadly. It always has been, and it still is. And that’s what killed people in 1918. The other clue is that it was an H1N1 virus. And so how do we know that? We know that because people like Jeff Taubenberger and his team have actually gone back to tissue samples from 1918. So here we have wax-embedded autopsy tissue samples taken from a victim of the Spanish flu. And Jeff Taubenberger and his team take slices of that material, and they amplify up that genome that we saw before. So we know exactly, genetically, what this thing looked like. In fact, the one complete genome that we have from 1918 comes from here. So this is a place called Brevig Mission, which is a little village in Alaska. And in 1918, the Spanish flu pretty much wiped out the whole village. And a lot of people ended up in mass graves. And this guy here, [? Johan ?] [? Holten, ?] is a virologist, who for years had been saying we might be able to get virus out of those graves. If they were buried in permafrost, you might have good preservation of material. He tried in the ’50s, at a time before genomics and amplifying gene sequences. What he needed to do then was actually get a live virus— couldn’t do it, not possible. Everything’s kind of dead and fragmented. So he went back in the ’90s on a solitary trip. I thought this was funny. So he carried just one tool, a pair of garden clippers borrowed from his wife, without permission. And let me just go back.

There they are. OK? So from this mass grave, via Mrs. Holten’sgarden clippers, to the big screen. That’s not just a flu genome. That is the Spanish flu genome from that victim from that mass grave in Alaska. Those A’s, C’s, G’s and T’s are the virus that killed that person and so many others in 1918. And I can remember when the first full gene came out in 1998. And it was very exciting because finally we had insight into why this thing killed so many people, and killed young people. We had the full gene of the H protein, this most important gene of all. And it was published. And it was kind of like hmm, there’s nothing here. There’s no smoking gun. There’s nothing you can see in the actual genome sequence that really tells us anything about why it was so nasty. And it turns out you need more than just one genome. You need to put it into context of other ones. So those are the clues. But we’re left with really big questions. Did this virus jump in from a bird? Did it come from swine influenza and human influenza swapping genes? Some people interpret the data in that way. When did it happen? We don’t really know when this virus came together. And we don’t know where. We don’t even know which hemisphere this occurred in. There’s good agreement on a couple of points. One, we have no idea why young people died in such high numbers. And two, the Spanish flu wasn’t from Spain. So Spain was one of the only countries in Western Europe that didn’t have press censorship during World War I. So in the summer when things started heating up with the flu pandemic, they were publishing stories. And so everyone associated it with Spain. OK so back to what I do. So I look at genome sequences of this sort and compare them. And I’ll just take you through this 2009 example to warm you up for 1918. OK so the main tool here is called a phylogenetic tree, or an evolutionary tree. This is the first phylogenetic tree ever, in Darwin’s notebook from 1937. And what a phylogenetic tree does is describes a process of common ancestors splitting into independent lineages. And whether this is a species that gets separated into geographically isolated populations, and enough time passes, and enough DNA mutations occur independently, or whether you’re looking at descendants of a single female, my wife is in the audience and we have two daughters. So 10,000 years from now, you might find some people trace back to Hazel, and some people trace back to Iris. Whatever happened in terms of DNA mutations along each of those lineages was independent. And that forms the branches on the tree. So it’s a fundamental tool in biology. So fundamental, you might want to get it tattooed on your body. It’s a very fresh looking tattoo. When you do an images search of this, you get about a dozen different tattoos of people who independently thought, I’m going to be the only person in the world who has this tattoo. OK so here’s a tree. Here’s a real schematic phylogenetic tree, where if you looked at a bunch of human DNA sequences, they would trace back to a common ancestor about 100,000 years ago. If you compared them to chimpanzee DNA, there’s many more changes in each lineage before you get back to the common ancestor, say five million years ago. So now let’s switch to 2009 swine flu.

So the new flu that jumped into humans in 2009, we found traced back to an ancestor not too many months, maybe four months, before the first case was noticed. But if you compare the human virus to the closest pig viruses, and every one of the eight segments was a pig segment, so this thing jumped directly from pigs. You have to go back to about 10 years ago. And what that means is that the actual constellation of genome segments that formed that 2009 virus, formed many years before it jumped into humans, and spread cryptically in pigs. And our surveillance of pigs is obviously so bad that no one knew about it until it jumped into humans.

OK so how do you get those dates? This is kind of one of the main things I do in my research. And with viruses they evolve so quickly, that you can actually see evolution happen in real time. And you can liken it to a saguaro. And there are, I know, quite a few Canadians in the crowd. So just imagine a Douglas fir tree and you’ll be OK. If you just go out into your front yard and look at your cactus, it’s really hard to guess how old it is if you don’t have anything to go on. But if you had say, a picture from 1999, and then another one from 2009, now just with 10 years window of observation, you can start to get an idea of how fast the thing grows. And maybe you would extrapolate back and say, it’s about 100 years old. So what we do with viruses is the same. We sample them at different time points, calibrate what we call a molecular clock, and then you use that to— this is Salvador Dali’s painting called Bayesian phylogenetics and relaxed molecular clocks.

So if you do that with influenza, what I noticed about a year and a— so this project really kind of got rolling a year and a half ago. One night I had a whole bunch of these evolutionary trees out on my kitchen table. And I was actually using my daughter’s plastic ruler to just measure the kind of— the distance from the base of the cactus to the tip, and compare that to the date. The same thing with the cactus, except with these actual trees. And what really came out was that the virus evolves, the clock ticks at a very different rate, in viruses infecting say, pigs, on the one hand, or humans on the other, or horses, or birds. And you can’t just average across them. If you average across them, things go wrong. So I started collaborating with a couple of other people, [? Gwangju ?] [? Hahn, ?] who is my graduate student. [? Gwangju, ?] are you in the audience here? There he is, excellent PhD student. And so [? Gwangju ?] and I started gathering together all of the genome sequences of flu, more than 100,000 of these genome segments. And then teamed up with this guy here, Andrew Rambaut, who I worked with in Oxford. He’s one of the best all around biologists, and an incredible computer program. And we came up with a method that just kind of makes that simple correction of allowing the clock to tick at different rates. And he’s a very good biologist, especially considering he’s only 18 inches tall. He’s just absolutely amazing. The truth is I, of course, didn’t email him asking for a picture. I just did a Google Images search. When you Google images, Andrew Rambaut, you get phylogenetic trees, and two very low quality photos, including this one.

So those are my main collaborators. I also want to thank Tom Watts, who’s my lab manager, and has a hand in everything I do. And also my wife. So while I’ve been doing this project, I’ve been kind of immersed, living between about 1872 and 1918. And thinking about 50 million dead people, and particularly for the last six months. My main goal normally in life is to make my wife laugh. And I haven’t been doing that too much over the course of this project. But with this tool in hand, what we found, was basically everything that had been done in the last 10 or 15 years on evolutionary trees of flu, including stuff that we had done, was wrong.

And this is just a kind of figurative representation. But everything that you do without correcting for the different clock rates is kind of like looking through a fun house mirror. When you make the correction, everything snaps into place. And I know what you’re thinking, that it doesn’t look any better now than it did in the previous slide. But I’ll take you through it. It’s not as bad as it looks. OK, so here’s this Brevig Mission. This is the flu sequence from one genome segment from that 1918 patient. And you see, that’s 1918 there. And it’s nested in a little part of the family tree along with human viruses and pig viruses. And just a few years before 1918, it jumped from a bird, boom, right there. And it jumped from a bird that lived in the Western hemisphere. So this whole big family tree kind of splits into two main lineages. One— all these black lineages are from bird viruses. One from the Western hemisphere, another from the eastern hemisphere, and they trace back to a virus at this point here, that existed a few years after 1872, OK? So we’re getting somewhere now, at least for this segment. This thing seems to have jumped from a bird not too many years before 1918, and probably from the Western hemisphere. Also check out that horse lineage while I get a drink.

So this was a big surprise. This horse lineage, known from just a handful of sequences, first isolated by a woman in Prague in 1956. And it probably went extinct in the ’70s. So no one cares about. No one even includes it in analyses normally. But we were doing a comprehensive analysis, and we included it. And it turned out to be a kind of key to the story. We’ll get back to it in a second. Now let me take you from that one genome sequence to six. So now we’ve got six of the genome sequences, everything except that key G and N.

And all six of them have this, basically the same story— the human virus jumping from birds shortly before 1918. Boom, boom, boom, boom, boom, boom. Western hemisphere, Western hemisphere. So what this looks like is a single event. A single virus brought all of these avian influenza genome segments into the Spanish flu genome a few years before 1918. Let’s now zoom in on one of them. So here’s our 1918 sequence. There’s the related human ones that circulated up to 1957. Most of you in the crowd tonight would have been exposed to that.

But that 1918 virus jumped from a bird really close to 1918, and probably from Canada or the US. All of the closely related viruses are from North America. So now we’re getting even closer. Six of those genome segments seem to have jumped in from North America. And if you look at the species here, the conventional wisdom is that flu jumps from wild waterfowl. That’s the basic conventional wisdom about the reservoir of flu. That’s the natural home of flu viruses. And every once in a while an accident, an unnatural event happens, and one of those genes jumps into humans. Well it could have been a turkey, could have been a chicken. And starts to look a little bit more like the situation now, with H5N1, which is this virus that infects poultry worldwide. And about which there’s a tremendous amount of concern that it might make the jump. OK let’s now go back to the horses, OK?

Same pattern, horse horse horse horse horse. So at node one here, what you’re looking at is the ancestor of that big bird lineage that gave rise to most of the Spanish flu genome. And a horse lineage, and at node 1 there’s just two possibilities. Either that was a virus that was in a bird and jumped to a horse at that point. Or it was in a horse, and jumped into a bird.

So now the story is going to get even more strange. And believe me, when I was doing this research, and started making inferences about what happened and when it happened, I actually didn’t know about the great epizootic of 1872. And if actually, if you wouldn’t mind, can you raise your hand if you’ve heard about the great epizootic of 1872? Larry Venable’s showing off, OK. So he’s heard me give this talk before, I’m pretty sure. OK, I didn’t know about it either. An epizootic is just an epidemic in an animal. It’s a different name, OK. But in 1872, starting in Toronto in the fall, you had essentially Spanish flu for horses. You had a devastating outbreak of influenza in horses that swept from Toronto in all directions. By 1873 in March, it was out this way. And so the US cavalry and the Apaches were actually fighting each other on foot for several weeks. Because all their horses and mules were either sick or dead. And this was actually a historically extremely important event. This is at a time when horsepower came from horses and mules. And each one of these cities, like Boston, for instance, for several weeks, all the horses and mules were sick or dead. And it’s kind of like a science fiction movie where all of the cars, trains, and the electrical grid get knocked out for three weeks. It was pandemonium. It was— half of Boston, the oldest part of Boston burnt down, because they were trying to drag the fire engines with teams of young men instead of horses.

And it wasn’t just a historical curiosity, OK? This horse flu outbreak is connected to the bird flu outbreak that we see in that evolutionary tree historically. So just about a year ago, the same guy that generated the 1918 genome of the virus was doing some historical research with David Morens, who’s a historian. And they found this weird thing. In 1872 simultaneously with this horse flu outbreak, in city, after city, after city, there’s reports of what, for all the world, looks like highly pathogenic avian influenza. Except it’s five, six years before that disease was scientifically recognized in Europe. So this is a real headline from the New York Herald in November of 1872. «Swelled Heads on the Thanksgiving Gobblers & Henfluenza Devastating the Chicken Coops.» Very strange. OK this match between the evolutionary tree telling the story of a potential horse to bird jump of these virus segments, and the history. And it’s really hard to dismiss the coincidence. And what it looks like is instead of the conventional view of birds being at the center of the genetic diversity of the virus, that before the birds had it, horses had it. And they gave it to the birds. And so the majority of the genome of that 1918 virus started in horses, jumped probably, and almost certainly into domesticated birds, possibly then into wild birds in the Western hemisphere, North America, and then finally into us. And again, this is not abstract. So after I saw Joaquin this evening and shook his hand, and he told me he was feeling sick, after rushing to the bathroom and washing my hand, I thought well maybe he’s got flu. Chances are, some of us in the audience tonight have H3N2. That’s the seasonal flu that’s going around. Well five of its eight genome segments trace back to the 1918 virus. And six of its eight genome segments trace back to that horse in Toronto.

OK, the next part has absolutely nothing to do with influenza. But while I was digging around about the history surrounding the 1918 virus, and I was really interested in troop movements and horse movements from North America to the Western Front in 1918, I stumbled across this. And figured I got to share this with you guys. So apparently, there were lots of plans that never came to fruition for Germany to invade Canada. So Canada was fighting as part of the British empire in World War I for several years before the US entered the war. The US was still neutral. And one genius imagined through rose tinted glasses that there were probably 100,000 secret German military reservists living in the US. And that if he planned an invasion, they would probably be joined by maybe a quarter million German Americans. And just throw in 300,000 Irish Americans. They’ve got to hate the British enough to invade Canada. And this is the relatively sane part of the story. So the plan was this— to maintain secrecy, the army of 650,000 would dress as cowboys. So this also served as excellent comedy relief while I was doing this project. It kind of cleared my head. And this is not just some crazy thing off the internet. I found out two days ago when I was writing this talk that the paper was written by a guy I took a course from, Simon Fraser University, on World War Ii history. So this is real. I know that guy. And it was taken seriously enough that the foreign office actually had the lawyers look at it. And they said, you know, this is OK. You need some way for them to know who’s part of the secret plot. And dressing as cowboys seemed the obvious thing. No one’s going to suspect 650,000 cowboys with German accents. And so this was not, this never happened, but not because it was insane. But because the German foreign office was worried that the invasion might bother the United States, who was still neutral. OK. So this was basically just to clear our palate before the final part of the talk here. So back to flu. We’ve got a virus, most of its segments in 1918 have jumped in from a bird flu. The end of these two important proteins on the surface of the virus, the end looks a lot like those other six, looks like a bird virus that jumped in. But the H is a different story. So this H protein, its genetic diversity, if you compare it to all the other ones, is way older than all the other segments. And it’s a very simple interpretation. It’s the same thing that happened in 1957 and 1968. That an already existing virus in humans with an H1 that probably jumped in in about 1900, combined with an avian virus a few years before 1918. And that formed the genome. OK so that’s the story, that’s the basic. Those are the facts. And now how do we link those to what happened in 1918?

Well it turns out that evolutionary trees are not the only way that you can look back in time at what was happening with influenza. And before there was 1918, there was 1889, and before that 1847, and before that 1830, periodic pandemics, when we think new influenza H and N genome segments entered the population and became new human viruses. If you look at people born just after 1900, there’s a big spike in the anti-bodies of those people to H1. And that’s not what you would expect if H1 came in 1918. If it came in in 1918, you would think that babies born in 1916, -17, -18 would have that spike. So now we’ve got some triangulation going on. The gene sequences are telling us H1 probably jumped in in 1900. The antibodies in people’s blood is telling us people born during this time probably had H1. And you can go back further. So this pink line here is HeN8 antibodies in people born during the 1889 pandemic. Very, very clear signal that the virus in 1889 was H3N8. That’s not my discovery. That was already known and quite widely accepted. But check this out— N8 didn’t disappear in 1900 when I think H1 jumped in. It stuck around. And this was actually the best part of this whole project. I had worked out the kind of main parts of it. But I couldn’t, I still couldn’t figure out why, in 1918, you would have this switch where 20 to 40 year olds would suddenly be doing much worse with this new virus. And 88 year olds were suddenly doing much better with H1N1. It affected them less than the previous virus. And so I thought to myself, you know, the 1889 virus was probably H3N8. And so I wonder if the N8 was actually passed like a baton, as often happens with flu, from 1889 to 1900 virus, and then to the 1900 to 1918 virus. And I thought OK, I’m going to have to write a grant. And I’m going to have to learn how to do all this stuff that I don’t normally do. And then I thought, oh wait a minute. This is not how we work as scientists. And thought I’ll look through the scientific literature. And within about five minutes, I found a paper from March of 1973, a month before I was born. And in the abstract of the paper it said that it really seems like the N8 portion, the N8 antigens, persisted in a virus after 1900 until about 1915 or 1916. And that was like, OK, this is actually coming together. And if you— I’m just going to skip through that— if you put it all together, now you can start to fill in the layered immunity that would have existed in the population alive in 1918. Other people had already speculated that those old people did well because their childhood infection was H1N1. If that was the case, once they were re-infected 80 years later with another H1N1, their body retained that memory and was pumping out good H1N1 antibodies. 1900 may have been an H1N8 which would mean the cohort born between 1900 and 1918 would have had some protection. They would have had prior exposure to H1 so the 1918 virus wasn’t completely new.

It’s this little wedge here, these people who were primed as infants and children by the 1889 virus, which was probably H3N8. They were probably the worst off in terms of their childhood immunity. They probably had very little protection to the 1918 virus, probably kind of similar to a baby who’d never had any influenza before, have no antibodies at all. Their H3N8 were useless against H1 and N1. And that’s what it looks like when you’re in a boxing match against a virus, and you have no antibodies. It’s like having two arms tied behind your back, and boom. I also notice that Sonny Liston’s foot looks very long in this photo. That’s like the German cowboys, has nothing to do with anything. OK so if you just take this model, and you say what’s the sort of average cohort protection of people of different age groups?

Well 88 year olds, we think maybe they had H1N1, and so they had good protection in 1918 against H1N1. These guys here sort of had intermediate. They had some protection from H1, not from N8. And various reasons I’m not going to go into, these ones also may have had sort of half protection. But it’s this one , this cohort here, the babies of the 1889 pandemic, that were the worst off according to this model. And if you just compare that to who died more than seasonal flu, and less than seasonal flu in 1918, it looks pretty similar. Thank you, Ken [? Gaudet. ?] Old people, good protection, low mortality. 25 to 29 year olds, kind of naked, no help from their antibodies, two arms tied behind their back, high death rates. And it seems unlikely that it’s a coincidence that World War I packed 20 to 30 year olds into high densities in army camps, troopships, trains, the front lines. If this model is correct, what that was was essentially taking kind of dry tinder and packing it into very compact physical locations. And if that’s the case, then probably the emergence of H1N1 was both a cause and a consequence. World War I may have been an ecological disaster in terms of allowing an H1N1 virus to emerge. And once it did, it was most likely to infect people who had no protection. OK so what does this mean? First thing is this, which I tell my daughters on a daily basis. If that is what happened in 1918 it’s kind of reassuring. It means we can actually have a handle on the likelihood of a 1918-like pandemic happening again. And it’s not very likely. The next slide here shows a big difference from 1918. That’s about half a percent mortality per year due to infectious disease in 1918, now, negligible. So there’s just not as much pneumonia secondary bacterial infections and all of the other things that actually deliver the coup de grace when you’re infected with flu as there was back then. And we have things that we can do about it. We have vaccines against a whole bunch of bacterial pathogens, vaccines against flu, and antibiotics, modern medical interventions. If everything else was equal to 1918, and you put that virus, and you had that layered immunity, everything else except modern medicine, it’s very unlikely that things would unfold in anything like what they did in 1918. At the same time, it makes you kind of— and this is something that was raised by Joaquin during my practice talk— kind of makes you think about the benefit that we have in this current age, when we have antibiotics to streptococcus pneumoniae which kills so many people, and so forth. And so I’ve got juxtaposed do not panic with antibiotic resistant diseases pose apocalyptic threat. But actually this is a resource that you don’t want to squander on things that you don’t need antibiotics for. You want your antibiotics to work to prevent you from dying from things like this if they ever happen. OK so this perspective also gives us a handle on where things are going to go from this time forward in terms of new pandemics. Will H5N1 jump into humans? We don’t know. But if it does, it would seem that the people born in 1968, who their childhood infection was an H3N2, different, those individuals who are now about 40 years old, might be a kind of beachhead cohort like the 25 to 29 year olds were in 1918. So looking ahead you might want to think about targeting vaccines to those people if a H5N1 pandemic took hold. And you might be able to slow it down, or maybe even prevent it. H5N1 has been in the news quite a bit lately, this highly pathogenic avian virus. It’s not clear whether it can ever be transmitted from human to human. And there are scientists who are working with ferrets, which are the best mammalian model, that whatever happens in ferrets seems to happen in humans. And you’ll be happy to know that they have developed mammal transmissible H5N1.

Seasonal flu, OK so there’s H3N2 and H1N1. H3N2 is the severe virus. It’s the one that’s circulating this year. It kills about 25,000 people in an average year over the age of 65, H1N1 maybe 1,000. So it’s— H1N1 is extremely mild in the elderly. Why is that? No one knows. Well if you were born before, got to be careful here who I include in the elderly cohort. If you’re 65 or older, if you wouldn’t mind raise both your hands. Very brave. OK so that’s— keep them up if you don’t mind, and then look around. I’m not 65 or older. That’s our herd immunity to H1N1 if indeed this childhood infection is really important. And if you were born— if you are that age, look around now. Because that’s the childhood immunity to H3N2. So this might be, actually, really important, not just for pandemics, but for seasonal flu as well. And I should also just very briefly say that looking ahead several decades, eventually today’s 40 year olds will be 65. And at that point, there will be a lot of H3N2 childhood antibodies. And the prediction would be that H3N2 would stop being so severe. And H1N1, if it’s still around, would be more severe. OK wind things up here. The timing of those genomic analyses of maybe 1915. A few years before 1918, maybe this virus actually already existed and was cryptically circulating. This is interesting because other people had rediscovered a couple of old papers from the Lancet that were published before the 1918 pandemic occurred. Both of them talking about purulent bronchitis. So both of them were from army camps, one in ETOP in France, which was the main receiving hospital from the Western front, and one in Aldershot in England. And both of these in 1916 and ’17 witnessed what, in retrospect, looked a lot like Spanish flu. Really high mortality in young people, a lot of the heliotrope cyanosis, this ashen blue pallor. And at least this group here speculated a year later that that indeed was an early Herald wave of the 1918 virus. Adolph Abrahams was the doctor on this one. He was the brother of Harold Abrahams, the sprinter that you might remember from Chariots of Fire. This one here was the first of these two papers to be published. And there’s a guy named John Oxford in England who’s been speculating for years that maybe these were early Herald waves. And in a lot of his papers he kind of said, you know, we don’t have sample— we don’t have access to these samples because a few weeks after World War I ended, those hospitals in ETOP were demobilized, like, immediately. And so one night, again in the midst of this, some time after the 650,000 cowboys, I thought well, I wonder if there might be samples. And so I thought OK, I’m going to go through the author list. And I’m going to target the pathologist. Because he’ll be the one that would have kept samples, if there are any. And so I started just using the magic of Google. And I started searching to see if I could find what happened to this guy. He survived the war. He went on to live for quite a long time. I found that he had two sons, one of whom survived to adulthood, and his name was Charles Hammond. At about this point, it’s 2:00 in the morning. And my wife came out and said go to bed, you wacko. It’s 2:00 o’clock in the morning, and why are you doing this? But I pressed on, and found an obituary of this guy’s son. And that obituary was written by the husband of his granddaughter. And through a bit more internet sleuth work that bordered on cyber stalking I found what I thought was his e-mail address. And so 2:30 or 3:00 in the morning now, I send him an email. And say, you know first off, forgive me if you’re not the person who I thought you were. And this is going to sound crazy. But I’d really like to find these samples. They’d be, amazingly, they’d be kind of like the holy grail for influenza, 1918 influenza research. Are you the right guy? Do you have any information? And the next morning I got up and sat down and read my e-mail. And there was an e-mail saying I am that guy. What a pleasure to hear from you. And I think I have the samples that you’re looking for. So two boxes of slides were preserved from this guy. They were his teaching collection. And in those two boxes of slides, there are at least a handful of samples. March 18th, 19177, this is from that study.

The family sent me a picture of him. So now we’re stepping back to 1917. This is the guy that made that sample. And the bottom line, I went to London a while back. They took me out to dinner. They’re absolutely delightful, very generous people. And they said, you know, he would want you to look at this. And so even though these are kind of family treasures, that’s what we’re going to do. And so we’re trying to figure out ways of sampling these slides without destroying them. And the beat goes on. We’ll see what we get out— we don’t have any information yet. But I think with that, I will thank you for your attention.

Oh, one more.

Well you know that the lecture next week is going to be another Michael, and boy are you in trouble. So there’s a few questions that maybe Michael Worobey can answer?

Michael. Yeah, Rich. So on that one slide showing the strains circulating through humans, through time, at the end you had the H3N2 and you had that H1N1. And it said H1N1 lab escape? Yeah, so I thought someone might ask about that. So in 1957, the Spanish flu like H1N1 went extinct when H2N2 emerged. In 1968, H3N2 kicked out H2N2. But in 1977, there was a kind of mini pandemic, and H1N1 reemerged. And bottom line is if you look at the molecular clocks, it was frozen in time, not even since 1957, but it was 1950 H1N1 strain. And it’s virtually certain that it was an accidental escape, probably from an experimental strain from China or Russia. And so yes. The worst pathogen in human history was accidentally re-released and not too many people know about it until you guys. Yeah? Hi, I had a quick question about, as far as immunization today. So vaccinations, I hear, I could be completely wrong, but they only protect us from about five different strands of flu virus. So I was wondering exactly how the scientists determine which strands to protect us from, and how really effective they are for us. And is it really worth us taking the shots? Yeah, so first answer is yeah, it is worth taking them. Second is that there are two different variants, H3N2 and H1N1. And there’s also another variant called influenza B. So the normal vaccine has all three of those components. And it’s kind of glass half empty or half full. It protects maybe somewhere around 50%. And that’s bad compared to other vaccines. It’s good if it prevents you or someone you love from dying. There is one issue though. It does seem to have the worst efficacy in the elderly, which is the age that you would want it to have the best efficacy. And I’m hoping that part of the story is that childhood infection with what we call heterosubtypic, so the wrong subtype, might actually be interfering with the efficacy. And we might be able to ramp it up in those older people.

There’s one somewhere, yeah. Hi, so you mentioned earlier how with the bird flu researchers, they’re using ferrets as a model organism. Why is, like, what’s so special about ferrets? Just because I would never think of that. Yeah they just happen to have lungs that act a lot like ours. Mice are also used. Pigs are also used. But ferrets, they can be infected either through the nose or through the blood. And whatever happens in them seems to be the clearest analog to what happens in humans. And so that’s why they’re used. Okay, the last one. Can the virus completely destroy the host cells they’re in? Or does the H and the N just open the cell for the virus to pass in? It actually can leave the host cell intact for a time, keep it alive. But once it is taken over, it’s no good as a host cell any more. If it does stay alive, it’s only staying alive to pump out new virus particles. Michael, thank you for an extraordinary talk. My pleasure.


Michael Worobey, Professor, Ecology and Evolutionary Biology, The University of Arizona