Late one night Michael Worobey began poking around on the internet, looking for descendants of a World War I British military doctor named William Rolland.
Rolland, a pathologist, had written a report in 1917, the year before the start of the Spanish flu. It described cases of British soldiers in France who had contracted an unusually fatal respiratory illness. Worobey, an evolutionary biology professor with a particular interest in the 1918 pandemic, wanted to know whether any of Rolland’s samples might still be lying about a century later.
Within a few hours, he had found a possible contact and fired off an email. Across the Atlantic, 5,000 miles away, a retired family physician in England’s picturesque Lake District received it. He replied immediately.
When Worobey read it, he was thunderstruck. “I almost fell out of my chair, for real,” he recalled. “I actually did cry real tears.”
The University of Arizona scientist had connected with a man named Jim Cox. And Cox just so happened to have in his possession a collection of human tissue slides that Rolland had handed down through the generations. Those slides, it turns out, could now help rewrite the history of the 1918 Spanish flu — altering our understanding of when it began and how it spread.
It makes for one heck of a story, one that could enter a new chapter next spring when Worobey and his team start trying to unlock the secrets contained within 10 rectangular glass time capsules.
Worobey has had the slides for several years now, but has waited to look for evidence of old viruses in them while he and his team honed the technique they plan to use. The goal is to sample a bit of tissue from each slide, while preserving them to return to the owners, for whom they represent part of a precious family heirloom.
“We’ve messed around with similar slides to try to even just figure out what’s the best way of getting at the partial specimen without ruining the whole,” Worobey explained. “I don’t want to crack open the cover slip on any of these before I really know that I have the best possible method of doing the work.”
“It’s kind of one chance for a clean shot,” he admitted.
There’s no guarantee the specimens will contain any evidence that points to answers to Worobey’s questions. “But at the same time,” he said, “if these people were loaded with influenza virus, it’s certainly possible that we’ll get something.”
The Spanish flu pandemic, which swept the globe in a series of waves from 1918 to 1920, is the deadliest infectious disease outbreak in known history. An estimated 50 million people worldwide died from the infection. Puzzlingly, many of the dead were in their late 20s, in the prime of life.
The pandemic occurred before the age of molecular virology. Back then it wasn’t even known what caused influenza; influenza viruses were first detected in 1930. In the decades that followed, though, it was recognized that the pandemic was caused by an influenza A virus of the H1N1 subtype. In the late 1990s, U.S. army scientists found and deciphered the virus’s genetic code.
Still, questions remain about when and where the virus started circulating in people.
A prevailing theory, outlined in historian John Barry’s 2004 best-seller “The Great Influenza,” is that the virus emerged in Haskell County, Kan., in January 1918, after having jumped into humans from another host.
It circulated through that spring without causing unusually severe illness. By the following autumn, however, an influenza outbreak with a high attack rate and an off-the-charts lethality swamped hospitals and overwhelmed the capacity of undertakers to bury the dead across the United States and elsewhere.
But other theories exist.
“There are published claims for Kansas, for England, for France, and for China. And I don’t believe any of them are definitive. And we’ll never really know,” said Niall Ferguson, an Australian historian who has also studied the origins of the 1918 flu.
Another lingering mystery relates to the fact that the outbreak was unusually deadly among young adults. Even garden-variety influenza kills — every year between 12,000 and 79,000 people die from flu in the United States alone. But flu’s victims are typically the very young and the elderly, creating a U shape when fatalities broken down by age are charted on the X and Y axes of a graph.
The fatality chart for the 1918 pandemic had a distinctive W shape, with the inner peak capturing the deaths of young adults in their late 20s — a population that typically bounces back quickly from illness.
Pulling on little threads
Worobey has a penchant for finding things others might assume are lost to the sands of time.
He specializes in the origins of pathogens. A few years ago, he and some colleagues garnered major headlines with the publication of a scientific paper showing that a French Canadian flight attendant named Gaetan Dugas — who had been labeled as Patient Zero in the narratives that emerged from the early days of the AIDS epidemic in the United States — was not the person who brought the virus to North America.
Worobey and his team have developed techniques for mining old medical samples for evidence of infections. To explore the Patient Zero mystery, they excavated stored blood samples and pulled out enough genetic material to sequence the HIV virus that eventually took Dugas’ life. They then compared it to viral sequences retrieved from blood samples taken from gay men in New York and San Francisco before the existence of AIDS was known.
In the case of the 1918 flu, Worobey is convinced the virus responsible for the pandemic could not have jumped into humans from another host in 1918, a theory that stems from the virus’s explosive behavior once it started transmitting among people. He has already conducted research analyzing eight genes of the virus that points to a two-stage emergence of this virus.
“It was like this little thread that I kept pulling and pulling and pulling and more of this story unfolded,” Worobey told STAT in a recent interview.
He believes a flu virus with an H1 hemagglutinin — the surface protein that gives a flu virus the H portion of its name — likely started circulating among people in the late 1890s or early 1900s and later swapped genes with a avian flu virus to form the H1N1 virus known to have caused the 1918 pandemic.
He also thinks the pandemic H1N1 virus had been sickening people for a few years before 1918, which is not the current view of history.
But how to test the theory?
In his research, Ferguson, the Australian historian, had come across a couple of journal articles published in The Lancet in 1917. One was the account from Rolland and two colleagues about an epidemic at a British military hospital in France. The second was from Dr. Adolphe Abrahams, who wrote to report similar illnesses at a military hospital at Aldershot, southwest of London.
Rolland’s Lancet article piqued Worobey’s curiosity. “Purulent Bronchitis: A study of cases occurring amongst the British troops at a base in France,” as it was titled, described an outbreak of severe respiratory illness at the military hospital at Étaples, in northwestern France.
“Patients suffering from this unusually fatal disease present a symptom complex so distinctive as to constitute a definite clinical entity,” they wrote in the journal article. Many of the men developed cyanosis — bluish skin resulting from a lack of oxygen.
Cyanosis was also widely remarked in reports of patients suffering from the Spanish flu. And a number of scientists and historians have speculated over the years that the outbreak at Étaples in the winter of 1916-1917 was actually part of the pandemic.
Worobey noticed Rolland was identified as a pathologist. Pathologists take and keep tissue samples for teaching purposes, he knew. Were Rolland’s samples preserved and, if so, where might they be?
That’s when Worobey turned to Google. He found Rolland’s obituary and noted he had a son, Charles. More digging turned up the son’s obituary. The warmth and affection in the account of Charles Rolland’s life made Worobey suspect the writer was a relative.
Worobey was on a roll.
“It was about 2 o’clock in the morning at this point and my wife got up to get a drink and said … ‘Go to bed, you weirdo,’” he recalled. “And I was like, ‘No, no, I’m making progress.’”
Within four hours or so of when he started his search, Worobey had found an email address for Cox, who had written the obituary. By the time Worobey’s laptop flipped shut for the night, he felt a bit like a cyber stalker.
In England, Cox was startled to receive the email from a stranger asking three questions. Had he written Charles Rolland’s obituary? Was Charles Rolland the son of Dr. William Rolland, who had written an article in The Lancet in 1917? And were there any pathology specimens of Rolland’s still in existence?
“It was an extraordinary email to receive,” said Cox, who is married to Charles Rolland’s daughter, Fiona. “And I think to his astonishment and mine, I was able to answer yes, yes, and yes.’’
The answer to the last question might well have been no. William Rolland’s widow was not one for hanging on to memorabilia, Cox said; she did not like clutter. But Charles Rolland had rescued the pathology slide collection from his father’s office after his death, and later gave the collection to Cox.
After a dinner in London, and later a visit with Jim and Fiona Cox at their home in the town of Keswick, the couple agreed to let Worobey work with 10 slides in the collection that date back to the respiratory patients at Étaples.
“It was miraculous that these ones were preserved because … these [military] hospitals were decommissioned within weeks of the end of the war,” Worobey said.
Study of the Rolland samples could prove that the Spanish flu pandemic didn’t actually begin in 1918 and didn’t begin in Kansas. The virus that infected Rolland’s soldiers in early 1917, in other words, might have been among the earliest signs of the pandemic.
Historian Barry said he hopes Worobey’s search turns up some evidence, but he’s not convinced, noting that back in 1918 the Étaples outbreak was well-known, but felt to be unrelated to the Spanish flu.
If Worobey finds that’s not the case, William Rolland’s pathology slides will rewrite influenza history. Even if they show that the soldiers were infected with a precursor H1 virus, our understanding of what happened in the lead-up to the Spanish flu will change.
“It would be really interesting to me if this horrible virus was really circulating under the radar for so long before the fall of 1918 when it really had its peak effects,” said Worobey. “It would tell us that there’s a whole lot we have to learn about the forces that keep pandemic viruses simmering at a low level before they explode.”
A theory of the case
There is one other mystery Worobey hopes to crack — the reason for the W-shaped death curve. He already has a theory.
When it comes to flu, there is increasing evidence that the body’s immune system responds best to the virus it first encounters in childhood. The phenomenon is called imprinting. The less related a later virus is from the one that caused that first infection, the less effective the immune response will be in fighting it.
Three decades before the 1918 Spanish flu, the world was hit by a pandemic caused by an H3N8 virus. That virus was likely to have shared no genes with the 1918 virus; furthermore, the viruses in those two pandemics belong to different sides of the influenza virus family tree, meaning that the antibodies generated by the H3 exposures wouldn’t even offer modest “cross-protection” against the virus in the Spanish flu pandemic.
In presentations Worobey gives and papers he’s published exploring the issue, he shows a graphic that makes a compelling case: The people who were born shortly after 1889 were in their late 20s in 1918. They were the people who died in unexpected numbers from the Spanish flu.
People just a little younger than the worst hit seemed to weather the storm more easily, leading Worobey to believe they may have been exposed to an earlier iteration of the Spanish flu virus.
Figuring out which flu viruses were circulating before 1918 should help explain the unusual pattern of severe infections caused by the Spanish flu, and it may also further bolster the imprinting theory.
Might the miraculously conserved Rolland slides deliver up a century-old flu virus? Worobey is trying to keep his expectations in check.
“In all likelihood, we’ll just get no result at all. That’s something to keep in perspective,” he admitted. “ But if we did find something, whatever we found relative to flu would be really interesting and informative.”