Flu Past and Present

Penn medical students raced to treat a flu epidemic in 1918 with few resources. Today, researchers are finding new ways to battle an old illness—because the threat of another major global pandemic is not as far in the past as we might think.

By Katharine Gammon

The summer of 1918 in Philadelphia was hot and sticky. Isaac Starr had just finished the second year of medical school at the University of Pennsylvania. The First World War was raging, but it still seemed far away in Europe.

Starr spent the early summer doing research at the Marine Biological Laboratory at Woods Hole, Massachusetts. Late in August he went on a hiking trip the White Mountains with his father—and during this trip, he first heard a news report about an epidemic of influenza in Spain.

When he returned to campus in September, Starr discovered a city abuzz with activity. A freighter of soldiers was pulling into port, and many of the men were sick. When the medical school started session again, the first lecture was about influenza—a departure from the usual schedule. Unfortunately, Alfred Stengel, MD, the professor who gave the lecture, had no advice for remedies—he didn’t think any existed. “For me and my classmates, knowledge of the disease we were to face so soon was limited to the contents of that one lecture,” Starr recalled in an essay published many decades later in the Annals of Internal Medicine.

A few days after that meager preparation, Starr and his classmates received grave news: An epidemic was judged to be developing in Philadelphia. With so many medical practitioners away in the armed forces, medical students’ services were needed in caring for the sick. School was closed for third- and fourth-year students.

Over the next few days, medical students and city workers constructed a temporary hospital with wooden partitions, based on the skeleton of Philadelphia’s Medical-Chirurgical Hospital at 18th and Cherry Streets, then recently shuttered to make way for construction of the Benjamin Franklin Parkway. The building had five floors, each containing about 25 beds. Then they waited.

Soon the patients poured in. At first, nobody on Starr’s floor was very ill—the patients had fever but little else wrong with them. But over a few days, their illness changed for the worse. The patients’ lungs filled with sticky phlegm, and they became short of breath. “After gasping for several hours they became delirious and incontinent, and many died struggling to clear their airways of a blood-tinged froth that sometimes gushed from their nose and mouth,” Starr writes. “It was a dreadful business.”

Before the flu epidemic ended, the illness swept the globe. An estimated 50 to 100 million people died worldwide—and one-third of the total human population was sickened. Philadelphia, the hardest-hit American city, got the brunt of the illness. Strangely, many of the people who were the sickest were in their twenties, even though, typically, babies and elderly have the worst flu outcomes—so the tragedy was compounded by the virtual loss of a generation, not to war, but to pandemic.

The influenza of 1918 ravaged the world and then went away, without explanation, and without reassurance that a similarly deadly global pandemic could not happen again. A century later, we know much more about the 1918 outbreak, about preventing influenza, and about surveillance that can identify the next pandemic before it takes hold. We also know that there is no use wondering if there will be another outbreak, but only when—and how much better prepared we will be to handle it when it strikes.

The Quest to Intervene

From the start, researchers fought to figure out what the illness was. On Sept. 21, 1918, just days after the first civilian flu cases were confirmed, Paul A. Lewis, MD, of the Henry Phipps Institute at the University of Pennsylvania, claimed to have determined the cause of the disease—a bacterium known as Pfeiffer’s B. influenzae. The Philadelphia Inquirer wrote that Lewis’s findings had now “armed the medical profession with absolute knowledge on which to base its campaign against the disease.” Of course, Lewis was wrong; influenza is caused by a virus. But in the pre-antibiotic era, that knowledge would have made little difference.

At first, life in the city and region went on as normal. With a patriotic fervor to support the troops, a rally for the Fourth Liberty Loan Campaign brought together 200,000 Philadelphians in the city’s streets on Sept. 28. Philadelphia raised vital funds for the war effort—but this success came with a big downside. Within three days of the event, 635 new civilian cases of influenza signaled the beginning of the deadliest period of illness in Philadelphia’s history.

Patients on the hospital wards were gasping for breath and dying. Starr reported for each eight-hour shift at 4 p.m., and found few familiar patients—most had died overnight and been carried away.

Starr Isaac

As a medical student in the fall of 1918, Isaac Starr (right) received just one lecture from Alfred Stengel, MD (left) about influenza before the pandemic struck Philadelphia.

“This happened night after night,” he recalls. He began to wonder if the people responsible for admissions were sending the sickest of the sick to his floor—the top floor. “The deaths in the hospital as a whole exceeded 25 percent per night during the peak of the epidemic,” he writes. “To make room for others the bodies were being tossed from the cellar into trucks, which when filled carted them away.”

Philadelphia’s city morgue, built to hold 36 bodies, was now faced with the arrival of hundreds. Soon, the entire city was quarantined to try and stop the disease’s spread.

The life of the city had almost stopped: Public assembly was forbidden, so there were no plays, movies, concerts, or church services. Schools were closed. Some stores and businesses stayed open, some did not.

By Oct. 4, there were 636 new cases and 139 deaths—just that day. With the city shut down, businessmen started to panic—after all, more cases meant more employee absences and fewer customers. The Bell Telephone Company ran a full-page notice in the newspapers, letting the public know that 27 percent of its operators were absent, and imploring them to avoid calling unless absolutely necessary.

In contrast with the quiet streets and empty buildings outside, Starr and his classmates struggled to keep up with the human tragedy inside the makeshift hospital. Starr had started off thinking of himself as a nurse, prepared to carry out orders from a doctor. To his surprise, he was the only medical professional his patients would see. He was alone in making decisions. While sick patients writhed around him, he made sure to wear a mask, gown, and wash his hands religiously. Very few doctors got sick in the hospital.

It wasn’t easy to find treatments. At the Philadelphia College of Pharmacy and Temple University, administrators decided to suspend classes so that pharmacy students could help fill prescriptions. Most were for whiskey: Since saloons were closed, alcohol was available only in drugstores. People began to try out home remedies like goose-grease poultices, sulfur fumes, onion syrup, and chloride of lime. 

Inside the hospital, the supplies weren’t much better. The hospital had tanks of oxygen but no effective way of administering it. Starr had two ideas for possible treatments: atropine, a nervous system blocker, and camphor oil, a stimulant. Starr was convinced that atropine was worthless, but he thought camphor helped a bit. From time to time, Starr found that a patient’s pulse would pick up after an injection—but the patients soon died nonetheless.

Of course, today’s medical system is completely different from the overcrowded, low-resource system of 1918. Hospitals today have good antibiotics, surveillance, and better supportive care. “So even if a serious virus came about, it’s likely we would see less overall mortality,” says Scott Hensley, PhD’06, an associate professor of Microbiology at the Perelman School of Medicine who studies human antibody responses to influenza and other viruses, and who has taken an interest in the 1918 pandemic. He notes that many of the deaths in 1918 were from secondary bacterial infections; if a similar outbreak happened today, the widespread use of antibiotics would limit mortality.

Liberty Loan parade 1918

The Fourth Liberty Loan Parade brought 200,000 Philadelphians
together on Broad Street south of City Hall just before the influenza pandemic intensified in the city. 

Ebbing Lautenbach, MD, MPH, MSCE’01, chief of the division of Infectious Diseases, Robert Austrian Professor of Medicine, and a professor of Epidemiology, adds that there are systems in place at the local, regional and federal level to identify infectious diseases early enough to deploy containment and prevention strategies. Lautenbach points to the recent example of Ebola. “The high mortality rates in West Africa were due in large part to a lack of public health infrastructure and limited resources in healthcare facilities there,” he says, adding that the high cost of supportive care meant facilities there were not equipped to care for infected patients. For Ebola cases in the U.S., the outcomes were very different. “In a country that has the best of modern medicine, as well as a robust public health system, our ability to screen and identify infected patients and then keep people alive while the body fights a pathogen is much greater,” he says.

Still, in a large-scale pandemic today, the sheer number of sick people could again overwhelm even a strong system. In 1918, the flu sickened around a third of the global population and killed between 5 and 10 percent of those who got sick, notes Gary Kobinger, PhD, a virologist at Quebec’s Centre Hospitalier de l'Université Laval, who is collaborating with a Penn team on a new vaccine strategy. “Even if less than 10 percent of the U.S. population had a severe illness and had to go to the hospital, I don’t know if we would be able to support 30 million people in ICUs.”

If we are lucky, the impact of most common circulating flu strains today should also be reduced by the availability of flu vaccines—but there is still that element of luck. On average, current flu vaccines are 60 percent effective against seasonal flu infections—and they require someone to get revaccinated every year, among other shortcomings (see sidebar, “Future Imperfect Prevention”).

But the shortcomings of seasonal flu vaccines are minor compared to their inadequacy against an emergent pandemic strain, one that might be similar to the flu of 1918: “Current seasonal flu vaccines likely would offer no protection against a new pandemic viral strain,” Hensley says. “A new vaccine would need to be created. During the 2009 H1N1 flu pandemic, a new vaccine was rushed into production but it was too late by the time that the vaccine was available to the public.”

Scott Hensley stands at the site at present day

Scott Hensley stands at the site of the Fourth Liberty Loan Parade at present day (Photo by Peggy Peterson)

Future Imperfect Prevention

Modern public health planning for the flu involves a bit of future-casting: Based on the virus that is circulating in the southern hemisphere in the spring, the Centers for Disease Control and Prevention (CDC) puts together four strains in a shot for that fall’s flu season in the northern hemisphere.

Sometimes, however, that prediction fails to recognize the important strains, as in 2014, when the flu vaccine was only 19 percent effective. Other times, the vaccine fails for other reasons. Hensley recently published research in the Proceedings of the National Academy of Sciences showing that the 2016-2017 seasonal flu vaccine had a poor rate because of a quirk of chicken eggs used to grow the vaccine. The vaccine acquired a mutation that completely changed how the human immune system recognizes it, explains Hensley. “The vaccine itself changed its properties as it was prepared,” he says.

The problem of chicken egg adaptation is not a new one—but in the 2016 vaccine, it was particularly bad. “The virus has evolved over the past years to just grow terribly in chicken eggs,” explains Hensley. “The mutation that occurred last year is sort of a massive one and there’s no way to get around it.” His research found that vaccine grown in insect cells instead of chicken eggs didn’t acquire the mutations.

One Shot, Forever

Universal flu vaccines could circumvent the guesswork involved in making flu vaccines, as well as the need for an annual shot. A universal vaccine could fight all strains, including pandemic strains—for decades. This is a goal that Drew Weissman, MD, PhD, a professor of Infectious Diseases at the Perelman School of Medicine, has in sight.

It might sound farfetched, but Weissman has been creating modified messenger RNA molecules to produce any protein that the body might need. He figured out a way to make the RNA invisible to the immune system, so it could deliver a therapeutic protein to an animal as a form of treatment. Because therapeutic proteins (for example for cancer or anti-inflammatory treatments) are the fastest-growing medicines in the world, this RNA approach has taken off in research in numerous directions.

While doing this work, Weissman found he could take a standard flu antigen and deliver it as an RNA to activate a universal flu response. And the response is large in the body: The titers of antibodies produced in animal models are about 25 times higher than those elicited by the standard vaccine that people get from their doctors, he says. RNA vaccines are also in the works for rabies, HIV, Zika, and some bacterial or parasitic infections.

An estimated 50 to 100 million people died worldwide from the 1918 influenza pandemic—and one-third of the total human population was sickened.

The flu work is still in early stages. Weissman and his team are working on animal trials with very old and young mice and monkeys to see if the vaccine protects them. Another issue to contend with in developing the approach as a potential human vaccine may be scaling up. Right now, his lab can make very small amounts—10 or 20 milligrams—of RNA. To immunize the world, it would take kilograms.

Hensley, who works with Weissman on this project, says he’s excited about the possibilities of RNA-based vaccines and that he’s amazed at high antibody responses the RNA-based vaccines elicit. He also points out that it takes a long time to make our current flu vaccine—but the RNA can be made quickly.

And time matters. No one knows when the next pandemic will arrive, but the experts agree: It’s not a matter of if humans will be hit with another devastating virus, but when. Flu is a fluid, adaptable virus, with reservoirs in pigs and birds, so there’s no telling from where the next virus will pop up. Influenza infections are the seventh leading cause of death in the U.S. and result in almost 500,000 deaths worldwide per year, according to the CDC.

The current technology to create flu vaccines—chicken eggs—takes 8 or 9 months to get a shot out to people. And that’s not fast enough, says Kobinger, the Canadian virologist. “If we have a new strain emerging, within three weeks it will be all over the continent, based on what we learned from H1N1 in 2009,” he says. Luckily, that outbreak was not a particularly deadly one—but it was lightning quick, spreading in weeks in North America and in three months around the world. “So how would you provide a vaccine in three weeks to have an impact on the first wave?” asks Kobinger. If a more severe virus came along that was equally swift, he says, “it would be catastrophic.”

Targeting the Nose

As quickly as the 1918 influenza outbreak began, it began to subside. After weeks of misery, Starr watched as the patients’ deaths on the top floor of the hospital started to wane. By the end of October, the number of patients decreased, public places reopened, and quarantines were lifted.

The type of flu was also milder as the weeks wore onward. “So, as mysteriously as it had come, the killer departed,” he writes. After about five weeks of working in the clapboard, temporary hospital surrounded by bodies, medical students went back to books and rotations. Slowly, life returned to normal. By the spring of 1919, it was estimated there were 12,191 flu deaths in Philadelphia alone—out of a population of 1.7 million.

"Current seasonal flu vaccines likely would offer no protection against a new pandemic viral strain."– Scott Hensley, PhD

But as the years and decades have unfolded since then, physicians like Starr, who earned his medical degree at Penn in 1920 and went on to join the faculty and served as dean after World War II, remained aware that another pandemic could occur and contemplated how, why, and where. Historians have looked back at the massive public gathering at the September 1918 Liberty Loan parade as one likely contributor to the spread of disease in Philadelphia. Perhaps they looked at other public gatherings in the aftermath and considered that, with every cough, sneeze, and droplet that flew through the air, another possible virus was upon them. We are never truly free from the threat of another viral pandemic.

But while droplets carry viruses, they also carry information—and they might be one way to stop future pandemics.

“It seems crazy that we’re developing a systemic response to block something like the flu—an infection around the nose,” says James Wilson, MD, PhD, director of the Orphan Disease Center at Penn, who is working on a gene-therapy flu vaccine that elicits a faster immune response than traditional vaccines in part because it intercepts the actual path of the virus. The infection gains entry into a body through breathing in someone’s cough or sneeze. You may not get sick for many days—but the virus is slowly amplifying in the nose. Eventually the virus gets to your lungs by getting inhaled through your nose.

“Our strategy was developed to prevent the virus from gaining entry into the lungs,” explains Maria Limberis, PhD, a research associate professor and executive director of the Comparative Medicine Program in the Penn Gene Therapy Program, who initially began working with Wilson as a postdoctoral fellow in his lab. “These viruses replicate quickly to bypass the immune system, infect the lungs, and cause disease.”

Spanish Flu

Public health officials knew in 1918 that droplets of saliva and
sneezes would spread disease.

This is a comparatively new entry strategy for Wilson, who had been working on gene-therapy approaches to fight HIV and other viruses for years before considering flu. The big idea is to take the gene encoding a therapeutic protein, clone it into an adeno-associated virus (AAV) vector, and inject the vector. That would program a patient’s cells to express the therapeutic protein. AAV is a huge change in the way vaccines work because it programs non-immune cells to express antibodies against a pathogen.

Then Bill Gates, who had taken a personal interest in Wilson’s work because of his long-standing desire to battle HIV, stepped in. He asked the researcher a provocative question: Could AAV be used to prevent flu? Wilson started to think about it. The problem was, muscle and liver cells, which he had targeted with gene therapies for HIV and other blood-borne viruses, wouldn’t work against the flu because it spreads through the air. Targeting airway cells seemed like a good pivot of the technology. Wilson enlisted Limberis to collaborate on the project because she was experienced with using AAV vectors in epithelial cells like those lining the nose from her work in gene therapy for cystic fibrosis. Kobinger, the Canadian virologist who also completed a postdoc in Wilson’s lab, rounded out the team.

Early Exposures

 The Great Equalizer

In the century that has transpired since the influenza pandemic in 1918, there is one enduring mystery for which our answers are incomplete. Why were so many young adults killed by the virus? That was an abnormal age distribution for an illness that usually kills babies and the elderly.

Scott Hensley has an idea of why that happened. When analyzing the pandemic H1N1 flu (another swine flu) from 2009—a particularly bad flu in terms of virulence—his lab found most people had some H1N1 immunity. But when they dissected the specific antibodies within individuals, the team found something strange: People of different ages mounted different types of immune response that recognized the virus in different places. Those responses were based on the type of flu that each person had first encountered in childhood.

"There is something magical about childhood," Hensley says. "Many different B cells [responsible for creating antibodies] are activated during initial childhood infections and some of these differentiate into memory B cells that hang around a long time. And when we're infected later in life, these memory B cells become reactivated and dominate our responses against new viruses." As a result, a person's antibody response narrows over time, and that can be dangerous: Single mutations in an evolving virus can prevent antibodies from binding.

This might explain the disproportionate deaths of young adults during the 1918 flu.

It's possible, says Hensley, that people born around the 1890s were exposed to a virus in childhood that made them more susceptible to the 1918 flu. He adds that children today get their first flu exposure from a vaccine and not from the live virus, so it's an open question how this first exposure will impact their antibody response in the future.

The problem of chicken egg adaptation is not a new one—but in the 2016 vaccine, it was particularly bad. “The virus has evolved over the past years to just grow terribly in chicken eggs,” explains Hensley. “The mutation that occurred last year is sort of a massive one and there’s no way to get around it.” His research found that vaccine grown in insect cells instead of chicken eggs didn’t acquire the mutations.

In 2012, the team started the work with a simple experiment. They took several known flu antibody sequences, cloned them into one of the AAV vectors, injected the vectors into mice, and challenged them with a common H1N1 flu.

Limberis remembers telling Wilson the experiment didn’t work—because the treated mice survived and the non-treated mice didn’t, and that was too perfect. “I couldn’t believe it would work so efficiently, especially the first time we tried it.” The team eventually reproduced the results and Limberis was convinced.

No one knows when the next pandemic will arrive, but the experts agree: It's not a matter of if humans will be hit with another devastating virus, but when.

From there, the team started working on more antibodies. They wanted to find out if their results were just an academic success, of if they could have applications in the real world. The team tested the technology against lethal doses of clinical diseases like H1N1 and a strain of the 1918 flu that had been reconstituted from human tissue by a member of the team Kobinger then headed in Winnipeg.

In 2013, they published the results in Science Translational Medicine. The mice and ferrets who received a single dose of an AAV vector expressing a broadly neutralizing flu antibody into their nasal passages were protected from the viruses, and the untreated animals were not.

Their work began to get attention from federal defense agencies for its potential application to protecting from bioweapons, and they collaborated with several programs, trying to find the ideal, broadly active flu antibodies to take the AAV technology into clinical trials. This team has recently formed a collaboration to license this technology to Janssen Pharmaceuticals, Inc. and is rapidly moving toward Phase 1 trials utilizing AAV vector gene therapy to deliver Janssen’s proprietary anti-influenza antibody. The trial will target people over the age of 65, who have a particularly bad outcome with flu—and for whom existing vaccines are only 20 percent effective.

Like Weissman’s RNA vaccine, the AAV vaccine is a hoped-for universal vaccine—one that might work against most or all flu strains without requiring the annual guesswork involved in predicting regular annual flu strains. “If this antibody was effective against every major strain in the last 100 years, it should be effective in the future,” Wilson says.  

Another advantage of the AAV vector vaccine is how quickly it works: Normally, it takes 2 to 3 weeks for a person’s immune system to get activated—but this method expresses antibodies within 24 hours.

If safe and effective, the AAV vector vaccine could thus also subsequently be considered as part of a pandemic flu response for the healthy population. (But don’t look to this work to create protection forever. When it goes into the nose and targets epithelial cells, it lasts 4 to 5 months in monkeys for safety reasons, Wilson says.)

Using AAV as a vehicle, a vaccine could get people protected against influenza within weeks rather than months required to grow traditional vaccines in chicken eggs. Kobinger explains that once researchers identified a new strain, they could sequence it within a week. And with that genetic information, scientists could make antibodies very quickly. “In theory, you get protected within two months rather than six or eight.” He could envision a future where someone could go to the pharmacy and get a dose to inject themselves within a few weeks of the start of an outbreak. This method could help in urgent pandemics beyond flu like Ebola, SARS, MERS—any disease that’s spread through the air.

Wilson Limberis

James Wilson, MD, PhD, and Maria Limberis, PhD, are testing gene therapy in the nose
to block inhaled flu viruses from taking hold.
(Photo by Peggy Peterson)

“There is good reason to believe that a future epidemic could be handled much more effectively than was the last,” Starr wrote in his Annals of Internal Medicine essay recalling his ordeal in 1918. He penned the piece in 1976, when the U.S. was facing the threat of a novel swine flu, feared at the time to be the next great pandemic. He notes that boring-but-important efforts like hand washing and gowns helped medical professionals stay healthy. And antibiotics and supportive care would keep many more people alive in future pandemics. Even more innovations that Starr could not yet imagine—RNA vaccines and AAV vector vaccines among them—still lie ahead. These could rapidly confer immunity against more flu strains and change the game—for people now and in the future.

But the big question for the future that no one can answer is when the next pandemic will happen.

“All bets are off the table because we can’t forecast this,” says Hensley. “We don’t know how pathogenic the next strain will be. We have come a long way but have much more still to learn if we want to be better prepared for the next pandemic.”

Read more about Isaac Starr's story on the Penn Medicine News Blog.


Thank you for reading! Please take a moment to complete a brief reader survey about Penn Medicine magazine.

envelope  Sign up here to subscribe to the email edition of Penn Medicine magazine.

Share This Page: