Autopsy of a Deadly Virus

Jeffery K Taubenberger. Natural History. Volume 125, Issue 8, September 2017.

Influenza viruses were unknown before 1918, the advent of the Spanish flu pandemic. Indeed, viruses of any kind, too small to be seen, were only beginning to register on scientists’ radars. Virus was just a word for a “filterable” disease agent, conceivably a liquid, that could pass through the finest of filters. Bacteria did not pass through such filters, and they could be seen with the available microscopes. The prevailing view was that the cause of the Spanish flu was a bacterium dubbed Pfeiffer’s bacillus, now known as Haemophilus influenzae. It often (though not always) could be detected in the respiratory tracts of the victims. Support for this explanation began to wane during the pandemic in favor of a virus, but it was not until the early 1930s that influenza A viruses-to which Spanish flu belonged-were isolated, first from swine and then from people.

Influenza virus infections in humans are generally acute (as opposed to chronic) and self-limited (run their course without specific treatment). However, severe symptoms of influenza, including hemorrhagic bronchitis, diffuse alveolar damage, and pneumonia, can develop within hours. Fulminant fatal influenza viral pneumonia occasionally occurs. But pneumonias associated with flu are usually caused by secondary bacterial infections.

Most people infected in the 1918 pandemic had typical, self-limited influenza, but an unusually high percentage experienced more severe disease and fatal outcomes. More than 2.5 percent of cases resulted in death, compared with a fatality rate of less than 0.1 percent in other influenza pandemics. Total global deaths are now estimated to have been about 50 million and may have been as high as 100 million. The progression to severe disease likely involved a combination of viral, host, and bacterial factors, interacting in ways we may never fully understand. But new insights have been gained by examining the genetic sequences of the 1918 virus and in pathogenesis studies using experimental animals.

Based on their genetic structure, human influenza viruses are classified as A, B, and C. Influenza type C viruses usually cause only mild infections and do not lead to epidemics. Our ordinary winter flus are caused by A and B viruses; in either type, a new variant can arise that sparks a dangerous seasonal or annual epidemic, but only A viruses may cause a pandemic. That is because B viruses are found only in people, while A viruses naturally infect hundreds of warm-blooded animal hosts, both birds and mammals. Their major reservoir is in wild aquatic birds, whose intestinal tract they usually infect without causing significant disease symptoms. Within that reservoir the viruses display enormous genetic diversity and can switch hosts, ultimately adapting to new host species, including domestic poultry, swine, horses, dogs, and humans.

We now know the 1918 virus was an influenza A virus of the family Orthomyxoviridae. For the technically minded, these are “enveloped, negative-sense, single-stranded RNA viruses with segmented genomes”; the takeaway is that such viruses are sophisticated enough to suppress the innate immune responses of the hosts they infect.

Influenza A viruses encode at least thirteen proteins, including two major surface proteins: hemagglutinin (abbreviated HA, by itself, or H in virus names), and neuraminidase (NA). Eighteen different HA and eleven different NA subtypes are known, with sixteen of the HAs and nine of the NAs consistently found in avian hosts in various combinations. These are labeled, for instance, H1N1 or H3N2. These combinations can arise in a host with mixed infections, in a process called reassortment. Within the many subtypes of influenza A virus, biologists can also identify and trace particular strains.

To reconstruct the 1918 pandemic viral genome, my team isolated fragments of viral RNA taken from the tissues of several original victims. In addition to sampling lung tissue from preserved autopsies, we were able to exhume (with community permission) several bodies from a mass burial in Alaska, where they had been preserved by permafrost. I should note that to protect the researchers and to prevent the escape of pathogens, our investigations are carried out in “high containment” labs engineered for handling dangerous pathogens. All work with the 1918 virus requires special United States government approval, continuous oversight, rigorous training, background checks, and so on.

Our analysis demonstrated that the causative agent of the 1918 pandemic was an avian-like H1N1 virus. As no human pre-1918 influenza A virus sequences are currently available, the origin of that virus, including timing of its emergence in humans and whether an intermediate host was involved, remains unresolved.

We have learned that the 1918 viral RNA polymerase-the protein that allows the viral RNA genome to be copied-is very similar to the polymerase of avian influenza A viral strains. We have also learned that it allows the virus to replicate to high concentrations. This would have contributed to the virulence of the 1918 virus but was not its principle basis. The main culprit was the virus’s HA surface protein. When we took a human influenza virus that ordinarily was harmless to mice (our experimental animals), and altered it to express the 1918 HA, it proved lethal. High concentrations of the virus replicated all along the respiratory tree, from the upper airways to the air sacs in the lungs, and the infection resulted in the death of surface epithelial cells throughout the respiratory system

Significantly, viruses expressing the 1918 HA surface protein also triggered a prominent inflammatory response from the host. The respiratory tree was infiltrated by inflammatory material largely consisting of neutrophils, white blood cells that kill bacteria and cells infected by viruses. This neutrophilic infiltrate caused further damage to the cells of the respiratory system and contributed significantly to lung injury. Tellingly, drugs that block neutrophil damage but do not affect virus replication proved effective in treating mice given a lethal dose of the 1918 virus, demonstrating that the host’s inflammatory response alone contributes significantly to disease progression following infection with a 1918 virus.

A third major reason why the 1918 influenza virus was so pathogenic was bacterial co-infection. Bacterial pneumonias following influenza virus infection were instrumental in the deaths of over 95 percent of 1918 flu victims. The viral infection contributed directly to the death of epithelial cells lining the airways of the respiratory tree, exposing the basal cells needed to repair and regenerate the damaged tissue. This gave bacteria, usually streptococcal or staphylococcal species, the opportunity to invade the lower respiratory tract and cause a serious pneumonia. In addition to that injury to the lung lining, co-infection with Streptococcus pneumoniae caused unique damage to the vascular system of the lungs. Autopsies conducted at the time of the pandemic revealed widespread blood-clot formation in lung tissues.

All three factors-pathogen virulence, host inflammatory response, and secondary bacterial infections-together help explain the extraordinary severity of the 1918 pandemic.

Although its origins have not been fully traced, the 1918 virus’s subsequent evolution is better understood. By the early 1920s direct descendants of the virus, evolving in response to population immunity, began to circulate in a pattern of seasonal endemic recurrences, and they continued to do so for nearly forty years. Then, in 1957, a 1918-lineage H1N1 virus reassorted with an avian H2N2 virus to cause a new pandemic. A subsequent reassortment of that new virus incorporated an H3 component to cause a pandemic in 1968. And in 1977, 1918-lineage H1N1 viruses returned to global circulation and co-circulated with H3N2 viruses.

During the 1918 pandemic, humans likely infected pigs with the H1N1 virus, and these classical swine influenza viruses continue to circulate today. This swine branch of the 1918 H1N1 lineage also underwent additional reassortment events with avian and other swine strains, and ultimately led to the emergence of an H1N1 virus that was pandemic among humans in 2009.

It is remarkable that direct descendants of the 1918 virus still circulate in swine and were pandemic in humans as recently as 2009. In addition, for the past century the founder 1918 virus and its progeny have continually donated genes to new influenza A virus strains to cause new pandemics, epidemics, and epizootics. In that sense, all cases of influenza A virus worldwide are connected to the 1918 virus, including H1N1, H2N2, and H3N2 viruses. The H3N2 are composed of key genes from the 1918 virus, updated by subsequently incorporated avian influenza genes that code for novel surface proteins. The exceptions are sporadic human infections from avian viruses, such as H5N1 and H7N9.

In a sense, the single founding virus that caused the 1918 pandemic can be held responsible for tens of millions of deaths from influenza in the century since the pandemic, in addition to the tens of millions who died during the pandemic itself.

Given our better understanding of why the 1918 virus was so devastating, it is critically important to recognize that other currently circulating avian influenza A virus strains share many of their features with the 1918 virus. It is possible that a future pandemic with the potential to cause morbidity and mortality on a 1918 scale could develop.

Avian strains expressing a handful of the eighteen known HA subtypes-H1, H6, H7, H10, and H15-share virulence properties with the 1918 virus. One is the current avian H7N9 strain, whose outbreaks have caused approximately 1,500 human infections in China since 2013, with a fatality rate of 38 percent. Ultimately, as we continue to probe the biological mysteries of the 1918 virus, it is crucial to apply the lessons learned to enhance surveillance and develop new therapies for severe influenza. A further goal is to move beyond the ever-changing annual flu shots and create broadly protective “universal” vaccines that could prevent any future pandemic.