杏吧原创

Imagine a disease as deadly as Ebola and as contagious as the common cold. Imagine it kills by punching holes in your lungs, starving you of oxygen so that your face turns a dark purple, and your feet, black. Now imagine that there鈥檚 no sure-fire cure.

Imagine all that, and you have some idea of just how destructive the flu virus can get. Three times in recent history, the virus-one of the simplest life-forms on the planet-has mysteriously transmogrified from a mere purveyor of aches and sneezes into a superbug that left ghost towns and mass graves in its wake. In the worst of those pandemics, Spanish flu swept the globe in 1918, killing roughly 40 million in a single year-only to disappear as suddenly as it had emerged, leaving no clues to what had made it so virulent.

Until now, that is.

In the past few years, an unlikely partnership between an expert on dolphin diseases and a retired pathologist has brought Spanish flu back from the grave. What they鈥檝e found has sent shivers down the spines of the flu experts who are trying to protect the planet from the next big pandemic. Not only are the textbook theories on what makes the flu virus turn nasty probably wrong, but predicting when the next lethal strain will strike looks tougher than anyone ever imagined. So great is the concern that in a bold attempt to beat the bug, flu specialists are even thinking the unthinkable: recreating a killer strain of the virus in the laboratory to see if they can find its weak spot.

For now, only one thing is certain-killer flu will strike again. 鈥淚t鈥檚 like the San Francisco earthquake, we know it鈥檚 going to come,鈥� says Graeme Laver, a virologist at the Australian National Laboratory in Canberra, who developed the anti-flu drug Relenza.

The flu virus鈥檚 success is due to its mutability (see 鈥淰iral sex鈥�). Each strain of the virus mutates at breathtaking speed. On top of that, every so often a whole new strain of flu jumps into humans from birds, or pigs that get it from birds. Those new strains can trigger human pandemics, but as the natural reservoir for influenza, the birds don鈥檛 normally get sick. Sometimes, however, the flu virus kills birds by making use of glitches in the genes that code for two proteins-neuraminidase (NA) and haemagglutinin (HA).

Those two proteins make up the large spikes on the surface of the virus. Inside a lung cell enzymes cut up those spikes, enabling the virus to escape to the next cell. When the NA and HA proteins contain the lethal glitches, enzymes from the liver, spleen, heart and other organs can also break down the spikes. This means the virus can attack almost every organ in the bird鈥檚 body, causing massive internal bleeding and death. Most flu experts had assumed that Spanish flu must share the same lethal mutations.

For the past few years, attempts to confirm that hypothesis have focused on exhuming the bodies of 1918 flu victims and extracting the virus鈥檚 genetic blueprint from their tissue. The exercise is fraught with difficulties. Intact tissue from corpses interred for 80 years is hard to come by. And while most creatures carry their genes in a sturdy double strand of DNA, the flu virus carries its genes on eight fragile single strands of RNA that disintegrate very easily.

When it comes to cracking Spanish flu鈥檚 secrets, however, Jeffery Taubenberger of the Armed Forces Institute of Pathology in Washington DC has a distinct advantage. He is a world expert in the arcane art of recovering genetic information from preserved or rotting tissue. Taubenberger learned his craft tracking down the virus responsible for the bloated, rotting dolphin carcasses that have been washing up on the coasts of North America since 1987.

Taubenberger, along with his technician Ann Reid, began their search for the killer flu in 1995 at the US National Tissue Repository at the AFIP. Originally built as a bomb shelter, the repository is home to row upon row of metallic shelves, stacked high with small brown cardboard boxes. Inside each box are thumbnail-sized pieces of human tissue preserved in formaldehyde and encased in blocks of transparent wax. There are tens of millions of samples, all taken from autopsies on soldiers as far back as the American Civil War in the 1860s. It鈥檚 the world鈥檚 largest library of illness and death.

A lung tissue sample from Private Roscoe Vaughan, who was 21 when he succumbed to Spanish flu, turned out to have exactly the right pathology. Under the microscope, his right lung tissue was pitted with lesions where the virus had replicated nearly 80 years earlier. 鈥淣o one had ever recovered RNA from a sample so old,鈥� says Taubenberger. 鈥淲e were lucky.鈥�

Mass grave

What Taubenberger and Reid saw when they first glimpsed the RNA sequence of the Spanish flu virus was so remarkable that they rushed out a report on the basis of that one patient (Science, vol 275, p 1793). The genes for both the NA and HA surface proteins in the soldier鈥檚 virus were completely normal. Taubenberger also found no trace of the virus in any part of the soldier鈥檚 body, apart from the lungs. Bang went the notion that mutations in the NA and HA genes were Spanish flu鈥檚 calling card. Still, it would take more than one victim to convince the establishment to abandon their favourite hypothesis.

Luck was on Taubenberger鈥檚 side. A retired pathologist called Johan Hultin heard about the research. Forty-six years earlier, Hultin had gone to Brevig, a missionary station in Alaska, to collect tissue from 1918 flu victims buried in the permafrost. Seventy-two of the 80 missionaries had died of the flu, and the remaining eight had buried them in a mass grave. Like everyone else, Hultin wanted to know what type of virus was capable of such devastation. He didn鈥檛 succeed then. But in 1997 Hultin returned to Brevig.

Among the many skeletons, Hultin discovered one well-preserved woman. Lucy, as he named her, had been fat, and 鈥渋t was the fat that kept her organs in near-perfect condition,鈥� he says. He removed both her lungs, as well as tissue from the liver, spleen and kidneys, and mailed them to Taubenberger in a plain brown box.

Back in the lab, Taubenberger and Reid got ample RNA from Lucy, and two other cases they鈥檇 found at the repository. There were no mutations in any of the NA and HA genes. 鈥淭he supervirulence of 1918 can鈥檛 be explained by a simple genetic variation in the surface proteins,鈥� Taubenberger says.

Then last month, Rod Daniels of the National Institute of Medical Research in London told New 杏吧原创 that he too could find no mutations in the NA and HA genes. Daniels was the virologist on a team that travelled to Norway in 1998 to retrieve tissue from frozen flu victims (New 杏吧原创, 8 June 1996, p 11). 鈥淲e all thought we knew the answer,鈥� says John Oxford of St Bartholomew鈥檚 Hospital in London, who works with Daniels. 鈥淏ut we didn鈥檛. That was the biggest surprise of all.鈥�

It was also bad news for the World Health Organization and the Centers for Disease Control and Prevention in Atlanta. The WHO and the CDC coordinate world surveillance of human flu viruses, looking out for anything new and dangerous. The trouble is they only look for genetic changes in the HA and NA surface proteins. On one level that makes perfect sense. HA and NA are the bits of the virus that the immune system can easily spot, and for that reason they constantly mutate to help the virus stay one step ahead. Knowing those changes allows vaccine makers to redesign each season鈥檚 vaccines to recognise the flu strain that is in circulation. But Taubenberger鈥檚 findings suggest that this kind of surveillance won鈥檛 help spot the next Spanish flu.

Taubenberger believes that the most likely culprit for the 1918 pandemic is a mutation in the polymerase gene in the core of the virus. When a new flu virus enters the human population, it must quickly adapt to its new host or die out. To adapt, it needs to replicate and mutate. Polymerase genes make the proteins that copy the virus鈥檚 genetic material, so the virus can replicate. A mutation in the polymerase gene could trigger a lot of genetic mistakes. Some would undoubtedly be fatal to the virus, but others could help the virus adapt to its new environment and become supervirulent.

Fast and deep

Such a change in the polymerase gene could explain why the 1918 strain appeared to multiply much faster than normal, says Taubenberger. 鈥淭his was a very fit virus. It reached deep into the lungs where regular flu never goes.鈥�

One way to find out if he鈥檚 right could be to compare the 1918 polymerase gene with the polymerase genes in the flu strains that caused the 1957 and 1968 pandemics. Which is exactly what Taubenberger and Reid intend to do when they finish sequencing the gene some time next year.

In the meantime, they are making full use of the 50 per cent of the genome they have. By scrutinising the make-up of surface proteins of the flu, and comparing them with hundreds of flu viruses from the 1930s onwards, the two researchers have discovered the possible origin of Spanish flu.

Unlike the 1957 and 1968 strains, which are related to bird viruses, Spanish flu is related to pig and bird virus, suggesting that it may have entered the human population from pigs. Reid points out, however, that the possibility that humans infected pigs during the 1918 pandemic hasn鈥檛 been ruled out (New 杏吧原创, 27 February 1999, page 27). 鈥淲e need to figure out the animal source of the 1918 pandemic. We need more information about how flu viruses move between and among species. And we need increased surveillance of wild waterfowl, domestic birds, pigs and horses,鈥� says Taubenberger.

Most flu experts agree that the only way to nip a pandemic in the bud is to know it鈥檚 coming. In Europe and the US-the possible birthplace of the 1918 pandemic-surveillance relies on farmers or vets spotting an outbreak of pig or chicken flu, and reporting these outbreaks to the CDC and the WHO.

In China, where the last two pandemics originated, the surveillance is more sophisticated. After the 1997 chicken flu outbreak in Hong Kong, the WHO increased funding to China so that the authorities could routinely test the droppings of healthy farm animals for different flu strains. That information is used to draw maps of flu trends to make it easier to spot a strain that has the potential to trigger a pandemic. But, argues Laver, more farms should be screened, as well as other species, including wild ducks and birds. 鈥淐hina is the hot spot, but it鈥檚 also a black box,鈥� he says. 鈥淲e don鈥檛 even know where to start looking for deadly strains, much less find them.鈥�

Virologist George Brownlee at the University of Oxford sees another way of using the 1918 genetic code to work out what made the virus so lethal. Once each of the eight gene segments have been sequenced, he says, it will be possible to recreate viral sex in the lab, inserting one segment at a time into an everyday human flu strain. By examining the hybrid viruses, or testing them in animals, you might work out which segments make the virus supervirulent. 鈥淸These experiments] are necessary before the CDC can change the way it monitors influenza,鈥� says Nancy Cox, chief of the Influenza Branch at the CDC.

Reconstructed killer

Robert Webster, a virologist who studies flu pathogenesis at St Jude Children鈥檚 Research Hospital in Memphis, has an even scarier proposition. He suggests recreating the complete 1918 virus in the lab. That idea will be discussed at a meeting at the National Institutes of Health near Washington DC this autumn. 鈥淯ltimately, someone will take it upon themselves to reconstruct the virus,鈥� predicts Oxford.

Even with proper containment there would be a tiny-some say negligible -risk of the virus escaping and wreaking the very death and destruction the experiment was designed to avert. But Webster and Brownlee maintain that recreating the virus and testing it in animals may be the only way to solve the riddle surrounding the Spanish flu. To this day, Brownlee point outs, our understanding of how the virus kills is practically nonexistent. Does it trigger a fever so extreme that the vital organs shut down? Or does it merely weaken the lungs so that that bacterial pneumonia can deal the death stroke? No one knows.

There may, of course, be one other way to study pandemic flu viruses-wait for the next one. 鈥淭o find out more about the cause of flu pandemics, including the 1918, we need another one,鈥� says Oxford. 鈥淯ntil then, we all wait with baited breath.鈥�