In 1954, a mysterious disease struck children in Manila. They were showing up at hospitals with internal bleeding. Their blood vessels were leaking.

Over the next few years, similar outbreaks cropped up every rainy season. And then in 1958, a massive outbreak hit Bangkok. More than 2,500 children were hospitalized. About 10 percent of them died.

That year, an American doctor, working on polio in Southeast Asia, began searching for the culprit. Eventually, he isolated a mosquito-borne virus — dengue — and, in the process, launched a 60-year-old medical mystery.

The four dengue viruses originated in monkeys and independently jumped to humans in Africa or Southeast Asia between 100 and 800 years ago. Dengue remained a relatively minor, geographically restricted disease until the middle of the 20th century.

Known as "breakbone fever" because of the joint pain it can bring, dengue had been causing for problems for decades, maybe even centuries. But it rarely caused hemorrhaging or death. Why had it all of a sudden become so dangerous? Had the virus mutated? Was their an additional virus — or environmental factor — boosting dengue's potency?

Or did a previous infection with dengue somehow make a person more vulnerable to this deadly form?

Over the past 50 years, scientists have accumulated more and more data pointing to the last hypothesis. Now a study, published Thursday in the journal Science, finally appears to nail down why.

"This is a rock-star study," exclaims Jean Lim, a virologist at Icahn Medical School at Mount Sinai, who wasn't involved in the study. "I think it will be a benchmark paper."

Of course, some scientists in the field are still skeptical and want still more proof, as Science's Jon Cohen reported.

But the findings also open up some intriguing theories about why Zika became such a threat in certain places of South America.

In the study, researchers at the University of California, Berkeley, followed about 6,600 children in Nicaragua where dengue circulates. For 12 years, the researchers drew the children's blood annually and measured their concentrations of dengue-binding antibodies — molecules the immune system makes to destroy viruses.

Then the researchers looked for connections between severe dengue cases and antibody levels.

"If a child developed dengue, we could go back to the banked antibody samples and say, 'OK, is there something about the child's antibody levels that are different than that of the healthy kids?' " says Eva Harris, an infectious disease researcher at the University of California, Berkeley, who led the study.

Now you would expect the presence of dengue antibodies in the kids' blood would protect them from new dengue infections. That's what antibodies do, after all! If you get infected with a virus, your immune system makes antibodies, which then hang out in the blood and fight off new infections of the same virus.

But with severe cases of dengue, the exact opposite turns out to be true, Harris and her team find. The antibodies actually backfire.

When antibody levels in the children's blood fell into a particular range, kids had an increased risk of developing hemorrhagic dengue.

And the effect was big. Inside the antibody "danger zone," kids were more than seven times more likely to develop severe cases of dengue compared to children who have never been infected with dengue.

In other words, prior dengue infection actually primes the immune system in a way that makes the next infection worse, if the antibody levels have fallen within that window.

When antibody levels were higher than the "danger zone," they didn't raise a child's risk but they also didn't decrease it, the study finds. "Actually having higher antibodies wasn't helping either," Harris says.

Scientists first proposed this idea — called antibody-dependent enhancement — back in the 1970s. And since then, many studies have demonstrated the phenomenon inside petri dishes and animal models. Researchers even have hypotheses about how the antibodies help the virus enter immune cells and eventually damage blood vessels.

"The problem is that you can see antibody-dependent enhancement with many viruses inside test tubes," Harris says. "And it's been unclear how these experiments translate to dengue in humans. That's been the crux of the problem."

"The study answers this age-old question about dengue," Harris says, "but raises new questions about Zika."

You see, dengue and Zika are closely related. And antibodies made for dengue can also bind to Zika (and vice versa).

So the big question: Is there a "danger zone" of dengue antibodies that would worsen a Zika infection? Or help the virus enter placental cells and eventually infect the fetus?

Copyright 2017 NPR. To see more, visit http://www.npr.org/.