One Vaccine. Two Antibodies. An Entire Viral Family Still Waiting.
We have a licensed Ebola vaccine. We have licensed Ebola antibodies. We have shown that modern medicine can dramatically reduce mortality from one of the most feared pathogens on Earth. Yet today’s outbreaks are caused by the wrong species of filoviruses and we are largely back where we started.
Filoviruses are a family of thread-shaped RNA viruses that kill with a brutality that is almost unmatched in the natural world. Ebola Zaire, the most lethal species, has killed as many as 90% of the people it has infected in some outbreaks. Marburg, discovered in Germany and Yugoslavia in 1967 when laboratory workers were exposed to infected monkeys imported from Uganda, has hit 88% in certain outbreak settings. These are not numbers that should invite complacency.
Filoviruses are not one virus. They are a family. Building a countermeasure against Ebola Zaire does not automatically produce protection against Sudan, Bundibugyo, or Marburg because the viral glycoproteins differ substantially. The challenge is analogous to developing a highly effective key for one lock and discovering it doesn’t fit the others.
And yet, if you survey the full landscape of this viral family today, you find the following: one licensed vaccine for one species, one licensed therapeutic antibody for that same species, one approved antibody cocktail also targeting only that species, and nothing — not a single approved countermeasure — for Marburg, Ebola Sudan, Ebola Bundibugyo, or Ebola Tai Forest. Viruses capable of killing more than half the people they infect.
The 2026 DRC outbreak, now over 1,500 cases and still spreading, is caused by Bundibugyo ebolavirus — a species with no licensed countermeasures whatsoever. Uganda has confirmed Marburg cases alongside Ebola cases. These outbreaks are burning right now, in real time, with nothing purpose-built to stop them.
How did we get the tools we have, and why do we have them only for one species?
The Cold War built the first case for taking filoviruses seriously as a weapons threat
To understand where filovirus countermeasures came from, you have to start not with a public health crisis but with a military intelligence discovery.
The Soviet Union ran an enormous, secret biological weapons program called Biopreparat for most of the Cold War, in direct violation of the 1972 Biological Weapons Convention — a treaty the Soviets had signed. Defectors who came forward after the Soviet collapse described a program that, at its peak, employed tens of thousands of scientists working across dozens of facilities. They were not working on defensive measures. They were developing agents intended to kill, including, according to multiple accounts, experiments aimed at weaponizing Ebola.
The realization that state actors had been weaponizing agents like anthrax and researching filoviruses changed the strategic calculus. These were no longer theoretical public health concerns. They were national security threats.
Non-state actors were engaged too. Aum Shinrikyo, the Japanese doomsday cult that carried out the 1995 Tokyo subway sarin attack, had earlier dispatched members to active Ebola outbreaks in Africa with the explicit goal of collecting blood from infected patients for use in a weapon. They failed to turn it into a viable agent, but the attempt was real and documented.
Then came September 11, 2001, and weeks later, anthrax-laced letters were mailed through the U.S. postal system. Five people died. The government was forced to confront a national security gap it had largely not addressed: there were almost no medical countermeasures for the biological threats its own intelligence community had identified as genuine dangers. Anthrax, smallpox, plague — and Ebola.
The response was to build an infrastructure specifically designed to fill those gaps. Congress created Project BioShield. BARDA — the Biomedical Advanced Research and Development Authority — was established under the 2006 Pandemic and All Hazards Preparedness Act as an advanced development agency, its explicit purpose being to bridge the "valley of death" between early-stage research and licensed products. Pharmaceutical companies do not naturally invest in countermeasures for threats that may never materialize into commercial markets. The government needed to substitute as buyer and funder.
The ZMapp monoclonal antibody cocktail — the experimental treatment that made international news when it was administered to infected American missionaries Kent Brantly and Nancy Writebol during the 2014 West Africa outbreak — came directly out of this biodefense investment pipeline. The antiviral and antibody candidates that entered trials during 2014 to 2016 were the products of a funding stream that traced its origins to the post-9/11 bioterrorism recognition that Ebola could be used as a weapon. The market incentive that produced these medicines was, at its core, a weapons threat.
The 2014 West Africa outbreak created the pressure that produced a licensed vaccine
The 2013–2016 West Africa Ebola epidemic changed everything. More than 28,000 cases. More than 11,000 deaths. The world's attention focused on the problem in a way it never had during smaller, self-limiting outbreaks in remote Congo.
The vaccine that emerged was Merck's Ervebo — rVSV-ZEBOV-GP — a replication-competent viral-vectored vaccine that works by inserting the Ebola Zaire glycoprotein into the backbone of vesicular stomatitis virus platform.
Originally developed by the Public Health Agency of Canada, licensed to NewLink Genetics, and then sublicensed to Merck, the vaccine was tested in the remarkable ring vaccination trial in 2015. The logic of ring vaccination, which goes back to the smallpox eradication campaign which DA Henderson led, is that you identify a case, vaccinate everyone in that person's immediate network of contacts, and create a vaccine wall around cases. In the Guinea trial, rings were randomized to receive vaccine immediately or after a 21-day delay. Among those vaccinated immediately, there were zero Ebola cases after 10 days. In the delayed groups, cases continued to appear. The efficacy signal was strong enough that the trial was redesigned to eliminate the delayed arm.
The FDA approved Ervebo in December 2019. It was the first licensed Ebola vaccine in history — more than 40 years after the virus was first identified in 1976 outbreaks in what was then Zaire and South Sudan.
The vaccine was deployed at scale during the 2018–2020 North Kivu DRC outbreak, the second-largest Ebola outbreak ever recorded at the time. It delivered under extraordinarily difficult conditions: active armed conflict, deep community distrust, treatment units being burned down, healthcare workers being killed. The vaccine demonstrated effectiveness in a real-world emergency, not just a trial.
But Ervebo only works against Ebola Zaire. The glycoprotein it displays to the immune system is specific to one species. Present it with Ebola Sudan, with Bundibugyo, with Marburg — and the immunity it generates is not reliable.
The European Medicines Agency granted conditional marketing authorization in July 2020 to a second Ebola Zaire vaccine — Janssen's two-dose regimen (sold as Zabdeno and Mvabea), which uses an adenovirus vector for the prime dose followed by a modified vaccinia Ankara booster — though like Ervebo it covers only Zaire ebolavirus, and its two-dose schedule makes rapid ring-vaccination deployment more logistically demanding than the single-shot Merck product.
Ebanga, Inmazeb, and the PALM trial proved that targeted antibodies could turn a death sentence into survivable disease
The second major breakthrough came not from the vaccine pipeline but from the therapeutic antibody pipeline, and it grew directly from a question that arose from the field for decades: could we find, in the immune systems of Ebola survivors, antibodies potent enough to serve as treatments?
The 1995 Kikwit outbreak in the Democratic Republic of the Congo was one of the more extensively studied Ebola events prior to 2014. Researchers collected blood samples from survivors, and those samples went into archives. Years later, NIH researchers, working with the archived samples, identified a single antibody — designated mAb114 — that bound to the Ebola Zaire glycoprotein with remarkable potency and neutralized the virus in preclinical models. This antibody was licensed by Ridgeback Biotherapeutics and eventually approved by the FDA in December 2020 under the name Ebanga.
What established it was the PALM trial conducted during the 2018–2020 DRC outbreak. PALM was a randomized controlled trial that compared four treatment arms: ZMapp (the old standard), remdesivir, mAb114 (Ebanga), and REGN-EB3 (Inmazeb, a three-antibody cocktail from Regeneron). The trial enrolled patients presenting to Ebola treatment units in North Kivu.
The result was decisive. ZMapp achieved 49% mortality at 28 days. REGN-EB3 achieved 34% and mAb114 35%. Both antibody therapies dramatically outperformed ZMapp, and both outperformed remdesivir. The PALM trial stopped the ZMapp and remdesivir arms early once the antibody superiority was established.
The FDA approved two antibody-based therapeutics for Ebola Zaire in late 2020: Inmazeb — a cocktail of three monoclonal antibodies developed by Regeneron and Ebanga, a single monoclonal antibody from Ridgeback Biotherapeutics.
Ebanga is now, alongside Inmazeb, the standard of care for Ebola Zaire. These are genuine life-saving medicines. A disease that killed 90% of patients in early outbreaks now kills closer to a third of patients who reach treatment coupled to supportive care.
But the same species-specificity problem applies here. Ebanga was derived from an Ebola Zaire survivor. Its target, the glycoprotein it binds, is specific to Zaire. It will not work for Sudan, for Bundibugyo, for Marburg. The antibodies are as species-specific as the vaccine.
For every other filovirus, we start from scratch
The current DRC outbreak is caused by Bundibugyo ebolavirus, first discovered in 2007 in western Uganda during a 149-person outbreak. Before the current crisis, the world had seen it in exactly two prior events. Its glycoprotein is genetically divergent enough from Zaire that the existing antibodies and vaccines — all designed around the Zaire surface protein — do not offer meaningful protection.
There is no Bundibugyo vaccine. There is no Bundibugyo monoclonal antibody therapy. The diagnosis itself was delayed by weeks because the standard rapid field tests are calibrated for Zaire and Sudan — the outbreak was circulating for an estimated four weeks before the causative strain was even identified.
Into this gap, several efforts are now moving. A clinical trial (Partners) has begun in the DRC enrolling patients to test remdesivir in combination with a Mapp Biopharmaceutical monoclonal antibody (MBP134) — a descendant of the ZMapp lineage — that may offer cross-reactive activity against Bundibugyo, though the data on cross-reactivity remain preliminary. An Oxford/Serum Institute of India vaccine candidate built on the same adenovirus-vector platform as the AstraZeneca COVID vaccine is advancing toward field deployment, potentially within months. There is also an mRNA vaccine in development from Moderna and another candidate similar to the Merck vaccine in development from IAVI. These vaccine candidates are receiving funding from CEPI. BARDA is developing an antibody product that may have cross-filovirus activity, exploring whether a more conserved region of the glycoprotein can serve as a target that would work across multiple species. These are promising efforts happening while an active outbreak spreads through four DRC provinces.
Marburg has no countermeasures and a confirmed 2026 case in Uganda
Marburg is worth treating separately, because it is a different genus within the Filoviridae family.
An 18-month-old child in Uganda tested positive for Marburg in July 2026. There is a possible second case. The case count is small. The concern is not the current count, but what it reveals: a toddler is almost never the index case in a Marburg outbreak. Marburg typically enters human populations through contact with infected Egyptian fruit bats, in caves or mineshafts. A child that young almost certainly didn't have that contact independently, which means there is an upstream transmission chain we have not yet traced.
Marburg vaccine development has been underway for years, principally led by IAVI, using a cAd3 (chimpanzee adenovirus type 3) vector platform. Phase 1 and Phase 2 trials have been completed with promising immunogenicity data. This is the same platform used in some Ebola Sudan vaccine candidates, and the science behind it is sound. But there is no licensed Marburg vaccine. There is no licensed Marburg therapeutic. If the Uganda case is the beginning of something larger, the response toolkit is supportive care: fluids, electrolytes, oxygen, preventing secondary infections. It is not nothing — improved supportive care has meaningfully reduced filovirus mortality even without targeted agents — but it is far from what the science is capable of producing.
The pattern and the lesson
Step back and you can see the through-line clearly. Filovirus countermeasure development was kicked into motion by a Cold War weapons threat and accelerated by post-9/11 biodefense investment. The money that funded ZMapp, that funded the early vaccine candidates, that sustained the research pipelines long enough to reach the PALM trial — nearly all of it traces back to the recognition that Ebola was a potential weapon and that the United States had no way to protect against it.
It is challenging to generate the sustained investment needed to develop countermeasures against low-frequency, high-mortality pathogens with no commercial market. Biodefense funding was critical in the development of Ebola Zaire medical countermeasures.
The result is a genuine achievement: one licensed vaccine, two licensed therapeutic antibodies, all for Ebola Zaire.
The gap is everything else in the family. Ebola Sudan has killed more than half the people it has infected in every documented outbreak and has no licensed countermeasures. Ebola Bundibugyo is causing the largest active Ebola outbreak in 2026 and has no licensed countermeasures. Marburg can kill 88% of its victims and has no licensed countermeasures.
This is not a scientific mystery. The biology of filovirus glycoproteins is well understood. The antibody technology that produced Ebanga and Inmazeb is the same technology that could produce Marburg and Sudan antibodies. Platform vaccines — viral-vectored and mRNA, — can be adapted to new glycoprotein targets in months rather than years. Scientists know how to do this. The knowledge exists.
What doesn't exist, or hasn't existed consistently enough, is the sustained political will and funding infrastructure to take a known but infrequent threat all the way to licensure when the commercial market is negligible and the outbreak won't last long enough to run a Phase 3 trial. BARDA was created specifically to address that market failure, and when it has been allowed to function — the PALM trial, the 2018-2020 vaccine deployment — it has worked.
The family-wide countermeasure approach I've been arguing for for years is not radical. It is simply the application of what we learned from Ervebo and Ebanga to the rest of the Filoviridae family. Build the antibodies. Run the platform vaccine trials. Do as much as possible between outbreaks rather than during them. Civilization will be categorically safer in a world where these technologies are prioritized versus one in which they aren't.
Filoviruses are not unconquerable. The success of Ervebo and Ebanga proves the opposite. The lesson of the last twenty years is that scientific capability is no longer the limiting factor. The limiting factor is whether we choose to build the tools before the outbreak arrives. Nature will continue generating threats. The question is whether human beings will continue applying reason, resources, and foresight to meet them.