Coronavirus vaccine developers wary of errant antibodies

Concerns persist that COVID-19 vaccines could cause antibody-dependent enhancement, which can potentiate viral entry into host cells and worsen disease.
Ann Arbor, Michigan, USA.

Search for this author in:

In late May, CanSino Biologics published the first data for its COVID-19 vaccine expressing spike protein. The vaccine generated neutralizing antibodies in many recipients and appeared safe, but the company, like others in this space, remains on alert for a dangerous phenomenon known as antibody-dependent enhancement (ADE). “It’s important to talk about it [ADE],” says Gregory Glenn, president of R&D at Novavax, which launched its COVID-19 vaccine trial in May. But “we can’t be overly cautious. People are dying. So we need to be aggressive here.”

There are mounting theoretical concerns that vaccines generating antibodies against SARS-CoV-2 may bind to the virus without neutralizing it. Should this happen, the non-neutralizing antibodies could enhance viral entry into cells and viral replication and end up worsening infection instead of offering protection, through the poorly understood phenomenon of ADE. ADE “is a genuine concern,” says virologist Kevin Gilligan, a senior consultant with Biologics Consulting, who advises thorough safety studies. “Because if the gun is jumped, and a vaccine is widely distributed that is disease enhancing, that would be worse than actually not doing any vaccination at all.”

The degree of ADE vaccine risk for SARS-CoV-2 is unknown.

Following vaccination, antibodies that bind the SARS-CoV-2 virus may still fail to neutralize it, and could even facilitate infection.Reprinted with permission from A. Iwasaki and Y. Yang, Nat. Rev. Immunol. 20, 339–341 (2020).

“Most experts who look at it acknowledge the potential risk but don’t see compelling evidence for it in humans right now,” says James Crowe, director of the Vanderbilt Vaccine Center. Some view coronavirus ADE as a laboratory phenomenon, at least so far. “The theoretical possibility [of ADE] is there,” says Rafi Ahmed, director of the Emory Vaccine Center. “The more difficult question becomes, how do you balance this concern with moving forward? There’s not an easy answer.”

ADE was first described in 1977 by virologist Scott Halstead, then at the University of Hawaii. Halstead studied dengue, caused by four related but antigenically distinct viruses known as ‘serotypes’. Halstead noted that animals previously infected with one serotype did worse if infected again with a different serotype than if never previously infected at all. He found antibodies to be responsible for this phenomenon. Instead of neutralizing the virus, the antibodies facilitated viral entry into host cells in vitro. In his publication, Halstead presciently wrote: “Non-neutralising antibody presumably provides a specific molecular ‘ride’ for an infectious dengue virion into the interior of a receptive phagocytic cell.”

Halstead’s speculation was later proven correct, not just for dengue, a flavivirus, but for other viruses, including coronaviruses, in cell culture experiments. The coronavirus uses its spike protein to bind host receptors and infect airway cells. But it also can infect immune cells, typically macrophages. Anti-spike antibodies ideally bind and neutralize the virus, either by blocking spike binding to its receptor or by preventing fusion of the virus with the cell. But if antibodies fail to do this because they bind unproductively to the virus, these non-neutralizing antibodies may instead, using the Fc region at the base of the antibody, latch onto host macrophage Fc receptors and ferry the virus into these host cells, at least in vitro. Once inside, the viruses replicate and then exit to further spread infection to neighboring cells and tissues.

Poor-quality antibodies that bind the virus without neutralizing it are one reason vaccine candidates fail, and, in theory, could also cause ADE, delivering virions to host cells.

Can SARS-CoV-2 infect macrophages? “I don’t know,” answers Stan Perlman, a coronavirus expert at the University of Iowa. “Those data are not out there yet.” The virus can infect macrophages in vitro, but that doesn’t necessarily translate to in vivo human biology. “In tissue culture cells you can make almost any virus show this, by jigging the right cells, the right amount of antibody, and the right virus. But that [is] not really meaningful, because of the fact that it’s so artificial.”

The risk that a COVID-19 vaccine will cause ADE and facilitate infection “is completely theoretical,” says Novavax’s Glenn. “It’s a lab, an animal phenomenon that’s been observed. [But] in RSV it’s a real clinical event.” In a 1960s vaccine trial, 80% of infants and young children immunized against RSV (respiratory syncytial virus, an RNA virus in the family Paramyxoviridae) ended up hospitalized, and 2 children died. Only 5% of the children in the control group required hospitalization. Vaccinated children carried high titers of non-neutralizing antibodies, implicating ADE in the tragic outcome, although a TH2 (inflammatory) T cell response also probably contributed.

Since then, some vaccine studies in animals, as well as human trials where vaccinated people had worse outcomes, suggest ADE at work. But definitive evidence is lacking. “We can’t say they were not due to ADE, but there’s not evidence that they were,” says Ahmed. “There are many other ways that an immune response can cause damage.”

For coronavirus vaccines, the ADE story is similarly inconclusive. Cats vaccinated with spike protein against feline coronavirus died much faster than unvaccinated cats and carried more anti-spike antibodies, implicating ADE. But macaques vaccinated against SARS-CoV-1 did not show enhanced infection or disease.

In humans, limited data for vaccines against earlier coronaviruses did not show ADE. Two vaccines against the closely related SARS-CoV-1 appeared safe in early trials. A DNA vaccine against the MERS coronavirus in macaques protected all eight monkeys from infection symptoms, and that MERS vaccine was well-tolerated in a 2016 phase 1 human trial. MERS vaccine co-developer David Weiner, director of the Vaccine & Immunotherapy Center at the Wistar Institute, co-developed a very similar COVID-19 vaccine that Inovio Pharmaceuticals took into human trials in early April.

“We don’t see it [ADE] with flu, we don’t see it with so many other vaccines that we have,” says Ahmed. “We shouldn’t dismiss it completely, but then if you look at the broader world of vaccinology, this has not been seen.” But Gilligan says some earlier SARS and MERS vaccine candidates didn’t advance because of signs of ADE.

ADE is a concern for monoclonal antibodies (mAbs) too, at least in theory, because mAbs that fail to neutralize the virus also could facilitate viral entry into cells. Many companies are working on such mAbs for either COVID-19 prevention or therapy. AstraZeneca, for example, is partnering with Vanderbilt, which has identified, expressed and evaluated virus-neutralizing antibodies from B cells recovered from convalescents. AstraZeneca will take the best ones forward into clinical trials.

ADE is part of the mAb discussion. “There’s not a lot of obvious epidemiologic evidence for it yet,” says Vanderbilt’s Crowe. Nevertheless, he says, “we will address this.” Crowe’s group will test for ADE by seeing whether its monoclonal antibodies can deliver viruses into macrophages, compared with antibodies whose Fc region has been inactivated. (An antibody lacking Fc can’t bind macrophage Fc receptors and thus can’t cause ADE. That’s why some companies are advancing mAbs with the Fc region knocked out.) Still, “the animal models are so inadequate right now, it would be very difficult to prove whether [ADE] is common or uncommon,” Crowe says.

For mAbs, the risk is probably low. “You’re already characterizing it as a highly neutralizing antibody,” says Ahmed. Convalescent antibody passive immunization trials, now underway, could carry some ADE risk, Ahmed says. “[But] I would be less concerned with that than with the vaccine that might be inducing suboptimal response.”

Vaccine companies are fully aware of ADE. “All the groups have been discussing this, everyone’s on the same page, everyone wants a safe [COVID-19] vaccine,” says Weiner. Inovio went forward into the clinic after showing safety in primates, and is reassured by the strong human safety record of its MERS vaccine. But vaccine trials are being approved “without the usual amount of preclinical safety/tox data available for guidance,” says Gilligan.

Given the urgency, the US Food and Drug Administration and European Medicines Agency may have little choice but to give the go ahead to enter the clinic. Novavax launched a phase 1/2 trial of its vaccine, a stable prefusion SARS-CoV-2 spike protein trimer nanoparticle vaccine with a saponin-based adjuvant, in late May. The company is experienced with viral vaccines: its nanoparticle flu vaccine recently met its phase 3 endpoints. The SARS-CoV-2 virus “is not new territory,” Glenn says, adding that Novavax and other vaccine companies can assay for whether antibodies can block spike protein–receptor binding, or neutralize the virus, or both, which should prevent infection.

And ADE? “One has to address it,” Glenn says, but he points out that ADE is only a theoretical COVID-19 vaccine risk. “It’s hypothesized from animal studies,” he says. “I think the way forward is to take it seriously, do some animal studies, show you don’t have enhancement, monitor during the trial. My view is, we can’t let this issue keep us from getting to a vaccine that will save millions of lives.”

But Anne De Groot, CEO of EpiVax, argues for a more nuanced approach. EpiVax develops T-cell-directed vaccines, using computational tools to select viral T cell epitopes. The company found SARS-CoV-2 sequences that would appear human to T cell receptors.

This could be a problem for any vaccine. “The T cells either don’t respond as strongly, because they think that the [coronavirus] spike is human, or that sequence can trigger a regulatory T cell response,” says De Groot. Such T cells can interfere with beneficial antibody production and “you don’t get neutralizing antibodies, you get the kind of antibodies that actually potentially enhance infection.” In other words, ADE.

Showing vaccine safety in primates does not guarantee safety in humans, De Groot points out, because primates express different major histocompatibility complex (MHC) molecules, which alters epitope presentation and the immune response.

Other vaccine makers are also building a strong T-cell response into their vaccines, although not an epitope-specific response. Such T cells address ADE to some extent because they will clear infected cells, even in the presence of disease-enhancing antibodies. But to prevent ADE, De Groot also argues for removing epitopes responsible for stimulating regulatory T cells, which most vaccine makers are not doing. EpiVax partners Entos Pharmaceuticals and Immunomic Therapeutics are developing such T cell epitope-specific COVID-19 vaccines.

For now, the risk of ADE remains theoretical. “At this stage, what the world needs is a vaccine that’s effective,” says Ahmed. “I think the priority will be, do you get a neutralizing antibody, do you get a reasonably good T-cell response in the vaccine. And I think one would have to move forward with that vaccine as soon as we can.”

doi: 10.1038/d41587-020-00016-w

Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.