Four COVID-19 vaccine candidates induce immune responses in preliminary studies, but questions remain

Preliminary data from four recently published COVID-19 vaccine trials suggest that the experimental injections may be safe. Encouragingly, the vaccines were able to induce antibodies that target the novel coronavirus, SARS-CoV-2, in most people who volunteered to get the shots. Their effectiveness, however, remains to be determined.

The reports provide the first published, peer-reviewed data for two of the most closely watched COVID-19 vaccines: one developed by the US firm Moderna and one from the University of Oxford. The Chinese company CanSino Biologics also published the first data from the first Phase II study of any COVID-19 vaccine trial. And Pfizer and the German firm BioNTech joined the frenzy as well, releasing a preprint—a report that has not been peer-reviewed—that outlines results from a small study of the firms’ jointly developed vaccine.

It’s too early to know if these vaccines will work, but researchers who are closely tracking their development are cautiously optimistic. “These results are promising,” says Hildegund Ertl, a vaccine scientist and immunologist at the Wistar Institute. “We have four vaccines that clearly give immune response in most individuals. It is hard to say which one is the best, but it could be worse. We could have four vaccines that give you nothing.”

Experts warn against overinterpreting these preliminary studies, which come with a basketful of caveats. First, although the vaccines don’t appear to be dangerous, they are not without side effects, including fatigue, fevers, and headaches. Second, it is difficult to make meaningful comparisons since the four groups used different methods to measure the antibody responses. And finally, no one knows how strong of an immune response a vaccine needs to create for protection, or how long that immune response will last.

We won’t know if these vaccines are protective until they are tested in large Phase III studies designed to answer that question, says Angela Rasmussen, a virologist at Columbia University. “These immune responses alone don’t demonstrate that any of these vaccines are effective at protecting recipients from SARS-CoV-2 infection or COVID-19 disease.”

The data dump comes at a pivotal moment, as these groups prepare to recruit thousands of volunteers for their Phase III clinical trials. Moderna and Pfizer plan to start efficacy trials in the US this month. Oxford’s vaccine is already in efficacy trials in Brazil, South Africa, and the UK. A 30,000-person study of that vaccine could begin in the US soon too.

The US government has invested in or preordered three of these four vaccines as part of its Operation Warp Speed program to distribute 300 million COVID-19 vaccines by early 2021. In April, the government awarded up to $483 million to Moderna to develop and manufacture its vaccine. In May, it signed a deal worth up to $1.2 billion with AstraZeneca, which is developing Oxford’s vaccine, for at least 300 million doses of the vaccine. And on July 22, the US government ordered 100 million doses of Pfizer’s vaccine for $1.95 billion, with the potential for up to 500 million more doses.

“They are all marching at top speed into efficacy trials,” Ertl says. “That’s really what we need, especially with this new virus where we don’t have a clue what’s protective.”

The four groups are developing vaccines based on different technologies, but they all have the same goal: tricking our bodies into making the SARS-CoV-2 spike protein and providing target practice for this crucial part of the coronavirus. Moderna and Pfizer are encoding instructions for the spike protein in mRNA, which is delivered into our cells via lipid nanoparticles. CanSino and Oxford have created adenoviral vector vaccines, which use a genetically engineered cold virus to shuttle DNA encoding the spike protein into our cells.

All four groups reported that most or all people developed antibodies to the SARS-CoV-2 spike protein 2–4 weeks after vaccination. There are many ways to quantify these antibody levels. The simplest method measures the antibodies that bind to the spike protein. A similar method measures antibodies that bind to a particularly important region of the spike protein, called the receptor binding domain (RBD), which the virus uses to infiltrate human cells.

Moderna reported levels of these kinds of antibodies that are “through the roof,” Ertl says, although she cautions that the values can vary widely based on how researchers conduct the experiment. Researchers incubate antibodies from a blood sample with artificial antigens—in this case, the spike protein—and see how many matches are made. “If you want to impress your investors, you can load the plate with antigen, keep the incubation running as long as you can, and then you get very high titers,” she says. “You cannot compare these studies. Everyone uses a different assay for antibodies.”

There’s another problem with these measurements. Antibodies that bind the spike protein, or even the RBD, aren’t guaranteed to prevent infection. Only a handful of antibodies that bind at just the right spot, so-called neutralizing antibodies, will do that.

To measure neutralizing antibodies, researchers must employ more costly and time-consuming experiments that either use an artificial version of SARS-CoV-2, called a pseudovirus, or live SARS-CoV-2 itself. The live virus tests are the most reliable, but also the most difficult to do, says Norbert Pardi, an mRNA vaccine scientist at the University of Pennsylvania. “Pseudovirus tests are OK, but not ideal,” he adds.

Even then, the level of neutralizing antibodies needed for protection is unknown. As a benchmark, Moderna, Pfizer, and Oxford all measured neutralizing antibody levels in plasma from people who have recovered from a natural COVID-19 infection. “This is a very good comparator,” Pardi says.

The levels of neutralizing antibodies found in this convalescent plasma ranged wildly, from undetectable in a few individuals to extraordinarily high in others. Moderna and Pfizer’s studies showed that one shot of their mRNA vaccines was not enough to induce levels of neutralizing antibodies equal to the average levels found in convalescent plasma. A booster shot, however, could raise the levels 2–3 times higher than the average in plasma, depending on the dosage.

Oxford’s vaccine, tested in 1,077 people, did not induce as strong of an antibody response. The researchers decided to try out a booster shot in a tiny group of 10 people. The booster raised binding antibody and neutralizing antibody levels close to the averages found in convalescent plasma. CanSino didn’t compare its vaccine data to convalescent plasma.

Researchers studying convalescent plasma have noted that the average levels of antibodies developed after a SARS-CoV-2 infection are not particularly high, and some studies suggest that the antibody levels decline over time. For that reason, some researchers say an ideal COVID-19 vaccine should induce antibody levels that significantly exceed average levels in convalescent plasma.

Ertl applauds the Oxford study for including a more detailed analysis of convalescent plasma antibody levels, sorting them by people who had severe, mild, or asymptomatic infections. People who had severe infections had the highest levels of neutralizing antibodies, and she would like to see the vaccines induce antibodies closer to those levels.

The longevity of the antibody response is another factor that could make or break a vaccine. The vaccines from Moderna and Pfizer induced high levels of antibodies soon after the second shot, but levels began to decline rapidly 2 weeks later. “If this is going to be the trend, then this is concerning,” Pardi says.

The antibody levels appeared more stable for the Oxford vaccine, but it is too soon to know if the antibody levels for any of these vaccines will continue to decline or will reach a plateau, Pardi explains. CanSino’s study reports data only up to 4 weeks after the vaccination, Pfizer and BioNTech’s study reports data up to 6 weeks, and Moderna and Oxford’s studies report data for up to 8 weeks.

So far, most groups have largely focused on antibody responses, but it is important to watch for T cell responses, too, Ertl says. Some T cells can help boost the antibody response, and others can help destroy cells that are already infected with a virus. All four groups reported the presence of a T cell response, but it was not as strong as Ertl would have liked to see. “I was disappointed with the T cell data,” she says.

Safety and side effects could present another sticking point. None of these four studies were large enough to determine if the vaccines cause rare reactions. But all four vaccines commonly caused systemic side effects, and higher dosages were linked to more frequent and more severe side effects.

For Oxford’s vaccine, which the university licensed to AstraZeneca, some 70% of volunteers reported fatigue and 68% had headaches. Muscle ache, malaise, chills, and fevers were each present in about half the volunteers.

For Pfizer’s vaccine, chills, fever, joint pain, and muscle pain were increasingly common at higher dosages, and more frequent after a booster shot. After the second injection of the medium dosage of Moderna’s vaccine, 80% experienced fatigue, 80% had chills, and about half had a headache, muscle pain, nausea, or fever. One person, in the low-dosage group, didn’t get their second shot due to hives on both of their legs.

Although most of these side effects were deemed mild or moderate, there’s a chance that severe side effects could become more pronounced as the vaccines are tested in larger trials. “These vaccines may be rougher than some other vaccines,” Ertl says. “But if it is a choice between COVID and having a fever for a day, it is a no-brainer.”

CanSino’s 500-person Phase II trial shows that age and other differences between individuals will play a role in how well the vaccines work and how severe the side effects are. Participants who were 55 years of age or older had significantly reduced antibody responses to the vaccine.

The firm’s vaccine has another hurdle. About half the people in the study had preexisting immunity to the adenovirus used to deliver the SARS-CoV-2 spike gene, which significantly reduced the ability of the vaccine to induce neutralizing antibodies. Oxford’s vaccine didn’t face this problem, since it was based on an adenovirus that normally infects chimpanzees, not humans.

CanSino’s trial also suggests that the people who may benefit most from the vaccine may also experience the worst reactions to it. Some 9% of those who received the high dosage of the firm’s vaccine experienced severe reactions—typically a fever that resolved within 3–4 days without drugs. Almost all those people had no preexisting immunity to the adenovirus used in the vaccine.

The other three vaccine trials were limited to people aged 18–55, and almost all the participants were white. Pardi says that it will be important for the upcoming Phase III trials to be more inclusive of people of older ages and other ethnicities.

“None of these vaccine studies give a clear picture of vaccine efficacy,” Rasmussen says. “I encourage people to hold off on making too many conclusions based on these data alone.”