A version of this article appeared on the website COVID-Explained.

Recently, Moderna announced a significant milestone in their quest for a COVID-19 vaccine (more on the particulars of this below). In response, Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, said he expected 100 million COVID-19 vaccine doses by early 2021. This would be completely transformative. But developing and manufacturing new vaccines typically takes years—decades in many cases. There’s reason to hope that a COVID-19 vaccine could come sooner, but as hopeful as anyone, including Fauci, may be, there are still many unknowns.

No one has a crystal ball. But we can help you understand what’s happening in COVID-19 vaccine research, what we do know, and why there’s so much that we don’t.

Even the most innovative COVID-19 vaccine candidates would work on the same basic principles as all vaccinations. To understand that, here’s a brief review of some immunology fundamentals. (For more, see this Path of the Virus explainer.)

When your immune system meets a viral threat, it begins fighting with a generic response, which is not specific to any particular virus. But during the first one to two weeks after infection, your adaptive immune cells kick in. These cells learn to recognize specific structures of the virus, called antigens, and train to target the virus once they recognize it. Adaptive immune cells do many things, but one of the major ways they protect you is by producing antibodies. You’ve probably heard quite a lot about antibodies lately. They’re large proteins created to target and stick to the antigens on the virus, and kill it.

Importantly, some adaptive immune cells become “memory cells,” long-lived cells that remain in your body, ready to quickly ramp up a fight against re-infections of the same pathogen so that your body doesn’t have to start from scratch next time. These memory cells, along with antibodies which also stick around in your body, are key to viral immunity.

Vaccines take advantage of your immune system’s memory by simulating an infection in a safe manner so your body will produce memory cells and antibodies that will be ready if an actual infection occurs. To do this, vaccines have to mimic the virus so the immune cells can undergo the learning process.

The most traditional method to train the immune system is with killed or weakened virus; the virus is recognizable enough that the immune system responds, but too weak to make you sick. Measles, polio, and some flu vaccines use weakened or killed viruses. This method is how humans have been making vaccines for the past 100 years, and it works for many diseases. But these vaccines take a lot of effort to manufacture—it would be hard to make enough for all 8 billion of us at once.

Other vaccine approaches introduce the antigen—the part of the virus that antibodies target—to your body without the rest of the virus. There are a few versions of these. Viral vector vaccines use a live, genetically engineered virus to introduce viral DNA into our cells, which then hijacks our own cells to produce viral antigens. This is a relatively new method—two Ebola vaccines have been developed using viral vectors, but these are the only vaccines that have used this method—and, while promising, it has some limitations. (For example, if you’ve already been infected with the virus that is used as a vector, this will not work.)

Nucleic acid vaccines work on the same principle, but with a different delivery vehicle. Rather being delivered by a virus, the antigen-producing snippet of DNA is snuck in mechanically. One method is an oil-like structure that can pass through cell membranes without disrupting them; another, called electroporation, uses electric shocks to briefly open up small holes in our cells, allowing DNA to enter. In either case, once the DNA is inside, our cells begin reading the DNA instructions and produce antigens.

No vaccines for any diseases have yet been approved with this approach; however, several COVID-19 vaccine candidates using this approach have shown promising results (including the vaccine from Moderna about which Fauci is so optimistic). One of the biggest advantages of nucleic acid vaccines is the impressive speed at which they can be designed, allowing researchers to quickly produce potential vaccine candidates.

Finally, a protein-based vaccine takes a more direct approach, providing your body with the antigen itself. Antigen manufacturing (by a vaccine company rather than by your body’s cells) can be a slower process, and protein-based vaccines also require a component called an adjuvant, which sculpts how your immune system responds to the antigen (i.e. the type of antibodies made and how long your immune system will remember the antigen). Adjuvants are well understood chemicals, but adding additional ingredients means more development challenges. When done well, this approach has been shown to be highly successful. The Hepatitis B vaccine uses this method, as do some flu vaccines.

However, all of those vaccines—indeed, all vaccines of all kinds—were developed more slowly than the COVID vaccine timeline. (For an overview of the usual timeline of vaccine development in the U.S., see here.) Vaccine development and testing usually takes about 10 to 15 years. (For some diseases, like HIV, TB, and malaria, years of development still haven’t led to effective vaccines.) In the case of COVID-19, the timeline is likely to be dramatically accelerated. The FDA is determining which vaccines can go on to initial clinical trials (without preclinical animal testing) often based on previous safety and efficacy data of the type of vaccine that is being proposed. Additionally, multiple initiatives, including ones from the Bill & Melinda Gates Foundation and the NIH, have been launched to further speed up development and access to vaccines.

Currently, there are more than 135 COVID-19 vaccines in development, with 16 of them in human trial phases. (Phase I tests safety and optimal vaccine dosage in a small number of healthy volunteers; Phases II and III test efficacy in larger and larger groups of people, including higher-risk patients. Each phase can take six months or a year, and very few vaccines usually make it through all three successfully.) For a continuously updated tracker of vaccine candidates, check out BioRender, the NIH website, and the New York Times Vaccine Tracker. There are many different groups developing vaccines and it can be difficult to predict which candidates will succeed, as it is common for many vaccines to fail due to problems with safety, efficacy, or other issues. Here are a few candidates that are at the forefront of the clinical trial phases and may be worth watching:

China recently announced via social media that a vaccine candidate by the Wuhan Institute of Biological Products and the Beijing Institute of Biological Products may be ready by late 2020 or early 2021, one of the most optimistic predictions yet. These researchers are employing a killed/weakened virus and are currently in Phase II of trials. While this inactivated virus method often provides the most robust immune response (since the virus vaccine better represents the actual SARS-CoV-2), the development process is typically longer than other approaches. Extra care may be needed to verify its safety.

The U.S. company Moderna is developing a nucleic acid vaccine that delivers mRNA instructions for the SARS-CoV-2 spike protein to our cells. On May 22, it reported results showing that people who were given the vaccine produced antibodies against the virus. This is a huge first step in showing the vaccine could be effective. The results were widely interpreted as positive and sent stock prices surging. Fauci said he regards these results as “quite promising” and said this candidate could conceivably be ready by the end of the year. However, because the data hasn’t been published yet, many researchers are waiting to see the details needed to properly evaluate these claims and this would be the first ever human mRNA vaccine. Moderna has begun Phase II trials in 600 people.

A group at the Jenner Institute of Oxford University in the U.K. is employing a type of adenovirus as a viral vector to make their COVID-19 vaccine. They are currently following up from Phase I trials, which began in April and ended in late May. In preclinical trials, their candidate protected six monkeys from pneumonia, but it is hard to draw strong conclusions from these studies since the monkey model of COVID-19 only replicates mild disease. Before publishing the results from its Phase I study, the Jenner Phase II/III trial has already begun to enroll subjects. In this combination study to test how many doses the vaccine would require (Phase II) and its efficacy in a variety of age groups (Phase III), Oxford plans to enroll over 10,000 individuals. This is the largest human clinical trial to date on a COVID-19 vaccine candidate and it is moving at light speed compared to normal vaccine development.

The speed of vaccine development overall for COVID-19 is unprecedented. Because of the desire to go fast, most of the development has focused on newer technologies, not the traditional weakened virus approach. These newer technologies have to pick a particular antigen to focus on, and all of the candidates in development target the same protein in SARS-CoV-2: the spike protein, or S-protein. The differences across technology are largely in how they deliver instructions to your cells for how to make it.

We’ve only known this virus for five months now, and the biomedical community is trying to make a lot of educated guesses based on what we’ve learned from other coronaviruses. Scientists are developing S-protein based vaccines because research on SARS and MERS suggests that an antibody response to S-protein could be protective, and we know that the S-protein plays a key part in the induction of antibodies. But we never developed a SARS vaccine—the epidemic ended before a vaccine was ready to be tested—and we don’t have human vaccines for any other coronaviruses. We don’t know for sure that a S-protein based vaccines is going to work.

At least some scientists are worried that we have all of our eggs in one S-protein basket, so to speak. We are relying on an assumption that the immune response to the S-protein works—and while this is an assumption based on promising data, it still could be wrong. Some enlightening research published in the last month has shown that the human body forms antibody and T-cells responses not only in response to the S-protein but also to other important proteins of the SARS-CoV-2 virus. If we discover that raising a vaccine-mediated immune response to S-protein doesn’t provide protection (regardless of the different delivery technologies), research would be set back months, or even years.

Because of this concern, ImmunityBio and NantKwest are developing an adenovirus vector vaccine that incorporates the S-protein and a second protein on surface of the virus, nucleocapsid protein. They aren’t as far ahead in development as Moderna and the rest, but they are planning a Phase I trial this summer and were selected for Operation Warp Speed for expedited development. Stay tuned this summer and fall to see how these new approaches pan out.

Many different companies are racing toward the same goal through different approaches, which provides some (but not perfect) hedging against failure of a single candidate. A process that usually takes years is being condensed into months. Vaccine companies have partnered with manufacturing companies before they’ve seen positive results to ensure high production capabilities.

But as different vaccine candidates are being accelerated through the pipeline, companies and government agencies must still ensure safety in addition to efficacy.

It is important to stress that any predictions about when these vaccines will be ready, not to mention how well they will work, are guesses in the dark. We all want a solid timeline to grasp and hope for, and “we just don’t know” is an unsatisfactory answer—but we really just don’t know. It’s possible that the successful vaccine candidate has not yet been created. Efficacious vaccines are often not safe. Safe vaccines might not be effective. Vaccines that work in monkeys often don’t work at all in humans.

It is also important to note that as the number of COVID-19 cases decreases, efficacy of the various vaccine candidates will be harder to judge because it will be hard to tell if the vaccine is working to protect people who’ve received it, or if they’re just not being exposed to the virus thanks to social distancing and other efforts to reduce transmission.

Will this extend the timescales of the trials? Will the currently accelerated trials compromise safety? How long will it take to scale up manufacturing capabilities and produce enough doses? The speed of progress is encouraging, but we’ll need more time to answer these questions for certain.

Future Tense
is a partnership of
Slate,
New America, and
Arizona State University
that examines emerging technologies, public policy, and society.