William Haseltine

The COVID-19 pandemic is not over. It will not pass simply because we have grown tired of it. But it is equally misguided to think we are no better off now than we were during the early days of 2020. The past three years have yielded an immense amount of research on, and careful analysis of, SARS-CoV-2. This growing foundation of knowledge has allowed us to develop effective vaccines and antiviral drugs to combat the virus’ spread. Still, this does not warrant complacency: our current vaccines are not infallible, immunity wanes. We need to continue pushing ahead, developing next-generation vaccine technology that provides long-lasting, broadly-neutralizing immunity.

Here, I provide an overview of some of the latest data on waning immunity followed by a discussion of promising advances in the field of vaccinology.

Waning immunity: Infection, serious disease, and hospitalization

The gold standard for any vaccine is protection against the acquisition of infection — if you can’t get infected, you can’t get sick. But for most respiratory viruses, this is a high bar to set. Diseases like polio and measles grant us lifelong, “sterilizing” immunity after infection. By extension, so do the vaccines that protect against them. But unlike polio or measles, infection with a respiratory virus does not provide us with lasting immunity.

This is why, for example, we have yearly flu seasons, with reinfections a common occurrence. Vaccine-induced immunity wanes accordingly. COVID-19 is no exception.

Recent data from Qatar suggests that pre-omicron infections are only 35.5% effective at protecting against symptomatic reinfection with the BA.4 or BA.5 subvariants. This plummets to 27.7% when asymptomatic infections are taken into consideration as well. And making it clearer still, reinfection with the same variant has also been documented.

Many expected immunity from infection to fade. The hope was that immunity against severe disease and hospitalization would remain. In this scenario, those who received a “full” course of the vaccine — generally understood to mean two doses of any of the first-generation vaccines — would still be susceptible to infection, but it would only be mild infection.

A new report released by the U.S. Centers for Centers for Disease Control and Prevention (CDC) suggests otherwise.

The CDC researchers collected data through the Influenza and Other Viruses in the Acutely Ill (IVY) network, consisting of 21 large hospitals spread across 20 different cities in 18 states. Founded in 2019, the initial purpose of IVY was to track influenza vaccine effectiveness among patients admitted to the intensive care unit (ICU). Following the outbreak of COVID-19, the initiative expanded to also enroll hospitalized COVID-19 patients. Their latest analysis is based on a group of 4,730 adult, immunocompetent patients enrolled between December 26, 2021 and August 31, 2022.

Enrolled patients were then split into four different groups, depending on their vaccination status: 1) those who had not been vaccinated before contracting the virus, 2) those who had received two doses of the original mRNA vaccines at least fourteen days before contracting the virus, 3) those who received two doses plus a booster dose of the original mRNA vaccines at least seven days before disease onset, and 4) those who received two doses plus two booster shots of the original mRNA vaccines, again at least seven days before disease onset.

The team of scientists found that, in those who received two doses, vaccine effectiveness against hospitalization during the period of BA.1/BA.2 predominance hovered around 63%. After 150 days, vaccine effectiveness dropped down to 34%. In the group that received two doses plus a booster dose, vaccine effectiveness against hospitalization started at 79% before dropping to 41% after 120 days (Table 1).

Vaccine effectiveness against hospitalization during Omicron
Table 1.
Vaccine effectiveness against hospitalization during Omicron BA.1/BA.2 wave, 7 to 150 days since last dose and 150+ days since last dose. Source: Access Health International.

A similar waning of protection against hospitalization was seen during the BA.4/BA.5 wave of infections. Here, vaccine effectiveness in the two-dose group began at 83% but plummeted to a mere 37% after 150 days. After two doses and a booster shot, vaccine effectiveness stood at 60% before dropping down to 29%. And vaccine effectiveness after two doses and two booster shots stayed at roughly 60% for the duration of 120 days (Table 2).

Vaccine effectiveness against hospitalization during Omicron BA.4/BA.5 wave
Table 2. Vaccine effectiveness against hospitalization during Omicron BA.4/BA.5 wave, 7 to 150 days since last dose and 150+ days since last dose. Source: Access Health International.

The problem of fading immunity is not new, but it does force us to reevaluate if our current approach to vaccination is the best approach.

Solutions? Boosters and “universal vaccines”

Broadly speaking, available solutions can be split into two camps: short-term and long-term.

In the short term, people will need to continue getting their boosters at intervals of around four to six months. For those older than 55, it should be closer to the four month mark. Part of this approach is ensuring that the booster shots are adapted to target the dominant variant in circulation.

Although we do not yet have an exhaustive understanding of the biomarkers associated with immune protection, known as “correlates of protection”, there is enough evidence to suggest that neutralizing antibody titers play a key role. These antibodies home in on the spike protein and bind to it, preventing the virus from entering our cells.

The higher the number of neutralizing antibodies, the higher the protection against infection. But neutralizing antibody titers begin to drop off significantly—up to a five-fold decrease—within three to four months of vaccination. And crucially, high levels of neutralizing antibodies are only as useful as they are “accurate.” All the antibodies in the world won’t do any good if they can’t bind to the relevant antigen.

It is for this same reason that we need to update our flu vaccines every year, to make sure the antigens in the vaccine match those of the latest strains in circulation. When a mismatch does occur, we end up with a more severe flu season. SARS-CoV-2 is no different.

Although still sparse, preliminary data on the bivalent, FDA-approved booster shots indicate that accounting for this mismatch can help reduce the virus’ ability to evade our immune system and to cause serious damage. A press release from Pfizer reports that people aged 55 and up enjoyed a four-fold increase in neutralizing antibody titers against Omicron BA.4/BA.5 compared to those who received a booster shot of the original COVID-19 mRNA vaccine. Compared to those who received no booster shot, neutralizing antibody titers a month after vaccination were 13 times higher in adults older than 55. Younger adults also saw the benefits, with a nine-fold increase in antibody titers.

And in a study posted on the preprint server bioRxiv, researchers discovered that the bivalent booster shot also remains effective against the newest members of the Omicron family, BA.2.75.2 and BQ.1.1.

The issue? Over time people will inevitably get tired of needing two or three booster shots a year. Vaccine uptake will dwindle and we will be back to square one. We are already seeing this happen, with only around 10% of the US population having opted to get the latest bivalent booster.

It is clear we need to continue improving our COVID-19 vaccines so that they provide broad, long-lasting immunity. This is easier said than done, but promising advances are underway.

Targeting highly-conserved regions of the viral genome

One way of achieving broad, variant-resistant immunity is by targeting highly conserved regions of the viral genome.

Current mRNA vaccines work by exposing our bodies to the SARS-CoV-2 spike protein, which the virus depends on to bind and eventually enter our cells. Antibodies that block the spike protein can block infection. The issue is that the spike protein is prone to mutation — its structure can change a lot without sacrificing functionality. By extension, vaccines based only on the spike protein risk losing efficacy when confronted with new variants.

The nucleocapsid protein is a structural protein with an integral role in viral assembly and the packaging of genetic material (Figure 1). It is 90% conserved between SARS-CoV-1 and SARS-CoV-2, compared to 76% for the spike protein. These two features combined make it a very promising target for vaccine design.

Figure 1. A schematic diagram of the SARS-CoV-2 viral particle
Figure 1. A schematic diagram of the SARS-CoV-2 viral particle, with the nucleocapsid protein clearly marked. From: “SARS-Cov-2 nucleocapsid protein is decorated with multiple N- and O-glycans” Supekar et al. 2020

Added to this is the fact the nucleocapsid protein has been shown to elicit a powerful T cell response. Where B cells produce antibodies that can bind to viral particles before they enter cells, T cells are in charge of destroying host cells already infected with the virus; the stronger the T cell response, the better the body can contain the spread of the virus and clear the infection. Analyses of people infected with SARS-CoV-1 have indicated that N-specific T cell immunity can be very long lasting, with some individuals retaining memory T cells up to 17 years after initial infection. These same T cells managed to recognize the SARS-CoV-2 nucleocapsid protein, mounting a quick and specific immune response.

Indeed, research published in Science Translational Medicine corroborates this, suggesting an mRNA vaccine that targets both the SARS-CoV-2 spike protein (S) as well as the nucleocapsid protein (N) may offer stronger and broader protection than current, spike-only vaccines.

Hajnik et al. separated mice into three groups: a control group, which received a saline solution; a test group, which received the bivalent mRNA-S+N vaccine; and a second test group, which received an mRNA vaccine containing only the spike protein (mRNA-S). The vaccines were administered intramuscularly in two doses, an initial prime and a boost three weeks later. They then infected the mice with a mouse-adapted SARS-CoV-2 strain two weeks after administration of the booster shot. They did the same for the hamsters.

Mouse lung chart
Figure 2. A comparison of viral RNA copies in the mouse lungs (log10 viral copies per milligram) between different groups at 2 DPI. From: Hajnik et al. 2022

Both the mRNA-N+S vaccine and the mRNA-S vaccine managed to successfully control the infection, with almost no detectable infectious virus in the lungs (Figure 2).

The vaccines also worked well against the Delta virus in hamster models, but importantly, the mRNA-N+S vaccine outperformed the spike-only vaccine. Where the mRNA-S vaccine managed to reduce lung viral RNA copies 57-fold compared to mock, the mRNA-S+N vaccine managed to do so by 770-fold. Both vaccines protected against lung damage, including bronchiolitis and interstitial pneumonia.

To further test the breadth of the immune response elicited by the mRNA-N+S vaccine, Hajnik et al. exposed hamsters to the Omicron variant (BA.1). At two micrograms, the spike-only vaccine induced modest viral clearance from the lungs; a 12-fold reduction in viral RNA copies two days after infection. The mRNA-N+S vaccine, in contrast, managed to completely clear viral RNA copies from the lungs by day two of the infection.

The same held true for viral titers, with four out of five hamsters having no detectable levels of infectious virus.

More for less: Self-amplifying RNA technology

Another promising advance in the fight against COVID-19, and its many variants, is the adoption of self-amplifying RNA technology.

Where traditional COVID-19 mRNA vaccines are based only on the genetic sequence that encodes the spike protein of SARS-CoV-2, self-amplifying RNA vaccines include a sequence of alphavirus RNA that encodes four non-structural proteins. Once inside a host cell, these non-structural proteins come together to form a molecule called an RNA replicase — a transportable “photocopier” that prints out multiple copies of the mRNA sequence (Figure 3). This means each RNA sequence included in the vaccine is able to make copies of itself, increasing the amount of antigen produced. It also extends the duration of mRNA translation over a longer period of time; vaccine-derived mRNA is usually degraded after a day or two, limiting protein expression to two to three days at the most. However, self-amplifying RNA technology can extend this timeframe — both mRNA and protein production — to up to almost a month.

Conventional mRNAs
Figure 3. A)
Conventional mRNAs encode the vaccine immunogen and flanking 5′ and 3′ UTRs. An antigen or immunotherapy is translated from the nonreplicating transcript. B) Self-amplifying RNA encodes 5′ and 3′ CSE sequences, the nsP1-4 genes, a subgenomic promoter, and the vaccine immunogen. Following in situ translation, the nsP1-4 proteins form an RdRP complex which recognizes flanking CSE sequences and amplifies vaccine-encoding transcripts. This results in an accumulation of the antigen or immunotherapy within the cell. From: “Self-amplyfying RNA vacinnes for infectious diseases” bloom et al. 2020

Self-amplifying RNA technology comes with a number of benefits. For one, because each RNA sequence will produce multiple copies of itself, manufacturers can use a much smaller amount per vaccine dose. The smaller dose enables multiple different strands of mRNA — each encoding its own antigen — to be included in one vaccine. The dose-sparing approach also significantly reduces the cost of production; the price of the vaccine drops accordingly, making it more accessible to low- and middle-income nations.

Like mRNA vaccines, the manufacture of self-amplifying RNA is entirely synthetic. Manufacturing does not require growth in any living cell. It is entirely a chemical process. Production is rapid and flexible and can be adjusted rapidly to respond to new demands. The manufacturing infrastructure and quality control are greatly simplified as compared to traditional vaccines.

Since self-amplifying mRNA allows more of the target antigen to be produced, over a longer period of time, it gives our immune system a better chance to learn what it needs to develop a highly targeted immune response. Longer exposure to an antigen allows the antibody response to mature, to make higher-affinity, more-potently neutralizing antibodies, as well as antibodies that may elicit cellular cytotoxicity. Moreover, longer antigen exposure is important for generating memory T cells required for long-lasting memory.

Intradermal delivery: Boosting cellular immunity

How a vaccine is administered, despite not being part of the vaccine per se, contributes to the immunogenicity and safety of the vaccine all the same. This includes both the equipment as well as the route of administration.

Our skin is made up of three layers: the outermost protective layer called the epidermis, a middle layer called the dermis, and a fatty layer at the bottom called the hypodermis or subcutaneous tissue. Most COVID-19 vaccines are injected into muscle tissue, which lies well beneath the skin. This has been the standard approach for a long time. But intradermal delivery holds a few distinct advantages.

The dermis layer of skin is tightly packed with immune cells, including antigen-presenting cells (APCs). These play a vital role in the initiation and modulation of our immune system, including our adaptive immune response. In particular, various different types of dendritic cell latch onto antigens and present these to naïve T cells, helping to shape and mature a targeted cellular immune response. Since intradermal injection lets the antigen be produced in close proximity to these antigen presenting cells, it may lead to a quicker and more fine-tuned adaptive immune response. A recent study comparing intradermal and intramuscular administration of the SARS-CoV-2 receptor binding domain (RBD) confirmed as much, describing improved T cell responses after injection into the dermis.

Intradermal administration comes with the added benefit of needle-free delivery options, such as jet injectors. These use springs or compressed gas to produce a narrow, high-pressure stream of fluid that penetrates the skin and delivers the vaccine. The lack of a needle helps improve uptake, especially among younger and needle-averse populations. Needle-free delivery also nullifies the possibility of needle reuse and subsequent cross contamination, which continues to be a commonplace issue in many parts of the globe, with up to 1.3 million deaths a year attributed to such practices.

Structure of the skin info graphics illustration
Credit: Paladjai/Getty Images


Current mRNA vaccines work well, protecting against both infection and severe disease, but they do so only for a short period. Antibody titers begin to wane as early as three months post-vaccination, and with them wanes the protection against immunity. The CDC data indicate that protection against serious disease and hospitalization drops off soon after, even against “milder” variants such as omicron.

Our vaccination schedule needs to adapt accordingly: for optimal protection, booster doses should be sought every four to six months. Those over 55 should lean towards a four month interval, those younger than 55 may be able to get away with a six month interval. In both cases, consistency is key.

Given enough funding and research, these boosters may eventually be replaced by longer-lasting, broadly-neutralizing vaccines. The technological foundations for this transition are in development as we speak, with advances in antigen design, antigen expression, and antigen delivery. Importantly, these factors are not mutually exclusive — progress in one can be combined with progress in another, with the benefits of each “stacking” on top of each other.

Although a cautious optimism may be justified, we would do well to remember that developing vaccines that induce sterilizing immunity — total protection from infection — is extremely difficult. Despite decades of research on influenza, for example, we still lack vaccines that offer such protection against the viruses. Human Immunodeficiency Virus (HIV) is another example; after more than forty years, we still do not have any successful vaccines. But the fight against HIV proves that even in the absence of vaccines, progress can be made. Pre-exposure prophylaxis (PrEP) medication helps prevent infection in individuals at risk of exposure, and antiretroviral treatment (ART) helps those who already have HIV bring levels down to barely perceptible levels, allowing them to live long, healthy lives.

A similar strategy, revolving around highly-active antiviral drugs, may be necessary to control COVID-19 while we work towards vaccines capable of inducing sterilizing immunity.


William R. Haseltine, PhD, is chair and president of the think tank ACCESS Health
International, a former Harvard Medical School and School of Public Health professor and founder of the university’s cancer and HIV/AIDS research departments. He is also the founder of more than a dozen biotechnology companies, including Human Genome Sciences.

Also of Interest