The unforgettable first flu

Does the first flu you ever caught affect how you will cope with all future flu infections you might encounter, including pandemic flu? This fascinating hypothesis, originally proposed by Thomas Francis in the 1940s and 50s, has been revisited in the past few years and its implications summarised by Decan Butler in Nature. The original hypothesis suggested that individuals develop a stronger antibody response to the first encountered strain than to those encountered later in life. This phenomenon is now better defined at a molecular level and is called “imprinting”, which is described in detail by Guthmiller and Wilson.

What is imprinting?

Imprinting (also referred to as antigenic seniority) describes the activation of a memory response following exposure to a novel influenza virus. When babies encounter their first flu (typically before turning 3-years-old), their immune system activates B cells with affinity for the flu antigens; mature, and at once the infection is under control, a memory B cell response is established, where the best performers hide in a resting state, ready to reactivate at the next encounter. In the next encounter, generally with a different influenza strain, two responses play a role: a biased memory B cell response targeting conserved epitopes, and a naïve B cell response targeting novel epitopes; the first being faster than the second. The memory response is often directed against conserved epitopes on the “head” of HA, but also against the highly conserved epitopes in the receptor binding domain (RBD) within the HA head and stalk. This is both good news and bad news.

The bad news first. During secondary activation upon encounter with a novel flu strain, this memory response is preferentially activated and will be directed against shared epitopes, limiting the activation of a more tailored response. This means that subsequent exposure to drifter influenza strains induces a progressively narrower breath of circulating antibodies, generally with activity against highly accessible epitopes. By occupying the easily accessible sites, these antibodies create sterical hindrance, thereby preventing more selective antibodies from recognizing the more specific or conserved regions. Besides eliciting a suboptimal response, this also creates a strong selective pressure for the virus to mutate or to develop novel glycosylation sites that can mask previously recognized regions.

Now the good news. Imprinting also allows activation of B cells with specificity for conserved epitopes, providing protection upon exposure to novel drifted and shifted viruses, and it is because of a lack of memory response that the first influenza is particularly severe for babies. This response against conserved antigens is going to become stronger following repeated exposure to the same antigen, a phenomenon called “affinity maturation”.

Imprinting at work

The importance of imprinting can be demonstrated by analyses of past pandemics. The most severe flu pandemic we know about was the Spanish flu of 1918. This was an H1N1 strain, a Group 1 strain of influenza A. The people worst affected by the 1918 Spanish flu were young adults, while older people suffered less than what would be expected. The current hypothesis is that the young people that fell victim to this pandemic were imprinted with H3N8, a group 2 strain with little antigenic resemblance to Group 1. Equally, the 2009 H1N1 influenza was less severe for old people, likely due to their exposure to other H1N1 strains that followed the 1918 outbreak. More recently, young people were worse affected by the H5N1 pandemic (a Group 2 strain), but less affected by the H7N9 pandemic (a Group 1 strain), while the opposite was true for older people. Intriguingly, young people were more likely to have been primed by the Group 1 strains H1N1 and H2N2, while old people were more likely to have been primed by the Group 2 strain H3N2.

Of course, it doesn’t come as a surprise that exposure to a similar strain confers protection against new exposure; what is really fundamental about imprinting is that it doesn’t matter (or at least not as much) how many different strains you have encountered before: what matters is which was the very first, as this is what’s going to affect your response to every other strain. Cohort studies in China following 150 people from 7- to 81-years-old also reach the same conclusion: while also multiple exposures to different strains – as well as vaccines – do have an impact, the first infection is the one that matters the most.

Is this relevant to vaccines?

While there is strong evidence that vaccinated people deal with seasonal flu much better than their unvaccinated peers, with milder episodes and fewer hospitalization cases, protection against the latest seasonal strains was about 36%. A mismatch between the circulating strains and those included in the seasonal vaccines can often account for low vaccine efficacy. More recently, vaccine production in eggs has also been blamed, as this approach introduces mutations not present in the original and circulating strains. But could imprinting also determine vaccine effectiveness? And importantly, can we harness this information to design more effective vaccines?

According to the imprinting theory, repeated annual vaccinations are bound to raise milder B cell responses compared to previous infection (or also to the first vaccination) providing a partial explanation to the observation that first-time vaccines are more protected. Also, as with any other infection, vaccination efficacy can be hampered by epitope masking and by a weak response to easily accessible antigens. However, multiple vaccinations have an important advantage: each vaccination contributes to the phenomenon of affinity maturation, making the response against conserved antigens stronger and stronger.

The race for a universal flu vaccine has never been so hot, and in light of these new immunological insights, it is important to ask what can be done better. As the most conserved epitopes are in the less accessible RBD, administration of these portions as a free antigen, unmasked by the rest of the virus, can be a valid strategy and “Headless” vaccines have been shown to induce a broadly cross-reactive response against different strains. However, given the plasticity of RNA viruses, it wouldn’t be too surprising if this was to select for escape mutation in the stalk region. Chimeric vaccines have also been suggested, where the stalk comes from the seasonal flu strain and the head from pandemic or more exotic strains.

Outlook

While a molecular understanding of imprinting has shed light on some of the population dynamics behind influenza infection, it is less clear how we can use this knowledge to design better vaccine or to predict the impact of different strains on different age groups.

The ability to sequence at the single cell level is going to be critical to understanding which epitopes are recognized by different B cells and their neutralization potential. This, in turn, can help design vaccines presenting the more immunogenic and the most conserved epitopes in a non-hindered manner to induce a strong and long-lasting immune response.

As the major antigenic differences appear to be between Class 1 and Class 2 strains, a vaccination strategy might need to identify (by testing or using epidemiological records) the most likely imprinting class and provide immunogens able to induce effective immunity against the other class, while at the same time encouraging affinity maturation of the B cells recognising the other class. The feasibility of this approach is still unclear, but it seems likely that the traditional systems of seasonal strains identification and growth in eggs will have to be replaced by molecular biology techniques.

Would this be enough to provide protection against multiple flu strains? We won’t know until we try.

At Virology Research Services we have experience with the molecular biology of RNA viruses and vaccine design. If your research involves the development of novel vaccination strategies or testing innovative hypotheses against influenza, contact us to find out how we can help.