Influenza is a highly contagious respiratory disease that causes significant symptoms and mortality worldwide, often in outbreaks and epidemics. Currently, annual vaccination is the best preventative measure to combat influenza, but constant mutations and evolutions of different strains pose significant challenges to vaccine development.  Influenza vaccines typically target around four different specific strains of the flu (out of thousands that exist) and due to this scientists are forced to predict which strains of influenza will be most prominent yearly. The RNA patterns of influenza strains and data from the past 20 years informing on transmission rates, mutations, and circulation all aid in discerning the predictability of mutations, the specific strains that will circulate in a given year, and the overall effectiveness of the countermeasures such as vaccines. 

Introduction

Influenza, mostly known as “flu”, is a highly prevalent and contagious viral infection of the respiratory passages that causes significant illness and mortality worldwide and is predominantly caused by the influenza virus A or B. Around 3%-11% of the US population each flu season contract influenza with around 8% of the US population, on average, contracting influenza and getting sick [2]. The annual vaccine remains the most effective preventative measure against influenza, but the constant evolution and mutation of influenza strains pose significant challenges to the development of effective vaccines. Despite scientists’ efforts to predict which strains will be prominent in any given year, predicting the specific strains that will circulate remains a relatively uncertain process. Scientists protect against a flu A (H1) virus, a flu A (H3) virus, a flu B/Yamagata lineage virus, and a flu B/Victoria lineage virus every year. They select these strains based off of which flu strains are making people sick prior to the start of the flu season, the extent to which those viruses are spreading prior to the start of the flu season, how well the previous season’s vaccine protected against those flu viruses, and the ability of the strains chosen for the vaccine to provide cross-protection against a range of flu related viruses of the same type or subtype [3]. In this research paper, I aim to explore the cyclic nature of influenza strains and their effects on the effectiveness of the influenza vaccine. I hypothesize that by analyzing the RNA patterns of influenza strains utilizing past data on circulation, mutations, and prevalence, it will be possible to develop more effective countermeasures against influenza strains on a yearly basis. 

History

[4] The influenza virus has caused recurrent epidemics every 1-4 years for at least the recent centuries [5]. The greatest occurrence of an influenza pandemic was in 1918, which was caused by a new H1N1 flu strain that resulted in the death of approximately 50 million people [6]. The first flu vaccines were developed in the 1930s and 1940s by researchers and they were approved for use on people who are not in the US military in 1945 [7]. In the 2009 pandemic, the influenza A (H1N1) virus had a strain with a combination of gene segments not previously reported in swine or human influenza virus strains [8]. 

Etiology/Epidemiology

Influenza is a single-stranded RNA virus from a family of viruses titled “Orthomyxoviridae” that has diverse antigenic characteristics. This family has three main types of viruses, A, B, and C. With regard to influenza, type C generally causes sporadic upper respiratory symptoms while types A and B are responsible for epidemics and outbreaks [9]. Additionally, type A influenza is typically  responsible for pandemics while types B and C only affect humans.  All viruses contain glycoproteins and an RNA gene. For the flu virus, the two most important glycoproteins are hemagglutinin (H or HA) and neuraminidase (N or NA), as both of these have important roles in the pathogenesis of the disease. 

Hemagglutinin facilitates viral attachment and entry into host cells by binding to receptors on the surface of target cells. HA also mediates the fusion of the viral envelope with the host cell membrane, which allows for genetic material to enter the cell and begin infection. HA is the primary target of antibodies produced by the host immune system, so HA is extremely important in influenza’s ability to combat antibodies - mutation of hemagglutinin means that the virus can evade preexisting immunity and leads to new influenza strains. Neuraminidase, on the other hand, is responsible for many actions after the virus has entered the host cell. NA is responsible for releasing newly formed viral particles from the host cell and spreading the virus to neighboring cells as a result. NA heavily aids in viral motility, which affects viral infectivity as a whole [10]. 

Closely related flu viruses have similar antigenic properties (and similar hemagglutinin and neuraminidase), which means that the antibodies that the immune system creates to combat one virus can combat all of these closely related viruses. However, mutation can cause these antigenic properties to change, which means that the antibodies cannot combat the virus successfully, resulting in a loss of protection. 

16 hemagglutinins and 9 distinct neuraminidases  have been recognized in influenza type A which allows scientists to divide the virus into subtypes such as H1N1 or H3N2 based on the combination patterns of the specific H or N proteins [11]. Human influenza viruses can only be made up of certain subtypes, including H1, H2, H3, and N1, N2, while other forms of influenza can be made up of nearly all of the different subtypes [12].  These subtypes can be further divided based on geographical origin, natural host species, year of isolation, and strain number. For example, a H1N1 strain  originating in ducks in Alberta, Canada could be labeled H1N1/A/duck/Alberta/35/76. The influenza B virus is very similar to type A, but has fixed antigenic characteristics and therefore has no subtypes. Influenza type B changes much more gradually than influenza type A, which is why its hemagglutinin and neuraminidase do not change. However, the influenza B virus still has small antigenic variabilities that allow scientists to classify strains into two antigenically distinguishable lineages [13]. These two lineages can be distinguished due to the different binding preferences of influenza hemagglutinin to the cellular receptor. The Victoria lineage can bind to cell receptors with a larger variety of sialic acid residues than the Yamagata lineage can [14]. 

The epidemiologic pattern of influenza is based on factors such as the changing nature of the antigenic properties of the virus, transmissibility power of the virus, and the susceptibility of the population. The strength of epidemics and its mortality rate are strongly affected by the susceptibility of the population, which some argue is one of the most important factors because of this [15]. 

Pandemic strains of influenza are often caused by animal and human-type reassortment due to the high rates of reassortment on the influenza virus’ genome and the segmented pattern of its gene. These reassortments result in antigenic shifts, which are major changes in the antigenic characteristics of hemagglutinin and neuraminidase (H and N). Minor changes of these proteins are referred to as antigenic drifts. Antigenic shifts are more commonly associated with influenza epidemics and pandemics, while antigenic drifts are associated with localized outbreaks [16]. For example, the 2009 Influenza pandemic would be associated with an antigenic shift, whereas a small outbreak in a city such as Lubbock, Texas would be associated with  an antigenic drift. Due to the fact that influenza type B only affects humans, it does not undergo antigenic shift like influenza type A does, and is only gradually changed over time through antigenic drift [17]. This is what allows type A to cause major pandemics, as it mutates much more and much faster than other types of influenza. 

Influenza typically attacks young people although older adults have a high mortality rate. Mortality is also high in cases with those with high-risk medical conditions such as cardiovascular disease, metabolic disease, or extreme age [18]. Historically, the most severe pandemics had high mortality rates among young adults. 

Methods

To focus on finding the efficiency of the vaccine and working to improve the influenza vaccine, I chose to look for patterns in strains of type A that affect humans and evaluate the strains targeted yearly in the vaccine. Type A is responsible for all pandemics and typically contributes to larger epidemics, so it is important to understand the genetic shifts and potential patterns in type A. In addition, I chose to focus on the strains that affect humans because these are the only strains that the vaccines realistically need to target. As a result, I evaluated strains H1N1, H1N2, H2N1, H2N2, H3N1, and H3N2 from type A influenza. These are the only strains in influenza type A that are human viruses. I also wanted to consider all of the strains protected in vaccines in order to accomplish my goal of evaluating and improving vaccine effectiveness, so I focused on H1N1 and H3N2 from type A, and type B Yamagata and Victoria. I  consulted the CDC (Center for Disease Control)’s archives on vaccine efficiency and flu burden to evaluate how detrimental certain flu strains are over a 10 year period and how effective the vaccine was at reducing the effects among different age groups and different strains.  I also used datasets from the National Center for Biotechnology Information in order to look for specific antigenic changes and to evaluate RNA sequences to look for patterns and potential predictability in mutations. 

The vaccine effectiveness is calculated through how well vaccines perform in real-world conditions, whereas vaccine efficacy is calculated through random controlled trials (clinical trials) and observational studies. I consolidated information regarding the efficacy and effectiveness of vaccines from 2009 to present in order to attempt to understand the issues with current/past vaccines and how they can be improved upon. I chose to evaluate vaccine effectiveness over the time period of 2009-2020 as 2009 is the earliest that detailed vaccine effectiveness data can be provided. The CDC does not keep records from before 2009 of vaccine effectiveness and the vaccines’ impact on different age groups, so I evaluated data starting in 2009 to keep the factors I was considering consistent and incorporate a holistic review. 

However, I evaluated prevalent strains and their amino acids from 2000-2020 to have a larger dataset in order to more accurately identify trends. The National Center for Biotechnology Information offers in depth information on strains, RNA sequencing, amino acids, and proteins involved in almost every strain of the flu. I used a programming tool, BLAST (Basic Local Alignment Search Tool) to compare sequences of different strains and consider how their amino acids compare to not only each other, but to the strains targeted in the vaccine for every year. BLAST can determine the similarities and inform on which sequences/segments are similar and which are different. In analyzing this data, I chose to go through every year from 2000-2020 and compare the most popular strains of that year with each other and with the strains targeted in the vaccine. 

Collecting all of the data required synthesizing from more sources than I initially thought I would need. I used the CDC in order to find information on vaccine effectiveness, vaccine efficacy, and influenza burden from 2009-2020, and I used the National Center for Biotechnology Information to directly compare amino acid sequencing and segments. The CDC’s influenza archive on past surveillance reports allowed me to find the most prevalent strains of every year and the ones targeted in the vaccine. I acquired all of my sequencing information and detailed data on strains through NCBI Influenza Virus Sequence Database, which allowed me to organize data with much more ease. 

Evaluation

[19] According to the CDC, the vaccine effectiveness has been higher than 50% only three times since 2009. While the data from 2020-2021 cannot be accurately reported due to the COVID-19 pandemic, the data from other time periods over the past 10 years clearly shows the low effectiveness of the vaccine, especially in recent years. In the early 2010s, vaccines were close to 50% effectiveness, but the vaccine severely underperformed in 2014-2015, resulting in 19% effectiveness. After the 2014-2015 influenza season, the vaccine effectiveness has been around 30-40%. This can be further explained through trends in the H1N1, H3N2, Yamagata, and Victoria strains. The line graph shows correlation between the vaccine’s effectiveness against the H3N2 strain and the vaccine’s overall effectiveness as a whole. This can be further explained through the fact that the H3N2 strain typically has the lowest effectiveness in all age groups when compared to other strains. For example, in 2018-2019, the vaccine’s effectiveness in combating H3N2 was 9% whereas its effectiveness against H1N1 was 44%. H3N2 also typically makes up a majority of influenza cases worldwide, except in a few seasons, so the prevalence of H3N2 combined with the lack of vaccine effectiveness against H3N2 ensures that the overall effectiveness of the vaccine is low. The effectiveness of the vaccine against H1N1 is also somewhat inverse to the effectiveness of the vaccine against H3N2, implying that the vaccine only protects against one of the type A strains effectively every year. 

This led me to primarily focus on H3N2 and analyzing its mutations, as it was the least protected against. I aimed to compare the similarities in the RNA sequencing between popular strains and the strain used in the vaccine using BLAST to get a calculated percent of how similar the strains were. However, I quickly learned that small differences can have major impacts. In 2014-2015, the vaccine was only 9% effective in cases involving H3N2. The strain of H3N2 used in the vaccine that year was A/Texas/50/2012, which accounted for around 36% of the total H3N2 viruses that year. The other 64% was A/Switzerland/9715293/2013, which was antigenically and genetically distinguishable from the strain used in the vaccine. Considering the low effectiveness of the vaccine, I had hypothesized that the sequencing of the two strains would be very different due to antigenic shifts and as a result the vaccine would be rendered ineffective. However, analysis showed that the RNA sequencing for the hemagglutinin for the two strains was actually 98.2% similar. The length of the sequence was 566 and the antigenic and genetic differences were the result of an accumulation in mutation between 140 and 180. Additionally, the RNA sequencing for the neuraminidase for the two strains was 99.36% similar. I decided to continue evaluating the strains and focus on potential patterns in where the mutations occur or the clusters in which they occur. 

These sequencing differences, combined with the prevalence of different influenza strains in a year, can help identify what leads to influenza pandemics. In 2009, the vaccine protected against A/Brisbane/10/2007, but 99.5% of the cases were A/California/07/2009. Additionally, these two strains had significant differences, although they were related, resulting in antigenically and genetically different strains as the hemagglutinin was only 79% similar. 

Almost all of the strains I evaluated were very similar to each other, but an accumulation in mutations resulted in distinct classifications of different strains. On average, the hemagglutinin was much more likely to have differences than the neuraminidase and typically had more severe differences than the neuraminidase. This is important because hemagglutinin is the main target of a host’s antibodies, so mutations in hemagglutinin mean that the virus may not be identified properly. Additionally, the lack of change in neuraminidase implies that the virus’s infectivity remains relatively the same across a majority of the strains. However, most of the sequencing for different strains under one subtype of influenza (i.e. H1N1 or H3N2) have almost identical amino acid sequencing, especially at the beginning and the end of the sequence. 

Originally, I had intended to synthesize all of this data in order to increase the efficiency of current prediction models to increase the accuracy of the vaccine as a result. While this was still a priority, the research I did into RNA sequencing brought a new topic up: conserved epitopes. 

Conserved epitopes are regions of viral proteins that remain relatively constant across different strains of a virus [20]. With regard to influenza, these epitopes are present in critical proteins that the virus needs to infect host cells and replicate. While strains becoming antigenically different can pose a large threat to vaccines, the small regions of proteins that remain unchanged in all of these strains bring significant implications to the table. By targeting these conserved epitopes, vaccines can induce an immune response that recognizes and neutralizes a broader range of influenza strains, even as the virus undergoes genetic changes over time.

The significance of conserved epitopes lies in their ability to confer cross-reactivity. Cross-reactivity means that an immune response generated against one strain of the virus can also recognize and provide some level of protection against other strains that share similar conserved epitopes. This is particularly important in influenza, as the virus can undergo antigenic drift or shift, leading to the emergence of new strains. By targeting conserved epitopes, vaccines can provide a level of protection against both current and future strains of influenza, reducing the overall impact of the disease. 

Conclusion and Future Work

I aimed to analyze trends in influenza strains - specifically type A strains that affect humans and the strains targeted in the yearly vaccine - and  observed a strong correlation between the vaccine’s effectiveness against the H3N2 strain and the vaccine’s overall effectiveness, which led me to focus closer on this specific strain. However, this led me to discover that the vaccine’s low effectiveness was not solely attributed to significant genetic variations. This suggests that factors beyond the similarity of protein sequencing contribute to the effectiveness of the vaccine, showcasing the need for a more comprehensive understanding of influenza strains. Through the use of my extensive dataset, which included data from the CDC and the National Center for Biotechnology Information, I was able to analyze and compare various strains of influenza over multiple years. This research provides valuable insights into the potential predictability of influenza strains and the need for improvements in the creation of the vaccine. The analysis of the data highlights the importance of continuous monitoring and analysis of influenza strains, especially those with low vaccine effectiveness such as H3N2. By identifying patterns in mutations and clusters of changes, we can anticipate potentially harmful strains in the future and enhance current predictive capabilities. 

Based on the data analysis conducted in this study, it is evident that conserved epitopes play a significant role in reducing the effects of influenza. The analysis of samples from multiple strains over a variety of years shows that these conserved epitopes remain relatively unchanged across different strains, implying that targeting these in vaccine development could lead to broader protection against various strains, including those that heavily mutate during the flu season. Furthermore, conserved epitopes offer the potential for universal influenza vaccines. Traditional seasonal influenza vaccines need to be updated regularly to match the circulating strains, as they primarily target the more variable regions of the virus and typically target only a few strains. In contrast, vaccines that target conserved epitopes have the potential to provide longer lasting immunity and have countless benefits as they could potentially reduce vaccine production and distribution needs and mitigate the impact of influenza. 

The implications of these findings are significant. By gaining a better understanding of trends in influenza strains and the limitations of current vaccines, we can work towards developing more effective vaccines that target prevalent strains more accurately. This has the potential to greatly reduce the spread and mortality of the disease, as well as minimize the burden on healthcare systems during influenza seasons. Through my exploration of patterns in genetic shifts and the analysis of extensive datasets, I have identified areas in which current vaccines fall short and highlighted areas that have potential for more accurate prediction of harmful influenza strains. The identification and targeting of conserved epitopes in influenza can lead to vaccine development that will achieve broader protection against multiple strains. This can reduce the spread and impact of the disease, ultimately contributing to an overall safer public health. 

References 

[1] Centers for Disease Control and Prevention. (2019, October 16). Images of influenza viruses. Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/resource-center/freeresources/graphics/images.htm 

[2] Tokars, J. I., Olsen, S. J., & Reed, C. (2017, December 1). Seasonal incidence of symptomatic influenza in the United States. OUP Academic. Retrieved August 23, 2022, from academic.oup.com/cid/article/66/10/1511/4682599 

[3] Centers for Disease Control and Prevention. (2022, November 3). Selecting viruses for the seasonal influenza vaccine. Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/prevent/vaccine-selection.htm 

[4] Centers for Disease Control and Prevention. (2018, August 10). Past pandemics | Pandemic Influenza (Flu). Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/pandemic-resources/basics/past-pandemics.html 

[5] Moghadami, M. (2017, January). A narrative review of influenza: A seasonal and pandemic disease. Iranian journal of medical sciences. Retrieved from ncbi.nlm.nih.gov/pmc/articles/PMC5337761/ 

[6] Johnson, N. P. A. S., & Mueller, J. (2002). Updating the accounts: global mortality of the 1918-1920 ‘Spanish’ influenza pandemic. Bulletin of the History of Medicine, 76(1), 105–115. doi:10.1353/bhm.2002.0022

[7] Mayo Foundation for Medical Education and Research. (2021). History of flu (influenza): Outbreaks and vaccine timeline. Mayo Clinic. Retrieved from mayoclinic.org/coronavirus-covid-19/history-disease-outbreaks-vaccine-timeline/flu 

[8] Johnson, N. P. A. S., & Mueller, J. (2002). Updating the accounts: global mortality of the 1918-1920 ‘Spanish’ influenza pandemic. Bulletin of the History of Medicine, 76(1), 105–115. doi:10.1353/bhm.2002.0022

[9] Poon LL, Song T, Rosenfeld R, Lin X, Rogers MB, Zhou B, et al. (2016) Quantifying influenza virus diversity and transmission in humans. Nat Genet. 48:195–200. doi: 10.1038/ng.3479.

[10] Fuller TL, Gilbert M, Martin V, Cappelle J, Hosseini P, Njabo KY, et al. (2013) Predicting hotspots for influenza virus reassortment. Emerg Infect Dis. 19:581–8. doi: 10.3201/eid1904.120903

[11] Benton, D. J., Wharton, S. A., Martin, S. R., & McCauley, J. W. (2017). Role of neuraminidase in influenza A(H7N9) virus receptor binding. Journal of Virology, 91(11). doi.org/10.1128/jvi.02293-16 

[12] Shao, Wenhan et al. (2017, August 7) Evolution of Influenza A Virus by Mutation and Re-Assortment. International journal of molecular sciences vol. 18,8 1650. doi:10.3390/ijms18081650

[13] Kanegae Y, Sugita S, Endo A, Ishida M, Senya S, Osako K, et al. (1990) Evolutionary pattern of the hemagglutinin gene of influenza B viruses isolated in Japan: cocirculating lineages in the same epidemic season. J Virol. 64:2860–5.

[14] Reina, J. (2022, February 18). [the Victoria and Yamagata lineages of influenza B viruses, unknown and undervalued]. Revista espanola de quimioterapia : publicacion oficial de la Sociedad Espanola de Quimioterapia. pubmed.ncbi.nlm.nih.gov/35180825 

[15] Moghadami M, Moattari A, Tabatabaee HR, Mirahmadizadeh A, Rezaianzadeh A, Hasanzadeh J, et al. (2010) High titers of hemagglutination inhibition antibodies against 2009 H1N1 influenza virus in Southern Iran. Iran J Immunol. 2010;7:39–48.

[16] Centers for Disease Control and Prevention. (2022, December 12). How flu viruses can change: "drift" and "shift". Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/about/viruses/change.htm 

[17] Centers for Disease Control and Prevention. (2022, December 12). How flu viruses can change: “drift” and “shift.”Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/about/viruses/change.htm

[18] Centers for Disease Control and Prevention. (2022, October 24). Key facts about influenza (flu). Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/about/keyfacts.htm 

[19] Centers for Disease Control and Prevention. (2022b, December 22). CDC Seasonal Flu Vaccine Effectiveness Studies. Centers for Disease Control and Prevention. Retrieved from cdc.gov/flu/vaccines-work/effectiveness-studies.htm 

[20] Staneková, Z., & Varečková, E. (2010). Conserved epitopes of influenza A virus inducing protective immunity and their prospects for Universal Vaccine Development. Virology Journal, 7(1). doi.org/10.1186/1743-422x-7-351 

[21] Zeigler, D. F., Gage, E., & Clegg, C. H. (2021). Epitope-targeting platform for broadly protective influenza vaccines. PLOS ONE, 16(5). doi.org/10.1371/journal.pone.0252170 

[22] Ren, J., Ellis, J., & Li, J. (2014). Influenza A HA’s conserved epitopes and broadly neutralizing antibodies: A prediction method. Journal of Bioinformatics and Computational Biology, 12(05), 1450023. doi.org/10.1142/s0219720014500231