Introduction

Influenza is an infectious respiratory disease that poses a significant public health concern, contributing to substantial morbidity and mortality worldwide. According to the World Health Organization (WHO), annual seasonal epidemics result in approximately 1 billion infections, 3 to 5 million cases of severe illness, and 290,000 to 650,000 deaths each year1. Pandemics caused by influenza type A viruses occur every 10 to 50 years, typically emerging from a new influenza A strain that is antigenically distinct from previously circulating viruses, for which the population has no pre-existing immunity, resulting in increased severity and mortality during pandemics2. The deadliest influenza pandemic, which occurred in 1918, caused over 40 million deaths worldwide3.

Influenza viruses belong to the Orthomyxoviridae family and are classified into four types: A, B, C, and D. Influenza A viruses (IAV) and Influenza B viruses (IBV) are responsible for seasonal human epidemics, presenting significant public health challenges due to their high transmissibility, potential to cause severe respiratory disease, and the substantial morbidity and mortality associated with infection. IAV are highly adaptable pathogens that infect a wide range of species, including avian, swine, humans, and waterfowl (their primary reservoirs)4. These viruses show a notable capacity to cross species barriers5. IBVs are primarily human-specific and have not been associated with pandemics yet, while they often co-circulate with IAVs, accounting for approximately 20% of global influenza cases6. Influenza C viruses (ICV) generally cause mild respiratory illness in humans and are not linked to large-scale outbreaks7. Recently discovered influenza D viruses (IDV) primarily infect cattle and pigs, with no confirmed human infections thus far8.

IAVs are enveloped, negative-sense, single-stranded RNA viruses with a segmented genome comprising eight viral RNA (vRNA) segments that encode 12 viral proteins (Fig. 1A)9. These essential components are enclosed within a lipid envelope, which features two primary glycoproteins, hemagglutinin (HA) and neuraminidase (NA), along with the membrane ion channel protein M2. Beneath the lipid membrane lies a structural layer formed by the matrix 1 (M1) protein. Inside the viral particle, additional components include the nuclear export protein (NEP), non-structural protein (NS1), and viral ribonucleoprotein (vRNP) complexes. The vRNP complexes consist of vRNAs coated with viral nucleoproteins (NP) and are associated with a viral RNA-dependent RNA polymerase (RdRp) heterotrimeric complex composed of PB2, PB1, and PA proteins.

Fig. 1: Influenza A virus (IAV) virion structure, genome organization, and the plasmid-based reverse genetics system.
figure 1

A Left: Virion Structure of IAV. The IAV virion is composed of a lipid envelope containing glycoproteins hemagglutinin (HA), neuraminidase (NA), and the M2 ion channel. Beneath the envelope lies a layer of matrix protein (M1) and nuclear export protein (NEP, not shown). Inside the virion, the viral ribonucleoproteins (vRNPs) are composed of viral RNA (vRNA) coated with nucleoprotein (NP) and associated with the polymerase proteins PB2, PB1, and PA. Right: Genome Organization of IAV. The IAV genome consists of eight segmented vRNAs: PB2, PB1, PA, HA, NP, NA, M, and NS. Each segment contains a coding region (box) flanked by untranslated regions (UTRs, short lines) at both ends. The packaging signals (Ψ) are composed of both noncoding and coding regions located at the terminal ends of each RNA segment, and are necessary for vRNP incorporation into new virions. B Plasmid-based reverse genetics for influenza A virus. The reverse genetics system involves cloning cDNA copies of each vRNA segment into plasmid vectors. These plasmids are co-transfected into permissive cells, enabling the rescue, production, and propagation of infectious virus particles.

Full size image

Given the global health threat posed by influenza viruses, the development of effective vaccines to counter both seasonal outbreaks and pandemics is crucial, particularly for protecting vulnerable populations10. The annual influenza vaccine remains the primary strategy for preventing seasonal influenza outbreaks, primarily by stimulating antibody responses against HA and NA, which are essential for viral entry and release11,12. Current vaccination approaches generally rely on two main types: inactivated influenza vaccines (IIVs) and live attenuated influenza vaccines (LAIVs), each with distinct strengths and limitations13,14. IIVs are biologically inert vaccines that include inactivated viruses or viral proteins and have been widely used for decades. They are generally safe and effective against homologous strains and are recommended for individuals aged six months and older. IIVs stimulate robust humoral immune responses against HA and NA antigens of circulating strains, generating HA-specific antibodies that neutralize the virus and prevent infection, and NA-specific antibodies that help mitigate the severity of illness15,16. However, IIVs induce limited cell-mediated immunity, particularly in vulnerable populations such as the elderly17. Furthermore, IIVs are administered intramuscularly and often require large amounts of HA antigen per dose, making their production vulnerable to shortages during periods of high demand or disruptions in egg supply caused by avian influenza outbreaks18.

Live attenuated vaccines are particularly effective because they induce both humoral and cellular immune responses that closely mimic those triggered by natural viral infections, potentially offering long-term protection without causing symptomatic disease or enabling viral transmission19,20. Following intranasal administration of LAIV, most side effects are typically mild and transient. These include runny nose, sore throat, and cough, which generally do not require medical intervention and serious adverse events are rare21.

Compared to IIVs, LAIVs elicit broader T cell responses and stronger mucosal immunity, both of which are critical for cross-protection, particularly in younger individuals or those with limited prior exposure to influenza20,22,23. The mucosal immunity induced by LAIVs, including secretory IgA production, provides protection at the site of viral entry against subsequent infection and help limit transmission24. In addition, LAIVs induce cross-reactive CD8⁺ T cells, which play a key role in protecting against heterologous influenza strains23.

Attenuation is a key process in the development of LAIVs, in which the virulence of a pathogen is reduced while its capacity to elicit an effective immune response is preserved. Traditionally, LAIVs are attenuated through cold-adaptation, in which the virus undergoes serial passages at progressively lower temperatures, ultimately restricting replication to the cooler environment of the upper respiratory tract and eliciting protective immunity25,26,27. This method underpins intranasally administered commercial LAIVs in several countries, such as FluMist, which combine a cold-adapted, temperature-sensitive, and attenuated (ca/ts/att) master donor strain with contemporary HA and NA gene segments28. However, due to variable performance across countries, populations, and influenza seasons, concerns about FluMist’s efficacy and safety have been raised, especially for its use in vulnerable groups20. LAIV is not recommended for infants under 2 years of age due to the risk of increased wheezing and hospitalization. Mild viral shedding has been observed post-vaccination, while documented transmission is rare and has not been associated with clinically significant illness29. As a result, its use is restricted to healthy individuals aged 2 to 49 years30.

Early studies suggested that LAIV offered superior protection compared to IIV in children; therefore, it was recommended for healthy children aged 2 to 8 years in 2014 by the Centers for Disease Control and Prevention (CDC)31,32,33,34,35,36. However, during the 2013–2014 influenza season, while IIV provided significant protection against the predominant 2009 pandemic H1N1 strain, LAIV4 (FluMist Quadrivalent) showed poor effectiveness in children, resulting in its suspension between 2016 and 201837,38,39. Subsequent investigations attributed this to poor replicative fitness and limited thermostability of the A/California/7/2009 vaccine strain. Structural instability in the HA protein—possibly involving residues such as E47 in the stalk domain—rendered the virus more sensitive to temperature fluctuations during storage and distribution, potentially compromising vaccine potency37,40. Additionally, prior vaccination with either LAIV or IIV may have led to pre-existing immunity, suppressing replication of the live virus and thereby limiting the induction of protective immune responses38. LAIV4 was subsequently reformulated with a more thermostable and replication-competent strain and was again reintroduced for the 2018–2019 season. This case highlights the importance of continuous evaluation of strain selection and vaccine performance.

Despite the widespread use of current influenza vaccines, their effectiveness typically remains below 60% and can drop to as low as 10%, with substantial variability across seasons and populations41,42. Moreover, LAIVs do not provide lifelong immunity, as influenza A viruses evolve rapidly through genetic reassortment and antigenic drift, accumulating mutations in the HA and NA proteins43. These changes facilitate the emergence of novel variants with enhanced replication, transmission, or immune evasion capabilities, leading to frequent mismatches between vaccine strains and circulating viruses, thereby compromising protective immunity44,45,46,47,48. Consequently, the efficacy of seasonal vaccines often declines over time, posing significant global health and economic challenges and necessitating frequent reformulation. In addition, traditional attenuation methods for LAIVs raise safety concerns, as there is still a risk of unpredictable virulence reversion49. Another major challenge lies in the empirically based and time-consuming development and manufacturing processes of current influenza vaccines, which limit the production capacity. These challenges underscore the urgent need for next-generation vaccine platforms capable of overcoming limitations related to manufacturing efficiency, antigenic variability, and safety.

Innovations in biotechnology and synthetic biology have provided new tools for the rational design of next-generation live attenuated influenza A vaccines. These novel approaches utilize precise genomic modifications of viral genomes, enabling a faster and more controlled pathway to developing safe and effective live attenuated vaccines50. Here, we review next-generation attenuation strategies to develop live attenuated influenza A vaccines, highlighting the transformative potential of genetic engineering and synthetic biology to revolutionize vaccine development (Fig. 2, Table 1). We also offer perspectives on how these advances may help address future global health challenges.

Fig. 2
figure 2

Representative rational design strategies for developing live attenuated influenza A vaccines.

Full size image
Table 1 Overview of novel LAIV platforms
Full size table

Rational design strategies for live attenuation vaccine development

Reverse genetics for influenza vaccine development

Over 80% of seasonal and pandemic influenza vaccines are produced using egg-based systems, with inactivated influenza vaccines being the majority51. However, egg-based methods can introduce adaptive mutations that may reduce vaccine effectiveness. They are also vulnerable to disruptions in egg supply and offer limited manufacturing flexibility, potentially resulting in delays or shortages in vaccine deployment. To address these challenges, cell culture–based production systems have emerged as a more scalable and adaptable alternative for vaccine manufacturing52,53.

Plasmid-based reverse genetics is an instrumental technique in modern virology and represents a foundational advancement enabling both cell-based live vaccine production and rational LAIV design54. Reverse genetics allows precise manipulation of the viral genome and has significantly facilitated the generation of genetically modified LAIVs. For virus production, plasmids containing full-length complementary DNA (cDNA) encoding the segmented RNA genome are transfected into continuous cell lines that support efficient viral replication55,56. For example, a commonly used reverse genetics system for influenza A virus employs eight plasmids, each encoding one of the virus’s eight vRNA segments. Upon transfection, these plasmids initiate transcription of viral RNA and translation of viral proteins, leading to the production of infectious influenza A virus (Fig. 1B). The virus is then harvested from the culture supernatant and propagated for further analysis. By modifying these plasmids, researchers can introduce specific mutations, deletions, or insertions into the viral genome. Meanwhile, rescuing rationally attenuated vaccine strains in engineered specialized cell lines enables rapid virus replication; such systems are essential for large-scale vaccine manufacturing. These advancements have significantly enhanced the rational design of attenuated influenza viruses, paving the way for the development of safer and more effective vaccine candidates to support future vaccination strategies.

NS1-deficient or truncated influenza viruses

The NS1 protein of influenza A virus is a nonstructural RNA-binding protein that has multiple interactions within the infected cell, all of which favor viral multiplication. NS1 antagonizes the host’s type I interferon (IFN) antiviral defense response and facilitates the translation of viral mRNA, playing a pivotal role in the virus’s pathogenicity57,58. Consequently, recombinant IAV strains with truncated NS1 or those lacking the NS1 gene (ΔNS1) replicate poorly in IFN-competent hosts, serving as promising candidates for LAIV. Several studies have explored these NS1-modified strains as vaccine candidates across multiple species, including equine, swine, avian, canine, and human models59,60,61,62,63,64,65,66,67,68,69,70,71,72. Using reverse genetics, stop codons or gene deletions were introduced into the NS gene to abolish the expression of a fully functional NS1 protein, generating IAV strains with deleted or truncated NS1 in IFN-deficient cells. Following immunization, the inability of NS1-deficient viruses to inhibit IFN production results in a rapid and transient induction of antiviral responses at the site of infection in IFN-competent hosts. This contributes to virus attenuation in vivo, while the remaining viral components elicit robust immune responses and provide protective immunity against the wide-type (WT) virus.

In 1998, García-Sastre et al. used the IAV reverse genetics system to delete the gene segment encoding the NS1 protein in the influenza A/PR/8/34 virus (PR8) without compromising the expression of the nuclear export protein (NEP)73. The resulting mutant viruses lacking the NS1 protein (ΔNS1) were attenuated in MDCK cells but replicated efficiently in IFN-deficient Vero cells. Subsequent studies investigated different truncation sites within NS1 to balance attenuation and immunogenicity. For example, Steel et al. developed NS1-deficient vaccine candidates for the highly pathogenic avian influenza A/Vietnam/1203/04 (VN1203, H5N1) virus with various C-terminal truncations of NS1 (expressing the first N-terminal 126, 99 or 73 amino acids, NS1 1–126, NS1 1–99, and NS1 1–73) (Fig. 3)72. Additional modifications, including the removal of the polybasic cleavage site in HA and mutations in PB2 (K627 or E627), were introduced to enhance safety. These recombinant viruses grew to high titers in 10-day-old embryonated chicken eggs and exhibited significant attenuation in IFN-competent cell cultures and mouse models, inducing high levels of beta interferon. Notably, variations in NS1 truncation sites had minimal effects on viral growth kinetics. Vaccination with a single dose of NS1-truncated recombinant viruses provided complete protection against lethal challenge with a mouse-adapted H5N1 virus. In chickens, a single dose of a representative strain encoding NS1 1–99 conferred protection against both homologous and heterologous highly pathogenic avian influenza viruses.

Fig. 3: Schematic diagram of gene segments of the NS1-deficient virus.
figure 3

The WT NS segment encodes the full-length NS1 protein and the NEP protein. For the engineered NS1-truncated virus, stop codons were introduced into the NS gene segment after the first 126, 99, or 73 amino acids of the NS1 protein, while the NEP protein remained unaffected.

Full size image

NS1-deficient or truncated IAVs showed great promise and broad applicability as potential LAIV candidates. The risk of reversion to a WT or virulent phenotype is minimal due to gene deletions and limited in vivo replication capacity, enhancing vaccine safety. Some NS1-truncated strains induce a stronger immune response and provide better protective efficacy compared to inactivated virus vaccines in aged mice74,75. The tolerability and immunogenic potential of monovalent ΔNS1 vaccines have been evaluated in clinical trials. A phase I/II trial of a trivalent ΔNS1 vaccine candidate, including H1N1, H3N2, and B strains, demonstrated safety and high antibody induction71,76,77. Importantly, production of NS1-deficient or truncated influenza viruses requires IFN-deficient cells (e.g., Vero cells) or systems with underdeveloped IFN responses, such as 5–6-day-old chicken embryonated eggs. This strategy was also successfully applied to generate live vaccines for influenza B virus78,79. However, a recent study provided evidence that a swine LAIV encoding a truncated NS1 reassorted with a circulating wild-type strain, indicating that viral reassortment is possible with NS1-deficient LAIV. This finding underscores the need for further studies and routine viral surveillance to assess and mitigate potential risks80,81.

M2-deficient influenza viruses

The M2 protein of influenza A virus is a multifunctional component crucial for viral replication and assembly. The M2 protein comprises three structural domains: an extracellular domain, a transmembrane domain, and a cytoplasmic tail domain. The transmembrane domain functions as an ion channel82. The cytoplasmic tail plays a pivotal role in viral assembly, contributing to particle formation and budding83,84,85. Disruption of M2 protein has been explored in several studies as a promising approach for LAIV development.

Early studies demonstrated that deletions within the M2 cytoplasmic tail caused a growth defect in the influenza A/WSN/33 (WSN, H1N1) strain of IAV, highlighting the importance of this domain for efficient viral production83,84,85. Building on this knowledge, in 2008, T. Watanabe and colleagues engineered M2 cytoplasmic tail mutants to attenuate the highly pathogenic influenza A/Vietnam/1203/04 (VN1203, H5N1) virus86. By introducing stop codons to abolish gene expression within the M2 protein’s cytoplasmic tail, the researchers generated a mutant virus with an 11 amino acid tail deletion that exhibited growth kinetics comparable to the WT virus in M2-overexpressing cells but significantly attenuated in vivo (M2del11, Fig. 4). To enhance safety, they also modified the HA cleavage site, replacing it with a sequence from an avirulent avian virus to reduce protease cleavage efficiency. The resulting virus provided protective immunity against both homologous and heterologous H5N1 strains in mice. Following this success, the team extended their strategy to the 2009 pandemic H1N1 virus, producing a mutant strain that elicited robust immunity and fully protected mice from lethal viral challenges87.

Fig. 4: Schematic diagram of gene segments of the M2-deficient virus.
figure 4

The M gene of the M2del11 virus (VN1203, H5N1) has stop codons introduced to abolish the expression of 11-amino-acid (aa) in the cytoplasmic tail. The M gene of the M2KO virus (PR8) has stop codons introduced downstream of the M1 open reading frame, abolishing the expression of the M2 transmembrane and cytoplasmic tail domains. For the M2SR virus, two stop codons were introduced downstream of the M1 open reading frame, followed by deletion of nucleotides in the transmembrane domain.

Full size image

In 2009, S. Watanabe and colleagues developed the M2 ‘knockout’ (M2KO) virus using the A/Puerto Rico/8/34 (PR8, H1N1) strain88. This mutant was generated by introducing stop codons that abolished the expression of both the M2 transmembrane and cytoplasmic tail domains. The resulting M2KO virus showed reduced virulence in cell cultures and animal models while inducing robust protective immune responses in vivo, demonstrating its potential as a live attenuated influenza vaccine.

Further advancements were achieved in 2016 with the development of the M2-deficient single-replication (M2SR) virus89. This virus, derived from the PR8 strain, contains stop codons as in the M2KO design, along with gene deletions in the M2 transmembrane domain. The M2SR virus retains the ability to infect cells and express all viral proteins except the M2 protein. Due to the lack of functional M2, it cannot produce progeny virus, limiting its replication to a single cycle. For virus manufacture, M2SR viruses can be generated in a cell line that stably expresses the complementing M2 protein.

The M2SR virus demonstrated strong cross-protective immunity against diverse influenza strains. It elicited strong systemic and mucosal antibody responses and protected mice and ferrets against both homologous and heterologous influenza strains, including H3N2 and H5N1 strain90. Its heterosubtypic protection possibly involves cross-reactive, non-neutralizing antibodies targeting conserved regions in HA2 (stalk) or NA, cross-reactive T cells or cytotoxic CD8⁺ T cells. Also, M2SR induced only low to moderate cytokine responses without triggering the high levels of cytokines as observed during severe or lethal influenza infections which could lead to lung damage89,91.

Comparative studies of M2SR and FluMist showed that when comparing immunogenicity in ferrets with pre-existing immunity to naïve animals, M2SR immunogenicity was not reduced, whereas immune responses to FluMist were significantly reduced in pre-infected ferrets compared to naïve ferrets92. This suggests the immunological advantages of M2SR over FluMist. Furthermore, the M2SR vaccine conferred protection against drifted H1N1 and H3N2 strains in ferrets with pre-existing immunity, highlighting its versatility92.

The M2SR influenza A vaccine has since progressed to clinical trials, showing promising safety and providing enhanced mucosal and serum antibody responses in adults93,94,95. Participants demonstrated durable serum antibody responses against diverse influenza strains, suggesting broad cross-protection. Importantly, the M2SR design strategy was extended to develop BM2-deficient, single-replication vaccines against influenza B virus96. These findings highlight the potential of M2-deficient influenza viruses as innovative LAIV candidates with improved safety and cross-protective efficacy.

Viral genome rearrangement for LAIV development

The viral genome is a highly organized system, and disrupting its natural arrangement can result in inefficient gene translation, leading to compromised viral packaging and reduced replication. Genome rearrangement of influenza A virus has emerged as a promising strategy for developing LAIVs. This technique enables the generation of recombinant influenza viruses with attenuated replication or the creation of virus vectors that stably express foreign genes to enhance immunogenicity. In this approach, the viral genome is rearranged to incorporate foreign genes—such as green fluorescent protein (GFP), or different influenza HA subtypes—which can be effectively cloned into new chimeric virions using influenza packaging signals and expressed in recombinant viruses97,98,99,100. For viral genome rearrangement, the ORF of the foreign protein either replaces an original viral protein ORF, is included downstream of the original ORF via a 2A autoproteolytic cleavage site, or is expressed in an additional segment, sometimes accompanied by the deletion of the original viral protein ORF and complemented by viral protein-expressing cells. These strategies often result in the creation of immunogenic, replication-deficient recombinant viruses, holding potential to develop novel attenuated vaccine candidates.

For example, the Kawaoka group developed a replication-incompetent PB2-knockout (PB2-KO) IAV, where the PB2 ORF was replaced with a GFP gene (Fig. 5A)101,102. This virus replicated efficiently only in PB2-expressing cell lines, while being highly attenuated and stimulating immune responses in vivo. The authors also demonstrated the functional expression of different subtypes of HA and NA using the PB2-KO virus as a vector103,104. Similarly, HA-knockout IAV vaccine candidates were developed by replacing the HA ORF with GFP using various virus strains as the backbone, including the pandemic influenza A/California/04/2009 (pH1N1) and A/HKx31(X31, H3N2) strains (Fig. 5B)105,106,107.

Fig. 5: Schematic diagram of gene segments in recombinant LAIV candidates.
figure 5

A–C Schematic representations of the PB2 (A), HA (B), and NA (C) gene segments from different rearranged viruses, where the original viral open reading frames (ORFs) have been replaced by foreign gene ORFs. D PB1 mutant segment from a recombinant strain containing nine RNA segments expressing an additional foreign HA. The PB1 ORF is flanked by the NA packaging signal, alongside a ninth chimeric segment with the foreign HA gene from the HK strain, flanked by the PB1 packaging signal. E Rearranged H9N2 influenza A virus segments expressing an additional foreign HA. The NEP protein is positioned downstream of the PB1 gene with an autoproteolytic cleavage site (2A), and the foreign HA gene is expressed downstream of a truncated NS1 gene (NS1 1–99) with a 2A site. F Rearranged M and NS gene segments that enable independent expression of the M1/M2 and NS1/NEP proteins with 2A sites. Schematic of the Ms, NSs, and Ms/NSs gene segments. G NEP gene ORF is rearranged downstream of a truncated NS1 gene (NS1 1–128) with a 2A site.

Full size image

In 2013, Masic et al. developed an eight-segment swine influenza A/Swine/Saskatchewan/18789/02 (SK02, H1N1) virus that harbored both H1 and H3 HA subtypes (Fig. 5C)108. This was achieved by removing the NA ORF and replacing it with the ORF of H3 HA from the A/Swine/Texas/4199-2/98 (H3N2) strain, flanked by NA packaging signals. This resulted in an NA knockout virus, with its NA activity supplemented exogenously. The resulting H1-H3 bivalent virus was highly attenuated in pigs, elicited robust immune responses, and provided complete protection against infections with both H1 and H3 subtypes109.

In another study, Gao et al. rearranged H1 and H3 HA subtypes into the surface of a recombinant influenza A/Puerto Rico/8/34 (PR8, H1N1) strain containing nine RNA segments (Fig. 5D)99. Instead of using PB1 packaging signals, the PB1 ORF was flanked by the packaging signals of NA. The H3 HA ORF from influenza A/Hong Kong/1/68 (A/HK/1/68) was flanked by PB1 packaging signals and constructed into a ninth segment, enabling its expression alongside the H1 HA in the PR8 viral envelope. Vaccination with this recombinant virus resulted in significant protection against both H1 and H3 strains, demonstrating the potential of this approach to develop bivalent vaccines. Additionally, it was shown that the incorporation of the ninth segment could also be achieved by manipulating the packaging signals of the PB2 or PA segments, further emphasizing the adaptability and versatility of this genome rearrangement method.

Genome rearrangement has also been employed to develop LAIVs targeting highly pathogenic IAVs with pandemic potential, such as H5N1 and H9N2. For instance, Pena et al. engineered an avian influenza A/Guinea fowl/Hong Kong/WF10/99 (GFHK99, H9N2) virus, which also expresses the HA of the highly pathogenic A/Vietnam/1203/04 (VN1203, H5N1) strain (Fig. 5E)110. This was achieved by inserting the NEP ORF downstream of the PB1 gene via a foot-and-mouth disease virus (FMDV) 2A autoproteolytic cleavage site, enabling the co-translational release of NEP from the PB1-2A chimeric protein. The H5 HA ORF was then cloned downstream of either a full-length or truncated NS1 gene (NS1-99), with a 2A site in between. The resulting recombinant virus, H9N2-H5, successfully expressed both H9 and H5 HA proteins, confirming the incorporation of the H5 HA gene into the rearranged vector. Due to decreased polymerase activity, the H9N2-H5 virus was highly attenuated and exhibited significantly reduced titers and replication compared to the WT virus or non-rearranged viruses with the same NS1 deletion. In ferrets and mice, the rearranged vaccine demonstrated protective efficacy against both H5N1 and H9N2 strains. Importantly, because the H5 ORF was expressed using packaging signals from the NS gene, the likelihood of reassortment of the H5 HA was minimized, thereby enhancing the safety of the vaccine candidate.

Viral protein genome rearrangement has also been used to independently express the M1/M2 protein in the M gene and the NS1/NEP protein in the NS gene, generating attenuated virus vaccine candidates. In 2016, Nogales et al. rearranged the M and NS genes of influenza A/Puerto Rico/8/1934 (PR8, H1N1) virus, inserting a porcine teschovirus 1 (PTV-1) 2A autoproteolytic cleavage site to separate the overlapping ORFs of M1/M2 and NS1/NEP (Fig. 5F)111. Three types of rearranged viruses were created: one with independent M1/M2 (split M segment, Ms), another with independent NS1/NEP (split NS segment, NSs), and a third combining both (Ms/NSs). Viruses with the rearranged M segment displayed a temperature-sensitive phenotype, replicating efficiently at a permissive temperature (33 °C) while remaining attenuated in vivo. These viruses retained strong immunogenicity and conferred protection against a lethal challenge with the WT virus, suggesting their potential as safe and protective LAIV candidates.

In 2021, Lee et al. developed recombinant swine influenza viruses expressing the HA and NA genes of A/Swine/Texas/4199-2/98 (H3N2) and the six internal genes of the bat influenza A virus bat/Guatemala/164/2009 (Bat09, H17N10)64. Additionally, the NS1 gene was truncated at the C-terminal (NS1 1-128) followed by a 2A autoproteolytic cleavage site and an ORF encoding NEP to separately express NEP from NS1 (Fig. 5G). Pigs vaccinated with this strain showed significantly reduced nasal virus shedding, viral replication in the lungs, and associated lesions. Furthermore, the vaccine induced stronger mucosal IgA responses and a higher number of IFN-γ-secreting cells. This study further supports the potential of NS1-truncated LAIVs and genome rearrangement as promising strategies for controlling influenza infections across different species, including swine and potentially other animal hosts.

Collectively, these studies highlight the potential of viral genome rearrangement as a strategy for developing safe and effective LAIVs. By disrupting the native organization of the viral genome and enabling the expression of foreign genes to achieve attenuation, this approach supports the development of multivalent vaccines and the targeting of highly pathogenic strains, with broad applicability across species. However, a major limitation of this strategy is its dependence on detailed knowledge of viral gene functions and genome packaging mechanisms. Furthermore, the need for specialized production cell lines and extensive empirical optimization may extend vaccine research and development timelines.

Modification of the HA protein cleavage site for LAIV development

The influenza HA protein of IAV is cleaved by trypsin-like serine proteases in the host into two subunits. This cleavage facilitates the fusion of the viral and host cellular membranes, a crucial step for viral entry and necessary for the virus to gain infectivity112,113. In 2005, Stech et al. engineered an elastase-dependent mutant of the influenza A/WSN/33 (WSN, H1N1) virus. The trypsin-specific cleavage site (Arg-Gly) in the HA protein was modified to an elastase-sensitive site (Val-Gly), enabling cleavage by pancreatic elastase, which is absent in the host under natural conditions (Fig. 6)114. The resulting mutant virus, WSN-E, was generated in the presence of elastase and exhibited growth kinetics similar to the WT virus when elastase was provided. In contrast, in normal cells lacking elastase, WSN-E was resistant to trypsin cleavage and could not be processed into its functional HA form, thereby exhibiting attenuation. In mouse models, the mutant virus showed reduced replication but elicited immune responses comparable to the WT virus and conferred protection against influenza challenge. As the virus lacked sufficient time and population size to mutate back to a trypsin-cleavable form, the authors concluded that the mutant had a low reversion rate.

Fig. 6: Schematic representation of gene segments of the HA cleavage site-modified virus.
figure 6

Arginine at the HA cleavage site of the WT virus was mutated to valine to generate the WSN-E mutant virus, which shifts the cleavage dependency from trypsin to elastase.

Full size image

This strategy has also been applied to generate highly attenuated elastase-dependent swine influenza virus (SIV) mutants, which induced robust cell-mediated and mucosal antibody responses, providing strong protection against both homologous and heterologous influenza A virus challenges in pigs115,116,117. These studies highlight the potential of modifying influenza viral gene segments to generate elastase-dependent mutants as live-attenuated vaccines, demonstrating that rational genetic mutations can be used to precisely manipulate viral genomes and proteins for vaccine development.

MicroRNA-based attenuation for LAIV development

Efficient translation of viral RNA is essential for viral replication, making the disruption of RNA function a promising approach for viral attenuation. MicroRNAs (miRNAs) are endogenous non-coding RNAs that regulate gene expression through the RNA-induced silencing complex (RISC), where miRNAs bind to complementary mRNAs and lead to translational repression or mRNA degradation118. The incorporation of miRNA recognition elements (MREs) into viral genomes has emerged as a promising approach for achieving viral attenuation and has been applied to develop attenuated vaccines against several viruses, including influenza virus, poliovirus, Dengue virus and flavivirus119,120,121.

In 2009, Perez et al. engineered an attenuated influenza A/Puerto Rico/8/34 (PR8, H1N1) strain by introducing MREs into the viral genome (Fig. 7)122. They incorporated MREs targeted by miR-93, a miRNA highly expressed in human and murine lung tissue but absent in chickens, into the viral NP open reading frame (ORF). The incorporation of near-perfect MREs was performed without altering the physical properties of the amino acids at these sites. Reassortant viruses expressing the HA and NA of influenza A/Vietnam/1203/04 (VN1203, H5N1) strain with MRE-containing NP segments were also generated. These modifications resulted in highly attenuated mutant viruses in mammalian hosts, while allowing good replication in embryonated eggs. Mechanistic studies confirmed that attenuation was miR-93-specific and resulted from translational repression. The engineered H1N1 strains and H5N1 reassortants elicited robust antibody responses in vivo and provided complete protection against lethal viral challenges.

Fig. 7: Schematic of the microRNA-based virus gene segments.
figure 7

MicroRNA recognition elements (MREs) were introduced either into the ORF of the NP gene or downstream of the NP stop codon and upstream of the packaging signals.

Full size image

Subsequent studies incorporated additional miRNA target sites into various influenza viral gene segments, including the insertion of miR-21, miR-192 and miR-142 target sites into the NP gene, miR-192 target sites into the HA gene, and miR-let-7b target sites into the PB1 gene123,124,125,126,127,128. Interestingly, some of these later studies positioned multiple MREs downstream of the viral protein stop codon and upstream of the duplicated packaging sequence (Fig. 7). A potential concern with miRNA-based attenuation is the risk of escape mutants. Although such mutants were not detected in these studies, future research is encouraged to further enhance the vaccine safety, such as by targeting multiple gene segments using either the same or distinct species-specific miRNAs. These movements may also improve the overall attenuation of the viral mutants.

Collectively, these findings underscore the versatility of miRNA-mediated attenuation as a powerful tool for developing LAIVs, leveraging tissue- or species-specific miRNA expression to precisely control viral replication.

Codon or codon pair deoptimization for LAIV development

Influenza viruses rely on the host’s protein synthesis machinery to translate viral mRNAs into viral proteins, which then assemble into new viral particles alongside newly synthesized viral ribonucleoproteins (vRNPs). Downregulating the translation and expression efficiency of viral proteins can slow viral assembly and lead to attenuation. The codon deoptimization strategy reduces viral protein translation efficiency by exploiting codon usage bias, which refers to the preference for specific codons to encode the same amino acid. Suboptimal synonymous codons, which are underrepresented in human cells, are introduced into the viral genome, reducing translation efficiency without altering the viral protein sequence or function.

Viral genome codon usage deoptimization has been successfully applied to generate an attenuated poliovirus with reduced viral replication and infectivity129,130. Later, Nogales et al. applied this strategy to deoptimize the influenza A virus genome, introducing synonymous codon-deoptimized mutations into the NS gene segment of the influenza A/PR/8/34 (PR8, H1N1) strain (Fig. 8)131. The most attenuated strain, NScd virus, which included 113 (32%) codon substitutions, exhibited reduced virulence in mice while replicating efficiently in the MDCK cell line. This deoptimized virus retained its immunogenicity, providing protection against both homologous and heterologous influenza challenges in mice, demonstrating its potential as a safe live attenuated vaccine. In another study, Fan et al. introduced over 300 silent codon-deoptimized mutations into the eight gene segments of the influenza A/Brisbane/59/2007 (BR59, H1N1) virus132. The resulting virus showed significant attenuation in both mammalian cells and mice, while enabling robust growth in embryonated eggs. This deoptimized virus elicited strong immune responses and provided protection against both homologous and heterologous viral challenges.

Fig. 8: Schematic of the codon-deoptimized (cd) virus gene segments.
figure 8

Codon-deoptimized mutations (red boxes) were introduced into the NS gene, which encodes both the NS1 and NEP proteins.

Full size image

Another strategy, codon pair deoptimization, leverages codon pair bias, which refers to the uneven use of synonymous codons for adjacent amino acids within genomes133. The Wimmer group developed the ‘synthetic attenuated virus engineering’ (SAVE) method, which systematically introduces suboptimal or rarely used codon pairs into the viral genome, reducing translation efficiency and attenuating the virus without altering protein sequences134,135. In 2010, Mueller et al. applied SAVE to deoptimize codon pairs in the NP, PB1, and HA genes of influenza A/PR8/34 (PR8, H1N1) strain using a computational algorithm135. A single intranasal dose of this codon pair-deoptimized virus protected mice from the WT virus challenge. In 2013, Yang et al. applied SAVE to deoptimize the NA and HA genes of influenza A PR8 virus, achieving viral attenuation, long-lasting immunity, and heterologous protection136. Similarly, Stauft et al. deoptimized the HA and NA genes and developed CodaVax-H1N1, a vaccine candidate for the 2009 H1N1 pandemic virus, which demonstrated strong efficacy in animal models and is currently in phase I/II trials137.

Codon and codon pair deoptimization have enabled the rapid development of live attenuated vaccines with reduced viral replication and pathogenicity in hosts without compromising immunogenicity. These viruses are highly resistant to virulent reversion due to the extensive mutations, making them much safer. Additionally, this approach requires minimal prior knowledge of viral protein functions, which is crucial for responding to infectious disease outbreaks, and can be further accelerated with computational tools and advancements in de novo DNA synthesis138. Codon and codon pair deoptimization have been successfully applied to other pathogens, including respiratory syncytial virus (RSV), arenavirus, foot-and-mouth disease virus and severe acute respiratory syndrome coronavirus-2 virus (SARS-CoV-2), yielding safe and effective live attenuated vaccine candidates139,140,141,142,143. Several of these candidates, such as RSV CodaVax (RSV), and CoviLiv (SARS-CoV-2), have entered clinical trials144,145. However, as noted by the authors, optimizing the balance between attenuation and immunogenicity remains a challenge, potentially increasing production timeline and costs. Despite these challenges, genome deoptimization remains a promising strategy for the rapid, robust, and safe development of vaccines.

Genetic code expansion strategy for LAIV development

The efficient synthesis and expression of viral proteins are crucial for the viral life cycle. A novel approach to attenuating viruses involves controlling viral replication by manipulating the biosynthesis of essential viral proteins through genetic code expansion. This method introduces reassigned codons —encoding unnatural amino acids (UAAs) within an orthogonal translation system—into viral genes146. In normal cells, viral protein translation is halted at the UAA codon site, leading to the production of dysfunctional viral proteins and attenuation. For vaccine production, an engineered transgenic cell line with an orthogonal translation system is required, where a specific tRNA synthetase/tRNA pair recognizes the UAA, enabling continued translation at the UAA codon site and ensuring the expression of functional viral proteins147. This technique provides precise control over viral replication. Consequently, replication-incompetent, highly attenuated viruses can be generated as promising candidates for live attenuated vaccines. Genetic code expansion has been successfully applied in engineering lentiviral vectors, adeno-associated virus 2 (AAV2) vectors, hepatitis D virus (HDV), and in the development of live attenuated vaccines for human immunodeficiency virus type 1 (HIV-1), Zika virus, and pseudorabies virus (PRV)148,149,150,151,152,153,154.

Zhou et al. employed this strategy to engineer influenza A/WSN/1933 (WSN, H1N1) virus harboring premature termination codons (PTCs) by introducing amber stop codons at permissive sites within viral protein genes (Fig. 9)155. The sites of mutation were evaluated across the viral genome. To enhance safety and minimize the risk of codon reversion or viral escape, four amber codon mutations were introduced into non-envelope genes (PA, PB2, PB1, and NP). This resulted in the replication-incompetent PTC-4A strain, which retained the native conformation of surface antigens to ensure immunogenicity. To generate PTC viruses, manufacturing must be conducted in specialized transgenic cell lines engineered with the orthogonal translation system with the respective tRNA/tRNA synthetase pair and supplemented with the corresponding exogenous UAA. In such system, PTC-4A viral proteins are efficiently translated, allowing virus propagation at levels comparable to the WT strain. In contrast, in conventional cells lacking the orthogonal translation system or UAA, PTC viruses are fully infectious but replication-deficient. Mice inoculated with PTC-4A showed no influenza-associated symptoms, and in guinea pigs, no transmission occurred between inoculated and uninfected animals, further supporting the safety profile of the PTC virus. Moreover, the PTC virus exhibited an inhibitory effect on WT virus propagation through genetic reassortment which neutralized replicating WT viruses. PTC-4A elicited robust immune responses in vivo, providing complete protection against homologous WT virus challenge and cross-protection against antigenically distinct strains. Later, this virus was repurposed as a delivery vector for the development of lung cancer vaccines156.

Fig. 9: Schematic representation of the generation of premature termination codon (PTC) influenza viruses using the genetic code expansion strategy.
figure 9

PTC-viruses are generated in transgenic cell lines that express an orthogonal tRNA (tRNACUA), tRNA synthase (pylRS), and a gene encoding an amber codon–containing GFP (GFP39TAG). In the presence of an unnatural amino acid (UAA), the PTC-viruses replicate effectively in transgenic cells. In conventional cells without the orthogonal translation machinery, the insertion of PTCs halts the translation of relative viral proteins, which subsequently attenuates the PTC-viruses.

Full size image

The genetic code expansion strategy enables the generation of PTC-harboring, replication-incompetent viruses with enhanced safety profiles and strong immune responses while minimizing the risk of viral reversion. These findings highlight the potential of the genetic code expansion strategy to develop PTC viruses, which could be used as prophylactic vaccines and therapeutic tools for controlling actively replicating viruses.

Transposon mutagenesis-based strategy for LAIV development

Previous viral attenuation strategies have demonstrated that genetic modifications in influenza gene segments can attenuate the virus. An alternative approach utilizes phenotype-based screening of genetically modified strains to identify attenuated candidates. In 2017, Wang et al. used a transposon mutagenesis system in combination with the reverse genetics system to rapidly generate and screen viral clones with attenuated properties in vivo (Fig. 10)157.

Fig. 10: Schematic diagram of the generation of attenuated viruses through transposon mutagenesis and identification of mutant candidates based on in vivo growth screening.
figure 10

Constructs from transposon mutagenesis of the M gene were co-transfected with the remaining WT viral gene plasmids. The resulting mutant virus library was subjected to in vivo screening and genotyping to identify insertion mutations and attenuated mutant strains.

Full size image

Transposons are widely used for constructing mutants by facilitating gene disruption or allele replacement, thus enabling the investigation of gene function and protein structure-function relationships158. The Mu-transposon-mediated insertional mutagenesis platform, previously applied to functionally profile the Hepatitis C virus genome, was adapted by Wang et al. to develop live influenza vaccines159. 15-Nucleotide sequences were introduced into the M gene of influenza A/WSN/1933 (WSN, H1N1), and the resulting mutant M gene library was combined with the seven complementary segments of WT WSN using reverse genetics and co-transfected into HEK293T cells to generate a mutant virus library. This library was injected into mice for in vivo growth screening, followed by high-throughput genotypic analysis to identify regions of mutation associated with distinct viral growth profiles. Notably, mutants from Cluster C, which had insertions in the M2-exclusive region, exhibited in vivo attenuation. Further evaluation of a candidate strain, W7-791, which had an amino acid insertion in the cytoplasmic tail of M2, showed that it retained the attenuated phenotype after multiple passages and did not revert to the WT form. In animal models, W7-791 induced significantly less severe pathology than the WT virus while eliciting robust immune responses. In vivo, W7-791 induced both humoral and cellular immunity, including cross-protective CD8+ T cell responses, with one single dose conferred protection against lethal homologous and heterologous influenza challenges, including H1, H3, and H5 strains.

This study highlights the feasibility of using transposon mutagenesis to generate attenuated influenza virus libraries and identify promising LAIV candidates through high-throughput screening. The approach offers a robust platform for developing vaccines with minimal prior knowledge of viral gene functions required. It is particularly useful for profiling less-understood or emerging viruses.

Interferon-sensitive influenza viruses

The influenza virus has evolved mechanisms to suppress interferon (IFN) production and function to facilitate viral replication. Inhibiting these IFN-modulating functions can sensitize the virus to IFN responses and lead to viral attenuation. The influenza NS1 protein is a well-characterized IFN modulator, and deleting NS1 (ΔNS1) has shown promise as a vaccine candidate in clinical trials160. In addition to NS1, other viral proteins, including components of the polymerase complex (PB2, PB1, PA, NP) and M1/M2, may also modulate IFN responses.

In 2018, Du et al. developed an approach to eliminate IFN-modulating functions across the influenza A/WSN/33 (WSN, H1N1) genome without compromising replication (Fig. 11)161. They developed a high-throughput genomic system combining saturation mutagenesis with next-generation sequencing to assess viral replication fitness and IFN sensitivity across the genome162. By comparing viral mutants with and without IFN selection, they identified mutations that significantly increased IFN sensitivity compared to the WT virus. Based on these, they engineered a hyper-interferon-sensitive (HIS) virus with eight mutations—three in PB2, three in M1, and two previously identified IFN-inducing mutations in NS1. The HIS virus replicated efficiently in IFN-deficient cells but was highly attenuated in IFN-competent cells and in vivo. It also induced higher IFN responses than the WT virus or the virus with the mutation only in NS1, demonstrating the independent roles of PB2 and M1 in modulating IFN sensitivity. In mice and ferrets, the HIS virus transiently activated the IFN system, induced robust immune responses, and provided protection against homologous and heterologous influenza challenges.

Fig. 11: Schematic diagram of interferon (IFN)-sensitive viruses.
figure 11

Mutations were introduced into the viral genome to engineer a hyper-interferon-sensitive (HIS) virus with its IFN-modulating functions removed. These IFN-sensitive HIS viruses induce a strong IFN response upon infection of conventional cells, leading to viral attenuation.

Full size image

This approach highlights the potential of IFN-sensitive mutations in developing safe, effective live attenuated vaccines. Furthermore, the HIS virus was later shown to be effective as an in situ vaccination (ISV) in murine models of non-small cell lung cancer (NSCLC)163. The HIS ISV strategy induced a strong type I IFN response, activated T lymphocytes, and overcame resistance to anti–PD–1 therapy. This enhanced survival and promoted systemic tumor-specific immunity in LKB1-deficient NSCLC models. The authors suggested that similar genomic analyses could be applied to other immune components, such as cytokines, natural killer cells, and T cells, to optimize virus-based vaccines by targeting viral immune evasion mechanisms. This offers a promising strategy for future vaccine development and immunotherapy.

Small molecule controlled intein-based influenza virus attenuation

Influenza A virus replication depends on virus-encoded proteins, making post-translational inactivation of viral protein function a promising strategy for viral attenuation. Inteins are protein segments that can excise themselves from a host protein, allowing the ligation of the remaining protein segments164. Small-molecule-dependent inteins have been developed to provide controllable splicing using cell-permeable small molecules. This approach offers a powerful method for the post-translational regulation of protein structure and function165. For example, a 4-hydroxytamoxifen (4-HT)-dependent intein splicing system has been used to develop vesicular stomatitis virus (VSV) with 4-HT-regulated viral replication in vivo166.

Chen et al. applied this 4-HT-regulated intein system to control the replication of influenza A virus and developed attenuated vaccine candidates (Fig. 12)167. They inserted a 4-HT-dependent intein sequence into the PA RNA segment to construct the recombinant virus S218, which exhibited 4-HT-regulated viral replication. In the presence of 4-HT, the intein is spliced, restoring functional PA protein and enabling viral replication at levels comparable to the WT virus. In the absence of 4-HT, the intein remained unspliced, inactivating the PA protein and attenuating the virus. In mouse models, S218 was highly attenuated with no influenza-related clinical symptoms and demonstrated a favorable safety profile. Vaccinated mice showed robust humoral, mucosal, and cellular immune responses and protection against WT influenza virus challenge.

Fig. 12: Schematic representation of the generation of 4-HT-dependent influenza viruses using the intein system.
figure 12

In conventional cells, the virus expresses an unspliced intein in the viral PA protein, resulting in viral attenuation. The addition of 4-hydroxytamoxifen (4-HT) specifically induces the excision of the intein, leading to efficient viral replication.

Full size image

This study underscores the potential of 4-HT-regulated intein-based systems for developing live attenuated vaccines, offering precise attenuation control while ensuring safety and immunogenicity. However, concerns remain about unintended intein activation by estrogen receptor modulators and potential immune or toxic responses to intein byproducts. To mitigate these risks, the authors proposed incorporating an E3-specific degron to eliminate the intein via the ubiquitin-proteasome system.

Proteolysis-targeting (PROTAR) live attenuated vaccine

Viral replication relies on the stability and function of viral proteins, and manipulating their stability can be an effective strategy for controlling viral replication and developing live attenuated vaccines. One promising approach involves the targeted degradation of essential viral proteins through the host cell’s ubiquitin-proteasome system (UPS), which provides a post-translational method for modulating the viral life cycle and achieving viral attenuation168. The UPS controls protein quality by tagging substrate proteins with a polyubiquitin chain via a cascade of enzymatic reactions involving E1, E2, and E3, marking them for proteasomal degradation. Inspired by proteolysis-targeting chimeras (PROTACs)—bifunctional small molecules that induce protein degradation via the UPS—researchers have started to utilize this endogenous machinery for LAIV development169,170,171,172.

In 2022, Si et al. demonstrated the feasibility of this strategy in a proof-of-concept study, showing that incorporating proteasome-targeting degrons (PTDs) into the IAV genome could achieve controlled viral attenuation (Fig. 13)170. A PTD, recognized by the E3 ligase von Hippel–Lindau (VHL), was engineered to the C-terminus of the influenza A viral M1 protein via a tobacco etch virus cleavage site (TEVcs) linker. This design enabled conditional cleavage of the PTD in TEV protease (TEVp) expressing MDCK cell lines, allowing functional M1 protein expression during vaccine production and efficient viral replication. In conventional cells, the PTD is recognized by VHL E3 ligase, triggering proteasomal degradation of the M1 protein and leading to viral attenuation. Mechanistic studies confirmed that M1 protein degradation was dependent on the PTD, VHL, and proteasome. In vivo studies showed that the M1-PTD virus was significantly attenuated and elicited broad, robust immune responses, protecting against both homologous and heterologous IAV strains. This study demonstrates the potential of leveraging host protein degradation machinery for the development of live attenuated vaccines.

Fig. 13: Schematic representation of the design of proteolysis-targeting (PROTAR) attenuated viruses.
figure 13

Proteasome-targeting degrons (PTDs) are incorporated into viral proteins. In conventional cells, the PTDs are recognized by their respective E3 ubiquitin ligases, which polyubiquitinate the viral proteins, targeting them for degradation by the proteasome, leading to proteasomal viral protein degradation.

Full size image

Building on this, the group expanded the concept to develop proteolysis-targeting (PROTAR) LAIVs by incorporating a broader range of PTDs171. The researchers engineered 22 distinct PTDs into the C-terminus of the M1 protein via a TEVcs linker, each recognized by a different E3 ligase. This resulted in a library of PROTAR virus strains that require one of the 22 E3 ligases for proteasome-mediated viral protein degradation and viral attenuation. These PROTAR viruses exhibited efficient replication in TEVp-expressing cells while remaining attenuated to varying extents in conventional cells, providing the opportunity to fine-tune the replication and attenuation of the engineered vaccine candidates. Mutations in the PTDs, knockdown of the corresponding E3 ligases, or inhibition of the proteasome restored M1 protein levels in conventional cells, confirming that PROTAR virus attenuation is dependent on PTDs, E3 ligases, and the UPS. In animal models, PROTAR viruses elicited broad-spectrum immune responses and provided cross-protective immunity against diverse IAV strains. This systematic approach highlights the versatility and adaptability of leveraging the UPS for live attenuated vaccine development.

To expand the incorporation sites of PTDs into internal regions of viral proteins, the researchers developed a next-generation PROTAR vaccine platform, referred to as PROTAR 2.0172. This platform enables flexible incorporation of PTDs at various genomic loci of influenza viruses. The researchers engineered a library of viral proteins by inserting the VHL E3 ligase PTD at 60 potential sites across eight viral proteins, prioritizing surface-accessible and loop regions to facilitate E3 binding while minimizing disruption to viral function and propagation. Four PROTAR 2.0 strains targeting VHL were successfully rescued in VHL-knockout (KO) cells, achieving comparable replication kinetics to WT viruses, while exhibiting significant UPS-dependent attenuation in conventional cells. Using a similar approach, the researchers identified four additional strains incorporating PTDs targeting the SCFβ-TrCP E3 ligase. To further enhance the safety of PROTAR 2.0 vaccines, the team incorporated multiple PTDs into one virus, generating optimized vaccine candidates: PROTARVHL/VHL, PROTARβ-TrCP/β-TrCP, and PROTARVHL/β-TrCP. In animal models, a single intranasal dose of these PROTAR 2.0 candidates elicited broad immune responses and provided complete protection against both homologous and heterologous IAV challenges. This approach also demonstrated applicability in influenza A/PR/8/34 (PR8) virus and influenza B viruses, suggesting the generalizability of the PROTAR 2.0 vaccine strategy across different influenza virus strains.

These studies collectively support the establishment and advancement of the PROTAR vaccine strategy, enhancing its generalizability and translational potential170,171,172,173. The versatility of this approach is highlighted by the existence of over 600 E3 ligases in the human proteome174. However, as the authors noted, because PROTAR vaccines rely on the host UPS for attenuation, it is important to assess their safety in individuals with compromised UPS function or those undergoing proteasome inhibitor therapy. Nonetheless, alternative protein degradation machineries, such as the endosomal, lysosomal, and autophagic pathways, offer promising complementary pathways for targeted protein degradation175. The PROTAR strategy can also be adapted for designing live attenuated vaccines against other viral pathogens, providing opportunities to develop next-generation vaccines with enhanced safety and efficacy.

Concluding remarks

Global epidemics and pandemics, such as those caused by influenza and SARS-CoV-2, have underscored the profound impact of infectious diseases on public health and the global economy. Vaccination remains the most cost-effective strategy for influenza prevention and control. Live attenuated vaccines represent a promising modality, yet current platforms face significant limitations. Frequent mutations in circulating strains lead to reduced efficacy over time and necessitate vaccine reformulation. Traditional LAIV development has largely relied on empirical, time-consuming approaches that often lag behind the rapidly evolving viral landscape or provided limited protection.

Rational live vaccine design, enabled by advances in synthetic biology, offers a transformative solution. Unlike traditional trial-and-error methods, modern genetic tools allow precise, targeted modifications of viral genomes. This enables the development of live vaccines that are both safely attenuated and capable of inducing broad protective immunity, while offering a more predictable and efficient path to vaccine optimization and production. These strategies show strong potential for overcoming the limitations of traditional LAIVs.

While this review focuses primarily on influenza A, several rational attenuation strategies—such as NS1 truncation, single-replication M2-deficient viruses, and the PROTAR strategy—have been successfully extended to influenza B. Additional strategies like HA cleavage site modification may also be applicable, given the conserved nature of this site across influenza A and B viruses176. In contrast, influenza C and D viruses differ significantly in genome structure and surface glycoproteins. They possess only seven RNA segments and lack HA and NA, encoding a hemagglutinin-esterase-fusion (HEF) protein that combines receptor binding, destruction, and membrane fusion functions. Due to limited sequence homology, HA-based attenuation strategies are not applicable to influenza C and D177,178,179. However, the HEF protein has conserved features, such as glycosylation sites and key fusion or esterase residues, that may serve as engineering targets. Broadly applicable methods, including miRNA-mediated attenuation, codon and codon pair deoptimization, and the PROTAR strategy, also show promise for extension to these viral types.

For live vaccine development, safety remains a top consideration. Rational attenuation allows for precise control of viral replication and improved genetic stability, reducing pathogenicity while preserving immunogenicity. Importantly, some strategies minimize the risk of reversion to virulence, which is a key concern with traditional LAIVs, especially in vulnerable populations. For example, the genetic code expansion strategy enhances safety by making viral replication strictly dependent on the presence of unnatural amino acids, effectively preventing uncontrolled replication in the host while maintaining strong immune responses. Similarly, NS1/M2 modification, genome rearrangement, and codon deoptimization all yield low-reversion viral strains.

Concerns around viral reassortment remain relevant, since co-infection with circulating WT viruses could generate novel strains. Although this risk cannot be fully eliminated, several rational strategies incorporate built-in genetic safeguards. These include codon deoptimization, miRNA-based attenuation, genetic code expansion, and the PROTAR platform—many of which target essential polymerase genes, ensuring any reassortants are either replication-defective or remain attenuated.

Rational design also holds promise for improving LAIV efficacy. Traditional LAIVs with cold-adapted donor backbones are susceptibility to pre-existing immunity and antigenic drift. Novel rational design strategies that do not rely on cold-adapted backbones can replicate more effectively in the upper respiratory tract and provide robust protection even in individuals with pre-existing immunity, thereby addressing some of the limitations observed with FluMist. Moreover, direct attenuation of the circulating strains allows for more precise antigenic matching and accelerates development timelines. Expression of conserved or multivalent antigens, and targeting internal viral proteins like NP or polymerase, can improve cross-protection and reduce dependence on antigenic match of HA and NA180. Strategies such as miRNA-based attenuation and codon deoptimization permit fine-tuning at the nucleotide level while preserving protein function, allowing a better balance between attenuation and immunogenicity.

Future advancements in LAIV manufacturing are expected to enable large-scale, high-quality production and more rapid responses to emerging threats. Cell-based vaccine production offers advantages over egg-based systems—including greater standardization, scalability, and suitability for egg-allergic individuals. Technologies such as single-use bioreactors and fixed-bed systems allow for high-density cell growth in closed, sterile environments, greatly improving biosafety and yield181,182. Maintaining vaccine quality remains critical and involves ensuring genetic stability, limiting viral passages, and sequencing seed viruses to monitor antigenic drift183,184. Thermostable formulations, such as lyophilized, spray-dried, or foam-dried versions, can significantly extend shelf life and reduce the cold chain dependence of LAIVs185,186,187. Furthermore, due to variability in immunogenicity across age groups, rational LAIV design may be tailored to specific populations through adjustments in adjuvant composition or dosing46. These strategies support the development of high-yield, cost-effective production systems that can ensure a reliable vaccine supply, especially during pandemics. Together with improved safety and efficacy, these advances may ultimately support expansion of the approved age range of LAIVs (e.g., beyond 2 to 49 years).

Clinical and regulatory challenges also remain. The lack of appropriate animal models and well-defined immune correlates complicates the evaluation of LAIVs. Emerging tools such as human organ-on-a-chip and organoid models offer more physiologically relevant alternatives and may support translational studies and regulatory approval171,188,189. Inclusion of high-risk populations in clinical trials and long-term safety monitoring might also be essential for maximizing public health impact.

In conclusion, rational design enables a more precise, efficient, and adaptable approach to vaccine development, allowing for the creation of vaccines that are not only safer and more broadly protective but also easier to manufacture. As rational vaccine design and attenuation strategies continue to evolve, next-generation LAIVs are expected to play an increasingly central role in the control of influenza. Ongoing research into the molecular mechanisms of attenuation, immunogenicity, and cross-protection will further accelerate their development. Through successful clinical translation, we anticipate that these engineered LAIVs will overcome many of the limitations of traditional approaches and provide rapid and scalable solutions to emerging infectious disease threats.