Molecular determinants of influenza virus pathogenesis in mice

mRNA vaccines encoding computationally optimized hemagglutinin elicit protective antibodies against future antigenically drifted H1N1 and H3N2 influenza viruses isolated between 2018-2020

Background

The implementation of mRNA vaccines against COVID-19 has successfully validated the safety and efficacy of the platform, while at the same time revealing the potential for their applications against other infectious diseases. Traditional seasonal influenza vaccines often induce strain specific antibody responses that offer limited protection against antigenically drifted viruses, leading to reduced vaccine efficacy. Modern advances in viral surveillance and sequencing have led to the development of in-silico methodologies for generating computationally optimized broadly reactive antigens (COBRAs) to improve seasonal influenza vaccines.

Methods

In this study, immunologically naïve mice were intramuscularly vaccinated with mRNA encoding H1 and H3 COBRA hemagglutinins (HA) or wild-type (WT) influenza HAs encapsulated in lipid nanoparticles (LNPs).

Results

Mice vaccinated with H1 and H3 COBRA HA-encoding mRNA vaccines generated robust neutralizing serum antibody responses against more antigenically distinct contemporary and future drifted H1N1 and H3N2 influenza strains than those vaccinated with WT H1 and H3 HA-encoding mRNA vaccines. The H1 and H3 COBRA HA-encoding mRNA vaccines also prevented influenza illness, including severe disease in the mouse model against H1N1 and H3N2 viruses.

Conclusions

This study highlights the potential benefits of combining universal influenza antigen design technology with modern vaccine delivery platforms and exhibits how these vaccines can be advantageous over traditional WT vaccine antigens at eliciting superior protective antibody responses against a broader number of influenza virus isolates.

Differential interferon responses to influenza A and B viruses in primary ferret respiratory epithelial cells

Influenza B viruses (IBV) cocirculate with influenza A viruses (IAV) and cause periodic epidemics of disease, yet antibody and cellular responses following IBV infection are less well understood. Using the ferret model for antisera generation for influenza surveillance purposes, IAV resulted in robust antibody responses following infection, whereas IBV required an additional booster dose, over 85% of the time, to generate equivalent antibody titers. In this study, we utilized primary differentiated ferret nasal epithelial cells (FNECs) which were inoculated with IAV and IBV to study differences in innate immune responses which may result in differences in adaptive immune responses in the host. FNECs were inoculated with IAV (H1N1pdm09 and H3N2 subtypes) or IBV (B/Victoria and B/Yamagata lineages) and assessed for 72 h. Cells were analyzed for gene expression by quantitative real-time PCR, and apical and basolateral supernatants were assessed for virus kinetics and interferon (IFN), respectively. Similar virus kinetics were observed with IAV and IBV in FNECs. A comparison of gene expression and protein secretion profiles demonstrated that IBV-inoculated FNEC expressed delayed type-I/II IFN responses and reduced type-III IFN secretion compared to IAV-inoculated cells. Concurrently, gene expression of Thymic Stromal Lymphopoietin (TSLP), a type-III IFN-induced gene that enhances adaptive immune responses, was significantly downregulated in IBV-inoculated FNECs. Significant differences in other proinflammatory and adaptive genes were suppressed and delayed following IBV inoculation. Following IBV infection, ex vivo cell cultures derived from the ferret upper respiratory tract exhibited reduced and delayed innate responses which may contribute to reduced antibody responses in vivo.IMPORTANCEInfluenza B viruses (IBV) represent nearly one-quarter of all human influenza cases and are responsible for significant clinical and socioeconomic impacts but do not pose the same pandemic risks as influenza A viruses (IAV) and have thus received much less attention. IBV accounts for greater severity and deaths in children, and vaccine efficacy remains low. The ferret can be readily infected with human clinical isolates and demonstrates a similar course of disease and immune responses. IBV, however, generates lower antibodies in ferrets than IAV following the challenge. To determine whether differences in initial innate responses following infection may affect the development of robust adaptive immune responses, ferret respiratory tract cells were isolated, infected with IAV/IBV, and compared. Understanding the differences in the initial innate immune responses to IAV and IBV may be important in the development of more effective vaccines and interventions to generate more robust protective immune responses.

Effects of Extended-Release Buprenorphine on Mouse Models of Influenza

Mice are widely used as small animal models for influenza infection and immunization studies because of their susceptibility to many strains of influenza, obvious clinical signs of infection, and ease of handling. Analgesia is rarely used in such studies even if nonstudy effects such as fight wounds, tail injuries, or severe dermatitis would otherwise justify it because of concerns that treatment might have confounding effects on primary study parameters such as the course of infection and/or the serological response to infection. However, analgesia for study-related or -unrelated effects may be desirable for animal welfare purposes. Opioids, such as extended-release buprenorphine, are well-characterized analgesics in mice and may have fewer immune-modulatory effects than other drug classes. In this study, BALB/c and DBA/2 mice were inoculated with influenza virus, and treatment groups received either no analgesics or 2 doses of extended-release buprenorphine 72 h apart. Clinical signs, mortality, and influenza-specific antibody responses were comparable in mice that did or did not receive buprenorphine. We therefore conclude that extended-release buprenorphine can be used to alleviate incidental pain during studies of influenza infection without altering the course of infection or the immune response.

Innate immune control of influenza virus interspecies adaptation

Influenza virus pandemics are caused by viruses from animal reservoirs that adapt to efficiently infect and replicate in human hosts. Here, we investigated whether Interferon-Induced Transmembrane Protein 3 (IFITM3), a host antiviral factor with known human deficiencies, plays a role in interspecies virus infection and adaptation. We found that IFITM3-deficient mice and human cells could be infected with low doses of avian influenza viruses that failed to infect WT counterparts, identifying a new role for IFITM3 in controlling the minimum infectious viral dose threshold. Remarkably, influenza viruses passaged through Ifitm3-/- mice exhibited enhanced host adaptation, a result that was distinct from passaging in mice deficient for interferon signaling, which caused virus attenuation. Our data demonstrate that IFITM3 deficiency uniquely facilitates zoonotic influenza virus infections and subsequent adaptation, implicating IFITM3 deficiencies in the human population as a vulnerability for emergence of new pandemic viruses.

Effects of life-stage and passive tobacco smoke exposure on pulmonary innate immunity and influenza infection in mice

Limited data are available on the effects of perinatal environmental tobacco smoke (ETS) exposure for early childhood influenza infection. The aim of the present study was to examine whether perinatal versus adult ETS exposure might provoke more severe systemic and pulmonary innate immune responses in mice inoculated with influenza A/Puerto Rico/8/34 virus (IAV) compared to phosphate-buffered saline (PBS). BALB/c mice were exposed to filtered air (FA) or ETS for 6 weeks during the perinatal or adult period of life. Immediately following the final exposure, mice were intranasally inoculated with IAV or PBS. Significant inflammatory effects were observed in bronchoalveolar lavage fluid of neonates inoculated with IAV (FA+IAV or ETS+IAV) compared to PBS (ETS+PBS or FA+PBS), and in the lung parenchyma of neonates administered ETS+IAV versus FA+IAV. Type I and III interferons were also elevated in the spleens of neonates, but not adults with ETS+IAV versus FA+IAV exposure. Both IAV-inoculated neonate groups exhibited significantly more CD4 T cells and increasing numbers of CD8 and CD25 T cells in lungs relative to their adult counterparts. Taken together, these results suggest perinatal ETS exposure induces an exaggerated innate immune response, which may overwhelm protective anti-inflammatory defenses against IAV, and enhances severity of infection at early life stages (e.g., in infants and young children).

Key amino acids of M1-41 and M2-27 determine growth and pathogenicity of chimeric H17 bat influenza virus in cells and in mice

Based on our previous studies, we show that M gene is critical for viral replication and pathogenicity of the chimeric H17 bat influenza virus (Bat09:mH1mN1) by replacing bat M gene with those from human and swine influenza A viruses. However, the key amino acids of M1 and/or M2 proteins responsible for virus replication and pathogenicity remain unknown. In this study, the Eurasian avian-like M gene from the A/California/04/2009 pandemic H1N1 virus significantly decreased viral replication in both mammalian and avian cells in the background of chimeric H17 bat influenza virus by replacing the PR8 M gene. Further studies revealed that the M1 was more crucial for viral growth and pathogenicity in contrast to the M2, and amino acid residues of M1-41V and M2-27A were responsible for these characteristics in cells and in mice. These key residues of M1 and M2 proteins identified in this study might be important for influenza virus surveillance and used to produce live attenuated vaccines in the future. Importance The M1 and M2 proteins influence the morphology, replication, virulence and transmissibility of influenza viruses. Although a few key residues in M1/M2 proteins have been identified, whether other residues of M1/M2 proteins involved in viral replication and pathogenicity need to be discovered. In the background of chimeric H17 bat influenza virus, the Eurasian avian-like M gene from A/California/04/2009 significantly decreased viral growth in mammalian and avian cells. Further study showed that M1 was implicated more than M2 for viral growth and pathogenicity in vitro and in vivo, and the key amino acid residues of M1-41V and M2-27A were responsible for these characteristics in cells and in mice. These key residues of M1 and M2 proteins could be used for influenza virus surveillance and live attenuated vaccine application in the future. These findings provide important information for knowledge on the genetic basis of virulence of influenza viruses.

VirPreNet: a weighted ensemble convolutional neural network for the virulence prediction of influenza A virus using all eight segments

MOTIVATION

Influenza viruses are persistently threatening public health, causing annual epidemics and sporadic pandemics. The evolution of influenza viruses remains to be the main obstacle in the effectiveness of antiviral treatments due to rapid mutations. Previous work has been investigated to reveal the determinants of virulence of the influenza A virus. To further facilitate flu surveillance, explicit detection of influenza virulence is crucial to protect public health from potential future pandemics.

RESULTS

In this article, we propose a weighted ensemble convolutional neural network (CNN) for the virulence prediction of influenza A viruses named VirPreNet that uses all eight segments. Firstly, mouse lethal dose 50 is exerted to label the virulence of infections into two classes, namely avirulent and virulent. A numerical representation of amino acids named ProtVec is applied to the eight-segments in a distributed manner to encode the biological sequences. After splittings and embeddings of influenza strains, the ensemble CNN is constructed as the base model on the influenza dataset of each segment, which serves as the VirPreNet's main part. Followed by a linear layer, the initial predictive outcomes are integrated and assigned with different weights for the final prediction. The experimental results on the collected influenza dataset indicate that VirPreNet achieves state-of-the-art performance combining ProtVec with our proposed architecture. It outperforms baseline methods on the independent testing data. Moreover, our proposed model reveals the importance of PB2 and HA segments on the virulence prediction. We believe that our model may provide new insights into the investigation of influenza virulence.

AVAILABILITY AND IMPLEMENTATION

Codes and data to generate the VirPreNet are publicly available at https://github.com/Rayin-saber/VirPreNet.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Hemagglutinin Stability and Its Impact on Influenza A Virus Infectivity, Pathogenicity, and Transmissibility in Avians, Mice, Swine, Seals, Ferrets, and Humans

Genetically diverse influenza A viruses (IAVs) circulate in wild aquatic birds. From this reservoir, IAVs sporadically cause outbreaks, epidemics, and pandemics in wild and domestic avians, wild land and sea mammals, horses, canines, felines, swine, humans, and other species. One molecular trait shown to modulate IAV host range is the stability of the hemagglutinin (HA) surface glycoprotein. The HA protein is the major antigen and during virus entry, this trimeric envelope glycoprotein binds sialic acid-containing receptors before being triggered by endosomal low pH to undergo irreversible structural changes that cause membrane fusion. The HA proteins from different IAV isolates can vary in the pH at which HA protein structural changes are triggered, the protein causes membrane fusion, or outside the cell the virion becomes inactivated. HA activation pH values generally range from pH 4.8 to 6.2. Human-adapted HA proteins tend to have relatively stable HA proteins activated at pH 5.5 or below. Here, studies are reviewed that report HA stability values and investigate the biological impact of variations in HA stability on replication, pathogenicity, and transmissibility in experimental animal models. Overall, a stabilized HA protein appears to be necessary for human pandemic potential and should be considered when assessing human pandemic risk.

Identification and Characterization of Novel Antibody Epitopes on the N2 Neuraminidase

The influenza virus neuraminidase (NA) is becoming a focus for novel vaccine designs. However, the epitopes of human anti-NA antibodies have been poorly defined. Using a panel of 10 anti-N2 monoclonal antibodies (MAbs) that bind the H3N2 virus A/Switzerland/9715293/2013, we generated five escape mutant viruses. These viruses contained mutations K199E/T, E258K, A272D, and S331N. We found that mutations at K199 and E258 had the largest impact on MAb binding, NA inhibition and neutralization activity. In addition, a natural isolate from the 2017-2018 season was found to contain the E258K mutation and was resistant to numerous antibodies tested. The mutation S331N, was identified in virus passaged in the presence of antibody; however, it had little impact on MAb activity and greatly decreased viral fitness. This information aids in identifying novel human MAb epitopes on the N2 and helps with the detection of antigenically drifted NAs.

IMPORTANCE The influenza virus neuraminidase is an emerging target for universal influenza virus vaccines. However, in contrast to influenza virus hemagglutinin, we know little about antibody epitopes and antigenic sites on the neuraminidase. Characterizing and defining these sites is aiding vaccine development and helping to understand antigenic drift of NA.

Genetic influences on viral-induced cytokine responses in the lung

Infection with respiratory viruses such as influenza, respiratory syncytial virus and coronavirus provides a difficult immunological challenge for the host, where a balance must be established between controlling viral replication and limiting damage to the delicate lung structure. Although the genetic architecture of host responses to respiratory viral infections is not yet understood, it is clear there is underlying heritability that influences pathogenesis. Immune control of virus replication is essential in respiratory infections, but overt activation can enhance inflammation and disease severity. Cytokines initiate antiviral immune responses but are implicated in viral pathogenesis. Here, we discuss how host genetic variation may influence cytokine responses to respiratory viral infections and, based on our current understanding of the role that cytokines play in viral pathogenesis, how this may influence disease severity. We also discuss how induced pluripotent stem cells may be utilised to probe the mechanistic implications of allelic variation in genes in virus-induced inflammatory responses. Ultimately, this could help to design better immune modulators, stratify high risk patients and tailor anti-inflammatory treatments, potentially expanding the ability to treat respiratory virus outbreaks in the future.

Development of Human Vectored Brucellosis Vaccine Formulation: Assessment of Safety and Protectiveness of Influenza Viral Vectors Expressing Brucella Immunodominant Proteins in Mice and Guinea Pigs

In this paper, we first used recombinant influenza viral vector (rIVV) subtype H5N1 expressing from the open reading frame of NS1 80 and NS1 124 amino acids of Brucella outer membrane proteins (Omp) 16 and 19, ribosomal L7/L12, and Cu-Zn superoxide dismutase (SOD) proteins to develop a human brucellosis vaccine. We made 18 combinations of IVVs in mono-, bi-, and tetravalent vaccine formulations and tested them on mice to select the safest and most effective vaccine samples. Then, the most effective vaccine candidates were further tested on guinea pigs. Safety of the rIVV-based vaccine candidate was evaluated by a mouse weight-gain test. Mice and guinea pigs were challenged with the virulent strain B. melitensis 16M. The protective effect of the rIVV-based vaccine candidate was assessed by quantitation of Brucella colonization in tissues and organs of challenged animals. All vaccine formulations were safe in mice. Tested vaccine formulations, as well as the commercial B. melitensis Rev.1 vaccine, have been found to protect mice from B. melitensis 16M infection within the range of 1.6 to 2.97 log₁₀ units (P < 0.05). Tetravalent vaccine formulations from the position of NS1 80 amino acids (0.2 ± 0.4), as well as the commercial B. melitensis Rev.1 vaccine (1.2 ± 2.6), have been found to protect guinea pigs from B. melitensis 16M infection at a significant level (P < 0.05). Thus, tetravalent vaccine formulation Flu-NS1-80-Omp16+Flu-NS1-80-L7/L12+Flu-NS1-80-Omp19+Flu-NS1-80-SOD was chosen as a potential vaccine candidate for further development of an effective human vaccine against brucellosis. These results show a promising future for the development of a safe human vaccine against brucellosis based on rIVVs.

ANP32B Deficiency Protects Mice From Lethal Influenza A Virus Challenge by Dampening the Host Immune Response

Deciphering complex virus-host interactions is crucial for pandemic preparedness. In this study, we assessed the impact of recently postulated cellular factors ANP32A and ANP32B of influenza A virus (IAV) species specificity on viral pathogenesis in a genetically modified mouse model. Infection of ANP32A^−/− and ANP32A^+/+ mice with a seasonal H3N2 IAV or a highly pathogenic H5N1 human isolate did not result in any significant differences in virus tropism, innate immune response or disease outcome. However, infection of ANP32B^−/− mice with H3N2 or H5N1 IAV revealed significantly reduced virus loads, inflammatory cytokine response and reduced pathogenicity compared to ANP32B^+/+ mice. Genome-wide transcriptome analyses in ANP32B^+/+ and ANP32B^−/− mice further uncovered novel immune-regulatory pathways that correlate with reduced pathogenicity in the absence of ANP32B. These data show that ANP32B but not ANP32A promotes IAV pathogenesis in mice. Moreover, ANP32B might possess a yet unknown immune-modulatory function during IAV infection. Targeting ANP32B or its regulated pathways might therefore pose a new strategy to combat severe influenza.

Murine Model for the Study of Influenza D Virus

A novel genus within the Orthomyxoviridae family was identified in the United States and named influenza D virus (IDV). Bovines have been proposed to be the primary host, and three main viral lineages (D/OK-like, D/660-like, and D/Japan-like) have been described. Experimental infections had previously been performed in swine, ferrets, calves, and guinea pigs in order to study IDV pathogenesis. We developed a murine experimental model to facilitate the study of IDV pathogenesis and the immune response. DBA/2 mice were inoculated with 10⁵ 50% tissue culture infective dose (TCID₅₀) of D/bovine/France/5920/2014 (D/OK-like). No clinical signs or weight loss were observed. Viral replication was observed mainly in the upper respiratory tract (nasal turbinates) but also in the lower respiratory tract of infected mice, with a peak at 4 days postinfection. Moreover, the virus was also detected in the intestines. All infected mice seroconverted by 14 days postinfection. Transcriptomic analyses demonstrated that IDV induced the activation of proinflammatory genes, such as gamma interferon (IFN-γ) and CCL2. Inoculation of NF-κB-luciferase and Ifnar1^-/- mice demonstrated that IDV induced mild inflammation and that a type I interferon response was not necessary in IDV clearance. Adaptation of IDV by serial passages in mice was not sufficient to induce disease or increased pathogenesis. Taken together, present data and comparisons with the calf model show that our mouse model allows for the study of IDV replication and fitness (before selected viruses may be inoculated on calves) and also of the immune response.IMPORTANCE Influenza D virus (IDV), a new genus of Orthomyxoviridae family, presents a large host range and a worldwide circulation. The pathogenicity of this virus has been studied in the calf model. The mouse model is frequently used to enable a first assessment of a pathogen's fitness, replication, and pathogenesis for influenza A and B viruses. We showed that DBA/2 mice are a relevant in vivo model for the study of IDV replication. This model will allow for rapid IDV fitness and replication evaluation and will enable phenotypic comparisons between isolated viruses. It will also allow for a better understanding of the immune response induced after IDV infection.

The effect of mutations derived from mouse-adapted H3N2 seasonal influenza A virus to pathogenicity and host adaptation

Elucidating the genetic basis of influenza A viruses (IAVs) is important to understand which mutations will determine the virulence and the host range of mammals. Here, seasonal H3N2 influenza was adapted in mice by serial passage and four mutants, each carrying amino acid substitutions related to mouse adaptation in either the PB2, HA, NP, or NA protein, were generated. To confirm the contribution of each gene to enhanced pathogenicity and mouse adaptation, mice were inoculated with the respective variants, and virulence, replication, histopathology, and infectivity were examined. The virus harboring HA mutations displayed increased infection efficiency and replication competence, resulting in higher mortality in mice relative to those infected with wild-type virus. By contrast, the NP D34N mutation caused rapid and widespread infection in multiple organs without presenting virulent symptoms. Additionally, the PB2 F323L mutation presented delayed but elevated replication competence in the respiratory tract, whereas the S331R mutation in NA showed no considerable effects on mouse adaptation. These results suggested that mouse-adapted changes in HA are major factors in increased pathogenicity and that mutations in NP and PB2 also contribute to cross-species adaptability. Our findings offer a better understanding of the molecular basis for IAV pathogenicity and adaptation in a new host.

Rule-based meta-analysis reveals the major role of PB2 in influencing influenza A virus virulence in mice

Background

Influenza A virus (IAV) poses threats to human health and life. Many individual studies have been carried out in mice to uncover the viral factors responsible for the virulence of IAV infections. Nonetheless, a single study may not provide enough confident about virulence factors, hence combining several studies for a meta-analysis is desired to provide better views. For this, we documented more than 500 records of IAV infections in mice, whose viral proteins could be retrieved and the mouse lethal dose 50 or alternatively, weight loss and/or survival data, was/were available for virulence classification.

Results

IAV virulence models were learned from various datasets containing aligned IAV proteins and the corresponding two virulence classes (avirulent and virulent) or three virulence classes (low, intermediate and high virulence). Three proven rule-based learning approaches, i.e., OneR, JRip and PART, and additionally random forest were used for modelling. PART models achieved the best performance, with moderate average model accuracies ranged from 65.0 to 84.4% and from 54.0 to 66.6% for the two-class and three-class problems, respectively. PART models were comparable to or even better than random forest models and should be preferred based on the Occam’s razor principle. Interestingly, the average accuracy of the models was improved when host information was taken into account. For model interpretation, we observed that although many sites in HA were highly correlated with virulence, PART models based on sites in PB2 could compete against and were often better than PART models based on sites in HA. Moreover, PART had a high preference to include sites in PB2 when models were learned from datasets containing the concatenated alignments of all IAV proteins. Several sites with a known contribution to virulence were found as the top protein sites, and site pairs that may synergistically influence virulence were also uncovered.

Conclusion

Modelling IAV virulence is a challenging problem. Rule-based models generated using viral proteins are useful for its advantage in interpretation, but only achieve moderate performance. Development of more advanced approaches that learn models from features extracted from both viral and host proteins shall be considered for future works.

Influenza Viruses in Mice: Deep Sequencing Analysis of Serial Passage and Effects of Sialic Acid Structural Variation

Influenza A viruses have regularly jumped to new host species to cause epidemics or pandemics, an evolutionary process that involves variation in the viral traits necessary to overcome host barriers and facilitate transmission. Mice are not a natural host for influenza virus but are frequently used as models in studies of pathogenesis, often after multiple passages to achieve higher viral titers that result in clinical disease such as weight loss or death. Here, we examine the processes of influenza A virus infection and evolution in mice by comparing single nucleotide variations of a human H1N1 pandemic virus, a seasonal H3N2 virus, and an H3N2 canine influenza virus during experimental passage. We also compared replication and sequence variation in wild-type mice expressing N-glycolylneuraminic acid (Neu5Gc) with those seen in mice expressing only N-acetylneuraminic acid (Neu5Ac). Viruses derived from plasmids were propagated in MDCK cells and then passaged in mice up to four times. Full-genome deep sequencing of the plasmids, cultured viruses, and viruses from mice at various passages revealed only small numbers of mutational changes. The H3N2 canine influenza virus showed increases in frequency of sporadic mutations in the PB2, PA, and NA segments. The H1N1 pandemic virus grew well in mice, and while it exhibited the maintenance of some minority mutations, there was no clear evidence for adaptive evolution. The H3N2 seasonal virus did not establish in the mice. Finally, there were no clear sequence differences associated with the presence or absence of Neu5Gc.IMPORTANCE Mice are commonly used as a model to study the growth and virulence of influenza A viruses in mammals but are not a natural host and have distinct sialic acid receptor profiles compared to humans. Using experimental infections with different subtypes of influenza A virus derived from different hosts, we found that evolution of influenza A virus in mice did not necessarily proceed through the linear accumulation of host-adaptive mutations, that there was variation in the patterns of mutations detected in each repetition, and that the mutation dynamics depended on the virus examined. In addition, variation in the viral receptor, sialic acid, did not affect influenza virus evolution in this model. Overall, our results show that while mice provide a useful animal model for influenza virus pathology, host passage evolution will vary depending on the specific virus tested.

Identifying genetic determinants of complex phenotypes from whole genome sequence data

BACKGROUND

A critical goal in biology is to relate the phenotype to the genotype, that is, to find the genetic determinants of various traits. However, while simple monofactorial determinants are relatively easy to identify, the underpinnings of complex phenotypes are harder to predict. While traditional approaches rely on genome-wide association studies based on Single Nucleotide Polymorphism data, the ability of machine learning algorithms to find these determinants in whole proteome data is still not well known.

RESULTS

To better understand the applicability of machine learning in this case, we implemented two such algorithms, adaptive boosting (AB) and repeated random forest (RRF), and developed a chunking layer that facilitates the analysis of whole proteome data. We first assessed the performance of these algorithms and tuned them on an influenza data set, for which the determinants of three complex phenotypes (infectivity, transmissibility, and pathogenicity) are known based on experimental evidence. This allowed us to show that chunking improves runtimes by an order of magnitude. Based on simulations, we showed that chunking also increases sensitivity of the predictions, reaching 100% with as few as 20 sequences in a small proteome as in the influenza case (5k sites), but may require at least 30 sequences to reach 90% on larger alignments (500k sites). While RRF has less specificity than random forest, it was never <50%, and RRF sensitivity was significantly higher at smaller chunk sizes. We then used these algorithms to predict the determinants of three types of drug resistance (to Ciprofloxacin, Ceftazidime, and Gentamicin) in a bacterium, Pseudomonas aeruginosa. While both algorithms performed well in the case of the influenza data, results were more nuanced in the bacterial case, with RRF making more sensible predictions, with smaller errors rates, than AB.

CONCLUSIONS

Altogether, we demonstrated that ML algorithms can be used to identify genetic determinants in small proteomes (viruses), even when trained on small numbers of individuals. We further showed that our RRF algorithm may deserve more scrutiny, which should be facilitated by the decreasing costs of both sequencing and phenotyping of large cohorts of individuals.

Pathogenesis and Transmission of Genetically Diverse Swine-Origin H3N2 Variant Influenza A Viruses from Multiple Lineages Isolated in the United States, 2011-2016

While several swine-origin influenza A H3N2 variant (H3N2v) viruses isolated from humans prior to 2011 have been previously characterized for their virulence and transmissibility in ferrets, the recent genetic and antigenic divergence of H3N2v viruses warrants an updated assessment of their pandemic potential. Here, four contemporary H3N2v viruses isolated during 2011 to 2016 were evaluated for their replicative ability in both in vitro and in vivo in mammalian models as well as their transmissibility among ferrets. We found that all four H3N2v viruses possessed similar or enhanced replication capacities in a human bronchial epithelium cell line (Calu-3) compared to a human seasonal influenza virus, suggestive of strong fitness in human respiratory tract cells. The majority of H3N2v viruses examined in our study were mildly virulent in mice and capable of replicating in mouse lungs with different degrees of efficiency. In ferrets, all four H3N2v viruses caused moderate morbidity and exhibited comparable titers in the upper respiratory tract, but only 2 of the 4 viruses replicated in the lower respiratory tract in this model. Furthermore, despite efficient transmission among cohoused ferrets, recently isolated H3N2v viruses displayed considerable variance in their ability to transmit by respiratory droplets. The lack of a full understanding of the molecular correlates of virulence and transmission underscores the need for close genotypic and phenotypic monitoring of H3N2v viruses and the importance of continued surveillance to improve pandemic preparedness.IMPORTANCE Swine-origin influenza viruses of the H3N2 subtype, with the hemagglutinin (HA) and neuraminidase (NA) derived from historic human seasonal influenza viruses, continue to cross species barriers and cause human infections, posing an indelible threat to public health. To help us better understand the potential risk associated with swine-origin H3N2v viruses that emerged in the United States during the 2011-2016 influenza seasons, we use both in vitro and in vivo models to characterize the abilities of these viruses to replicate, cause disease, and transmit in mammalian hosts. The efficient respiratory droplet transmission exhibited by some of the H3N2v viruses in the ferret model combined with the existing evidence of low immunity against such viruses in young children and older adults highlight their pandemic potential. Extensive surveillance and risk assessment of H3N2v viruses should continue to be an essential component of our pandemic preparedness strategy.

Genetic characterization and pathogenic potential of H10 avian influenza viruses isolated from live poultry markets in Bangladesh

Fatal human cases of avian-origin H10N8 influenza virus infections have raised concern about their potential for human-to-human transmission. H10 subtype avian influenza viruses (AIVs) have been isolated from wild and domestic aquatic birds across Eurasia and North America. We isolated eight H10 AIVs (four H10N7, two H10N9, one H10N1, and one H10N6) from live poultry markets in Bangladesh. Genetic analyses demonstrated that all eight isolates belong to the Eurasian lineage. HA phylogenetic and antigenic analyses indicated that two antigenically distinct groups of H10 AIVs are circulating in Bangladeshi live poultry markets. We evaluated the virulence of four representative H10 AIV strains in DBA/2J mice and found that they replicated efficiently in mice without prior adaptation. Moreover, H10N6 and H10N1 AIVs caused high mortality with systemic dissemination. These results indicate that H10 AIVs pose a potential threat to human health and the mechanisms of their transmissibility should be elucidated.

Development and application of bioluminescence imaging for the influenza A virus

Influenza A viruses (IAVs) cause seasonal epidemics and intermittent pandemics which threaten human health. Conventional assays cannot meet the demands for rapid and sensitive detection of viral spread and pathogenesis in real time cannot be used for high-throughput screens of novel antivirals. Bioluminescence imaging (BLI) has emerged as a powerful tool in the study of infectious diseases in animal models. The advent of influenza reverse genetics has enabled the incorporation of bioluminescent reporter proteins into replication-competent IAVs. This review briefly describes the current development and applications of bioluminescence in the study of viral infections and antiviral therapeutics for IAVs. BLI is expected to substantially accelerate the basic and applied research of IAV both in vitro and in vivo.

A Single Amino Acid in the Polymerase Acidic Protein Determines the Pathogenicity of Influenza B Viruses

Influenza B virus (IBV) is one of the human respiratory viruses and one of the targets of seasonal vaccination. However, the bifurcation of two antigenically distinct lineages of IBVs makes it difficult to arrange proper medical countermeasures. Moreover, compared with pathogenicity-related molecular markers known for influenza A virus, little has been known for IBVs. To understand pathogenicity caused by IBVs, we investigated the molecular determinants of IBV pathogenicity in animal models. After serial lung-to-lung passages of Victoria lineage B/Brisbane/60/2008 (Vc_BR60) and Yamagata lineage B/Wisconsin/01/2010 (Ym_WI01) viruses in BALB/c mice, we identified the mouse-adapted Vc_BR60 (maVc_BR60) and Ym_WI01 (maYm_WI01) viruses, respectively. To find a molecular clue(s) to the increased pathogenicity of maVc_BR60 and maYm_WI01, we determined their genetic sequences. Several amino acid mutations were identified in the PB2, PB1, PA, BM2, and/or NS1 protein-coding regions, and one concurrent lysine (K)-to-arginine (R) mutation in PA residue 338 (PA K338R) was found in both maVc_BR60 and maYm_WI01 viruses. When analyzed using viruses rescued through reverse genetics, it was shown that PA K338R alone could increase the pathogenicity of both IBVs in mice and viral replication in the respiratory tracts of ferrets. In a subsequent minireplicon assay, the effect of PA K338R was highlighted by the enhancement of viral polymerase complex activity of both Vc_BR60 and Ym_WI01 viruses. These results suggest that the PA K338R mutation may be a molecular determinant of IBV pathogenicity via modulating the viral polymerase function of IBVs.IMPORTANCE To investigate molecular pathogenic determinants of IBVs, which are one of the targets of seasonal influenza vaccines, we adapted both Victoria and Yamagata lineage IBVs independently in mice. The recovered mouse-adapted viruses exhibited increased virulence, and of the various mutations identified from both mouse-adapted viruses, a concurrent amino acid mutation was found in the PA protein-coding region. When analyzed using viruses rescued through reverse genetics, the PA mutation alone appeared to contribute to viral pathogenicity in mice within the compatible genetic constellation between the IBV lineages and to the replication of IBVs in ferrets. Regarding the potential mechanism of increased viral pathogenicity, it was shown that the PA mutation could upregulate the viral polymerase complex activity of both IBV lineages. These results indicate that the PA mutation could be a newly defined molecular pathogenic determinant of IBVs that substantiates our understanding of the viral pathogenicity and public health risks of IBVs.

Prion protein protects mice from lethal infection with influenza A viruses

The cellular prion protein, designated PrP^C, is a membrane glycoprotein expressed abundantly in brains and to a lesser extent in other tissues. Conformational conversion of PrP^C into the amyloidogenic isoform is a key pathogenic event in prion diseases. However, the physiological functions of PrP^C remain largely unknown, particularly in non-neuronal tissues. Here, we show that PrP^C is expressed in lung epithelial cells, including alveolar type 1 and 2 cells and bronchiolar Clara cells. Compared with wild-type (WT) mice, PrP^C-null mice (Prnp^0/0) were highly susceptible to influenza A viruses (IAVs), with higher mortality. Infected Prnp^0/0 lungs were severely injured, with higher inflammation and higher apoptosis of epithelial cells, and contained higher reactive oxygen species (ROS) than control WT lungs. Treatment with a ROS scavenger or an inhibitor of xanthine oxidase (XO), a major ROS-generating enzyme in IAV-infected lungs, rescued Prnp^0/0 mice from the lethal infection with IAV. Moreover, Prnp^0/0 mice transgenic for PrP with a deletion of the Cu-binding octapeptide repeat (OR) region, Tg(PrPΔOR)/Prnp^0/0 mice, were also highly susceptible to IAV infection. These results indicate that PrP^C has a protective role against lethal infection with IAVs through the Cu-binding OR region by reducing ROS in infected lungs. Cu content and the activity of anti-oxidant enzyme Cu/Zn-dependent superoxide dismutase, SOD1, were lower in Prnp^0/0 and Tg(PrPΔOR)/Prnp^0/0 lungs than in WT lungs. It is thus conceivable that PrP^C functions to maintain Cu content and regulate SOD1 through the OR region in lungs, thereby reducing ROS in IAV-infected lungs and eventually protecting them from lethal infection with IAVs. Our current results highlight the role of PrP^C in protection against IAV infection, and suggest that PrP^C might be a novel target molecule for anti-influenza therapeutics.

Influenza A virus (IAV) is an enveloped, negative sense, single-stranded RNA virus, causing seasonal epidemic outbreaks of influenza. Anti-influenza agents targeting viral molecules, such as neuraminidase inhibitors, are currently available. However, these agents have accelerated emergence of mutant IAVs that are resistant to these agents among human populations. Development of new types of anti-influenza agents is awaited. We show that the cellular prion protein PrP^C has a protective role against lethal infection with IAVs through the octapeptide repeat (OR) region by abrogating lung epithelial cell apoptosis induced by reactive oxygen species (ROS) in infected lungs. We also show that PrP^C could reduce ROS in IAV-infected lungs through the OR region by maintaining Cu ion homeostasis and thereby activating Cu/Zn-dependent superoxide dismutase, SOD1. These results highlight the protective role of PrP^C in IAV infection. Elucidation of the exact mechanism underlying the PrP^C-mediated protection against IAV infection would be important for further understanding the pathogenesis of IAV infection and could be useful for development of new types of anti-influenza therapeutics.

A Hitchhiker's guide to humanized mice: new pathways to studying viral infections

Humanized mice are increasingly appreciated as an incredibly powerful platform for infectious disease research. The often very narrow species tropism of many viral infections, coupled with the sometimes misleading results from preclinical studies in animal models further emphasize the need for more predictive model systems based on human cells rather than surrogates. Humanized mice represent such a model and have been greatly enhanced with regards to their immune system reconstitution as well as immune functionality in the past years, resulting in their recommendation as a preclinical model by the US Food and Drug Administration. This review aims to give a detailed summary of the generation of human peripheral blood lymphocyte-, CD34⁺ haematopoietic stem cell- and bone marrow/liver/thymus-reconstituted mice and available improved models (e.g. myeloid- or T-cell-only mice, MISTRG, NSG-SGM3). Additionally, we summarize human-tropic viral infections, for which humanized mice offer a novel approach for the study of disease pathogenesis as well as future perspectives for their use in biomedical, drug and vaccine research.

PPARgamma Deficiency Counteracts Thymic Senescence

Thymic senescence contributes to increased incidence of infection, cancer and autoimmunity at senior ages. This process manifests as adipose involution. As with other adipose tissues, thymic adipose involution is also controlled by PPARgamma. This is supported by observations reporting that systemic PPARgamma activation accelerates thymic adipose involution. Therefore, we hypothesized that decreased PPARgamma activity could prevent thymic adipose involution, although it may trigger metabolic adverse effects. We have confirmed that both human and murine thymic sections show marked staining for PPARgamma at senior ages. We have also tested the thymic lobes of PPARgamma haplo-insufficient and null mice. Supporting our working hypothesis both adult PPARgamma haplo-insufficient and null mice show delayed thymic senescence by thymus histology, thymocyte mouse T-cell recombination excision circle qPCR and peripheral blood naive T-cell ratio by flow-cytometry. Delayed senescence showed dose-response with respect to PPARgamma deficiency. Functional immune parameters were also evaluated at senior ages in PPARgamma haplo-insufficient mice (null mice do not reach senior ages due to metabolic adverse affects). As expected, sustained and elevated T-cell production conferred oral tolerance and enhanced vaccination efficiency in senior PPARgamma haplo-insufficient, but not in senior wild-type littermates according to ELISA IgG measurements. Of note, humans also show increased oral intolerance issues and decreased protection by vaccines at senior ages. Moreover, PPARgamma haplo-insufficiency also exists in human known as a rare disease (FPLD3) causing metabolic adverse effects, similar to the mouse. When compared to age- and metabolic disorder-matched other patient samples (FPLD2 not affecting PPARgamma activity), FPLD3 patients showed increased human Trec (hTrec) values by qPCR (within healthy human range) suggesting delayed thymic senescence, in accordance with mouse results and supporting our working hypothesis. In summary, our experiments prove that systemic decrease of PPARgamma activity prevents thymic senescence, albeit with metabolic drawbacks. However, thymic tissue-specific PPARgamma antagonism would likely solve the issue.

Adaptive mutations of neuraminidase stalk truncation and deglycosylation confer enhanced pathogenicity of influenza A viruses

It has been noticed that neuraminidase (NA) stalk truncation has arisen from evolutionary adaptation of avian influenza A viruses (IAVs) from wild aquatic birds to domestic poultry. We identified this molecular alteration after the adaptation of a 2009 pandemic H1N1 virus (pH1N1) in BALB/c mice. The mouse-adapted pH1N1 lost its eight consecutive amino acids including one potential N-linked glycosite from the NA stalk region. To explore the relationship of NA stalk truncation or deglycosylation with viral pathogenicity changes, we generated NA stalk mutant viruses on the pH1N1 backbone by reverse genetics. Intriguingly, either NA stalk truncation or deglycosylation changed pH1N1 into a lethal virus to mice by resulting in extensive pathologic transformation in the mouse lungs and systemic infection affecting beyond the respiratory organs in mice. The increased pathogenicity of these NA stalk mutants was also reproduced in ferrets. In further investigation using a human-infecting H7N9 avian IAV strain, NA stalk truncation or deglycosylation enhanced the replication property and pathogenicity of H7N9 NA stalk mutant viruses in the same mouse model. Taken together, our results suggest that NA stalk truncation or deglycosylation can be the pathogenic determinants of seasonal influenza viruses associated with the evolutionary adaptation of IAVs.

Pandemic 2009 H1N1 Influenza Venus reporter virus reveals broad diversity of MHC class II-positive antigen-bearing cells following infection in vivo

Although it is well established that Influenza A virus infection is initiated in the respiratory tract, the sequence of events and the cell types that become infected or access viral antigens remains incompletely understood. In this report, we used a novel Influenza A/California/04/09 (H1N1) reporter virus that stably expresses the Venus fluorescent protein to identify antigen-bearing cells over time in a mouse model of infection using flow cytometry. These studies revealed that many hematopoietic cells, including subsets of monocytes, macrophages, dendritic cells, neutrophils and eosinophils acquire influenza antigen in the lungs early post-infection. Surface staining of the viral HA revealed that most cell populations become infected, most prominently CD45^neg cells, alveolar macrophages and neutrophils. Finally, differences in infection status, cell lineage and MHC class II expression by antigen-bearing cells correlated with differences in their ability to re-stimulate influenza-specific CD4 T cells ex vivo. Collectively, these studies have revealed the cellular heterogeneity and complexity of antigen-bearing cells within the lung and their potential as targets of antigen recognition by CD4 T cells.

The significance of avian influenza virus mouse-adaptation and its application in characterizing the efficacy of new vaccines and therapeutic agents

Due to the increased frequency of interspecies transmission of avian influenza viruses, studies designed to identify the molecular determinants that could lead to an expansion of the host range have been increased. A variety of mouse-based mammalian-adaptation studies of avian influenza viruses have provided insight into the genetic alterations of various avian influenza subtypes that may contribute to the generation of a pandemic virus. To date, the studies have focused on avian influenza subtypes H5, H6, H7, H9, and H10 which have recently caused human infection. Although mice cannot fully reflect the course of human infection with avian influenza, these mouse studies can be a useful method for investigating potential mammalian adaptive markers against newly emerging avian influenza viruses. In addition, due to the lack of appropriate vaccines against the diverse emerging influenza viruses, the generation of mouse-adapted lethal variants could contribute to the development of effective vaccines or therapeutic agents. Within this review, we will summarize studies that have demonstrated adaptations of avian influenza viruses that result in an altered pathogenicity in mice which may suggest the potential application of mouse-lethal strains in the development of influenza vaccines and/or therapeutics in preclinical studies.

The matrix gene of influenza A H5N1 in Egypt, 2006-2016: molecular insights and distribution of amantadine-resistant variants

Large-scale sequence analysis of Matrix (M) gene and its coding proteins M1 and M2 was performed for 274 highly pathogenic avian influenza viruses H5N1 circulated in Egypt from 2006 to 2016. The aim is to study the amantadine-resistant markers distribution and to estimate the evolutionary rate. 246 viruses were obtained from the Global Initiative on Sharing All Influenza Data base, and 28 additional viruses were sequenced. Maximum clade credibility (MCC) phylogenetic tree revealed that the M gene has evolved into two different lineages. Estimated Evolutionary analysis showed that the M2 protein possessed higher evolutionary rates (3.45 × 10^-3) than the M1 protein (2.73 × 10^-3). M gene encoding proteins revealed significant markers described to be associated with host tropism and increase in virulence: V15I, N30D, and T121A in M1 and L55F in M2 protein. Site analysis focusing attention on the temporal and host distribution of the amantadine-resistant markers was carried out and showed that vast majority of the M2 amantadine-resistant variants of clade 2.2.1.1 (n = 90) is N31 marker, in addition to G27 (n = 7), A27 (n = 5), I27 (n = 1), and S30 (n = 1). In 2010-2011, amantadine resistant frequency increased considerably resembling more than half of the resistant variants. Notably, all viruses of clade 2.2.1.1 possessed amantadine-resistant marker. However, almost all current circulating viruses in Egypt of clade 2.2.1.2 from 2014 to 2016 did not carry any amantadine-resistant markers.

Amino acid substitutions occurring during adaptation of an emergent H5N6 avian influenza virus to mammals

Avian influenza viruses (AIVs) are known to cross species barriers, and emergent highly pathogenic H5N6 AIVs pose a serious threat to human health and the poultry industry. Here, we serially passaged an H5N6 virus 10 times in BALB/c mice. The pathogenicity of the wild-type 6D2 (WT-6D2) and mammal-adapted 6D2 strain (MA-6D2) were compared. The viral titer in multiple organs and the death rate for MA-6D2 were significantly higher than for WT-6D2. We provide evidence that the mutations HA A150V, NA R143K and G147E, PB2 E627K, and PA A343T may be important for adaptation of H5N6 AIVs to mammals.

Stepwise adaptation of murine cytomegalovirus to cells of a foreign host for identification of host range determinants

Ever since their first isolation 60 years ago, cytomegaloviruses have been recognized as being highly species specific. They replicate only in cells of their own or a closely related host species, while cells of phylogenetically more distant hosts are usually not permissive for viral replication. For instance, human cytomegalovirus replicates in human and chimpanzee fibroblasts but not in rodent cells, and murine cytomegalovirus (MCMV) replicates in cells of mice and rats but not in primate cells. However, the viral and cellular factors determining the narrow host range of cytomegaloviruses have remained largely unknown. We show that MCMV can be adapted stepwise to replicate in cultured human retinal pigment epithelial (RPE-1) cells and human fibroblasts. The human RPE-1 cells used for the initial adaptation step showed a pronounced contact inhibition and produced very low level of interferon-β transcripts upon cytomegalovirus infection, suggesting that these cells provide a particularly favorable environment for adaptation. By whole genome sequencing of the 230 kbp viral genomes of several adapted mutants, a limited number of mutations were detected. Comparison of several human cell-adapted MCMV clones and introduction of specific mutations into the wild-type MCMV genome by site-directed mutagenesis allows for the identification of viral host range determinants and provides the basis for elucidating the molecular basis of the cytomegalovirus host species specificity.