Scientists Use Immune Informatics to Sketch a First Vaccine Blueprint Against Human Metapneumovirus

Human metapneumovirus, a pathogen first identified in 2001, has quietly become one of the most consequential respiratory viruses in the world, yet it still has no licensed vaccine. A new computational study published in Frontiers in Virology reports the design of two multi-epitope vaccine candidates against HMPV that incorporate the human immune molecules BAFF and APRIL as built-in immunostimulatory components. The work, led by Reham M. Alahmadi and colleagues at King Saud University, King Fahad Medical City, Shendi University, and the Children’s Hospital Affiliated to Shandong University, applies a reverse vaccinology strategy in which candidate vaccines are assembled entirely from computationally predicted immune targets before any laboratory work begins. The authors stress that their findings are predictions requiring experimental validation, but the study offers one of the most detailed blueprints yet for a vaccine against a virus estimated to cause roughly 11 million episodes of lower respiratory tract infection in children under five each year.

The clinical burden of HMPV is substantial and often underestimated. The virus accounts for approximately 4 to 13 percent of hospitalizations associated with lower respiratory tract infections in children under five, and it is recognized as the second most common viral cause of pediatric lower respiratory disease after respiratory syncytial virus. Globally, HMPV is estimated to produce nearly 502,000 hospitalizations and around 11,300 deaths annually in young children, placing its overall impact somewhere between that of RSV and influenza. Severe disease also strikes older adults and immunocompromised individuals. Part of the difficulty in developing a vaccine lies in the virus’s own biology: HMPV interferes with antiviral signaling pathways, suppresses interferon production, and impairs immune cell function, which helps explain why natural infection produces weak, short-lived protective immunity and why reinfections are common.

To overcome this, the research team cast a deliberately wide net across the viral proteome. They retrieved sequences of all eight major HMPV proteins from the CAN97-83 lineage: the fusion (F) and attachment (G) glycoproteins on the viral surface, the small hydrophobic (SH) protein, the matrix proteins M2-1 and M2-2, the nucleoprotein (N), the phosphoprotein (P), and the large RNA-dependent RNA polymerase (L). Each protein was screened for predicted antigenicity using the VaxiJen v2.0 server. The nucleoprotein scored highest at 0.5286, followed by the phosphoprotein, the polymerase L (0.4770), fusion F (0.4711), attachment G (0.4710), M2-1 (0.4163), SH (0.3781), and M2-2 (0.3630). Including internal and replication proteins alongside the surface glycoproteins reflects evidence that both classes contribute to humoral and cellular immunity, making a multi-antigen approach potentially more protective than designs targeting surface proteins alone.

From these proteins, the team predicted three classes of immune epitopes. Cytotoxic T lymphocyte (CTL) epitopes, which would be presented on MHC class I molecules to CD8-positive killer T cells, were identified with NetCTL v1.2 and the Immune Epitope Database (IEDB) resources. Helper T lymphocyte (HTL) epitopes, presented on MHC class II to CD4-positive helper cells, were predicted with NetMHCIIpan and IEDB tools using 10 to 15 amino acid peptides. Linear B-cell epitopes, the targets of antibody responses, were predicted with the IEDB B-cell prediction tool. Every candidate peptide then passed through a safety and quality filter: VaxiJen confirmed predicted antigenicity above a threshold of 0.4, AllergenFP excluded potential allergens, ToxinPred screened out toxic peptides using a threshold of 0.9, and the IFNepitope server flagged peptides likely to induce interferon-gamma, a cytokine central to antiviral immunity.

The final vaccine constructs each contain 24 non-overlapping epitopes: eight CTL, eight HTL, and eight B-cell epitopes, one of each type drawn from every analyzed HMPV protein. The epitopes were stitched together with peptide linkers chosen for their conventional roles in multi-epitope design: alanine-alanine-tyrosine (AAY) linkers between CTL epitopes to support proteasomal processing, glycine-proline-glycine-proline-glycine (GPGPG) linkers between HTL epitopes, and lysine-lysine (KK) linkers between B-cell epitopes. The core epitope region spans 315 amino acids. What distinguishes this study from earlier HMPV immunoinformatics work is the addition of the human B-cell activating factor (BAFF, a 285-amino-acid TNF family member) or the proliferation-inducing ligand APRIL (250 amino acids) at the N-terminus of each construct, connected through an EAAAK linker. Both molecules naturally regulate B-cell survival, differentiation, and antibody production, and prior experimental vaccine studies have shown that BAFF and APRIL can enhance germinal-center reactions and antigen-specific antibody responses in animal models.

The physicochemical profiles of the two constructs suggest they could, in principle, be manufactured as recombinant proteins. The BAFF-adjuvanted construct comprises 533 amino acids with a molecular weight of 57.8 kilodaltons, while the APRIL-containing version is 496 amino acids and 53.63 kilodaltons. Both are predicted to be stable, with instability indices of 32.98 and 31.76 respectively, values comfortably below the threshold of 40 that flags instability. High aliphatic indices (80.28 and 81.69) and negative grand average hydropathicity scores (-0.455 and -0.425) indicate thermostability and hydrophilicity, the latter suggesting compatibility with aqueous solubility. Both constructs were predicted to be antigenic (VaxiJen scores of 0.6776 for the BAFF version and 0.7243 for the APRIL version) and non-allergenic and non-toxic by the screening tools. Secondary structure predictions from the GOR4 and PSIPRED servers showed a balanced mix of alpha-helices, beta-strands, and random coils, with coils predominating, a pattern the authors interpret as supporting both structural stability and exposed, accessible epitopes.

To probe how the constructs might engage the innate immune system, the team performed protein-protein molecular docking against two Toll-like receptors. TLR4 was chosen because previous experimental studies have implicated it in innate responses to HMPV, while TLR1 was included as an exploratory target, with the caveat that TLR1 normally functions as a heterodimer with TLR2. Using structures generated with AlphaFold and the HADDOCK docking platform, the analysis placed each vaccine construct along the leucine-rich repeat domain of the receptors, with multiple predicted contacts. In the TLR4 complex, hydrogen bonds were predicted between Gln185 and Gln187 of the construct and Asn622 and Ser623 of TLR4, along with electrostatic, hydrophobic, and salt-bridge contacts involving residues such as Glu30, Pro27, and Asp50 on the vaccine side and Arg154, Lys184, and Arg337 on the receptor. The TLR1 complex similarly showed hydrogen bonding between construct residue Gln178 and receptor residues Asn518 and Ser515. The authors are explicit that docking demonstrates only structural compatibility, not binding affinity, receptor activation, or signaling.

An immune simulation with the C-ImmSim server, run over 100 time steps under a defined human HLA background, painted a theoretically favorable response profile. Following a single modeled immunization, IgM antibodies appeared first and peaked around days 10 to 12, followed by a class-switched response dominated by IgG1. Interferon-gamma showed the strongest cytokine response, peaking near day 4, with a prominent early IL-2 peak, a pattern consistent with a Th1-oriented cellular response relevant to antiviral defense. Simulated B-cell populations rose to roughly 470 to 480 cells per cubic millimeter by day 6, memory B cells accumulated progressively, and the helper T-cell compartment expanded to about 5,500 cells before declining, with memory T-helper cells increasing to approximately 400 to 450. Population coverage analysis using the IEDB tool predicted that the selected HLA-restricted T-cell epitopes would cover 100 percent of individuals in all eight geographic populations examined, from Central Africa to East Asia to Europe and the Americas, though the number of epitope-HLA combinations recognized per person varied by region.

The study’s authors and its framing are careful about the limits of the work, and rightly so. Every result, from antigenicity scores to docking poses to simulated cytokine kinetics, is a computational prediction resting on algorithmic assumptions rather than biological measurement. The predicted non-allergenic and non-toxic profiles do not establish clinical safety, and the incorporation of BAFF and APRIL carries its own cautionary note: excessive or dysregulated signaling through this pathway is associated with persistence of autoreactive B cells and autoimmune disease, so these endogenous molecules should not be assumed safer than conventional adjuvants. The BAFF and APRIL receptors that would actually mediate their effects, BAFF-R, TACI, and BCMA, were not docked and remain targets for future study. The authors call for molecular dynamics simulations, binding free-energy calculations, recombinant expression, in vitro immunological assays, and ultimately in vivo studies of immunogenicity, safety, and protective efficacy. Until those experiments are done, the constructs stand as rigorously prioritized candidates, an early but structurally detailed step toward closing a significant gap in respiratory virus prevention.

Subject of Research: Immunoinformatics-guided design of BAFF- and APRIL-adjuvanted multi-epitope vaccine candidates against human metapneumovirus

Article Title: Immunoinformatics-guided design and in silico evaluation of BAFF- and APRIL-adjuvanted multi-epitope vaccine candidates against human metapneumovirus

Article References: Immunoinformatics-guided design and in silico evaluation of BAFF- and APRIL-adjuvanted multi-epitope vaccine candidates against human metapneumovirus. (n.d.). https://doi.org/10.3389/fviro.2026.1949694

Image Credits: AI Generated

DOI: 10.3389/fviro.2026.1949694

Keywords: human metapneumovirus, immunoinformatics, reverse vaccinology, multi-epitope vaccine, BAFF, APRIL, Toll-like receptors, molecular docking, vaccine design, interferon-gamma, HLA population coverage, immune simulation