Does ligand valency change how Fc-conjugated antivirals fight influenza?

Valency becomes a controllable design variable

The paper asks whether attaching more copies of a small-molecule neuraminidase inhibitor to an antibody Fc domain does more than strengthen binding. To test this, the authors built a click-chemistry platform that links zanamivir, oseltamivir, or peramivir to Fc in defined valencies. Using zanamivir as the model ligand, they produced mono-, di-, and trivalent conjugates (AVC-Z1, AVC-Z2, AVC-Z1H, AVC-Z3) plus a short-linker control.

The measured drug-to-Fc ratios were 4.2, 4.0, 11.8, and 3.7 for AVC-Z1, AVC-Z2, AVC-Z1H, and AVC-Z3 respectively, corresponding to roughly 4.2, 8.0, 11.8, and 11.1 zanamivir molecules per conjugate. That matters because AVC-Z1H carries about as many zanamivir units as AVC-Z3 but arranges them differently, letting the study separate raw ligand count from spatial arrangement.

Higher valency tightens NA binding and inhibition

Surface plasmon resonance showed affinity for influenza neuraminidase (NA) rising with valency. Against N2, equilibrium dissociation constants were 23.00 nM for AVC-Z1, 10.09 nM for AVC-Z2, 8.33 nM for AVC-Z1H, and 6.71 nM for AVC-Z3; AVC-Z3 also dissociated most slowly. All conjugates bound more tightly than free zanamivir.

The functional readout followed the same order. AVC-Z3 inhibited PR8 NA 17-, 10-, and 5-fold better than AVC-Z1, AVC-Z2, and AVC-Z1H, and retained activity against the oseltamivir-resistant H275Y NA variant. In cell-based assays across five strains, potency ranked AVC-Z3 > AVC-Z2 > AVC-Z1H > AVC-Z1 > zanamivir, with AVC-Z3 median IC50 values of 10.98 nM against A/H1N1 and 7.75 nM against A/H3N2. The short-linker control AVC-Z1H-S, with a similar drug-to-Fc ratio, was markedly weaker, so linker geometry mattered alongside valency.

Cross-linking virions explains the valency advantage

Binding strength alone does not explain why AVC-Z3 beat AVC-Z1H, which carries a comparable zanamivir load. Dynamic light scattering showed AVC–NA complexes growing with valency, and AVC-Z3 produced NA aggregates around 100 nm versus roughly 20 nm for AVC-Z2 and AVC-Z1H. Electron microscopy confirmed that AVC-Z3 cross-linked NA proteins and viral particles more extensively.

The authors attribute this to geometry: the linkers in the trivalent conjugates span 23.6 to 37.3 Å, a range compatible with bridging NA tetramers on separate virions. Time-of-drug-addition experiments supported a dual-stage effect. AVC-Z3 reduced extracellular viral titres whether added at infection or afterward, and it lowered titres when virus was pretreated (P = 0.0052) or co-added (P = 0.0308), while pretreating cells alone did not block entry. Fluorescence imaging showed aggregated virus on MDCK and A549 cell surfaces. The paper therefore frames the mechanism as a shift from neuraminidase suppression to physical viral immobilization, coupling entry blockade with restricted release.

The peramivir conjugate adds persistence to cross-linking

The authors then swapped the ligand, choosing derivatization sites intended to preserve NA binding. The oseltamivir conjugate AVC-O3 failed this test: its IC50 against NA was 442.10 nM and cellular EC50 values exceeded 600 nM, which the authors attribute to steric hindrance from the C5 PEG chain, so it was dropped.

AVC-P3, the trivalent peramivir conjugate, kept potent activity with an NA IC50 of 3 nM against wild-type and H275Y mutant NA, where peramivir lost more than 10-fold activity against the mutant. Its KD was 1.43 nM and 3.29 nM against NA from A/Darwin/6/2021 and A/Sydney/5/2021, and it induced virus aggregation like AVC-Z3. In mice, AVC-Z3 had a half-life of 141.2 h versus 34.3 h for Fc protein alone, and AVC-P3 reached 225.02 h; in humanized FcRn mice AVC-P3 lasted 253 h versus 163 h for CD388 with a YTE mutation. These are pharmacokinetic measurements in specific mouse models, not human outcomes.

In vivo protection spans subtypes but stays in mice

In mouse infection models, AVC-P3 outperformed the comparison groups. The abstract reports that its in vivo efficacy surpassed other AVCs, baloxavir, and CD388 in H1N1, H3N2, H5N1, and influenza B models. In the H1N1 and H3N2 models, AVC-P3 reduced lung IL-6 and γ-IFN levels and improved lung histology relative to PBS, with reported P values below 0.05 for the comparisons shown.

A sequential-infection experiment sharpened the contrast with baloxavir. Baloxavir lowered lung viral load best in the first round, but treated mice began dying after a second infection on day 15, whereas AVC-Z3 and AVC-P3 at 3 mg/kg gave 100% protection through day 29. AVCs also reduced lung viral copies per gram of tissue by 100- to 1000-fold versus untreated and baloxavir groups. All of this comes from mouse models, with female mice used throughout and dosing given intravenously, so the translation to human prophylaxis or treatment remains untested here.

Valency, not just potency, should guide conjugate design

The paper's central claim is that ligand valency is a determinant of antiviral mechanism, not merely of binding strength. The clearest internal comparison is AVC-Z3 versus AVC-Z1H: similar zanamivir loading, different spatial arrangement, and a large gap in aggregation and antiviral activity. That makes linker length and geometry part of the design specification rather than an afterthought.

For readers weighing this against existing Fc-conjugated antivirals, the relevant priority is the combination shown by AVC-P3: cross-linking activity retained from the trivalent zanamivir design, plus a half-life longer than AVC-Z3 and than CD388 with a YTE mutation in the mouse systems tested. The open question is whether that combination holds in humans, since the efficacy and pharmacokinetic evidence here comes from mouse models and the paper does not report human data.

Sources used in this answer