A universal dual mechanism immunotherapy for the treatment of influenza virus infections

Abstract

Seasonal influenza epidemics lead to 3-5 million severe infections and 290,000-650,000 annual global deaths. With deaths from the 1918 influenza pandemic estimated at >50,000,000 and future pandemics anticipated, the need for a potent influenza treatment is critical. In this study, we design and synthesize a bifunctional small molecule by conjugating the neuraminidase inhibitor, zanamivir, with the highly immunogenic hapten, dinitrophenyl (DNP), which specifically targets the surface of free virus and viral-infected cells. We show that this leads to simultaneous inhibition of virus release, and immune-mediated elimination of both free virus and virus-infected cells. Intranasal or intraperitoneal administration of a single dose of drug to mice infected with 100x MLD₅₀ virus is shown to eradicate advanced infections from representative strains of both influenza A and B viruses. Since treatments of severe infections remain effective up to three days post lethal inoculation, our approach may successfully treat infections refractory to current therapies.

Fig. 1. Dual mechanisms of action of neuraminidase-targeted anti-influenza therapy.

a Both influenza virus and influenza virus-infected cells express a viral neuraminidase (orange receptor) on their membrane surfaces. b Binding of a zanamivir-dinitrophenyl conjugate (zan-DNP) to these neuraminidases not only inhibits neuraminidase activity and thereby suppresses virus budding from the host cell, but also decorates both the free virus and infected host cell with the highly immunogenic hapten, dinitrophenyl (DNP). This painting of the virus and infected host cell with zan-DNP then recruits naturally occurring anti-DNP antibodies that are present in the blood of virtually all humans, resulting in opsonization of the virus/virus-infected cell and the consequent immune-mediated clearance of the virus/infected cell. c zan-DNP was synthesized by conjugating zanamivir (targeting ligand) with 2,4-dinitrophenyl group (hapten) via a polyethylene glycol (PEG) chain (linker).

Fig. 2. Binding affinity and specificity of zan-DNP for viral neuraminidases.

a Crystal structure of N1 neuraminidase of influenza A virus complexed with zanamivir (PDB code: 3TI5). Note that the C-7 hydroxyl of zanamivir is exposed to solvent and available for linkage to DNP. b Structure of zanamivir-rhodamine conjugate (zan-rhodamine). c–f Binding of zan-rhodamine (direct binding curve, left panel), zanamivir (competitive binding curve, central panel), and zan-DNP (competitive binding curve, right panel) to neuraminidase expressed on MDCK cells infected with influenza virus A/Wisconsin/629-D00015/2009 (H1N1)pdm09 (c), A/Netherlands/22/2003 (H3N2) (d), B/Florida/4/2006 (Yamagata Lineage) (e) or B/Brisbane/60/2008 (Victoria Lineage) (f). All the virus-infected cells were used 24 h post-infection for binding studies. Data are presented as mean values ± SD (n = 3).

Fig. 3. Anti-DNP antibodies recruiting, complement dependent cytotoxicity (CDC), and antibody-dependent cellular cytotoxicity (ADCC) of neuraminidase expressing cells induced by zan-DNP.

a and c Confocal micrographs of binding of biotinylated anti-DNP antibodies (visualized by streptavidin-PE) to the surface of A/Puerto Rico/8/1934 (H1N1) (a) and A/Aichi/2/1968 (H3N2) (c) virus-infected MDCK cells mediated by zan-DNP (n = 3). Nuclear staining is shown in blue, and streptavidin-PE staining is shown in red. b, d Binding of anti-DNP antibodies to influenza A(H1N1) (b) and A(H3N2) (d) virus-infected MDCK cells determined by quantitating the fluorescence of cell-bound streptavidin-PE in the presence of anti-DNP antibodies as a function of the concentration of zan-DNP (n = 3). All the virus-infected cells were used 24 h post-infection for the above studies. e Flowcytometry analyses of recruitment of anti-DNP antibodies to virus-infected cells mediated by zan-DNP in vivo (n = 3). A/California/07/2009 (H1N1)pdm09 virus-infected mice (3 mice/group) were first given zan-DNP and anti-DNP antibodies (3 days post-infection). 12 h later, the cells digested from the lung of virus-infected/uninfected mice were stained first with anti-hemagglutinin antibodies and then labeled with dye conjugated secondary antibodies against anti-hemagglutinin antibodies and anti-DNP antibodies. Left: selected flow cytometry plots; Right: summary graph. Statistical difference of antibody stained positive cells (%) between indicated two groups was analyzed by two-sided t test (*P < 0.05, **P < 0.01, ***P < 0.001, see exact P values in Supplementary Table 1). f CDC mediated killing of N1 neuraminidase transduced and nontransduced HEK 293 cells mediated by zan-DNP in the presence of anti-DNP antibodies and complement-preserved rabbit serum (n = 3). g Antibody-dependent cellular cytotoxicity (ADCC) of N1 neuraminidase transduced and nontransduced HEK 293 cells mediated by zan-DNP in the presence of anti-DNP antibodies and human FcγRIIIa expressing effector cells (n = 3). ADCC activity is proportional to the amount of firefly luciferase produced by effector cells (detected as the luminescence intensity). The coding sequence of N1 neuraminidase for the transduction of HEK 293 cells is from strain A/Puerto Rico/8/1934 (H1N1); Data are presented as mean values ± SD (n = 3).

Fig. 4. Therapeutic efficacy of zan-DNP in protecting mice from lethal influenza virus infections.

DNP-KLH immunized BALB/c mice were challenged with 100x MLD₅₀ of the different strains of influenza virus indicated above. Mice were treated with the indicated test articles by intranasal administration (IN) using the dosage regimens indicated above each panel. Mice were counted as dead when they lost either 25% of their initial weight or became moribund. Body weight curves (left panels) and survival curves (right panels) for each treatment are shown. (a) compares the efficacy of zan-DNP with its component parts (5 mice/group for zan-DNP and zanamivir treatment groups, 8 mice/group for the rest of the groups). (b) examines the effect of different delays between virus inoculation and zanamivir or zan-DNP administration on weight loss and animal survival (5 mice/group). (c) investigates the potency of a single dose of either zanamivir or zan-DNP on the same parameters in mice infected with viruses indicated above (5 mice/group). Statistical differences between PBS and drug treatment groups were determined by two-sided log-rank test (*P < 0.005, see exact P values in Supplementary Table 1). Body weight change (%) are presented as mean values ± SD.

Fig. 5. Examination of the properties of parenteral administration of zan-DNP.

a Structure of zanamivir-^99mTc conjugate (zan-^99mTc). b Binding of zanamivir-^99mTc to N1 neuraminidase (expressed on A/Puerto Rico/8/1934 (H1N1) virus-infected MDCK cells) in the absence (red curve) and presence (blue curve) of 100-fold excess zanamivir (competitor) (n = 3). c Biodistribution of intravenously injected (IV) zan-^99mTc in A/Puerto Rico/8/1934 (H1N1) virus-infected mice (3 mice/group). d SPECT/CT images showing the locations of intravenously injected zan-^99mTc in A/Puerto Rico/8/1934 (H1N1) virus-infected mice (left panel) or virus-infected mice which concurrently received 100-fold excess free zanamivir (right panel) (3 mice/group). a: lung; b: kidney; c: bowels. e Weight loss and survival curves of virus-infected mice treated with zan-DNP or zanamivir. DNP-KLH immunized BALB/c mice (5 mice/group) were challenged with 100x MLD₅₀ of the influenza virus strains indicated above each panel. Mice were treated with a single dose of zan-DNP by intraperitoneal administration (IP) 24 h post-infection. Mice were counted as dead when they lost either 25% of their initial weight or became moribund. Statistical differences between PBS and drug treatment groups were determined by two-sided log-rank test (*P < 0.005, see exact P values in Supplementary Table 1). Data are presented as mean values ± SD.

Fig. 6. Therapeutic efficacy of co-administration of zan-DNP plus exogenous anti-DNP antibodies in protecting nonimmunized mice from a lethal influenza virus infection.

a Nonimmunized BALB/c mice were challenged with 100x MLD₅₀ of influenza virus A/Puerto Rico/8/1934 (H1N1) prior to treatment with a single dose of 1.5 μmol/kg zan-DNP (intranasal administration, IN) plus exogenous anti-DNP antibodies (1–10 mg/kg, intravenous administration, IV) 24 h post-infection (3 mice/group for 3 mg/kg and 1 mg/kg anti-DNP antibody treatment groups, 5 mice/group for the rest of the groups). Control groups include (i) substitution of zanamivir for zan-DNP, (ii) omission of anti-DNP antibodies, and (iii) treatment with PBS. b Nonimmunized BALB/c mice (5 mice/group) were challenged with the indicated strains of influenza virus prior to treatment with a single dose of 1.5 μmol/kg zan-DNP or zanamivir (IN) plus 10 mg/kg exogenous anti-DNP antibodies (IV) 24 h post-infection. Mice were counted as dead when they lost either 25% of their initial weight or became moribund. Statistical differences between PBS and drug treatment groups were determined by two-sided log-rank test (*P < 0.005, see exact P values in Supplementary Table 1). Body weight change (%) are presented as mean values ± SD.