Development of a stable liquid formulation of live attenuated influenza vaccine

Abstract

Vaccination is the most effective means of preventing influenza. However, the cost of producing annual seasonal influenza vaccines puts them out of reach for most developing countries. While live attenuated influenza vaccines are among the most efficacious and can be manufactured at low cost, they may require lyophilization to be stable enough for developing-country use, which adds a significant cost burden. The development of a liquid live attenuated seasonal influenza vaccine that is stable for around a year-the duration of an annual influenza season-would significantly improve not only the production output but also the use and accessibility of influenza vaccines in low-resource settings. In this study, potential stabilizing excipients were screened and optimized using the least stable influenza vaccine strain presently known, H1N1 (A/California/07/2009), as a model. The stability-conferring properties of the lead formulations were also tested with a Type B strain of influenza virus (B/Brisbane/60/2008). Stability was also evaluated with higher titers of influenza virus and exposure to agitation and freeze-thaw stresses to further confirm the stability of the lead formulations. Through this process, we identified a liquid formulation consisting of sucrose phosphate glutamate buffer with 1% arginine and 0.5% recombinant human serum albumin that provided storage stability of one year at 2-8°C for the influenza A and B strains tested.

Keywords: Influenza vaccine; Vaccine stabilization.

Fig. 1

Screening of excipients. Formulations were prepared with LAIV H1N1 (A/California/07/2009) in SPG buffer and additional excipients were prepared at a titer of 2 × 10⁶ log₁₀ TCID₅₀/mL. The formulations were stored at 2–8 °C (A) and 25 °C (B), and their titer was measured by TCID₅₀. Formulations were tested for up to 6 weeks or until a titer loss greater than 1 log was observed. N = 1 for each formulation tested by TCID₅₀ in triplicate. Error bars represent the standard deviation of three TCID₅₀ replicates. Abbreviations: BSA; bovine serum albumin.

Fig. 2

Formulation optimization. Formulations of LAIV H1N1 (A/California/07/2009) in SPG buffer and additional excipients were prepared at a titer of 2 × 10⁶ log₁₀ TCID₅₀/mL. The formulations were stored at at 2 °C–8 °C (A), 25 °C (B), and 33 °C (C), and their titer was measured by TCID₅₀. Formulations were tested for up to 20 weeks or until a titer loss greater than 1 log was observed. N = 1 for each formulation tested by TCID₅₀ in triplicate. Error bars represent the standard deviation of three TCID₅₀ replicates. Abbreviations: SPG, sucrose phosphate glutamate; BSA, bovine serum albumin; HPMC, hydroxypropyl methylcellulose; PVP, polyvinylpyrrolidone.

Fig. 3

Formulation validation. Lead formulations were evaluated with an alternate LAIV strain, B/Brisbane/60/2008. Formulations of LAIV H1N1 or type B were prepared in SPG buffer or SPG buffer with 1% arginine at a titer of 2 × 10⁶ log₁₀ TCID₅₀/mL. The formulations were stored at 2–8 °C (A), 25 °C (B), and 33 °C (C), and their titer was measured by TCID₅₀. Formulations were tested for up to 52 weeks or until a titer loss greater than 1 log was observed. N = 1 for each formulation tested by TCID₅₀ in triplicate. Error bars represent the standard deviation of three TCID₅₀ replicates.

Fig. 4

Confirmation of lead formulations containing increased LAIV titer. Formulations of LAIV H1N1 (A/California/07/2009) or type B (B/Brisbane/60/2008) in SPG buffer and selected excipients were prepared at a titer of 1 × 10⁷ log₁₀ TCID₅₀/mL. The formulations were stored at at 2–8 °C (A), 25 °C (B), and 33 °C (C), and their titer was measured by TCID₅₀. Formulations were tested for up to 52 weeks or until a titer loss greater than 1 log was observed. N = 1 for each formulation tested by TCID₅₀ in triplicate. Error bars represent the standard deviation of three TCID₅₀ replicates. (D) Summary of time (in weeks) to 1 log loss for the lead formulations identified represents the lower 95% confidence interval from the average of all three experiments lead formulations were evaluated. NT = not tested.

Fig. 5

Effect of agitation and freeze–thaw stresses on LAIV titer. Formulations of LAIV H1N1 or Type B in SPG buffer and selected excipients were prepared at a titer of 2 × 10⁶ log₁₀ and 1 × 10⁷ log₁₀ TCID₅₀/mL. Agitation was performed by shaking vials horizontally for 24 h at 200 rpm at ambient temperature (20–25 °C). Freeze–thaw was performed by freezing vials at −20 °C for 2 h, then thawing at 20–25 °C for 1 h for a total of 3 freeze–thaw cycles. After the stress tests were completed, the formulations were tested by TCID_50. Error bars represent standard deviation of 3 formulation vials tested in triplicate.