Simon Gubbins ¹ Basudev Paudyal ¹ Barbara Dema ² Ashutosh Vats ¹ Marta Ulaszewska ² Eleni Vatzia ¹ Elma Tchilian ^1* Sarah Catherine Gilbert ²

• ¹ The Pirbright Institute, Woking, United Kingdom
• ² Pandemic Sciences Institute, University of Oxford, Oxford, England, United Kingdom

Whereas the intranasally delivered influenza vaccines used in children affect transmission of influenza virus in the community as well as reducing illness, inactivated influenza vaccines administered by intramuscular injection do not prevent transmission and have a variable, sometimes low rate of vaccine effectiveness. Although mucosally administered vaccines have the potential to induce more protective immune response at the site of viral infection, quantitating such immune responses in large scale clinical trials and developing correlates of protection is challenging. Here we show that by using mathematical models immune responses measured in the blood after delivery of vaccine to the lungs by aerosol can predict immune responses in the respiratory tract in pigs. Additionally, these models can predict protection from influenza virus challenge despite lower levels of blood responses following aerosol immunization. However, the inclusion of immune responses measured in nasal swab eluates did not improve the predictive power of the model. Our models are an important first step, providing proof of principle that it is feasible to predict immune responses and protection in pigs. This approach now provides a path to develop correlates of protection for mucosally delivered vaccines in samples that are easily accessed in clinical trials.