The epidemiological success of seasonal and pandemic influenza viruses depends on their ability to transmit efficiently from person to person. The size of infectious virus–containing particles released by infected individuals and the relative importance of fomites and airborne transmission is a matter of debate. Additionally, the contributions of viral, host and environmental factors on the transmissibility of influenza viruses are poorly understood. For the last decade, investigators have relied upon animal models such as ferrets and guinea pigs to assess this complex process. Conventional studies use small numbers of animals with exposure lasting 14 d that includes contact transmission, in which the infected donor and naive recipient animals are co-housed, or airborne transmission, where perforated panels separate them (Fig. 1).

Figure 1: Transmission of influenza viruses in animal models and humans.
figure 1

The design of animal transmission studies is not standardized, and they include either contact or airborne transmission. In humans, influenza viruses are transmitted in shared spaces such as classrooms, healthcare settings, work spaces and public transport, by direct or indirect contact via contaminated surfaces (fomites) and by respiratory droplets. These droplets include a wide range of particle sizes that are released from different parts of the respiratory tract; large particles fall out of the air quickly and can be deposited onto surfaces. Certain parameters that probably influence animal transmission by respiratory droplets could be modified to better mimic transmission in humans.

In a recent study, Killingley et al.1 sought to characterize influenza transmission in humans using the human challenge model that has been used to study disease progression and vaccine efficacy2,3,4. The authors examined the transmission of influenza from donor volunteers who were experimentally infected by intranasal administration of an influenza virus to healthy subjects who shared living conditions in a quarantine facility. The donors and recipients were separated except during exposure events that took place over several hours. The clinical signs and symptoms in the experimentally infected volunteers, which are known to be milder than in natural infection5, were runny nose, stuffy nose and cough. The authors assessed the deposition of influenza virus in the environment, on inanimate surfaces and in the air, which was sampled and size fractionated to assess the size of exhaled particles containing influenza virus. This proof-of-concept study establishes the feasibility of studying influenza transmission in humans and highlights several important considerations for future studies in human subjects, such as a low incidence of illness and lack of severe symptoms in infected donors. Notably, for bench researchers, this study also suggests crucial ways by which animal models of influenza transmission can and should be refined.

Although 77% of the donors were infected, the virus only transmitted to 25% of recipients; in contrast, the transmission efficiency of seasonal influenza viruses is 100% in animal models6,7,8. There are potential explanations for this discrepancy. First, humans often have prior experience with circulating influenza viruses, whereas influenza-naive animals are used in animal studies. Preexisting immunity is known to prevent or limit virus replication and might therefore limit the spread of influenza. During the 2009 H1N1 pandemic, elderly individuals were protected from severe disease because of prior exposure to H1N1 viruses that circulated before 1950 (ref. 9). Similarly, prior infection of donor guinea pigs with a seasonal influenza virus decreased the transmissibility of the pandemic H1N1 virus to recipient animals8. In addition to antibody-mediated immunity, T cells have been implicated in limiting disease symptoms and early viral clearance in humans4; these factors could reduce the likelihood of transmission to contacts. However, the role of memory T cells in transmission has not been explored in animal models owing to a lack of appropriate immunological reagents.

Second, human exposure events are limited to a few hours, as in the study by Killingley et al.1 in which exposure events included sharing meals, playing games and watching TV. Transmission studies in animals, however, involve continuous exposure, with the recipient and donor animals housed a few feet apart for 14 d. Possible modifications of animal studies could better mimic human transmission events (Fig. 1).

Influenza viruses can transmit between humans via direct or indirect contact and by respiratory droplets, but their relative contribution is unknown10,11. Contact transmission refers to viruses transferred by skin contact with secretions from an infected person, by way of contaminated hands or contact with a contaminated inanimate object, to mucous membranes of the eyes or nose. Respiratory droplet transmission is caused by a spray of particles that include large (>5 μm) droplets and small (<5 μm) particles. Large droplets can be deposited on inanimate objects and onto mucous surfaces, whereas small particles remain suspended in the air, can travel more than 1 m from the source to the recipient and can be inhaled into the lower respiratory tract (Fig. 1). Respiratory droplet rather than contact transmission is held as the gold standard for assessing the transmissibility of influenza viruses in ferrets and guinea pigs because animal contact behavior differs from humans, and influenza viruses that are epidemiologically unsuccessful in humans are able to transmit efficiently between ferrets via contact12. Contact transmission studies in animals require co-housed animals to share water bottles and food, and social and playful animals, such as ferrets, maintain close contact with their cage mates. Disregarding contact transmission in animal models might result in the loss of some potentially important information about influenza transmission. Therefore, the importance of this mode of transmission needs to be studied in the human challenge transmission model. Understanding the relative contribution of each mode of transmission will help tailor the use of nonpharmaceutical interventions, such as hand washing, social distancing measures and cough etiquette, and infection control measures, such as bed spacing and personal protective equipment, to limit the spread of influenza.

Air samplers developed by the US National Institute for Occupational Safety and Health fractionate aerosols by size, and the presence of viral nucleic acid can be determined by PCR13. Analysis of aerosols produced in clinical settings and from individuals who cough suggests that humans predominantly release submicron particles containing influenza14,15. Respiratory droplet transmission in ferrets correlated with release of submicron particles containing influenza RNA, and this release was linked to specific viral properties16. Infected ferrets also exhale aerosol plumes similar to humans, during tidal breathing and induced sneezing17. However, the duration and distribution of aerosol release may differ between intranasally inoculated animals and naturally infected humans. It is important to understand the similarities and differences between aerosol generation in humans and animal models.

To take the findings from the clinical study back to the bench, the design of transmission studies in animal models can be modified by changing environmental conditions, assessing aerosol release, using partially immune animals and mimicking human contact behavior (Fig. 1). Human exposure events, such as a short bus ride (30-min exposure), a long flight (8-h exposure), or intermittent daily exposure to a co-worker (1 h exposure every day for 5 d) can be mimicked. In addition, the timing of exposure of naive animals to experimentally infected animals can be limited to exposure early or late in the course of infection. Although such modifications will probably decrease the transmission efficiency in animal models and will require the use of larger numbers of animals per study, the experiments will better reflect the transmissibility of viruses in humans. A new, alternative approach to studying transmission without animals is the use of simulated coughing and breathing machines, which showed that a properly fitted N95 respirator or surgical mask was effective in blocking entry of the virus into the mouth18.

Understanding the mechanism and molecular basis of influenza transmission is essential for assessing the pandemic potential of novel influenza viruses and for the development of public health intervention strategies for influenza. Influenza virus transmission can be interrupted by the use of nonpharmaceutical interventions as well as by vaccination and antiviral prophylaxis if implemented early2,19. Human challenge studies will extend our understanding of human transmission events in controlled settings and should guide modification of animal models; the latter are of great value, as they allow us to tease apart molecular and viral properties that contribute to transmission using engineered viruses. A two-pronged approach that combines clinical research with the use of animal models will allow scientists to understand how we catch the flu.