DiscussionIn most cases, influenza

Discussion
In most cases, influenza A viruses must be adapted through
serial passage in mice before they replicate to high titers or cause
disease in this model host (22, 23). Previous work has indicated
that exceptions to this rule include the reconstructed 1918
influenza virus (24) and some recent H5N1 isolates (20, 25). We
have confirmed these findings and, in addition, determined that,
despite the productive infection of mice by 1918 influenza or
A
Vietnam
1203
04 viruses, these pathogens are not transmitted
from infected mice to naı¨ve contacts. Furthermore, even the
highly mouse-adapted WSN virus was not transmitted from
mouse to mouse. These results are contrary to those reported in
a series of studies spanning from 1962 to 1968, which describe
influenza virus transmission among experimentally infected
mice (12, 26–29). The fact that, in these early studies, CFW or
NCS mice were inoculated by exposure to aerosol may account
for the discrepancy. However, even in these experiments, the
efficiency of transmission was low (5–62.5%) (27, 29), and
success with a given strain appeared to depend on the mouse
rather than any property of the infection (12). Based on these
data and our own findings, it is clear that mice are not an ideal
model to mimic the transmission pattern of influenza viruses
among humans.
In the absence of a feasible animal model, the question of
which viral and host factors influence influenza virus spread has
been largely neglected. However, recent fears of a pandemic
have brought this issue to the forefront of the influenza field;
indeed, the containment of epidemic and pandemic influenza
depends on understanding what enables an influenza virus to
transmit from person to person at a molecular level. To facilitate
work toward such understanding, we propose the guinea pig as
a mammalian model system for influenza virus transmission
studies. We have shown that an unadapted human influenza
virus replicates to high titers in the respiratory tract of guinea pigs and that the infection is passed from infected to sentinel
guinea pigs. Furthermore, the presence of a physical barrier, or
even 91 cm of intervening space, between infected and uninfected
animals did not preclude transmission, indicating that not
only contact but also large droplet and
or aerosol transmission
occurred. Further studies assessing transmission over larger
distances are required to determine whether true aerosol spread
of influenza virus occurs in the guinea pig model. Although
symptoms were not observed in the infected Hartley strain
guinea pigs, this apparent lack of disease is also observed in a
proportion of human influenza cases (30). Furthermore, influenza
virus infection of inbred strain 13 guinea pigs (obtained
from the U.S. Army Medical Research Institute of Infectious
Diseases, Fort Detrick, MD) by means of droplet spread resulted
in severe disease in these animals, including weight loss, lethargy,
hair loss, and hypothermia (data not shown). It will be interesting
to compare the proficiency of these two guinea pig strains in
transmitting influenza viruses.
The ferret is well established as an animal model in influenza
virus research because of its natural susceptibility to infection
and the manifestation of influenza symptoms similar to those
seen in human hosts (15, 31). In addition, transmission of
influenza virus infection between ferrets has been reported to
occur efficiently (13, 14, 31). Nevertheless, reasons of practicality
limit the use of ferrets in academic research, with the result that
the number of influenza virus studies performed in mice far
outweighs that in ferrets, despite the superiority of the ferret
model. Occasionally, hamsters have been used as an experimental
model for influenza virus infection (32–34), and, in one
report, they were shown to be competent for transmission.
Hamster-to-hamster spread was not observed between adjacent
cages, and the consistency with which transmission occurred
varied with the virus strain that was analyzed (34).
It is not yet clear why influenza virus transmits efficiently from
guinea pig to guinea pig but not from mouse to mouse. The lower
susceptibility of mice to infection with most influenza virus
strains likely contributes to this difference but does not account
for the lack of transmission of mouse-adapted viruses, such as
WSN. Perhaps the process of adaptation to the murine host
concomitantly reduces the intrinsic transmissibility of an influenza
virus. Our data indicated that, for all strains tested,
influenza viruses grew to higher titers in the lungs of mice than
in the nose, whereas the opposite was true for Pan
99 virus in
the guinea pig; thus, the predominant site of viral replication in
a given host may dictate whether transmission occurs. Transmission
experiments combining mice and guinea pigs as either
donors or receivers of influenza virus should shed some light on
the reason that mice fail to propagate infection; furthermore, the
physiological and
or virological basis for this observation may in
turn provide clues as to why human hosts do not transmit avian
influenza viruses.
Indeed, a guinea pig model of influenza virus transmission
holds great potential for the elucidation of both host and viral
factors contributing to influenza virus spread. Once the determinants
of transmission have been mapped, effective strategies
for the containment of influenza outbreaks at their inception
might be developed.