The present study revealed several interesting findings regarding infAv receptors and infAv distributions in pigs. The SA-α-2,6-terminal saccharide receptor was found to be present in high amounts in all areas of the respiratory tract of all examined pigs regardless of their infection status. In contrast, the SA-α-2,3-terminal saccharide receptor was present only on the luminal surface of bronchiolar, but mostly alveolar epithelial cells in the lower respiratory tract.
Most human and swine influenza A viruses show highest affinity for SA-α-2,6-terminal saccharide receptors while most avian influenza A viruses show highest affinity for SA-α-2,3 receptors [9, 13–16, 19, 29]. Mapping of the distribution of receptors in the respiratory tract of different infAv hosts is therefore highly relevant for clarifying host susceptibility to various infAv strains. Earlier studies of the respiratory tract of humans have shown that humans predominantly possess the SA-α-2,6 receptor and ducks expresses mainly the SA-α-2,3 receptor [30, 31]. However, humans also possess SA-α-2,3 in the lower part of the respiratory tract and domestic birds like chicken and quail have also been shown to express both receptors [26, 31–33]. In our study we were able to confirm that chickens possess SA-α-2,6 receptors in rather large amounts in intestines, trachea and to a lesser extent in lungs.
Previous studies have reported that pigs have both the SA-α-2,6 and the SA-α-2,3 receptors in the trachea and therefore pigs have been regarded as the primary mixing vessel for new infAv subtypes [8, 9]. In the present study we could not detect the SA-α-2,3-terminal saccharide in the upper respiratory tract of pigs. Our findings are in accordance with the results of recent studies by Nelli et al.  and Van Poucke et al. , the latter examined ex vivo explants of tissue from the respiratory tract of pigs and found the same receptor distribution in pig tissue as we did. The findings are also in concordance with similar studies in AIV infected humans showing that it is mostly the lower areas of the lung that are infected with AIV, resulting in pneumonia [33–35]. Taken together, these recent studies indicate that the "pig as mixing vessel" theory is less substantial than thought previously and that other species such as chickens and even humans could act as mixing vessels [26, 31, 32].
There are several possible explanations for the discrepancies in receptor distribution found in different studies. One obvious source of variables is the choice of lectins used for the receptor staining. Thus, Nicholls et al.  showed how lectins from different manufacturers may vary in specificity and sensitivity. In addition, both isoforms of MAAI and II recognise SA-α-2,3, but are different in the way they recognises the inner sugar structures. In order to detect all SA-α-2,3 receptors it is important to use both isoforms. Furthermore, since lectins are ubiquitously distributed in different tissues and constitute a component of mucus it is important to be able to differentiate respiratory epithelial cells from other lectin signal positive components in the respiratory tract. We have tried to meet these uncertainties in our scoring system by using relatively wide percentage ranges of the distribution in staining and by specifically evaluating the epithelial cells throughout the respiratory tract. This allowed us to identify the non-specific staining. Furthermore, the selection of pigs for the study may also be important since the receptor distribution may be dependent on various factors such as age, infection status, pig breed etc. However, this has not been documented.
Factors other than the distribution of receptors, may affect the host specificity of a given infAv [20, 33] and it has been shown that some infAv's are able to adapt to a given host species resulting in a change in the receptor specificity during infection [20, 36, 37]. However, the receptor recognition is a crucial step for an infecting virus and therefore it has a large impact on the successful attachment and further replication of virus.
It is likely that the risk of infection is dependent on both the amount of receptors present on the site of infection and the infection dose. In our experimental conditions very heavy exposure will of course increase the chance of infection. This is also seen in human infections with H5N1 where the majority of the infected individuals were exposed to high doses of infAv, probably allowing some of the viruses to enter the lower areas of the respiratory tract where the receptor preferred by the avian viruses is expressed [38, 39].
One of our observations was that endothelial cells were SNA and MAAI lectin positive. This finding has also been observed in receptor studies in humans . InfAv's predominantly infect the respiratory epithelial cells and are therefore not in contact with endothelial cells. However, in a damaged epithelium surface infAv may get in contact with endothelia cells. It is, however, most likely that other cell factors such as the availability of specific proteases impact the success of infecting e.g. the endothelia cells.
Interestingly, in parallel sections comparing infAv antigen positive cells with lectin staining it was demonstrated that the lectin staining was markedly diminished in areas where infAv antigen positive cells were present. A likely explanation could be that the infAv neuraminidases have cleaved off the sialic acids during the infection. Afterwards there could be a delay in re-establishing the receptors on infected cells or it could be a defence mechanism prompted by the host cell. To our knowledge this has not been shown earlier but is interesting because it would significantly impact the risk of a single cell being infected with different infAv subtypes and thereby reduce the risk of generating re-assorted viruses. Further studies should be performed on this issue.
The two different SIV isolates used in the present study were found to infect nose, trachea, bronchial, bronchiolar and alveolar type I and II epithelial cells, but the infections were mainly located in the bronchial epithelial cells. The distribution of the SIV antigen positive cells had a lobular distribution for both H1N1 and H1N2 infected pigs, but more lobules were affected in H1N2 infected pigs. The lung lesions and the distribution of SIV found in this study were similar to other SIV infection studies [40–44].
Compared with the SIV infected pigs, only limited infection with very few affected areas was seen in pigs infected with the avian H4N6 strain and the infection with AIV was confined to the lower respiratory tract where especially the alveolar type II epithelial cells, but also a few epithelial cells of bronchioles were infected. Low-grade infection after experimental inoculation with AIV in pigs is in agreement with other studies comparing AIV and SIV infections in pigs [45, 46]. The predilection of an AIV for swine alveolar type II alveolar cells has to our knowledge not been described earlier, but alveolar type II cells have been found to be the primary cell type infected in the lungs of fatal human cases of HPAI H5N1 AIV infections [34, 35, 47]. This may indicate a preference of AIV for alveolar type II epithelial cells in both pigs and humans. Our study indicates that there is a predominance of SA-α-2,3-terminal saccharides on the alveolar type II epithelial cells compared to type I cells and this could explain why the avian virus is confined to alveolar type II epithelia cells. This also opens the possibility of using the pig as a model for the study of the pathogenesis of AIV infection in humans.