In this study, we developed ways to harvest macrophages from the lungs, blood and bone-marrow of pigs and to freeze them for later use. This approach provides a convenient basis for analysis of the genetic variation in host pathogen interaction using in vitro challenge models. Morphology, viability, TNF production and expression of surface markers (CD14 and CD16) were unaltered by cryopreservation. Our analysis allowed us to compare 5 different breeds and 3 different compartments (bone-marrow, blood and lungs) in terms of their gene expression profiles and responses to LPS.
The basal gene expression pattern in alveolar macrophages (AM) was clearly distinct from the patterns in monocyte-derived or bone marrow-derived macrophages, regardless of breed. The differences include a relatively high basal expression of IDO1, CXCL2, CCL24 and IL1B. Interestingly, we also found that the genes encoding VEGFA and its receptor FLT1 were also highly expressed in AM. Alveolar macrophages do express the receptor for the macrophage growth factor, CSF-1 (CSF1R). However, unlike most tissue macrophages, in the mouse they do not depend upon continued CSF1R signalling . In the op/op CSF-1 deficient mouse, lung macrophage numbers correct with age  and Flt1 signalling has been attributed a role in age-dependent correction of the bone phenotype in op/op mice . We suggest that VEGF might have specific roles in alveolar macrophage homeostasis.
Pattern recognition receptors also distinguished the macrophage populations regardless of breed background. The high levels of lectin-like receptors noted previously  could contribute to elimination of particulate material in the airways, including bacteria and fungi. AM expressed more TLR4 (Figure 5), suggesting that AM would target mostly bacteria detection. In contrast to AM, the BMDM and MDM expressed more TLR3, TLR7, TLR8 and TLR9 as well as RNA intracellular receptors DDX58 (RIG-1) and IFIH1 (MDA5) suggesting a co-regulated cluster of genes involved in virus detection. There is, of course, a fundamental difference between the AM and the two populations derived by cultivation in CSF-1, the MDM and BMDM. The latter cells expressed cell-cycle-related genes and may also be cell cultured adapted. For the purpose of genetic studies, the culture systems have the advantage that they largely eliminate the effect of in vivo environment including health status, and this is reflected in the relatively consistent gene expression profiles. Nevertheless, Fejer et al.  have recently emphasised the fact that the phenotype of alveolar macrophages in mice can be replicated in vitro to some extent by cultivation of bone marrow cells in GM-CSF, as opposed to CSF-1.
We also compared the inflammatory response in 5 different breeds and identified a small set of genes that could contribute to different disease resistance between breeds. Landrace pigs expressed substantially less IL-8R beta (CXCR2) than the other breeds (Duroc, Hampshire and Piétrain). Ait-Ali et al.  reported that Landrace alveolar macrophages were more resistant to PRRS virus replication and released large amounts of TNF and IL8 into the supernatant. It remains to be seen whether differential expression of the IL8 receptor contributes to this biology. The number of genes differentially expressed between the breeds was relatively small (Figure 6). There was much greater variation between individuals within breeds, which also urges caution upon studies of breed differences based upon relatively small group sizes. Amongst the differences was the apparent absence of expression of SLA6 (Figure 8) and highly variable expression of SLA-DOA. These differences might be associated with polymorphic variation in miRNA recognition sites reported elsewhere . Significant levels of protein-coding polymorphism have already been reported amongst pig pattern intracellular and extracellular recognition receptors [19, 32]. It is striking that such genes are DR between the macrophage populations, and highly variable between individuals. It remains to be determined whether such variation is heritable and can be linked to SNP markers to allow selective breeding. The method we have applied herein, which can be performed on blood and does not require pathogen challenge of the animal, could potentially permit in vitro screening of breeding animals for optimal innate immune responsiveness.
In keeping with our earlier findings, now applied to a much larger data set and macrophages from multiple sites including monocyte-derived macrophages, pigs and humans share innate immune responses that are absent from rodents, and conversely, mice induce pathways that are not shared with large animals [5, 16]. The index genes for these classes of genes are IDO1, expressed only in human and pigs, and NOS2A, expressed only in mouse. Using Biolayout Express
3D, we found clusters of genes that share the same expression patterns with these index genes (Figure 3). Analysis of the draft pig genome has highlighted numerous candidate genes underlying human pathology . The findings herein emphasise the applications of understanding pig innate immunity for biomedical research as well as improved livestock production and animal health .