Mouse models of human diseases are potentially important tools with which to investigate mechanisms involved in the pathology and identify new treatments. Currently a number of animal models of human AAA are available [4
]. Many of these models require significant interventions such as exposing the aorta and subsequently infusing or painting elastolytic solution on a segment of the artery to induce weakening and inflammation. Currently the angiotensin II infusion model is most commonly used, possibly due to the relatively ease with which aneurysms are produced and the lack of requirement for what might be considered artificial manipulation of the aorta. To our knowledge this is the first study to examine whole genome expression in relation to aneurysm formation within this model. Our findings demonstrate that aneurysms forming in ApoE-/-
mice have marked influx by a range of inflammatory cells and upregulation of cytokines which have all been previously demonstrated in human AAA [8
]. These findings confirm those from more selective previous studies within this model [5
]. Importantly they suggest the value of this model for investigating the inflammatory and cytokine aspects of human AAA. We also demonstrated the upregulation of a range of chemokines, cytokines and proteolytic enzymes, such as Ccl4, Ccl8, Il6 and Mmp2 which have been previously implicated in human AAA [9
]. The upregulation of the latter genes in AAAs was validated using real time PCR. We identified genes up or downregulated in the aortas of mice resistant to aneurysm formation suggesting potentially protective and pathological roles of these genes respectively in AAA. Sclerostin expression is decreased within human AAA biopsies compared to controls, however the significance of this finding is unclear [9
]. We found increased expression of the sclerostin gene (Sost) in the aortas of mice resistant to aneurysm formation compared to both mice with aneurysms and saline controls, suggesting that sclerostin may play a role in inhibiting aortic dilatation. The ability of sclerostin to antagonise transforming growth factor beta, which has been linked to aortic aneurysm development, could be of significance in an aortic protection role for this gene [25
]. Acta2 (vascular smooth muscle actin) expression was increased in aortas resistant to aneurysm formation compared to other groups. Acta2 encodes the single most abundant protein in vascular smooth muscle cells, and mutations in the human gene are associated with ascending thoracic aortic aneurysms and dissections [26
]. Increased expression of Acta2 and other genes encoding smooth muscle cell proteins (Cald1, Dstn), in aortas protected from aneurysm, highlight the importance of vascular smooth muscle cells in maintaining vascular integrity. We observed downregulation of the genes encoding kininogen (kng1), Apolipoprotein CI (Apoc1) and the neutrophil-associated leucine-rich alpha-2-glycoprotein 1 (Lrg1) amongst others, in aortas of mice resistant to aneurysm suggesting that they may be pathological genes. The role of these in aneurysm formation warrants further investigation.
We also investigated genes that might underlie the predilection for suprarenal aneurysm formation within this model, highlighting downregulation of a number of genes which may explain the susceptibility of the suprarenal aorta to aneurysm formation. Urocortin has been shown in vitro
and within animal models to inhibit the effects of the renin-angiotensin system, thus downregulation of Ucn3 (urocortin 3) within the suprarenal aorta of ApoE-/- mice may be relevant to its predilection for aneurysm formation following angiotensin II infusion [11
]. We also demonstrated downregulation of Lama3 (laminin alpha 3) within the suprarenal aorta. This extracellular matrix protein has been demonstrated to be present in reduced concentrations within some human AAAs and therefore may also be relevant to the preponderance of the suprarenal aorta to aneurysm formation [27
]. In vitro
laminin also plays important roles in modulating inflammation and MMP production suggesting its relevance to aneurysm formation [29
]. Further studies will be required to investigate these candidate genes for example employing deficient mice.
A small number of previous studies have used expression arrays to examine gene transcripts associated with AAA in rodent models or patient samples [9
] (Table ). There are a number of issues to consider when interpreting the results of the present study and those previously conducted, such as sample sizes, controls used, model employed (in the case of animal studies) and adjustment for multiple testing. Only one of the previous studies was a whole genome expression assessment, similar to the current study, with other studies examining much smaller numbers of transcripts. The sample sizes employed in all studies means that the power to detect differentially expressed genes is limited; therefore all current studies would have been expected to generate a large number of false negatives. We have therefore concentrated on identifying genes which were identified in the current study and also in these previous investigations and highlighted these throughout our tables (a summary of our findings in comparison with previous studies is provided in table ). A number of chemokines were identified in all studies underlining the findings of the current investigation. Other common mechanisms identified to be associated with AAA included pro-inflammatory cytokines and matrix proteolysis. The design of the current experiment has some advantages over previous studies in the use of controls employed. We utilised two control sample types within study 2, i.e. aortas from mice receiving saline infusion but importantly also from mice that received angiotensin II and did not develop aneurysms. This has enabled us to try and isolate effects associated with AAA rather than just those due to angiotensin II. Findings in previous rodent studies may have related in part simply due to the effects of elastase [32
]. Controls for human expression array studies are problematic for a number of reasons, including the large amount of patient to patient variations, and the effects of factors that are impossible to control for in small samples such as differences in medication taken and co-morbidities [9
Comparison of results from this study and those from other expression arrays for AAA.
This study is the first to carry out a whole genome expression analysis within the angiotensin II- Apo E-/-
model. We used an explorative approach using relatively small numbers of arrays (28 in total were included in the array studies). We were not adequately powered to adjust for multiple testing given the number of transcripts examined. Unlike some other microarray studies we have used real time PCR, IHC and ELISA to confirm some of our findings rather than relying on RNA assessment alone [9
]. Numerous studies have now demonstrated the validity of microarray platforms therefore we felt it would be useful to demonstrate the functional relevance of altered RNA expression [9
]. Our findings suggest the value of this mouse model as one to investigate the role of inflammation, cell recruitment and proteolysis in AAA.