Illuminating the Importance of Studying Interventions on the Propagation Phase of Experimental Mouse Abdominal Aortic Aneurysms

Abdominal aortic aneurysms (AAAs) are a critical public health problem for which the only effective treatment is surgical intervention by either open surgical repair or endovascular stent placement. While there is a critical need for medical therapies, the development of pharmaceutical approaches is hindered by our incomplete understanding of disease mechanisms.¹ To inform mechanisms, there has been a reliance on data derived from animal models of the disease. Before 2000, the field was bereft of animal models to provide mechanistic insight. In early 2000s, three mouse models of AAAs were developed. These were: 1. Angiotensin II (Ang II) infusion through subcutaneously implanted mini osmotic pumps;² 2. Aortic perfusion of elastase into the infrarenal aorta,³ with subsequent modification of topical infrarenal application;⁴ 3. Periaortic infrarenal application of CaCl₂.⁵ All three models recapitulate at least some facets of the human disease, including lumen dilation and pathological characteristics such as elastic fragmentation, inflammation, and extracellular matrix remodeling. These mouse models have been widely employed to investigate effects of genetic or pharmacological interventions on AAA development. In the vast majority of these studies, an intervention has been applied prior to the stimulus to initiate AAA formation. However, in the clinical setting, an intervention is provided when the disease is already established. Since it is unclear whether the mechanisms of AAA initiation are the same as those propagating the disease, there is a need for studies to determine the efficacy of interventions on the established disease. To provide insight into the efficacy of intervention on established AAAs, in this issue of the journal, Phie et al.⁶ systematically reviewed 4,758 published studies that used the 3 major mouse AAA models. Only 35 of these studies determined effects of an intervention on mice after a stimulus to induce AAAs was started. This study applied standard criteria from rigor and reproducibility guidelines to provide an overall assessment of the effectiveness of drugs reducing AAA expansion.

Of the 35 studies deemed eligible to be included in the analysis, the majority of the studies used AngII-induced AAA model (n=28), with smaller numbers of studies provoking AAAs by elastase (n=6), or calcium chloride (n=3). The initial determination of the study was the risk of bias that provides an index of whether the data supported the stated conclusions. In this regard, most studies provided information on age, sex, strain of mice, and group sizes. However, a significant number of publications did not include precise reports on modality or clear descriptions on whether studies were based on a sample size calculation, or executed using randomization and blinding. The analysis concluded that the risk of bias was low in less than 50% of the studies. This conclusion provides a clear signal that this research field needs to improve data quality.

Following determination of bias, meta-analysis of these 35 studies was performed by groupings based on the pharmacological classification of drugs. The stated conclusions were that anti-inflammatory drugs, protease inhibitors, stem cells, antiplatelet and anticoagulant drugs, and the renin-angiotensin inhibitors led to smaller aortic diameters. Anti-inflammatory drugs also reduced aortic rupture rate. The overall conclusion was that the literature has low quality data for a spectrum of drugs being effective in limiting progression of AAAs.

Several deficiencies in experimental design were highlighted by this meta-analysis. One of the most important is that approximately half of the 35 studies did not provide aortic diameter data prior to administering an intervention. Mouse models of AAAs are associated with variances in incidence and AAA size. Therefore, it is important to measure aortic diameters prior to drug administration by an imaging system in live animals, such as ultrasonography.⁷ Also, it is optimal for a pre-specified size limit to exclude non responders, while mice with measurable AAAs be stratified into vehicle versus drug groups (an example is shown in Figure 1). Unfortunately, most reported studies did not measure aortic dimensions prior to administration of an intervention, which potentially impacted data interpretation.

Figure 1:

An example of experimental design for testing an intervention on established abdominal aortic aneurysms (AAAs). Ultrasound is used to determine maximal luminal diameter of the abdominal aorta. Prior to (baseline) and after AAA induction, maximal luminal diameter is measured to determine whether AAA (50% increase of luminal diameter) is established. Mice with < 50% increase of luminal diameter will be excluded for further intervention. Vehicle and drug groups will be stratified based on luminal diameters after AAA induction.

Among the 35 studies,⁶ 19 initiated the intervention within the first 8 days after start of Ang II infusion or elastase or CaCl₂ application. It is unclear whether persistent luminal dilatation or pathological changes have formed within 8 days because the reported studies did not provide AAA evaluation prior to administering the intervention. Irrespective, this meta-analysis⁶ suggests that early timing of an intervention is the optimal predictor of whether it will be effective in reducing AAA diameters. The finding that earlier interventions have a more pronounced impact on AAAs is consistent with the desire of treating AAAs in human when they are small and asymptomatic. However, it remains challenging to align the time course of disease progression observed in a mouse model to that of the human disease. Without sufficient pathological data of small AAAs in humans, it will be difficult to know whether an early start of the intervention in the current models replicates the major pathophysiology of a small AAA in human.

Given the multiple issues for translating therapies that were efficacious in mouse AAAs into the human disease, the meta-analysis of Phie et al.⁶ highlights the need to focus on elements of experimental design including: 1. Sex, age, and mouse strains. 2. Confirmation that AAAs are established prior to starting a drug; a commonly used technique is ultrasonography. 3. Minimizing the risk of intra- and inter-personal bias by following the ARRIVE guidelines, including prospective inclusion/exclusion criteria, dose justification, sample size calculation to ensure appropriate power, randomization, and blinding of data execution and analysis. Based on the assumption that mechanisms of AAA propagation are different from AAA initiation, it will be important to encourage studies at this phase of the disease, as highlighted by the analyses of Phie at al.⁶