This protocol describes the consecutive implantation of an osmotic pump to induce abdominal aortic aneurysm by angiotensin II release in apolipoprotein E (ApoE) deficient mice and of a vascular access port with a jugular vein catheter for repeated drug treatment. Monitoring of aneurysm development by 3D ultrasound is effectively conducted despite dorsal implants.
Since pharmaceutical treatment options are lacking in the clinical management of abdominal aortic aneurysm (AAA), animal models, in particular mouse models, are applied to advance the understanding of the disease pathogenesis and to identify potential therapeutic targets. Testing novel drug candidates to block AAA growth in these models generally requires repeated drug administration during the time course of the experiment. Here, we describe a compiled protocol for AAA induction, insertion of an intravenous catheter to facilitate prolonged therapy, and serial AAA monitoring by 3D ultrasound. Aneurysms are induced in apolipoprotein E (ApoE) deficient mice by angiotensin II release over 28 days from osmotic mini-pumps implanted subcutaneously into the mouse back. Subsequently, the surgical procedure for external jugular vein catheterization is conducted to allow for daily intravenous drug treatment or repeated blood sampling via a subcutaneous vascular access button. Despite the two dorsal implants, the monitoring of AAA development is readily facilitated by sequential semi-automated 3D ultrasound analysis, which yields comprehensive information on the expansion of aortic diameter and volume and on aneurysm morphology, as illustrated by experimental examples.
An abdominal aortic aneurysm (AAA) is a pathological dilatation of a vessel due to inflammatory and tissue-destructive processes in the aortic wall that may ultimately lead to rupture and patient death. Despite considerable achievements in surgical AAA repair, a conservative drug treatment to block the progression of aneurysm expansion and potentially lower the risk of rupture is missing to date. Animal models have been developed to elucidate triggers and mediators of the disease and to test novel approaches to therapy. Mouse models of AAA are widely applied and cover the different observations from human tissue. Due to their pathomechanistic differences, often more than one model is applied to investigate the particular function of molecules/pathways or the efficacy of potential therapeutic drugs1,2. Among the most commonly used models of AAA induction is angiotensin-II (Ang-II) administration in apolipoprotein E deficient (ApoE KO) mice3, which has more chronic-like pathogenesis in comparison to models that rely on aneurysm formation from an acute insult to the aortic wall4,5. Thus, the Ang-II model seems particularly suited for monitoring disease progression and was recently shown to closely resemble human AAA disease in regard to metabolic and inflammatory responses6. Notably, the Ang-II model features not only AAA development but also thoracic aneurysm formation, as well as aortic dissection with intramural thrombus formation.
Treatments aimed at targeting the progression of already established AAA rather than preventing the initiation of the disease may have higher translational value as patients present with a pre-existing condition that requires treatment7,8. For a comparable experimental design, aortic size needs to be monitored before and after AAA induction to define a threshold of disease development and potentially stratify mice into treatment groups.
The mode of drug administration depends on the uptake and stability of the respective substance. Intraperitoneal (i.p.) injections are most often utilized due to their ease of application, not requiring anesthetic, and the lack of injection volume constraints9. Pharmacokinetics have to be considered, however, when choosing the route of administration, as substances administered i.p. are primarily absorbed through the hepatic portal circulation and may undergo liver metabolism before reaching circulation, which could result in varying plasma concentrations depending on the first pass effect10. Intravenous (i.v.) injection yields the highest bioavailability of substances, and the challenge of repetitive i.v. access can be circumvented by the use of catheters and vascular access ports for daily administration11,12,13. With respect to the AAA setting, drug distribution in circulation facilitates direct aneurysm exposure at defined concentrations.
Here, we describe a workflow for inducing AAA in the Ang-II mouse model via the subcutaneous implantation of an osmotic pump, for daily i.v. drug treatment via a vascular access port connected to a catheter inserted in the external jugular vein, as well as for the monitoring of aneurysm size via 3D ultrasound14 despite the presence of two dorsal implants.
Animal experiments were approved by the local ethics committee and the Austrian Ministry of Science (BMWFW-66.009/0355-WF/V/3b/2016), conforming to the European Directive 2010/63/EU on the protection of animals used for scientific purposes and the Austrian Animal Experiment Act 2012. Humane endpoints were set as follows: loss of ≥15% body weight, avoiding food and/or water intake, reduced activity (hypokinesia) or dyskinesia, or prolonged shaking, scratching, labored respiration, or hunched posture despite pain/symptom management. If necessary, an animal is euthanized under deep anesthesia, i.e., an overdose cocktail of ketamine (approx. 100 mg/kg) and xylazine (approx. 5 mg/kg), or by cervical dislocation. For surgical procedures, aseptic technique and sterile/clean gloves are used throughout.
1. Pump implantation
2. Jugular vein catheterization
NOTE: This surgical procedure requires a microscope with 8x-10x magnification.
3. 3D ultrasound
4. Ultrasound analysis
Representative results show the development and progression of the suprarenal aneurysms as monitored by ultrasound at baseline, day 8, and day 27 (Figure 1A). A trichrome stain (Figure 1B) of the day 27 aorta in Figure 1A further illustrates the morphology of the formed aneurysm with wall dissection and intramural thrombus. Aortic volume (mm3) was determined over a stretch of 12 mm14, and maximum aortic diameter was additionally measured from the EKV images. A threshold of 125% volume growth from baseline to day 8 was set for defining initial aneurysm development. Based on data collected over 2 years (2020-2021, n = 157), only 9% of animals failed to form an AAA according to this cutoff. However, 35% of the mice experienced aortic ruptures (thoracic or abdominal) prior to catheter implantation on day 9, thus resulting in a total of 56% of the remaining animals with established AAA disease amenable to stratification into treatment groups (Figure 1C). Of note, among our historic PBS controls (n = 21), aneurysms developed to varying degrees (range: 128%-314%, mean 199% ± 55% SD aortic volume growth at day 8). Importantly, an inverse relation was observed between the initial expansion and further disease progression, i.e., 55% of fast progressing aneurysms (>200% volume growth at day 8) did not progress further until day 27, while 80% of the other aneurysms (>125% and <200% volume growth at day 8) continued to expand until the end of the experiment (Figure 1D).
As recently reported14,17, the described methods have been successfully established, validated, and implemented, e.g., to document the therapeutic effect of a histone citrullination inhibitor (GSK484, for the inhibition of neutrophil extracellular trap formation) in blocking the progression of established AAA. ApoE deficient mice received Ang-II at 1000 ng/kg/min by subcutaneously implanted osmotic pumps over 28 days. Animals were stratified 1:1 to GSK484 (0.2 µg/g/day) or PBS treatment based on the aortic volume measured on day 8 and underwent the jugular vein catheterization procedure on day 9. Drug injections were conducted daily in a volume of 10 µL/g of mouse weight until the end of the study17. Figure 2 shows exemplary (n = 2/group) ultrasound results (time course of absolute and relative volume or diameter expansion), revealing that GSK484 treatment inhibited AAA progression, while the aneurysms continued to enlarge in control mice.
Figure 1: AAA formation and progression in the Ang-II mouse model as detected by 3D ultrasound. (A) The suprarenal aorta was monitored by 3D ultrasound at baseline (BL), day 8 (d8), and day 27 (d27) after Ang-II pump implantation. Volume was measured over a 12 mm stretch of the suprarenal aorta (157 frames) based on a 3D reconstructed image. The maximum aortic diameter was determined from EKV images. (B) Trichrome stain of a transverse section of the day 27 aorta after mouse sacrifice and organ collection. The presence of an aorta dissection is indicated by L1/L2 (lumen 1 and lumen 2), and the intramural thrombus is denoted by * in A and B. (C) Incidence rate of AAA (>125% aortic volume growth from BL) at day 8 and aortic ruptures within the first 9 days (thoracic or abdominal) from a data set collected over 2 years (n = 157). (D) Progression frequency from day 8 to day 27 of initially fast forming (>200% aortic volume growth from BL to day 8) versus moderately growing (>125% and <200% aortic volume growth from BL to day 8) aneurysms in PBS control-treated mice (n = 21). Please click here to view a larger version of this figure.
Figure 2: Exemplary results from inhibition of histone citrullination to block AAA progression in the Ang-II model by intravenous injection of GSK484 or PBS via vascular access button. (A) Aortic volume (mm3) as measured over a 12 mm stretch of the suprarenal aorta. (B) Calculated aortic volume growth from baseline (BL = 100%). (C) Maximum aortic diameter as determined from EKV images. (D) Calculated aortic diameter growth from BL. GSK484 data were extracted from a previously published study17. Please click here to view a larger version of this figure.
The Ang-II model is one of the most commonly used mouse models of AAA due to its low technical demands and particular features resembling human disease3,6. The surgery time is about 10 min per animal, and the subcutaneous pump implantation is well tolerated by the mice if the subcutaneous pocket is sufficiently wide and placed low on the animal's back, away from the incision site, so as not to interfere with wound healing. When the skin is tight around the pump, tissue irritation may occur, which can cause inflammation and scabbing and potentially disrupt the pump's mechanism of release by osmotic pressure. Measuring the volume of Ang-II remaining in the pump at the time of animal sacrifice gives insight as to whether the Ang-II was successfully released over the 28 days.
The Ang-II model has recently been proposed to be well suited to study aortic aneurysm and dissection progression as it exhibits resemblance with human features of both6. Importantly, testing drug candidates to block aortic expansion and influence remodeling would match the current clinical demand. In our experimental setting, a cutoff for aneurysm formation was defined prior to the treatment start based on 125% volume growth on day 8 in relation to baseline, which accounts for the natural variation in absolute aorta size in mice. The threshold and time point were derived from an initial time course that confirmed aorta wall destruction in histology (data not shown) and resulted in 35% ruptures and 56% observed AAAs prior to catheter implantation. While a minimum threshold of established disease was applied for study inclusion, it was subsequently observed that a high extent of initial aorta expansion may also limit experimental applicability. Aneurysms that progressed rapidly to >200% volume by day 8 did not further grow beyond that size in 55% of cases (Figure 1D). This has to be taken into account during experimental design and sample size calculation, as it could mask a treatment's true effect. Another facet of this model is the frequent aortic ruptures (thoracic or abdominal), occurring at rates of 20%-40% and mostly within the first 10 days after Ang-II pump implantation3,18,19. Thus, by choosing the start of treatment to be day 9, a high rate of established aneurysms was achieved, and the jugular vein catheterization was essentially performed on mice that were expected to survive to the end of the experiment (only 3/24 mice in our historic control group ruptured post day 9), thus conserving time, effort, and cost.
Apart from the aorta ruptures, which constitute a severe condition, the concurrent implantation of the catheter with vascular access button and the osmotic pump was well tolerated by the mice, with no notable effect on mobility or behavior post recovery from surgery. The jugular vein catheterization procedure should take about 30 min for trained researchers. The duration of exposure to (isoflurane) anesthesia should be kept to a minimum, and the animal breathing rate has to be closely monitored to prevent breathing depression, which may lead to a fatal outcome if not resolved20. Blood loss after puncturing the jugular vein for catheter insertion – leading to animal death if major – could potentially occur when the jugular vein is not properly ligated cranially or a side branch feeding into the isolated area of the vessel is not closed off. In that case, pressure with a cotton swab should be applied to the puncture site until blood leakage slows or stops, then the catheter insertion and ligation should be done as quickly as possible; a small piece of the collagen wound dressing may be temporarily utilized to aid with hemostasis.
Catheter patency is one of the most important factors, as catheter disconnection from the vein or the access button results in improper drug delivery where the drug leaks into the subcutaneous space. Following the manufacturer's recommendation of a minimum of 3 mm overlap between the catheter and metal connecter, only one case of catheter disconnection at the button side (indicated by the injected liquid leaking from the incision site at the button) was recorded over 3 years in this model (2020-2021, n = 73), which was fixed by opening the wound and re-establishing the connection in surgery. In addition, a catheter patency failure rate of around 10% in our historic PBS control group (2/21) was experienced due to either catheter occlusion (making it impossible to inject), catheter disconnection from the vein (indicated by apparent swelling in the neck during injection), or wound healing complications. These issues may be connected to self-inflicted injuries, i.e., mouse scratches or bites. Notably, drug treatments that interfere with wound healing may raise failure rates. Troubleshooting steps to improve the patency rate include increasing the length of the catheter inserted in the vein, ensuring ligatures are tightly knotted around the catheter and vein, and applying the positive pressure technique following the manufacturer's recommendation, as described in step 2.12.10., while injecting. Catheter patency should, additionally, be verified at the time of animal sacrifice by dissection and visual inspection under the microscope. Of note, the daily volume of injected drug solution has to be carefully considered. As plasma volume regulates blood pressure, the injection volume may affect AAA expansion, and, hence, control animals need to receive the sham procedure with carrier volume. Based on our experience (and unpublished observations), a daily amount of up to 250 µL of PBS seems to be well tolerated. Finally, similar to the pump implantation, skin irritation can occur around the implanted vascular access button. If inflammation accompanied by devitalized or necrotic tissue is observed, wound debridement should be carried out by removing non-viable tissue (necrotic tissue will often separate naturally from the wound), and the skin should be sutured if needed; if inflammation and necrosis are extensive, the animal's welfare and humane endpoints have to be considered according to guidelines.
Single and dual dorsal implantation of the osmotic pump and/or the VAS did not interfere with the ultrasound signal nor with securing the mouse in an appropriate position on the ultrasound stage. The automated acquisition of 157 frames over 12 mm to render a 3D image of the aorta for volume measurement is a simple and fast procedure14, which only requires ensuring the aorta is clear of interference over the area of interest. One pitfall in this context is applying too much pressure with the transducer while attempting to clear the image of interference, which may interrupt the automated measurement if the breathing rate is affected by the compression of the ribs when images of the cranial end of the abdominal aorta are recorded. Diameter is traditionally measured in images acquired using B-mode by the operator manually searching for the area of maximum diameter while conducting the ultrasound analysis. An advancement on the B-mode images is the EKV images, which can resolve small aortic motions to produce a high-quality, slowed-down image of the pulsating aorta. Furthermore, the maximum aortic diameter can be determined from the acquired 3D frames, where the 157 images offer a comprehensive overview of the aorta taken at systole (due to the set ECG trigger).
In conclusion, the presented compiled protocol provides a reliable and reproducible workflow for i.v. drug administration in a mouse model of Ang-II induced AAA and for monitoring aortic size by 3D ultrasound. The time points of monitoring and operation can be adjusted to the specific needs, and the jugular vein catheterization can be performed separately for any experimental setup requiring delivery of specific substances via i.v. injections. The VAS can alternatively be used for repeated blood sampling if a catheter lock solution is used to prevent clotting. The described 3D ultrasound procedure may be adapted to measure the infrarenal aorta, where aneurysms develop upon acute insult in elastase or CaCl2-based mouse models of AAA. While 3D ultrasound acquisition holds the advantage of giving an overview of the affected aorta region and aneurysm morphology, the image acquisition is more time-consuming and, hence, might be more cost-intensive. Another limitation of the protocol that should be acknowledged is the need for the animals to be anesthetized briefly for intravenous injections, while intraperitoneal administration is generally performed on conscious mice.
The authors have nothing to disclose.
We would like to thank Prof. Podesser's and Prof. Ellmeier's teams (Dept. of Biomedical Research and Core Facility for Laboratory Animal Breeding and Husbandry, Medical University of Vienna) for assistance in the animal experiments. The AAA trichrome staining was kindly performed by Monika Weiss and Prof. Peter Petzelbauer (Dept. of Dermatology, Medical University of Vienna). This work was supported by the Austrian Science Fund [SFB project F 5409-B21]. Marc Bailey is personally supported by the British Heart Foundation [FS/18/12/33270].
4-0 Polysorb sutures | Covidien | GL-46-MG | Braided absorbable suture CV-23 Taper |
6-0 Silk sutures | Ethicon | 639H | PERMA-HAND Silk |
ALZET 2004 osmotic pumps | DURECT Corp | 298 | Osmotic mini pumps |
Angiotensin-II | Bachem | 4006473.0100 | Angiotensin II acetate |
Aquasonic Clear Ultrasound Transmission Gel | Parker Labs | PUSG-0308 | Ultrasound gel |
Betadona Wound Spray | Mundipharma | Wound disinfectant spray (povidone-iodine spray) | |
Betaisodona Solution | Mundipharma | 15973 | Wound disinfectant solution (povidone-iodine solution) |
Catheter for mouse femoral vein/artery | Instech Laboratories Inc | C10PU-MFV1301 | 1 to 3Fr, 10.5 cm, collar @1.2 cm. Fits 22 G |
Hair removal cream | |||
Handling tool | Instech Laboratories Inc | VABMG | Handling tool for magnetic mouse Vascular Access Buttons |
HYLO NIGHT Eye Oinment | URSAPHARM | 538922 | Eye lubricant cream |
Needles and syringes of various sizes | 1 mL and 5 mL syringes, 27 G and 30 G needles | ||
Olympus SZ51 Stereo microscope | Olympus Corporation | Dissection and inspection microscope | |
PinPort injectors | Instech Laboratories Inc | PNP3M-50 | Injector for vascular access button |
Protective aluminum cap | Instech Laboratories Inc | VABM1C | Protective aluminum cap for magnetic 1 channel mouse VAB |
Signa Electrode Ultrasound Gel | Parker Labs | PE-1560 | Electrode gel |
Small electric shaver | |||
Surigcal and microsurgical equipment | |||
Suprasorb C | Lohmann & Rauscher | 20482 | Collagen wound dressing |
Vascular access button (VAB) | Instech Laboratories Inc | VABM1B/22 | Vascular Access Button for mouse, magnetic, 1 channel 22 G, injector |
Vevo 3100 Imaging System | FUJIFILM VisualSonics Inc | 51073-51 | Ultrasound system |
Vevo Lab 5.6.1 software | FUJIFILM VisualSonics Inc | Ultrasound analysis software | |
Vevo MX550D transducer | FUJIFILM VisualSonics Inc | Linear Array Transducer For Vevo 3100 system | |
Vevo Mouse Handling Table | FUJIFILM VisualSonics Inc | 11436 | Mouse heating, mouse core temperature capture and ECG pads for physiological monitoring |