Ketosis Prevents Abdominal Aortic Aneurysm Rupture Through C-C Chemokine Receptor Type 2 Downregulation and Enhanced MMP Balance

Abdominal aortic aneurysms (AAAs) are prevelant with aging, and AAA rupture is associated with high mortality. There is currently no effective medical therapy for AAA rupture. Previous work demonstrated that the monocyte chemoattractant protein (MCP-1) / C-C chemokine receptor type 2 (CCR2) axis critically regulates AAA inflammation, matrix-metalloproteinase (MMP) production, and extracellular matrix (ECM) stability. Here we similarly observed that Ccr2−/− mice have significantly reduced AAA expansion and rupture. We therefore hypothesized that a dietary modulation of the CCR2 axis may therapeutically impact AAA risk of rupture. Since ketone bodies (KBs) can trigger repair mechanisms in response to inflammation, we specifically evaluated whether systemic ketosis in vivo can reduce CCR2 and AAA progression. Male Sprague-Dawley rats underwent surgical AAA formation using porcine pancreatic elastase (PPE), and received daily β-aminopropionitrile (BAPN) to promote AAA rupture. Animals with AAAs received either a standard diet (SD), ketogenic diet (KD), or exogenous KBs (EKB). Animals recieving KD and EKB reached a state of ketosis, and had significant reduction in AAA expansion and incidence of rupture. Ketosis also led to significantly reduced aortic CCR2 content, improved MMP balance, and reduced ECM degradation. In summary, this study demonstrates that ketosis plays a crucial role in AAA pathobiology, and provides the impetus for future clinical studies investigating the potential benefit of ketosis for prevention of AAA expansion and rupture.


Introduction
Abdominal aortic aneurysm (AAA) formation, expansion, and rupture results from a complex series of biomolecular processes 1,2 . AAAs are often asymptomatic during their formation and expansion stages, but lead to a high risk of morbidity and mortality when they spontaneously rupture 3,4 . Unfortunately, there is currently no effective medical therapy to alleviate AAA expansion and the eventual risk of rupture. Invasive surgery is the only available management for AAAs that meet the traditional aortic diameter criteria or are rapidly expanding in diameter 5 . Given the limited medical treatment options for individuals with small AAAs that do not yet meet criteria for surgical repair, expectant management is usually the only remaining option 6 . Therefore, clinical stabilization of small AAAs remains a large unmet need, and a longer-term management strategy for individuals with AAA disease can be a tremendous value add 7 .
In ammation plays an essential role in AAA disease progression 8 . From AAA formation due to wall microdissections, subsequent expansion, and eventual rupture, the release of in ammatory mediators within the aortic wall leads to a cascade of biomolecular signals that lead to the activation of matrix metalloproteinases (MMPs). Activated MMPs consequentially lead to extracellular matrix degradation [9][10][11][12] , and leading to futher AAA expansion. The C-C chemokine receptor type 2 (CCR2) mediates tra cking of leukocytes to site of aortic tissue in ammation following initial and repeated injury 13 . Our team previously demonstrated that CCR2 content in AAAs, as demonstrated by positron emission tomography (PET) / computed tomography (CT) imaging in a preclinical rodent model, highly correlates with the incidence of AAA expansion and rate of rupture 14,15 . However, it remains unknown whether CCR2 content is essential for AAA rupture, and whether modulation of CCR2 can therefore help alleviate disease progression.
Various oral diets are known to impact immune function and in ammation 16 . Ketogenesis in particular can dramatically impact anti-in ammatory signaling, and is reported to promote vascular tissue repair [16][17][18] . As a natural physiologic process that leads to the production of ketone bodies (KBs) such as acetoacetate (AcAc), beta-hydroxybutyrate (βHB) and acetone, ketogenesis not only serves as alternative fuel source for organ systems, but it also activates signaling cascades that can impact various cell functions. Although high fat diets are linked to increased AAA expansion and aortic plaque formation 19,20 , recent studies suggested that a high fat -low carbohydrate ketogenic diet, as well as exogenous ketone body supplementation, can reduce tissue in ammation and ameliorate the risk of vascular injury and atheroprogression 21,22 . It is unknown whether these potential bene ts are limited to atherosclerosis, or whether ketosis can also have a broader impact on degenerative aortopathies such as AAAs. Therefore, we hypothesized that nutritional ketosis, either in the form of a ketogenic diet or exogenous ketone body supplementation, can impact CCR2-mediated in ammation and improve MMP balance in aortic tissue to reduce the risk of AAA progression, aortic wall in ammation, extracellular matrix (ECM) content, and the incidence of AAA rupture.

Animals
Male Sprague-Dawley rats (200-300g) were obtained from Charles River Laboratories (Wilmington, MA; see supplemental methods). Male Ccr2-/and Ccr2+/+ mice on a C57BL/6 background were obtained from The Jackson Laboratory (Bar Harbor, ME; see supplemental methods). All animals were housed at 21°C in a 12/12 hour light/dark cycle and had access to food and water ad libitum. Anesthesia was administered with a mixture of ~ 1.5% iso urane and oxygen for all procedures. The core body temperature was continously monitored and maintained with a heating pad at 37°C. Use of all animal experiments were performed in accordance with relevant guidelines and regulations, were approved by the Institutional Animal Care and Use Committee (IACUC) at Washington University School of Medicine in St. Louis, and reported in accordance with ARRIVE guidelines. At the conclusion of studies, live animals were sacri ced appropriately using anesthetic agents and cervical dislocation.

Induction of AAA
Rats were induced to develop infrarenal AAAs via an established model using porcine pancreatic elastase (PPE; 12 U/mL) as previously described 23 . Ventral abdominal wall laparotomy is performed, and the infrarenal abdominal aorta was exposed. (Supplemental Fig. 1). A customized polyethylene catheter (Braintree Scienti c, Braintree, MA) was introduced through an infrarenal aortotomy, and elastase was infused into the isolated aortic segment for 30 minutes. The exposed aortic segment was dilated to a maximal diameter, and constant pressure was maintained with the use of a syringe pump. Using a video micrometer, the baseline maximum aortic diameter was measured. After 14 days, all rat aortas were reexposed via ventral abdominal laparotomy, maximal aortic diameters were measured, and aortic tissue was harvested for further analysis (Supplemental Fig. 1).
Promoting AAA Rupture As previously described, β aminopropionitrile (BAPN) is reported to promote AAA tissue in ammation by day 6, and AAA rupture between days 7 and 14, but unlikely to cause rupture after day 14 if it has not already occurred 15 . We therefore promoted AAA rupture with daily administration of BAPN on a speci c cohort of rats (RAAA) starting 3 days before PPE exposure. Through drinking water, 300mg BAPN was administered daily (0.3% BAPN in 25mL water consumed per day by a 250g rat) 24 . AAA growth was monitored for 1 week (6-7 days) or 2 weeks (14 days). At the 1 or 2-week timepoints, rats were sacri ced, AAA diameters were evaluated, and aortic tissue was harvested for further analysis. Rats that developed ruptured AAAs during the study period promptly underwent necropsy to con rm and analyze the pathology (Supplemental Fig. 1). Tissue harvested at week 1, prior to rupture, was mainly used to assess AAA tissue in ammation (Supplemental Fig. 2).

Animal Diets
We evaluated four different dietary interventions. First, control groups in the AAA (n = 5) and RAAA (n = 12) cohorts were fed with a standard chow diet (SD). Second, experimental groups in the AAA (n = 6) and RAAA (n = 8) cohorts were fed a very high fat diet with almost no carbohydrate, also known as a classic ketogenic diet one week prior to PPE exposure to induce a 'priming' keto-adapted status 25 , before AAA induction (Supplemental Fig. 3). Ketogenic diet was then maintained prospectively in these groups following AAA formation. Third, experimental groups in the RAAA (n = 9) cohorts were separately started on ketogenic diets as a 'treatment' intervention 3 days after AAA induction. Lastly, experimental RAAA rats were administered a SD along with exogenous ketone body (EKB) supplementation starting 3 days after PPE exposure: RAAA + EKB (n = 10). As previously described 26 , this EKB supplementation was performed with daily intragastric gavage of 1,3-Butanediol (BD; 5g per kg dose; Prod # B84785-100ML, St. Louis, MO) and animals achieved a ketosis state only for 8 hours per day (Supplemental Fig. 3).

Synthesis and Radiolabeling of DOTA-ECL1i
The ECL1i peptide (LGTFLKC) was synthesized from D-form amino acids by CPC Scienti c (Sunnyvale, CA). DOTA-ECL1i was prepared following our previous report. Copper-64 (64Cu, t1/2 = 12.7 hour) was produced by the Washington University Cyclotron Facility. The DOTA-ECL1i conjugate was radiolabeled with 64CuCl2 (64Cu-DOTA-ECL1) as described, and the radiochemical purity was determined by radio- Animal PET/CT Imaging and Image Analysis Dynamic PET scan and corresponding CT images were obtained using Inveon MM PET/CT (Siemens, Malvern, PA) at 45 to 60 minutes after a tail vein injection of 64Cu-DOTA-ECL1i (selective CCR2-targeting radiotracer; 11.1 MBq per rat) to minimize the effect of blood retention on AAA uptake. To localize tracer uptake, a CT contrast agent (1.0 mL, eXIA 160XL, Binitio, Canada) was administrated via tail vein after PET imaging. Contrast-enhanced CT (Bin of 2, 90 mm axial FOV, 60 kV, 500 µA, 500 ms exposure time, 10 ms settle time, no magni cation, pixel size: 80-100 µm) was performed. The AAA uptake was calculated as standardized uptake value (SUV) in 3-dimensional regions of interest from PET images without correction for partial volume effect using Inveon Research Workplace software (Siemens). Dynamic (0-90 minutes) 18F-uorodeoxyglucose (41.1 MBq per rat) PET was also performed in AAA rats at week 1 and 2 post-PPE exposure. Only a speci c number of each group of rats received PET/CT imaging and analysis.

Ultrasound Aortic Assessments
Noninvasive ultrasound (GE, 12 MHz Zonare, Mountain View, CA), was used to evaluate serial aortic maximum diameter measurements. Relative to baseline aortic diameter prior to PPE exposure, the percentage increase in aortic diameter was evaluated at 1-and 2-weeks post-PPE exposure. As previously described, aortic aneurysms were de ned as > 100% increase in the aortic maximum diameter relative to baseline diameter 23,27 . Blood βHB Assessments Animal state of ketosis was evaluated via whole blood D-βHB (Keto-MoJo blood ketone meter; Keto-Mojo, Napa, CA, USA) concentrations 28 . Tail vein puncture was used for blood sample, which was tested on day 0 pre-PPE exposure, and then 1-and 2-weeks following AAA induction. βHB values > 0.5 mmol/L were indicative of ketosis.

Animal Weight
Animal whole body weights were evaluated at day 0 pre-PPE exposure, and 1 and 2 weeks followed AAA induction. Body weight was evaluated by the difference between the values at the baseline (Day 0) and the values at week 1 and 2 respectively and then divided by the baseline to assess difference. All these absolute numbers were then multiplied by 100 to present it as the percentage of difference in weight throughout the time of the study.

Histology and Immunostaining
Aortic tissue was harvested from all animals. AAA tissue was xed in Histochoice (VWR), and para n embedded. Para n blocks were sectioned at 5 µm, and depara nized. Processing for antigen retrieval was performed with Sodium Citrate solution, pH 6.0, for 10 min. Tissue sections were blocked with 10% serum, and sections were incubated with primary antibody anti-CD68, 1:100 [Bio-Rad, MCA341GA]. Sections were then incubated with anti-mouse secondary antibodies conjugated with HRP (Cell Signaling), DAB peroxidase substrate kit (Vector Laboratories), and counter stained with hematoxylin, imaged using an Olympus uorescent microscope system. To evaluate AAA tissue morphology and pathology, tissue sections were also evaluated using Hematoxylin and Eosin (H&E) and Mason Trichrome (MT). AAA wall tissue-stained sections were then analyzed and quanti ed by Image J software via color deconvolution and shown as percentage of stained area for speci c regions of interest (ROI).

MMP Activity Zymography
For each AAA tissue sample, 25ug of protein was loaded on wells of 10% Gelatin Zymogram electrophoresis gels. Gels were then incubated in Zymogram renature buffer for 30 min, followed by 36 hours of Zymogram development buffer at 37°C. Gels were then stained with Coomassie Brilliant Blue R-25 solution from BioRad for 30 min, followed by destaining buffer (20% Methanol, 20% Acetic acid, 60% DI water) until MMP bands were visualized. Gels were scanned on BioRad Chemi doc and analyzed using ImageJ software.

Statistics
All data are presented as the mean ± SD. Most group comparisons were performed using unpaired t test. For comparisons that included one endpoint in more than one animal/diet groups, an ordinary one-way ANOVA with multiple comparison was performed. For comparisons that included more than one endpoint in more than one animal/diet group, we utilized a two-way ANOVA with multiple comparison. Data was considered statistically signi cant with p ≤ 0.05. Kaplan-Meier curve was generated to assess the survival of BAPN-exposed animals. GraphPad Prism 9 (La Jolla, CA) was used for all statistical analyses and graphical data representations. In certain circumstances outlier data points were excluded from the analysis if they met the pre-determinated criteria of the outlier was more than (1.5x Interquartile Range (IQR)) above the third quartile (QR) or below the rst quartile (Q1). MT cross section staining's were analyzed using ImageJ software by color deconvolution, adjust threshold and region of interest assessment of the AAA wall.

CCR2 is Essential for AAA Formation and Rupture
To evaluate whether CCR2 plays an essential role in AAA pathology, we evaluated whether AAA progression and/or rupture were impacted in mice that underwent whole body gentic knockdown of Ccr2. Age-matched male wildtype (Ccr2+/+) and Ccr2-/adult mice received angiotensin II osmotic pump administration to promote AAA formation 14,29 , and daily BAPN administration to promote AAA rupture 30 (Supplemental and 47% higher survival compared to Ccr2+/+ (Supplemental Fig. 4D). PET/CT with 64 Cu-DOTA-ECL1i (selective CCR2-targeting PET radiotracer) demonstrated signi cantly reduced CCR2 content in the aorta of Ccr2-/mice and no AAAs were observed (p < 0.001; Supplemental Fig. 4E & F). These data con rmed that CCR2 is essential for AAA formation and rupture, and led us to next evaluate whether an easy to implement dietary intervention can reduce CCR2 and ameliorate AAA pathology.

Ketosis Attenuates AAA Formation and Content of MMP9 in Aortic Tissue
To evaluate the impact of a ketogenic diet on AAA pathology, we rst compared to animals fed a standard diet (SD), to animals maintained on a ketogenic diet prior to AAA induction (KDp; Fig. 1A). KDp achieved a state of sustained ketosis from day 0-14 ( Fig. 1B), and caused a moderate decrease in weight gain by week 1 and 2 (p < 0.001; Fig. 1C). By week 2 there was a substantial 42% decrease in AAA diameter in KDp animals (p = 0.008; Fig. 1D and Supplemental Fig. 5A). Aortic wall media demonstrated equivalent masson trichrome (MT)-stained collagen between animals maintained on SD and KDp (Fig. 1E-G). Zymography analysis of harvested aortic tissue at week 2 demonstrated a modest decrease in pro and total-MMP9 in KDp animals ( Fig. 1H-J). These data suggested that diet-induced ketosis can inhibit AAA expansion, and that this may in part be due to a decrease in aortic wall total and pro-MMP9.

Sustained Ketosis Reduces AAA Expansion and CCR2 Content in Rupture-Prone Animals
To evaluate whether in a AAA rupture model dietary ketosis can impact CCR2 content and AAA pathology, we then evaluated a separate cohort of animals that received either SD or KDp after AAA induction with PPE, and daily BAPN administration to promote AAA rupture ( Fig. 2A). By day 14, animals that survived Similarly, total MMP 2, known to promote AAA expansion 36 , was reduced in KDp animals (p < 0.001; Fig. 3C & D). Content of MMP9 and Tissue Inhibitor of Metalloproteinases 1 complex (MMP9/TIMP1; known to prevent MMP9 over-activation) was also signi cantly increased in KDp animals (p = 0.008; Fig. 3E & F). Correspondingly, AAA tissue in KDp animals demonstrated equivalent levels of total MMP-9 ( Fig. 3G), and signi cantly reduced TIMP1 compared to SD animals (p = 0.03; Fig. 3H). Overall, these data demonstrate that sustained ketosis with KDp decreases active MMP9 while increasing MMP9/TIMP1 stabilizing complex in AAA tissue. Finally, we also observed a signi cant positive correlation between active MMP9 and CCR2 content in the AAA tissue in both SD and KDp animals (p = 0.03 and p = 0.39 respectively; Fig. 3I & 3J). These ndings suggest that CCR2 content in AAA tissue may be responsible for activating MMPs, and therefore resulting in a higher incidence of AAA rupture.

Impact of Ketosis That is Initiated 'Therapeutically' After AAA Formation
Animals treated with an abbreviated course of KD, therapeutically initiated 3 days post-AAA formation with PPE (KDt; Fig. 4A), also led to a state of ketosis (Fig. 4B). Animals treated with supplemental exogenous ketone bodies by oral daily gavage (EKB; Supplemental Fig. 2 & Fig. 4A) also led to ketosis, but only for 8hour per day (Fig. 4B). Similar to KDp animals, KDt and EKB animals also had reduced weight gain at both week 1 and 2 when compared to SD animals (p < 0.001; Fig. 4C & Supplemental Fig. 2). Although, AAA rupture rate was reduced in KDt and EKB animals compared to SD animals (22% reduction with KDt, p = 0.03, and 40% reduction EKB, p = 0.12: Fig. 4D & E), the relative decrease in rupture was not as much as KDp animals (Fig. 2E & F). AAA absolute diameter and percentage of aortic diameter increase were also signi cantly reduced at both week 1 and 2 in EKB animals while only signi cantly reduced at week 2 in KDt animals ( Fig. 4F & Supplemental Fig. 5C). These ndings demonstrate that KDt and EKB therapeutic regimens lead to reduced AAA expansion and risk of rupture.
KDt and EKB animals also demonstrated increased AAA wall media Collagen content (p = 0.08 and p = 0.02, respectively; Fig. 4G-J), and reduced CCR2 immunostaining (p = 0.06 and p < 0.05, respectively; Fig. 4K-N). No difference was observed in CD68 immunostaining across groups (Fig. 4O). Equivalent levels of pro-MMP9 were observed among both treatment groups (Fig. 4P & S), but active MMP9 was signi cantly decreased in KDt and EKB animals (p = 0.02 and p = 0.001, respectively; Fig. 4Q & S). Total MMP2 was also notably attenuated in KDt animals (p < 0.001), but not in EKB animals (Fig. 4R & S). These data suggest that even an abbreviated therapeutic course of ketosis following AAA formation can help stabilize AAAs, preserve aortic wall collagen content, reduce CCR2 tissue content, and promote MMP balance.

Discussion
To our knowledge, our study is the rst to demonstrate that Ccr2 is essential for the incidence of AAA rupture, and that diet-induced ketosis can also signi cantly decrease AAA progression and the risk of rupture. Using previously validated, pre-clinical murine and rat models for AAA 14,15 , and different ketogenic supplementation strategies, we provide a robust and comprehensive assessment of the impact of dietary ketosis on AAA formation and the risk of rupture. We also speci cally demonstrate that administration of either a ketogenic diet (KDp or KDt) or an oral ketone body supplementation (EKB) can reliably induce systemic ketosis, signi cantly reduced aortic wall CCR2 and pro-in ammatory cytokines, increase collagen content in the AAA media, and promote an MMP balance that minimizes elastin degredation (Fig. 5).
Animals that received KDp demonstrated the most notable decrease in AAA expansion and risk of rupture.
Animals that received KDt and EKB supplements also demonstrated differences in AAA progression, but not to the same extent. There was also mild to moderate variability in the KDt and EKB values of CCR2, CD68, and MMP content in AAA tissue. Administration of BAPN was reliable in inducing AAA rupture and did not appear to confound the impact of ketosis on AAA expansion and risk of rupture. Additionally, our complementary studies demonstrated that ketosis can impact pro-in ammatory CCR2-mediated signaling mechanisms that can lead to AAA progression. Therefore, this pre-clinical study demonstrates that a lowrisk, and relatively easy dietary intervention, can potentially alter the course of AAA disease progression, and provides important insights that can be easily translated to human patients with AAAs who lack an effective medical management strategy.
Endogenous ketone body production mainly occurs in the liver, and results in a high glucagon/insulin ratio leading to an increased serum free fatty acids production in the circulation 37 . This naturally occurs during periods of fasting, where βHB is released into the bloodstream as a byproduct of enzymatic degradation of ketone bodies within the mitochondrial matrix and is converted into ATP through oxidative phosphorylation 38 . βHB rises to a few hundred micromolar (µM) concentrations within 12-16 hours of fasting, 1-2 mM after 2 days of fasting 39 , and 6-8 mM with prolonged starvation 40 . Ketogenic diets modify a host's systemic energy metabolism to mimic the biochemical impact of starvation by signi cantly increasing serum βHB levels, lowering blood glucose, and increasing fatty acid concentrations 41 . These regimens were originally introduced as a treatment for refractory epilepsy in children and have now become popular for weight loss programs, patients with diabetes, obesity, various types of cancer, and among high performance athletes [42][43][44][45][46] . Standard ketogenic diets that are devoid of carbohydrates can lead to elevated βHB serum levels that are consistently > 2 mM 46 . Recent studies demonstrate that βHB can serve as an important signaling mediator that can inhibit histone deacetylases 47 , blunt tissue oxidative stress 48,49 , active G-protein-coupled receptors 50,51 , and regulate in ammatory mediators such as prostaglandin D2 52 , interleukins 53 , nuclear factor kappa B (NF-κB) 54 , and NLRP3 in ammasome 55 . Similarly, our study shows that animals with high serum βHB have blunted tissue in ammation and CCR2 content, which in part likely contributes to reduced pathological AAA expansion and risk of rupture.
Uniquely, our study administered three different ketosis regimens: two types of ketogenic diets (KDp and KDt), and an oral supplement regimen (EKB). KDp included a 1-week priming period prior to AAA formation, that imitates the phenomenon of keto-adaptation that occurs in humans who are maintained longer-term on a ketogenic diet 56 . This regimen aided in determining whether a ketosis primer can have a 'protective' impact against AAA formation and expansion. On the other hand, KDt was designed to evaluate the potential 'therapeutic' impact of ketosis on expansion and rupture of AAA post-induction with PPE. This regimen would hypothetically be similar to how medical management would be prescribed in humans with small AAAs that do not yet meet the traditional size criteria for operative intervention. In the course of this study, we observed that animals tolerated both KDp and KDt, and that both were successful in inducing a sustained systemic state of ketosis. Interestingly, both regimens yielded signi cant reductions in AAA expansion and incidence of rupture relative to animals that received SD. However, the longer-term KDp regimen appeared to have a more protective impact, and a more impressive reduction of CCR2 content in AAA tissue. These ndings suggest that the length of diet-induced ketosis may be an important variable in the extent of reduction of AAA tissue in ammation and risk of rupture.
With the recent advent of EKB supplements, oral regimens have been increasingly utilized to manipulate levels of circulating blood ketone body concentrations in humans for various health bene ts 57 . While most studies involving EKB supplementation have traditionally focused on its impact among high-performance athletes 58 , these supplements are increasingly being studied as remedies for conditions such as epilepsy, heart failure, diabetes, and sepsis-related muscle atrophy 59 . Our study evaluated the use of EKB to induce ketosis in animals with AAAs that are prone to rupture. Interestingly, we observed that EKB not only decreased AAA tissue in ammation (Supplemental Fig. 10), but also reduced AAA expansion and incidence of rupture (Fig. 4). The impact of EKB on CCR2 content and AAA rupture was variable from KDp and KDt, and we suspect this is because EKB only induced intermittent ketosis (limited to 8 hours per day).
Nonetheless, these ndings are the rst to show that oral supplementation with ketone bodies can indeed serve as a minimally invasive method for the potential medical management of AAAs, and is a compelling topic for further exploration in future human clinical trials that completement prior efforts 60-62 .
Our study results also suggest that ketosis has a multifaceted impact on aortic wall structure and function.
In ammation is the major molecular mediator of AAA disease progression (Fig. 5). Previous studies demonstrated that excessive aortic wall in ammation can inhibit reparative signaling, wall brosis, and collagen deposition, which can in turn accelerate AAA expansion and lead to a higher risk of rupture 63 .
Tissue macrophages are known to promote AAA disease, in particular subsets that highly express CCR2 12 .
We as well as others, also previously demonstrated that genetic or molecular targeting of CCR2 can reduce AAA progression [13][14][15] . Here we provide further compleing evidence that CCR2 content indeed correlates with AAA disease progression, and that systemic ketosis in vivo can signi cantly reduce its both CCR2 content as well as downstream pro-in ammatory cytokines in AAA tissue.
Previous studies investigating the in ammasome in AAA tissue, demonstrated that TNFα and RANTES are both up-regulated in expanding AAA wall tissue 64,65 . Inhibition of TNFα appears to decrease aortic wall MMP activity, reduce ECM disruption, and decrease aortic diameter in a murine pre-clinical AAA model 66 . In another study, Empagli ozin, a sodium-glucose cotransporter 2 inhibitor that increases plasma ketone bodies 67,68 , was found to reduce aortic aneurysm diameter and aortic wall RANTES in Apo E -/-mice 69 .
Similarly, in our study we observed that diet-induced ketosis can signi cantly decrease aortic wall proin ammatory cytokines TNFα and RANTES, as well as increase aortic wall Collagen content. Although the direct mechanism of action for this is yet to be fully elucidated, we suspect that the molecular interplay between macrophage and other pro-in ammatory cell types may be playing a critical role in the immune modulation of these processes and AAA progression 70,71 .
A central pathological feature of AAA disease progression is excessive and aberrant extracellular matrix (ECM) remodeling. This results from increased MMP activity, which promotes rapid ECM breakdown and disruption of the integrity of the aortic wall 72,73 . Previous work demonstrates that MMP2 plays a central role in the formation and early expansion of AAAs, while MMP9 is more related to late AAA expansion and risk of aneurysm rupture 30,74,75 . Synergistic activation of both MMP2 and MMP9 provides an unfavorable environment that can accelerate AAA dilation and lead to a higher risk of aneurysm rupture 76 . Previous studies also demonstrate that ketosis, high serum βHB, and signaling via NF-Kβ, play key roles in suppressing MMP-9 expression in colonic tissue 77 . Our studies extend on this molecular mechanism of action, and demonstrate that ketosis and elevated serum βHB can also signi cantly attenuate both active MMP9, and total MMP2 in aortic tissue. In fact, a CCR2 antagonist has shown to downregulate MMP-9 expression in lung cancer cells, therefore mitigating cellular motility and metastatic invasion 78 . These results may help explain why we observed a notable decrease in MMP-9 content in AAA tissue from animals with ketosis.
TIMPs are endogenous speci c inhibitors of MMPs produced by vascular smooth muscle cells (VSMCs) as well as other cell types in AAA tissue 79 , which inhibit zymogenesis of pro-MMPs and reduces overall MMP activation. Given their central role in maintaining the dynamic balance in ECM turnover in aortic wall tissue, the role of TIMPs in AAA progression continues to be an area of intense investigation 35 . Our study demonstrates that while nutritional ketosis decreases the content of free TIMP1, it signi cantly increases the content of the stabilizing TIMP1/MMP9 complex in AAA tissue. This data suggests that complexed TIMP1 leads to a reduction in active MMP9 content, therefore decreasing AAA wall ECM degradation, further aneurysm expansion, and the overall risk of rupture (Fig. 5).
Our study also demonstrated a mild-moderate, but non-signi cant, increase in AAA tissue TGFβ content in animals treated with ketogenic diets (Fig. 5). TGFβ belongs to a superfamily of growth factors that regulate many cellular functions such as cell growth, adhesion, migration, differentiation, and apoptosis 80 .
TGFβ content appears to be signi cantly reduced in human AAA tissue 81 . A recent study demonstrated that ketosis promoted TGFβ-induced myocardial brosis and Collagen 1 and 3 deposition in spontaneously hypertensive rats 82 , suggesting that TGFβ up-regulation was deleterious in this setting.
However, in aortic tissue, TGFβ appears to have a bene cial role. For example, administration of TGFβ neutralizing antibodies appeared to promote excessive monocyte-macrophage in ltration within murine and rat AAA tissue 34,83 , while overexpression or administration of TGFβ1 signi cantly increased aortic wall collagen deposition 84 , and collagen synthesis in normal arteries 85 . This in part explains our observation that animals receiving a ketogenic diet had signi cantly increase aortic wall Collagen 1 deposition, which correlated with higher aortic tissue TGFβ content.
We acknowledge that there are some limitations in our study. First, all our data is derived from pre-clinical rodent models that are not necessarily representative of human AAA pathophysiology. However, the rat AAA rupture model was previously validated and shown to be the most reliable and consistent AAA rupture model currently available. Second, our studies did not systematically evaluate arterial blood pressure. This would have required sophisticated in dwelling sensors and the use of continuous telemetry. While such monitoring systems are feasible for shorter experimental protocols, our 2-3-week experimental protocol would have greatly complicated the experimental design and led to several confounding variables. We therefore elected to instead serially monitor AAA endpoints via ultrasound, which provided reliable and reproducible data. Third, our study used a single composition for the ketogenic diet intervention. We acknowledge that this is not fully representative of the wide variety of lipid and oil-based ketogenic diets consumed by humans, but this was selected to maintain consistency and adherence within all rodent study groups.
In conclusion, this study demonstrates that a ketogenic diet and EKB supplementation strategy that can signi cantly reduce AAA expansion and reduce the incidence of AAA rupture. Importantly, a ketogenic priming period appears to also be further protective, while EKB appears to be less effective than other dietary regimens. Ketogenic diets reduced CCR2 content, promoted MMP balance, and attenuated ECM degradation in AAA tissue. These ndings provide the impetus for future pre-clinical and clinical studies geared to determine the role of ketosis as a medical management tool for human patients with AAAs that do not yet meet the criteria for surgical intervention.

Disclosures
The authors declare that they have no competing interests.

Supplemental Methods
Supplemental Figures 1-10 Data Availability The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. and 3±0.5) respectively (p < 0.01). (C) Percent body weight difference in SD vs KDp animals at week 1 (13±5 vs 2±1.3) and at week 2 (23±5 vs 8±2) respectively (p < 0.001). (D) Percent aortic diameter in SD vs KDp animals at week 1 (154±48 vs 137±42; p = ns) and at week 2 (332±129 vs 140±152; p = 0.008) respectively (aneurysms were de ned by a >100% increase in the aortic diameter compared with pretreatment measurements). (E) AAA collagen staining quanti cation for SD and KDp at week 2 (33±4 vs 34±2; p = ns) respectively. (F and G) Trichrome staining of abdominal aortas (cross-section of tissue slides) with 5x magni cation for SD and KDp animals. Areas with blue staining signify areas with higher collagen deposition. (H) Zymogram demonstrating pro and active MMP9 levels were measured by integrated optical density (IOD).. (I) Pro MMP-9 levels for SD and KDp at week 2 (4.8±3x10 3 vs 2±0.9x10 3 ; p = ns) respectively. (J) Total MMP-9 levels for SD and KDp at week 2 (4.8±3x10 3 vs 2.4±0.9x10 3 ; p = ns) respectively. Data presented as mean ± standard deviation. ns > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001 using Student's t test. No outliers were observed in the analyses, and all data was included in the gure.    One outlier data point in the SD group was excluded based on pre-de ned criteria prior to analysis (see methods). (P) α-SMA protein content expressed as a ratio to GAPDH content in AAA tissue of SD and KDp animals (3.4±1 vs 3.8±0.7; p = ns). (Q) Representative Western blots of collagen 1, α-SMA, TGFβ-1, GAPDH and Caveolin 1 in AAA tissue. Data presented as mean ± standard deviation. ns > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001 using Student's t test. Impact of therapeutic ketosis on AAA risk of rupture. (A) Animals underwent exposure to PPE to develop AAAs and were also treated with BAPN to promote AAA rupture. Following AAA induction, animals received a ketogenic 'treatment' via an oral diet (KDt) or exogenous supplement (EKB). (B) Ketosis (βHB whole blood levels > 0.5 mM/L) in SD, KDt, and EKB animals at week 1 (0.2±0.1, 1.8±0.9 and 1±0.02, respectively; p < 0.05) and at week 2 (0.2±0.1, 2.7±1.1, and 1±0.02 respectively; p < 0.01) analyzed using two-way ANOVA. Ketosis impacts AAA expansion and risk of rupture. AAA expansion and risk rupture is in uenced by CCR2, which in turn recruits CD68+ pro-in ammatory macrophages, and also leads to cytokine release, and MMP activation. Vascular smooth muscle cells (VSMCs) production of TIMP1 can complex with MMP9 to help balance out the rate of MMP-medicated ECM degradation. Decreased TIMP1/MMP9 complex can lead to higher ECM degradation and AAA expansion. CCR2-mediated release of TNFa, RANTES, IL-10, IL-17A, and