Inhibition of VEGF-A or its receptor activity suppresses experimental aneurysm progression in the aortic elastase infusion model

Abstract

Objective:

We examined the pathogenic significance of VEGF-A in experimental abdominal aortic aneurysms (AAAs) and the translational value of pharmacologic VEGF-A or its receptor inhibition in aneurysm suppression.

Approaches and Results:

AAAs were created in male C57BL/6J mice via intra-aortic elastase infusion. Soluble VEGFR-2 receptor extracellular ligand binding domain (delivered in Ad-VEGFR-2), anti-VEGF-A mAb, and sunitinib were used to sequester VEGF-A, neutralize VEGF-A and inhibit receptor tyrosine kinase activity, respectively. Influences on AAAs were assessed using ultrasonography and histopathology. In vitro transwell migration and quantitative RT-PCR assays were used to assess myeloid cell chemotaxis and mRNA expression, respectively.

Abundant VEGF-A mRNA and VEGF-A-positive cells were present in aneurysmal aortae. Sequestration of VEGF-A by Ad-VEGFR-2 prevented AAA formation, with attenuation of medial elastolysis and smooth muscle depletion, mural angiogenesis and monocyte/macrophage infiltration. Treatment with anti-VEGF-A mAb prevented AAA formation without affecting further progression of established AAAs. Sunitinib therapy substantially mitigated both AAA formation and further progression of established AAAs, attenuated aneurysmal aortic MMP2 and MMP9 protein expression, inhibited inflammatory monocyte and neutrophil chemotaxis to VEGF-A, and reduced MMP2, 9 and VEGF-A mRNA expression in macrophages and smooth muscle cells in vitro. Additionally, sunitinib treatment reduced circulating monocytes in aneurysmal mice.

Conclusions:

VEGF-A and its receptors contribute to experimental AAA formation by suppressing mural angiogenesis, MMP and VEGF-A production, myeloid cell chemotaxis and circulating monocytes. Pharmacological inhibition of receptor tyrosine kinases by sunitinib or related compounds may provide novel opportunities for clinical aneurysm suppression.

Graphical Abstract

Introduction

Abdominal aortic aneurysm (AAA) is a common and lethal disease of the infrarenal aorta. Although increasingly identified by ultrasound screening programs worldwide, no pharmacological therapies have proven effective in limiting AAA progression or death from aneurysm rupture¹.

Aortic mural neovessel formation, also known as angiogenesis, is a pathological hallmark of experimental and clinical AAAs^2-4. In human aneurysmal aortae, neovessels are more prevalent in aneurysmal segments adjacent to areas of rupture, and localize within areas infiltrated by leukocytes^5-8. Drugs and hemodynamic conditions that limit experimental AAA progression also attenuate mural angiogenesis^{4, 9-18}. The density of aortic endothelial cell progenitors, critical for angiogenesis, is positively correlated with aneurysm enlargement¹⁸.

Vascular endothelial growth factor (VEGF)-A has a central role in promoting angiogenesis¹⁹. VEGF-A contributes to the proliferation, survival and migration of endothelial cells and their progenitors²⁰. VEGF-A also mediates the migration of circulating inflammatory monocytes, tissue macrophage precursors, to target organs. Elevated VEGF-A levels and increased VEGF-A-producing cells have been consistently documented in experimental and clinical AAAs^{7, 21-25}.

Administration of exogenous VEGF-A was previously found to augment AAA formation in the angiotensin II-infused hyperlipidemic mouse model²⁶. In an additional mouse AAA model created by abluminal application of calcium chloride to the infrarenal aorta, treatment with VEGFR-1 fusion protein, an investigational agent for sequestering VEGF-A, suppressed AAA formation²². Although no single experimental model recapitulates all the pathologic features of clinical AAA disease, these models both entail significant limitations²⁷. For example, medial elastic lamellae and smooth muscle cell (SMC) populations are largely preserved in both models, except for non-circumferential foci of elastin disruption present in the angiotensin II model. And to date, no clinically relevant angiogenesis inhibitors have been tested in AAA suppression experiments, regardless of modeling platforms employed.

In this study, the significance of VEGF-A activity in experimental AAA pathogenesis was examined by VEGF-A sequestration, as well as treatment with an anti-VEGF-A mAb antibody or the receptor tyrosine kinase (RTK) inhibitor sunitinib, in the intra-aortic porcine pancreatic elastase (PPE) infusion model.

Materials and Methods

The authors declare that all supporting data are available within the article and its online supplementary materials.

AAA creation

Intra-aortic PPE infusion was used to create AAAs in 10-12 weeks old male C57BL/6J mice (The Jackson Laboratory, Bar Harbor, Maine) as previously described^{11, 12, 14}. Sex is a significant determinant of AAA risk (approximately 4-fold greater prevalence in males). Since loss of function approaches were exclusively employed in these experiments, male mice were used to maximize the ability to discern a difference between treatment and control. And single sex mice were exclusively employed to minimize variability in outcome within and between groups. In order to maximize the generalizability of these results to the human condition, however, further experiments will need to include sufficient numbers of mice of both sexes²⁸.

Briefly, at laparotomy, the infrarenal aorta was isolated from the level of the left renal vein to the iliac bifurcation and temporarily controlled with 6-0 silk suture. Heat-tapered P-10 tubing was inserted into the controlled segment, and 30 μl of type I PPE (1.5 units/ml freshly prepared in phosphate-buffered saline (PBS) (please see Major Resources Table in the Supplemental Material) was infused for 5 min under constant pressure via an infusion syringe pump. Following infusion, the residual PPE solution and tubing were withdrawn and aortotomy closed with 10-0 nylon suture.

Anesthesia was maintained via 2% inhaled isoflurane. All procedures were performed under sterile conditions with surgical magnification. Following the procedure, mice were covered and housed in separate cages with free access to chow and water. All experimental procedures and animal care were conducted in compliance with the Stanford Laboratory Animal Care and Biosafety Guidelines and approved by the University Administrative Panels on Laboratory Animal Care (protocol #11131) and Biosafety (protocol #1118).

Analysis of CCL2, MMP and VEGF-A mRNA expression

Real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) was used for mRNA quantification. cDNA was synthesized from total RNA isolated from aortae or cultured cells and amplified using SYBR^® GreenER qPCR SuperMix Univerisal on the Stratagene^® MX300P system (Aligent Technologies, Inc, Santa Clara, CA). PCR primers were TAA AAA CCT GGA TCG GAA CCA AA and GCA TTA GCT TCA CAT TTA CGG GT for CCL2; GAT GTC GCC CCT AAA ACA GAC and CAG CCA TAG AAA GTG TTC AGG T for MMP2; GGA CCC GAA GCG GAC ATT G and GAA GGG ATA CCC GTC TCC GT for MMP9; GCA CAT AGG AGA GAT GAG CTT CC and CTC CGC TCT GAA CAA GGC T for VEGF-A; and TAT TGG CAA CGA GCG GTT CC and GGC ATA GAG GTC TTT ACG GAT GT for β actin.

Messenger RNA levels are presented as fold changes relative to aorta from mice without PPE or PBS infusion or cells without vehicle or sunitinib treatment. All reagents purchased from, or synthesized at, Life Technologies, Grand Island, NY.

Treatment with adenoviruses, anti-VEGF-A monoclonal antibody or sunitinib

An adenovirus encoding mouse VEGFR-2 (flk1) extracellular ligand binding domain fused to mouse IgG2a Fc (Ad-VEGFR-2 (flk1)) and a control adenovirus encoding a mouse IgG2a Fc alone (Ad-Fc) were constructed and propagated as reported previously²⁹. Infection with Ad-VEGFR-2 produces soluble VEGFR-2 extracellular ligand binding domain in mice, which binds and systemically sequesters VEGF-A, thus eliminating VEGF-A available for binding to its cellular receptors. Ad-VEGFR-2 or Ad-Fc (10⁹ PFUs) was intravenously injected to each mouse 3 days prior to PPE infusion. A single injection of Ad flk1-Fc transfected hepatocytes to produce long-lasting (>3 weeks) and high levels of circulating soluble VEGFR-2 ligand-binding domain, with a plasma concentration of 2-9 mg/ml, sufficient for systemic VEGF inhibition in several experimental settings^29-31.

In an additional mouse cohort, a cross-species reactive anti-VEGF-A neutralizing mAb (Clone G6-31, Genentech Inc., South San Francisco, CA) or mouse IgG as negative control, were administered intraperitoneally at a dose of 5 mg/kg every other day³². Treatment with VEGF-A mAb or mouse IgG was either initiated 3 days prior to (progression study), or 4 days following (regression study), PPE infusion, and then continued through 14 days following AAA creation in all mice.

In a third mouse cohort, VEGF receptor signaling was inhibited by administration of the multiple RTK inhibitor sunitinib (Major Resources Table). Sunitinib was prepared in vehicle containing 0.5% carboxymethyl cellulose sodium, 1.8% NaCl, 0.4% Tween 80 and 0.9% benzyl alcohol immediately prior to use, and administered via oral gavage. Treatment with sunitinib or vehicle alone was also initiated either 3 days prior to (progression study with dosing at 4, 20 and 100 mg/kg/day) or 4 days following (regression study with dosing at 4 and 100 mg/kg/day), PPE infusion, and continued through 14 days following AAA creation in both groups. Doses were derived from both human and experimental published data^33-35. By allometric scaling, 4 mg/kg/day in the mouse is comparable to the clinical dose administered to 70 kg cancer patients (SUTENT® (sunitinib malate) capsules, FDA reference ID: 2950085).

Aortic diameter measurement

Aortic diameter, as a surrogate for AAA development and progression, was measured in vivo via transabdominal ultrasonography at 40 MHz (Vevo 770, Fujifilm VisualSonics, Toronto, ON, Canada)¹². These measurements were performed on the day prior to (day 0), and 3, 7 and 14 days after, PPE infusion. An AAA was defined as a ≥ 50% increase in aortic diameter over the baseline.

Histopathological analyses

All histological analyses were conducted on aortic sections prepared from mice 14 days following PPE or PBS infusion. For tissue immunostaining, mice were perfused with 4% paraformaldehyde (for paraffin sections) or PBS (for frozen sections) via left ventricle. Aortae were harvested, embedded in paraffin or OCT media, and sectioned (thickness: 4 and 8 μm for paraffin and frozen sections, respectively). Standard biotin-streptavidin peroxidase procedures were employed as previously described¹². Immunostaining was performed on paraffin sections for VEGF-A antibody and on frozen sections for all other staining.

Primary antibodies used for tissue immunostaining were a cross-species reactive rabbit anti-VEGF-A polyclonal antibody; a rabbit or goat anti-mouse SMC α actin polyclonal antibody; goat polyclonal antibodies against MMP2 and MMP9; rat mAbs against mouse CD11b, CD45R/B220 and CD31; a hamster mAb against mouse CD3 mAb; isotype-matched control antibodies (rat IgG2a, rat IgG2b and hamster IgG); and normal rabbit or goat IgG. Other reagents included biotinylated anti-rabbit, rat or hamster IgG antibody and streptavidin-peroxidase conjugates; Alexa Fluor 488 donkey anti-rabbit IgG antibody, Alexa Fluor 594 donkey anti-rat IgG antibody and Alexa Fluor 594 donkey anti-goat IgG antibody; and the peroxidase substrate AEC or DAB (VEGF-A staining only) kit. Sections were counterstained with hematoxylin and coverslipped. Sources, clone number, catalog number and working solution concentrations for all tissue immunostaining reagents, as well as flow cytometric reagents, are summarized in the Major Resources Table.

To colocalize VEGF-A with CD11b⁺ myeloid cells and SMCs, aortic paraffin sections were sequentially incubated with rabbit VEGF-A and rat CD11b (or goat SMC α actin) antibodies, Alexa Fluor 488 donkey anti-rabbit IgG antibody (to detect VEGF-A) and Alexa Fluor 594 donkey anti-rat IgG antibody (to detect CD11b) (or Alexa Fluor 594 donkey anti-goat IgG antibody to detect SMCs), PBS, mounted with fluorescence mounting media, and imaged on florescence microscopy.

Aortic monocytes/macrophages (CD11b⁺), T cells (CD3⁺) and B cells (B220⁺) were tabulated as positive cells per aortic cross-section (ACS), while mural angiogenesis was quantified as CD31⁺ neovessels per ACS. Medial elastin was stained using the Elastica Van Gieson (EVG) technique. Both medial elastin degradation and SMC loss were graded as I (mild) to IV (severe) according to EVG and SMC α actin staining pattern (Supplemental Table I) as described previously¹³. Aortic immunostaining for MMP2 and MMP9 expression was quantitated using NIH ImageJ (Ver 1.49, http://imagej.nih.gov/ij) and data were presented as the percentage of positively stained area to total aortic cross sectional area.

In vitro stimulation of macrophages and aortic SMCs for the production of MMPs, CCL2 and VEGF-A

Bone marrow-derived macrophages were generated in vitro by exposure of whole bone marrow cells to M-CSF (20 ng/ml) for one week³⁶. To isolate aortic SMCs, endothelial and adventitial layers were sequentially removed from harvested aortae. Aortic media was digested with HBSS buffer containing type II collagenase (1 mg/ml) and elastase (744 units/ml) for 1 h³⁷. Residual cells were cultured in DMEM/F12/20% FBS media for one week until confluent. Bone marrow-derived macrophages and confluent SMCs were activated with LPS (20 ng/ml) and VEGF-A (40 ng/ml for 48 and 36 h, respectively (Major Resources Table), and assayed for expression of mRNAs for MMPs, CCL2 and VEGF-A^{36, 38}.

In vitro leukocyte chemotaxis assay

Chemotaxis was assayed in RPMI-1640/10% FBS media using the 5-μm pore, 24-well transwell plate according to the manufacturer’s instructions (Costar, Corning, NY). Briefly, mononuclear cells were isolated from bone marrow and preincubated with sunitinib (40 ng/ml) or vehicle alone for 30 min. Subsequently 10⁶ cells (100 μl media) and a chemoattractant (20 ng/ml CCL2 or 40 ng/ml VEGF-A, 600 μl media) were added to the top and bottom wells, respectively (Major Resources Table). Following 3 h incubation at 37°C in 5% CO₂ atmosphere, migrated cells in the bottom wells were harvested, counted, stained with fluorochrome-conjugated, cell-type (myeloid, inflammatory and non-inflammatory monocyte, and neutrophil)-specific mAbs, and analyzed via flow cytometry (see Flow Cytometric Analysis Section below). Data for migrated cells are presented as the absolute numbers for total leukocytes, myeloid cells, inflammatory and non-inflammatory monocytes, and neutrophils.

Flow cytometric analyses

Flow cytometry was used to quantify myeloid cells, inflammatory and non-inflammatory monocytes, and neutrophils in peripheral blood and in vitro cell migration assays. For blood samples, whole blood was collected from mouse tail vein and suspended in 10 mM EDTA/PBS buffer. Samples were stained with differential fluorochrome-conjugated mAbs against CD11b, Ly-6C and Ly-6G (Major Resources Table) or followed by lysis of erythrocytes with red blood cell lysis buffer for blood samples. Staining data were acquired on a BD FACS Calibur or LRSII flow cytometer with CellQuest or FACSDiva software (BD Biosciences, San Jose, CA) and analyzed using FlowJo software (Ver 10.0.8r1, FlowJo LLC, Eugene, OR). Inflammatory monocytes and neutrophils were identified as CD11b⁺Ly-6C^high and CD1b⁺Ly-6G⁺ cells, respectively. All data were presented as the percentage of inflammatory monocytes or neutrophils in total leukocytes.

Statistical Analysis

All continuous data are presented as mean and standard deviation (SD). The D’Agostino-Pearson omnibus normality test was used to determine whether the data were normally distributed. For normally distributed data, the Student’s t-test or two-way repeat measures ANOVA (to determine influence by row and column factors as well as their interaction), followed by the Newman-Keuls post-hoc test (to determine the significance between two treatment groups at an indicated time point), were used to determine the significance between groups. For data failing normal distribution, differences were tested using the non-parametric Mann-Whitney test. Difference in aneurysm incidence between two groups was tested by Kaplan-Meier analysis. All statistical analyses were performed using Prism version 6.0h, Graphpad Software, Inc, La Jolla, CA. P<0.05 was considered statistically significant. Sample size and statistical power are summarized in the Supplemental Table II.

Results

Increased VEGF-A expression in experimental AAAs

To evaluate the expression of pro-angiogenic VEGF-A in PPE-induced experimental AAAs, aortic paraffin sections were prepared from mice 14 days following intra-aortic PPE (aneurysmal) or PBS (non-aneurysmal) infusion. Sections were stained for VEGF-A using a rabbit anti-VEGF-A polyclonal antibody. As compared to PBS-infused non-aneurysmal aortae (Fig. 1A), VEGF-A-expressing cells were substantially increased in PPE-infused aneurysmal aortae (Fig. 1B). Normal rabbit IgG, a negative control for anti-VEGF-A antibody, did not stain the sections from non-aneurysmal (not shown) or aneurysmal (Fig. 1C) aortae. High power magnification views of VEGF-A staining in non-aneurysmal and aneurysmal aortae, respectively, were illustrated in Figs. 1D & 1E. In two-color immunofluorescence staining, most VEGF-A-positive staining colocalized with aortic wall CD11b⁺ myeloid cells in aneurysmal lesions (Fig. 1F), whereas very few VEGF-A-positive cells were observed in the residual medial SMC population (Fig. 1G). Additionally, real-time qRT-PCR confirmed elevated VEGF-A mRNA levels in aneurysmal, as compared to non-aneurysmal, aortae (Fig. 1H). These results indicate that VEGF-A expression is augmented in PPE-induced experimental AAAs.

Figure 1. Aortic mural VEGF-A expression is increased in experimental AAAs.

Male C57BL/6J mice underwent transient intra-aortic infusion of PPE (aneurysm) or PBS (control, non-aneurysm). Fourteen days thereafter, mice were sacrificed, and aortae were harvested for tissue immunostaining or real-time qRT-PCR analyses. (A-E): Representative VEGF-A immunostaining in control (A, D) and aneurysmal (B, E) aortic paraffin-embedded sections with a rabbit anti-VEGF-A polyclonal antibody. Neither aneurysmal (C) nor control (not shown) aortae were stained with normal rabbit IgG (a negative control for anti-VEGF-A antibody). Original magnification: 10X for A, B & C; and 40X for D & E. (F, G): Two-color immunofluorescence staining of VEGF-A with myeloid cell (F, CD11b) or SMC (G, SMC α actin) markers on aortic paraffin aneurysmal sections. (H): Mean and SD of aortic VEGF-A mRNA levels measured via real-time qRT-PCR assays. mRNA levels: fold changes relative to normal aortae from mice without PPE or PBS infusion. Non-parametric Mann-Whitney test, *P<0.05 compared to PBS group. n=5-7 mice per group.

VEGF-A sequestration via soluble VEGFR-2 suppresses aneurysm development

To probe the potential causality of VEGF-A expression in experimental AAA pathogenesis, mice were injected with Ad-VEGFR-2 or Ad-Fc (as the negative control) intravenously 3 days prior to AAA creation. Ad-VEGFR-2 transfer promotes soluble VEGFR-2 extracellular ligand-binding domain levels in plasma (2-9 mg/ml) sufficient for physiologically significant in vivo VEGF-A sequestration in several experimental settings^29-31. Hematocrit levels were sufficiently elevated in Ad-VEGFR-2-treated mice 2 weeks following adenovirus infection, indicating effective VEGF-A sequestration by this method (Supplemental Fig. I).

In Ad-Fc-injected mice, aortae enlarged remarkably and significantly from days 3 to 14 following PPE infusion (Figs. 2A & 2B). In Ad-VEGFR-2-injected mice, AAA formation was nearly completely abolished. All Ad-Fc mice, as compared to no Ad-VEGF-R2 mice, developed AAAs (defined as a 50% or greater increase in aortic diameter over the baseline level) within 14 days following PPE infusion (Fig. 2C).

Figure 2. Ad-VEGFR-2 gene transfer suppresses experimental AAAs.

Male C57BL/6J mice were inoculated intravenously with Ad-VEGFR-2 or Ad-Fc (n=5-7 mice per group). Three days after adenovirus inoculation, AAAs were created by transient intra-aortic infusion of PPE. Aortic diameter was subsequently monitored by serial transabdominal ultrasonographic examination. AAA histological examination was performed at sacrifice (14 days after PPE infusion). (A): Representative infrarenal aortic ultrasound images prior to, and 14 days following, PPE infusion. (B): Infrarenal aortic diameters for individual mice as well as the mean and SD for each treatment group at the baseline level (day 0) and indicated days following PPE infusion. Two-way ANOVA followed by Newman-Keuls post-test, *P<0.05 compared to Ad-Fc group at corresponding time point. (C): AAA-free mice. AAAs: ≥ 50% increase in aortic diameter over the baseline level. Kaplan-Meier test, **P<0.01 compared to Ad-Fc group. (D): Representative histological staining images for medial elastin (EVG), SMCs (SMC α- actin), mural monocytes/macrophages (CD11b) and angiogenesis (CD31). Original magnification: 10X. (E-G): Quantification (mean and SD) of elastin degradation and SMC depletion severity score (E), mural angiogenesis (F), and aortic accumulation of monocytes/macrophages (Mono/Mac, CD11b), T cells (CD3) and B cells (B220). Non-parametric Mann-Whitney test, *P<0.05 compared to Ad-Fc group.

On histological analysis, Ad-VEGFR-2 treatment significantly attenuated medial elastin degradation and SMC depletion (Figs. 2D & 2E). Mural angiogenesis, as quantified by CD31 immunostaining, was significantly diminished in Ad-VEGFR-2-treated mice (Figs. 2D & 2F). Additionally, mural CD11b⁺ monocyte/macrophage infiltration (but not T or B cells) was significantly reduced (Figs. 2D & 2G).

Taken together, these experimental results support the hypothesis that VEGF-A mediates aneurysm formation at least in part via promotion of aortic mural angiogenesis and monocyte/macrophage accumulation.

Anti-VEGF-A mAb suppresses aneurysm development

VEGFR-2 also binds VEGF-subtypes C, D and F with varying but generally low affinity. To isolate VEGF-A specific effects in aneurysm pathogenesis, mice were intraperitoneally injected with a cross-species reactive anti-VEGF-A blocking mAb every other day beginning 3 days prior to PPE infusion. By ultrasonic measurement, aortic diameter increased progressively and substantially from day 3 following PPE infusion in control mAb-treated mice (Figs. 3A & 3B). In comparison, anti-VEGF-A mAb treatment substantially abrogated PPE-induced aortic enlargement (Figs. 3A & 3B). Less than 20% of anti-VEGF-A mAb-treated mice developed AAAs within 14 days following PPE infusion, as compared to 100% of control mAb-treated mice (Fig. 3C), underscoring the specific role of the “A” subtype in experimental aneurysm formation and progression in this model.

Figure 3. Neutralization of VEGF-A prevents AAA initiation but does not limit the progression of established AAAs.

Anti-VEGF-A or its isotype control mAb was given to male C57BL/6J mice every other day starting 3 days prior to (e.g. “progression”), or 4 days following (“regression”), PPE infusion. Influences on AAA progression and regression were assessed by ultrasonography-aided serial measurements of maximal infrarenal aortic diameters and histological analyses at sacrifice (14 days after PPE infusion). Five to seven mice were used in each group. (A): Representative infrarenal aortic ultrasound images prior to, and 14 days following, PPE infusion. (B): Infrarenal aortic diameters for individual mice as well as the mean and SD for each treatment group at the baseline level (day 0) and indicated days following PPE infusion. Two-way ANOVA followed by Newman-Keuls post-test, *P<0.05 and **P<0.01 compared to control mAb group at identical time points. (C): The percentage of AAA-free mice. AAA: ≥ 50% increase in aortic diameter over the baseline level. Kaplan-Meier test, **P<0.01 compared to control mAb group. (D): Infrarenal aortic diameters for individual mice as well as the mean and SD for each treatment group in mice injecting VEGF-A mAb or control mAb for 10 days starting on day 4 following PPE infusion. (E): Quantification (mean and SD) of elastin degradation and SMC depletion scores (left), mural angiogenesis (middle), and aortic accumulation of monocytes/macrophages (Mono/Mac, CD11b), T cells (CD3) and B cells (B220) (right). Non-parametric Mann-Whitney test, *P<0.05 and **P<0.01 compared to control mAb group.

Humanized anti-VEGF-A mAb has considerable clinical utility in the treatment of specific solid-organ tumors. To further evaluate its potential translational value in AAA disease, mice were treated with an anti-VEGF-A or a control mAb starting day 4 following PPE infusion, the time point at which most AAAs form in this model. Influence on further aortic enlargement was assessed by ultrasound measurements of aortic diameters following mAb treatment. No significant difference in aortic enlargement was noted at either day 7 or 14 between VEGF-A and control mAb treatments (Fig. 3D). These results suggest that mechanisms of AAA suppression likely vary in different phases of aneurysm progression, and that anti-VEGF-A mAb, effective in suppressing the initiation phase, has less promise as a potential clinical therapy for existing disease.

On EVG and SMC α actin staining, anti-VEGF-A mAb treatment initiated prior to AAA induction substantially mitigated medial elastin degradation and SMC depletion (Fig. 3E, left panel). In mice with existing aneurysms, similar but less significant effects were noted; as expected, the density of mural CD31-expressing neovessels in the aortae from mice treated with either VEGF-A mAb regimen was also significantly reduced (Fig. 3E, middle panel). Both VEGF-A mAb treatment strategies markedly reduced mural accumulation of leukocytes, including CD11b⁺ monocytes/macrophages, CD3⁺ T cells and B220⁺ B cells, as compared to control mAb treatment (Fig. 3E, right panel). Additionally, VEGF-A mAb treatment initiated prior to aneurysm creation exhibited slightly more inhibition on mural angiogenesis and leukocyte infiltration than that accomplished via later treatment, although this difference did not reach statistical significance.

These results indicate that VEGF-A modulates AAA initiation by regulating mural leukocyte accumulation and angiogenesis.

The RTK inhibitor sunitinib limits AAA initiation and progression

Sunitinib, a FDA-approved RTK inhibitor, targets multi-RTKs, with a high affinity for VEGFR-2 and lower affinity for platelet-derived growth factor receptors, fibroblast growth factor receptor and c-Kit³⁹. Sunitinib was administered to mice at 4, 20 or 100 mg/kg/day via oral gavage beginning 3 days prior to PPE infusion and continued for a total of 17 days. As expected, vehicle-treated mice underwent substantial aortic enlargement from day 3 through day 14 following PPE infusion (Figs. 4A & 4B). In comparison, sunitinib administration, in all dosages tested, substantially limited aortic enlargement in response to PPE. Whereas all vehicle-treated mice developed AAAs within 14 days following PPE infusion, no AAAs developed in sunitinib-treated mice (Fig. 4C).

Figure 4. Sunitinib limits experimental AAAs.

Male C57BL/6J mice were daily treated with vehicle or varying doses of sunitinib starting 3 days prior to PPE infusion for a total of 17 days. AAA formation and progression were assessed via serial ultrasound measurements of infrarenal aortic diameters and histology at sacrifice. Four to seven mice were used in each group. (A): Representative aortic ultrasound images prior to (day 0), and 14 days following, PPE infusion in mice treated with vehicle or 100 mg/kg sunitinib. (B): Infrarenal aortic diameters for individual mice as well as the mean and SD for each treatment group at the baseline level (day 0) and indicated days following PPE infusion. Two-way ANOVA followed by Newman-Keuls post-test, **P<0.01 compared to vehicle group at same time point. (C): AAA-free mice. AAA: ≥ 50% increase in aortic diameter over the baseline level. Kaplan-Meier test, **P<0.01 compared to vehicle group. (D-F): Quantification (mean and SD) of medial elastin degradation and SMC depletion scores (D), mural angiogenesis (CD31), and aortic accumulation of monocytes/macrophages (Mon/Mac, CD11b), T cells (CD3) and B cells (B220). Non-parametric Mann-Whitney test, **P<0.01 compared to vehicle group.

On histological analysis (Figs. 4D-4F & Supplemental Fig. II), pretreatment with sunitinib substantially mitigated medial elastin degradation and SMC depletion as assessed by EVG and SMC α actin staining, respectively (Figs. 4D & Supplemental Fig. II). A marked reduction in mural neoangiogenesis was also present, as indicated by the reduced density of CD31-expressing mural capillary vessels in sunitinib-treated mice (Figs. 4E & Supplemental Fig. II). Treatment also attenuated mural accumulation of leukocytes, including CD11b⁺ monocytes/macrophages, CD3⁺ T cells and B220⁺ B cells (Figs. 4F & Supplemental Fig. II). However, no dose-response relationship was observed for individual histologic parameters.

Thus, sunitinib treatment is highly effective in suppressing aortic aneurysms in a dose-independent manner at least within the dose range tested in this experimental AAA modeling system.

Sunitinib limits the progression of existing AAAs

To examine its effectiveness in limiting the progression of existing AAAs, in additional mouse cohorts, sunitinib or vehicle was administered beginning 4 days following AAA creation. As shown in Figs. 5A & 5B, further aortic diameter enlargement was significantly limited in mice receiving either 4 or 100 mg/kg sunitinib as compared to those treated with vehicle alone. Similar to sunitinib pretreatment, no dose-response effect was noted for further aneurysmal enlargement (Fig. 5B).

Figure 5. Sunitinib limits the progression of established AAAs.

Male C57BL/6J mice were daily treated with vehicle or sunitinib (100 or 4 mg/kg) beginning on day 4 following PPE infusion for a total of 10 days. Aneurysm expansion was assessed via serial ultrasound imaging of infrarenal aorta and histology at sacrifice. There were five to eight animals in each group. (A): Representative aortic ultrasound images prior to (day 0), and 10 days after, initiating vehicle or 100 mg/kg sunitinib treatment. (B): Infrarenal aortic diameters for individual mice as well as the mean and SD for each treatment group administered sunitinib or vehicle daily for 10 days beginning day 4 after AAA creation. Two-way ANOVA followed by Newman-Keuls post-test, **P<0.01 compared to vehicle group at corresponding time point. (C-E): Quantification (mean and SD) of medial elastin degradation and SMC depletion (C), mural angiogenesis (CD31) (D), and aortic accumulation of monocytes/macrophages (Mon/Mac, CD11b), T cells (CD3) and B cells (B220) (E). Non-parametric Mann-Whitney test, *P<0.05 and **P<0.01 compared to vehicle group.

Consistent with the effects on aneurysmal expansion, medial elastin fibers and SMC cellularity were well preserved in sunitinib-treated mice (Fig. 5C & Supplemental Figure II). Expectedly, the density of CD31-expressing mural vessels was substantially and significantly lower in sunitinib-treated mice (Fig. 5D & Supplemental Figure II). The numbers of mural leukocytes, including CD11b⁺ monocytes/macrophages, CD3⁺ T cells and B220⁺ B cells, were also significantly reduced (Fig. 5E & Supplemental Figure II). Similar to its influence on aortic diameter, no dose-response effect was noted for these histological parameters at least within the dose range tested in this study.

Regardless of its mechanism of action, sunitinib was highly effective in suppressing both aneurysm initiation and further progression of established AAAs in the PPE-infusion modeling system.

Sunitinib treatment downregulates MMP message and protein

To further investigate potential mechanisms of sunitinib-mediated AAA suppression, we performed immunostaining on aortic frozen sections from differentially treated mice using antibodies against MMP 2 and MMP 9. As shown in Supplemental Fig. III, substantial MMP 2 & 9 staining was present in the adventitia and SMC-depleted media in vehicle-treated mice. Consistent with diminished aortic leukocyte infiltration and well-preserved aortic media, only slight mural MMP 2 & 9 staining was present in sunitinib-treated mice. Normal goat IgG, the negative control for both MMP antibodies, did not stain aortic sections from mice with either treatment. Both MMP2 & 9 protein expression, as quantified by area of positive staining, was significantly reduced in sunitinib-treated mouse aortae (Supplemental Fig. III).

Further, we assessed whether sunitinib treatment per se inhibits MMP production by SMCs and macrophages (Supplemental Fig. III). In bone marrow-derived macrophages stimulated with LPS, sunitinib treatment was associated with a >50% reduction in mRNA expression for both MMP 2 & 9. In SMCs, MMP 9 mRNA was undetectable in sunitinib-treated mice regardless of the presence or absence of VEGF-A. Sunitinib treatment reduced MMP 2 mRNA levels by >60%.

Thus, sunitinib mediates AAA suppression in part by downregulating MMP 2 and MMP 9.

Sunitinib suppresses inflammatory cell chemotaxis in vitro

CCL2 and VEGF-A recruit circulating myeloid cells into tissues by binding cell surface CCR2 and VEGF-R1, respectively. Thus, we evaluated the effect of sunitinib treatment on chemotaxis of monocytes and neutrophils to CCL2 and VEGF-A in in vitro transwell migration assays.

As shown in Supplemental Fig. IV, pretreatment of bone marrow cells with sunitinib reduced chemotaxis of total bone marrow or myeloid cells, inflammatory monocytes or neutrophils to VEGF-A by >25%, without obvious effect on non-inflammatory monocyte chemotaxis. In contrast, sunitinib treatment did not impact transwell migration of total or individual subsets of myeloid cells to CCL2. In additional experiments, we found an 50% reduction in the expression of mRNA for VEGF-A, but not CCL2, in bone marrow-derived macrophages and aortic SMCs following stimulation with LPS and VEGF-A, respectively (Supplemental Fig. IV).

These results suggest that AAA suppression by sunitinib may result in part from reduced VEGF-A-mediated inflammatory cell chemotaxis.

Sunitinib reduces circulating monocytes

Flow cytometric analysis was used to quantify circulating inflammatory monocytes (CD11b⁺Ly-6C^high cells), the precursors for tissue macrophages, present in PPE-infused mice treated with either sunitinib or vehicle.

As shown in Fig. 6, the relative number of CD11b⁺Ly-6C^high cells (inflammatory monocytes) was substantially increased at 3 and 7 days following PPE infusion in vehicle-treated AAA mice. However, sunitinib treatment was associated with reduced relative number of circulating CD11b⁺Ly-6C^high cells one day following PPE infusion as compared to the baseline. Sunitinib treatment further blunted the increase in circulating inflammatory monocytes noted on days 3 and 7 following aneurysm creation. In contrast, there was only a transient increase in circulating neutrophils, identified as CD11b⁺Ly-6G⁺ cells, regardless of sunitinib-treatment status.

Figure 6. Sunitinib reduces circulating inflammatory monocytes but not neutrophils in experimental AAAs.

Male C57BL/6J mice were treated with vehicle alone or 4 mg/kg sunitinib beginning 2 h prior to PPE infusion and daily thereafter for a total of 7 days. Whole blood samples collected from tail veins prior to, and on indicated days following, PPE infusion were stained with mAbs against CD11b, Ly-6C and Ly-6G. Stained samples were analyzed by flow cytometry. (A): Inflammatory monocytes (CD11b⁺Ly-6C^high). (B): Neutrophils (CD11b⁺Ly-6G⁺). Data presented as mean and SD of the percentages of inflammatory monocytes and neutrophils in total leukocytes gated by forward and side scatters. Two-way ANOVA followed by Newman-Keuls post-test, **P<0.01 compared to vehicle group at corresponding time point. n=6-7 mice per group.

These results suggest that at least some of the anti-aneurysmal effects associated with sunitinib therapy in this model may be associated with influences on circulating inflammatory monocyte titer.

Discussion

Three distinct VEGF-A inhibition strategies, including Ad-VEGFR-2 gene transfer-mediated VEGF-A sequestration, VEGF-A mAb-mediated VEGF-A neutralization and VEGF-A receptor inhibition by sunitinib, all effectively suppress aortic aneurysmal enlargement and mural degradation following PPE infusion in mice. Sunitinib, but not VEGF-A mAb, was remarkably effective in limiting both AAA formation and progression. All three strategies attenuated aortic mural angiogenesis and leukocyte accumulation. Sunitinib was also effective in reducing MMP2, MMP9 and VEGF-A mRNA expression by macrophages and SMCs, chemotaxis of inflammatory monocytes and neutrophils to VEGF-A, and circulating inflammatory monocyte titers. These findings underscore the specific role of the “A” subtype of VEGF and its receptor in experimental AAA pathogenesis in this model, as well as identify RTK inhibition as a promising translational pharmacologic strategy for AAA disease suppression.

VEGF-A plays a critical role in multiple pathological conditions¹⁹. Prior studies have shown increased VEGF-A expression by vascular SMCs and infiltrative leukocytes in human and experimental AAA tissue^{7, 21-25}. The current study identified CD11b-positive cells as a major source for VEGF-A in AAAs, with a very limited contribution from residual medial SMCs. Aortic VEGF-A levels are known to correlate with size and risk of rupture for aortic aneurysms ^{6, 40, 41}. The present findings validate and extend the significance of VEGF-A expression and activity in AAA pathogenesis.

VEGF-A binds to, and signals via, VEGFR-2 to promote angiogenesis^{20, 42}. VEGF-A also recruits VEGFR-1-expressing myeloid cells to target tissues for either tissue damage or repair⁴². In prior experiments in the angiotensin II/hyperlipidemic mouse model, recombinant VEGF-A treatment, starting 1 or 2 weeks following angiotensin II infusion, modestly augmented aneurysm enlargement without affecting incidence, rupture rate or mortality²⁶. The angiotensin II/hyperlipidemic mouse modeling system, however, demonstrates specific pathologic features not present in isolated infrarenal human AAAs, such as predisposing focal aortic dissections in areas of eventual aneurysmal enlargement ²⁷. In the current experiments, based on a modeling system with greater pathologic fidelity to the human condition, minimal effect on established aneurysm progression was noted following VEGF-A mAb administration. Taken together, both these gain- and loss-of-function experimental results, obtained in complementary experimental modeling systems, would seem to diminish the translational potential of humanized VEGF-A mAb in clinical AAA management.

RTK inhibition may hold more translational promise for AAA suppression. Monocytes/macrophages exert multifaceted roles in AAA disease⁴³. Aortic mural monocyte/macrophage density is regulated by neovessels (the portal for cellular ingress), tissue gradients of chemoattractant factors such as CCL2 and VEGF-A, and the number of circulating monocytes^{44, 45}. In two sets of in vitro experiments, sunitinib treatment partially inhibited migration of myeloid cells, particularly Ly-6^high inflammatory monocytes and Ly-6G⁺ neutrophils, towards VEGF-A as well as attenuated VEGF-A mRNA expression by activated macrophages and SMCs. No similar response was seen to CCL2. Thus, inhibiting the expression of, and monocyte chemotaxis to, VEGF-A may serve as an alternative mechanism for AAA suppression by sunitinib, likely by limiting circulating monocyte migration to inflamed aorta as well as lessening VEGF-A-mediated angiogenesis.

Sunitinib treatment also reduced circulating inflammatory monocytes, thus diminishing the population of monocytes available for transendothelial migration and aortic localization. Mobilization of inflammatory monocytes from bone marrow to the circulation is mediated by the chemokine CCL2 and its receptor CCR2^46-48. Sunitinib was previously reported to reduce CCL2 production in certain in vitro and in vivo conditions^{49, 50}. In the present experiments, sunitinib treatment did not alter monocyte chemotaxis to, or macrophage- or SMC-derived mRNA levels of, CCL2. The finding that sunitinib partially impaired monocyte chemotaxis to VEGF-A raises the possibility that VEGF-A may contribute to bone marrow inflammatory monocyte mobilization in this experimental construct and in part account for reduced circulating inflammatory monocytes following sunitinib treatment.

While a dose-response effect was not observed across the range of doses tested, even at 4 mg/kg/day (substantially less than that recommended for clinical cancer chemotherapy (US FDA Reference ID: 3675896)) early and delayed AAA suppression was apparent in this modeling system. Indeed at either 4 or 100 mg/kg/day, some slight reduction of existing aneurysms was apparent, although the relevance of this finding to clinical efficacy is uncertain. Thus, sunitinib, with well-established pharmacokinetics, could conceivably be considered a viable candidate agent for medical suppression of clinical AAA disease, although its side effect profile, acceptable for anti-oncologic applications, might preclude feasibility in “worried-well” AAA patients with perhaps a 70% chance of lifetime surgical intervention for the disease (https://www.sutent.com/possible-side-effects).

Recruitment of inflammatory leukocytes via luminal expression of adhesion molecules and chemokines is an essential contributor to AAA pathogenesis. In these experiments, attenuated aortic mural angiogenesis coincided with markedly reduced aortic leukocyte density regardless of the angiogenesis inhibition strategy applied. In human aneurysmal aortae, neovessels express ICAM-1, VCAM-1 and E-selectin⁵¹. In prior experiments, genetic deficiency of P-selectin or ICAM-1 suppressed or nearly completely abrogated experimental AAAs^52-54. These results serve to further highlight the significance of neovessel adhesion molecules and their respective leukocyte cognate receptors in aneurysm pathogenesis.

This study has several limitations. Other non-RTK activity-dependent, anti-angiogenic properties of sunitinib include inhibition of hypoxia inducible factor-1α (a transcription factor regulating the expression of VEGF-A and other pro-angiogenic factors) expression and activity, promotion of anti-inflammatory macrophage phenotypic differentiation and soluble VEGFR-2 production^{33, 55, 56}, and depletion of vascular pericytes^{56, 57}. Most of these effects also influence angiogenesis and, potentially, AAA pathogenesis^58-60. Isolating and quantifying the specific mechanism(s) by which sunitinib treatment reduced experimental AAA production was beyond the scope of these experiments. Similarly, examining the recruitment of VEGFR-2-expressing endothelial progenitor cells to peripheral tissues⁶¹, another potential pro-angiogenic mechanism associated with VEGF-A, was not feasible in this experimental construct.

As a practical matter, no experimental design will comprehensively assess the functional relationships between monocytes/macrophages, aortic mural angiogenesis, and the relative contributions of VEGFR-1, 2 and related RTKs in VEGF-A neutralization/sequestration- or sunitinib-mediated AAA suppression with currently available models. Finally, uncertainty remains regarding the validity and fidelity of experimental aneurysm models, including the PPE infusion model, to human AAA disease. Our results in the PPE infusion AAA model, however, correlate well with those previously obtained in alternative AAA modeling systems, further strengthening the likely translational value of the observed results.

In conclusion, these results further support the hypothesis that angiogenesis in general, and VEGF-A-mediated angiogenesis in particular, promote the pathogenesis and progression of experimental AAAs. Pharmacologic inhibition of RTK or analogous mechanisms holds promise as a potential translational strategy for suppression and medical management of clinical AAA disease.

Highlights.

• Sequestering or blocking VEGF-A suppressed elastase infusion-induced aortic aneurysm enlargement.

• Receptor tyrosine kinase inhibitor sunitinib suppressed initiation of experimental AAAs as well as further progression of established aneurysms.

• Inhibition of VEGF-A or its receptor activity preserved medial elastin and SMCs and ameliorated mural monocyte/macrophage infiltration and neoangiogenesis.

• Sunitinib treatment suppressed MMP2 and MMP9 protein expression in aneurysmal lesions as well as attenuated MMP mRNA expression by both activated macrophages and SMCs in vitro.

• Sunitinib treatment inhibited myeloid cell migration towards VEGF-A and reduced VEGF-A gene expression in activated SMCs and macrophages.

• Sunitinib treatment attenuated the increase in circulating inflammatory monocytes following experimental AAA creation.

Abbreviations

AAA

abdominal aortic aneurysm

Ad

adenovirus

EVG

Verhoeff’s Van Gieson

Mon/Mac

monocyte/macrophage

PPE

porcine pancreatic elastase

RTK

receptor tyrosine kinase

SMC

smooth muscle cell

VEGF

vascular endothelial growth factor

VEGFR

vascular endothelial growth factor receptor