Lenvatinib halts aortic aneurysm growth by restoring smooth muscle cell contractility

Abdominal aortic aneurysm (AAA) is a disease with high morbidity and mortality, especially when ruptured. The rationale of this study was to evaluate the repurposing of lenvatinib, a multi–tyrosine kinase inhibitor, in limiting experimental AAA growth targeting vascular smooth muscle cells (VSMCs) and angiogenesis. We applied systemic and local lenvatinib treatment to elastase-induced murine aortic aneurysms, and RNA profiling identified myosin heavy chain 11 (Myh11) as the most deregulated transcript. Daily oral treatment substantially reduced aneurysm formation in 2 independent mouse models. In addition, a large animal aneurysm model in hypercholesterolemic low-density lipoprotein receptor–knockout (LDLR–/–) Yucatan minipigs was applied to endovascularly deliver lenvatinib via drug-eluting balloons (DEBs). Here, a single local endovascular delivery blocked AAA progression successfully compared with a DEB-delivered control treatment. Reduced VSMC proliferation and a restored contractile phenotype were observed in animal tissues (murine and porcine), as well as AAA patient-derived cells. Apart from increasing MYH11 levels, lenvatinib reduced downstream ERK signaling. Hence, lenvatinib is a promising therapy to limit aortic aneurysm expansion upon local endovascular delivery. The tyrosine kinase inhibitor was able to positively affect pathways of key relevance to human AAA disease, even in a potentially new local delivery using DEBs.


Suppl.
: Murine AngII-AAA data: Oral gavage of lenvatinib (dark red; n=6) continuously from day 4 (ultrasound at day 3) after AAA induction by Angiotensin II infusion. (A) Absolute aortic diameters and volumes in mm and mm 3 for the murine AngII and lenvantinib experiments (n=5 for AngII; n=6 for AngII + Lenva systemic; *=p<0.05; mean + SEM; Ang II induction started by implantation of osmotic minipump at day 0). (B) shows the relative aortic diameter and volume for the same experiment (relative to baseline aortic diameter/volume). (p-value calculations and detailed measurements in Suppl. Table I).
Suppl. Figure 4: PPE MTA pathway analysis: (A) KEGG-pathway gene overrepresentation analysis of significantly up-regulated genes (adjusted p-value < 0.05, Fc < -2) at day 7 post PPE compared to sham (saline) controls, showing baseline transcriptomic characteristics of the model at the timepoint of lenvatinib treatment initiation. Top 15 significantly affected pathways are shown (adjusted p value is given in shades of blue). (B) Subset of contractile marker gene expression analysis of PPE mice aortic tissue on day 28 after aneurysm induction compared to untreated control aortas based on own MTA array data (upper panel). Similar analysis of PPE mice aortic tissue on day 7 after aneurysm induction compared to sham-treated (saline) aortas (lower panel) (*=p<0.05).
Suppl. Figure 5: Lenvatinib effect on AAA patient-derived SMCs #3: (A) Migration (left panel) proliferation (middle panel) and apoptotic (right panel) assays of AAA patient #3-derived SMCs exposed to lenvatinib. (B) A timely resolution of gene expression analysis at 6h, 24h, and 48h, assessing early and late effects of lenvatinib treatment on primary human AAA patient-derived VSMCs. (AAA control=treatment with 0.1% DMSO; *=p<0.05; ****=p<0.0001; for list of genes, primers and assays used, a detailed explanation is presented in Suppl. Table 4 Figure 4C. (B) Additionally, the quantification of the western blot images (MYH11 normalized to ß-actin) is presented for all conditions (control treatment 6h/48h and lenvatinib 6h/48h). (C) Representative images of collagen matrices from the collagen contractility assay after treatment with lenvatinib reveals a time-dependent contraction of the matrix from 0-24h. (D) The gene expression analysis at 24h time point of the contractility assay shows upregulation of contractile VSMC genes (i.e. ACTA2, TAGLN and MYH11) upon lenvatinib treatment (*=p<0.05; mean ± SEM) A subset of these genes is presented in Figure 4.
Suppl. Figure 10: ERK1-2/pERK1-2 complete Western Blot images: (A) The two membranes show the staining for pERK1-2/ERK1-2 for all 3 AAA patient-derived cells (#1-3) and the healthy donor control cells. Total ERK1-2 and ß-actin as loading control are shown as an inlay. One cell isolate (patient derived cells #1) is highlighted and presented in Figure 4D. (B) Additionally, the quantification of the Western Blot images (pERK1-2 normalized to ERK1-2) is shown for all conditions (control 6h/48h and lenvatinib 6h/48h). Figure 11: PPE-AAA in Yucatan minipigs: (A) 12-months-old LDLR-/-(bottom) and wildtype (top) Yucatan minipigs demonstrating an obese phenotype upon gross appearance. (B) Pigs are placed on their right side and a left retroperitoneal approach exposes the infrarenal aorta from the lower renal artery ($) to the aortic trifurcation. After ligation of the lumbar arteries, a Satinsky clamp (arrow) is applied, and the isolated segment is pressure-perfused with elastase via a blunt 5G needle (#) secured by an additional vessel loop tourniquet (asterisk). For the angioplasty (7 days after AAA induction), femoral access is achieved by a longitudinal incision into the right groin (bottom image), and a 5F introducer sheath is inserted after distal ligation of the vessel (top image). (C) Abdominal B-mode ultrasound is used to measure diameters of the infrarenal aorta, based on leading-edge technique at baseline (top), and after 1 and 4 weeks after aneurysm induction. (D) HE staining indicates non-calcified plaque formation with foam cells and cholesterol depositions in the endothelium and media of LDLR -/compared to wildtype aorta. Note the cell-rich infiltrate at the border of plaques. (E) Intraluminal saline perfusion (grey) instead of PPE (dark blue) does not lead to aortic diameter enlargement. Endovascular control (DMSO as vehicle only) coated balloon (light blue) treatment at day 7 after AAA induction showed a trend towards smaller diameter, however, only lenvatinib coated balloon angioplasty (red) resulted in a significant diameter reduction compared to PPE (*=p<0.05; mean ± SEM; PPE AAA induction at day 0; scale bar 200μm; Adv=adventitia; Lu=lumen; Med=Media).

Suppl.
Suppl. Figure 12: Minipig low and high magnification double IF and conventional IHC: (A,B) Porcine double staining for aSMA and MYH11 of the aortic media shown at 20x magnification in relation to higher magnifications using digital zoom (Lumen always oriented to the right). The high magnification images are also shown in Figure 6 B. (C, upper panel) Aortas of LDLR-/-minipigs present a wall structure with parallel orientated elastic fibers in the media and plaque formation in the intima/media (HE and EvG stains) (C, middle panel) VSMCs in the media are positive for aSMA and MYH11, without indicating CD31 positivity. Upon PPE-AAA induction, elastic fibers become disrupted, and the parallel-layered structure of the media is vanished. Vast positivity for aSMA remains, whereas only a few cells show MYH11. (C, highlighted window) CD31-positive neovessels are detected in the media and the thickened adventitia. (C, lower panel) In animals with PPE-induced aneurysms and local lenvatinib-coated balloon treatment, a more organized and parallel-layered structure is recovered. Elastic fibers remain disrupted, but cells show positivity for aSMA and MYH11. CD31 positivity is not observed (magnification 5x; scale bar 100μm) Suppl. Figure 13: Complete Western Blots and quantification from primary porcine SMC culture and pig tissue lysates: The two membranes show the staining for MYH11 (255kDa) after 24/48h incubation of primary porcine SMCs with lenvatinib or control treatment, respectively (A). ß-actin (45kDa) is used for normalization and presented as an inlay. The 48h-time point is highlighted and presented in Figure 5F. (B) Additionally, the membranes and quantification of the staining intensity comparing MYH11 and ß-tubulin (55kDa) in lysates from pigs treated with the PPE + Lenva-DEB versus PPE only show a trend of higher MYH11 protein abundance. This membrane is also shown in Figure 5E.

Online data supplement -Supplementary Material and Methods
Mouse PPE experiments: S. study approval section for ethics declaration. Anesthesia and PPE Procedure: Anesthesia was induced with isoflurane, while analgesia was provided with Temgesic (4mg/kg) before and after surgery. The aorta was isolated, and a temporary proximal and distal sutures were placed. Aneurysms were then induced with porcine pancreatic elastase (PPE; SigmaAldrich, Stockholm, Sweden) at a final concentration of 2 U/mL for 10 min by inserting a microcatheter into the aortic lumen. The aorta was then flushed with warm saline, and a single knot suture (10-0) was placed to seal the catheter entry site. After layered closure of the abdomen, the animal was allowed to recover on a heated mat. Murine ultrasound: B-mode ultrasound was performed under general anesthesia with isoflurane using a Vevo 770 system and a Vevo Imaging Station (both Visualsonics, San Antonio, TX, USA). Our protocol for diameter assessment has been described before.(1) (detailed ultrasound data in Suppl. Table I) Euthanasia and tissue processing: In all experiments presented in this study, animals were sacrificed 4 weeks after aneurysm induction using CO 2 . Operations and images were performed using the operating microscope LEICA EZ4HD with a built in camera and the Leica Acquire Version 3.1 software (both Leica Microsystems, Buffalo Grove, IL, USA). Tissue samples were collected at the time of animal sacrifice. The dilated part of the aorta was carefully dissected and then cut in half. One part was kept for formalin fixation for histochemistry, and the second part was placed in RNAlater (SigmaAldrich, Stockholm, Sweden) and stored at -80°C awaiting further analysis. Control tissue was non-dilated infrarenal aorta from an additional six animals (also 10 week old male C57BL/6J wild type mice (Taconic, Biosciences, Hudson, NY, USA) sacrificed explicitly for this purpose.

Mouse AngII experiments:
S. study approval section for ethics declaration. Anesthesia and AngII Procedure: 18 mice were purchased from Charles River (Sulzfeld, Germany). Ang-II (BACHEM, Switzerland) was dissolved in saline for the required dose of 1000 ng/kg/min at a pump distribution rate of 0.25 µl/h. The Alzet osmotic pumps (Model 2004, DURECT Corporation) were filled with Ang-II solution and conditioned over night at 37°C in saline. Animals were anesthetized and the Alzet osmotic pump was inserted into a subcutaneous pocket lateral of the spine, with the flow moderator facing downwards. Murine ultrasound: For ultrasound measurements of aneurysms, the following settings of the Vevo 2100 Imaging System (FUJIFILM VisualSonics Inc., Toronto, ON, Canada) were applied: gain 30 dB, image depth 9 mm, image width 8 mm. Respiratory gating was set to 25% delay and a window of 50%; T1 50 ms were used for ECG trigger. After localization of the left renal artery, the MS550 transmitter was moved 6 mm cranially for automated scan of the suprarenal aorta. 157 imaging frames were generated over a scan distance of 12 mm with 0.076 mm step size. The suprarenal aortic volume was calculated in mm3 (over the monitored distance of 12 mm) with Vevo Lab 3.1.1 software; images of ECG-gated kilohertz visualization were used to additionally determine the maximum diameter of the aorta (inner-to-inner wall) at maximal blood flow as described before.(2) Independent measurements by three observers were averaged in volume and diameter calculations (Suppl. Fig.2). Euthanasia and tissue processing: please see section above on PPE experiments.

Mouse RNA extraction:
Approximately 1mm long pieces of mouse aortic tissue were homogenized in Qiazol using Multi-Gen probes (Pro-Scientific, Stockholm, Sweden). RNA extraction was performed using the Qiagen miRNeasy Micro Kit (Qiagen, Sollentuna, Sweden) following the manufacturer's protocol. Total RNA concentration was assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). In the MTA analysis (Affymetrix, Santa Clara, CA, USA), samples from n=6 PPE, n=7 PPE+Lenvatinib, and n=5 control aortas were included with RIN numbers between 5.7-9.8. The MTA procedure was performed by the Center for Molecular Medicine (CMM) core facility at Karolinska Institutet, Stockholm, Sweden, according to the manufacturer's protocol.
Mouse Immunohistochemistry: 5µm sections were stained for different targets with an overnight protocol with primary antibodies (Suppl. Table 3). Endogenous peroxidase activity was quenched with hydrogen peroxide for 10 min (Merck, Darmstadt, Germany). Target staining was done with the Vectastain ABC and the Peroxidase Substrate kit (both Vector Laboratories, Burlingame, CA, USA) after species-specific secondary antibody incubation. Counterstaining was performed using Mayer's hematoxylin (Carl Roth, Karlsruhe, Germany). Negative control, including incubation with phosphate buffered saline instead of primary antibody, was done for each antibody. Slides were then imaged with a Keyence BZ9000 microscope (Keyence, Kyoto, Japan) and scanned with a NanoZoomer 2.0-HT Digital slide scanner C9600. Pictures were analyzed with the NDP.view2 software (both Meyer Instruments, Hamamatsu, Japan). Image analysis was performed with AZ AnalyserII software (Keyence), as well as Imaris (Oxford Instruments, Zurich, Switzerland) and Fiji ImageJ for high power field (HPF) analysis of cell counts. Mouse OCT frozen sections were fixed in 4% paraformaldehyde PBS buffered solution and incubated with primary Ki-67 antibodies (Abcam, ab16667) 1:200 overnight. To visualize staining TSA Plus Cyanine 5 Evaluation Kit (PerkinElmer, NEL745E001KT) was applied in 1:50 dilution. Sections were mounted with ProLong Gold (ThermoFisher), and images were taken using a confocal Leica SP5 microscope. For each mouse aorta, 2 slides were stained for the respective antibody and out of these, the most representative image was chosen for presentation. Minipig PPE-AAA experiments: S. study approval section for ethics declaration. Anesthesia: The basic anesthesia regime included: Sedation by combination of ketamine 15mg/kg, azaperone 2mg/kg and atropine 0.1mg/kg intramuscularly. An intravenous catheter was placed in the lateral auricular vein. Induction of anesthesia by propofol (4-8 mg/kg) and oral intubation with a 7.0-8.5 mm cuffed endotracheal tube. Anesthesia was maintained by continuous infusion of propofol (2.5-7mg/kg/h). Pigs were mechanically ventilated in volume-controlled mode (6-10ml/kg) with 40-50% oxygen (1-2l/min) and breathing rate 11-13x/minute. Parameters were adjusted to maintain normocapnia (end-tidal CO2: 35-45mmHg). Saline was infused at a maintenance rate of ~10 ml/kg/h. Intraoperative monitoring included reflex status, heart rate and peripheral arterial oxygen saturation. External warm-air supply was provided. A single-shot cefuroxim (750mg) was administered before skin incision. For multimodal analgesia during the surgery, metamizole (50mg/kg), carprofen (4mg/kg) and fentanyl (0.001-0.01mg/kg) were administered. Ultrasound: Ultrasound was performed immediately after anesthesia before any further procedure by 2 independent examiners (A Busch and/or C Knappich) using a GE Logiq S7 system equipped with a 4-5MhZ abdominal ultrasound probe (GE, Frankfurt, Germany). The aortic trifurcation and the renal arteries were identified and aortic diameter was acquired using the leading edge method in transverse and longitudinal sections at the point of maximum diameter (Suppl. Figure 9C; Suppl. Table II).

(3)
Postoperative analgesia: After each surgery, carprofen (4mg/kg) orally for at least 5 days and a single dose of buprenorphine (0.005-0.1mg/kg) shortly before the end of anesthesia were given to each pig. After the first surgery (induction of aneurysm), the administration of buprenorphine was continued for at least 2 days.
Minipig tissue collection and processing: Directly after euthanasia, the incisional wounds were reopened and access to the infrarenal aorta was gained in order to identify the area of the aneurysmatic lesion. Samples were collected from the AAA, the surgically untouched pararenal aorta (no additional animals were sacrificed for control tissue sampling), the common carotid, the iliac and the coronary arteries, the aortic valve, the left ventricle myocardium, a jugular lymph node, liver, kidney, spleen and lung parenchyma, as well as whole blood in EDTA and serum. All samples were divided for RNA isolation and snap frozen in liquid nitrogen and formalin fixation.

Primary human AAA cells RNA isolation and qPCR:
Cells were placed in 6-well plates (triplicates each) and treated with 0.1µM lenvatinib in OptiMEM (Thermo Fisher, USA, Lenvatinib-treatment) or 0.1% DMSO in OptiMEM (control-treatment) for 6, 24 and 48h. Cells were washed with PBS and harvested with 350µl RLT Lysis Buffer (Qiagen, Netherlands). Total RNA was isolated using the RNeasy Mini Kit (Qiagen, Netherlands) according to manufacturer´s instruction. RNA concentration and purity was assessed using a NanoDrop system. Next, first strand cDNA synthesis was performed using the High-Capacity-RNA-to-cDNA Kit (Applied Biosystems, USA), following the manufacturer's instruction. Quantitative real-time TaqMan PCR was then performed using primers for the following genes: Tagln, Cnn1, Myocd, Itga8, Col3A1, Cald1, S100A1, Kdr, Vcan, Hif1a, Acta2, Smtn and Mmp2 (detailed primer description in Suppl. Table 4). qPCR was run on a QuantStudio5 Cycler (Applied Biosystems, USA), using 384 well plates. Gene expression was normalized to Rplp0 and GAPDH, and quantified with the 2^ΔΔCt method. For Myh11, SYBR Green based quantitative real-time PCR was performed. PCR was run on a QuantStudio5 Cycler, using 384 well plates. Gene expression was normalized to Rplp0 and GAPDH and quantified with the 2^ΔΔCt method.

Cytoskeleton staining (integrity) of primary human cells:
Human primary SMCs isolated AAA biopsies (from three different patients) were stained by immunofluorescence for the typical SMC markers (SM22-SMA-CNN1) as well as for F-actin with phalloidin and compared to commercially available SMC (from healthy donor  Table 3), following the protocol described previously for minipig sections. Slides (including immunohistochemistry) were scanned with an Aperio AT2 (Leica, Wetzlar, Germany), and images were taken with the Aperio ImageScope software (Leica).

Human immunofluorescence (IF) double staining:
4µm sections of paraffin-embedded human aortic samples were mounted on precoated SuperFrost Plus slides. Antigen retrieval and blocking of peroxidase activity was performed as described previously for IHC. In addition, slides were blocked with 5% horse serum for 30 min and all antibodies were diluted in 5% horse serum. Two primary and appropriate secondary antibodies (Suppl . Table IV) were added subsequently to the slides, with the primary antibody of interest (MYH11) being incubated overnight at 4°C, and the following primary antibody for co-staining (aSMA) for 1h at room temperature. Secondary antibodies were incubated for 1h each. Autofluorescence quenching, counterstaining, and imaging were performed as described in the pig double IF section.

Human tissue RNA extraction and qPCR:
Tissues were cut in ~50mg pieces on dry ice. Tissue was homogenized in 700µl Qiazol lysis reagent and total RNA was isolated using the miRNeasy Mini Kit (Qiagen, Netherlands) according to manufacturer´s instruction. RNA concentration and purity were assessed using the NanoDrop system described above. RIN number was assessed using the RNA Screen Tape (Agilent, USA) in the Agilent TapeStation 4200. Next, first strand cDNA synthesis was performed using the High-Capacity-RNA-to-cDNA Kit (Applied Biosystems, USA), following the manufacturer's instruction. Quantitative real-time TaqMan PCR was then performed using primers for the following genes: Tagln, Cnn1, Myocd, Itga8, Col3A1, Cald1, S100A1, Kdr, Vcan, Hif1a, Acta2, Smtn and Mmp2 (Suppl.  ). The blots were blocked with 5% BSA in Tris-buffered saline+0.1% Tween-20 for 1h, followed by overnight incubation with the primary antibody against the un-phosphorylated Kinase in TBS-T+5%BSA. After washing with TBS-T, blots were incubated with anti-mouse and rabbit HRP (horseradish peroxidase)-conjugated secondary antibody and visualized using Enhanced Chemiluminescence (Pierce Fast Western Blot Kit ECL Substrate, Thermo Fisher, USA) in the Intas ECL Chemocam Imager (Intas, Germany) using the Intas ChemoStar Imager Software (Intas, Germany). Blots were stripped using Restore Plus Western Blot Stripping Buffer (Thermo Fisher) and incubated overnight with the primary antibody against the phosphorylated Kinase in TBS-T+5%BSA. After washing with TBS-T, blots were again incubated with anti-mouse and rabbit HRP (horseradish peroxidase)-conjugated secondary antibody and visualized using Enhanced Chemiluminescence (Pierce Fast Western Blot Kit ECL Substrate, Thermo Fisher, USA) in the Intas ECL Chemocam Imager (Intas, Germany), using the Intas ChemoStar Imager Software (Intas, Germany). The blots were quantified using Fiji ImageJ Software.
Western Blot for human MYH11: 10µg of protein of each sample was denatured and reduced at 70°C for 10min, then separated in a NuPage 3-8% Tris Acetate-Gel and transferred onto Trans-Blot Turbo Mini-Size LF-PVDF Membranes (BioRad, USA). The blots were blocked with 5% BSA in Tris-buffered saline+0.1%Tween-20 for 1h, followed by overnight incubation with the primary Rabbit Anti-smooth muscle Myosin heavy chain 11 antibody (ab53219, Abcam, United Kingdom) in TBS-T + 5% BSA.
After washing with TBS-T, blots were incubated with anti-mouse and rabbit HRP (horseradish peroxidase)-conjugated secondary antibody and visualized using Enhanced Chemiluminescence (Pierce Fast Western Blot Kit ECL Substrate, Thermo Fisher) in the Intas ECL Chemocam Imager (Intas, Germany) using the Intas ChemoStar Imager Software (Intas, Germany). Blots were stripped using Restore Plus Western Blot Stripping Buffer (Thermo Fisher, USA) and incubated with the loading control primary antibody against ß-Actin (A1978-200µl, Sigma Aldrich, USA) in TBS-T+5% BSA for 1h. After washing with TBS-T, blots were again incubated with anti-mouse and rabbit HRP (horseradish peroxidase)-conjugated secondary antibody and visualized using Enhanced Chemiluminescence (Pierce Fast Western Blot Kit ECL Substrate, Thermo Fisher, USA) in the Intas ECL Chemocam Imager (Intas, Germany), using the Intas ChemoStar Imager Software (Intas, Germany).
Porcine Tissue: Protein-Isolation, Concentration Measurement and WB: AAA tissue of 25mM Lenvatinib-treated (n=4) and control-treated pigs (n=3) was cut in ~50mg pieces on dry ice. Tissue was homogenized in 200µl Tissue Extraction Reagent I (Thermo Fisher, USA). After homogenization with the Bio-Gen PRO200 Homogenizer and Multi-Gen 7XL Probes (Pro Scientific, USA), samples were centrifuged for 20 min, 14.000 rpm at 4°C and the supernatant was frozen down at -80°C in aliquots of total protein lysate. Total protein concentration was measured using the Pierce BCA Protein Assay Kit (Thermo Fisher) following the manufacturer´s instruction. 15µg of protein of each sample was denatured and reduced at 70°C for 10 minutes, then separated in a NuPage 3-8% Tris Acetate-Gel and transferred onto Trans-Blot Turbo Mini-Size LF-PVDF Membranes (BioRad, USA). The blots were blocked with 5% BSA in Tris-buffered saline+0.1% Tween-20 for 1h, followed by overnight incubation with the primary Rabbit Anti-smooth muscle Myosin heavy chain 11 antibody (ab53219, Abcam, United Kingdom) in TBS-T+5%BSA. After washing with TBS-T, blots were incubated with anti-mouse and rabbit HRP (horseradish peroxidase)-conjugated secondary antibody and visualized using Enhanced Chemiluminescence (Pierce Fast Western Blot Kit ECL Substrate, Thermo Fisher, USA) in the Intas ECL Chemocam Imager (Intas, Germany) using the Intas ChemoStar Imager Software (Intas, Germany). Blots were stripped using Restore Plus Western Blot Stripping Buffer (Thermo Fisher, USA) and incubated with the loading control primary antibody against ß-Tubulin in TBS-T+5%BSA for 1h. After washing with TBS-T, blots were again incubated with anti-mouse and rabbit HRP (horseradish peroxidase)-conjugated secondary antibody and visualized using Enhanced Chemiluminescence (Pierce Fast Western Blot Kit ECL Substrate, Thermo Fisher, USA) in the Intas ECL Chemocam Imager (Intas, Germany), using the Intas ChemoStar Imager Software (Intas, Germany).