Anti-PCSK9 antibodies inhibit pro-atherogenic mechanisms in APOE*3Leiden.CETP mice

LDL-cholesterol (LDL-C) is a causal pathogenic factor in atherosclerosis. Monoclonal anti-proprotein convertase subtilisin/kexin type 9 (PCSK9) neutralizing antibodies are novel potent LDL-lowering drugs which reduce cardiovascular events. To characterize their effect on atherogenesis, APOE*3Leiden.CETP mice were fed a high cholesterol/high fat diet (WTD) or normal chow (NC) for 18 weeks. Mice on WTD were injected with the human anti-PCSK9 antibody mAb1 (PL-45134, 10 mg*kg−1 s.c.) or 0.9% saline every 10 days. PCSK9 inhibition decreased total cholesterol in serum of APOE*3Leiden.CETP mice and prevented the development of atherosclerosis. The plaque area in the aortic root was reduced by half and macrophage infiltration determined by Ly6c and Mac-3 staining was ameliorated. PCSK9 inhibition decreased markers of inflammation in mononuclear cells (Il-6, Tnfa mRNA), and in serum (CXCL-1,-10,-13; complement factor C5a) compared to control WTD fed animals. The number of circulating Sca-1/VEGF-R2 positive endothelial progenitor cells of the peripheral blood and spleen-derived diLDL/lectin double positive circulating angiogenic cells was increased. To conclude, the PCSK9-mediated anti-atherosclerotic effect involves the upregulation of pro-regeneratory endothelial progenitor cells, a reduction of inflammation and change of plaque composition.

Circulating endothelial progenitor cells (EPC) derived from the bone marrow and are released into the blood to form new blood vessels or to aid in the repair of damaged vessels [17][18][19][20] . After recruitment to sites of ischaemia and endothelial damage, EPC integrate into the endothelium and positively regulate endothelial cell growth and neovascularization 19,21,22 . Inflammation, metabolic and vascular risk factors and type 2 diabetes have been shown to downregulate EPC numbers and impair their function [22][23][24][25][26] . In contrast, statin-mediated lipid lowering or physical activity improve EPC numbers and cell function [27][28][29] . In a large prospective clinical study, CD34/ KDR-positive cells were shown to predict cardiovascular events 18 . Within the EPC population, circulating angiogenic cells (CAC) can be distinguished 19 . CAC derive from myeloid haematopoietic cells and share features with monocytes and exert their angiogenic effects via paracrine and signaling mechanisms.
Here, we aimed to characterize the vascular effects of PCSK9 inhibition in APOE*3Leiden.cholesteryl ester transfer protein (CETP) mice, a mouse model for familial dysbetalipoproteinemia with human-like lipoprotein metabolism [30][31][32] , on atherosclerosis development, inflammation and EPC and CAC number.

Discussion
The present study demonstrated that the treatment with the anti-PCSK9 antibody mAb1 reduced atherosclerosis development, macrophage infiltration and cardiovascular inflammation and increased the number of EPC and CAC in APOE*3Leiden.CETP mice fed a Western type diet.
In line with previous studies, we found that inhibition of PCSK9 reduces serum cholesterol levels and further attenuates atherogenesis and plaque inflammation 13,[36][37][38][39] . Studies in Pcsk9 knock-out mice showed increased hepatic LDLR levels that were linked to reduced circulating LDL cholesterol 40 . The association between serum cholesterol level and atherosclerotic plaque size observed in the present study indicates a causal link of serum cholesterol and plaque formation 39 . Furthermore, there is a strong association between the severity of coronary atherosclerosis and adverse cardiac events 41,42 . The prevention of atherogenesis in anti-PCSK9 antibody treated mice was accompanied by reduced collagen deposition. This is in accordance with findings of Nicholls et al. who found no alterations in fibrous components in atheroma after evolocumab therapy in coronary artery disease patients 41 . However, an in vivo study with alirocumab has shown improved lesion composition in atherosclerotic mice by increasing collagen content 39 .
In addition to the reduction in plaque lipid load, the present study demonstrated that PCSK9 inhibition also affects plaque composition by reducing the amount of activated monocyte-derived macrophages in the atheroma. This suggests that the LDL-C lowering effect by PCSK9 inhibition may also lead to reduction in monocyte transmigration through the arterial wall as well as to reduced differentiation into mononuclear phagocytes. Those phagocytes can transform into cholesterol-laden "foam cells" as they ingest the accumulated normal and modified lipoproteins. Alirocumab therapy was further shown to ameliorate monocyte adhesion and trafficking as well  www.nature.com/scientificreports www.nature.com/scientificreports/ as lesion composition in APOE*3Leiden CETP mice 39 . This is in line with our findings that PCSK9 inhibition downregulates chemokine levels, such as CXCL-1, -10, -13, which can be expressed by activated endothelial cells, smooth muscle cells as well as emigrated leukocytes. They are key players involved in the recruitment and infiltration of immune cells into the vessel wall and exacerbate atherosclerosis progression 43,44 . Besides ameliorated chemokine levels C5a is reduced in PCSK9 antibody treated mice. C5a is one of the most potent inflammatory chemoattractants 45 . It activates endothelial cells and regulates the secretion of various cytokines and chemokines such as Tnfa, IL-6 and CCL-2 -12,-13 46 .
Noteworthy, monoclonal antibodies only target PCSK9 in the circulation, whereas the pathophysiological relevance of intracellular PCSK9, such as in intra-atheroma macrophages, requires better understanding 47 . Further, our study revealed the novel observation that LDL-C lowering with anti-PCSK9-antibody significantly increased the number of circulating EPC and CAC that are considered markers of endothelial and vascular health and are associated with positive clinical outcomes 18 . Additionally, PCSK9 expression is positively associated with apoptosis in vascular endothelial cells, tumor cells, and neurons 48,49 suggesting a detrimental role for PCSK9 on endothelial repair and vasculogenesis. Indeed, Chao et al. 50 reported that PCSK9 serum levels correlated with apoptosis in circulating endothelial cells which may provide the explanation for our observation.
In conclusion, our data show that PCSK9 inhibition prevents the development of atherosclerotic plaques. This effect is mediated by several anti-atherogenic mechanisms that involve the up-regulation of pro-regenerative endothelial progenitor cells, reduction of inflammation in circulating cells, in the serum and in the plaque as well as changes of the plaque composition. These data provide important mechanistic explanations for the observed reduction of clinical outcomes with the human monoclonal antibodies against PCSK9 that have been reported in the meantime 13,14 .

Methods
Animals and anti-PCSK9 antibody treatment. APOE*3Leiden.CETP transgenic mice were fed ad libitum a western type diet (WTD: 21% fat, 19.5% casein, 1.25% cholesterol) or normal chow (NC) (Ssniff, Soest, Germany) for a period of 18 weeks. Mice on WTD were injected with the human anti-PCSK9 antibody mAb1 (PL-45134, 10 mg*kg −1 subcutanously (s.c.), n = 10) or 0.9% saline (control, n = 12) every 10 days for 18 weeks. The antibody was generated as described previously 33,36 and provided by Amgen. Mice were euthanized with ketamine and xylazine 51 . The experiments were approved by the Universität des Saarlandes and complied with national guidelines (directive 63/2010 of the European Parliament) as well as the ARRIVE guidelines for reporting experiments involving animals 52 .
Serum cholesterol quantification. Venous blood was obtained before and after treatment with anti-PCSK9 antibody mAb1 or 0.9% saline. Cholesterol concentration was measured in serum using the standard curve based LabAssay ™ Cholesterol Kit (Wako, Neuss, Germany) according to the manufacturer's instructions.
Staining of frozen aortic root sections. Atherosclerotic plaques and collagen deposition in the aortic root was quantified as described previously 51 . Atherosclerotic plaque area is expressed as % of total aortic sinus area. Picric sirius red staining (0.1%) was used to analyze collagen deposition in total aortic sinus as well as in atherosclerotic plaque area. Pro-inflammatory macrophage infiltration into the arterial wall was determined corresponding secondary ImmPRESS ® anti-rat HRP conjugated antibody (Vector Laboratories, Burlingame, CA, USA) or anti-rat Alexa 594 antibody (Dianova, Hamburg, Germany), respectively. For immunofluorescent α-SMA staining, the primary antibody from Abcam (1:250, Cat.ab5694, Abcam, Cambridge, UK) and the corresponding anti-rabbit Alexa Fluor 647 conjugated secondary antibody (Invitrogen, Carlsbad, USA) were used. Stained sections were fully digitalized at 20 × magnification using a digital slide scanner (Pannoramic Scan II, 3D HISTECH Ltd., Budapest, Hungary). Images of stained tissue slices were captured from slide scanner data sets (Pannoramic Viewer, version 1.15.4., 3D HISTECH Ltd., Budapest, Hungary) in TIFF format at 10 × or 20 × magnification and were quantitated by ImageJ using the plugin color deconvolution as described in 53 . Percentage (%) of area was calculated from grey-scaled pictures.
Quantification of EPC by flow cytometry. Mouse blood was used to characterize and quantitate EPC by flow cytometry as described previously 51 . Mouse equivalent surface markers Sca-1 (FITC conjugated), VEGFR2 (PE-conjugated) 20,54 and appropriate isotype controls (IgG2a κ FITC and PE, BD Pharmingen) were used. FACS analysis was performed on a FACS Calibur instrument (BD) and Cell Quest software version 6.0 (BD Biosciences, Heidelberg, Germany).
Cytokine array. To analyse serum cytokines and chemokines, we used the membrane-based Proteom Profiler Array ™ (R&D, Minneapolis, USA) and followed manufacturer's instruction. Multiple exposure times were applied (1-10 min) and pixel density was quantitated by ImageJ after background substraction and normalization to positive control on the membrane. Data are expressed as fold over WTD fed mice which was set at 1.