Low human and murine Mcl-1 expression leads to a pro-apoptotic plaque phenotype enriched in giant-cells

The anti-apoptotic protein myeloid cell leukemia 1 (Mcl-1) plays an important role in survival and differentiation of leukocytes, more specifically of neutrophils. Here, we investigated the impact of myeloid Mcl-1 deletion in atherosclerosis. Western type diet fed LDL receptor-deficient mice were transplanted with either wild-type (WT) or LysMCre Mcl-1fl/fl (Mcl-1−/−) bone marrow. Mcl-1 myeloid deletion resulted in enhanced apoptosis and lipid accumulation in atherosclerotic plaques. In vitro, Mcl-1 deficient macrophages also showed increased lipid accumulation, resulting in increased sensitivity to lipid-induced cell death. However, plaque size, necrotic core and macrophage content were similar in Mcl-1−/− compared to WT mice, most likely due to decreased circulating and plaque-residing neutrophils. Interestingly, Mcl-1−/− peritoneal foam cells formed up to 45% more multinucleated giant cells (MGCs) in vitro compared to WT, which concurred with an increased MGC presence in atherosclerotic lesions of Mcl-1−/− mice. Moreover, analysis of human unstable atherosclerotic lesions also revealed a significant inverse correlation between MGC lesion content and Mcl-1 gene expression, coinciding with the mouse data. Taken together, these findings suggest that myeloid Mcl-1 deletion leads to a more apoptotic, lipid and MGC-enriched phenotype. These potentially pro-atherogenic effects are however counteracted by neutropenia in circulation and plaque.

Nevertheless, it is now believed that neutrophils are instrumental in plaque development and later destabilization, as we and others have shown [14][15][16] . Warnatch et al. demonstrated that Neutrophil Extracellular Traps (NETs) prime macrophages to produce inflammatory cytokines, resulting in increased atherogenesis 17 . Dissection of the contribution of neutrophils to atherosclerotic plaque progression has however remained difficult, since other cell populations are also affected when neutrophils are experimentally manipulated. Deletion of Interferon Regulatory Factor 8 (IRF8) for example results in neutrophilia which exacerbates atherosclerosis in IRF8 −/− ApoE −/− chimeric mice, but was at the same time accompanied by changes in monocytes and several dendritic cell populations 18,19 . In addition, neutrophils have thus far not been depleted during the complete course of atherosclerosis. Neutrophil depleting antibodies such as the anti-polymorphonuclear leukocyte (PMN) antibody cannot be sustained for more than 4 weeks in mice without affecting monocyte counts 16 .
As Mcl-1 is vital for neutrophil survival, whilst presumably having mild effects on macrophage apoptosis, we hypothesized that myeloid Mcl-1 deletion could serve as an efficient neutropenia model during the full pathogenesis of atherosclerosis. Moreover, although several Bcl-2 family members have been investigated in the context of atherosclerosis [20][21][22][23] , the role of Mcl-1 in disease progression has not been assessed thus far. To this end, we studied effects of specific deletion of Mcl-1 in the lysozyme M expressing myeloid subsets neutrophils and macrophages on early and advanced atherosclerosis using bone marrow transplantation of Mcl-1 fl/fl LysMCre or wild-type (WT) bone marrow into low density lipoprotein receptor-null (LDLr −/− ) mice. Our study shows that myeloid Mcl-1 deletion indeed has a profound impact on neutrophil survival. However, Mcl-1 deficient macrophages were seen to accumulate more lipids in vitro and in vivo and were subsequently more sensitive to apoptosis. Moreover, we show that Mcl-1 is implicated in the fusion of macrophages. These combined effects however counteracted each other affecting only viable cell plaque composition, but not plaque growth.

Mcl-1 deletion altered neutrophil levels and characteristics.
We first quantified Mcl-1 gene expression during atherogenesis. Mcl-1 levels in collar-induced carotid artery lesions of LDLr −/− mice gradually increased during lesion development and in particular in advanced plaques, six weeks after collar induction (Fig. 1A). Of note, this increase was not validated at the protein level. Mcl-1 was mostly expressed in activated macrophages (M1-or M2-macropahges), as compared to other cell types (Fig. 1B). Mcl-1 was also detectable in human atherosclerotic plaques, and its expression did not differ between stable and unstable plaque (Fig. 1C). However, Pearson correlation analysis revealed that Mcl-1 expression did correlate with pathogenic plaque traits, with more traits in unstable plaques, suggesting an involvement in the disease process ( Compatible with the notion that Mcl-1 is essential for neutrophil survival 7 , circulating and splenic neutrophil numbers were sharply reduced by 80% and 86%, respectively in Mcl-1 fl/fl LysMcre mice 9 . Circulating neutrophils were depressed in Mcl-1 −/− chimeras both at baseline (82% depletion) and even more so under hyperlipidemic conditions (91% depletion) ( Fig. 2A). Likewise, neutrophil content in Mcl-1 −/− atherosclerotic lesions was decreased, albeit to a lower extent than in blood (Fig. 2B,C), hinting to an enhanced adhesive capacity or faster turnover of residual neutrophils in circulation. Considering that an elevated CXCR4/ CXCR2 balance is associated with regress to the bone marrow and that CXCR4 is an established measure of neutrophil ageing 16 , we examined neutrophil phenotype. CXCR4 expression on circulating and peritoneal residual neutrophils was increased (Fig. 2D,E, respectively), suggesting hyperactivation and increased SDF1 migratory capacity. Moreover, responsiveness of remaining neutrophils to the potent neutrophil chemokine CXCL1 was blunted, concordant with the reduced CXCR2 expression by pre-apoptotic neutrophils 16 . Peritoneal neutrophil influx 2 hours after i.p. injection of CXCL1 was prominent in WT transplanted mice, whereas Mcl-1 −/− transplanted mice only showed a minor, non-significant, increase in peritoneal neutrophils (Fig. 2F,G). Of note, neutrophil recruitment was paralleled by stromal egress of neutrophils into circulation in WT, but not Mcl-1 −/− mice (data not shown). Taken together, these results confirm Mcl-1 as a crucial neutrophil survival factor, also under hyperlipidemic conditions, and demonstrate that Mcl-1 myeloid deletion can be used as a genetic tool to induce a long-lasting, severe neutropenia in atherosclerosis.
Myeloid Mcl-1 deletion increased plaque apoptotic cell content but did not affect atherosclerotic lesion size. Despite its profound effects on circulating neutrophils, and on plaque neutrophil content, myeloid Mcl-1 deletion did neither alter early nor advanced plaque area, necrotic core size, or plaque macrophage content as compared to controls (Fig. 3A-D). We did observe an increase in plaque apoptosis by 71% and 77% in atherosclerotic lesions of Mcl1 −/− mice fed a WTD for 5 and 10 weeks, respectively, compared to WT mice (Fig. 3E), suggesting that Mcl-1 not only plays an important role in the survival of neutrophils, but also of other myeloid plaque-resident cells, such as plaque macrophages and foam cells.  . All data is presented as mean ± SEM. *p < 0.05. formation in human atherosclerosis, we quantified MGCs in human plaques. Cathepsin K + multinucleated cells were frequently found in both stable and unstable plaques (Fig. 6A,B). Interestingly, their presence inversely correlated with Mcl-1 gene expression in unstable plaque, and only in unstable lesions associated significantly with lesion hemorrhages and calcifications (P value = 0.044 and 0.055, respectively, Fig. 6C,D). Thus, our results unveil a hitherto unknown link between Mcl-1 and MGC formation, which is influenced by hyperlipidemia, an association potentially preserved in human atherosclerotic lesions as well.

Discussion
In this study, we evaluated the effects of myeloid Mcl-1 deletion and its accompanying neutropenia on atherosclerosis progression. In addition to extreme neutropenia, Mcl-1 deficiency resulted in increased macrophage apoptosis and lipid handling, and triggered multinucleated giant cell formation. First, we found that myeloid Mcl-1 deletion dramatically reduced neutrophil numbers both in circulation and in atherosclerotic lesions. This is in keeping with extensive data on the vital role of Mcl-1 in neutrophil survival 9,10,26,27 which as we now Our work thus provides a mouse model for continuous neutropenia, an important advantage to models used in earlier studies addressing neutrophil contribution to atherogenesis. Though Zernecke et al. demonstrated that CXCR4 blockade aggravated atherosclerosis due to increased neutrophil recruitment to the plaque, they were unable to extend the neutrophil depletion beyond a 4 week period 16 . They did show that plaques of neutrophil depleted mice were smaller and had a lower neutrophil and macrophage content, but did not evaluate plaque apoptosis or necrotic core size. Similarly so, CCL3 −/− LDLr −/− bone marrow chimeras with 50% less neutrophils, developed smaller atherosclerotic lesions 30  Although Mcl-1 loss was previously shown to have no effect on monocyte and macrophage development in wild type mice 9,10 , Mcl-1 −/− macrophages were more sensitive to apoptosis upon an infection 12 or phagocytic challenge 10 . We found that the apoptotic cell content in advanced aortic root lesions (10 weeks of WTD) was increased by 44% in mice with myeloid Mcl-1 deficiency. As neutrophils are only scarcely present in advanced lesions 14 and most apoptotic cells were located in the central atheroma (data not shown), the high apoptotic cell density is likely to reflect dying LysM + plaque macrophages. Our work thus identifies Mcl-1 as a major survival protein in atherosclerotic lesions. Atherosclerotic lesion burden was however unaltered in Mcl-1 −/− BM recipients, as were necrotic core, macrophage and collagen content. Similar results were obtained when studying plaque initiation five weeks after WTD. Our results correspond with those from Thorp et al. 21 , who showed increased macrophage apoptosis, but unchanged lesion burden in Bcl-2 flox -LysMCre ApoE −/− mice that are deficient in macrophage and neutrophil Bcl-2 21 . In turn, hematopoietic Bim deficiency, a pro-apoptotic Bcl-2-family member, had no impact on macrophage apoptosis and lesion burden in LDLr −/− mice 23 . Thus, Mcl-1 −/− deletion in macrophages led to higher apoptosis level in advanced plaques, however it is clear from the above that this not always translates into a pro-atherogenic plaque progression.
In addition to an increased sensitivity to oxLDL induced cell death, Mcl-1 −/− macrophages showed augmented lipid accumulation after incubation with oxLDL and VLDL. In keeping, we observed elevated foam cell levels in vivo in the peritoneal cavity of Mcl-1 −/− BM compared to WT BM recipients. These findings seem to contrast with  35 . In addition, Samokhin et al. 36 showed that mice fed a Paigen diet displayed a 4-fold increase in MGC number in atherosclerotic lesions 36 . In our study, we observed a higher macrophage fusion capacity in both Mcl-1 −/− peritoneal and BMDMs, which increased even more upon oxLDL or VLDL stimulation. Furthermore, Mcl-1 −/− deficiency in BMDMs seemed to promote MGC formation, independently of lipid uptake. In line with our in vitro findings, Mcl-1 −/− atherosclerotic plaques had a 4-fold increase in MGC presence as compared to WT lesions. Additionally, the presence of MGCs in human unstable plaques correlated negatively with Mcl-1 gene expression and significantly associated with hemorrhages and calcifications in the lesion. The exact role of MGCs in atherosclerosis is not yet fully understood, however it was previously shown that MGCs facilitate vascular smooth muscle cell migration in the context of atherosclerosis by producing cathepsin K and destroying the elastin fibers 36 . Although not providing conclusive evidence, our findings support such pro-atherogenic role of MGCs. To our knowledge, we are the first to implicate Mcl-1 in the fusion of macrophages. Possibly, this is related to increased oxidative phosphorylation capacity for energy production in Mcl-1 −/− deficient BMDMs (data not shown), however the exact mechanism by which Mcl-1 induces MGC formation remains to be investigated. In

Mcl-1 gene expression during atherogenesis.
Twenty male LDLr −/− mice were fed a Western-type diet (WTD) two weeks prior to surgery and throughout the experiment. Atherosclerotic carotid artery lesions were induced by perivascular collar placement 37 . Carotid Mcl-1 gene expression was analyzed prior to and two, four, six and eight weeks after collar placement (n = 4, per timepoint).The mice were anaesthetized and perfused with phosphate buffered saline (PBS) after which both common carotid arteries were isolated. After dissection of the adventitia, plaque containing segments were excised based on macroscopic examination, snapfrozen and stored at −80°c. Two to three atherosclerotic plaques were pooled per sample and total RNA was isolated using Trizol reagent (Invitrogen). Gene expression was analyzed by real time PCR (qPCR) using ABI PRISM 7700 Sequence Blood cell analysis and flow cytometry. Blood samples were taken by tail bleeding immediately before BMT, prior to (week 0) and after four weeks of WTD feeding (week 4) and at the time of sacrifice (week 5 or week 10). Peritoneal leukocytes were isolated at the time of sacrifice by peritoneal lavage with 10 ml PBS. Whole blood and peritoneal lavage samples were analyzed using a Sysmex blood cell analyzer (XT-2000i). For flow cytometry, WBC and peritoneal leukocytes were stained with fluorescently labelled antibodies against F4/80, CD19, CD4, CD71 and CD11b (eBioscience) and Gr1, CD8 and CXCR4 (BD Pharmingen). Fluorescence-activated cell sorting (FACS) analysis was performed on a FACSCalibur with CellQuest software (BD Biosciences). tissue harvesting and analysis. Two hours before sacrifice, Mcl-1 −/− or WT mice (n = 5) received intraperitoneal injections of CXCL1 (200 ng/ml in 1 ml PBS) or PBS control. The mice were anesthetized and perfused with PBS. Cryosections of the aortic root tissue were stained with hematoxylin and eosin (HE) or Oil Red O. Lesion size was quantified using a Leica DMRE microscope with camera and Leica Qwin Imaging software (Leica Ltd). MGCs were defined as macrophages with two or more round nuclei on the HE slides and quantified by an animal pathologist. Immunohistochemical stainings were performed for macrophage (MOMA-2, Sigma) and vSMC (α-smooth muscle actin, Sigma) content. Apoptotic cell content was quantified using terminal deoxytransferase dUTP nick-end labeling (TUNEL) kit (Roche Diagnostics). Human atherosclerotic plaque collection. Stable and unstable human carotid atherosclerotic plaques segments (classified according to Virmani et al. 40 ) were collected from the same symptomatic patient (n = 22/23) undergoing carotid endarterectomy in Maastricht University Medical Centre (MUMC, The Netherlands) and Zuyderland Medical Center (Sittard, the Netherlands). Collection, storage, and use in the Maastricht Pathology Tissue Collection (MPTC) were approved by medical ethical committee (16-4-181) and in accordance with the "Code for Proper Secondary Use of Human Tissue in the Netherlands" (http://www.fmwv.nl). Atherosclerotic plaque segments were alternatively snapfrozen for RNA and microarray analysis, or formalin-fixed for paraffin-embedding.
Human atherosclerotic plaque histology. The human plaque sections adjacent to the snapfrozen segment were classified to determine plaque type according to Virmani et al. 40 . HE staining was used to quantify plaque size, lipid core size and hemorrhage. Alizarin red staining was done to measure the percentage of calcification.
Human atherosclerotic plaque immunohistochemistry. All stainings were performed on adjacent plaque sections for vascular endothelial marker CD31 (Dako), macrophage marker CD68 (Dako), T-cell marker CD3 (Dako) and MGC marker (cathepsin K 41 ). MGCs were defined as cathepsin-K positive cells having 2 or more round nuclei. MGC were quantified in all sections and averaged per patient. Relative abundance of MGC was calculated by dividing the number of MGC by that of CD68 positive cells.
Human atherosclerotic plaque RnA extraction and transcriptomics. RNA was isolated by Guanidium Thiocyanate lysis followed by Cesium Chloride gradient centrifugation, and then purified using the Nucleospin RNAII kit. 750 ng of biotinylated cRNA per sample was hybridized to Illumina Human Sentrix-8 V2.0 BeadChip ® and washed according to the Illumina standard procedure. Scanning was performed on the Illumina BeadStation 500. Raw expression data were extracted from the images using default settings and without normalization.
Statistics. Values are expressed as mean ± standard error of the mean (SEM). All statistical analyses were performed using Prism (GraphPad Software). Statistically significant differences (p < 0.05) were evaluated using the Student's t-test unless stated otherwise. Pearson correlation analysis was performed to assess the