Typically, neutrophils were viewed as a homogenous short-lived population of terminally differentiated cells with limited transcriptional capacities, possessing specific killing abilities against invading pathogens. In recent years however, mounting evidence demonstrates that neutrophils are a heterogenous group of cells, presenting vast phenotypic and functional heterogeneity (1). Along this line, we described for the first time a new and novel subpopulation of blood derived neutrophils which spontaneously develop in prolonged culture conditions (2–4). Freshly isolated blood neutrophils in culture are known to have a very short lifespan among leukocytes and are constitutively committed to apoptosis. Notably, in culture, most neutrophils become apoptotic within 24 hrs (5, 6). Yet, by maintaining neutrophils in prolonged culture conditions, without additional growth factors or cytokines, a small number of Annexin-V negative/LC3B+ neutrophils remain viable and display an extended life span. Within 3 days they become immensely vacuolated, acquiring an excessively large size by internalizing granules, cellular residues and debris of non-viable neutrophils, thus, becoming giant phagocytes (Gφ) (2, 3). Moreover, unlike other types of giant phagocytic cells which are multinucleated heterokarions, arising from cell-cell fusion, these neutrophil-derived Gφ have a single nucleus (7). They also differ from monocytic macrophages by their loose attachment to the surface of the tissue culture dish and their movement abilities in culture differ from those of macrophages, as demonstrated by time lapse-photography (2). In addition to these unique morphological features, these cells are devoid of monocytic (CD14, CD163, CD16) or dendritic (CD1c, CD141) cells markers while exhibiting phenotypic markers and functions of neutrophils. Their neutrophilic origin is evident by the expression of the specific neutrophil granules marker - CD66b, azurophilic granules markers - CD63, and myeloperoxidase (MPO). Also, Gφ express additional neutrophilic markers including neutrophil elastase (NE), CD11b and CD15. Functionally, these Gφ take-up latex beads and opsonized zymosan, express the NADPH oxidase gp91-phox and p22-phox subunits and generate reactive oxygen species (ROS) in response to stimulation with PMA (2–4). However, unlike freshly harvested neutrophils, Gφ also intensively express the scavenger receptors CD68 and CD36, and take-up oxidized low density lipoprotein (oxLDL). An additional unique feature of Gφ in culture is their ability to generate ROS in response to stimulation with oxLDL. The presence of LC3B-coated vacuoles and LC3B aggregates in Gφ indicate that autophagy is a constitutive trait in these cells.

Treating Gφ with various inhibitors affected Gφ development in culture. Inhibiting NADPH oxidase by diphenyl iodide (DPI), ROS production by N-acetylcysteine (NAC), phagocytosis by cytochalasin B, or inhibiting PI3K pathway - all abolished their formation in culture. Additionally, inhibiting autophagy or culturing neutrophils in the presence of GM-CSF/IL-4 also abolished their development. These latter findings indicate that functional phagocytosis, NADPH-oxidase, ROS signaling as well as autophagy are essential for their development in culture conditions (2, 4).

However, based on their phenotype and functions studied thus far, it is unclear whether neutrophil-derived Gφ bear any significant relevance to physiological or pathological conditions in vivo.

The human carotid plaque represents an in vivo easily attainable atherosclerotic tissue which enables to analyze the composition of inflammatory cells in human arterial disease. In earlier studies, most cellular analyses of carotid plaques primarily focused on macrophage foam cells and T cells. However, in the last decade the contribution of neutrophils to atherosclerotic plaque destabilization and rupture emerged as an important factor (8, 9). In a comprehensive study by Ionita et al. high neutrophil numbers were associated with a large lipid core in human atherosclerotic carotid plaques and more so in symptomatic patients. These increases in neutrophil numbers were also associated with histopathologic features of rupture-prone atherosclerotic lesions, suggesting plaque destabilization (10). Hosokawa et al. identified infiltration of esterase activity and CD66b+ neutrophils in seven human atherosclerotic carotid plaques which appeared morphologically activated and expressed oxidant activity (11). Of note, Naruko et al. have shown that neutrophil infiltration in plaques was actively associated with acute coronary events. Moreover, all patients who had died from acute myocardial infarction (AMI) had neutrophils within their plaques, whereas in coronary lesions of patients who had died of non-cardiovascular disease, neutrophils were extremely rare (12).

In the current study we sought to investigate whether this cell phenotype that spontaneously developed in culture conditions, might be present in human arterial atherosclerotic disease, such as the carotid plaques of patients undergoing endarterectomy, and might be associated with inflammatory/anti-inflammatory conditions in vivo. Neutrophils are difficult to study, particularly in tissues, due to their propensity for activation and trans-differentiation (13, 14). Nonetheless, the presence of cells resembling neutrophil-derived Gφ was confirmed based on their specific CD markers combined with their morphology. These cells, like Gφ in culture, expressed the specific neutrophil marker CD66b, neutrophil elastase (NE) myeloperoxidase (MPO) and CD63. Additional markers which were previously identified in Gφ in culture included CD68 and CD36 scavenger receptors, 3-nitrotyrosin (3-NT) and lipids. In parallel, because of their foam cell morphology monocytes/macrophages were also identified by their exclusive CD163 marker. Due to the morphological similarities between CD66b+ and CD163+ expressing cells, CD66b+/CD163+ co-localization experiments were performed. In several plaques CD66b/CD163 co-localization was not observed, in other plaques a low co-localization was noted, and in few plaques CD66b+/CD163+ co-localization was high. Thus, indicating the presence of several populations of CD66b/CD163 expressing cells in human carotid plaques.

It is commonly known that neutrophils are short-lived in ex vivo conditions (18). Yet, in earlier studies we described the spontaneous development of a small sub-population of neutrophil-derived Gφ with extended life span in culture (2, 3). We were intrigued by their vastly enlarged size observed in vitro after a few days in culture. Thus, we divided fresh neutrophils into two sub-groups, each labeled with a different PKH stain (green or red), mixed them together and followed their fate daily, microscopically. As expected, most of the neutrophils died within 24 hrs., however, the few neutrophils that remained viable (0.01- 0.02%), avidly internalized remnants of dead neutrophils and cellular debris. Subsequently, transforming within 3-5 days in culture from classical small-sized neutrophils with a multi-lobed nucleus into large phagocytic cells with a rounded nucleus. They also internalized oxidized LDL, zymosan, and latex upon stimulation, and generated ROS. These newly developed neutrophils resembled macrophage-foam cells morphologically but expressed neutrophil specific markers (CD66b, CD63, NE, MPO), and were negative for monocytic or dendritic markers. Importantly, FACS analysis revealed that the cultures consisted of pure neutrophils. Moreover, by mixing and co-culturing autologous neutrophils and monocytes, fewer Gφ developed in culture.

In the current study we sought to investigate the potential relevance of Gφ to inflammatory-atherosclerotic conditions in vivo , using the human carotid plaque as a model. Several important findings emerged from this study. First, the presence of CD66b+ neutrophils and 163+ macrophages was observed in lipid-rich areas of the plaque. Second, Both CD66b+ and CD163+ expressing cells also co-expressed CD68 scavenger receptors and both had similar morphology, size, and lipid content which is characteristic of foam cells. Third, CD66b+/CD163+ neutrophil/macrophage co-localization was also noted. This co-expression varied widely between the plaques, ranging from zero in some of the plaques, to low in most plaques. A high CD66b+/CD163+ co-localization was also found in few plaques. Fourth, the CD66b+ cells in the plaques also co-expressed additional specific neutrophil markers visualized by confocal immunofluorescence double staining. They co-localized with NE, MPO, and CD63 (data not sown), as well as with CD36, which is another scavenger receptor that was previously shown to co-localize with CD66b+ cultured Gφ. Fifth, CD66b+ neutrophils and CD163+ macrophages co-expressed oxidative-nitrosative stress activity detected by co-localization of 3-NT+/NE+ and 3-NT+/CD163+ cells, respectively. Sixth, the presence of VEGF was demonstrated by neutrophils co-expressing CD66b+/VEGF+ and macrophages co-expressing CD163+/VEGF.

We hypothesize that CD66b+ cells which did not co-localize with CD163+, resemble the Gφ which developed in pure neutrophil cultures and represent a type of an inflammatory/anti-inflammatory-activated neutrophil phenotype. Yet, other neutrophils that co-expressed CD66b+/CD163+ to varying degrees, exhibit neutrophil/macrophage plasticity and trans-differentiation to monocytes/macrophages or monocytes/macrophages hybrids. Evidently, these hybrid cells might have combined neutrophil-macrophage properties. The latter co-localization findings, in particularly, demonstrate the complexity and the difficulties in studying intricate tissues such as carotid plaques, and likely each plaque represents a different stage in development and severity.

The neutrophil is the most abundant white blood cell in the circulation, yet, it received little attention in the pathophysiology of atherosclerosis (19, 20). However, in the last decade the evidence is gradually mounting as to its involvement and significance in various stages of plaque formation, development, and rupture (18, 21). In earlier studies, CD66b+ neutrophils’ localization within plaques, and their numbers were strongly associated with histopathologic features of rupture-prone plaques, implicating their involvement in plaque destabilization (10). A massive neutrophil infiltration was found in culprit lesions of autopsy specimens from acute coronary syndrome (ACS) patients as compared to the rare neutrophil infiltration in patients who died from non-cardiovascular disease (12). Also, in seven severely atherosclerotic vulnerable plaques investigated, infiltrating neutrophils appeared activated and were generating oxidants (11).

More recent studies, however, demonstrated that neutrophils accelerate all stages of atherosclerotic plaque development including plaque initiation, progression, instability, and rupture, staring with the onset of atherosclerosis through damage to the arterial wall caused by oscillating shear forces, hyperlipidemia, and inflammatory cytokines. These promote endothelial cell activation and adhesion by myeloid cells, consequently infiltrating into the arterial intima. The accumulation of immune cells, lipoproteins and cell debris in the arterial intima contribute to atherosclerotic plaque development and instability, promoting plaque rupture and cardiovascular complications (21). Specifically, neutrophils were shown to be the first to extravasate to sites of inflammation and to recruit and regulate monocyte entry into atherosclerotic lesions, thereby accelerating foam cell formation (22). Neutrophils and their secretory products also affected macrophage fate and function (21, 23), and were also shown to accumulate in injured areas of the plaque and to surround damaged vasculature, promoting tissue repair by inducing angiogenesis and revascularization (24, 25). Collectively, these studies clearly demonstrate the significance of CD66b+ neutrophils in all stages of plaque development and tissue repair.

Unlike the studies reported thus far, implicating neutrophils in various stages of plaque development, we focused our attention on identifying the presence of neutrophils resembling the Gφ that were previously reported to develop spontaneously in culture conditions in vitro, by characterizing their morphology and the markers co-expressed in plaque neutrophils.

In the current study, neutrophils expressing CD66b+/NE+/MPO+ were shown to reside in lipid-rich areas of the plaques as were CD163+ foam cell macrophages, and both cell types co-expressed the scavenger receptors CD68 and CD36. Nonetheless, in some of the plaques also varying sums of CD66b+/CD163+ co-localization was noted. However, CD66b+ cells looked like macrophages-foam cells morphologically and expressed CD68 and CD36 scavenger receptors which are considered exclusive as macrophages markers. Thus, the co-localization of both scavenger receptors - CD68 and CD36 with CD66b+ expressing cells might have facilitated their transformation into foam cells, similarly to Gφ in culture. It is therefore likely that the rooted belief that CD68 is exclusively expressed by cells of the monocytic lineage and macrophage-foam cells, lead some investigations in earlier studies to identify plaque macrophages based solely on the expression of CD68, consequently, ignoring the possibility that other cell types present in plaques might express similar phenotypes (26, 27). This approach might have resulted in underestimation of the contribution of neutrophils to plaque pathology (27).

To clarify whether CD66b+ and CD163+ cells in carotid plaques are distinct cell populations, CD66b/CD163 co-localization was performed. Substantial differences were noted between plaques. In a third, co-localization was negative, suggesting that CD66b+ and CD163+ represent distinct cell populations. In other plaques, co-localization varied, but was mostly low, whilst, in few plaques it was highly positive. Hence, indicating that CD66b+/CD163+ co-localization is a gradual process of neutrophil to macrophage (or vice versa) trans-differentiation. Evidently, neutrophil plasticity and trans-differentiation are increasingly being recognized. Neutrophils were shown to express features of other myeloid cells including eosinophils, antigen presenting cells (APCs), dendritic cells (DCs) as well as monocytes (13). Trans-differentiation of neutrophils to APCs was found at inflammatory sites of autoimmune arthritis (28). Long-lived neutrophils with phenotypic plasticity, trans-differentiated into neutrophil-dendritic cells hybrids (N-DCs), expressing properties of neutrophils and DCs, accompanied with a potent anti-fungal activity (29). In rheumatoid arthritis, N-DCs were found concurrently with CD66b+ neutrophils co-expressing features of APCs. Moreover, increased ROS generation in the synovial fluid of these patients promoted trans-differentiation of neutrophils to N-DCs (30). Also, mouse bone marrow resident cells with neutrophilic phenotype were able to differentiate to monocytes when cultured with M-CSF (31). Taken together these recent findings support a high neutrophil plasticity and trans-differentiation. Yet, we cannot rule out the possibilities that there is a macrophage to neutrophil trans-differentiation or that CD66b+/CD163+ co-expressing cells in the carotid plaques might have resulted from macrophages phagocytosing dead neutrophils and adopting their markers.

Of note, in carotid plaques, smooth muscle cells (SMCs) were shown to trans-differentiate and acquire macrophage-like phenotype, contributing to atherosclerotic plaque pathogenesis due to their phenotypic plasticity and the ability to mimic macrophages. This trait resulted in underestimating SMCs’ involvement in plaque pathology (32, 33). Moreover, these latter findings support a fundamental mechanism of myeloid cells (and neutrophils in particularly), in which they rapidly adapt to their immediate tissue environment or to diverse biological microenvironmental stimuli in health and disease through plasticity and trans-differentiation (13).

Correlation analysis revealed that CD66b+ and CD163+ cells were significantly positively correlated with CD68. While CD66b+ cells were negatively correlated with extracellular lipid and all lipids, CD163+ cells were significantly positively correlated with intracellular lipids, hence becoming foam cells. Given that CD66b expression in neutrophils is positively correlated with the expression of CD68, and negatively with all lipids indicates that also CD66b+ might become foam cells in lipid rich areas of the plaques, and thus, represent a supportive role by acting as a backup system when the macrophage clearing system is insufficient and/or overwhelmed (34). For instance, by eliminating and protecting the arterial wall from excessive lipid accumulation.

Among other factors, increased oxidant activity in plaques is associated with endothelial dysfunction (19) and neutrophil trans-differentiation into N-DCs (30). Macrophages co-expressed 3-NT+/CD163+ while co-expression in neutrophils was detected in 3-NT+/NE+ cells. Both cell types were significantly positively correlated with quantitative expression of 3-NT, indicating that in carotid plaques, nitrosative stress is associated with both. Interestingly, cells co-expressing CD66b+/CD163+ were highly positively correlated with 3-NT, as were neutrophils that trans-differentiated into N-DCs (30). Activated macrophages are known to produce NO through iNOS. However, the data are scarce for neutrophils. Yet, few studies show that all three NOS isoforms are present and generate NO in adult human neutrophils (35). Moreover, MPO, a major factor in oxidizing products in neutrophils was shown to produce nitrating oxidants in vivo (36). In addition, 3-NT levels were shown to decrease in stable plaques and increase in unstable plaques (37). In a study identifying CD68+ cells as representing only macrophages, carotid 3-NT and NOSII co-localized in late stages of atherosclerosis, suggesting that NOS expression is involved in oxidation/peroxidation of lipids promoting carotid formation (27). Hence, 3-NT+/NE+ neutrophils in carotid plaques can contribute to free radicals by oxidizing/peroxidizing lipids, and thus, facilitate foam cell formation through scavenger receptor regulation.

A great number of cytokines and chemokines were identified over the last few years in neutrophils, signifying their involvement in inflammatory and immune processes as well as in angiogenesis and repair (38). In earlier studies, neutrophils were shown to synthesize and store molecules with angiogenic activity including VEGF (39), which participates in angiogenesis and neovascularization, to promote plaque destabilization and thrombeolitic events (40). However, VEGF can have dual effects, detrimental as well as beneficial, by protecting endothelial cells (41). Besides VEGF, intraplaque neutrophils were also shown to produce IL-8 and elastase which are all crucial for the development and progression of plaques (42). Herein, we confirmed that both CD163+ and CD66b+ cells in plaques co-expressed VEGF, as illustrated in Figure 7 .

The limitations of this study should be acknowledged. In the immuno-histological staining experiments, we could not perform co-localization with two antibodies simultaneously due to limitations in the kit’s procedure. However, consecutive sections were analyzed basically at the same area of the plaque (see Figure 1 ). Immunofluorescence and co-localization experiments with confocal microscopy, also compensated to some extent for this limitation.

The number of plaques investigated was relatively small. However, it should also be acknowledged that these experimental procedures for identifying the markers investigated on a large-scale study are tedious, time consuming and very expensive. Moreover, unlike tissue culture studies or studies with inbred animal strains, each plaque is a complex tissue, with a unique morphology and cellular composition which was endarterectomized at different stages of its development and exposed to different environmental/microenvironmental signals. Undoubtedly, large-scale studies are needed to further clarify the characteristics and functions of CD66b+ and CD66b/CD163 co-expressing cells and their contribution to plaque pathology. It is still an open question – weather it is neutrophil to macrophage trans-differentiation or vice versa, or do they represent neutrophil-macrophage hybrids? Or possibly CD66b+/CD163+ cells are macrophages phagocytosing dead neutrophils and adopting their markers? However, since neutrophils are the first to arrive to plaques, and to facilitate many monocyte functions they might be the ones that respond swiftly to signals and acquire macrophage markers.

In surveying the literature, we could not find a report demonstrating CD66b+ neutrophils resembling the ones reported herein. Neutrophils that are large foam cells with a rounded nucleus but positive for NE/MPO/CD63, for the non-neutrophil markers, CD68 and CD36 scavenger receptors, and concurrent neutrophil-macrophage CD66b+/CD163+ co-expression to varying extents in many of the plaques. Some of the earlier studies presenting photomicrographs of CD66b+ positive cells in human carotid plaques specimens were stained only by immunohistochemistry, validating the presence of neutrophils solely based on their specific CD66b+ markers. In these studies, the resolutions presented for CD66b+ cells were illustrated by very small magnifications, or by molecular methods, making it impossible to appreciate their size or morphology (10–12, 38, 42–44) as described herein. The CD66b+/CD163+ cells’ co-localizations depicted here might represent a neutrophil-macrophage hybrid cell or trans-differentiation of one type of cell to the other. Since CD66b+ cells and lipids were significantly inversely correlated, we hypothesize that the presence of CD66b+ cells in lipid-rich areas might represent a supportive role in helping macrophage foam-cells to combat inflammation and eliminate and protect the arterial wall from excessive lipid accumulation. On the other hand, 3-NT which promotes oxidation of lipids, could facilitate carotid plaque formation. Of note, identifying these unique subpopulations of neutrophils in plaques and in additional in vivo inflammatory conditions may provide valuable insights into the regulation of inflammatory responses during the initiation and progression of atherosclerotic plaque development. Moreover, the in vitro studies described earlier can be utilized to further facilitate the understanding of inflammatory/atherosclerotic neutrophil biology.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by The Local Humans Rights Committee of RAMBAM Medical Center (RMB-0175-14), according to the declaration of Helsinki. The patients/participants provided their written informed consent to participate in this study.

LL led the project. LL and AH participated in the conceptual framework of the project. AH recruited the patients. AH and ES-o performed endarterectomy. ES-o collected and organized patients’ clinical data. LL analyzed and interpreted the data, wrote and edited the manuscript and edited the proof. All authors contributed to the article and approved the submitted version.