**8. Roles of macrophages in the development and progression of atherosclerosis**

The major functions of lesional macrophages include removal of excessive lipids such as oxidized LDL and glycolaldehyde-LDL. Macrophages internalize oxidized lipids through scavenger receptors and become lipid-laden foam cells. Activated macrophage foam cells produce potent chemoattractants, such as MCP-1, which recruit additional monocytes/ macrophages from the circulating blood, accelerating the positive feedback loop of vascular inflammation. The imbalance between uptake and removal causes excessive cholesterol ester accumulation and apoptosis in lesional macrophages. Due to the lack of negative feedback mechanisms of scavenger receptor expression, macrophages cannot limit the uptake of lipids and largely depend on the cholesterol efflux to maintain cellular lipid homeostasis [42]. In early atheroma, neighboring macrophages take up apoptotic macrophages (efferocytosis) [90-92]. As atherosclerosis progresses, however, efferocytosis becomes impaired [93], leading to secondary necrosis. Due to the macrophage release of their cellular contents (e.g., debris, oxidized lipids, proinflammatory mediators), secondary necrosis amplifies pro-inflammatory response and develop the necrotic core within the lesions.

The balance between proinflammatory and anti-inflammatory populations of accumulating macrophages may determine the development of atherosclerotic plaques [53]. Active proinflammatory responses of macrophages may destabilize atherosclerotic plaques. The pro‐ duction of MMPs by macrophages may degrade collagen in the fibrous cap and make plaques susceptible for plaque rupture and thrombosis. Fukumoto et al. and Deguchi et al. used genetically altered mouse strains to provide the first in vivo direct evidence for the role of collagenases of the MMP family for the loss of fibrillar collagen within the intima [94, 95].

As mentioned, emerging data have proposed the heterogeneity of macrophages. A subpopu‐ lation of T lymphocytes (Th1 cells) secretes such as IFNγ, IL-2, IL-12, and TNFα and promotes the activation of macrophages toward a pro-inflammatory phenotype (M1). Th2 cytokines (e.g., L-4 and IL-13) induce an alternative form of activation toward a non/anti-inflammatory (M2) phenotype. The balance of such macrophage polarization (M1/M2 balance) may affect plaque outcome [71]. The high M1/M2 ratio in atherosclerotic plaques may induce lesion formation and plaque vulnerability [80, 96]. The evidence has identified switching of macro‐ phage phenotypes from M1 to M2 during the regression of atherosclerosis or in response to anti-inflammatory therapies [84, 97]. Proinflammatory M1 macrophages also induce SMC proliferation [98].

Alternatively activated M2 macrophages may generally exert anti-atherogenic effects. M2 cells suppress Th1 inflammatory responses. TGF-β released by M2 macrophages may inhibit the recruitment of inflammatory cells and the development of atherosclerosis [99]. M2 macro‐ phages also release IL-10, which inhibits the production of inflammatory cytokines from T lymphocytes and other macrophages. M2 macrophages suppress inflammatory milieu by clearing apoptotic cells and tissue debris [100, 101]. During the repair process after tissue injury, M2 cell may promote fibrosis [102], [103]. This action may potentially be beneficial in plaque instability via thickening the fibrous cap [104].

While the M1/M2 paradigm has clear relationships between stimulators and downstream effects and has thus served as a useful mechanistic model, the evidence suggests that macro‐ phages are more diverse. In particular, M2 may further contain various forms of macrophage activation, e.g, M2a to M2d [104]. Mox macrophages develop in response to atherogenic phospholipids and have lower phagocytotic and chemotactic capacity than do conventional M1 and M2 cells [104, 105]. Mhem cells, induced by intraplaque hemorrhage, often associate with atherothrombotic complications [105, 106]. CXC chemokine ligand 4 induces M4, a recently proposed subtype of atherogenic macrophages [107]. The heterogeneity of macro‐ phages thus seems to be more complex than previously proposed. Furthermore, in vivo functional significance of each macrophage subpopulation remains incompletely understood. Recently, new terms more specific to each stimulator were proposed, e.g. M(IFNγ), M(LPS), M(IL-10), and M(IL-4) [87].
