Acidification of the intimal fluid: the perfect storm for atherogenesis

Abstract

Atherosclerotic lesions are often hypoxic and exhibit elevated lactate concentrations and local acidification of the extracellular fluids. The acidification may be a consequence of the abundant accumulation of lipid-scavenging macrophages in the lesions. Activated macrophages have a very high energy demand and they preferentially use glycolysis for ATP synthesis even under normoxic conditions, resulting in enhanced local generation and secretion of lactate and protons. In this review, we summarize our current understanding of the effects of acidic extracellular pH on three key players in atherogenesis: macrophages, apoB-containing lipoproteins, and HDL particles. Acidic extracellular pH enhances receptor-mediated phagocytosis and antigen presentation by macrophages and, importantly, triggers the secretion of proinflammatory cytokines from macrophages through activation of the inflammasome pathway. Acidity enhances the proteolytic, lipolytic, and oxidative modifications of LDL and other apoB-containing lipoproteins, and strongly increases their affinity for proteoglycans, and may thus have major effects on their retention and the ensuing cellular responses in the arterial intima. Finally, the decrease in the expression of ABCA1 at acidic pH may compromise cholesterol clearance from atherosclerotic lesions. Taken together, acidic extracellular pH amplifies the proatherogenic and proinflammatory processes involved in atherogenesis.

Atherosclerosis is a chronic disease of the inner layer of the arterial wall, the intima. The disease involves slow accumulation of lipoprotein-derived lipids into the intimal layer where they become modified and, as a result, trigger a maladaptive immune response characterized by infiltration of monocyte-derived macrophages into the arterial wall. Mature atherosclerotic lesions are gradually formed via chronic inflammation, tissue remodeling, smooth muscle cell proliferation, fibrosis, and calcification.

Some areas of the arterial tree are more prone to atherosclerosis than others, due to their relatively thick intima (1–3). One major factor explaining the relationship between intimal thickness and susceptibility to atherosclerosis is the absence of capillaries and lymphatics (4), which restricts the supply of oxygen and nutrition of intimal cells (5). Thus, hypoxia develops easily in the deep intima, and is further exacerbated by the additional increase in intimal thickness that occurs during atherogenesis (6, 7). Due to hypoxia, intimal cells become more dependent on glycolysis for energy production. Importantly, activated immune cells, including macrophages, favor energy production by glycolysis even in a normoxic environment (8). Similarly, proliferation of smooth muscle cells has been shown to involve stimulation of glycolysis (9). As cells with a glycolytic phenotype produce and secrete more protons than cells using oxidative phosphorylation, both aerobic and anaerobic glycolysis cause acidification of the extracellular fluid.

Indeed, acidic pH is found in various inflammatory sites (10–13), where local acidosis can affect the ongoing immune response (14, 15). The extrusion of intracellular protons is important for the activity of immune cells, because, by extruding excess intracellular acid, the cells not only protect themselves from intracellular acidification, but also deliver protons to the extracellular milieu to facilitate various cellular functions in a paracrine or autocrine fashion. For example, extrusion of intracellular protons allows sustained activity of NADPH oxidase, an enzyme present on the plasma membrane of phagocytes involved in mounting a key bactericidal response, the oxidative burst (16, 17). Acidosis greatly enhances the receptor-mediated uptake of opsonized bacteria into macrophages thereby promoting efficient clearance of an infection (18), and it also boosts antigen presentation by these cells via enhancing fluid-phase endocytosis and increasing the expression of molecules involved in antigen presentation (19, 20). Extracellular acidity is also needed for the hydrolytic activity of lysosomal enzymes secreted by the phagocytes, and thereby tends to augment the extracellular destruction of bacteria. By allowing the secreted lysosomal cathepsins to retain their activity, extracellular acidity also facilitates the movement of immune cells into the site of action (21).

Similar to other inflammatory sites, acidic extracellular pH is also found in atherosclerotic lesions. In the report by Naghavi et al. (22), pH values as low as 6.8 were measured using a microelectrode in the subendothelial areas of human carotid plaques, and differences as great as 1.0 pH units were noted within most plaques. In addition, visualization of the plaques with two pH-sensitive fluorescent dyes indicated that the pH may reach values even below 6. In this review, we describe results from various experimental settings providing strong supportive evidence that an acidic microenvironment in the arterial intima could have a direct role in atherogenesis. Importantly, in the context of atherosclerosis, many of the acidity-induced physiological cellular functions that have evolved to aid, e.g., in the killing of bacteria, become maladaptive and actually may aggravate atherogenesis, particularly by affecting several key elements of the intimal lipoprotein metabolism involved in the progression of atherosclerosis.

MECHANISMS OF LOCAL ACIDIFICATION IN ATHEROSCLEROTIC PLAQUES

Tissue hypoxia and acidification may be linked via enhancement of glycolytic cellular metabolism at the hypoxic areas. Hypoxic cells have been visualized in human carotid atherosclerotic lesions using the hypoxia marker pimonidazole that becomes reductively activated by intracellular redox enzymes at oxygen tensions ≤10 mmHg and forms adducts with thiol groups of proteins (23, 24). Immunohistochemical staining of the pimonidazole adducts showed strong hypoxia in macrophages near the deep intimal core regions of the lesions and, furthermore, the hypoxic macrophages colocalized with nuclear staining of hypoxia-inducible transcription factor 1α (HIF-1α), a major regulator of cellular response to hypoxia that induces, e.g., the metabolic switch to glycolysis (25). Consistent with the data from human lesions, hypoxia was detected in macrophages present in the lipid core region of advanced rabbit atherosclerotic plaques, and, moreover, high lactate levels were measured in the same areas indicating induction of glycolysis (26).

As stated above, several types of activated immune cells, including the macrophages abundant in atheromas, favor energy production by glycolysis even in a normoxic environment [reviewed in (8, 27)]. This phenomenon of aerobic glycolysis, known as the Warburg effect, was first described in cancer cells by the world-renowned biochemist Otto Warburg (28). Curiously, it seems that aerobic glycolysis is specifically upregulated by proinflammatory activation of immune cells, not by anti-inflammatory activation of immune cells; classically activated proinflammatory M1 macrophages and T helper cells display a glycolytic phenotype, whereas alternatively activated M2 macrophages and regulatory T cells with anti-inflammatory properties are characterized by enhanced oxidative phosphorylation (27). Basal metabolism in resting mouse peritoneal macrophages is predominantly glycolytic though metabolic flux is slow, and the rate of glycolytic flux is greatly increased by classical M1 type activation (LPS + IFN-γ) and innate activation (LPS or other TLR agonists alone), but not by alternative M2 type activation [interleukin (IL)-4/IL-13; IL-10] (31). The importance of HIF-1α and glycolysis in macrophage energy metabolism under normoxia is further highlighted by the drastic decrease in steady state ATP levels in HIF-1α-deficient macrophages (32) and by the marked increase in the expression of glycolytic enzymes during differentiation of human monocytes into macrophages (33). Similar to macrophages, proinflammatory activation of dendritic cells through toll-like receptors and of T lymphocytes through the T cell receptors triggers a HIF-1α-dependent increase in glycolytic metabolism (34, 35). The rationale behind the strong induction of aerobic glycolysis in activated macrophages and other immune cells most likely lies in the strong induction of various biosynthetic pathways and proliferation in these cells; glycolysis is not only a rapid source of ATP, but a high glycolytic rate also promotes accumulation of glycolytic intermediates that are mainly fed into the pentose phosphate pathway for the production of amino acids, nucleotides, and NADPH (8). Another possible rationale for glycolytic energy production is the compensation of a shift in mitochondrial function from production of ATP toward production of mitochondrial reactive oxygen species (ROS), which was recently shown to have an important role in macrophage bactericidal activity (36).

Taken together, levels of both anaerobic and aerobic glycolysis are likely to increase in the intimal cells during lesion development, resulting in increased production of lactate and H⁺. The excess H⁺ and lactate are secreted from the cells through the activity of several pumps, exchangers, and transporters, which locally decrease the extracellular pH (Fig. 1). Indeed, both increased lactate concentrations (26) and extracellular acidification (22) are observed in atherosclerotic lesions.

Fig. 1.

Enhanced glycolysis leads to extracellular acidification. After proinflammatory activation, macrophages produce energy predominantly through glycolysis in both hypoxic and normoxic conditions (the “Warburg effect”), which leads to the formation and secretion of lactate and H⁺ via the activity of various pumps, exchangers, and transporters located in the plasma membrane.

EFFECT OF LOCAL ACIDOSIS ON IMMUNE FUNCTIONS OF MACROPHAGES

It has been known for decades that macrophages are able to adapt to and survive the local acidosis that develops at acute inflammatory sites (37). Thus, macrophages, the most abundant immune cells in atherosclerotic lesions, will remain viable despite the acidic microenvironment of the atherosclerotic intima and are subject to acidosis-induced modulation of their immune functions. Of note, extracellular acidity also decreases the intracellular pH of macrophages (38), which greatly amplifies the number of pathways potentially modulated by extracellular acidosis. Because macrophages have a central role in triggering and maintaining the inflammatory reaction in the arterial wall throughout all stages of atherogenesis, any modification in their function would profoundly affect several mechanisms at play in lesion progression. Here, we focus on the effects of acidic extracellular pH on cells of the monocyte-macrophage lineage (summarized in Fig. 2). For a review of the effects of pH on lymphocytes and neutrophils, we commend the excellent review by Lardner (14).

Fig. 2.

The effects of local extracellular acidity on macrophage immune functions. Acidic environment enhances fluid phase endocytosis, Fcγ-receptor-mediated phagocytosis, clearance of apoptotic cells, and antigen presentation. Extracellular acidity also activates the NLRP3 inflammasome, which results in the secretion of two potent proinflammatory cytokines, IL-1β and IL-18.

One of the central features emerging from studies of macrophages in acidic environments is the modulation of various cellular uptake mechanims by pH. At pH 6.5, the avidity of heat-aggregated human IgG immune complexes to monocyte and macrophage Fcγ receptors (FcγRs) was double that at pH 7.3 (39). Furthermore, preincubation of macrophages at pH 6.0 led to enhanced FcγR-mediated phagocytosis of IgG-opsonized latex beads and bacteria at pH 7.4 without increasing the binding of particles to the cell surface (18). Thus, exposure to an acidic environment enhances FcγR-mediated uptake both via increased binding to FcγR and via increased activity of the internalization machinery. In a recent study, expression of the phosphatidylserine receptor, stabilin-1, on macrophages was shown to be increased at pH 6.8, resulting in enhanced clearance of apoptotic cells (40). Finally, extracellular acidosis enhanced fluid phase endocytosis by macrophages and dendritic cells, accompanied by increased expression of molecules involved in antigen presentation, including major histocompatibility complex I and CD86 (19, 20).

Macrophage-generated ROS are an important component of the bactericidal activity of the cells; however, excessive ROS production may cause oxidative damage to the producing cell and its surroundings. Macrophage superoxide production in response to phorbol myristate acetate was decreased both by extracellular and intracellular acidosis (41, 42). On the other hand, acidic pH increases the rate of superoxide dismutation into hydrogen peroxide, and protonation of superoxide anions at acidic pH generates a species with increased reactivity (43). Regarding atherosclerosis, the net effect of the apparently two-way modulatory influences of extracellular acidosis on ROS production are reflected by increased iron-catalyzed extracellular LDL oxidation by macrophages, which could be explained by enhancement of the hydrogen peroxide-dependent Fenton reaction producing hydroxyl radicals (44). In addition to promoting oxidative modifications, acidic pH may also promote nitrosative damage, because an acidic environment induces the expression of the inducible nitric oxide synthase in macrophages, resulting in nitrite accumulation in the culture medium (45).

Activated macrophages secrete a plethora of proinflammatory cytokines and mediators that contribute to the inflammatory reaction in the atherosclerotic intima. Bellocq et al. (45) have shown that 2 h culture in medium adjusted to pH 7.0 activates nuclear factor (NF)-κB, the key inducer of proinflammatory cytokine expression in rat and mouse macrophages via a positive feedback loop of TNF-α secretion and autocrine signaling. In contrast, other studies have shown that low pH (pH 7.0–5.5) inhibits LPS-induced TNF-α secretion (mediated by NF-κB) in mouse and rabbit, but not in human macrophages (46–49). Accordingly, an inhibitory effect by low pH on LPS-induced NF-κB signaling was found in mouse, but not in human macrophages (46, 49). Differences in the study setups and the pH range used likely explain the contrasting results obtained in mouse macrophages. Human macrophages, based on Gerry and Leake (49), seem less sensitive to modulation of NF-κB activity by low pH, but more data on NF-κB target binding at a wider range of acidic pH values is required before firm conclusions can be made.

Recently, we and others have shown that acidic pH stimulates the secretion of the key proinflammatory and proatherogenic cytokine IL-1β in primary human monocytes and monocyte-derived macrophages (38, 50). IL-1β production is tightly regulated; the inactive procytokine, pro-IL-1β, is only expressed following specific stimulation, and cleavage of the procytokine by caspase-1 protease is required for biological activity. Monocytes constitutively express active caspase-1, and thus IL-1β is proteolytically activated and secreted immediately after induction of pro-IL-1β expression (51), as exemplified by acidic pH-induced pro-IL-1β expression and secretion (50). In contrast, caspase-1 in macrophages is in the inactive pro-form and, therefore, pro-IL-1β expression and caspase-1 activation are both required for secretion of mature IL-1β (51). We showed that extracellular acidity has no effect on pro-IL-1β expression in macrophages. However, when macrophages were stimulated with LPS to produce pro-IL-1β, extracellular pH 6.0–7.0 triggered activation of caspase-1 via the NLRP3 inflammasome and high-level secretion of mature IL-1β, as well as that of IL-18, another caspase-1 target cytokine (38). We also found synergy between low pH and cholesterol crystals, another activator of the NLRP3 inflammasome (52, 53), in induction of IL-1β secretion (38). Confirming the strong proinflammatory potential of acidic environment, a very recent microarray study compared macrophage gene expression at pH 7.4 and 6.8 and found 353 differentially expressed genes that showed marked enrichment of pathways related to inflammation and immune responses (54).

How might these changes in macrophage immune function relate to atherogenesis in the arterial wall? As discussed above, extracellular acidosis increases the FcγR-mediated uptake of immune complexes. Immune complexes of modified LDL and corresponding antibodies have been found in atheromas, and their uptake to macrophages via FcγRs triggers secretion of proinflammatory mediators and may enhance foam cell formation (55). Of note, acidosis also increases LDL oxidation by macrophages, potentially contributing to the pool of immunogenic modified LDL species in the lesions. On the other hand, enhancement of apoptotic cell clearance by macrophages at low pH may be a more anti-atherogenic effect. Importantly, low pH triggers the secretion of the potent proinflammatory cytokines, IL-1β and IL-18, in macrophages. For example, IL-1β increases the expression of adhesion molecules on endothelial cells to attract more inflammatory cells into the lesion and IL-18 contributes to induction of highly proatherogenic IFN-γ in T cells (56). Low pH can also boost antigen presentation in macrophages and thus contribute to induction of adaptive immune responses in the lesions. Thus, acidosis promotes many key immune functions of macrophages with a predominantly proinflammatory effect. Although these effects may be beneficial for efficient clearance of acute inflammation, they seem maladaptive in the context of atherosclerosis. In atherosclerosis, the local inflammatory response develops mainly under sterile conditions and excessive unresolved inflammation promotes lesion development and instability of the plaques.

ACIDIC EXTRACELLULAR pH INCREASES THE RETENTION OF ATHEROGENIC LIPOPROTEINS

Atherosclerosis is characterized by the extra- and intracellular accumulation of lipoprotein-derived lipids. Low extracellular pH affects many of the processes involved in lipid accumulation (Fig. 3). Upon entering the subendothelial arterial intima, LDL particles encounter a dense extracellular matrix network rich in proteoglycans, collagen, and elastin, with which the LDL particles tend to interact. Especially important is the interaction with proteoglycans (57, 58), which initiates LDL retention in the intima (59, 60), particularly at the atherosclerosis-prone sites, where the proteoglycan composition favors the retention of apoB-containing lipoproteins (61, 62). In addition to LDL, other apoB-containing lipoproteins (chylomicron remnants, VLDL, and IDL) also bind to proteoglycans, albeit less tightly (63–65), and can therefore contribute to lipid accumulation in the intima (66). Recently, Mendelian randomization studies have provided strong supportive evidence for the causative roles of both LDL and triglyceride-rich lipoproteins in the development of cardiovascular disease (67).

Fig. 3.

The effects of extracellular acidity on extracellular and intracellular cholesterol accumulation. Extracellular acidity enhances retention, modification, and aggregation of LDL, and so promotes both extra- and intracellular cholesterol accumulation. Extracellular acidity also remodels HDL particles with generation of preβ-HDL. However, in acidic environments, preβ-HDL is prone to degradation by acidic proteases. Moreover, acidity decreases the expression of the ABCA1 transporter and the secretion of apoE by macrophage foam cells, so decreasing cholesterol efflux from these cells.

The affinity of lipoproteins for proteoglycans is quite low at neutral pH, but acidic pH significantly enhances the binding of all three atherogenic apoB-100-containing lipoproteins (VLDL, IDL, and LDL) to human aortic proteoglycans (68, 69). The lipoprotein-proteoglycan interaction is mediated by certain positively charged sequences in apoB which contain lysine and arginine residues and negatively charged sulfate and carboxyl groups in the proteoglycan glycosaminoglycan chains. At acidic pH, additional sequences of apoB are likely to be important for the interaction: because the pK_a of histidine side chains is ∼6.0, the positive charge of the histidine residues increases as the pH decreases around this pH, histidine residues of apoB-100 may also be involved in the interaction with proteoglycans. In contrast, the negative charge of the sulfate and carboxyl groups in the proteoglycans are unlikely to be affected by the degree of acidification found in the arterial intima, because they have a pK_a of <2.0 and 3.02–4.37, respectively (70). Thus, the intimal acidity, even if reaching a pH value of 6, would not result in protonation of the sulfate and carboxyl groups, and the net negative charge of the proteoglycans will be preserved, thereby allowing interactions with positively charged lipoproteins and other molecules. Interestingly, proteoglycans may contribute to extracellular acidification, mediated by the attraction of H⁺ to their negatively charged sulfate and carboxyl groups (71). Although the negatively charged groups attract H⁺ ions and other cations, they do not bind the ions at the slightly acidic pH values present in the arterial intima. This can cause differences in the distribution of the ions and lower the pH in the vicinity of the proteoglycans.

Lipoproteins may be retained in the arterial intima also via bridging molecules, such as LPL (72–74). Thus, LPL binds to proteoglycans via ionic interactions of high affinity and to lipoproteins via hydrophobic interactions. There are no studies in which the effect of acidic pH has been studied on this bridging property of LPL. However, the binding of LPL to heparin is enhanced at mildly acidic pH values (75) and LPL is able to bind to lipid droplets at least at pH values as low as pH 6.5 (76). Together, these pieces of information suggest that acidic pH would most likely also enhance this type of lipoprotein retention in the arterial intima.

ACIDITY ENHANCES MODIFICATION OF apoB-CONTAINING LIPOPROTEINS

LDL particles isolated from atherosclerotic lesions show signs of various modifications, such as oxidation, proteolysis, and lipolysis, which are indicative of oxidative, proteolytic, and lipolytic enzyme activities in the lesions (77). Such modified LDL particles are often aggregated and/or fused into large lipid droplets. Acidic extracellular pH may increase lipoprotein modification (71, 78), which may in turn further decrease the extracellular pH (see below). Acidic proteases, such as cathepsins, are found extracellularly in normal and atherosclerotic intima, and they efficiently proteolyze apoB-100 leading to LDL particle fusion (79–81). Moreover, proteolysis of apoB-100 sensitizes the LDL particles to lipases, thereby promoting their lipolytic modification (82, 83). Interestingly, activated macrophages secrete cathepsins together with H⁺ ions and so, by acidifying their local microenvironment, provide optimal conditions for activity of these secreted proteases (84). Finally, when macrophages latch onto large aggregates of LDL, they create partially sealed compartments on the surface of these aggregates, drop the pH, and secrete lysosomal acid lipase (LAL). This enzyme then hydrolyzes the cholesteryl esters of the aggregated lipoproteins into unesterified cholesterol and fatty acids. Unesterified cholesterol can enter the cells and the fatty acids can further acidify the microenvironment (85, 86). This process is analogous to the way osteoclasts, macrophage-like cells, normally latch onto bone and degrade it through the formation of sealed compartments at low pH (85).

The free cholesterol generated from lipoproteins or other extracellular lipid deposits by the secreted acidic LAL may contribute to the nucleation and growth of cholesterol monohydrate crystals in the lesions. Large cholesterol crystals are easily visible by microscopic examination of atherosclerotic arteries, and they are a hallmark of advanced atheromas; however, smaller crystals are also found in about one third of intermediate lesions that lack necrotic cores (87, 88). Using a new microscopic technique, cholesterol microcrystals were detected in the aortic wall of apoE-deficient mice after just 2 weeks of high cholesterol diet, coinciding with the first appearance of macrophages (52). As discussed above, cholesterol crystals were recently shown to elicit an inflammatory response via the NLRP3 inflammasome, but the mechanisms of crystal nucleation in atherosclerotic lesions in vivo remain elusive. Electron microscopy of human atherosclerotic lesions has shown that cholesterol crystal growth occurs predominantly in the matrix-embedded extracellular lipid deposits of the deep arterial intima, whereas most macrophage foam cells reside in the more superficial intimal layer (89). Although less frequent, cholesterol crystals were also found inside macrophage foam cells in human lesions (89), and thus, the original site of crystal nucleation could not be defined with certainty. However, more recent in vitro studies have shown that cholesterol crystal nucleation can occur both within macrophage foam cells (87, 90–93) and on the surface of enzyme-modified LDL particles (94), of which the latter mechanism may indeed be amplified by secreted LAL in an acidic environment. Acidity may also directly affect cholesterol crystallization and the interaction of cholesterol crystals with each other, lipoproteins, or cell membranes (95–97).

Phospholipase A₂ (PLA₂) enzymes hydrolyze sn-2 ester bonds in glycerophospholipids, yielding lysophospholipids and FFAs, which have been shown to have proinflammatory effects (98–102). In addition, as noted above, the FFAs may contribute to extracellular acidification. Several secreted PLA₂ enzymes with different substrate specificities and lipolytic activities are present in human atherosclerotic arteries (98, 103, 104). PLA₂-V is one of the enzymes that may modify lipoproteins in the intima (105, 106), and at acidic pH it is more active against lipoproteins (69). The activity of the PLA₂s is largely determined by their ability to bind to substrate membranes (107, 108) and it is sensitive to changes in the lipid membrane induced, e.g., by acidic pH. The electric charge distribution at the membrane interface of zwitterionic phosphatidylcholine molecules is determined by interaction of the phosphate and amine groups with counterions, such as protons and hydroxide ions (109). At neutral pH, the lipid head groups associated with hydroxide ions are the predominant form, meaning a net negative charge, but the increased proton concentration of acidic pH tends to enhance the association of the protons with the head groups and so enhance their charge-neutral character (109, 110). This may enhance the interaction between LDL and PLA₂-V and so explain the observed acidity-increased activity of PLA₂ toward LDL (69). Acidity also affects the fate of the products of PLA₂ activity. At neutral pH and physiological albumin concentration, most of the FFAs and lysophospholipids generated through PLA₂-dependent hydrolysis of LDL particles are readily scavenged by albumin; however, at acidic pH the ability of albumin to bind the lipolytic products is decreased, and more of the lipolytic products remain in the LDL particles (111). This likely results from the property of albumin to preferentially bind the FFAs in their anionic form (112, 113); at acidic pH the FFAs are largely uncharged.

SMase hydrolyzes sphingomyelin on the surface of lipoprotein particles into ceramide and phosphocholine. Secretory SMase, a product of the acid SMase gene, is able, at neutral pH, to hydrolyze PLA₂-digested and proteolyzed LDL or apoC-III-enriched LDL, as well as LDL extracted from human atheromata (82, 114). Digestion induced by secretory SMase is promoted at acidic pH (82, 114, 115). In vitro, SMase-modified LDL particles promptly aggregate, the aggregate size increasing by synergy with chondroitin-sulfate-rich proteoglycans and LPL (72). Aggregate size after SMase digestion also increases as the pH drops, and these aggregates at pH 5.5 can ultimately span several micrometers (116). Consistent with these in vitro results, some of the ceramide-containing LDL particles in human atherosclerotic lesions become large micron-sized aggregates (117). Ceramide efficiently displaces cholesterol from lipid bilayers into the crystalline phase, thus promoting cholesterol crystal nucleation (118, 119). Therefore, the combined actions of SMase and cholesterol esterase on LDL may be important for cholesterol crystallization (87, 94), particularly in acidic environments, where these enzymes are most active.

Oxidative modification of lipoprotein particles leads to formation of lipid peroxides, which further decompose into aldehydes that react with the protein components of the particle. Acidic pH enhances oxidation of LDL by iron, nitric oxide, and myeloperoxidase [reviewed by Leake (71)]. Interestingly, acidic pH induces the aggregation of oxidized LDL (120), but, unexpectedly, LDL oxidized at acidic pH is less cytotoxic than LDL oxidized at neutral pH (121). In contrast to the proteolytic and lipolytic modifications that enhance LDL-proteoglycan binding, oxidation decreases the binding of LDL to proteoglycans due to the neutralization of positively charged lysine residues in apoB-100 (122). However, oxidized LDL particles can bind to human aortic proteoglycans under acidic conditions despite the oxidation-induced decrease in their affinity for proteoglycans (68).

ACIDITY INCREASES LDL UPTAKE BY MACROPHAGES

The appearance of macrophage-derived foam cells in the intima is the hallmark of developing atherosclerotic lesions. Foam cells are formed when macrophages take up modified apoB-containing lipoproteins via various mechanisms, including scavenger receptor-mediated uptake and phagocytosis (124). Also, aggregated LDL bound to the components of the extracellular matrix produced by smooth muscle cells are readily taken up by macrophages (72). Extracellular acidosis enhances the uptake of native and modified LDL particles by macrophages through increases in the levels of cell surface proteoglycans, and of LDL-proteoglycan binding (69, 125). These proteoglycans are most likely heparan-sulfate-rich proteoglycans of the syndecan family (126, 127). Extracellular acidosis may also accelerate foam cell formation by enhancing lipoprotein modifications that promote lipoprotein uptake by macrophages (69, 79, 116, 120). In addition, Howard Kruth has proposed a model of foam cell formation that does not involve LDL modification or macrophage receptors. Thus, when macrophages are incubated with high LDL concentrations comparable to those found the intimal interstitial fluid (128), foam cells are formed as a result of fluid phase pinocytosis of unmodified LDL particles (129). Fluid phase pinocytosis by macrophages has been reported to be increased at acidic pH (19, 20). Thus, most of the proposed mechanisms involved in the uptake of lipoproteins by macrophages and foam cell formation are augmented at acidic extracellular pH (Fig. 3)

ACIDITY DECREASES CHOLESTEROL EFFLUX FROM MACROPHAGE FOAM CELLS AND INDUCES HDL REMODELING

By promoting cholesterol efflux from macrophage foam cells and by inducing anti-inflammatory effects in macrophages and endothelial cells, HDL particles are also thought to possess strong atheroprotective functions in vivo (130). Although an inverse relation between plasma HDL-cholesterol levels and the rate of atherosclerosis progression has been documented in experimental animal models and in human population studies by aid of imaging of atherosclerotic lesions, recent clinical and genetic studies have failed to confirm the hypothesis of plasma HDL-cholesterol level being per se a determinant of, at least, the final atherothrombotic events in humans (131). Such failures to therapeutically modify atherogenesis in humans have actually led to a conservative skepticism regarding the benefits of HDL-oriented therapies and may have their root cause in our incomplete comprehension of the high complexity of the HDL particles. Indeed, it currently appears that the capability of HDL to prevent atherosclerosis depends on both quantitative and qualitative features of their proteome and lipidome, which ultimately translates itself into functional differences not detected by simply measuring the plasma levels of HDL-cholesterol (132). HDLs are also capable of mediating intercellular communication among different types of cells by a mechanism that involves the transfer of endogenous miRNAs among the body compartments, which may partly explain the high versatility of HDL function (133).

Cholesterol efflux induced by HDL initiates reverse cholesterol transport (RCT), which transfers peripheral cholesterol to the liver for ultimate excretion in the gut (134). Because reduction of the cholesterol pool in arterial macrophages can prevent progression of atherosclerosis or even induce its regression, the particular path of RCT that initiates in the macrophage foam cells (macrophage-RCT) is the most relevant fraction of the total body RCT regarding atherosclerosis (135). In human macrophage foam cells, enhanced cholesterol efflux in response to LXR activation appears to be entirely dependent upon the lipid transporter ABCA1 (136), which promotes cholesterol efflux to lipid-free apoA-I and to the nascent lipid-poor preβ-migrating HDL subpopulation (preβ-HDL) (137). Importantly, cholesterol loading induces a compensatory response in macrophages by upregulating ABCA1 mRNA and protein expression (138).

Various conditions present in the atherosclerotic intima, such as hypoxia (139), inflammation (140), and oxidative stress (141), reduce the ABCA1-mediated cholesterol efflux from macrophage foam cells. We found that acidic pH also reduces ABCA1-mediated efflux of cholesterol from cultured human macrophage foam cells to apoA-I (142). Because the α-helical content and secondary structure of lipid-free apoA-I are not affected by low pH (109), our finding strongly suggests that acidity impairs the function of ABCA1 rather than the function of apoA-I. Consistent with this speculation, we found that impaired cholesterol efflux from macrophages cultured in medium with a pH value of pH 5.5 or 6.5 is accompanied by progressive reduction in the levels of the ABCA1 protein. In accord with this in vitro finding, the ABCA1 protein level is significantly reduced in whole extracts of human carotid atheromas (143, 144), and, moreover, ABCA1 mRNA was found not to be expressed in the foam cells within the necrotic core of advanced plaques in human atherosclerotic aortas (145), where the intimal fluid most likely has an acidic pH. It is plausible to assume that acidification of the extracellular fluid has additional effects on the activity of ABCA1 in cholesterol efflux; this may be through alterations in the physical properties of ABCA1 and the spatial geometry of the plasma membrane, which is thought to be the main location of ABCA1-mediated apoA-I lipidation (146), or through effects on the electrostatic interaction between ABCA1 and apoA-I (147). Moreover, secretion of apoE, which also stimulates ABCA1-mediated lipid efflux (148), is reduced in macrophages incubated at acidic pH (142). Taken together, these findings support the notion that the cholesterol efflux-mediating activity of ABCA1 in macrophages is reduced by the low extracellular pH found in advanced atherosclerotic lesions (9). Because the interaction of ABCA1 with apoA-I inhibits the expression of inflammatory cytokines in macrophages (149), the lack or low activity of ABCA1 in acidic microenvironments may also exacerbate atherogenesis by leading to enhanced proinflammatory responses in the lesional macrophages.

A small labile pool of apoA-I constantly recycles on and off HDL particles during metabolic remodeling in vivo (150, 151), thereby exchanging apoA-I molecules with the preβ-HDL pool. In this regard, we have shown that an acidic pH in vitro promotes remodeling of the mature HDL particles resulting in formation of preβ-HDL and fusion of the α-migrating HDL (152). Such remodeling was initiated by unfolding of the apolipoproteins on the surface of HDL particles, which was followed by the release of apoA-I from the particles, resulting in the generation of unstable apoA-I-deficient HDL particles that then fuse. However, it is important to note that because lipid-poor apoA-I is extremely sensitive to proteolysis, the proteases known to be secreted by intimal cells will easily render it nonfunctional (153). Indeed, we have found that various acidic cathepsins found in atherosclerotic lesions (81, 79) also effectively degrade apoA-I both in lipid-free and lipid-poor forms with loss of their cholesterol efflux-inducing activity (155). Thus, the production of HDL-derived lipid-free or lipid-poor apoA-I in atherosclerotic plaques may increase as a result of acidification of the intimal fluid, but such generated apoA-I species may also be degraded when acidic proteases gain in function in the low pH microenvironment (Fig. 3).

Based on the above-described fragmentary and apparently two-way processes regarding the effects of acidity on HDL-dependent mechanisms which regulate cholesterol efflux from macrophage foam cells, the envisioned scenario of such acidity-dependent effects on atherogenesis remains undefined. Thus, while acidity induces remodeling of HDL and ensuing generation of lipid-poor apoA-I species, the lipid-poor species of apoA-I may easily be lost due to extracellular degradation by acidic proteases. Moreover, the expression of ABCA1 in macrophage foam cells at acidic pH is low or absent, and so may compromise cholesterol clearance from foam cells in an arterial segment with an acidic extracellular pH (142). Because the first of the three acidity-induced processes tends to increase, and the two latter ones tend to decrease cholesterol efflux, it is impossible even to predict the net effect of acidity on cholesterol efflux from macrophage foam cells in an acidic environment in vivo.

CONCLUDING REMARKS AND FUTURE PERSPECTIVES

By virtue of its ability to enhance extracellular and intracellular lipid accumulation and to promote proinflammatory processes in macrophages, extracellular acidity has emerged as a novel and potentially crucial element of atherogenesis. Extracellular acidity modifies various proinflammatory functions of macrophages, e.g., by triggering the secretion of potent proinflammatory cytokines by macrophages via activation of the inflammasome pathway (38). Acidity also aggravates extra- and intracellular accumulation of cholesterol in the atherosclerosis-prone regions of the arterial tree. In contrast to the rather straightforward scenario of a proatherogenic role of acidity on intimal accumulation of cholesterol derived from apoB- 100-containing lipoproteins, HDL metabolism in the acidic intimal fluid appears to be complex and unpredictable in its potential outcomes. Such entanglement of HDL metabolism in advanced human atherosclerotic lesions, if present, could be one of the root causes of the failure in designing a clinically relevant HDL-based strategy for anti-atherogenic therapies.

Atherosclerotic lesions contain a heterogeneous mixture of macrophage phenotypes, and some of them may confer more resistance to acidity and hypoxia than others. Recent findings indicate that macrophage proliferation within the plaque plays an important role in the regulation of the size of the macrophage population in atherosclerotic lesions (156). This raises the intriguing possibility that the macrophage population in the plaque evolves over time through selection of acid-resistant macrophages, analogous to the evolution of an acid-resistant cell population in solid tumors (157). There may be an additional level of selection if smooth muscle cells in the atherosclerotic lesion are sensitive to acid-induced cellular toxicity. Indeed, extracellular acidosis inhibits proliferation and migration of vascular smooth muscle cells, and also increases their susceptibility to apoptosis (158). Thus, in response to the selection pressure from the microenvironmental pH, the lesion could gradually become enriched with acid-resistant macrophages and depleted of smooth muscle cells, a population imbalance that is found in the rupture-prone subset of atherosclerotic lesions (159). Such a scenario invites us to envision that acidic pH in atherosclerotic lesions not only promotes atherogenesis, but may also contribute to the often lethal atherothrombotic complications of the disease.

After having defined the acidic intimal environment as a perfect soil for atherogenesis, we need to ask: how can it be changed back to neutral? Obviously, we do not know the answer, but are compelled to find it. It is of great interest to note that recent developments in nanoparticle-based therapy of cancer are exploiting the acidic extracellular environment of a tumor for targeted drug delivery to cancer cells (160). Importantly, recent understanding of the similarities between cancer cells and inflammatory cells has unraveled the role of AMP-activated protein kinase as an inhibitor of glycolysis that boosts oxidative phosphorylation (8, 27, 161). The AMP-activated protein kinase is activated by certain drugs and xenobiotics, most notably by the type 2 diabetes drug, metformin, and by the classic anti-inflammatory drug, salicylate (8, 27, 161). This new information, coupled with the developing technologies for acid-dependent drug delivery, might guide us when searching for new therapeutics and anti-atherogenic strategies. The prospects of a successful novel proton-lowering strategy are increased when considering that attenuation of the rate of glycolysis in macrophages may be associated with a phenotypic shift from a proinflammatory into an anti-inflammatory macrophage (8).