Evidence that the initial up-regulation of phosphatidylcholine biosynthesis in free cholesterol-loaded macrophages is an adaptive response that prevents cholesterol-induced cellular necrosis. Proposed role of an eventual failure of this response in foam cell necrosis in advanced atherosclerosis.

Macrophages in atherosclerotic lesions accumulate free cholesterol (FC) as well as cholesteryl ester and appear to have high rates of phospholipid (PL) synthesis and increased PL mass. Previous short term (i.e. </=24 h) studies with cultured macrophages have shown that these cells respond to FC loading by up-regulating phosphatidylcholine biosynthesis. We propose that this response is adaptive by keeping the FC:PL ratio in the macrophages from reaching toxic levels. We further propose that one cause of macrophage necrosis, a prominent and important event in atherosclerosis, is an eventual decrease of this adaptive response. To explore these ideas, cultured macrophages were loaded with FC for up to 4 days and assayed for phosphatidylcholine biosynthesis, FC and PL mass, and cytotoxicity. For the first 24 h, cellular phosphatidylcholine biosynthesis and FC and PL mass increased 3-4-fold, and thus the FC:PL molar ratio was prevented from reaching very high levels; at this point, there were no overt signs of cytotoxicity. Over the next 24-48 h, however, phosphatidylcholine biosynthesis, and then phosphatidylcholine mass, began to decrease. Initially, the macrophages remained healthy and continued to accumulate FC, but eventually these macrophages, but not unloaded macrophages, became necrotic (swollen organelles and disrupted membranes). Lipoprotein dose studies indicated a close relationship between the onset of macrophage necrosis and the FC:PL ratio. To test further the causal nature of these relationships, cellular FC and PL mass were independently manipulated by using high density lipoprotein3 (HDL3) to decrease cellular FC and choline depletion to decrease cellular PC. As predicted by our hypotheses, HDL3 protected FC-loaded macrophages from necrosis, whereas choline depletion accelerated cytotoxic changes. These findings support the idea that the initial increase in phosphatidylcholine biosynthesis in FC-loaded macrophages is an adaptive response that prevents cholesterol-induced macrophage necrosis. We propose that an eventual failure of the PL response in foam cells may represent one cause of macrophage necrosis in advanced atherosclerotic lesions.


Macrophages in atherosclerotic lesions accumulate free cholesterol (FC) as well as cholesteryl ester and appear to have high rates of phospholipid (PL) synthesis and increased PL mass. Previous short term (i.e. 24 h) studies with cultured macrophages have shown that these cells respond to FC loading by up-regulating phosphatidylcholine biosynthesis.
We propose that this response is adaptive by keeping the FC:PL ratio in the macrophages from reaching toxic levels. We further propose that one cause of macrophage necrosis, a prominent and important event in atherosclerosis, is an eventual decrease of this adaptive response. To explore these ideas, cultured macrophages were loaded with FC for up to 4 days and assayed for phosphatidylcholine biosynthesis, FC and PL mass, and cytotoxicity. For the first 24 h, cellular phosphatidylcholine biosynthesis and FC and PL mass increased 3-4-fold, and thus the FC:PL molar ratio was prevented from reaching very high levels; at this point, there were no overt signs of cytotoxicity. Over the next 24 -48 h, however, phosphatidylcholine biosynthesis, and then phosphatidylcholine mass, began to decrease. Initially, the macrophages remained healthy and continued to accumulate FC, but eventually these macrophages, but not unloaded macrophages, became necrotic (swollen organelles and disrupted membranes). Lipoprotein dose studies indicated a close relationship between the onset of macrophage necrosis and the FC:PL ratio. To test further the causal nature of these relationships, cellular FC and PL mass were independently manipulated by using high density lipoprotein 3 (HDL 3 ) to decrease cellular FC and choline depletion to decrease cellular PC. As predicted by our hypotheses, HDL 3 protected FC-loaded macrophages from necrosis, whereas choline depletion accelerated cytotoxic changes. These findings support the idea that the initial increase in phosphatidylcholine biosynthesis in FC-loaded macrophages is an adaptive response that prevents cholesterol-induced macrophage necrosis. We propose that an eventual failure of the PL response in foam cells may represent one cause of macrophage necrosis in advanced atherosclerotic lesions.
Cholesterol-loaded macrophages are prominent features of atherosclerotic lesions (1)(2)(3), and there is increasing evidence that these cells play an important role both in early atherogenesis and in the clinical progression of advanced lesions (4 -6). Although cholesteryl ester accumulation in lesional macrophages (foam cells) is often emphasized, these cells also accumulate large amounts of FC, 1 particularly in advanced atherosclerosis (7)(8)(9)(10). In this light, we have been interested in elucidating biological responses of macrophages to FC loading. One such response is the post-translational activation of the phosphatidylcholine (PC) biosynthetic enzyme, CTP:phosphocholine cytidylyltransferase (CT), which leads to an increase in PC biosynthesis and in cellular PC mass (11,12). This response is likely to be physiologically important, since increases in PC biosynthesis and mass have been noted to occur in lesional macrophages in vivo (13)(14)(15)(16)(17).
We have hypothesized that this PC response is initially adaptive (11,12), since it would keep the cellular FC:PL ratio from getting too high and causing damage to cells (see Ref. 18). For example, membranes enriched with FC demonstrate inhibition of several membrane-bound enzymes (19 -22), and cholesterol crystals may accumulate in such cells (14). A corollary of our hypothesis is that an eventual blunting of this PC response would lead to cellular necrosis, and this scenario may be one cause of the necrosis of macrophages that is known to occur in advanced lesions (5,6,23,24). Macrophage necrosis has been proposed to play an important role in plaque destabilization and thus clinical progression of lesions (5,6).
The goal of the present study was to test these ideas using FC-loaded cultured macrophages. In our previous studies, the macrophages were FC-loaded for no longer than 24 h, at which point the PC biosynthetic response was still increasing, and the cells appeared healthy (11). In the present study, we have cultured these cells for longer periods, and we found that the PC biosynthetic response, but not the accumulation of FC, began to decrease after 24 h of culture. As predicted by our hypotheses, this event caused an increase in the cellular FC:PL ratio, and the cells subsequently showed signs of necrosis. Furthermore, removal of cellular FC prevented cytotoxicity, whereas premature blunting of the PL response accelerated cytotoxicity. These findings support the idea that the initial increase in PC biosynthesis in response to FC loading in macrophages is adaptive and raise the possibility that an eventual blunting of this response may lead to foam cell necrosis in advanced atherosclerotic lesions.

EXPERIMENTAL PROCEDURES
Materials-The Falcon tissue culture plasticware used in these studies was purchased from Fisher. Tissue culture media and reagents were obtained from Life Technologies, Inc., and fetal bovine serum (FBS) was purchased from Gemini Bioproducts (Calabasas, CA). Choline-deficient medium, generously provided by Dr. Martin Houweling (University of Alberta), was made by adding the following supplements to 5 liters of DMEM, formula 79 -5141: NaHCO 3 (18.5 g), D-glucose (22.5 g), L-leucine (525 mg), sodium pyruvate (550 mg), L-methionine (150 mg), and L-arginine (420 mg). Lipoprotein-deficient serum (LPDS) was prepared from fetal bovine serum by preparative ultracentrifugation (density, 1.21 g/ml) (25 (26) was generously provided by Dr. John Heider of Sandoz, Inc. (East Hanover, NJ). Stock solutions (10 mg/ml) were prepared in dimethyl sulfoxide; the final dimethyl sulfoxide concentration in both treated and control cells was 0.05%. All organic solvents were purchased from Fisher. Cholesterol (Ͼ99% pure) was from Nu-Chek-Prep, Inc. (Elysian, MN). All other chemicals and reagents were purchased from Sigma.
Cells-Monolayer cultures of J774.A1 macrophages (from the American Type Culture Collection) were grown and maintained in spinner culture with high glucose Dulbecco's modified Eagle's medium (DMEM) containing 10% (v/v) FBS, penicillin (100 units/ml), streptomycin (100 g/ml), and glutamine (292 g/ml) as described previously (27,28). Resident mouse peritoneal macrophages were harvested from female ICR mice (25-30 g); concanavalin A-elicited macrophages were harvested from the peritoneum of female ICR mice 3 days after the intraperitoneal injection of 40 g of concanavalin A in 1 ml of sterile PBS (29). For each experiment, the cells were plated in 22-mm dishes at an approximate density of 10 6 cells/dish (or 35-mm dishes at 2 ϫ 10 6 cells/dish) in DMEM, 10% FBS and then incubated at 37°C in an atmosphere containing 5% CO 2 . After 2 h of incubation, the monolayers were washed with warm PBS and then incubated with DMEM, 10% LPDS alone or containing acetyl-LDL, 58035, or acetyl-LDL plus 58035, as indicated in the individual figure legends. These media were replaced with fresh media of the same composition every 12 h.
Lipoproteins and Nonlipoprotein Cholesterol-LDL (density, 1.020 -1.063 g/ml) and HDL 3 (density, 1.125-1.21 g/ml) from fresh human plasma were isolated by preparative ultracentrifugation (25). Acetyl-LDL was prepared by treating LDL with acetic anhydride as described by Goldstein and co-workers (30). All concentrations of lipoproteins are given in terms of their protein content with bovine serum albumin as a standard. All lipoprotein preparations were stored under argon at 4°C and used within 4 weeks. Nonlipoprotein cholesterol was added to media from a 20 mg/ml stock in ethanol; the final concentration of ethanol (also added to control wells) was 0.25%. Cholesterol-phosphatidylserine liposomes were made by mixing 2.8 mg of phosphatidylserine and 1.5 mg of cholesterol (1:1 molar ratio) in chloroform; the solvent was completely removed by evaporation under nitrogen followed by lyophilization. Three ml of PBS were added to the dried lipids, and the suspension was sonicated under argon for 15 4-min bursts at 4°C using a tapered microtip on a Branson 450 sonicator (setting number 3).
[ 3 H]Choline Labeling Experiments-Incorporation of [ 3 H]choline into phosphatidylcholine in intact cells was determined as described previously (11). In brief, macrophages incubated under various conditions were pulsed for 1 h with 2 Ci of [ 3 H]choline/ml (80.0 Ci/mmol). Lipid extracts of the cells were then separated by TLC in a solvent system of chloroform/methanol/acetic acid/water (50:25:8:4, v/v), and the radioactivity in phosphatidylcholine was quantified. Although the incorporation of labeled choline into PC can be affected by pool sizes of precursors, we have previously shown that cellular FC loading does not affect these pool sizes (11).
Cellular Lipid and Cytotoxicity Assays-Cellular FC and PL mass measurements were carried out exactly as described by Shiratori et al. (11). Cell protein content was measured by the method of Lowry et al. (31), and cell DNA content was determined using a minor modification (11) of the method of Labarca and Paigen (32). Cells were loaded with 1 Ci of [ 3 H]adenine and then assayed for adenine release as described by Warner et al. (33). Lactate dehydrogenase activity was assayed in medium and cells as reported previously (34), except a kit purchased from Sigma (catalogue number 500) was used to measure enzyme activity; the assay measures the unreacted pyruvic acid using a colorimetric assay.
Phase Microscopy-Cells in 22-mm tissue culture dishes were viewed using a Leitz Diavert inverted phase microscope and photographed using a Nikon camera attached to the microscope.
Electron Microscopy-Cells were plated on 35-mm tissue culture dishes and incubated for 2 days as indicated. At the end of incubations, the cells were prepared for electron microscopy as described previously by our laboratory (35). Briefly, the macrophages were washed three times with ice-cold 0.13 M sodium phosphate buffer, pH 7.4 (Buffer C), and then fixed for 30 min at 4°C in Buffer A containing 3% glutaraldehyde. After three 5-min washes in Buffer A, the cells were subjected to two sequential 45-min incubations with 1% OsO 4 in Buffer C. These two incubations and all subsequent procedures were performed at room temperature in Buffer C. After the OsO 4 incubations, the samples were treated with tannic acid and p-phenylenediamine, as described by Guyton and Klemp (36). The cells were then dehydrated using graded ethanol washes and embedded in Epox (Ernest F. Fullam, Inc., Latham, NY). The embedded samples were removed from the plastic dishes and sectioned on an MT 6000 ultramicrotome. The sections were stained with uranyl acetate and lead citrate and examined with a JEOL 1200 EX electron microscope.
Statistics-Unless otherwise indicated, results are given as means Ϯ S.D. (n ϭ 3); absent error bars in these figures signify S.D. values smaller than the graphic symbols.

Effects of Prolonged FC Loading of Macrophages on PC Biosynthesis, Cellular Lipid Content, and Cell
Viability-FC loading of cultured macrophages over a 24-h period leads to an increase in PC biosynthesis due to post-translational activation of the PC biosynthetic enzyme, CTP:phosphocholine cytidylyltransferase (CT) (11,12). This response results in an increase in cellular PC mass (11). To explore the effects on PC biosynthesis of more prolonged FC loading, such as occurs in atheromata (7-10), macrophages were incubated with 50 g of acetyl-LDL/ml plus the ACAT inhibitor 58035 (26) for up to 80 h ( Fig.  1). The data show that after the initial increase in PC biosynthesis, there is a decline in the rate of this reaction. We have hypothesized that the initial increase in PC biosynthesis is an adaptive response that prevents the cellular FC:PL ratio from reaching potentially toxic levels (11,12,18). Thus, in the context of this hypothesis, the data in Fig. 1 predicted two outcomes: first, that the FC:PL ratio would rise to high levels in macrophages loaded with FC for long periods; and second, that the macrophages would eventually show signs of cytotoxicity. The data in panel A of Fig. 2 show that cellular PL mass rose steeply during the first 24 h of FC loading, as described previously (11), but then increased at a slower rate and eventually declined, which is consistent with the PC biosynthesis data in Fig. 1. PL subclass analysis indicated that the major species contributing to this trend was PC (data not shown). FC mass, however, continued to rise at the same rate until the 3rd day of FC loading (panel B). These changes in cellular lipids resulted in a 3-fold increase in the FC:PL ratio by day 4 of FC loading (panel C).
The second prediction of our hypothesis, namely that the macrophages would show signs of cytotoxicity after prolonged FC loading, was investigated using both morphological and biochemical methods. Morphologically, we noted that the macrophages became rounded and vacuolated and began to detach from the plate at day 2 of FC loading. Phase micrographs of control and 3-day FC-loaded macrophages are shown in Fig. 3. Note that the macrophages incubated with 58035 alone (panel B) or acetyl-LDL alone (panel C) showed similar density and morphology as control macrophages (panel A). In contrast, many of the macrophages that were FC-loaded by incubation with acetyl-LDL plus 58035 had detached from the plate, and most of those that remained were fragmented and rounded (panel D). Electron micrographs of control macrophages and those loaded with FC for 1 and 3 days are shown in Fig. 4. Two major points are revealed by these data. First, the 1-day FC-loaded cells (panel B) appear morphologically healthy compared with unloaded macrophages (panel A), although the loaded cells did contain intracellular membrane whorls, as described previously (11). These whorls most likely are major sites for the increased PL mass in FC-loaded cells. Apparently healthy cells were also noted at 30 h of FC loading (see also below), indicating no overt cytotoxicity at the time the PC biosynthesis response first began to decrease (see Fig. 1). Second, the 3-day FC-loaded cells that did remain attached to the plate (panel C) showed signs of necrosis, including swollen cytoplasm and disrupted organelles. Careful inspection of electron micrographs of many 2-and 3-day FC-loaded macrophages consistently showed these signs, whereas properties of apoptosis (e.g. fragmented and condensed chromatin, condensed organelles, and multiple plasma membrane blebs (37)) were not seen at these times or at 1 day of FC loading. In addition, a biochemical marker of apoptosis, cleavage of 116-kDa poly-(ADP-ribose) polymerase to an 80-kDa signature fragment by an interleukin 1␤-converting enzyme-like protease (38), was not observed at any stage during FC loading of macrophages (data not shown). Quantitative biochemical data supporting the morphological observations of both cell detachment and cytotoxicity are shown in Fig. 5. Contents of cellular protein (panel A) and DNA (panel B) decreased by more than 50% in 3-day FC-loaded macrophages, consistent with a loss of cells from the monolayer. In addition, there was 2-3-fold increase in the release of intracellular [ 3 H]adenine (panel C) and LDH (panel D) from these cells, consistent with cellular necrosis. As with the morphological observations, these changes were not seen at 30 h of FC loading, the initial time of decrease of PC biosynthesis (Fig. 1). In summary, the data in Figs. 1-5 are consistent with the notion that a decline in the initial induction of PC biosynthesis in FC-loaded macrophages leads to a toxic FC:PL ratio, which, in turn, results in cellular necrosis.
To further establish the relationship between the cellular FC:PL ratio and cytotoxicity in FC-loaded macrophages, we incubated the cells with three different concentrations of acetyl-LDL, all in the presence of 58035 (Fig. 6). At various time points, we measured cellular FC and PL contents and looked for signs of necrosis using both morphological (phase microscopy) and biochemical (cellular protein) criteria, as described above. At 10 g of acetyl-LDL/ml plus 58035, the cellular FC:PL ratio remained only slightly elevated for up to 4 days, and overt necrosis did not begin until after the 5th day (Fig. 6, panel A). At 25 g of acetyl-LDL/ml plus 58035, the FC:PL ratio increased to a level greater than that seen with the lower lipoprotein concentration, and necrotic changes began approximately 2 days earlier, at day 3 (panel B). At the highest concentration of acetyl-LDL (50 g/ml) plus 58035, the FC:PL ratio rose steeply, and necrosis began by day 2 (panel C). Although the absolute FC:PL ratio at which necrosis was triggered in J774 macrophages varied somewhat between experiments, we usually found that elevation above 0.4 for Ͼ24 h led to these changes. Thus, there appears to be a close correlation between the FC:PL ratio in macrophages and the onset of cellular necrosis.
The experiments described up to this point were conducted with J774 macrophages, a dividing cell line. To examine these issues in a nondividing macrophage, we studied both resident and immunologically primed mouse peritoneal macrophages. The primed macrophages, which are defined as being able to respond to subsequent stimulation with lipopolysaccharide, were prepared by injecting the peritoneal cavities of mice with concanavalin A (Con A) 3 days before harvesting the peritoneal macrophages (29). Unforeseen differences in acetyl-LDL-induced cholesterol loading between these two states of nondividing macrophages provided us an opportunity to test our ideas. As shown in panel A of Fig. 7, the Con A-elicited macrophages (closed squares) accumulated more FC than the resident cells (open squares) when incubated with acetyl-LDL plus 58035. The increase in PL mass, however, was similar in both cell types (panel B), and so the FC:PL ratio increased substantially more in the Con A-elicited cells (panel C). Morphological observations (not shown) and cellular protein mass data (panel D) revealed that the Con A-elicited cells eventually showed signs of necrosis and loss of adherence, whereas the resident cells remained healthy and adherent throughout 3 days of observation. Con A-elicited macrophages that were not loaded with FC maintained a FC:PL ratio of 0.3 and showed no overt signs of cytotoxicity throughout the 3-day period. Note that peritoneal macrophages appear to be more resistant to FC-mediated toxicity than J774 macrophages, since a ratio of 0.7 in resident cells did not lead to necrotic changes during this time. Nonetheless, the overall principle of a relationship between the FC:PL ratio and cytotoxicity is established with these nondi-  Effect of Separate Manipulations of Cellular FC Content and PL Response in FC-loaded Macrophages-To further define the causal relationships between PL metabolism, the FC:PL ratio, and cellular necrosis in FC-loaded macrophages, we conducted one set of experiments in which FC mass, but not PL mass, was altered, and another set of experiments in which the PL response, but not FC mass, was manipulated. In the first set of experiments, macrophages were loaded with FC for 36 h and then chased in medium in the absence or presence of HDL 3 , an inducer of cellular cholesterol efflux (39). At the time of addition of the HDL 3 (36 h), the macrophages had accumulated both FC and PL as expected (Table I), and the cells showed no signs of necrosis (compare panel D of Fig. 8 with the control macrophages in panel A of Fig. 8). The macrophages chased in the absence of HDL 3 continued to accumulate a little more FC, presumably due to the uptake of residual surface-bound acetyl-LDL and the processing of prelysosomal acetyl-LDL, whereas those chased in the presence of HDL 3 lost substantial amounts of FC as expected (Table I). The PL content of these macrophages, however, was very similar (Table I). As shown in Fig. 8, the macrophages chased in the absence of HDL 3 showed signs of toxicity (compare panels E (2 days) and F (3 days) with the control macrophages in panels B and C), whereas those chased in the presence of HDL 3 remained healthy (panels G and H of Fig. 8). Thus, as predicted by our hypothesis, preventing the rise in the FC:PL ratio prevented cytotoxicity.
In the second set of experiments, our goal was to manipulate the PL response of FC-loaded macrophages. To accomplish this goal, we took advantage of the previous finding that incubating cells in choline-depleted medium partially inhibits PC biosynthesis (40). To establish this point in our system, we first analyzed the PL content of control and choline-depleted macrophages, both under cholesterol-deficient and FC-loading conditions. For these experiments, it was necessary to load the cells with a source of cholesterol that did not contain choline-phospholipids, and we chose nonlipoprotein cholesterol, which is known to modestly up-regulate PC biosynthesis in macrophages (11). As shown in Table II, choline deficiency resulted in a relative decrease in the PL content of both unloaded and FC-loaded macrophages compared with the respective choline-replete cells. The phase micrographs of these cells are shown in Fig. 9. Importantly, choline depletion in the absence of FC loading did not result in cytotoxicity under the conditions of our experiment (compare the macrophages in panel B with the control cells in panel A). When loaded with nonlipoprotein cholesterol, the choline-replete macrophages remained healthy (panel C), whereas the choline-depleted cells showed definite signs of cytotoxicity, including detachment, rounding, and fragmentation (panel D); cellular protein measurements confirmed this finding (not shown). Similar data were obtained with macrophages loaded with cholesterol-phosphatidylserine liposomes as another source of cholesterol without choline (compare the choline-depleted cells in panel F with the cholinereplete cells in panel E). Thus, partially blunting the PL response in FC-loaded macrophages promotes cytotoxicity, as predicted by our hypothesis. 2

DISCUSSION
As summarized in Fig. 10, the findings in this report support the hypothesis that the initial rise in PC biosynthesis in FCloaded macrophages is an adaptive response that keeps the FC:PL ratio from rising to toxic levels. This idea likely explains the lag in the onset of FC-mediated cytotoxicity observed in our study as well as that of Warner et al. (33), which did not look at cellular PL metabolism. With prolonged FC loading, however, this adaptive response fails, the FC:PL ratio rises to cytotoxic levels, and macrophage necrosis ensues. Although we did not 2 Theoretically, the data in Fig. 9 could be explained by choline deficiency increasing cellular FC content instead of, or in addition to, its proven effect in decreasing cellular PL content. Unfortunately, cellular FC measurements were difficult to interpret, since a substantial amount of both cholesterol added directly to medium, in the form of microcrystals (62), as well as cholesterol-phosphatidylserine liposomes stick to the cell surface. Thus, the FC:PL ratio could not be compared with the previous experiments using acetyl-LDL as the source of cholesterol. Nonetheless, there was no evidence that choline deficiency increased cellular FC. For example, the FC contents of marcophages not exposed to exogenous cholesterol were 44.8 Ϯ 0.1 (choline-replete) and 40.9 Ϯ 0.1 (choline-depleted) nmol/mg cell protein; in cells exposed to cholesterol added directly to medium, the FC contents of control and choline-depleted cells were 280.5 Ϯ 4.8 and 264.8 Ϯ 8.7 nmol/mg cell protein, respectively. study the exact cause of macrophage death in these studies, high cellular FC levels are known to inhibit several critical membrane enzymes, including Na ϩ /K ϩ -ATPase activity (19), Ca 2ϩ /Mg 2ϩ -ATPase activity (20), carnitine palmitoyltransferase activity (21), and alkaline phosphatase activity (22). Furthermore, as Small (14) points out, excessive accumulation of cholesterol monohydrate crystals could lead to lysosomal rupture and cellular necrosis. In the case of membrane enzyme inhibition, the cause of FC-induced toxicity is probably related to perturbations of membrane fluidity, which could be compensated by increases in membrane phospholipid content (18). Cholesterol crystal formation would also be expected to be prevented by increases in cellular PL (14). Interestingly, mouse peritoneal macrophages appear to require a higher FC:PL ratio than J774 macrophages to trigger cytotoxic changes (Fig. 7). Investigation into the mechanism of this relative resistance may shed additional light on how cells adapt to excess FC.
The FC-loaded macrophages used in our studies are noted to have intracellular membrane whorls (see Ref. 11 and Fig. 4, panel B), which probably represent the sites where most of the increased PL mass in these cells accumulate. Furthermore, filipin-labeling studies have shown that much of the FC that accumulates in cultured macrophages incubated with acetyl-LDL plus 58035 is localized in perinuclear lysosomes, presumably in lysosomal membranes (12). These findings suggest that our experimental model may reflect physiological events, since lesional macrophages in vivo have both intracellular membrane whorls (10) and accumulate FC in lysosomes (41,42). These observations raise questions, however, about how the increased phospholipid might protect the cells from FC-mediated toxicity. First of all, where in the cell does the accumulation of FC cause problems? Many of the enzymes inhibited by excess FC are localized in the plasma membrane (see above), and Warner et al. (33) have shown that FC export from the lysosomes is necessary for FC-mediated toxicity in macrophages. Thus, even though the bulk of FC appears to be in lysosomes, a critical amount of excess FC is probably in other cellular membranes, particularly the plasma membrane, and this localization is likely important for FC-mediated toxicity. Interestingly, we have shown that PC synthesis is still stimulated when lysosomal FC export is blocked (12), suggesting that lysosomal FC, while itself not initially toxic, may be the signal to "warn" the cell to protect other membranes from ensuing FC enrichment and damage. Eventually, massive lysosomal FC accumulation after very prolonged FC loading may result in cholesterol crystallization (43); consistent with this idea, cholesterol crystallization appears to be a relatively late effect of FC loading, since we have not observed it in our cells even after 3 days of incubation with acetyl-LDL plus 58035.
How does the presence of intracellular membrane whorls pertain to the proposed ability of increased cellular PL to protect macrophages from FC-mediated toxicity? One possibility is that the membrane whorls serve as a "sink" for excess cellular cholesterol; for example, a critical amount of excess cholesterol from the sites of sensitive membrane enzymes (e.g. plasma membrane) might be transferred to the whorls, thus preventing inhibition of these enzymes (see above). Another idea is that the whorls represent a storage form of "excess" phospholipid in FC-loaded macrophages; in this scenario, phospholipid would be transferred from the whorls to membranes in the cells that have a high FC:PL (cf. Ref. 44). If such cholesterol or phospholipid transfer reactions are, in fact, found to play a role in the adaptive response of macrophages to FC loading, a defect in these transfer reactions might contribute to or accelerate FC-mediated cytotoxicity.
Two important issues related to our studies but not addressed herein are the mechanisms of FC accumulation in vivo and the mechanism of blunting of the PC biosynthetic response with prolonged FC loading. Regarding the first issue, cells normally possess several mechanisms to prevent the accumulation of excess FC. These include cellular cholesterol efflux, cholesterol esterification, down-regulation of LDL receptors and of endogenous cholesterol biosynthesis, and cholesterol

FIG. 7. PL and FC mass and cell monolayer protein content in FC-loaded resident and concanavalin A-elicited mouse peritoneal macrophages. Monolayers of resident (open squares)
and concanavalin A-elicited (closed squares) mouse peritoneal macrophages were incubated for the indicated times in DMEM, 10% LPDS containing 50 g of acetyl-LDL/ml plus 5 g of 58035/ml. The cells were then assayed for cellular FC content (panel A) and cellular PL content (panel B); from these data, the cellular FC:PL ratio was calculated (panel C). The cells remaining on the monolayer were then assayed for protein content (panel D) as in Fig. 5; the 100% values for the resident and Con A-elicited cells were 0.29 and 0.13 g/well, respectively. As mentioned in the text, Con A-elicited macrophages that were not loaded with FC remained morphologically healthy throughout the 3-day period and showed much less loss of cells from the monolayer. metabolism (e.g. bile acid synthesis in hepatocytes) (39,45,46). For macrophages internalizing large amounts of lipoproteincholesterol by means other than the LDL receptor (e.g. via the scavenger receptor), several of these mechanisms are irrelevant, including receptor down-regulation, down-regulation of cholesterol biosynthesis, and metabolism of cholesterol into bile. In lesions, cholesterol efflux may be impeded due to inaccessibility of the cells to inducers of efflux, or it may be overwhelmed by the large amount of cholesterol in the cells. Similarly, the esterification pathway may be saturated or inhibited; for example, oxysterols in oxidized LDL may prevent traffick-ing of lipoprotein-cholesterol to ACAT (47), or ACAT itself may become dysfunctional in advanced foam cells.
With respect to PC biosynthesis, the initial up-regulation in response to FC loading is due to post-translational activation of the enzyme CTP:phosphocholine cytidylyltransferase (CT) (12); specifically, FC-mediated induction appears to involve a cellsignaling event involving the dephosphorylation of CT and probably other cellular proteins (12). The cause of the eventual blunting of the PC biosynthesis response in our cell-culture model and whether it is related to a decline in CT activity (e.g. through changes in CT phosphorylation state) require further FIG. 8. Phase micrographs of control and FC-loaded macrophages subsequently incubated in the absence or presence of HDL 3 . Panels A-C, monolayers of J774 macrophages were incubated in DMEM, 10% LPDS alone and viewed by phase microscopy at day 1 (A), day 2 (B), and day 3 (C). Panels D-H, other monolayers of J774 macrophages were incubated for 36 h in DMEM, 10% LPDS containing 50 g of acetyl-LDL/ml plus 5 g of 58035/ml. The cells were then washed with PBS and incubated an additional 60 h with DMEM, 10% LPDS containing 58035 in the absence or presence of 300 g of HDL 3 /ml (see Fig. 8). The cells were viewed at day 1 before the differential chase (D), at day 2 (minus HDL 3 ϭ E; plus HDL 3 ϭ G), and at day 3 (minus HDL 3 ϭ F; plus HDL 3 ϭ H). The original microscopy magnification was ϫ 40. Bar, 10 m. investigation. According to our working hypothesis, blunting of the PC biosynthesis response is an early event that subsequently causes a rise in the FC:PL ratio and FC-induced macrophage necrosis. We base this idea on the finding that PC biosynthesis started to decline at 24 h of FC loading (Fig. 1), while the first signs of necrosis were seen 24 h later. It is possible, however, that the blunting of the PC biosynthesis response itself is a very early result of subtle FC-induced toxicity, which then causes an escalation in the rise of the cellular FC:PL ratio. Assaying the sensitivity of CT and the other membrane-bound PC biosynthetic enzymes to changes in membrane FC content in vitro may shed light on this idea. In vivo, it is also possible that certain cytokines blunt the PC biosynthesis response to FC loading. For example, there is evidence that tumor necrosis factor-␣, which is synthesized and secreted by macrophages in response to a variety of factors (48,49) and is known to be present in atherosclerotic lesions (50, 51), can inhibit cellular PC biosynthesis in alveolar cells (52). We feel that the major impact of this work is that it provides  Macrophages encounter and internalize atherogenic lipoproteins in the subendothelial matrix of developing lesions, thereby accumulating excess cellular cholesterol. In early lesions, much of this cholesterol is esterified by the ACAT pathway, leading to the cholesteryl ester (CE)-filled foam cell. In later lesions, however, there is a progressive increase in cellular FC content, perhaps due to a disruption in cholesterol transport to ACAT, a failure of ACAT itself, or an increase in hydrolysis of cholesteryl ester stores by neutral cholesteryl ester hydrolase (NCEH). The increase in cellular FC, particularly lysosomal FC, leads to the post-translational activation of CT via a signaling pathway involving protein dephosphorylation. Activation of CT, in turn, results in an increase in cellular PC mass, the bulk of which appears to be stored in intracellular membrane whorls. We hypothesize that this response is adaptive by keeping the FC:PL ratio from reaching cytotoxic levels. We further propose that this initially adaptive response eventually diminishes, perhaps due to deactivation of CT (e.g. by cytokines), leading to an increase in the cellular FC:PL ratio and, ultimately, cholesterolinduced cellular necrosis. Refer to text and Refs. 11, 12 for details. at least one plausible explanation for macrophage necrosis in atherosclerotic lesions. Necrosis of macrophages in advanced atherosclerosis is thought to be an important event in lesion progression (5,6,23,24). For example, degradative enzymes released from these cells might contribute to plaque rupture and eventual acute thrombosis (5,6). The mechanism of cellular necrosis in atheromata, however, is not known. In addition to our hypothesis, other possibilities include nutrient deprivation, loss of growth factors, free radical and oxidized lipid injury, and toxic cytokines (for example, see Ref. 53). Heinle (54) attempted to test the importance of oxygen deprivation and glycogen depletion in lesional cell necrosis in the absence or presence of lipids by comparing two types of rabbit carotid lesions, proliferative lesions alone and proliferative lesions with fatty deposits. Although both lesions had a similar degree of glucose and glycogen depletion and of lactate accumulation, necrosis was only associated with the fatty lesions. This led the author to conclude that some factor other than or in addition to hypoxia is necessary for lesion necrosis and that the factor must be related to lipids in the lesions. This conclusion is consistent with our hypothesis as well as with hypotheses implicating other toxic lipids. In this regard, we did consider the possibility that the generation of lysophosphatidylcholine (lyso-PC), which can be cytotoxic (55,56), contributed to the necrosis of FC-loaded, PL-rich macrophages. In preliminary studies, we found that lyso-PC levels are only slightly elevated before the onset of necrosis, but they are markedly elevated in the medium after necrosis. Thus, although we have no evidence that lyso-PC generation causes necrosis of FC-loaded macrophages, it may be an important consequence of this event due to released cellular phospholipase A 2 acting on accumulated PC. This lyso-PC may then accelerate the necrosis of neighboring cells or contribute to specific atherogenic effects thought to be signaled by this lipid (cf. Refs. 57-59). Our hypotheses predict that manipulations resulting in a decrease in the PL content of lesional cells might accelerate lesion necrosis and progression. Investigators several decades ago studied the vessels of rats on a partially choline-deficient diet, which is known to partially decrease the phosphatidylcholine (PC) content of cells in culture (40) and in liver in vivo (60). Interestingly, when these rats were fed cholesterol and saturated fat, there was evidence of accelerated atherosclerosis in the aorta and coronary arteries, including increased lesion necrosis (Ref. 61 and references therein). Although the etiology of the accelerated atherosclerosis in this situation is not known and may be multifactorial, it is tempting to speculate that at least part of the mechanism may be related to our hypothesis about the cytotoxicity of increased FC:PL ratios. The current goal of our laboratory is to test our hypotheses in vivo by genetically manipulating arterial wall PL metabolism using induced-mutant mouse models. By crossing these mice to currently available mouse models of atherosclerosis, we hope to be able to show the important role of macrophage PL metabolism in macrophage necrosis and lesion progression.