Mast Cell Tryptase Degrades HDL and Blocks Its Function as an Acceptor of Cellular Cholesterol
Objective— In human atherosclerotic lesions, degranulated mast cells are found in the vicinity of macrophage foam cells. Mast cell granules contain tryptase, a tetrameric serine protease requiring glycosaminoglycans for stabilization. No endogenous inhibitors have been described for tryptase, and the physiological functions of the enzyme are poorly understood. Here, we investigated the effects of human tryptase on the integrity of high density lipoprotein (HDL)3 and on its ability to release cholesterol from cultured mouse macrophage foam cells.
Methods and Results— Incubation of HDL3 with tryptase led to degradation of its apolipoproteins. Tryptase predominantly degraded a quantitatively minor subfraction of HDL3 that is lipid poor, exhibits electrophoretic pre-β mobility, and contains either apolipoprotein A-I or apolipoprotein A-IV as its sole apolipoprotein. Moreover, tryptase caused functional changes in HDL3 by destroying its ability to promote high-affinity efflux of cholesterol from macrophage foam cells, ie, the pre-β-HDL-dependent component of the process. Human aortic proteoglycans increased the ability of tryptase to proteolyze HDL3, suggesting that the proteoglycan-rich extracellular matrix of the arterial intima provides an appropriate environment for the extracellular actions of tryptase.
Conclusions— By depleting pre-β-HDL, mast cell tryptase may impair the initial step of reverse cholesterol transport and will then favor cellular accumulation of cholesterol during atherogenesis.
A key feature of atherosclerosis is the accumulation of cholesterol in the macrophages of the arterial intima.1 According to current concepts, cholesterol is taken up by the intimal macrophages when they ingest modified LDL, whereas HDL can remove cholesterol from the macrophages and carry it back into the circulation.2 Thus, for cholesterol to accumulate in the intimal macrophages, uptake must exceed release of cholesterol by these cells.3 One reason for such an imbalance is the inability of the HDL particles to remove cholesterol efficiently from intimal macrophages.
HDL removes cellular cholesterol by unspecific and specific mechanisms, ie, by aqueous diffusion and by interaction with cell-surface proteins, respectively.4,5⇓ The former process, ie, in the nonspecific diffusion-mediated pathway, does not essentially require apolipoproteins. If there is any role of apolipoproteins in this pathway, it would be enhancement of the plasma enzyme lecithin-cholesterol acyltransferase (LCAT), which operates as a driving force in the net efflux of cholesterol from cells to α-HDL. By promoting esterification of the cholesterol in plasma HDL, it channels cholesterol into the core of the spherical α-migrating HDL and so expands the size of the core and improves the cholesterol acceptor properties of α-HDL. The latter process, ie, the apolipoprotein-mediated removal of cellular cholesterol, is mediated by lipid-free apolipoproteins and small discoidal lipid-poor HDL subfractions, which, in contrast to the major α-migrating and lipid-rich component of HDL, exhibit pre-β mobility when analyzed by agarose gel electrophoresis. In this pathway, the lipid acceptors interact with cell membrane proteins, notably with the ATP-binding cassette transporter A1, which leads to efficient net removal of phospholipids and cholesterol from the cells.6
After the initial observation that apoA-I in pre-β-HDL is very sensitive to proteolytic degradation,7 we have shown that several proteases known to be present in the human arterial intima can degrade the apoA-I of HDL3 in vitro, thereby impairing cellular cholesterol efflux. These proteases include certain matrix metalloproteases,8 plasmin and kallikrein,9 and chymase, a mast cell protease.10 In human atherosclerotic lesions, mast cells are often present in the vicinity of foam cells.11 The mast cells are filled with cytoplasmic secretory granules that contain an abundance of 2 species of neutral serine proteases, tryptase and chymase. Indeed, as in other tissues, all the mast cells in the human arterial intima contain tryptase, and in addition, a variable fraction of them also contains chymase.11 On mast cell activation, these proteases are secreted in their active forms. Accordingly, mast cells are potentially a significant source of neutral proteases in the extracellular fluid of the arterial intima. We have previously demonstrated that mast cell chymase efficiently proteolyzes the minor pre-β-HDL subpopulations containing only apoA-I or only apoA-IV, thereby destroying the high-affinity component of cholesterol efflux induced by HDL3 and plasma.10,12⇓ However, the ability of the other major mast cell protease, tryptase, to degrade plasma lipoproteins has not been investigated so far.
Tryptase is a unique serine protease that is enzymatically active only in its tetrameric form and requires glycosaminoglycans, notably the heparin proteoglycans of mast cells, for preservation of its tetrameric structure and its enzymatic activity.13–15⇓⇓ Tryptase and heparin proteoglycans are colocalized within mast cells and are cosecreted on degranulation. Importantly, the mast cell heparin proteoglycans tend to prevent monomerization of the enzyme after secretion, thereby preserving its activity in the extracellular fluid.16 However, nothing is known about the potential of the proteoglycans present in the human arterial intima to regulate the activity of tryptase.
In the present study, we investigated whether human tryptase can degrade HDL. In addition, we studied the effect of tryptase on the ability of HDL3 to promote high-affinity efflux of cholesterol from macrophage foam cells and the influence of the proteoglycans isolated from the human aorta on these effects.
Please refer to the expanded Methods section in the online supplement (which can be accessed at http://atvb.ahajournals.org).
β-Tryptase was purified from human lung tissue and was also produced as a recombinant protein. Proteoglycans were isolated from the intima and media of human aortas. HDL3 (1.125 to 1.210 g/mL) was isolated from freshly isolated normolipidemic human plasma by sequential ultracentrifugation. 3H-labeled HDL3 was treated with tryptase as described in the figure legends. For measurement of cholesterol efflux, 3H-cholesterol-loaded mouse peritoneal macrophages were incubated with the indicated concentrations of HDL3, after which the radioactivity in the cell-free medium was determined.
Proteolysis of HDL3 by Tryptase
Because purified tryptase is an unstable tetrameric enzyme that requires aminoglycans for its stabilization, initial incubations of tryptase with HDL3 were performed in the presence of heparin. HDL3 (1 mg/mL), in which the apolipoproteins had been radiolabeled, was incubated with tryptase in the presence of heparin (mass ratio of heparin to tryptase 5:1) for up to 20 hours, and the formation of small peptides was measured (trichloroacetic acid-soluble radioactivity). As shown in Figure 1A, 10 μg/mL of tryptase in the assay produced a modest degree of degradation of HDL3 (slightly <1%) during the 20-hour incubation. However, increasing the concentration of tryptase to 30 μg/mL resulted in progressive degradation of HDL3 (up to ≈8%) during the incubation period. This concentration of the enzyme was selected for further experiments. As shown in Figure 1B, increasing the mass ratio of heparin to tryptase from 5:1 to 10:1 did not increase the rate at which tryptase degraded HDL3. However, in the presence of a lower heparin concentration (mass ratio 1:1), HDL3 degradation was slower, particularly during the latter part of the incubation, an effect comparable to that found when incubations were carried out in the absence of heparin (mass ratio 0:1). In general, in the presence of heparin, the rate of HDL3 degradation was most rapid during the first hour of incubation and then leveled off to reach a maximal value of 10% at 20 hours of incubation. When proteoglycans isolated from the human aortic wall were added instead of heparin, the rate of HDL3 degradation was even higher than that in the presence of heparin, reaching maximally 20% in the presence of a proteoglycan-to-tryptase mass ratio of 10:1 (Figure 1C). Finally, recombinant tryptase also degraded HDL3 (data not shown), thus excluding the possibility that the tryptase preparations from lung tissues were contaminated with other lung-derived proteases, particularly chymase.
Effect of Tryptase on HDL3-Induced Cholesterol Efflux
We next determined the effect of tryptase on the ability of HDL3 to promote the efflux of cellular cholesterol from macrophage foam cells. For this purpose, HDL3 was preincubated with or without tryptase for 20 hours at 37°C, after which the enzymatic activity of tryptase was blocked by the addition of leupeptin, and the mixtures were added to macrophage foam cells. As expected, intact control HDL3 led to a dose-dependent and saturable increase in 3H-cholesterol efflux (Figures 2A and 2B). Preincubation of HDL3 with tryptase reduced its ability to remove cholesterol from the cells. At HDL3 concentrations <50 μg/mL, the inhibitory effect of tryptase on the high-affinity component of the cholesterol efflux process became evident. This effect of tryptase on cholesterol efflux was observed with heparin (Figure 2A) and arterial proteoglycans (Figure 2B) as stabilizers of tryptase. The concentration of HDL3 necessary to achieve half of the maximal efficiency of the efflux process increased by ≈4-fold after tryptase treatment in both cases.
Proteolysis of ApoA-I by Tryptase
Next, we studied whether the effect of tryptase treatment on the ability of the HDL3 particles to act as cellular cholesterol acceptor was related to the degradation of the main HDL apolipoprotein, apoA-I. Aliquots of HDL3 that had been incubated with tryptase (30 μg/mL) at 37°C for 20 hours were analyzed by 15% SDS-PAGE under nonreducing conditions. As shown in Figure 3A, a minor polypeptide band with a molecular mass of ≈26 kDa appeared below the apoA-I band, and moreover, the intensity of the peptides in the size range of 10 to 15 kDa became much stronger (proteolytic products are marked with arrows). Densitometric analysis of the gel demonstrated that ≈25% of apoA-I had been degraded by tryptase treatment. We also studied the effect of tryptase treatment on lipid-free apoA-I by incubating apoA-I (1 mg/mL) with tryptase (30 μg/mL) for 30 minutes (Figure 3B). SDS-PAGE revealed that the tryptase-treated apoA-I contained, in addition to intact apoA-I, a proteolytic fragment with an estimated molecular mass of ≈26 kDa.
Effect of Tryptase on HDL3 Subpopulations
Using Superose 12 gel filtration chromatography, we next studied the size distribution of HDL3 particles before and after incubation with tryptase (Figure 4A). Incubation of HDL3 with or without tryptase was performed in the presence of heparin (10:1); the presence of the added glycosaminoglycan did not influence the elution profile of HDL3 when the eluants were monitored for protein (not shown). After incubation for 20 hours in the absence of tryptase, HDL3 (control HDL3) consisted of a predominant population of larger particles and a minor population of smaller-sized particles. Analysis of the fractions by agarose gel electrophoresis showed that particles in the main peak were α-migrating and those in the minor peak were pre-β-migrating (not shown). Incubation of HDL3 with tryptase did not change the size of the α-HDL peak but fully depleted the pre-β-HDL peak (Figure 4A). Moreover, tryptase treatment generated small peptides that had apparent molecular masses <7 kDa and that were found to be devoid of phospholipids; ie, they were essentially lipid-free apolipoprotein fragments. In addition, we found that proteolysis of HDL3 with tryptase decreased the protein-to-phospholipid mass ratio of the α-HDL from 0.55 to 0.42, reflecting partial loss of apolipoproteins.
Because the major effect of tryptase was observed in the minor pre-β-migrating component of HDL3, we directed our attention to the specific effect of tryptase on the small subpopulations of lipid-poor pre-β-HDL. To prevent undesirable shifts in the charge produced by the glycosaminoglycans during the electrophoretic analysis and given the fact that tryptase degraded HDL3 even in the absence of added aminoglycans (see Figure 1B), incubations were carried out in the absence of heparin. We observed that the anti-apoA-I gave a strong signal with the pre-β-HDL band, reflecting a higher avidity of the antibody against apoA-I in the pre-β-HDL particles than in the α-HDL particles. Moreover, the development time of the spots was adjusted to optimize visualization of the pre-β band. This may also have led to an underestimation of the α-HDL present in the samples. As shown in Figure 4B, on agarose gel electrophoresis and apoA-I immunoblotting of the tryptase-treated HDL3, specific depletion of pre-β-HDL from the HDL3 preparation was observed. Because tryptase depleted the pre-β-migrating HDL3, we further studied its effect on the minor subspecies of lipid-poor HDL, which contain either apoA-I or apoA-IV as the only proteins. This analysis was performed by using nondenaturing 2D gradient gel electrophoresis (Figure 4C). In accordance with the above results demonstrating the effect of tryptase on the pre-β-migrating fraction of HDL, we found that tryptase selectively degraded those apoA-I-containing particles (upper gels) and apoA-IV-containing particles (lower gels) that exhibited pre-β mobility.
In the present study, we demonstrate that human mast cell tryptase depletes HDL3 of the minor subpopulation of pre-β-HDL that contains only apoA-I or only apoA-IV and thereby prevents the high-affinity component of HDL3 from promoting cholesterol efflux from macrophage foam cells.
In the enzymatically active tryptase tetramer, the active sites of the 4 monomers are directed toward a central pore whose narrow pore openings considerably restrict the interaction with macromolecular inhibitors and substrates.13–15⇓⇓ In particular, the plasma anti-proteases and most other proteinase inhibitors are too bulky to fit into the narrow central pore, thus leaving the tryptase secreted by the degranulating mast cells in an enzymatically active state in the extracellular space. On the other hand, the position of the 4 active sites within the central pore of the tetramer allows access of only the substrates that are small and flexible or that present protruding regions to be cleaved from their surfaces.
ApoA-I of HDL3, albeit a polypeptide with 245-amino acid residues, was significantly degraded by tryptase. ApoA-I is a protein that has been predicted to contain a series of amphipathic α-helices, which adopt different conformations according to the degree of apoA-I lipidation.17 Monomeric apoA-I in solution is considered to be loosely folded and ellipsoidal in shape,18 whereas 2 discoidal models have been proposed for apoA-I in nascent HDL particles and are characterized as small unilamellar bilayers surrounded by apoA-I monomers.19,20⇓ It has been suggested that apoA-I in the small and lipid-poor pre-β-HDL exposes surface domains that are more susceptible to protease attack than apoA-I in large spherical α-HDL.7 Such exposure would also explain the susceptibility of apoA-I in the pre-β-HDL to tryptase-mediated degradation. Finally, our findings (ie, those demonstrating that tryptase-dependent proteolysis of HDL3 leads to simultaneous losses of the pre-β-HDL particles and of the high-affinity component of cholesterol efflux) strongly support the notion that the protease-sensitive domains of apoA-I in the pre-β-HDL particles are essential for mediating the unidirectional and high-affinity efflux of cellular cholesterol promoted by HDL3.10,12,21⇓⇓
In contrast to the pre-β-HDL particles, the fully lipidated (spherical) species of HDL in the HDL3 fraction, ie, the α-HDL particles, essentially promote a bidirectional and nonspecific aqueous diffusion mechanism, in which no specific apolipoprotein domains are required, although the presence of the scavenger receptor BI facilitates this process by tethering HDL particles on the surface of macrophages. Thus, limited proteolysis of apoA-I in the α-migrating particles does not modify their size or integrity (Figure 3A) and does not significantly reduce net cholesterol efflux from macrophage foam cells in vitro. Interestingly, we have found that limited proteolysis of HDL3 does not modify the ability of apoA-I as an activator of LCAT.21 However, in our culture system, the HDL3-dependent efflux of cellular cholesterol does not depend on LCAT activity (data not shown), a finding that suggested a minor role for α-HDL in the net cholesterol efflux at low HDL3 concentrations (in the high-affinity range). The fact that there is no LCAT activity in the extracellular fluid of the human arterial intima suggests that the net efflux of cellular cholesterol may largely depend on the maintenance of physiologically active lipid-poor species of HDL, primarily their pre-β-migrating subpopulations.
Tryptase is stored in and released from mast cells together with heparin proteoglycans that stabilize the tetrameric structure of the enzyme and hence the enzymatic activity for a prolonged period of time.15 In the present study, we used commercial heparin glycosaminoglycan to stabilize the isolated tryptase tetramer during incubation periods of up to 24 hours. In accordance with a previous report,22 the glycosaminoglycan-to-tryptase mass ratios ranging from 10:1 to 5:1 produced maximal stabilization, which was reflected in degradation rates that were higher than those with mass ratios of 1:1. Interestingly, tryptase degraded HDL3 more efficiently in the presence of human aortic proteoglycans than in the presence of commercial heparin. This result shows that (like the mast cell-derived heparin proteoglycans) the arterial proteoglycans can also stabilize the tryptase tetramer or may modulate the activity of tryptase, as has previously been reported for mast cell chymase.23 Unexpectedly, even in the absence of exogenous arterial proteoglycans, tryptase caused progressive degradation of HDL3 during a 20-hour incubation, suggesting that the high HDL3 concentrations (1 mg/mL) sustained the enzymatic activity of tryptase against the apolipoproteins of HDL3.
The regulation of tryptase in vivo is not known in detail. A recent study16 that used a physiological system in which tryptase was secreted from stimulated lung mast cells as a tryptase-proteoglycan complex revealed that this complex was stable only if another proteoglycan fraction, secreted by the mast cells but not complexed with tryptase, was also present. Removal of this fraction led to rapid inactivation of tryptase. The results suggest that in vivo after lung mast cell stimulation, the secreted proteoglycans that are not complexed with tryptase become separated from the secreted tryptase-proteoglycan complexes as they diffuse into the large volume of extracellular fluid. Thus, in human tissues, such as the lungs, after mast cell stimulation and exocytosis, the extracellular activity of tryptase is likely to be limited spatially to the microenvironment of each stimulated parent mast cell and limited temporally to a definite and short period of time. Our present observation suggests that in contrast to other tissues, in the subendothelially located proteoglycan-rich layer of the arterial intima, exocytosed tryptase may remain active for prolonged periods of time in a tissue volume that extends beyond the immediate microenvironment of the stimulated mast cell. In plasma, however, other proteins such as antithrombin III compete with tryptase for the binding of heparin, resulting in a rapid dissociation of the tryptase tetramer into enzymatically inactive monomers (that are detected by the ELISAs clinically used). Thus, the arterial intima but not the plasma compartment provides an appropriate environment for the degradation of apolipoproteins by tryptase that may result in a decrease of pre-β1-HDL in the intima rather than in the plasma.
We observed that tryptase is capable, although less so than HDL3, of degrading other plasma lipoproteins, such as the apoB-100-containing lipoproteins VLDL, IDL, and LDL (data not shown). This observation further supports the view that plasma apolipoproteins may be the physiological substrates for human tryptase. However, the apoA-I and apoA-IV present in the small lipid-poor pre-β-migrating HDL appear to be the most susceptible apolipoproteins, thus leading to specific and full depletion of these subpopulations of particles by tryptase under conditions in which the apolipoproteins of other lipoproteins are degraded only to a small degree.
In summary, the present results reveal that mast cell tryptase can degrade HDL3, particularly lipid-free or poorly lipidated apoA-I and apoA-IV present in the minor subpopulation of pre-β-migrating particles, and thereby may alter the functional properties of HDL3 as an efficient acceptor of cellular cholesterol. Because (1) degranulated mast cells present in atherosclerotic lesions release tryptase, (2) the released tryptase is resistant to physiological protease inhibitors, and (3) tryptase is stabilized by arterial proteoglycans, it appears conceivable that in vivo, tryptase-secreting intimal mast cells may reduce the efflux of cholesterol from macrophages and thus contribute to the formation and maintenance of foam cells in atherosclerotic lesions. This novel finding further emphasizes our notion that stimulated mast cells found in human arterial intima11 may play a role in retarding the efflux of cellular cholesterol from macrophage foam cells in vivo.
Financial support was obtained from the European Commission to the consortium (MAFAPS) by grant QLG1-1999-01007 to A.v.E. and P.T.K., from the Sonderforschungsbereich 469 of the University of Munich to C.P.S., and from the Aarne Koskelo and Sigrid Juselius Foundations (Helsinki, Finland) to M.L. We thank Dr Armin Steinmetz (University of Marburg, Marburg, Germany) for providing anti-apoA-IV antibodies and Päivi Ihamuotila and Isa Schaukal for excellent technical assistance.
Received May 27, 2002; revision accepted September 26, 2002.
- ↵von Eckardstein A, Nofer J-R, Assmann G. High density lipoproteins and arteriosclerosis: role of cholesterol efflux and reverse cholesterol transport. Arterioscler Thromb Vasc Biol. 2001; 21: 13–27.
- ↵Kovanen PT. Atheroma formation: defective control in the intimal round-trip of cholesterol. Eur Heart J. 1990; 11 (suppl E): 238–246.
- ↵Rothblat GH, de la Llera-Moya M, Atger V, Weibel GK, Williams DL. Cell cholesterol efflux: integration of old and new observations provides new insights. J Lipid Res. 1999; 40: 781–796.
- ↵Kunitake ST, Chen GC, Kung S-F, Schilling JW, Hardman DA, Kane JP. Pre-β high density lipoprotein: unique disposition of apolipoprotein A-I increases susceptibility to proteolysis. Arteriosclerosis. 1990; 10: 25–30.
- ↵Lindstedt L, Saarinen J, Kalkkinen N, Welgus H, Kovanen PT. Matrix metalloproteinases-3, -7, and -12, but not -9, reduce high density lipoprotein-induced cholesterol efflux from human macrophage foam cells by truncation of the carboxyl terminus of apolipoprotein AI. J Biol Chem. 1999; 274: 22627–22634.
- ↵Lee M, von Eckardstein A, Lindstedt L, Assmann G, Kovanen PT. Depletion of preβ1LpA-I and LpA-IV by mast cell chymase reduces the cholesterol efflux ability promoted by plasma. Arterioscler Thromb Vasc Biol. 1999; 19: 1066–1074.
- ↵Kaartinen M, Penttilä A, Kovanen PT. Mast cells of two types differing in neutral protease composition in the human aortic intima: demonstration of tryptase- and tryptase/chymase-containing mast cells in normal intimas, fatty streaks, and the shoulder region of atheromas. Arterioscler Thromb. 1994; 14: 966–972.
- ↵Lindstedt L, Lee M, Castro GR, Fruchart C-J, Kovanen PT. Chymase in exocytosed rat mast cell granules effectively proteolyzes apolipoprotein AI-containing lipoproteins, so reducing the cholesterol efflux-inducing ability of serum and aortic intimal fluid. J Clin Invest. 1996; 97: 2174–2182.
- ↵Sommerhoff CP, Bode W, Pereira PJB, Stubbs MT, Stürzebecher J, Piechottka GP, Matschiner G, Bergner A. The structure of the human βII-tryptase tetramer: fo(u)r better or worse. Proc Natl Acad Sci U S A. 1999; 96: 10984–10991.
- ↵Segrest JP, Jones MK, Klon AE, Sheldahl CJ, Hellinger M, De Loof H, Harvey SC. A detailed molecular belt model for apolipoprotein A-I in discoidal high density lipoprotein. J Biol Chem. 1999; 274: 31755–31758.
- ↵Lee M, Uboldi P, Giudice D, Catapano AL, Kovanen PT. Identification of domains in apoA-I susceptible to proteolysis by mast cell chymase: implications for HDL function. J Lipid Res. 2000; 41: 975–984.
- ↵Schwartz LB, Bradford TR. Regulation of tryptase from human lung mast cells by heparin: stabilization of the active tetramer. J Biol Chem. 1986; 261: 7372–7379.
- ↵Sommerhoff CP, Ruoss SJ, Caughey GH. Mast cell proteoglycans modulate the secretagogue, proteoglycanase, and amidolytic activities of dog mast cell chymase. J Immunol. 1992; 148: 2859–2866.