Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

LTP I, often referred to as cholesteryl ester transfer protein (Tall et al., 1983), is a glycoprotein with a molecular weight of 74,000 that has been shown to facilitate the transfer of CE, TG and phospholipids between different plasma lipoprotein particles (Morton and Zilversmit, 1982; Morton and Zilversmit, 1983; Morton, 1990). Although, other investigators have presented evidence for separate TG and phospholipid transfer proteins (Rajaram et al., 1980; Jarnagin et al., 1987) that lack the capacity to transfer neutral lipids (Tall et al., 1983; Albers et al., 1984), LTP I remains the best characterized and understood of the lipid transfer proteins in plasma.

Because the human body appears to recognize lipophilic drug compounds as lipid-like particles (Wasan, 1996), it has been hypothesized that an increase in LTP I concentration and activity may facilitate the movement of lipophilic drugs, such as the antifungal agent AmpB, among different lipoprotein classes (Wasan et al., 1993, 1994). We have previously demonstrated that AmpB initially associates with HDL upon incubation in plasma (Wasan et al., 1993, 1994). However, when human plasma was supplemented with exogenous LTP I, AmpB redistributes from HDL to LDL (Wasan et al., 1994). This observation suggests that changes in LTP I concentration may regulate the distribution of AmpB among the different lipoprotein particles within human plasma. This notion was further supported by the work of Hughes et al. (1991) who hypothesized that the plasma lipoprotein distribution of another water insoluble compound, CSA, is determined by factors other than simple diffusion between the lipoprotein particles.

CSA is an effective immunosuppressant used in the treatment of a number of autoimmune diseases as well as in human transplantation (Macoviak et al., 1985; Keown, 1990; Wong et al., 1993). In addition, CSA has been shown to bind with lipoproteins upon incubation in human plasma (Awani and Sawchuk, 1985; Mraz et al., 1983; Sgoutas et al., 1986). We have further shown that changes in the total and plasma lipoprotein lipid concentration and composition influence the lipoprotein binding of CSA (Wasan et al., 1997). One of the proposed biological consequences of CSA binding to lipoproteins is the decrease in the drug's pharmacological effect. Several investigators have reported decreased pharmacological effects of CSA with hyperlipidemia (particularly in hypertriglyceridemia) (Nemunaitis et al., 1986; Kippel et al., 1992), and increased toxic effects of CSA with hypolipidemia (particularly hypocholesterolemia) (de Groen et al., 1987).

Investigations have demonstrated that the cellular uptake of CSA is mediated through HDL (Hughes et al., 1991) and LDL receptors (de Groen, 1988), although recent work has shown that lipoproteins may not serve as a vehicle for the cellular uptake of CSA into hepatic-derived cells (Rifai et al., 1996). However, Lemaire et al. (1988) have suggested that the drug's availability to tissue and hence, its pharmacological (or toxic) effects may depend on which lipoprotein the drug is bound. They have observed enhanced antiproliferative effect of CSA when it was bound to LDL which was not evident when the drug was bound to either VLDL or HDL (Lemaire et al., 1988; Pardridge, 1979). Furthermore, transplantation patients, many who are administered CSA, exhibit plasma dyslipidemias (i.e., lipid disturbances) including hypocholesterolemia and hypertriglyceridemia (Gardier et al., 1993; Arnadottir et al., 1991). In addition, these dyslipidemic plasmas have an elevation in LTP I concentration (Moulin et al., 1992). Thus, determining if LTP I facilitates the binding of CSA to certain lipoproteins may help to explain differences in CSA's pharmacological behavior after administration to hypocholesterolemic (de Groen et al., 1987) and/or hypertriglyceridemic patients (Nemunaitis et al., 1986; Kippel et al., 1992).

The objectives of this study were to determine if LTP I regulates the plasma lipoprotein distribution of CSA and by what mechanisms. We hypothesized that the transfer of CSA between HDL and LDL was a result of direct movement of CSA and/or the cotransport of CSA and CE by LTP I.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals and Plasma

Radiolabeled CSA ([mebmt--³H] cyclosporin A; specific activity, 7.39 mCi/mg) and radiolabeled CE ([1,2 (n)- ³H] cholesteryl oleate; specific activity, 71.9 mCi/mg) were purchased from Amersham Life Science (Buckinghamshire, England). Sodium bromide was purchased from Sigma Chemical Co. (St. Louis, MO). Normolipidemic fasted human plasma was obtained from the Vancouver Red Cross (Vancouver, British Columbia). Ten µl of 0.4 M EDTA pH 7.1 (Sigma) was added to 1.0 ml of whole blood. For all CSA plasma distribution studies, ³H-CSA was dissolved in a 100% ethanol solution. However, the volume of ethanol used did not modify lipoprotein composition or LTP I activity (data not shown).

Lipoprotein Separation

Ultracentrifugation. The plasma was separated into its HDL, LDL, VLDL, and LPDP fractions by ultracentrifugation (Ramaswamy et al., 1997; Havel et al., 1955). Briefly, human plasma (3.0 ml) samples were placed in centrifuge tubes and their solvent densities were adjusted to 1.006 g/ml by sodium bromide. After centrifugation (L8-80 M; Beckman, Toronto, Ontario, Canada) at 50,000 rpm for 18 hours at 4°C the VLDL-rich and VLDL-deficient plasma fractions were recovered. The VLDL-deficient plasma fraction was readjusted to a density of 1.063 g/ml and respun at 50,000 rpm for 18 hours at 4°C to separate the LDL-rich and VLDL/LDL-deficient plasma fractions. This fraction was readjusted to a density of 1.21 g/ml and respun at 50,000 rpm for 18 h at 4°C to separate the HDL-rich and LPDP fractions.

Affinity chromatography and ultracentrifugation. Lipoproteins were separated into the HDL/LPDP and LDL/VLDL fractions by the LDL-Direct cholesterol chromatographic column (Wasan et al., 1993). This chromatographic column is a heparin-manganese polyacrylamide matrix, which separates lipoproteins based on their apolipoprotein content. Any plasma components that contain apolipoproteins B or E are retained by the column while all other components are eluted. Once the gel matrix was fully hydrated with 1 ml of a preparatory solution (0.02% sodium chloride + 0.002% chloramphenicol), plasma samples (200 µl) were placed onto the column followed by an HDL eluting agent (1 ml containing 0.02% sodium chloride + 0.002% chloramphenicol). The flow through fraction, which contains HDL (1.2 ml), was collected. The LDL/VLDL eluting agent (containing 2.9% sodium chloride + 0.002% chloramphenicol) was next placed onto the column and the LDL/VLDL fraction (2.4 ml) was collected. Subsequently, LDL is separated from VLDL and HDL is separated from LPDP by density gradient ultracentrifugation.

Isolation and Purification of LTP I

LTP I was purified from human lipoprotein-deficient plasma as has been previously described (Morton and Zilversmit, 1982). Briefly, citrated human plasma was made lipoprotein-deficient by the dextran-MnCl₂ procedure of Burstein et al. (1970). LTP I was then partially purified by sequential chromatography on phenyl-Sepharose and carboxy-methylcellulose gel (CMC-52, Whatman Inc., Chifton, NJ). Purified LTP I (2.0 mg protein/ml), enriched 800-fold relative to lipoprotein-deficient plasma, was stored at 4°C in 0.01% disodium EDTA pH 7.4. The CMC fraction of LTP I was used in all experiments. Lipid transfer protein I-free CMC solution does not elicit any lipid or drug transfer activity (data not shown).

Radiolabeling of Plasma Lipoproteins

Human HDL and LDL were labeled by the lipid dispersion technique as previously described (Morton and Zilversmit, 1982, 1983). Briefly, human plasma was incubated with a lipid dispersion containing egg PC and ³H-CE (13.9 ng/ml) or ³H-CSA (1000 ng/ml) at 37°C for 20 to 24 hr in the presence of LTP I. Then the HDL and LDL fractions were isolated from the total lipoprotein precipitate by centrifugation as previously described and further purified by dialyzing against PBS solution (4 liters) for 18 hr at 4°C. The molecular weight cut-off of the dialysis tubing used was 1000. After dialysis these lipoprotein fractions were filtered through a 0.2-µ filter. The dialysis and filtration steps were performed to remove any radiolabeled CE or CSA, which has not been incorporated into the core of HDL and LDL.

HDL labeled with ³H-CE had a specific activity of 1.9 × 10³ µCi/10 µg HDL cholesterol while HDL labeled with ³H-CSA had a specific activity of 3.5 × 10² µCi/10 µg HDL cholesterol. Low-density lipoproteins labeled with ³H-CE had a specific activity of 1.6 × 10³ µCi/µg LDL cholesterol while LDL labeled with ³H-CSA had a specific activity of 1.6 × 10² µCi/10 µg LDL cholesterol.

Lipid and Drug Transfer Assays

Lipid (CE) and drug (CSA) transfers were performed within the T150 buffer and lipoprotein-deficient plasma as has been previously described (Wasan et al., 1994; Morton and Zilversmit, 1982; Pattnaik and Zilversmit, 1979). Typically, 10 µg (total cholesterol) of radiolabeled donor and unlabeled acceptor are incubated ± LTP I (1.0 µg protein/ml; concentration was determined from a dose response curve in fig. 2A) in T150 buffer or delapidated human plasma (delipidated human plasma was used as a LTP I source with a concentration of 1.0 µg protein/ml as determined by ELISA), pH 7.4 for 60 min (time was determined from a time response curve in fig. 2B) at 37°C. Lipid and drug transfer between donor and acceptor lipoprotein is then quantitated by scintillation counting. The fraction of lipid and drug transferred (kt) is calculated as described by Pattnaik and Zilversmit (1979): where D_o and A_t are the radioactivity's in the donor at time 0 and in the acceptor at time t, respectively. The constant k is the fraction of label transferred per unit time (t). Acceptor radioactivity in the absence of LTP I (usually < 2-3%) is subtracted before calculating kt values. Calculations assume steady-state conditions where all lipid and drug transfer is an exchange process. To minimize calculation errors due to mass transfer, all values will be determined from assays in which the extent of radio label transfer is small (<15%).

Quantification of CSA and Plasma Lipids

HDL, LDL, VLDL and LPDP fractions were analyzed for ³H-CSA against external standard calibration curves (corrected for quenching and luminescence) using radioactivity. Enzymatic assay kits from Sigma Diagnostics (St. Louis, MO) determined total and lipoprotein triglyceride and cholesterol concentrations.

Calculation of CE Molar Transfer Rates

To determine the molar transfer rates of CE between HDL and LDL a molecular weight of 654 was used for radiolabeled CE. It was assumed that 70% of total cholesterol within each lipoprotein fraction was esterified (CE) (Grundy, 1990). In addition, it was assumed that radiolabeled CE transferred in the same manner as cold CE. The absolute amount of radiolabeled lipid transferred was determined and the transfer rate in pmoles per hour was calculated as follows:

Experimental Design

To provide evidence that LTP I may facilitate the movement of CSA between lipoprotein fractions the lipoprotein distribution of CSA (1000 ng/ml) within human plasma which has been supplemented with exogenous LTP I was determined (fig. 1). In addition, to establish that HDL and LDL have the ability to sequester CSA and CE within their hydrophobic lipid core, radiolabeled CSA and CE were incubated in human plasma and the amount of radiolabeled CSA and CE incorporated into HDL and LDL were determined (table 1).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Percent recovery of Cyclosporine (CSA) in the high-density lipoprotein (HDL)/lipoprotein-deficient (LPDP) [ ] and very-low-density lipoprotein (VLDL)/low-density lipoprotein (LDL) [] fractions within human plasma supplemented with increasing amounts of exogenous lipid transfer protein I (LTP I). Cyclosporine at 1000 ng/ml was incubated in human plasma with or without additional supplementation of LTP I for 60 min at and the percentage of CSA recovered in each of these fractions was determined by radioactivity. Endogenous LTP I concentration was 1.0 µg protein/ml for all test samples. After incubation the plasma was separated into its HDL/LPDP and LDL/VLDL by affinity chromatography and the percentage of CSA recovered in each of these fractions was determined by radioactivity. Data was expressed as mean ± S.D. (n = 6). *P < .05 vs. HDL/LPDP or LDL/VLDL fraction at LTP I = 0.

To investigate the hypotheses that the direct movement of CSA and/or the cotransport of CSA and CE is facilitated by LTP I, the experimental conditions by which the LTP I studies will be investigated under were established (table 2; fig. 2). Furthermore, two strategies that involved the supplementation and inhibition of LTP I were used to test the aforementioned hypotheses.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Net percent transfer (% kt) of radiolabeled cholesteryl ester (3H-CE) from high-density lipoproteins (HDL) to low-density lipoproteins (LDL) after the incubation of 3H-CE-enriched HDL and cold LDL (10 µg total cholesterol content in each lipoprotein fraction) in T150 buffer (A) for 120 min at 37°C which contains an increasing amount of exogenous lipid transfer protein I (LTP I) or (B) for 15, 30, 60 or 120 min at 37°C which contains 1.0 µg protein/ml of exogenous LTP I. Data was expressed as mean ± S.D. (n = 6).

The first strategy was to incubate CSA-enriched HDL or LDL in 50 mM Tris-HCl, 150 mM NaCl, 0.02% sodium azide, 0.01% disodium ETDA (T150 buffer), pH 7.4 which contain a drug-free lipoprotein counterpart in the presence or absence of LTP I (1.0 µg protein/ml). Endogenous LTP I concentration within normolipidemic human plasma is usually 1 to 2 ug protein/ml (Morton and Zilversmit, 1982,1983). In a further experiment LTP I was coincubated with TP2 (4 µg protein/ml) a monoclonal antibody directed against LTP I (Hesler et al., 1988) (figs. 3 and 4). These experiments were designed to further confirm if the movement of CSA between lipoprotein particles was partially facilitated by LTP I and/or a result of nonfacilitated drug transfer rather than the influence of other plasma components (e.g., TG and phospholipid transfer proteins).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Cholesteryl ester (CE) and cyclosporine (CSA) percent transfer from HDL to LDL, in the presence or absence of a monoclonal antibody (TP2) directed against lipid transfer protein I (LTP I), after the incubation of radiolabeled CE- and CSA-enriched HDL with cold LDL (at 10 µg lipoprotein cholesterol) for 60 min at 37°C in T150 buffer that has been supplemented with LTP I (1.0 µg protein/ml) or delipidated human plasma that contains 1.0 µg protein/ml of LTPI. Data was expressed as mean ± S.D. (n = 6). *P < .05 vs. CE or CSA percent transfer without TP2.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Cholesteryl ester (CE) and cyclosporine (CSA) percent transfer from LDL to HDL, in the presence or absence of a monoclonal antibody (TP2) directed against lipid transfer protein I (LTP I), after the incubation of radiolabeled CE- and CSA-enriched LDL with cold HDL (at 10 µg lipoprotein cholesterol) for 60 min at 37°C in T150 buffer that has been supplemented with LTP I (1.0 µg protein/ml) or delipidated human plasma that contains 1.0 µg protein/ml of LTPI. Data were expressed as mean ± S.D.(n = 6). *P < .05 vs. CE or CSA percent transfer without TP2.

A second strategy was to incorporate CSA into HDL and LDL, reisolate these CSA enriched lipoprotein fractions and then incubate these lipoprotein particles in lipoprotein-deficient human plasma in the presence of a drug-free lipoprotein counterpart (e.g., ³H-CSA-HDL and CSA- free LDL) (figs. 3 and 4). The human plasma, which served as the source of LTP I in this experiment, contained a LTP I concentration of 1.0 µg protein/ml as determined by ELISA (data not shown). To confirm that the transfer of CSA is due to LTP I and not other endogenous plasma factors, TP2 (4 µg protein/ml; as determined by ELISA) (data not shown) was coincubated with CSA-enriched and -free lipoprotein particles in plasma (figs. 3 and 4). To assure that the antibody significantly inhibits LTP I activity, CE transfer between HDL and LDL in the presence and absence of TP2 was determined (table 2). These experiments were designed to directly measure the potential role of LTP I in mediating drug transfer vs. the ability of the drug molecules to spontaneously transfer among lipoprotein classes within human plasma.

For all the aforementioned experiments incubations were carried out for 60 min at 37°C. After each incubation the plasma and T150 buffer sample were separated into their individual lipoprotein constituents either by affinity chromatography only, affinity chromatography followed by density gradient ultracentrifugation or density gradient ultracentrifugation and assayed for CSA by radioactivity.

Statistical Analysis

Differences in drug distribution within plasma lipoproteins and differences in LTP I mediated CE and CSA transfer activity in the presence of different treatment groups were determined by a two-way analysis of variance (PCANOVA; Human Systems Dynamics, La Jolla, CA). Critical differences were assessed by Neuman-Keuls post hoc tests. Differences were considered significant if P < .05. All data are expressed as mean ± S.D.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Influence of LTP I on the plasma distribution of CSA. To assess the influence of LTP I on the lipoprotein distribution of CSA within human plasma, radiolabeled CSA (1000 ng/ml) was incubated in human plasma for 60 min at 37°C which had been supplemented with exogenous LTP I (0, .5, 1, 2 µg protein/ml). A significantly greater percentage of CSA was recovered in the HDL/LPDP fraction and a significantly lower percentage of CSA was recovered in the LDL/VLDL fraction in plasma supplemented with LTP I compared to plasma that was not supplemented with exogenous LTP I (fig. 1). The endogenous LTP I concentration for all human plasma used in these studies was 1.0 µg protein/ml as determined by ELISA (data not shown). Because most pharmacokinetic data reported by other investigators have shown that CSA serum concentrations resulting from the daily administration of CSA seldom exceeds 1000 ng/ml, this concentration was chosen for all incubation experiments (Brunner et al., 1990; Sgoutas et al., 1986; Awni et al., 1989). CSA concentrations of 250 and 500 ng/ml were also tested with similar results (data not shown).

CE and CSA transfer between HDL and LDL. To determine the ability of LTP I to promote the transfer of CE and CSA from HDL to LDL, radiolabeled CE- or CSA-enriched HDL and radiolabeled CE- or CSA-free LDL particles were incubated in T150 buffer (which contained 1.0 ug protein/ml of exogenous LTP I) or in LPDP (which contained 1.0 µg protein/ml of endogenous LTP I) for 60 min at 37°C. The percent transfer of CE from HDL to LDL was significantly greater than the percent transfer of CSA in both T150 buffer and human plasma (fig. 3). Furthermore, the percent transfer of CE and CSA was greater in human plasma than in T150 buffer (fig. 3). When the percent transfer of CE and CSA were determined in the presence of TP2 (4 µg protein/ml), the percent transfer of CE was significantly decreased in T150 buffer and human plasma compared to controls (fig. 3). However, the percent transfer of CSA was not significantly different in T150 buffer and human plasma compared to controls (fig. 3).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The objective of this study was to determine the influence of LTP I on the plasma lipoprotein distribution of CSA. Our data suggests that LTP I appear to have a direct role in the distribution of CSA among plasma lipoproteins. This is similar to our observations with AmpB (Wasan et al., 1994; Wasan and Lopez-Berestein, 1995). However, unlike AmpB, the transfer of CSA between HDL and LDL appears to be only partially dependent on LTP I-facilitated transfer of CE.

We have previously demonstrated that the distribution of AmpB among HDL and LDL after incubation in human plasma is facilitated by LTP I. However, once AmpB was incorporated into liposomes composed of negatively charged and neutral phospholipids, the ability of LTP I to transfer AmpB and ³H-CE from HDL to LDL diminished (Wasan et al., 1994; Wasan and Lopez-Berestein, 1995). We concluded from these studies that because AmpB interacts with free cholesterol and CE upon incubation in plasma (Wasan et al., 1993; Bolard et al., 1980), LTP I's ability to transfer AmpB between HDL and LDL was due to its ability to transfer CE between HDL and LDL and not due to the direct transfer of AmpB between lipoprotein fractions. In the case of CSA increases in LTP I concentration resulted in an increased percentage of CSA recovered in the HDL/LPDP fraction during short-term incubations (fig. 1). As LTP I is the protein which catalyzes the transfer exchange of CE from CE-rich lipoproteins (HDL and LDL) for TG from TG-rich lipoproteins (VLDL), these findings suggest that CSA plasma distribution may be related to its lipoprotein-lipid content.

In experiments that were designed to directly measure the potential role of LTP I to facilitate CSA transfer, LTP I-mediated percent transfer of CE among HDL and LDL particles were significantly different than that of CSA (figs. 3 and 4). The differences in the percent transfer of CE vs. CSA may be attributed to an ability of LTP I to transfer lipid and drug separately. Furthermore, differences could be attributed to the ability of HDL and LDL particles to accumulate a higher amount of CE than CSA (e.g., HDL sequesters approximately 1460 ng CE/ng CSA; LDL sequesters approximately 3564 ng CE/ng CSA). Our findings further suggest that HDL particles are much more effective at binding CSA than LDL particles (table 1).

Sgoutas et al. (1986) have proposed that the nature of CSA's association with HDL and LDL particles appears to be nonspecific and of low affinity and high capacity suggesting that CSA is physically dissolved within the lipoprotein-lipid component. Furthermore, because CSA appears to be only partially recognized by LTP I as an endogenous lipid compound, LTP I's ability to transfer CSA between HDL and LDL is only part of the story. This is supported by evidence that demonstrates the percent transfer of CSA from LDL to HDL is significantly greater in human plasma than in T150 buffer (fig. 4) regardless of whether LTP I activity was decreased or not. These findings suggest two possibilities, 1) the spontaneous transfer of CSA and/or 2) the facilitated transfer of CSA by other endogenous plasma factors (e.g., TG and phospholipid transfer proteins). However, the percent transfer of CSA from HDL to LDL, although significantly greater in human plasma than in T150 buffer (fig. 3), does not decrease when a significant reduction in LTP I-mediated CE transfer was observed. These findings further support the notion that the transfer of CSA from HDL to LDL is not LTP I-mediated and may be due to spontaneous and/or facilitated transfer by other endogenous plasma factors. In addition, these results suggest that LTP I may only be partially responsible for the greater capacity of HDL than LDL to accept CSA. Different physical-chemical characteristics of HDLs including lipid composition and overall particle charge may possibly explain CSA's preference to bind with HDL. Studies that investigate these characteristics are currently being completed in our laboratory.

When the molar transfer rates of CE were calculated a number of additional conclusions could be made. The CE molar transfer rate between HDL and LDL was not significantly different in human plasma vs. T150 buffer (fig. 5). However, the percent transfer of CSA from HDL to LDL and LDL to HDL was five to nine times greater respectively in human plasma than T150 buffer (figs. 3 and 4). These observations provide further evidence that CSA transfer is independent of CE transfer and is mediated by plasma factors other than LTP I. Furthermore, when LTP I-mediated transfer of CE between HDL and LDL was inhibited by TP2, only the transfer of CSA from LDL to HDL was significantly decreased. These results suggest that the transfer of CSA from LDL to HDL is only partially facilitated by LTP I, however, the transfer of CSA from HDL to LDL is not facilitated by LTP I but by other plasma factors and/or spontaneous transfer (fig. 6).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Cholesteryl ester (CE) molar transfer rates from HDL to LDL or LDL to HDL after the incubation of radiolabeled CE-enriched HDL or with cold LDL or HDL (10 µg total cholesterol) for 60 min at 37°C in T150 buffer which has been supplemented with lipid transfer protein I (LTP I) (1.0 µg protein/ml) or in delipidated human plasma that contains 1.0 µg protein/ml of LTP I. After incubation the T150 buffer or human plasma was partitioned into its HDL and LDL fraction using precipitation. Data were expressed as mean ± S.D. (n = 6) and the percentage of CSA recovered in each of these lipoprotein fractions was determined by radioactivity.

View larger version (9K): [in this window] [in a new window]   Fig. 6.   The transfer of CSA between HDL and LDL is LTP I dependent and independent. The transfer of CE between HDL and LDL is only LTP I dependent.

In conclusion we have determined that the distribution of CSA among lipoproteins is partially influenced by LTP I. Because many bone marrow transplantation patients exhibit lipid disturbances, including hypocholesterolemia and hypertriglyceridemia, these results may provide an explanation for the unpredictable and inconsistent pharmacokinetics and pharmacodynamics of CSA after administration. Future studies will investigate the pharmacological implications of CSA's predominant association with plasma lipoproteins