Anal. Chem. 2006, 78, 680-685
Articles
Remnant Lipoprotein Density Profiling by CsBiEDTA Density Gradient Ultracentrifugation Richa Chandra and Ronald D. Macfarlane*
Department of Chemistry, Texas A&M University, 3255 TAMU, College Station, Texas 77843
Remnant lipoproteins (RLPs) are now considered a strong marker of the triglyceride-rich lipoprotein (TRL) class for cardiovascular heart disease. The purpose of this research is to demonstrate the efficacy of a novel method that combines an established immunoseparation assay used to measure the RLP class in human serum with ultracentrifugal density gradient separation. These two methods are combined to obtain an RLP density profile. The immunoseparation effectively removes the non-RLP lipoproteins from serum. The RLPs obtained from the immunoseparation are separated into two density-distinct fractions by ultracentrifugal density gradient separation in CsBiEDTA. It is now clear that IDL is distinct in density and immunoreactivity from the two RLP classes isolated by the immunoseparation and ultracentrifugation. This methodology defines the RLP by density and measures their relative prevalence in the TRL class. When applied to clinical samples, variations in the RLP subclasses in different patients are examined. The differences in the RLP density profile are also examined in fasting and postprandial samples. The RLP density profile significantly increases in the postprandial state versus the fasting state. However, the overall quantity of TRL does not appreciably increase in the postprandial state. This work demonstrates the feasibility of measuring the postprandial state in clinical samples to provide insight into the clearance of RLP by the liver as well as the general atherogenicity of these particles. The major outcome of this research is a novel analytical method that couples immunoseparation and density gradient ultracentrifugation to separate and differentially profile the RLP subclass against its nascent counterparts in the TRL class. The National Cholesterol Education ProgramsAdult Treatment Panel III listed remnant lipoproteins (RLPs), a subclass of the lipoprotein population, as an emerging atherogenic risk factor in 2001.1 The purpose of this investigation is to develop an analytical method for the measurement of the RLP density profile * To whom correspondnece should be addressed. E-mail: macfarlane@ mail.chem.tamu.edu. Phone: (979) 845-2021. Fax: (979) 845-8987. (1) Expert Panel on Detection, E. and Treatment of High Blood Cholesterol in Adults. J. Am. Med. Assoc. 2001, 285, 2486.
680 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
in CsBiEDTA by density gradient ultracentrifugation. We are using a modified version of a differential immunoseparation method that was developed by Nakajima et al.2 An additional ultracentrifugal separation is applied to the immunoseparation to further separate the density-distinct fractions of the triglyceride-rich lipoprotein (TRL) remnants from each other and serum proteins. By comparing the density profiles of the TRL remnants and total serum lipoproteins, the fraction of TRL that consists of RLP can be determined. This analysis is applied clinically to various patient samples to examine differences between patients. In addition, fasting and postprandial samples of patients can be analyzed to qualitatively examine the plasma residence time of these dietderived particles. RLP particles can be divided into two subclasses: very lowdensity lipoprotein (VLDL) remnants (rVLDL) and chylomicron remnants (CMR).3 In this study, we examine the features of a class of RLP particles that are defined according to Nakajima’s assay for differential immunoseparation of monoclonal antibodies to apolipoprotein B-100 (apo B-100) and apolipoprotein A-1 (apo A-1). This assay is approved by the Food and Drug Administration and predominates recent literature.2,4-9 Recent evidence indicates that RLPs, the diet-derived fraction of TRL, are the atherogenic component of the TRL class.9-11 Triglyceride-rich lipoproteins consist of chylomicrons (CM), CMR, VLDLs, rVLDLs, and intermediate density lipoproteins (IDLs).12 The methods used to analyze RLP are a diverse array including gel electrophoresis, ultracentrifugation, capillary electrophoresis, (2) Nakajima, K.; Okazaki, M.; Tanaka, A.; Pullinger, C. R.; Wang, T.; Nakano, T.; Masakazu, A.; Havel, R. J. J. Clin. Ligand Assay 1996, 19, 177. (3) Marcoux, C.; Hopkins, P. N.; Wang, T.; Leary, E. T.; Nakajima, K.; Davignon, J.; Cohn, J. S. J. Lipid Res. 2000, 41, 1428. (4) Chan, D. C.; Watts, G. F.; Barrett, P. H.; Mamo, J. C. L.; Redgrave, T. G. Clin. Chem. 2002, 48, 278. (5) Karpe, F.; Boquist, S.; Tang, R.; Bond, G. M.; de Faire, U.; Hamsten, A. J. Lipid Res. 2001, 42, 17. (6) Masuoka, H.; Kamei, S.; Ozaki, M.; Kawasaki, A.; Shintani, U.; Ito, M.; Nakano, T. Intern. Med. 2000, 39, 540. (7) Campos, E.; Nakajima, K.; Tanaka, A. J. Lipid Res. 1992, 33, 369. (8) Jialal, I.; Devaraj, S. Clin. Chem. 2002, 48, 217. (9) Twickler, T. B.; Dallinga-Thie, G. M.; Cohn, J. S.; Chapman, M. J. Circulation 2004, 109, 1918-1925. (10) Havel, R. J.; Rifai, N.; Warnick, G. R.; Dominiczak, M. H. In Handbook of Lipoprotein Testing; AACC Press: Washington DC, 1997; p 451. (11) Havel, R. J. Am. J. Clin. Nutr. 1994, 59, 795. (12) Smith, D.; Watts, G. F.; Dane-Stewart, C.; Mamo, J. C. L. Eur. J. Clin. Invest. 1999, 29, 204. 10.1021/ac050775w CCC: $33.50
© 2006 American Chemical Society Published on Web 12/22/2005
and immunoprecipitation. Agarose gel electrophoresis and polyacrylamide gel electrophoresis (PAGE) are useful for detection of RLPs and their apolipoproteins.13 Ultracentrifugation allows for the partial separation of RLPs with significant contamination from other serum components.14,15 IDL is separated by ultracentrifugation between the densities of VLDL and LDL (1.006 < d < 1.019 g/mL). IDL isolated in this manner does not include the less catabolized and more triglyceride-rich RLPs.13 As described before, the immunoprecipitation assay developed by Nakajima et al. defines RLPs for this study. The assay contains the monoclonal JI-H antibody to apo B-100 as well as an antibody to apo A-1. The assay effectively removes all HDL, CM, LDL and most VLDL. VLDL remnants are nonimmunoreactive to the JI-H antibody. This antibody is reactive to apo B-100, a surface protein, on only LDL and some VLDL. Chylomicrons and their remnants are also defined by their differential immunoreactivity to this antibody since these particles contain apo B-48, a truncated version of apo B-100. The JI-H antibody does not recognize apo B-48 since it reacts with an epitope beyond the C-terminus. In addition, CMR can be separated from CM because they do not contain an apo A-1.2,8,11,14,16 The aim of this work is to show the efficacy of the modified version of Nakajima’s immunoseparation assay in the separation of RLPs from serum when used in conjunction with ultracentrifugal density gradient separation in CsBiEDTA. This method is applied to clinical samples to examine patient variability as well as the differences between fasting and postprandial samples. In this study, we examine the effective removal of the CM, VLDL, LDL, and HDL classes and the subsequent separation of the remaining RLPs by CsBiEDTA density gradient ultracentrifugation to obtain an RLP density gradient profile. RLP density profiles of three different subjects are presented here. The following protocol has also been successfully applied to remove CM, VLDL, LDL, and HDL classes and to obtain RLP density profiles in other subjects not presented here. EXPERIMENTAL SECTION Materials. The RLP immunoseparation gel containing antibodies to apo B-100 and apo A-1 on Sepharose beads and RLP buffer (5 mM tris-HCl) was purchased from Polymedco, Inc. (Cortlandt Manor, NY). NBD C6-ceramide was purchased from Molecular Probes (Eugene, OR). An aqueous solution of CsBiEDTA was synthesized as described previously.17 The 3-8% NuPAGE trisacetate sodium dodecyl sulfate (SDS)-PAGE minigels, Novex trisacetate SDS running buffer (20×), and NuPAGE LDS sample buffer (4×) were purchased from Invitrogen (Carlsbad, CA). The phosphorylase b, cross-linked molecular weight markers (Catalog No, 9012-69-5), 2-mercaptoethanol, and an apo B-100 standard were purchased from Sigma-Aldrich (St. Louis, MO). Subject Selection. Serum samples were selected from a serum library having the following features as determined by (13) Cohn, J. S.; Marcoux, C.; Davignon, J. Arterioscler., Thromb., and Vasc. Biol. 1999, 19, 2474. (14) Devaraj, S.; Vega, G.; Lange, R.; Grundy, S. M.; Jialal, I. Am. J. Med. 1998, 104, 445. (15) Lovegrove, J. A.; Isherwood, S. G.; Jackson, K. G.; Williams, C. M.; Gould, B. J. Biochim. Biophys. Acta 1996, 1301, 221. (16) Hirany, S.; O’Byrne, D.; Devaraj, S.; Jialal, I. Clin. Chem. 2000, 46, 667. (17) Hosken, B. D.; Cockrill, S. L.; Macfarlane, R. D. Anal. Chem. 2005, 77, 200-207.
lipoprotein density profiling: elevated dense TRL and IDL levels. Serum samples were collected from donors in Vacutainer-brand serum collection tubes (Beckton Dickinson Systems, Franklin Lakes, NJ). Serum was stored at -86 °C until needed. Immunoseparation Assay. The RLP gel was prepared according to instructions provided by the manufacturer with some modifications. The total volume of the assay was altered to 900 µL, and the assay was performed in 1.5-mL Eppendorf tubes. For each assay, 60 µL of serum and 2 µL of a 2 mg/mL NBD C6ceramide in DMSO were added to the RLP gel and RLP buffer prior to incubation. The tubes were then placed on an M-60 orbital mixer (Labnet International, Inc.) and incubated at 1400 rpm for 2 h. Following the incubation, the RLP was recovered by removal of 660 µL of the supernatant by pipet. Ultrafiltration. Ultrafiltration of the recovered RLP was performed in a 100 000 NMWCO centrifugal filter (Millipore) mounted in a 1.5-mL Eppendorf tube according to manufacturer’s instructions. Approximately 60 µL of RLP and serum proteins was recovered. Ultracentrifugation. The density gradient ultracentrifugation protocol using 0.240 M CsBiEDTA as the density gradient forming solute described previously was followed for both total serum and RLP samples with minor modifications.17 For serum density profiling, 60 µL of serum and 2 µL of the 2 mg/mL NBD C6-ceramide in DMSO were diluted to 140 µL in H2O followed by the addition of 1100 µL of the CsBiEDTA. A 1000-µL aliquot of this solution was ultracentrifuged at 120 000 rpm for 6 h at 5°C in an Optima TLX ultracentrifuge, TLA 120.2 fixed-angle rotor (Beckman-Coulter, Palo Alto, CA). Similarly for the RLP density profiling, 60 µL of the RLP (concentrated by ultrafiltration) was diluted to 140 µL in H2O followed by the addition of 1100 µL of the CsBiEDTA. A 1000-µL aliquot of this solution was also ultracentrifuged under the same conditions as the serum. Following the spin, 200 µL of H2O was layered on top of the tube contents for each sample. An image of the tube was obtained and analyzed using a digital Optronics Microfire Camera (Goleta, CA) with a MH-100 metal halide continuous light source. Two filters matching the excitation and emission characteristics of NBD C6ceramide from Schott Glass (Elmsford, NY) were used. A blueviolet filter (centered at 407 nm) and a yellow emission filter (centered at 570 nm) were used as the excitation and emission filters, respectively. The image of the tube following ultracentrifugation was then converted to a density profile using the described method.18 This entire protocol was applied in the same manner to all subjects to obtain the total serum and RLP density profiles. Gel Electrophoresis. SDS-PAGE was performed on a 3-8% tris-acetate gradient slab minigel (Invitrogen). For subjects 1 and 3, the two TRL fractions and the LDL fraction from subject 1 were collected from the polycarbonate tube by a freeze/slice method where the tube was first frozen in liquid nitrogen and then cut for collection. For subject 3, one tube was layered with 200 µL of a 50% methanol in water solution (F ) 0.925 g/mL) instead of water to evaluate the buoyant TRL fraction separated at this density (Figure 4, lane g, subject 3 buoyant TRL-50% CH3OH in H2O). All of the above fractions were then concentrated by ultrafiltration as described above according to manufacturer’s (18) Johnson, J. D.; Bell, N. J.; Donahoe, E. L.; Macfarlane, R. D. Anal. Chem. 2005, 77, 7054-7061.
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
681
instructions in 100 000 NMWCO centrifugal filters (Millipore). The fractions were then prepared for SDS-PAGE under reduction conditions in LDS sample buffer following the manufacturer’s instructions. A high molecular weight phosphorylase b standard and an apo B-100 standard were prepared in the same manner and used to determine identities of protein bands in the TRL fractions. Briefly, 10 µL of the sample was mixed with 9 µL of sample buffer and 1 µL of 10% (v/v) 2-mercaptoethanol in water. The samples were reduced by incubation for 10 min at 70 °C. Following the electrophoresis, the gel was silver stained by the standard nonfixing silver staining method of Blum et al.19 Figure 4 depicts the protein identification of the TRL fractions and the LDL fraction in comparison to the phosphorylase b and apo B-100 standards. RESULTS AND DISCUSSION Figure 1a is a density profile of serum components from subject 1 in a CsBiEDTA gradient. This profile shows the typical features of a serum density profile including serum proteins at a tube coordinate of 28 mm, HDL subclasses from 21 to 26 mm, LDL centered at 18 mm, IDL at 13 mm, and the meniscus containing buoyant and dense TRL at ∼10 mm. As shown in Figure 1b, the buoyant and dense TRL are resolved by layering 200 µL of H2O shifting the meniscus ∼4 mm. The more buoyant and hence larger TRL separate from the less buoyant, smaller TRL components upon layering of water. The two fractions at the top of the tube consist of CM and VLDL with the meniscus containing the more buoyant, triglyceride-rich portion leaving behind the denser and less buoyant counterparts of the TRL class. The more buoyant TRLs immediately follow the meniscus, whereas the denser TRLs remain behind because of lesser flotation velocity. The meniscus region does not conform to the density gradient of the CsBiEDTA in the rest of the tube. The density of this region is consistent with H2O and is labeled as such. Lipoprotein Particle Density Profiles of the Study Subjects. Figure 1b illustrates a density profile in CsBiEDTA of a normolipidemic subject (subject 1) with features of minimal dense TRL and IDL. The features of a typical lipoprotein density profile in a CsBiEDTA gradient include the meniscus (buoyant TRL), dense TRL, IDL, LDL, and HDL regions primarily. Fluorescence arising from the IDL fraction is detected above the background fluorescence at density values between the dense TRL and LDL fraction. Another unique characteristic of this profile is the apparent HDL2 and HDL3 subclasses of the HDL region. At the bottom of the tube at 28 mm, serum proteins, mainly albumin, separate from the lipoprotein classes. A discontinuity dividing this fraction originates from scattering from a seam on the surface of the polycarbonate tube and is not due to any serum component. Density Profile of RLP in CsBiEDTA. Figure 1c is the RLP profile of subject 1. This RLP profile shows an effective removal of the LDL and IDL subclasses by the RLP immunoseparation assay. However, ∼5% of the HDL is not removed by the immunoseparation. This may be due to oversaturation of the apo A1 antibody or possibly a unique immunoreactivity of subject 1’s HDL. Again, by adding H2O immediately following the ultracentrifugation spin, a portion of the more buoyant TRL migrates in the shifting meniscus at the density of water. Hence, two density(19) Blum, H.; Beier, H.; Gross, H. J. Electrophoresis 1987, 8, 93-99.
682 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
Figure 1. (a) Lipoprotein density profile (unlayered) for subject 1 in CsBiEDTA. Lipoprotein density profile (b) and RLP density profile (c) for subject 1 in CsBiEDTA following the layering of H2O.
distinct RLP fractions remain at the corresponding buoyant and dense TRL densities. The RLP at the meniscus contain both buoyant CMR and buoyant rVLDL as is evidenced by apolipoprotein content of the meniscus fractions for subjects 1 and 3 in the SDS-PAGE analysis presented in Figure 4. The RLP at the dense TRL density contain both dense CMR and dense rVLDL. In addition, the serum proteins that are nonreactive to the antibodies of the immunoseparation assay also separate from the RLP fractions. The percent of RLP in total TRL at each density is relatively minimal. The buoyant TRL at the meniscus consist of only 0.041% buoyant RLP, and the dense TRL consist of 17.809% dense RLP. The RLP that separates at the IDL density is insignificant as it is almost completely removed.
Figure 2. Lipoprotein density profile (a) and RLP density profile (b) for subject 2 in CsBiEDTA following the layering of H2O.
Figure 2a is an example of a profile that satisfies the criteria of elevated dense TRL and the presence of IDL (subject 2). These characteristics are typically good indicators of elevated RLP. Many features of this profile are similar to subject 1’s profile in Figure 1b. Both profiles exhibit lipoprotein fractions at the meniscus (buoyant TRL), dense TRL, IDL, LDL, and HDL densities. However, subject 2’s profile features elevated dense TRL and IDL in contrast to subject 1’s profile. Subject 2’s RLP profile in Figure 2b dramatically demonstrates the efficacy of the immunoseparation assay in its removal of all non-RLP lipoprotein fractions. Again, two density-distinct RLP fractions remain in the profile. The buoyant TRLs at the meniscus consist of 22.9% buoyant RLP content. The dense RLP at the dense TRL density is 28.6% of the nascent fraction. It is interesting to note that lipoprotein particles at the density value of IDL are not a significant feature of the RLP density profiles. The RLP fraction at the IDL density is only 3.5% of the nascent particles in subject 2’s serum. Postprandial versus Fasting Samples. The RLP immunoseparation assay and the ultracentrifugation protocol in CsBiEDTA were performed on fasting and postprandial samples from subject 3. The postprandial serum sample from this subject was obtained 5 h after ingestion of a high-fat content, fast food burger. The fasting serum sample was obtained prior to the meal. The RLP profiles for both samples are overlaid with the density profiles for the fasting in Figure 3a and postprandial serum samples in Figure 3b. Once more, the RLP fractions separate at density values corresponding to their nascent counterparts, and minimal RLP separates at the density corresponding to IDL. Subject 3’s RLP profiles reconfirm the effectiveness of the removal of non-RLP
Figure 3. (a) Lipoprotein (red line) and RLP density profiles (black line) for subject 3 in the fasting state in CsBiEDTA following the layering of H2O. (b) Lipoprotein (red line) and RLP density profiles (black line) for subject 3 in the postprandial state in CsBiEDTA following the layering of H2O. Table 1. Percentage of RLP at TRL and IDL Densities for Fasting and Postprandial Samples of Subject 3 samples
integrated area, %
Fasting Samples buoyant RLP/buoyant TRL dense RLP/dense TRL RLP-IDL (IDL + RLP-IDL)
13.5 12.6 0
Postprandial Samples buoyant RLP/buoyant TRL dense RLP/dense TRL RLP-IDL (IDL + RLP-IDL)
29.4 53.5 5.8
lipoproteins by the immunoseparation. In addition, there is a marked increase in dense RLP from the fasting to postprandial samples. An interesting finding is the fact that the overall dense TRL content does not appreciably increase from the fasting to postprandial state according to our results. The percentages of RLP/TRL at each density for the fasting and postprandial samples from subject 3 are tabulated in Table 1. These numbers indicate the extent of the increase in RLP in the postprandial state. The data provide the ratio of the remnant particle of interest to its corresponding nascent TRL particle plus the remnant itself. For example, the ratio of “buoyant RLP/buoyant TRL” signifies the Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
683
Figure 4. SDS-PAGE analysis of (a) phosphorylase b standard, (b) apo B-100 standard, (c) subject 1 buoyant TRL, (d) subject 1 dense TRL, (e) subject 1-LDL, (f) subject 3 buoyant TRL (water), (g) subject 3 buoyant TRL (50% CH3OH in H2O), and (h) subject 3 dense TRL in a 3-8% tris-acetate gel.
ratio of the buoyant remnant particles to the nascent buoyant TRL particles plus the buoyant remnant particle itself. This is definitive evidence of the presence of RLP in circulation following a fatcontaining meal. The residence and dynamics of clearance of the buoyant and dense RLPs in the hours following a meal can potentially be measured by this new methodology. This would provide insight into the clearance rate of RLP by the liver in a particular subject. It could also provide useful information in regard to the atherogenicity of these particles and in the assessment of an individual’s risk factor for developing coronary heart disease. Gel Electrophoresis Analysis of Apolipoprotein Content in TRL. The buoyant and dense TRL fractions were identified for their apolipoprotein content by SDS-PAGE analysis as illustrated in Figure 4. Lanes a and b are the phosphorylase b and apo B-100 standards, respectively. These standards were used to assign molecular weights to the apo B-100 and apo B-48 bands in the TRL and LDL fractions. The molecular weights corresponding to the phosphorylase b standard are listed to the left of the gel image in Figure 4. The apo B-100 standard has a molecular weight of 515 000 as shown in Figure 4, lane b. Lanes c and f show the apo B-48 and apo B-100 content of the buoyant TRL class in the meniscus separated by the layering of water in subjects 1 and 3, respectively. Lanes d and h show the apo B-48 and apo B-100 content in the dense TRL class in subjects 1 and 3, respectively. Lane e is the LDL fraction from subject 1 and was used as an experimental control. This fraction contains a large amount of apo B-100 and no apo B-48 as is expected of LDL particles. Finally, lane g is a meniscus separation of TRL from subject 3 by layering of 50% CH3OH in water (F ) 0.925). This separation shows great potential of further separating buoyant TRL 684 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
from dense TRL in that the apo B-100 content was reduced in this fraction in comparison to the meniscus separation of buoyant TRL in water (Figure 4, lane f). CONCLUSIONS The coupling of the immunoseparation assay with ultracentrifugal density gradient separation is an important analytical method. Even though the immunoseparation assay developed by Nakajima et al. was designed for the detection of RLPcholesterol, it is an excellent assay for the detection of the RLP density profile.2 One of the most interesting discoveries that arises from RLP density profiling is the fact that the IDL lipoprotein class is not immunoreactive to the immunoseparation assay, which separates out the proatherogenic component of TRL. This may mean that IDL does not have as much atherogenic potential as RLP. IDL is often described synonymously with RLP, and we now know that the two density-distinct RLP classes are distinct in their immunoreactivity and unique in density when compared to IDL. It is now possible to separate two density-distinct subclasses of RLP for further analyses. Furthermore, we prove the feasibility of studying the postprandial condition. It is of particular interest to study postprandial samples in which RLPs are elevated to examine their clearance and properties. This novel method will potentially provide insight into the disorders of RLP metabolism as well as provide a potential diagnostic tool in the clinical setting. ACKNOWLEDGMENT This work was supported by the NIH, Heart, Lung & Blood Institute (HL 068794). We thank I. Leticia Espinosa, Ronald R.
Henriquez, and Jeffery D. Johnson for their valuable assistance and support.
was applied to the serum samples of the subjects. This material is available free of charge via the Internet at http://pubs.acs.org.
SUPPORTING INFORMATION AVAILABLE Figure presenting RLP density profiles (s) and serum density profiles (s) of other subjects not included in the article. The same protocol described in the Experimental Section of the paper
Received for review May 5, 2005. Accepted November 10, 2005. AC050775W
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
685
