Lipopolysaccharide (LPS)-binding protein is carried on lipoproteins and acts as a cofactor in the neutralization of LPS.

Lipoproteins isolated from normal human plasma can bind and neutralize bacterial lipopolysaccharide (LPS) and may represent an important mechanism in host defense against gram-negative septic shock. Recent studies have shown that experimentally elevating the levels of circulating high-density lipoproteins (HDL) provides protection against death in animal models of endotoxic shock. We sought to define the components of HDL that are required for neutralization of LPS. To accomplish this we have studied the functional neutralization of LPS by native and reconstituted HDL using a rapid assay that measures the CD14-dependent activation of leukocyte integrins on human neutrophils. We report here that reconstituted HDL particles (R-HDL), prepared from purified apolipoprotein A-I (apoA-I) combined with phospholipid and free cholesterol, are not sufficient to neutralize the biologic activity of LPS. However, addition of recombinant LPS binding protein (LBP), a protein known to transfer LPS to CD14 and enhance responses of cells to LPS, enabled prompt binding and neutralization of LPS by R-HDL. Thus, LBP appears capable of transferring LPS not only to CD14 but also to lipoprotein particles. In contrast with R-HDL, apoA-I containing lipoproteins (LpA-I) isolated from plasma by selected affinity immunosorption (SAIS) on an anti-apoA-I column, neutralized LPS without addition of exogenous LBP. Several lines of evidence demonstrated that LBP is a constituent of LpA-I in plasma. Passage of plasma over an anti-apoA-I column removed more than 99% of the LBP detectable by ELISA, whereas 31% of the LBP was recovered by elution of the column. Similarly, the ability of plasma to enable activation of neutrophils by LPS (LBP/Septin activity) was depleted and recovered by the same process. Furthermore, an immobilized anti-LBP monoclonal antibody coprecipitated apoA-I. The results described here suggest that in addition to its ability to transfer LPS to CD14, LBP may also transfer LPS to lipoproteins. Since LBP appears to be physically associated with lipoproteins in plasma, it is positioned to play an important role in the neutralization of LPS.

Endotoxin and cytokines increase hepatic messenger RNA levels and serum concentrations of apolipoprotein J (clusterin) in Syrian hamsters.

Infection and inflammation induce alterations in hepatic synthesis and plasma concentrations of the acute phase proteins. Our results show that apolipoprotein (apo) J is a positive acute phase protein. Endotoxin (LPS), tumor necrosis factor (TNF), and interleukin (IL)-1 increased hepatic mRNA and serum protein levels of apo J in Syrian hamsters. Hepatic apo J mRNA levels increased 10- to 15-fold with doses of LPS from 0.1 to 100 micrograms/100 g body weight within 4 h and were elevated for > or = 24 h. Serum apo J concentrations were significantly increased by 16 h and further elevated to 3.3 times that of control, 24 h after LPS administration. Serum apo J was associated with high density lipoprotein and increased fivefold in this fraction, after LPS administration. Hepatic apo J mRNA levels increased 3.5- and 4.6-fold, with TNF and IL-1, respectively, and 8.2-fold with a combination of TNF and IL-1. Serum apo J concentrations were increased 2.3-fold by TNF, 79% by IL-1, and 2.9-fold with a combination of TNF and IL-1. These results demonstrate that apo J is a positive acute phase protein.

Discrimination between subclasses of human high-density lipoproteins by the HDL binding sites of bovine liver.

The binding of human 125I-labeled HDL3 (high-density lipoproteins, rho 1.125-1.210 g/cm3) to a crude membrane fraction prepared from bovine liver closely fit the paradigm expected of a ligand binding to a single class of identical and independent sites, as demonstrated by computer-assisted binding analysis. The dissociation constant (Kd), at both 37 and 4 degrees C, was 2.9 micrograms protein/ml (approx. 2.9 X 10(-8) M); the capacity of the binding sites was 490 ng HDL3 (approx. 4.9 pmol) per mg membrane protein at 37 degrees C and 115 at 4 degrees C. Human low-density lipoproteins (LDL) and very-low-density lipoproteins (VLDL) also bound to these sites (Kd = 41 micrograms protein/ml, approx. 6.7 X 10(-8) M for LDL, and Kd = 5.7 micrograms protein/ml, approx. 7.0 X 10(-9) M for VLDL), but this observation must be considered in light of the fact that the normal circulating concentrations of these lipoproteins are much lower than those of HDL. The binding of 125I-labeled HDL3 to these sites was inhibited only slightly by 1 M NaCl, suggesting the presence of primarily hydrophobic interactions at the recognition site. The binding was not dependent on divalent cations and was not displaceable by heparin; the binding sites were sensitive to both trypsin and pronase. Of exceptional note was the finding that various subclasses of human HDL (including subclasses of immunoaffinity-isolated HDL) displaced 125I-labeled HDL3 from the hepatic HDL binding sites with different apparent affinities, indicating that these sites are capable of recognizing highly specific structural features of ligands. In particular, apolipoprotein A-I-containing lipoproteins with prebeta electrophoretic mobility bound to these sites with a strikingly lower affinity (Kd = 130 micrograms protein/ml) than did the other subclasses of HDL.

Radiation inactivation of binding sites for high-density lipoproteins in human liver membranes.

High-density lipoproteins (HDL) are involved in 'reverse cholesterol transport'. Whether or not cell-surface receptors for HDL exist and participate in this process remains controversial, and part of this controversy has centered on the nature of the HDL binding sites. We therefore used radiation inactivation to determine the molecular mass of the HDL binding sites in human liver membranes in situ. These binding sites, which shared all the characteristics of previously described putative HDL receptors, had a molecular mass of less than 10 kDa, indicating that they are probably not proteins. In addition, the binding of HDL to protein-free liposomes was characterized and was found to display the same affinity (KD = 5 micrograms protein/ml approximately 5.10(-8) M) as that to cell membranes, indicating that HDL binding to cell membranes may not require membrane proteins. These observations highlight an important application of radiation inactivation: the ability to demonstrate that something - in this case, a high-molecular-weight protein that accounts for the majority of the HDL binding activity in human liver membranes - is absent.

Mechanism of the heparin-induced increase in the concentration of free thyroxine in plasma.

The iv administration of heparin causes an increase in the plasma free T4 concentration, as determined by equilibrium dialysis. The mechanism and physiological consequences of this action of heparin are unknown. To explore the possibility that the heparin-induced increase in plasma free T4 is an in vitro artifact due to generation of FFA during equilibrium dialysis, we studied plasma samples from 10 subjects treated with iv heparin. In plasma from 4 of these subjects, free T4 concentrations measured by equilibrium dialysis did not increase above baseline values after heparin administration. In incubations performed in parallel with the equilibrium dialysis measurements, FFA concentrations in these plasma samples were found to increase, but in no subject did they exceed 2.5 meq/L after incubation. In contrast, in plasma from the other 6 subjects treated with heparin, free T4 concentrations rose markedly (by 130-520%) above baseline values after heparin administration. In all of these postheparin plasma samples, FFA concentrations were less than 2.8 meq/L before incubation, but rose during incubation by 80-270% to more than 3.8 meq/L. Treatment of these plasma samples with protamine to inhibit lipoprotein lipase and with specific antiserum to inhibit hepatic triglyceride lipase before equilibrium dialysis or incubation prevented, in parallel, the heparin-induced increases in FFA and free T4 concentrations. From these findings we conclude that the heparin-induced increase in free T4 is usually an in vitro artifact, and that most subjects receiving heparin have a normal plasma free T4 concentration in vivo. We also conclude that this in vitro artifact may account for many of the findings that led to the postulate of an inhibitor of T4 binding to plasma and intracellular proteins in heparin-treated patients and perhaps in patients with nonthyroid illness as well.

High-quality RNA from cells isolated by laser capture microdissection.

Laser capture microdissection (LCM) provides a rapid and simple method for procuring homogeneous populations of cells. However, reproducible isolation of intact RNAfrom these cells can be problematic; the sample may deteriorate before or during sectioning, RNA may degrade during slide staining and LCM, and inadequate extraction and isolation methods may lead to poor recovery. Our report describes an optimized protocol for preparation of frozen sections for LCM using the HistoGene Frozen Section Staining Kit. This slide preparation method is combined with the PicoPure RNA Isolation Kitfor extraction and isolation of RNA from low numbers of microdissected cells. The procedure is easy to perform, rapid, and reproducible. Our results show that the RNA isolated from the LCM samples prepared according to our protocol is of high quality. The RNA maintains its integrity as shown by RT-PCR detection of genes of different abundance levels and by electrophoretic analysis of ribosomal RNA. RNA obtained by this method has also been used to synthesize probes for interrogating cDNA microarray analyses to study expression levels of thousands of genes from LCM samples.

Cell-surface binding sites for high density lipoproteins do not mediate efflux of cholesterol from human fibroblasts in tissue culture.

The present investigation was designed to test the hypothesis that binding sites for high density lipoproteins (HDL3) on cell surfaces of peripheral tissues mediate cholesterol efflux from these cells. This hypothesis had been formulated to explain two observations: 1) HDL3 binding to peripheral cells and HDL3-mediated cholesterol efflux from these cells had both been found to saturate at similar unbound (free) HDL3 concentrations; and 2) both of these processes had been found to be similarly "up-regulated" by loading the cells with cholesterol. In the present study, however, we found that the "specific" binding of HDL3 to cholesterol-loaded human fibroblasts was saturated at a free HDL3 concentration of approximately 20 micrograms protein/ml, whereas efflux of cholesterol from these cells to HDL3 did not "saturate" even at a free HDL3 concentration of 2000 micrograms protein/ml. In addition, we found that the increase in cholesterol efflux caused by loading the fibroblasts with cholesterol was no greater when the acceptor particles were HDL3 than when albumin or phospholipid vesicles served as acceptors, despite a marked increase in HDL3 binding to these cells. Because HDL3 binding to these cells and HDL3-mediated cholesterol efflux from these cells do not saturate at similar free HDL3 concentrations, and because the cholesterol-induced increase in HDL3 binding is not accompanied by a similar increase in cholesterol efflux that is specific for HDL3, we conclude that the described HDL3 binding sites on human fibroblasts do not mediate cholesterol efflux.

Factors affecting the integrity of high density lipoproteins in the ultracentrifuge.

Because of reported losses of apolipoproteins from high density lipoproteins during ultracentrifugation, we studied several factors that could affect the integrity of these lipoprotein complexes. Alteration of temperature, rotor configuration, and composition of the tubes had little effect on loss of apolipoprotein A-I. Interestingly, the high ionic strengths commonly used in ultracentrifugal isolation of these lipoproteins were associated with the smallest loss of apolipoprotein A-I. Losses increased substantially as the ionic strength of the medium was decreased. After repeated ultracentrifugation, apolipoprotein A-I content of high density lipoproteins approached a limiting value of approximately 65% of the original serum value, but no apolipoprotein A-II was lost. Our results imply that the binding environments of these two apolipoproteins in high density lipoproteins differ. Further, they imply that apolipoprotein A-I may exist in more than one type of environment or in more than one form in high density lipoproteins.

Apolipoprotein A-I-containing lipoproteins, with or without apolipoprotein A-II, as progenitors of pre-beta high-density lipoprotein particles.

Apolipoprotein A-I-(apoA-I-) containing lipoproteins isolated by immunoaffinity chromatography can be divided into two general subfractions on the basis of the presence [Lp(AI + AII)] or absence [Lp(AI - AII)] of apoA-II. The Lp(AI - AII) subfraction can be further subfractionated into two subgroups with pre-beta mobility as well as those of alpha mobility. We have characterized the Lp(AI - AII) and Lp(AI + AII) subfractions after the removal of pre-beta high-density lipoproteins (pre-beta-HDL) to compare only the two subfractions with alpha mobility. The Lp(AI - AII) and Lp(AI + AII) of alpha mobility, while both heterogeneous subfractions, share many gross features in common. Both subfractions were predominantly spherical in shape, had similar conformation of apoA-I as investigated by circular dichroism and specific endoproteases, and had similar contents of phospholipids, phospholipid species, triglycerides, and cholesterol ester. However, there was significantly less protein (-10%) and more free cholesterol (+46%) in the Lp(AI - AII) subfraction than in the Lp(AI + AII) subfraction. We investigated the generation of pre-beta-HDL from both the Lp(AI - AII) and Lp(AI + AII) subfractions during incubation with low-density lipoproteins and cholesteryl ester transfer protein. We found that both Lp(AI - AII) and Lp(AI + AII) subfractions were capable of generating pre-beta-HDL-like particles. Our results suggest that the formation of pre-beta-HDL involves dissociation of apoA-I from both Lp(AI - AII) and Lp(AI + AII) subfractions. These results refine a model describing the cycling of apoA-I between pre-beta-HDL and alpha-HDL linked to the movement of cholesteryl esters through HDL.

Interconversion between apolipoprotein A-I-containing lipoproteins of pre-beta and alpha electrophoretic mobilities.

Apolipoprotein (apo) A-I-containing lipoproteins can be separated into two subfractions, pre-beta HDL and alpha HDL (high density lipoproteins), based on differences in their electrophoretic mobility. In this report we present results indicating that these two subfractions are metabolically linked. When plasma was incubated for 2 h at 37 degrees C, apoA-I mass with pre-beta electrophoretic mobility disappeared. This shift in apoA-I mass to alpha electrophoretic mobility was blocked by the addition of either 1.4 mM DTNB or 10 mM menthol to the plasma prior to incubation, suggesting that lecithin:cholesterol acyltransferase (LCAT) activity was involved. There was no change in the electrophoretic mobility of either pre-beta HDL or alpha HDL when they were incubated with cholesterol-loaded fibroblasts. However, after exposure to the fibroblasts, the cholesterol content of the pre-beta HDL did increase approximately sixfold, suggesting that pre-beta HDL can associate with appreciable amounts of cellular cholesterol. Pre-beta HDL-like particles appear to be generated by the incubation of alpha HDL with cholesteryl ester transfer protein (CETP) and either very low density lipoproteins (VLDL) or low density lipoproteins (LDL). This generation of pre-beta HDL-like particles was documented both by immunoelectrophoresis and by molecular sieve chromatography. Based on these findings, we propose a cyclical model in which 1) apoA-I mass moves from pre-beta HDL to alpha HDL in connection with the action of LCAT and the generation of cholesteryl esters within the HDL, and 2) apoA-I moves from alpha HDL to pre-beta HDL in connection with the action of CETP and the movement of cholesteryl esters out of the HDL. Additionally, we propose that the relative plasma concentrations of pre-beta HDL and alpha HDL reflect the movement of cholesteryl esters through the HDL. Conditions that result in the accumulation of HDL cholesteryl esters will be associated with low concentrations of pre-beta HDL, whereas conditions that result in the depletion of HDL cholesteryl esters will be associated with elevated concentrations of pre-beta HDL. This postulate is consistent with published findings in patients with hypertriglyceridemia and LCAT deficiency.

Complementarity of colestipol, niacin, and lovastatin in treatment of severe familial hypercholesterolemia.

OBJECTIVE: To compare the effectiveness of the ternary-drug combination of colestipol, niacin, and lovastatin with binary combinations of those drugs in treating patients with familial hypercholesterolemia. DESIGN: An open sequential study of serum lipoprotein responses in patients receiving diet alone (mean duration, 4 months); colestipol and niacin with diet (mean duration, 9 months); and colestipol, niacin, and lovastatin with diet (mean duration, 15 months). SETTING: Metabolic ward and lipid clinic of a university medical center. PATIENTS: Twenty-two patients with clinical characteristics of familial hypercholesterolemia (low-density-lipoprotein cholesterol, greater than 8.48 mmol/L; 21 of 22 with tendon xanthomas). INTERVENTIONS: Diet: less than 200 mg/d of cholesterol and less than 8% of total calories from saturated fat; colestipol, 30 g/d; lovastatin, 40 to 60 mg/d; and niacin, 1.5 to 7.5 g/d. MEASUREMENTS AND MAIN RESULTS: Mean total serum cholesterol and low-density-lipoprotein cholesterol levels of 4.86 +/- 0.62 mmol/L (188 +/- 24 mg/dL SD) and 2.89 +/- 0.54 mmol/L (112 +/- 21 mg/dL SD), respectively, were significantly lower during ternary-drug treatment than during colestipol-niacin treatment (p less than 0.003) or during treatment in which other possible binary combinations were given. The cholesterol content of very low-density-lipoproteins was lower and high-density-lipoprotein cholesterol levels higher during this phase than during the colestipol-niacin phase. CONCLUSIONS: Colestipol, lovastatin, and niacin are mutually complementary in treating hypercholesterolemia. This regimen produces reductions in serum cholesterol levels similar to those associated with regression of atheromatous plaques in animal studies.

Binding of transition metals by apolipoprotein A-I-containing plasma lipoproteins: inhibition of oxidation of low density lipoproteins.

We have found transition metals tightly bound to apolipoprotein A-I-containing lipoproteins [Lp(A-I)] isolated by selected affinity immunosorption from human serum. Prominent among the metal ions detected were iron and copper. By immunoblotting the proteins of Lp(A-I), we detected both transferrin and ceruloplasmin. The transferrin-containing Lp(A-I) particles, isolated by selected affinity immunosorption against transferrin, were larger (mean diameter of 14.2 nm) and had a higher protein content than most high density lipoproteins (HDL). Ultracentrifugally isolated HDL were found to contain much less transferrin, whereas transferrin was found associated with apolipoprotein A-I from the greater than 1.21-g/ml ultracentrifugal fraction. This suggests that the complex is not recovered in the classic HDL density interval because of its very high density. HDL inhibit copper-catalyzed oxidation of low density lipoproteins (LDL) in vitro. We have found that immunoisolated Lp(A-I) are an order of magnitude more effective in inhibiting the oxidation of LDL than ultracentrifugally isolated HDL, on the basis of protein mass. When the Lp(A-I) particles containing transferrin and ceruloplasmin were removed from the bulk of Lp(A-I), inhibition of the in vitro oxidation of LDL was significantly decreased.

Radiation inactivation of binding sites for high density lipoproteins in human fibroblast membranes.

Radiation inactivation and target analysis were used to determine the molecular mass of the binding sites for high density lipoproteins (HDL) on membranes prepared from human fibroblasts. These membrane binding sites shared characteristics with the previously described HDL binding sites on whole fibroblasts in tissue culture. They exhibited the same affinity for HDL, the same ligand specificity, and the same sensitivity to proteolytic agents. They were also up-regulated by cholesterol loading of the cells. Kinetics of HDL dissociation from membrane binding sites could not be described by a single exponential function, indicating that HDL probably bind to multiple classes of sites on fibroblast membranes. After exposure to ionizing radiation, these sites decreased in number as an apparent single exponential function of radiation dose, corresponding to an average molecular mass of 16,000 +/- 1,000 Da, which is smaller than any known cell-surface receptor protein. These data indicate that HDL binding sites on fibroblast membranes are not "classical" receptors in that they are kinetically heterogeneous and small in molecular mass.

Characteristics of human lipoproteins isolated by selected-affinity immunosorption of apolipoprotein A-I.

We have isolated high density lipoproteins (HDL) from human serum by a new strategy, selected-affinity immunosorption, avoiding perturbation from ultracentrifugation or polyanion precipitation. The principle of this strategy is to utilize a subpopulation of monospecific antibodies directed against apolipoprotein A-I (apo A-I), which was preselected for dissociation under mild conditions of elution. Particles containing more than 90% of the apo A-I contained in serum were isolated uncontaminated by any other serum proteins. The immunoaffinity columns have been cycled over 300 times without apparent diminution in capacity. The apo A-I-containing particles sequestered from serum by immunosorption were more polydisperse in diameter and contained more protein than in ultracentrifugally isolated HDL. They also contained minor apolipoproteins that were not observed in ultracentrifuged HDL. Furthermore, the apo A-I-containing particles had different electrophoretic mobilities from those of ultracentrifugally isolated HDL when run under nondenaturing conditions in polyacrylamide gels. The apo A-I-containing particles isolated by our procedure separate into discrete bands similar to the subspecies visualized when whole serum is subjected to electrophoresis, whereas ultracentrifuged HDL migrate as two broad featureless zones. This suggests that selected affinity immunosorption does not subject the lipoproteins to structural disruption like that which occurs during ultracentrifugation. The principle of selected-affinity immunosorption should be of widespread utility in the isolation of fragile biological complexes.

Physicochemical characterization of ten fractions of bovine alpha lipoproteins.

With the onset of milk production, serum concentrations of alpha lipoproteins in the dairy cow steadily increase, frequently attaining values greater than 1.5 g/dl. Since these lipoproteins comprise a highly polydisperse system, we have carried out studies to explore differences among bovine alpha lipoproteins in the density interval between 1.05 to 1.21 g/ml. Separation into ten fractions was achieved ultracentrifugally in an isopycnic gradient. Agarose gel electrophoresis showed that all but the bottom fraction contained alpha lipoproteins as either the major or sole lipoprotein class. Compositional analyses revealed an increasing percentage of both protein and phospholipid and a decreasing percentage of cholesterol with increasing fraction density. The esterified to unesterified cholesterol ratio ranged from 3 to 8 from the top to the bottom of the gradient. The densities of the particles obtained from the various fractions were calculated both from sedimentation velocity measurements and from compositional analyses. The resulting density values agreed well with the solution densities of these isopycnic gradient fractions. The major apoprotein of each fraction was apoA-I. Combining diffusion coefficient data obtained by intensity fluctuation spectroscopy with sedimentation velocity data, we were able to calculate molecular weights, frictional ratios, and Einstein-Stokes radii for three of the fractions. Results are discussed in terms of previously published data on bovine lipoproteins as well as other mammalian data.

Apolipoprotein A-I-containing lipoproteins with pre-beta electrophoretic mobility.

Among the apoA-I-containing lipoproteins isolated by selected-affinity immunosorption from human serum and plasma, we have identified a subpopulation which, unlike the bulk of high density lipoproteins, has pre-beta electrophoretic mobility. This pre-beta subpopulation can be observed directly in fresh plasma by immunoelectrophoresis. It contains phospholipid and free and esterified cholesterol, but protein constitutes 90% of its mass. Apolipoprotein A-I is the predominant apolipoprotein in this subpopulation; apolipoprotein A-II and the B lipoproteins are not detected. The protein moiety of this subpopulation exhibits markedly lower helicity than that of high density lipoproteins isolated by ultracentrifugation.

Identification of proteins associated with apolipoprotein A-I-containing lipoproteins purified by selected-affinity immunosorption.

The isolation of apolipoprotein A-I-containing lipoproteins [Lp(A-I)] by selected-affinity immunosorption minimizes the loss of associated proteins that occurs during the isolation of high-density lipoproteins (HDL) by sequential ultracentrifugation. We have used two-dimensional gel electrophoretic analysis to separate the proteins associated with Lp(A-I). Using a combination of amino acid sequencing of transblotted proteins and Western blotting with specific antisera, we have identified a number of associated proteins. The positions of the apolipoproteins (apo) A-I, A-II, A-IV, C-III, D, and E were located on the gels. Lecithin-cholesterol acyltransferase and cholesteryl ester transfer protein were identified in association with Lp(A-I) to a greater extent than found associated with HDL after centrifugation. In addition to those proteins previously identified in association with HDL, we detected a number of plasma proteins associated with Lp(A-I), namely, fibrinogen, haptoglobin, proline-rich protein (C4b-binding protein), and apolipoprotein J (SP40,40 sulfated glycoprotein). The co-isolation of these proteins with Lp(A-I) does not appear to be an artifact in that they have very low affinity for a sham column containing covalently bound preimmune goat IgG in place of the anti-apoA-I IgG. These findings suggest that in addition to apolipoproteins that exist largely in association with lipoproteins there is another class of proteins which exist in lipoprotein-associated form and in the dispersed state. Detection and identification of these lipoprotein-associated proteins may aid in the mechanistic determination of a number of observed functions attributed to HDL.

Enumeration of CD34+ hematopoietic progenitor cells in peripheral blood and leukapheresis products by microvolume fluorimetry: a comparison with flow cytometry.

There is increasing interest in both standardization and simplification of methods for enumeration of CD34+ hematopoietic progenitor cells (HPC) to facilitate cellular therapies and to improve interinstitutional comparison of clinical and laboratory results. We evaluated a novel method for CD34+ cell enumeration based on microvolume fluorimetry (MVF) compared with our laboratory's routine flow cytometric method on samples of peripheral blood and leukapheresis products. The MVF method is semiautomated and uses a 633-nm light from a helium-neon laser to scan fluorochrome-labeled cells held in stasis in a capillary known volume. The performance of the MVF assay for enumeration of CD34+ cells was found to be comparable to our routine flow cytometric assay in linearity and accuracy in the range of 5-1500 cells per microliter. Precision of MVF for replicate assays on the same instrument was demonstrated by coefficient of variation (CV) values of 8.4% at a CD34+ cell concentration of 284/microliters for a sample volume of 0.8 microliters, and 15.7% at 12/microliters for a sample volume of 3.2 microliter. Precision among three different instruments was demonstrated, using sample volumes of 1.6 microliters, by CV values of 44% at 6 cells/microliters and 4.6% at 733 cells/microliters. In a field sample evaluation, precision of the entire assay system for paired measurements on 0.8-microliter sample volumes was demonstrated by CV values of 50%, 31%, and 15% for peripheral blood samples with concentrations of 0-10, 10-20, and 20-100 CD34+ cells/microliters, respectively, and 6.3%, 8.1% and 6.5% for leukapheresis samples with concentrations of 0-100, 100-1,000, and 1,000-2,500 CD34+ cells/microliters, respectively. The MVF assay was easy to perform, required minimal technical training time, and had a turnaround time of 40 min, of which less than 10 min was actual technical time. These observations suggest that the MVF method for CD34+ cell enumeration may prove useful to clinical laboratories providing support for HPC collection, processing, and transplantation services that require relatively simple, rapid assays for product quality control or to guide real-time clinical decisions.

Isolation of subpopulations of high density lipoproteins: three particle species containing apoE and two species devoid of apoE that have affinity for heparin.

We have isolated and partially characterized five populations of lipoproteins from the pool of immunoisolated apoA-I-containing lipoproteins obtained from normal human plasma. The first three populations, each containing apoA-I and apoE, were isolated completely by sequential, selected affinity immunosorption against apoA-I and apoE. The lipoproteins isolated by this strategy fall into three morphologic groups; there are discs (LP-AI-E(1)), small spherical lipoproteins (LP-AI-E(2)), and large spherical lipoproteins (LP-AI-E(3)). The LP-AI-E(2) species was sufficiently abundant for detailed characterization. They have slightly larger diameters, and contain more lipid than the bulk of apoA-I-containing lipoproteins and they contain apoA-II:E heterodimers and apoE homodimers. Core lipids are enriched in triglyceride relative to cholesteryl esters. These lipoproteins compete with LDL equally, on a protein mass basis, for binding to human fibroblasts. After removal of apoE-containing lipoproteins from the pool of apoA-I-containing lipoproteins, we discovered two additional subpopulations of lipoproteins that bind to heparin. These lipoproteins, devoid of apoE, occur as populations of small, (LP-AI-HB(1)), and large, spherical lipoproteins, (LP-AI-HB(2). The heparin-binding lipoproteins were separated by gel permeation chromatography. The LP-AI-HB(1) population was of sufficient quantity for detailed study. These lipoproteins also had larger diameters than the bulk of HDL but their core lipids were enriched in cholesteryl esters rather than triglycerides. Three proteins associated with these lipoproteins were found to bind to heparin-Sepharose in the absence of lipid. The approximate molecular weights of these proteins are 40, 70, and 90 kDa. The 70 kDa molecule was found to be the SP 40,40 protein (apoJ).