Bound homocysteine, cysteine, and cysteinylglycine distribution between albumin and globulins

BACKGROUND: Major portions of homocysteine (Hcy), cysteine (Cys), cysteinylglycine (CysGly), and glutathione in serum are covalently bound to proteins via disulfides. Albumin has been considered the dominant binding protein. METHODS: Pooled serum and plasma from healthy adults were fractionated into albumin and globulins by affinity columns. Content of Hcy, Cys, CysGly, and glutathione was determined for serum and plasma fractions and purified proteins by an HPLC method before and after incubation with excess CysGly, Hcy, or glutathione RESULTS: Of protein-bound amino acids in pooled serum, 12% of Hcy, 21% of Cys, and 33% of CysGly were bound to globulins, with the remainder bound to albumin. Slightly higher proportions were bound to globulins in pooled plasma. Globulins had approximately 16% of total exchangeable disulfide and thiol groups in serum based on results of loading with CysGly. These results agree with expected abundance of unpaired Cys residues in globulins relative to albumin. Significant amounts of disulfide-linked amino acids were detected for HDL and alpha1-acid glycoprotein but not for transferrin. Exchange of disulfide-linked amino acids on exposure to excess Hcy or glutathione was much faster for albumin than for alpha1-acid glycoprotein. CONCLUSIONS: Approximately 10%-30%, of protein-bound Hcy, Cys, and CysGly are disulfide-linked to globulins. Amino acids disulfide-linked to albumin are rapidly exchangeable, while exchange of disulfide-linked amino acids from globulins, such as alpha1-acid glycoprotein, is much slower. Consequently, the pools of Hcy, Cys, and CysGly bound to albumin and globulin may represent kinetically and functionally distinct pools. Plasma concentrations of total Hcy and Cys, which are dominated by albumin-bound pools, may not reflect the abundance of functionally significant modifications of globulins.     

SKF 525A displaces drugs from serum alpha 1-acid glycoprotein binding sites

In vivo treatment of rats with beta-diethylaminoethyl-diphenylpropylacetate hydrochloride (SKF 525A), an inhibitor of hepatic monooxygenases, decreases the serum binding of oxprenolol and propranolol, which bind mainly to alpha 1-acid glycoprotein, but not that of phenytoin, which is bound to albumin. The effect was maximal 40 min after treatment and had disappeared after 6 hr, when enzyme inhibition was still present. A displacing effect was also observed when SKF 525A was added in vitro to serum of rats, dogs and humans and to human alpha 1-acid glycoprotein, whereas binding to human serum albumin was not affected. SKF 525A was found to be equipotent with tris(2-butoxyethyl)phosphate, a known displacer of binding of drugs from alpha 1-acid glycoprotein. When studying the pharmacokinetics or the effects of drugs after SKF 525A pretreatment, the possibility that altered protein binding may influence the findings should be considered.     

Alterations in serum protein binding and pharmacokinetics of l-propranolol in the rat elicited by the presence of an indwelling venous catheter

Preliminary investigations in the rat revealed some unexpected alterations in the pharmacokinetics of l-propranolol when single dose studies were performed on two separate occasions 7 days apart. The apparent volume of distribution of l-propranolol was found to be consistently lower on the second study day. Also, the systemic availability of orally administered propranolol (6 mg/kg) increased from 4.1 to 7.8% from the 1st to the 2nd study day, indicating a decrease in the presystemic clearance of the drug. These changes were associated with a 50 to 60% decrease in the serum unbound fraction of l-propranolol between the 2 study days. Further investigations revealed that the time-dependent increase in the serum protein binding of l-propranolol was elicited by the presence of an indwelling venous catheter which was surgically implanted for the purpose of blood sampling. In contrast, a slight decrease in the serum protein binding of an acidic drug, phenytoin, was observed after catheter implantation. These changes in serum drug protein binding were readily reversed when the indwelling catheter was removed. The total serum protein concentration was not affected by catheter implantation. Fractionation of serum proteins by electrophoresis revealed an increase in alpha 1- and gamma-globulin fractions, a slight decrease in serum albumin and no change in beta- and alpha 2-globulin fractions. Competitive protein binding studies with tris(2-butoxyethyl)phosphate, a specific ligand to alpha 1-acid glycoprotein (AAG), indicate that the increase in serum l-propranolol binding was largely due to an elevation in serum AAG. These observations also explain the opposite effect of catheterization on the serum protein binding of phenytoin, which is bound largely to serum albumin.(ABSTRACT TRUNCATED AT 250 WORDS)     

Precipitation and characterization of an aluminosilicate from AlCl3-Na2SiO3-HCl in serum, of interest for Alzheimer disease.

A precipitation experiment was performed with human serum to model aluminosilicate formation in brains of patients with Alzheimer disease. Aluminum and (or) silicate ions were added to serum in a 1:2 molar ratio at pH 7.4. Precipitates formed immediately and were left for 24 h at 37 degrees C before filtration. Silicate and aluminosilicate formed precipitates with human serum proteins albumin, transferrin, and IgG. In untreated samples, the IgG/albumin ratio increased slightly compared with the ratio in dried serum. Diethylbarbiturate-washed precipitates had a significantly lower protein content than did untreated ones. The IgG/albumin ratio increased considerably in the sample containing aluminosilicate. We conclude that IgG is the sodium dodecyl sulfate-soluble human protein most firmly bound to the aluminosilicate matrix. From 27Al magic-angle-spinning nuclear magnetic resonance (MAS NMR), a pronounced peak was found at 52.79 ppm and a minor peak at 0.53 ppm, suggesting that 4-coordinated aluminum predominates and that 6-coordinated aluminum is present in a smaller proportion. The 29Si MAS NMR spectrum shows a poorly ordered structure. The aluminosilicate formed also contains the cations Na+ > K+ > Ca2+ > Mg2+ and anions Cl- > PO4(3-). Rather than looking for aluminum toxicity to explain the effects of Alzheimer disease, one should consider that by precipitating such a composite phase, the balance of cations, anions, and proteins in human serum is changing.     

Effect of pH and citrate on binding of iron and gallium by transferrin in serum.

Although both Al and Fe are bound to transferrin in plasma, they are metabolized differently. Aluminum is less tightly bound to transferrin than is Fe, and might be released in circumstances in which Fe remains bound. The effect of pH ana citrate on the binding of 67Ga (a radiotracer used as an analog of Al) to transferrin in normal human serum was tested in the presence of physiological concentrations of CO2. At pH less than 6.8, Ga started to dissociate from transferrin; at pH 6, greater than 50% of the added 67Ga was present in a low-M(r) form. In contrast, almost all Fe remained bound to transferrin at pH values as low as 4.7. Citrate at concentrations as great as 100 mmol/L had no effect on binding of Fe, but the binding of 67Ga was markedly reduced at citrate greater than 1 mmol/L. Being bound to transferrin less strongly than Ga is, Al could dissociate even more readily, and loss of Al from transferrin in the kidney might explain why Al but not Fe is excreted in urine.     

Leucocyte-associated plasma proteins. Association of prealbumin, albumin, orosomucoid, alpha 1-antitrypsin, transferrin and haptoglobin with human lymphocytes, monocytes, granulocytes and a promyelocytic leukaemic cell line (HL-60)

The distribution of prealbumin, albumin, orosomucoid, alpha 1-antitrypsin, haptoglobin and transferrin, including their electrophoretic heterogeneous variants, was studied in isolated lymphocytes, monocytes and granulocytes and in a human promyelocytic cell line (HL-60) by crossed immunoelectrophoresis. Prealbumin, albumin and transferrin were present in lymphocytes, monocytes and granulocytes, whereas the cellular variants of orosomucoid and haptoglobin were present only in granulocytes. alpha 1-Antitrypsin was present in four electrophoretic variants which were differently distributed among the various cell types. Synthesis of alpha 1-antitrypsin by monocytes, granulocytes and HL-60 cells was demonstrated by 14C-leucine incorporation. The six plasma proteins could not be removed from intact cells by incubation with the respective antibodies at 0 degrees C, or iodinated by lactoperoxidase catalysed iodination at 23 degrees C. They were, however, readily solubilized by freeze-lysis of the cells, suggesting an intracellular localization. Compared to their plasma counterparts none of the proteins differed in their hydrophobic properties but the carbohydrate residues of orosomucoid, alpha 1-antitrypsin and haptoglobin were different. The pattern of disappearance of the proteins from the cells during incubation suggested that the localization of albumin and transferrin in relation to the cells differed from that of the other proteins.     

Differential binding of serum glycoproteins to lectins during hepatic regeneration in hepatocellular carcinoma and fulminant hepatic failure.

1. The concentrations of four serum glycoproteins, thyroxine-binding globulin, alpha 2-macroglobulin, alpha 1-antitrypsin and transferrin, as well as their reactivities with concanavalin A and lentil-lectin, were measured in patients with hepatocellular carcinoma or fulminant hepatic failure and in normal subjects. 2. Serum concentrations of thyroxine-binding globulin and alpha 1-antitrypsin were significantly greater in patients with hepatocellular carcinoma than in normal subjects, and the percentage lentil-lectin reactivity of these two proteins was markedly increased. 3. With the exception of transferrin, which did not bind to lentil-lectin, an enhancement of lentil-lectin reactivity was observed for the glycoproteins in serum from patients with fulminant hepatic failure. No difference in concanavalin A binding was found between the groups for any of the glycoproteins. 4. Altered fucosylation, as indicated by increased lentil-lectin binding, occurs in several glycoproteins arising in malignant and non-malignant conditions associated with abnormal hepatic regeneration.     

Detection of lysozyme and alpha 2-macroglobulin--lysozyme complexes by immunoblotting

An immunoblotting technique was developed to detect human lysozyme and lysozyme complexes in body fluids. The unoccupied binding capacity of proteins was demonstrated by addition of surplus lysozyme. The sensitivity of immunoblotting to the free enzyme in human albumin solution was less than 5 ng. In serum and pleural fluid, part of exogenous lysozyme was bound to alpha 2-macroglobulin (alpha 2-M). At high concentrations of lysozyme in leukemic sera, part of the enzyme formed an endogenous alpha 2-M complex. On the other hand, the formation of alpha 2-M complexes with exogenous lysozyme was especially striking in sera from nephrotic patients with elevated alpha 2-M. The findings corroborate with previous reports on lysozyme binding to purified alpha 2-M in vitro and suggest that the binding is concentration-dependent with respect to both reaction partners. In vivo the mechanism may provide a pathway for extrarenal lysozyme catabolism medicated by reticuloendothelial cells. No other binding proteins were seen in the present study: lysozyme did not bind to serum immunoglobulins in 35 samples with an immunoglobulin paraprotein, three samples with polyclonally elevated gamma-globulins, 20 other patient sera and 10 normal sera. Neither did lysozyme bind to urinary proteins in five samples from patients with myeloic leukemias nor in 10 samples from myeloma patients with urinary excretion of a monoclonal immunoglobulin light chain.     

The application of gel filtration,immunonephelometry and electrothermal atomic absorption spectrometry to the study of the distribution of copper-, iron- and zinc-bound constituents in human amniotic fluid

Gel filtration with Sephacryl S-300 was used to fractionate copper-, iron- and zinc-bound proteins and low molecular mass constituents in samples of amniotic fluid collected at delivery. The fractions were analysed for copper, iron and zinc by electrothermal (carbon furnace) atomic absorption spectrometry. The proteins associated with these metals were determined by immunonephelometry. The results showed that caeruloplasmin, transferrin and albumin are the major proteins that bind copper, iron and zinc, respectively. Less than 10% of the metals, except iron, recovered from the column was bound to low molecular mass constituents. Comparisons between the distribution of these elements in amniotic fluid and serum are shown and discussed.     

Quantitative immunological determination of 12 plasma proteins excreted in human urine collected before and after exercise

Urine was collected from 6 healthy male adults at rest and from 20 male adults after a marathon race (25 miles). The concentrated urines were quantitatively analyzed, by single radial immunodiffusion, for their content in 12 different plasma proteins: tryptophan-rich prealbumin, albumin, alpha(1)-acid glycoprotein, alpha(1)-antitrypsin, ceruloplasmin, haptoglobin, Gc-globulin, transferrin, hemopexin, beta(2)-glycoprotein I, gammaA-globulin, and gammaG-globulin.Albumin, gammaA-globulin, and gammaG-globulin represent the major part of the plasma proteins detected in normal urine excreted by humans at rest (12, 0.5, and 2.5 mg respectively, out of a total excretion of 17.5 mg of plasma proteins per 24 hr). The other plasma proteins were excreted at a lower rate (< 0.4 mg/24 hr). The relative content of tryptophan-rich prealbumin, alpha(1)-antitrypsin, Gc-globulin, transferrin, and gammaG-globulin was lower in normal urine than in normal serum, whereas that of alpha(1)-acid glycoprotein, beta(2)-glycoprotein I, and gammaA-globulin was higher. The ratio of gammaG-globulin to gammaA-globulin was 4.9:1. When plotted on a logarithmic scale, no direct relationship between the molecular weight of a protein and the value of its renal clearance could be observed.Strenuous exercise increased (up to 50-fold) the excretion of plasma proteins which represent 82% of the total proteins found in urine, instead of 57% in urine collected from humans at rest. There was particularly a significant rise of tryptophan-rich albumin, albumin, alpha(1)-acid glycoprotein, transferrin, gammaA-globulin, and gammaG-globulin (0.26, 127, 11.8, 3.3, 1.2, and 2.0 mug respectively, out of a total excretion of 167 mug of plasma proteins per min). The ratio of gammaG-globulin to gammaA-globulin was 16:1. After exercise, the renal clearance of proteins increased from 2 to 40 times, but, as for the urine of normal subjects at rest, no direct relationship between molecular weight and renal clearance could be observed.     

Serum iron, transferrin saturation, ferritin, and dietary data in age-related macular degeneration.

Iron (Fe) is a tightly metabolically controlled mineral and growth factor for all living cells. Iron not bound in erythrocyte hemoglobin is transported by the plasma iron transport protein transferrin (Tf) and bound within cells by ferritin. Apo-Tf and apo-hemopexin are also known to be made locally in the retina. Free Fe is cytotoxic, promotes oxidation/lipid peroxidation, has been implicated as a risk factor in cardiac disease, and is itself associated with age-related macular degeneration (ARMD), the leading cause of blindness in aging western societies. The authors evaluated Fe overload serum markers and dietary intake in patients with atrophic ARMD. After obtaining informed consent, an Fe panel consisting of serum Fe, total Fe binding capacity (TIBC), and ferritin was performed on 75 veterans (70 men, five women) with an average age of 75 years with a diagnosis of atrophic ARMD by combined criteria of International Retinal Classification and psychophysical/symptom abnormalities. Tf saturation was calculated by dividing serum Fe concentration by TIBC. Dietary iron with and without supplementation and vitamin C intake were determined for 86 patients using the Harvard School of Public Health/Department of Nutrition Food Frequency Questionnaire. Statistically significant correlations (P <0.1) were found between serum and dietary Fe (r = -.26), between serum Fe and serum ferritin (r =.34), and between dietary Fe and dietary vitamin C (r =.30). The data on mostly male geriatric veterans with atrophic ARMD indicate that single time-point assessment of systemic Fe status and dietary Fe is not useful. However, serial multiple-year data, correlating Fe markers with disease, may still be important. Also, because Fe transport proteins do not cross the blood-retina barrier, the local cellular toxic effects of Fe must also be considered.     

Determination of the concentration of immunoglobulins and other proteins in bronchoalveolar fluid of patients with various types of pulmonary pathology

The levels of immunoglobulins and other proteins (alpha 2-MG, alpha 1-AT, C3, albumin, transferrin and lactoferrin) were studied in the BAL of 60 patients with different types of pulmonary tuberculosis, 4 patients with sarcoidosis and 7 CNPD patients. The level of most proteins in BAL of the examinees was higher than that reported for healthy subjects. The highest protein levels were noted in CNPD and sarcoidosis patients. The diagnostic importance of the level of alpha 2-MG was established for sarcoidosis. 27 paired BAL-serum specimens from the same patients with pulmonary tuberculosis were investigated for analysis of the mechanisms of protein appearance in BAL. The protein/albumin ratio for most proteins was higher in BAL than in the respective serum. A relatively high level of proteins in the patients' BAL was probably determined by the activation of their local synthesis.     

Blood distribution of levocetirizine,a new non-sedating histamine H1-receptor antagonist,in humans

The aim of the present study was to determine (1) the extent of levocetirizine binding to human blood cells, plasma and individual plasma proteins; (2) the parameters for levocetirizine binding to individual plasma proteins both at their physiological concentrations and, for human serum albumin (HSA), at a lower saturating concentration; and (3) to simulate levocetirizine distribution in human blood using the information obtained at physiological haematocrit (H) for blood cells and at physiological concentrations for individual plasma proteins. The nature of the main binding sites of HSA, i.e. site I (warfarin) and site II (diazepam), preferentially involved in levocetirizine binding was also investigated. Over the range of therapeutic concentrations and multiples thereof, levocetirizine is extensively bound to blood components, the free fraction remaining constant (6.45%) and the fraction bound to blood cells and to plasma proteins accounting for 27.43 and 66.11%, respectively. The binding of levocetirizine to HSA in the presence of physiological concentrations of non-esterified fatty acids (NEFAs) is the main interaction of levocetirizine in blood (50.68% of overall blood binding). This interaction is fatty acid sensitive, with decreasing concentrations of NEFA increasing the amount of bound drug and vice versa. Levocetirizine is also bound to alpha1-acid-glycoprotein and high-density lipoproteins (5.17 and 6.89% of overall blood binding, respectively). The displacement of levocetirizine by diazepam is consistent with the binding of this drug to HSA at site II, as diazepam is a specific marker for this site. The binding of levocetirizine to HSA at site II being characterized by a low association constant, other drugs sharing the same site with high association constants cannot displace levocetirizine except at very high plasma concentrations. In any case, at therapeutic concentrations of levocetirizine and at physiological protein concentrations, the observation that none of the levocetirizine binding proteins is saturated suggests that very little or no variation of the free fraction will occur although a different distribution of its bound forms is possible.     

Identification and quantification of serum proteins secreted into the normal human jejunum

The in vivo transfer of serum proteins to the human intestinal lumen was characterized by crossed immunoelectrophoretic analyses of intestinal perfusates from four healthy volunteers. Serum proteins with molecular masses below 100 kDa and the immunoglobulins were found in human jejunal perfusates. Larger serum proteins were either absent (alpha and beta lipoproteins) or present in small amounts (alpha 2-macroglobulin, haptoglobulin and ceruloplasmin). These results demonstrate the existence of a selective transfer of serum proteins to the intestinal lumen under physiological conditions. The intestinal clearance rate was 0.1 ml serum per hour per 10 cm jejunum for albumin, prealbumin, alpha 1-antitrypsin, orosomucoid, transferrin and haemopexin. The rate of secretion of total protein to the jejunal lumen was 100 mg protein per hour per 10 cm jejunum. About 45% was due to immunoglobulins and further 10-15% due to the remaining serum proteins. It is suggested that the serum proteins pass through the epithelium by a transcellular mechanism.     

Pathogenesis of alcohol-induced accumulation of protein in the liver.

Alcohol feeding to rats produced hepatomegaly, associated with enlargement of the hepatocytes. The increase in liver dry weight was accounted for not only by fat but also by protein accumulation, primarily in microsomes and cytosol, with a selective increase in export proteins: concentrations of both immunoreactive albumin and transferrin were augmented in liver microsomes and cytosol of ethanol-fed rats. To investigate the mechanism of this protein accumulation, [14C]leucine was injected intravenously and its incorporation into both liver and serum proteins was measured after various time intervals. Rates of synthesis and export were assessed from protein labeling and specific activities of leucyl-tRNA. Synthesis of liver protein and proalbumin were enhanced by chronic ethanol feeding, but this was not associated with a corresponding rise in serum albumin output. Actually, there was a significant retention of the label in liver albumin and transferrin with delayed appearance in the serum of ethanol-fed rats. This indicated that, regardless of the changes in synthesis, the export of protein from the liver into the plasma was impaired. This alteration in export was associated with a decreased amount of polymerized tubulin in the liver of ethanol-treated animals. Thus, both enhanced protein synthesis and defective export contribute to the ethanol-induced accumulation of liver protein, and the decrease in liver microtubules represents a possible site for impairment of protein export.     

Characterization of zinc-binding proteins of plasma in familial hyperzincemia

Plasmas from three brothers (aged 35 to 45) with chronic hyperzincemia (325 to 525 micrograms/dl Zn) were analyzed and compared with pooled control plasma (104 microgram/dl Zn). The levels of copper, iron, total protein, albumin, and amino acids were similar in normal and hyperzincemic plasmas. Distribution of zinc among plasma components was determined chromatographically. Zinc eluted quantitatively in two distinct peaks from Sephadex G-100 gel filtration resin. The amount of the metal in fractions containing species larger than 100,000 molecular weight (peak I) was similar (35 to 45 micrograms/dl Zn) in normal and hyperzincemic plasmas. The additional complement of zinc in hyperzincemic plasma was localized within fractions containing zinc-binding proteins such as albumin, transferrin and HRG. That zinc was not associated with transferrin was determined by Affi-Gel affinity chromatography. The amounts of HRG in hyperzincemic plasmas were similar to or below those in control plasma. Zinc and albumin were selectively retained by anti-human albumin IgG coupled to protein A-Sepharose. In contrast, anti-human HRG IgG coupled to CM Bio-Gel A failed to bind plasma zinc. The findings indicate that (1) most available, protein-associated zinc in normal plasma is bound to albumin and (2) the additional complement of zinc in familial hyperzincemic plasma is associated with albumin. The biochemical basis for the enhanced binding of zinc by albumin in hyperzincemic plasma is unknown.     

Biochemical modifications in a case of analbuminemia (author's transl)]

A new case of analbuminemia is described for a six month old child of Algerian origin. The serum albumin concentration was 64 mg/l and its immunochemical action was identical to that of normal albumin. The system reacted by an increase of the synthesis of globulins. For the subject, the alpha1-antitrypsin, ceruloplasmin, haptoglobin, alpha2-macroglobulin, transferrin and immunoglobulins M contents were three times higher than the standard figures. However, it was possible to show that the presence of free bilirubin independent from proteins could be detected at a concentration of 17 mumol/l.     

An immunological study of nine proteins in CSF and serum of a group of epileptic patients.

The concentrations of nine proteins, alpha-1-acid glycoprotein, antitrypsin, prealbumin, transferrin, albumin, IgG, ceruloplasmin, IgA and alpha-2-macroglobulin, have been determined in the serum and CSF of two groups of patients, one control and one experimental, by an immunological method. The experimental group were patients suffering from grand mal epilepsy. The control group showed no detectable neurological disorder. In the group of grand mal epileptics, only prealbumin showed a significant elevation in CSF when compared with the control group. In contrast, the rest of the proteins are decreased with respect to the controls except for alpha 1-acid glycoprotein and transferrin. The results from this study also suggest that something more than an ultrafiltration process dependent upon molecular weight, is important in determining the concentration of some serum proteins in the CSF.     

EFFECT OF TURPENTINE-INDUCED INFLAMMATION ON THE DISPOSITION KINETICS OF PROPRANOLOL, METOPROLOL, AND ANTIPYRINE IN THE RAT

Plasma concentrations after oral administration of the high extraction drug propranolol are increased in patients and animals with inflammation. This could be due to increased serum propranolol binding, but also to decreased first-pass metabolism. We studied the pharmacokinetics of 3 drugs in control rats and in rats with turpentine-induced inflammation: propranolol, which is bound extensively to alpha 1-acid glycoprotein (alpha 1-AGP); metoprolol, another high extraction drug, but which is negligibly bound to alpha 1-AGP; and antipyrine, a low extraction drug, not bound to serum proteins. After IV administration of propranolol in rats with inflammation, systemic clearance, volume of distribution, and free fraction decreased, and the area under the curve (AUC) increased, whereas the half-life did not change. As the systemic clearance of a high extraction drug such as propranolol depends on hepatic blood flow only, a fall in hepatic blood flow or transition to a low extraction situation should be postulated. After oral administration of propranolol, the AUC was increased 20-fold in rats with inflammation; as the decrease in free fraction was only 4-fold, it can be concluded that a considerable decrease in hepatic intrinsic clearance was present. For metoprolol, in contrast to propranolol, after IV administration, no changes in pharmacokinetic parameters as a result of inflammation were observed. After oral administration, the AUC was increased about 4 times in rats with inflammation; as metoprolol is only negligibly bound to serum proteins, the increase in AUC can be attributed to a decrease in hepatic intrinsic clearance.(ABSTRACT TRUNCATED AT 250 WORDS)     

Protein clearances and selectivity determinations in childhood nephrosis: a reappraisal

To critically evaluate the clinical utility of determining specific proteins in patients with extensive proteinuria, we used immunonephelometric methods to measure albumin, transferrin, IgG, and alpha 2-macroglobulin in serum and in 24-h urine specimens from 37 children with idiopathic nephrotic syndrome. Renal biopsy demonstrated minimal change disease (I) in 15, focal glomerulosclerosis (II) in 15, and membranoproliferative glomerulonephritis (III) in seven patients. A three-group nonparametric rank test and three-group discriminant function analysis of the protein excretion and clearances of the four proteins we measured revealed significant differences in the excretion of IgG and the clearance of alpha 2-macroglobulin among the three groups of patients (p less than 0.05). Only patients with III had low serum complement C3 concentrations. Patients with I or II were best discriminated by differences in the excretion of transferrin and IgG, the clearance of alpha 2-macroglobulin, and the selectivity index (the clearance ratio of IgG/transferrin). These data indicate that measurement of specific urinary proteins and selectivity determinations may be helpful in predicting the type of histopathology and the prognosis of nephrotic children who have normal complement concentrations.