Immunoadsorption of human plasma with protein A-sepharose columns.

Four patients with refractory acute leukemia were treated a total of 31 times with an immunoadsorption system consisting of protein A-sepharose columns, a cell separator, and an elution monitor to test its safety and capacity to remove immunoglobulins. The procedure was tolerated well, and acutely reduced plasma IgG levels by approximately 18 percent. When the procedure was repeated two to three times per week for 3 weeks, IgG levels dropped by 30 to 40 percent, but they gradually returned to pretreatment levels after completion of the course of treatment. Single columns became saturated with IgG after approximately 1500 ml of plasma had passed through the columns. The use of multiple columns sequentially provided continuous extraction of immunoglobulin. One patient regained responsiveness to platelet transfusions after removal of platelet antibodies. These preliminary studies suggest that this immunoadsorption system is effective for specifically removing IgG and that it merits further clinical testing.

Further characterization of the in vitro tumoricidal activity of staphylococcal protein A.

Treatment of plasma or serum from leukemic patients with solid phase staphylococcal Protein A induced leukemic blast cell lysis in vitro, but this effect was relatively independent of the amount of immunoglobulin G (IgG) removed. Samples with approximately equal cytotoxic activity contained markedly different IgG levels, while samples with similar IgG levels had a wide range of tumoricidal activity. Assays of plasma samples collected during a perfusion of one plasma volume through a Protein A-Sepharose column indicated that the duration of the procedure had a greater effect on cytotoxic activity than did the amount of IgG removed. Neither added leukemic nor normal IgG significantly improved blast cell viability in treated serum. Cytotoxic activity was not dialyzable and concentrated in the Mr less than 100,000 fraction of samples separated by filtration. Treated cytotoxic serum samples did not have important Clq binding activity. These results suggest that the in vitro tumoricidal activity of solid phase Protein A is probably due to a toxic substance added to serum during immunoadsorption rather than to its immunoadsorptive capacity.

Systemic complications of inflammatory bowel disease.

Radiologic assessment of the sacroiliac joints should be part of every inflammatory bowel disease patient's workup; ankylosing spondylitis is 10 to 20 times more common in ulcerative colitis patients than in normal persons. Iritis, which occurs in 10 to 20% of ulcerative colitis patients, often precedes bowel symptoms. It may be necessary to use long-term, low-dose steroid therapy to control frequently recurring iritis.

Acute oligoarthritis associated with Clostridium difficile pseudomembranous colitis.

The abrupt onset of a sterile inflammatory oligoarthritis developed in a patient with active Clostridium difficile pseudomembranous colitis. The arthritis affected a hip and a knee. Leukocyte counts of synovial fluid obtained from the patient's left hip and knee were elevated. He was haplotyped as HLA-B27 antigen-positive. The colitis and arthritis promptly abated after treatment with oral vancomycin hydrochloride. Three other cases of arthritis associated with antibiotic-induced colitis were reviewed. It seems as if treatment of the colitis leads to resolution of the arthritis.

Specific removal of antibodies with an immunoadsorption system.

A continuous flow immunoadsorption system consisting of a cell separator, protein A-sepharose columns, and a semi-automatic elution component was developed to specifically remove circulating IgG. This system provides extensive absorption with an essentially unlimited column bed volume. Six dogs were treated a total of 19 times. In no case did fever, sepsis, or respiratory distress result from the treatment. Serial blood counts and tests of liver and renal function remained in the normal range. Ex vivo perfusion of one plasma volume caused an acute drop in IgG levels of approximately 50 percent. This was reflected in a similar decrease in specific antibody levels to sheep erythrocytes, bovine serum albumin, and canine parvovirus. Antibody kinetics following immunoadsorption were variable, but in several cases, antibody levels remained lowered. This immunoadsorptive system appears to be a safe and effective alternative to plasma exchange for removal of IgG antibodies.

The nature of IgG complexes in alcoholic liver disease.

Circulating immune complexes have been described in most liver diseases, including alcoholic liver disease, although their pathogenic significance remains unclear. Currently available immune complex assays do not distinguish immunoglobulin aggregates from antigen-antibody complexes. Immunoglobulin aggregate formation occurs in vitro at 37 degrees C in the presence of hypergammaglobulinemia and/or hypoalbuminemia, conditions common in liver disease. To determine if hypergammaglobulinemia and/or hypoalbuminemia could predispose to immunoglobulin aggregate formation in vivo, 25 patients with alcoholic liver disease were studied. Using sucrose density gradient fractionation followed by quantitation of IgG by radioimmunoassay, high molecular weight IgG complexes (greater than 11S) were frequently present in alcoholic liver disease sera, and correlated with the degree of hypergammaglobulinemia and/or hypoalbuminemia, and with 125I-C1q binding activity. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of these complexes revealed only IgG and nonspecific trapping of several normal serum proteins. A specific complex-associated antigen could not be identified. While small, undetectable quantities of true antigen-antibody complexes might also be present, our data suggest that IgG complexes in alcoholic liver disease may represent immunoglobulin aggregate formation in vivo.

The effect of serum immunoglobulin concentration on immune complex detection by polyethylene glycol.

The addition of 4% polyethylene glycol (PEG) to serum and quantitation of immunoglobulins in the dissolved precipitate has been advocated as a simple, reliable method for detecting circulating immune complexes. Because pathological sera, which often yield positive results in this assay, may contain increased concentrations of immunoglobulins compared to normal control sera, we have determined the relationship between total serum immunoglobulin concentration and the quantity of immunoglobulins precipitated by 4% PEG. When IgG was added to normal serum, the quantity and percentage of IgG in the precipitate was directly proportional to total serum IgG concentration. This concentration-dependent precipitation appeared to be unrelated to the presence of aggregates in the IgG preparation, the serum concentration of albumin, or interactions with serum complement. With normal serum, concentrated to yield a wide range of endogenous immunoglobulin concentrations, the percentage of IgG, IgM and IgA in the PEG precipitates was likewise directly proportional to the total serum concentration of these immunoglobulins. In view of these findings, this method is likely to give false-positive results in pathological sera containing increased immunoglobulin concentrations and is probably invalid as a means for detecting circulating immune complexes. However, with a final concentration of 2% PEG, successful discrimination between aggregated immunoglobulin and monomeric IgG may be achieved.

125I-C1q binding activity of soluble antigen-antibody complexes formed in vitro.

The ability of the 125I-C1q binding test to detect soluble antigen-antibody complexes formed in vitro was tested with two antigen-antibody systems. Using tetanus toxoid:anti-toxoid complexes, all of the increased 125I-C1q binding activity was due to soluble immunoglobulin aggregates present in the unheated tetanus immune globulin. Using BSA:anti-BSA complexes, maximum 125I-C1q binding activity was present in the soluble supernatants from the equivalence zone. No immune complexes were present in these supernatants and the increased activity appeared to be due to alterations in endogenous C1 during the formation of large, insoluble immune complexes. The 125I-C1q binding test readily detected large, insoluble BSA:anti-BSA complexes but may not detect small, soluble complexes which have been suggested to be important in disease pathogenesis.

Spontaneous aggregation of native immunoglobulins in hypoalbuminemic serum.

We have previously demonstrated that isolated IgG, when added to buffer or serum, undergoes aggregation during incubation at 37 degrees C. The addition of albumin inhibits aggregation. The present studies were designed to determine whether endogenous immunoglobulins in normal serum, present in low-normal concentration, would likewise aggregate in the presence of hypoalbuminemia. Albumin was removed from normal serum by absorption with Affi-gel Blue. Following incubation at 37 degrees C, high molecular weight IgG, IgM and IgA were detectable by density gradient ultracentrifugation. Reconstituting the samples with albumin prior to incubation prevented the formation of immunoglobulin aggregates. These data suggest that immunoglobulin aggregation might also occur in vivo in a variety of pathological conditions associated with hypoalbuminemia and/or hypergammaglobulinemia and might be incorrectly assumed to represent antigen-antibody complexes.

Cholinergic stimulation of immunoglobulin A secretion in rat intestine.

This study was undertaken to assess a possible role for cholinergic agents in the regulation of intestinal immunoglobulin A secretion. Intestinal loops, constructed in anesthetized rats, were perfused with phosphate buffered normal saline. Immunoglobulin A concentrations were measured by radioimmunoassay. When compared with the effect of normal saline in the same rats, intravenous injection of pilocarpine, 10 mg/kg, increased immunoglobulin A concentrations in perfusates from ileal loops (p less than 0.001). Qualitatively similar results were obtained with muscarine and bethanechol, from jejunal and colonic loops, and from unanesthetized rats. Immunoglobulin A concentrations increased four- to eightfold during maximal cholinergic stimulation. Atropine, 250 micrograms intravenously, completely blocked the effect of pilocarpine on immunoglobulin A secretion (p less than 0.005), and also inhibited basal intestinal immunoglobulin A secretion for 40 min after injection. As determined on 10%-40% sucrose density gradients, much of the immunoglobulin A secreted after cholinergic stimulation sedimented in the 11S range. These data indicate that intestinal secretion of immunoglobulin A is stimulated by the muscarinic effect of cholinergic agonists, and suggest that basal secretion of immunoglobulin A may be influenced by the parasympathetic nervous system.

Dissociation of IgG aggregates at low pH.

Previous studies have reported that antigen-antibody complexes dissociate at low pH while IgG aggregates do not, and it has been suggested that dissociation of high molecular weight IgG in serum at low pH supports the presence of circulating antigen-antibody complexes. Our studies confirm the observation that heat-aggregated IgG does not dissociate at low pH. By contrast, non-thermal IgG aggregates, formed in isolated IgG preparations or in hypoalbuminemic serum and which may also form in vivo, are almost totally dissociated at low pH. Consequently dissociation of high molecular weight IgG in serum at low pH does not distinguish antigen-antibody complexes from IgG aggregates.

Studies on the nature of intermolecular bonding in antigen-antibody complexes.

The formation and dissociation of BSA:anti-BSA complexes was studied under two dissociating conditions: (i) low pH, which disrupts Coulombic bonds, and (ii) low surface tension, which disrupts van der Waals bonds. Soluble complexes formed in marked (thirty-two times) antigen excess were almost totally dissociated at low pH. By contrast, complexes formed near equivalence or in the zone of antibody excess, while becoming smaller, were not dissociated at low pH. Lowering surface tension at neutral pH likewise resulted in smaller complexes but not in dissociation of antigen from antibody. Dissociation of complexes formed near equivalence or in antibody excess was achieved by simultaneously lowering pH and surface tension. These data suggest that complexes formed in marked antigen excess involve predominantly Coulombic bonding while complexes formed at other antigen:antibody ratios involve both Coulombic and van der Waals bonding.

Studies on the nature of heat-labile anti-complementary activity in normal human serum.

Heat-labile anti-complementary activity (ACA) appears in normal human serum during storage or heating as endogenous haemolytic activity disappears. Following gel filtration of unheated serum, two peaks of heat-labile ACA are present. The ACA of both whole and fractionated serum has previously been attributed to the presence of heat-labile immunoglobulin aggregates or immune complexes. Our data demonstrate that the heavy peak of ACA obtained by gel filtration does not bind to 125I-C1q or to Raji cells, and that its effect is abolished to C1INH, suggesting that it represents C1 rather than immunoglobulin aggregates or immune complexes. The lighter peak of ACA in fractionated serum has the functional characteristics of C1s and free C1s is demonstrable in fractions containing this activity. The ACA of whole serum likewise has functional characteristics of C1. The anti-complementary effect of C1 on guinea-pig complement would not be evident in the complement fixation assay until most endogenous haemolytic activity in human serum has been inactivated, either by heat or by storage. C1INH only partially inhibits this ACA in serum or in solutions containing isolated C1 in high concentrations. These observations indicate that heat-labile ACA in whole or fractionated sera is due to the presence of C1 and C1s and that this activity cannot be taken as evidence for the presence of immune complexes.

Evidence against the presence of circulating immune complexes in chronic inflammatory bowel disease.

Several investigators have reported the presence of circulating immune complexes in serum from patients with Crohn's disease and chronic ulcerative colitis. Because previous assays employed conditions which might have caused immunoglobulin aggregates to form in vitro, thus falsely suggesting the presence of immune complexes in vivo, we tested inflammatory bowel disease sera for immune complexes using four assays designed to minimize in vitro immunoglobulin aggregation. In three assays immune complexes were not detectable, while in a fourth, the Clq precipitin test, positive reactions occurred. These precipitin reactions did not have characteristics of immune complexes. Our data suggest that circulating immune complexes are either not present in patients with inflammatory bowel disease or that they occur infrequently or in low concentration.

Antibodies to viral gastroenteritis viruses in Crohn's disease.

Antibody prevalence and titer to rotavirus and Norwalk virus were studied in Crohn's disease patients and in age-, sex-, and time-matched controls. There were no significant antibody differences between the groups studied.