Platelet activating factor induces transient blood-brain barrier opening to facilitate edaravone penetration into the brain.

The blood-brain barrier (BBB) greatly limits the efficacy of many neuroprotective drugs' delivery to the brain, so improving drug penetration through the BBB has been an important focus of research. Here we report that platelet activating factor (PAF) transiently opened BBB and facilitated neuroprotectant edaravone penetration into the brain. Intravenous infusion with PAF induced a transient BBB opening in rats, reflected by increased Evans blue leakage and mild edema formation, which ceased within 6 h. Furthermore, rat regional cerebral blood flow (rCBF) declined acutely during PAF infusion, but recovered slowly. More importantly, this transient BBB opening significantly increased the penetration of edaravone into the brain, evidenced by increased edaravone concentrations in tissue interstitial fluid collected by microdialysis and analyzed by Ultra-performance liquid chromatograph combined with a hybrid quadrupole time-of-flight mass spectrometer (UPLC-MS/MS). Similarly, incubation of rat brain microvessel endothelial cells monolayer with 1 µM PAF for 1 h significantly increased monolayer permeability to (125)I-albumin, which recovered 1 h after PAF elimination. However, PAF incubation with rat brain microvessel endothelial cells for 1 h did not cause detectable cytotoxicity, and did not regulate intercellular adhesion molecule-1, matrix-metalloproteinase-9 and P-glycoprotein expression. In conclusion, PAF could induce transient and reversible BBB opening through abrupt rCBF decline, which significantly improved edaravone penetration into the brain. Platelet activating factor (PAF) transiently induces BBB dysfunction and increases BBB permeability, which may be due to vessel contraction and a temporary decline of regional cerebral blood flow (rCBF) triggered by PAF. More importantly, the PAF induced transient BBB opening facilitates neuroprotectant edaravone penetration into brain. The results of this study may provide a new approach to improve drug delivery into the brain.

99mTc-albumin can replace 125I-albumin to determine plasma volume repeatedly.

OBJECTIVE: Plasma volume assessment may be of importance in several disorders. The purpose of the present study was to compare the reliability of plasma volume measurements by technetium-labeled human serum albumin ((99m)Tc-HSA) with a simultaneously performed plasma volume determination with iodine-labeled human serum albumin ((125)I-HSA). MATERIALS AND METHODS: In 15 healthy volunteers, simultaneous plasma volume measurements with (99m)Tc-HSA and (125)I-HSA were performed after ½ hour in the supine position. Blood samples were obtained 10, 15, 20, and 30 minutes after the injection for accurate retropolation from the plasma counts to time zero to correct for leakage of the isotopes from the circulation. RESULTS: The mean difference (bias) between plasma volume measured with (125)I-albumin and (99m)Tc-albumin was 8 ml (0.1 ml/kg) with limits of agreement (bias ±1.96 SD) ranging from -181-196 ml (-2.3-2.5 ml/kg). The tracer disappearance rate was significantly higher with (99m)Tc-albumin (-23.1±7.1%/h) than with (125)I-albumin (-6.7±3.6%/h) (p <0.001). CONCLUSION: This study demonstrates that (99m)Tc-HSA can replace (125)I-HSA for single measurements of plasma volume in healthy volunteers. It needs to be emphasized however, that repeated blood sampling for 1/2 hour after injection of the tracer is required to correct for the disappearance of (99m)Tc and (99m)Tc-HSA from the circulation.

The concordance of MRI and quantitative autoradiography estimates of the transvascular transfer rate constant of albumin in a rat brain tumor model.

The apparent forward transfer constant, K transa, for albumin was measured in 9L cerebral tumors in 15 rats. An MRI study using gadolinium-labeled bovine serum albumin was followed by terminal quantitative autoradiography (QAR) using radioiodinated serum albumin. Look-Locker MRI estimates of T(1) followed gadolinium-labeled bovine serum albumin blood and tissue concentration. QAR and MRI maps of K transa were coregistered, a region of interest (ROI) that included the tumor and its surround was selected, and the two estimates of K transa from the ROI on QAR and MRI maps were compared by either mean per animal ROI or on pixel-by-pixel data using a generalized estimating equation. An ROI analysis showed a moderate correlation between the two measures (r = 0.57, P = 0.026); pixel-by-pixel generalized estimating equation analysis concurred (r = 0.54, P < 0.0001). The estimates of QAR with MRI of last time points (e.g., 25 min) showed a moderate correlation (ROI r = 0.55, P < 0.035; generalized estimating equation r = 0.58, P < 0.0001). Differences between the QAR and MRI estimates of K transa did not differ from zero, but the MRI 25-min estimate was significantly lower than the QAR estimate. Thus, noninvasive MRI estimates of vascular permeability can serve as a surrogate for QAR measures.

A model to measure lymphatic drainage from the eye.

Intraocular pressure (IOP) is the most important risk factor for glaucoma development and progression. Most anti-glaucoma treatments aim to lower IOP by enhancing aqueous humor drainage from the eye. Aqueous humor drainage occurs via well-characterized trabecular meshwork (TM) and uveoscleral (UVS) pathways, and recently described ciliary body lymphatics. The relative contribution of the lymphatic pathway to aqueous drainage is not known. We developed a sheep model to quantitatively assess lymphatic drainage along with TM and UVS outflows. This study describes that model and presents our initial findings. Following intracameral injection of (125)I-bovine serum albumin (BSA), lymph was continuously collected via cannulated cervical lymphatic vessels and the thoracic lymphatic duct over either a 3-h or 5-h time period. In the same animals, blood samples were collected from the right jugular vein every 15 min. Lymphatic and TM drainage were quantitatively assessed by measuring (125)I-BSA in lymph and plasma, respectively. Radioactive tracer levels were also measured in UVS and "other" ocular tissue, as well as periocular tissue harvested 3 and 5 h post-injection. Tracer recovered from UVS tissue was used to estimate UVS drainage. The amount of (125)I-BSA recovered from different fluid and tissue compartments was expressed as a percentage of total recovered tracer. Three hours after tracer injection, percentage of tracer recovered in lymph and plasma was 1.64% ± 0.89% and 68.86% ± 9.27%, respectively (n = 8). The percentage of tracer in UVS, other ocular and periocular tissues was 19.87% ± 5.59%, 4.30% ± 3.31% and 5.32% ± 2.46%, respectively. At 5 h (n = 2), lymphatic drainage was increased (6.40% and 4.96% vs. 1.64%). On the other hand, the percentage of tracer recovered from UVS and other ocular tissue had decreased, and the percentage from periocular tissue showed no change. Lymphatic drainage increased steadily over the 3 h post-injection period, while TM drainage increased rapidly - reaching a plateau at 30 min. This quantitative sheep model enables assessment of relative contributions of lymphatic drainage, TM and UVS outflows, and may help to better understand the effects of glaucoma agents on outflow pathways.

Cytokine and chemokine responses in serum and brain after single and repeated injections of lipopolysaccharide: multiplex quantification with path analysis.

Administration of the proinflammatory molecule lipopolysaccharide (LPS) alters transport rates for many peptides across the blood-brain barrier (BBB). We and others have previously shown that effects of LPS on BBB transport are highly dependent on the injection paradigm used, and timing of the study. Cytokine expression in both brain and serum compartments influences the BBB response to an inflammatory stimulus, and mediates changes in BBB transport. Here, we used multianalyte technology to simultaneously determine the responses of 13 cytokines and chemokines (G-CSF, GM-CSF, IL-1α, IL-1ß, IL-6, IL-10, IL-13, IP-10, KC, MCP-1, MIP-1α, RANTES, and TNF-α) in brain and blood to single and repeated injections of LPS and path analysis to determine the major relations among these analytes. Major findings are: (1) in comparison to measurements taken from a time course after a single injection of LPS, the three injection regimen of LPS produced significantly higher levels in brain for G-CSF, IL-1α, IL-6, MCP-1, MIP-1α, and TNF and in serum for G-CSF, IL-6, and GM-CSF and (2) path analysis distinguished direct from indirect correlations between analyte pairs, with MCP-1, IL-6, G-CSF, and KC mediating relations among these cytokines both within and between serum and brain compartments. These results suggest that potentiation of cytokine levels in brain and serum compartments could play important roles in the regulation of BBB transport, and that our novel application of an established statistical method can be used to assess direct correlations within multiplexed datasets.

Elevation of CSF albumin in old sheep: relations to CSF turnover and albumin extraction at blood-CSF barrier.

Albumin is the most abundant protein in both CSF and plasma, and albumin quotient is often used to assess the functions of brain barriers especially that of the blood-CSF barrier [i.e. the choroid plexus (CP) which also secretes CSF]. In this study, we took albumin as a model molecule to investigate ageing-related alterations in the CSF-CP system in sheep. We found significant ageing-related increases in the weight of lateral CP [122.4 +/- 14.0 mg in the young, 198.6 +/- 35.4 mg in the middle aged, 286.1 +/- 25.1 mg in the old (p < 0.05)], in the CSF albumin as well as the albumin quotient. Albumin protein spots in old CSF displayed wider on 2D western immunoblotting images, and had higher densities on images of 2D large gels stained with Pro-Q Emerald 488 compared to the young samples, suggesting ageing-related post-translational modification in the albumin. CSF secretion was reduced with age: 0.148 +/- 0.013 mL/min/g in the young, 0.092 +/- 0.02 mL/min/g in the middle aged, 0.070 +/- 0.013 mL/min/g in the old (p < 0.05). The (125)I-BSA extraction was not different among the sheep groups, nor was altered by temperature reduction, monensin, nocodazole, anti-transforming growth factor beta receptor II antibody, as well as unlabelled albumins. In conclusion, elevation of albumin in old CSF is associated with reduced CSF secretion by the CP, which size increases with age. (125)I-BSA extract, reflecting the extracellular space rather than the active albumin uptake in the CP, is not different between ages. These early changes in health ageing may result in the accumulation and modifications of CSF proteins leading to neurotoxicity.

The uptake and digestion of iodinated human serum albumin by macrophages in vitro.

Mouse peritoneal macrophages take up I*-HSA from their medium during in vitro cultivation. Conditions which promote I*-HSA uptake are the same as those which stimulate formation of pinocytic vesicles. Autoradiography of cells pulsed with (125)I-HSA showed that intracellular isotope is localized in perinuclear granules, or secondary lysosomes. Following a pulse of (125)I-HSA, intracellular radioactivity decreases and the amount of TCA-soluble isotope in the medium increases correspondingly. About 50% of the intracellular isotope is lost in 5 hr. The release of isotope from pulsed cells is not inhibited by parafluorophenylalanine, 2,4-dinitrophenol or by a reduction of the serum concentration of the medium. However, the processing of ingested (125)I-HSA is reversibly inhibited by reduced temperature. The TCA-soluble radioactive material excreted by pulsed macrophages was identified as monoiodotyrosine.

Receptors on immunocompetent cells. 3. Specificity and nature of receptors on dinitrophenylated guinea pig albumin- 125 I-binding cells of immunized guinea pigs.

Guinea pigs immunized with 2,4-dinitrophenyl-guinea pig albumin (DNP-GPA) possess lymphocytes which specifically bind sufficient DNP-GPA-(125)I to their surface to be detected by radioautography. These lymphocytes are present in the draining lymph nodes in a frequency of approximately 50/1000 lymphocytes in animals immunized 2-4 wk earlier with DNP-GPA in complete Freund's adjuvant. Nonimmunized animals have approximately 0.4 DNP-GPA antigen-binding cells (ABC) per 1000 lymphocytes. An increase in the frequency of DNP-GPA ABC in peripheral blood is detectable by 5 days after immunization, which is before the time that serum anti-DNP antibody is measurable. The receptors of these ABC are hapten specific in that free epsilon-DNP-L-lysine, at low concentration, inhibits the binding of DNP-GPA-(125)I; DNP bovine serum alumbin (DNP-BSA) is equivalent to DNP-GPA in the inhibition of binding of DNP-GPA-(125)I to ABC; and both DNP-GPA agarose beads and DNP-BSA agarose beads specifically adsorb DNP-GPA-(125)I ABC. Anti-immunoglobulin antisera, particularly anti-gamma(2) sera, inhibit the binding of DNP-GPA-(125)I to these cells implying that the receptors are immunoglobulin, primarily of the gamma(2) heavy chain class. DNP-GPA-(125)I ABC appear to represent precursors of antibody-secreting cells and have specificity characteristics which are very different from cells, of similarly immunized guinea pigs, which mediate a cellular immune response to DNP-GPA.

Receptors on immunocompetent cells. V. Cellular correlates of the "maturation" of the immune response.

During the course of the immune response to dinitrophenylated guinea pig albumin (DNP-GPA), a striking and parallel increase in avidity for epsilon-DNP-L-lysine occurs in the receptors on antigen-binding lymphocytes, antibody secreted by individual plaque-forming cells, and serum antibody molecules. A detailed analysis of the avidity distribution of antibody produced by plaque-forming cells indicates that this "immunologic maturation" is primarily due to a preservation of the high avidity subpopulation and a striking loss in the low avidity population rather than to sequential appearance of these cells. Moreover, the demonstration of the increased avidity of receptors of antigen-binding lymphocytes, which appear to be precursors of antibody-synthesizing cells, strongly suggests that the antigen-driven selectional process operates primarily on this cell type.

Pathogenetic mechanisms in experimental immune fever.

When rabbits sensitized to human serum albumin (HSA) are challenged intravenously with specific antigen, fever develops and two transferable pyrogens can be demonstrated in the circulation. The first appears prior to the development of fever and has properties consistent with soluble antigen-antibody complexes. These have been shown to be pyrogenic when prepared in vitro and to produce a state of febrile tolerance when repeatedly administered. The second pyrogen, demonstrable during fever in donor rabbits, appears to be similar to endogenous pyrogen described in other experimental fevers. It is postulated that the formation of antigen-antibody complexes constitutes an important initial phase of the febrile reaction in this type of immune fever.

The genetic control of antibody specificity.

The immune response to a synthetic polypeptide built on multichain polyproline, poly-L-(Tyr,Glu)-poly-L-Pro-poly-L-Lys [(T,G)-Pro--L], in the offspring of a cross between DBA/1 and SJL mice is under a genetic control superficially similar to the one operating for the immune response to a similar synthetic polypeptide built on multichain polyalanine, poly-L-(Tyr,Glu)-poly-D,L-Ala-poly-L-Lys [(T,G)-A--L], in the offspring of a cross between CBA and C57 mice. In both cases, the genetic control is a quantitative trait in which the major gene(s) is (are) dominant and the trait is not linked to any of the known structural genes coding for mouse immunoglobulin heavy chains. However, the genetic control of response to (T, G)-Pro--L, designated immune response-3 (Ir-3), is qualitatively different from the one operating for (T,G)-A--L [immune response-1 (Ir-1)] in that it is not linked to the histocompatibility-2 (H-2) locus. A study of the immune response to a related polypeptide built on multichain polyproline, poly-L-(Phe,Glu)-poly-L-Pro-poly-L--Lys [(Phe, G)-Pro--L], in the DBA/1 x SJL cross has shown a genetic control of antibody specificity. F(1) x DBA/1 backcross anti-(Phe, G)-Pro--L sera segregate in their ability to bind (T,G)-Pro--L, and there is no linkage of anti-(T,G)-Pro--L binding capacity with the H-2(s) allele of the SJL grandparent. F(1) x SJL anti-(Phe, G)-Pro-L sera segregate in their capacity to bind poly-L-(Phe,Glu)-poly-D,L-Ala-poly-L-Lys [(Phe, G)-A--L] and the ability to bind (Phe, G)-A--L is clearly linked to the H-2(q) allele from the DBA/1 grandparent. Thus, in mice all responding well to a given antigen [(Phe, G)-Pro--L], the specificity of the antibodies produced [i.e., anti-(Phe,G) or anti-prolyl] is genetically determined. Cross-inhibition of binding m (DBA/1 x SJL)F(1) anti-(Phe,G)-Pro--L antisera indicates that the anti-(Phe,G) and anti-prolyl specificities are a function of two separate and largely non-crossreacting antibody populations.

Receptors on immunocompetent cells. II. Specificity and nature of receptors on dinitrophenylated guinea pig albumin- 125 I-binding lymphocytes of normal guinea pigs.

Nonimmunized guinea pigs possess rare lymphocytes which bind sufficient 2,4-dinitrophenyl-guinea pig albumin-(125)I (DNP-GPA) to their surface to be detected by short-term radioautography. The cells occur in the lymph nodes, spleen, peripheral blood, and bone marrow with a frequency of approximately 40/100,000 lymphocytes, but are absent from the thymus. The receptors of these cells are largely specific for the haptenic group (epsilon-DNP-L-lysine) as shown by inhibition of DNP-GPA-(125)I binding with epsilon-DNP-L-lysine and with DNP bovine serum albumin (DNP-BSA). Furthermore, these cells specifically adsorb to agarose beads to which either DNP-GPA, DNP-BSA, or DNP-keyhole limpet hemocyanin (KLH) has been covalently linked. This hapten specific depletion of DNP-GPA-(125)I antigen-binding cells (ABC) correlates with a similar diminution in the capacity of adsorbed populations to transfer primary responsiveness to DNP-KLH to irradiated syngeneic recipients. Fluoresceinated anti-immunoglobulin binds to the surface of some guinea pig lymphocytes, and all DNP-GPA-(125)I ABC, as shown by a double-label technique. The great majority of DNP-GPA ABC and human gamma-globulin ABC possess surface Ig molecules of the gamma(2) heavy chain class. Preincubation of cell suspensions with anti-gamma(2) antibody markedly diminishes the number of DNP-GPA-(125)I ABC which are detected, strongly suggesting that the receptors of these cells are immunoglobulin molecules, most of which possess gamma(2) heavy chains. The specificity characteristics of DNP-GPA-(125)I ABC are strikingly different from those of cells mediating a cellular immune response to DNP-GPA, indicating major differences in the specificity and nature of the receptors of these cell types.

Immunoreactive basement membrane antigens in normal human urine and serum.

Using a sheep antiserum to human glomerular basement membrane (GBM), studies of urine from healthy adults showed the presence of two cross-reactive antigens. These antigens were purified partially by preparative electrophoresis and electrofocusing, and separated on G-200; both appeared to be acidic, of high molecular weight, and carbohydrate rich. Their immunologic relationship to human GBM solubilized by several techniques was deduced from lines of identity with the native GBM digests in double diffusion analyses. These antigens will combine with homologous anti-GBM antibodies and block their fixation to human kidney sections, and will evoke heterologous anti-GBM antibody production in the rabbit. Fractionation studies of normal human serum indicated the presence of trace amounts of basement membrane antigens in the circulation. Although the serum antigens appear immunologically identical to the urinary antigens, the precise anatomic structures from which both are derived is not certain. Demonstration of immunoreactive basement membrane antigens in the circulation provides a plausible source of immunogen for the potential development of anti-GBM antibody-mediated glomerulonephritis as well as a clue to a mode for reestablishment of tolerance in such an autoimmune disorder.

Complement-dependent release of immune complexes from the lymphocyte membrane.

Soluble antigen-antibody-complement complexes bound to mouse B lymphocytes are rapidly released from the cell membrane in the presence of normal serum from several mammalian species. The release is not the result of antigen-antibody dissociation or extensive degradation of the complexes. However, the released complexes have been altered because they will no longer bind to fresh lymphocytes. The release is not the result of lymphocyte damage mediated by complement. It is complement-dependent, and is generated either preferentially or exclusively via the alternate pathway, since it occurs in C4-deficient serum, is Mg(++) but not Ca(++) dependent, and requires C3 proactivator. C3 inactivator is not involved. The release activity of the serum, once generated, is unstable at 37 degrees C. The release of complexes from the lymphocyte membrane by serum provides a convenient assay for the functioning of the alternate pathway in the mouse and in other species.

Experimental chronic glomerulitis.

Three of 16 rabbits injected (intravenously) daily with crystalline bovine serum albumin (BSA) for periods in excess of 10 wk developed chronic glomerulonephritis. In vivo, animals with chronic proteinuria formed variable quantities of soluble complex after injection of antigen while animals without proteinuria exhibited rapid removal of the injected BSA. In vitro studies demonstrated that a major part of the antibodies produced by rabbits with chronic nephritis lacked precipitating properties. Interpretations of these observations were presented in the discussion. It is suggested that, in addition to quantity, quality of antibody plays an important role in the development of chronic serum sickness. Complexes formed with nonprecipitating antibody, which are less rapidly removed from circulation, would have a greater opportunity to deposit in glomeruli and induce inflammation.

The effect of cross-reacting antigens on the tolerant state.

Rabbits were rendered tolerant to human albumin (HA) and were then injected with azo and oxazolonated derivatives of human albumin. These injections were continued to a time at which all animals would have lost tolerance if they had not been injected. Injection of cross-reacting antigens prolonged the duration of tolerance, as judged by the mode of elimination of lightly iodinated human albumin (HA.(131)I). Different derivatives of HA differed in their capacity to prolong tolerance. Those neonatally injected rabbits which were immunized with cross-reacting antigens and lost tolerance, responded much more promptly to HA.(131)I than animals which were not immunized. Animals immunized with cross-reacting antigen which went on to eliminate HA.(131)I triphasically, usually had responded earlier by making antibodies. These antibodies contained a fraction which was reactive with HA, and which was usually equally well adapted to determinants on HA and on the cross-reacting antigen.

The effect of tolerance on the specificity of the antibody response and on immunogenicity. Antibody response to conformationally and chemically altered antigens.

Animals were rendered tolerant to human albumin and were then immunized with azo derivatives of human albumin which differed in the number of hapten groups per molecule and in the extent of conformational change. The incidence and specificity of the resulting antibody response was studied and the presence of antibody to azo groups and to conformationally altered protein determinants was demonstrated. Reactivity with the tolerance-inducing antigen was shown to be due to antibodies directed against conformationally altered protein determinants. The difference in the response of tolerant animals to hapten-poor and hapten-rich derivatives was attributed to the extent of conformational alteration. A genetic factor appeared to be implicated in the capacity of tolerant animals to respond to an antigen which cross-reacts with tolerance-inducing macromolecules.

Heterogeneity of antibody-forming cells. An electron microscopic analysis.

Cell suspensions from draining lymph nodes of immune and nonimmune rats were reacted in vitro with (125)I-labeled antigens. In light microscopic radioautographs of smears, 17% of the immunized cells were tagged by specific antigen; 2.0% of control cells were positive. In electron microscopic radioautographs, 90% of the labeled elements from immune donors were lymphocytes, blast and plasma cells; 10% were monocytes-macrophages or other elements, including naked nuclei. 15% of the labeled cells from control materials were lymphocytes and plasma cells, while 85% were monocytes-macrophages and naked nuclei. Within cell suspensions derived from immunized animals there were almost twice as many lymphocytes marked by isotope as plasma cells, and the lymphocytes ranged in morphology from mature monoribosomal elements to immature polyribosomal cells. Antibody-forming cells fixed labeled antigen at their surfaces. The monocyte-macrophage class was distinguished by a high mean grain count and by distribution of grains within cytoplasmic vacuoles and lysosomes.

The sensitization of rats by allografts transplanted to alymphatic pedicles of skin.

Pedicles of skin which lacked a lymphatic drainage were raised on the backs of rats in order to study the importance of afferent lymphatics in sensitization by skin allografts. Although allografts transplanted to the alymphatic pedicles enjoyed a prolonged survival, they contracted progressively from about 3 wk after transplantation and were reduced eventually to small scars. In contrast, autografts survived unchanged in size for the life-span of the pedicles which carried them. The slow contracture of the allografts was associated with sensitization of the host because test allografts applied orthotopically were destroyed with a second-set tempo. No regeneration of lymphatics from the long-standing pedicles could be demonstrated, and it was concluded that sensitization had occurred eventually through the blood, presumably by the process of peripheral sensitization. Allografts on skin pedicles could be destroyed rapidly by active or adoptive immunization, so it is probable that the level of sensitization to which they themselves gave rise was a low one. Although it is not disputed that afferent lymphatics are essential for the rapid destruction of skin allografts, it is clear that the absence of a lymphatic supply does not permanently exempt them from immunological attack in the rat.

Acute immune complex disease in rabbits. The role of complement and of a leukocyte-dependent release of vasoactive amines from platelets.

By depletion of C3 from rabbits undergoing acute experimental immune complex disease with an anticomplementary factor in cobra venom, it has been possible to demonstrate that deposition of the complexes in arteries and glomeruli does not require the complement components reacting after C2. Immunological reactions, in which platelets release their vasoactive amines, have been examined in rabbits undergoing immune complex disease. A correlation was obtained between the presence of a complement-independent reaction which required blood leukocytes, antigen and platelets, the deposition of immune complexes, and the induction of glomerulonephritis. C3 depletion did, however, have a marked alleviating effect on the severity of the arterial lesions. Neutrophil accumulation and the subsequent necrotizing arteritis were prevented. In contrast, the character and severity of the glomerulonephritis was not altered by depletion of later-acting complement components.