Biocompatibility of dialysis membranes: effects of chronic complement activation

The ability of three dialysis membranes (cuprophane, cellulose acetate, and polymethylmethacrylate) to activate complement was studied prospectively in ten chronic dialysis patients using new and reused membranes. Patients were dialyzed for 1 month with each type of membrane. New cuprophane membranes caused the most intense activation, while polymethylmethacrylate (PMMA) surfaces caused the least degree of complement activation. Reuse decreases the capacity of the cuprophane membrane to activate complement but does not significantly alter the capacity of cellulose acetate membranes. The extent of complement activation paralleled the ability of these membranes to induce neutropenia. Recurrent dialysis with new cuprophane and cellulose acetate membranes leads to a decrease in pre-dialysis and "rebound leukocytosis" neutrophil count, as well as a more intense activation of complement and an enhanced endogenous clearance of products of complement activation. The clinical sequelae of recurrent complement activation are discussed.     

Apoptosis of human polymorphonuclear neutrophils accelerated by dialysis membranes via the activation of the complement system.

BACKGROUND: Haemodialysis (HD) with bioincompatible cellulosic membranes like Cuprophan (CU) is considered to influence negatively the clinical outcome of acute and chronic renal failure. In this effect, apart from the disturbance of phagocytosis or oxygen species production by leukocytes, increased apoptosis also has been implicated recently. The objective of this study was to study the effect of HD membranes on apoptosis induction in polymorphonuclear neutrophils (PMN). METHODS: PMN from healthy donors and uraemic patients were isolated and apoptosis was induced by co-incubation with CU, Hemophan or polyamide hollow fibres in the presence of serum from healthy or uraemic humans. Apoptosis was quantified by flow cytometry using Annexin V-FITC and propidium iodide staining and was confirmed by the detection of DNA fragmentation on gel electrophoresis. The deposition of immunoglobulins (Ig) and complement factors on hollow fibres was detected by direct immunofluorescence. RESULTS: Heat inactivation or the depletion of complement components or Ig significantly reduced apoptosis, indicating its dependence on classical complement activation. The detection of IgG on hollow CU fibres and the restored acceleration of apoptosis by the appropriate replenishment of Ig-deficient sera additionally confirmed these findings. Inhibition experiments revealed that caspases were necessary mainly, but not exclusively, for apoptosis to occur after complement activation. Uraemia led to increased PMN apoptosis in the presence of bioincompatible, but not biocompatible, membranes. CONCLUSIONS: Our results suggest that the acceleration of PMN apoptosis in the presence of CU is mediated via an antibody-dependent activation of the classical complement pathway mobilizing both caspase-dependent and -independent pathways.     

Protein loss induced by complement activation during peritoneal dialysis

A variable loss of macromolecules during peritoneal dialysis has been noted in both humans and experimental animals. We investigated the potential role of the complement system for inducing protein loss during peritoneal dialysis, both to shed light on clinical variability of protein loss and to develop a model for quantitatively studying complement-induced microvascular protein leakage. Rats received intra-arterial injections of a fluorescent dye conjugated to rat serum albumin and underwent a 3.5-hr series of 15-min peritoneal dialysis exchanges. After the control exchanges, rats received either intra-arterial zymosan-activated rat serum, saline, unactivated rat serum, or endotoxin; other rats received an intraperitoneal injection of endotoxin, histamine, phenylephrine, or nitroprusside. The drainage volume from each exchange was measured, and the concentrations of labeled albumin, total protein, and urea were determined by spectroscopy. Zymosan-activated rat serum and endotoxin injections (both intraperitoneal and intra-arterial), each of which may activate the alternative pathway of complement, produced a dramatic increase in dialysate protein concentrations. In addition, histamine, which is a vasodilator but which may also be involved as a mediator of the activated complement system and/or endotoxemia, also produced an increase in dialysate protein concentrations. On the other hand, drugs which may alter peritoneal blood flow such as the vasodilator nitroprusside or the vasoconstrictor phenylephrine, did not affect dialysate protein concentrations. These data suggest that activation of the alternative pathway of complement may cause variation in protein loss during peritoneal dialysis and that in some situations, pharmacological control of this system could be an important therapeutic consideration.     

Modulation of complement activation on hemodialysis membranes by immobilized heparin.

To determine the effects of surface-associated heparin on the capacity of hemodialysis membranes to activate complement, cellulose acetate (CA) membranes that were untreated and CA membranes that had been coated with heparin (HCA) were incubated with C3-depleted serum repleted with radio-labeled C3. Next, the proteins in the supernatant and those eluted from the membranes were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. C3 activation was quantified by determining the radioactivity of the C3a-containing band in the gel. Total C3a generation (fluid phase C3a plus membrane-associated C3a) was three times greater in the presence of HCA compared with CA. Most (88%) of the C3a generated in the presence of HCA, however, was adsorbed onto the membrane surface. Consequently, there was more C3a in the CA supernatant than in the HCA supernatant. To determine the mechanism by which heparin enhanced alternative pathway activity, binding studies with radiolabeled factor B and factor H were performed. HCA bound 3.4 times more factor B and 20 times more factor H than did CA. The binding of these proteins, however, was not dependent on complement activation. Studies designed to test the functional activity of isolated factor H and factor B that had been adsorbed to the membrane showed that factor H was active on both CA and HCA, whereas factor B was active only on HCA. These data demonstrate that heparin immobilized onto CA hemodialysis membrane enhances C3 activation but produces low levels of C3a in the fluid phase because of high surface adsorption of the anaphylatoxin. Heparin appears to augment alternative pathway activity by favoring the interactions of factor B with other constituents of the amplification C3 convertase of the alternative pathway of complement.     

Phosphate removal in haemodialysis: Effect of two different dialysis membranes

Summary: Phosphate removal with two different dialysis membranes (a standard cellulose acetate membrane and a high performance cellulose diacetate membrane) were studied in ten haemodialysis patients. All patients were dialysed sequentially with two membranes (surface area was 1.5 m²) against bicarbonate buffered dialysis for 4 hours three times a week. With the diacetate membrane, the instantaneous clearances of urea and phosphate after 1 hour of haemodialysis were significantly higher than with the cellulose membrane. Also, the weekly total amount of urea and phosphate removal were significantly increased with the diacetate membrane (a 15% increase in urea and a 16% increase of phosphate). Although there was a significant increase in urea reduction ratio and significantly lower post dialytic plasma urea concentration with the diacetate membrane, these for phosphate did not reach statistical significance. These data suggest that the use of the diacetate membrane potentially offer clinical benefit. However, whether 16% of phosphate removal could improve clinical control of serum phosphate levels will need further investigation.     

Influence of dialysis with polyamide vs haemophan haemodialysers on monokines and complement activation during a 4-month long-term study.

BACKGROUND: Contact between blood and dialysis membranes activates mononuclear cells and the complement system. The extent of activation is dependent on the dialyser material used and is considered an index of biocompatibility. Polyamide dialysers consist of a synthetic membrane that claims high standards of biocompatibility. Haemophan dialysers represent membranes made of modified cellulose that are now broadly used for treatment in Europe and are already considered to be more biocompatible than the cuprophane membranes that were used as reference in most previous studies. METHODS: In a cross-over treatment study short-term as well as long-term effects of a polyamide dialyser with respect to monokine induction and complement activation were compared to a haemophan dialyser. RESULTS: Neither haemophan nor polyamide dialysers induced relevant changes in plasma monokine levels. However, in vitro challenge of mononuclear cells with lipopolysaccharide (LPS) unmasked a significantly stronger preactivation for the secretion of proinflammatory monokines during haemophan than polyamide dialysis. Unlike other monokines the production of the regulatory monokine IL-10 was mainly influenced by individual factors and correlated with the patient's immune status rather than the dialyser type used. Enhanced preactivation of monocytes in haemophan compared to polyamide dialysis was paralleled by an increased complement activation. Cellular preactivation and formation of terminal complement complex remained constant over the 4-month treatment period. CONCLUSIONS: Haemophan and polyamide dialysers do not induce changes in plasma cytokine levels both during short-term and long-term use. However, they significantly differ in complement activation as well as preactivation of monocytes. Preactivated monocytes are prone to secrete high amounts of proinflammatory cytokines when exposed to a second stimulus like endotoxin. Secretion of the regulatory cytokine IL-10 is not influenced by the dialyser type.     

Peritoneal dialysis, membranes and beyond

PURPOSE OF REVIEW: The peritoneal membrane provides the interface between dialysate fluid and blood for peritoneal dialysis patients. Functional properties of the peritoneal membrane have important clinical implications. This review will outline recent observations concerning structural changes in the peritoneal membrane and the impact on function and clinical outcomes. RECENT FINDINGS: Peritoneal membrane function - solute transport and ultrafiltration - is a complex process involving new blood vessel growth along with changes in the nature of blood vessels and the interstitial environment of these vessels. Advanced glycation end-products produced by reactive oxygen species in the dialysis fluid have been identified as an agent of tissue fibrosis. Nitric oxide and IL-6 also have important roles in peritoneal membrane injury. Gene polymorphisms associated with peritoneal membrane function have been identified. As the mechanisms of peritoneal membrane injury become better elucidated, targeted therapies are being developed. The role of biocompatible and nonglucose dialysis fluids needs to be further defined. SUMMARY: The peritoneal membrane is the lifeline for peritoneal dialysis patients. Our understanding of mechanisms of injury and functional responses continues to expand and will hopefully lead to therapies to improve the clinical outcomes for peritoneal dialysis patients.     

Blood substitution and complement activation

Complement activation was studied in 45 patients undergoing total hip arthroplasty under epidural anesthesia. The patients were randomly allocated to three groups. In Group I blood loss was replaced with microaggregate-poor erythrocyte concentrate (SAGM-ERC) plus 3% dextran-60 as plasma substitute, and postoperative analgesia was maintained with intramuscular ketobemidone. In Group II blood loss was replaced as in Group I, but epidural anesthesia was prolonged 12 h postoperatively and kept at a level of T4 with 0.5% bupivacaine. In Group III blood loss was replaced with non-frozen stored plasma plus SAGM-ERC, and postoperative analgesia was maintained with ketobemidone as in Group I. All groups received pre- and postoperative thrombo-prophylaxis with dextran. The plasma concentration of C3a-des-arginine (C3a-desArg) was measured by radioimmunoassay preoperatively, immediately after operation and 3, 6 and 18 h postoperatively. No significant differences in plasma C3 and C4 were found between the groups. C3a-desArg was significantly (P less than 0.01) increased up to 6 h postoperatively in Group III compared with both the preoperative value and Groups I and II. It is demonstrated that infusion of plasma can enhance or initiate endogenous complement activation. Blood component therapy with SAGM-ERC and 3% dextran-60, on the other hand, did not significantly increase the plasma level of C3a-desArg irrespective of the type of postoperative analgesia.     

Case control study on dialysis arthropathy: the influence of two different dialysis membranes: data from the EDTA Registry

In a retrospective case control study the prevalence of signs and symptoms of dialysis osteoarthropathy was analysed. Cases and controls had received over 9 years of maintenance haemodialysis uninterrupted by peritoneal dialysis or transplantation. The cases comprised 55 patients treated predominantly with polyacrylonitrile (AN69) dialysers. They were compared to a matched group dialysed exclusively with cellulosic membranes. Over 60% of all patients, cases and controls, showed one or more signs of disabling osteoarthropathy, with joint pains occurring more frequently in the older age groups. Twenty-seven of the 55 cases who had received less than 2 years of cellulosic membrane dialysis followed by 7-12 years of AN69 dialysis tended to have a lower prevalence of joint pains, carpal-tunnel syndrome and bone cysts. However, no statistically significant differences were obtained compared to the matched control group dialysed exclusively on cellulosic membranes (mostly cuprophane). The remaining 28 cases, who had been treated for more than 2 years with cellulosic membranes preceding the longer treatment period with polyacrylonitrile dialysers, showed a prevalence similar to that of their cellulosic controls. This study thus shows little, if any, influence of the two types of membranes on the prevalence of signs and symptoms of beta 2-microglobulin amyloidosis.     

Analysis of the complement C3 fragments associated with hemodialysis membranes

During hemodialysis with cuprophan membranes, bioactive peptides are generated because the alternative pathway of complement is activated. When cellulose acetate membranes are employed, complement activation is attenuated. The molecular basis for this improved biocompatibility is unknown. It has been postulated, however, that the complement activating potential of dialysis membranes is influenced by the availability of free hydroxyl groups which would provide an acceptor site for activated C3. To investigate this hypothesis, the forms of C3 associated with cellulose acetate and cuprophan membranes have been analyzed. By Western blot, the predominant form of C3 present on both types of membranes is C3c, a degradation product of C3 that lacks the thiolester necessary for covalent bonding. Minimal amounts of C3d (the region of C3 which contains the thiolester) were observed on both membranes; however, by ELISA, there was no difference in the amount bound to cellulose acetate compared to cuprophan. Further, membrane-associated C3d could be removed by urea, suggesting that it was not bound covalently. These studies indicate that the complement activating potential of dialysis membranes is not determined primarily by the availability of potential covalent binding sites for activated C3b.     

Effect of regional citrate anticoagulation on leukopenia, complement activation, and expression of leukocyte surface molecules during hemodialysis with unmodified cellulose membranes

BACKGROUND: Dialysis with complement-activating membranes is associated with leukopenia, which is related to an increased expression of adhesion molecules on leukocytes. Citrate chelates calcium and has been claimed to attenuate leukopenia. METHODS: In this study, the effects of citrate anticoagulation on leukocyte and granulocyte counts, complement activation, and the expression of CD11b, CD11c, and CD45 on the surface of granulocytes were evaluated during hemodialysis with unmodified cellulose membranes. Standard heparin was compared to citrate in three different schedules: citrate was infused to obtain a concentration of either 7 or 10 mmol/l blood. CaCl(2) was administered into the dialyzer outlet at 8. 25 mmol Ca(2+)/h (citrate 10 mmol/l) or at 11 mmol Ca(2+)/h (citrate 7 and 10 mmol/l) to reconstitute the calcium levels in the blood returning to the patient. RESULTS: The use of citrate at a high concentration (10 mmol/l) was associated with a blunted upregulation of CD11b, both at the inlet and at the outlet bloodline; for CD11c a reduced upregulation was observed on granulocytes harvested from the inlet bloodline. No effects of citrate were observed on leukopenia, granulocytopenia, or complement activation. A positive correlation between the decrease in systemic ionized Ca(2+) concentration and the increase in CD11b and CD11c expression was found. CONCLUSION: Citrate/CaCl(2) administration affects leukocyte adhesion molecule expression in a dose-dependent way; however, no significant effect could be demonstrated on leukopenia and complement activation.     

Interaction of prostaglandins, activated complement, and granulocytes in clinical sepsis and hypotension

Activated complement, thromboxane A2, prostacyclin, and activated granulocytes have been implicated in hemodynamic dysfunction after trauma, in sepsis, and in hypovolemic and septic shock. This study evaluated the interaction of plasma concentrations of complement components C3a and C5a, thromboxane B2 (TxB), prostaglandin 6-keto-F1 alpha (PGI), and granulocyte aggregation in clinical sepsis and hypotension. Forty-eight critically ill patients were followed clinically for as long as 10 days. Plasma C3a, C5a, TxB, and PGI were measured daily by the radioimmunoassay method. Granulocyte aggregation, the percentage of maximum aggregation of zymosan-activated plasma standard curves, was performed with patient plasma and normal human leukocytes. Patients were studied in four groups: group I, nonseptic, normotensive; group II, hypovolemic shock, group III, normotensive severe sepsis; and group IV, septic shock. Plasma from 12 normal adults was the control value. PGI, TxB, C3a, C5a, and granulocyte aggregation in patients were greater than that in the control subjects. Granulocyte aggregation was increased in groups III and IV versus groups I and II. C3a was increased in group IV versus groups II and III. C5a and TxB did not vary between groups. PGI was greatly increased in group IV compared with groups I through III. C3a and C5a decreased in nonsurvivors. PaO2/FiO2 ratios correlated directly with PGI and inversely with C3a and TxB/PGI. Plasma PGI and C3a are increased in septic shock. C3a and TxB/PGI imbalances are involved in hypovolemic and septic shock.