TY - JOUR
AU - Jacobs,, Enno
AB - Abstract Background. Neutrophil functions in haemodialysis (HD) patients are altered by uraemia and by HD procedure. We investigated details of the neutrophil dysfunction as its nature and origin is not well understood. This is reflected by conflicting results about neutrophil phagocytosis activity and by scarce data on the neutrophil killing capability in HD patients. Methods. Using a flow-cytometric test system we have measured simultaneously phagocytosis and the production of reactive oxygen species (ROS) of neutrophils and in parallel antimicrobial killing of yeast by neutrophils. 117 whole-blood samples of healthy controls and 50 pre- and 50 post-dialysis samples of HD patients, half of them with diabetes mellitus (DM), have been evaluated. We have constructed a model to account for the dependence on the stimulus-to-cell ratio and obtain means for phagocytosis and killing at different incubation times. Results. (i) HD patients have significantly lower neutrophil killing (20%) than healthy controls. (ii) Dialysis improves the killing capability by 10–15%, after dialysis the killing activity remains significantly (10%) below that of the controls. (iii) The percentage of neutrophils, which exhibit phagocytosis and produce ROS, does not differ significantly between HD patients and healthy controls. (iv) Age has no significant influence on phagocytosis and killing. Conclusion. The neutrophil killing capability is reduced in HD patients while the amount of neutrophils that phagocyte and produce ROS remains unchanged. Functional impairment of uraemic neutrophils is therefore mainly a result of their reduced capability to kill microorganisms intracellularly. antimicrobial killing, flowcytometry, haemodialysis, neutrophils, phagocytosis, uraemia Introduction As neutrophils play a crucial role in host defence against bacterial and fungal infections, neutrophil functions such as the production of reactive oxygen species (ROS) or phagocytosis have been studied extensively. This is especially true for patients undergoing haemodialysis (HD) for end-stage renal disease because infection is one of the leading causes of morbidity and mortality in these patients. In this context functional impairment of neutrophils in uraemia is believed to increase the susceptibility to infections [1]. The mechanisms responsible for reduced neutrophil functions are not well understood and have been attributed to iron overload, zinc deficiency, increased intracellular ionized calcium, anaemia, malnutrition, short time on dialysis and dialysis therapy per se [1,2]. Moreover, a number of uraemic toxins that affect neutrophil functions have been identified and characterized [3]. However, studies investigating phagocytosis or the production of reactive oxygen (ROS) in HD patients have produced conflicting results. Decreased, normal or increased levels of ROS-production and phagocytosis have been reported [3–13]. Studies about the influence of the dialysis membrane on the oxidative burst lead also to controversial results [10,12,14,15]. Some contradictions certainly can be attributed to different methods for the measurement of neutrophil functions. Many of the procedures used to isolate neutrophils cause artificial changes such as reduction of oxidative burst and chemotaxis. Moreover, it has been shown that the ratio of stimulating agents to cells, the type of stimuli, as well as incubation times influence measured neutrophil parameters [16]. This renders a comparison among published results difficult and may explain contradicting results. In contrast to phagocytosis and ROS-production, similar data from HD patients on neutrophil antimicrobial killing in conjunction with other neutrophil functions is lacking and no study is available where the neutrophil functions have been measured simultaneously. It should be recalled that the stimulation of ROS-production is not the only mechanism for killing microorganisms and therefore it is not uniquely related to neutrophil antimicrobial killing. Moreover, measurements of oxidative burst are heavily influenced by methodological procedures. Neutrophil function measurements in general are delicate since personal constants and day-to-day variations lead to large fluctuations in the data. Hence, large data sets are needed to render a comparison between HD patients and healthy controls significant [17]. In the present study the influence of incubation time and stimulus-to-cell ratio R on neutrophil functions has been quantified. To determine phagocytosis, ROS-production and antimicrobial killing of neutrophils from HD patients we use a recently described flowcytometric method. It allows one to measure these functions simultaneously directly from whole blood. With respect to ROS, it should be noted, however, that while the percentage of neutrophils producing ROS can be determined reliably with the present assay, this is not the case for the intensity of ROS (mean channel of fluorescence) [18]. We have modelled the ratio dependence of the neutrophil functions and use a non-linear regression to obtain reliable averaged values, which do not depend on the ratio. We have analysed the influence of uraemia and dialysis on neutrophil functions in increasing detail, comparing results of: (i) the control group with the group of HD-patients; (ii) the control group with the group of patients before and after dialysis; and (iii) discriminating between patients with and without diabetes mellitus (DM). Subjects and methods Characteristics of patients and healthy controls Fifty pre-dialysis and 50 post-dialysis venous blood samples of 25 HD patients entered the study. Patients have been evaluated twice on different days within 1 month. The mean age was 60.8 ± 15.6 years, 13 males and 12 females. Patients with acute or chronic infections, active immunological diseases, immunosuppressive or antibiotic therapy, a history of malignancy, haematological or liver disease were excluded. Twelve patients had DM, all of them took insulin regularly, the mean HbA1C was 7.1 ± 1.4%. The drug therapy included ACE-inhibitors in six patients and acetylsalicyl acid (100 mg) in four patients. No patient was treated with statins. All patients had been on maintenance HD for between 18 and 168 months (mean 80 ± 60 months). They underwent a regular dialysis schedule, three times a week for 4–5 h with a blood flow of between 250 and 300 ml/min. The Kt/V (single pool) was >1.3 in all patients. All patients were dialysed with biocompatible synthetic membranes [Asahi Polysulfone APS 650 and APS 900 (Diamed Medizintechnik GmbH), F40 and F60, Fresenius Polysulfone (Fresenius Medical Care)]. The dialysate was bicarbonate-based and contained 1.25 mmol/l calcium. Anticoagulation was performed according to routine heparine prescriptions with a heparine bolus followed by a constant heparine infusion. Venous blood samples were taken 3 days after the last dialysis directly before and 2–5 min after the end of a dialysis session. Twenty-three patients were treated with erythropoetin to achieve haemoglobin levels >10 g/dl (mean haemoglobin 6.78 ± 0.92 mmol/l = 10.92g/dl). The calcium phosphorus product is not considerably elevated (serum calcium 2.21 ± 2.6 mmol/l, serum phosphorus 1.78 ± 0.52 mmol/l), none of the patients had severe hyperparathyreoidism. Six patients received i.v. iron supplementation to avoid iron deficiency; the interval between the latest iron administration and the blood sampling was at least 3 days (serum iron 10.51 ± 3.31 μmol/l, serum ferritin 658 ± 406 μg/l). Leucocytes were in the normal range in all patients (6.61 ± 1.79/nl) and healthy volunteers. None of the patients or healthy volunteers had monocytosis. 117 venous blood samples were analysed from 49 healthy volunteers (22 male, 27 female) with an average age of 35.46 ± 14.3 years. None of the volunteers had acute or chronic diseases or took drugs regularly. Heparine (heparin-sodium without preservatives) was used as anticoagulant (5 IE heparin/1 ml blood) for all blood samples. All patients and healthy volunteers participating in this study gave informed consent approved by the Hospitals Ethical Committee. Study design We have determined phagocytosis, oxidative burst and killing activity of neutrophils together with monocytes as described recently [18]. The slightly modified protocol used for the present study is briefly described in the following. Cell labelling Culture of Candida albicans (DMS 1386) was performed in tryptic soy broth (Difco Laboratories, Detroit, MI). For fluorescence labelling of C.albicans, broth was centrifuged (500 g, 15 min), the pellet was re-suspended in 1 ml of phosphate-buffered saline/glucose [100 ml PBS (Sigma, St Louis, MO) containing 0.5 g glucose, pH 7.5] with 5 μl of Calcein-AM (Molecular Probes, Eugene, OR) and incubated for 50 min at 37°C in a thermomixer (Eppendorf-Nehler-Hinz GmbH, Germany) at 600 r.p.m. The stained yeast was washed in PBS, re-suspended in RPMI (RPMI-1640, Sigma) and stored in a refrigerator for maximal 7 days. The number of yeast per millilitre RPMI was determined with Neubauer’s counting chamber, the percentage of stained yeast was controlled by flow cytometry (>97%). For detecting burst reaction, blood was incubated with Dihydroethidium (DHE, Sigma, 20 μl of PBS containing 0.5 μg of DHE/1 ml blood) for 10 min at 37°C in a shaking water bath. ROS released by neutrophils oxydize the dye, which turns from colourless to red fluorescent. To identify neutrophils and monocytes, they were stained with orange fluorescence after the reaction had stopped (see below). The pellet of each tube was re-suspended with 100 μl of PBS containing 1.8 μl of CD13-RPE antibody (Coulter, Krefeld, Germany) and incubated at room temperature for 20 min. For labelling dead yeast, after lysis of neutrophils and centrifugation (see below) the pellets were each mixed with 800 μl of 2 μM ethidiumhomodimer-1-solution (EthD-1, Molecular Probes), incubated for 10 min at room temperature mixed and stored on ice. Preparation of blood samples RPMI solution containing C.albicans was diluted to achieve the desired ratio R between neutrophils and C.albicans. According to the chosen kinetics for incubation (0, 6, 10 and 30 min for phagocytosis/burst, 0, 15, 30, 60 and 120 min for killing) polypropylene tubes were prepared with 100 μl of RPMI and stored on ice. Blood (after incubation with DHE, see above) and microorganisms were mixed, aliquots of 200 μl were quickly dispensed into the test tubes (separate tubes for phagocytosis and killing assay). Hence, strictly speaking only phagocytosis and ROS active neutrophils are measured simultaneously, while killing is measured in parallel. For simplicity we call this type of measurement simultaneous with respect to phagocytosis, ROS and killing in the rest of the manuscript. Assay for phagocytosis and oxidative burst The tube for 0 min was immediately mixed with 400 μl of N-ethyl-maleinacidamide (NEM, Sigma) and stored on ice, the other tubes were incubated at 37°C in a thermomixer. At appropriate times the reaction was stopped by addition of 400 μl of ice-cold NEM; tubes are stored on ice. The tubes were then centrifuged together at 4°C at 250 g for 5 min (10 min for C.albicans/neutrophil ratio >2) to avoid attachment of microorganisms at the phagocyte surface. After staining neutrophils (see above) 1800 μl lysing buffer for erythrocytes (containing 8.27 mg ammonium chloride, 1 mg potassium hydrogencarbonate and 0.04 mg sodium–EDTA) was added to each tube and incubated at room temperature for 10 min. Tubes were again centrifuged (4°C, 250 g, 5 min), 1000 μl of the supernatant was removed. After vortexing the tubes, cells were stored on ice and measured with a maximum delay of 2 h. Assay for killing After incubation of the cells (37°C, thermomixer) the reaction was stopped by placing the test tubes on ice and mixing immediately with 400 μl lysing buffer for human cells, consisting of 2 μl Triton X-100 (Sigma) and 2 μl Tween-20 (Sigma) in distilled water. After incubation at 37°C for 25 min in the thermomixer at 600 r.p.m., the tubes were centrifuged at 4°C at 1000 g for 5 min (10 min for C.albicans/ratio >2). Dead yeast cells were then stained with EthD-1-solution as described above. Cells were measured within 2 h. Flow cytometry The samples were analysed with a FACSCalibur flow cytometer (BD Biosciences, Heidelberg, Germany) using Cellquest R software (BD, version 3.3). Green fluorescence (FL1) from calcein was collected through a bandpass filter of 530/30 nm, orange fluorescence (FL2) from CD13-RPE was detected with a band-pass filter of 585/42 nm and red fluorescence (FL3) from ethidium and Eth-D1 was analysed with a longpass filter of 670 nm. Data was collected using linear amplifiers for the forward scatter (FSC) and logarithmic amplifiers for the sideward scatter (SSC), FL1, FL2 and FL3. Data was displayed in single parameter histograms or in two parameter plots. Prior to each run, 11 samples containing blood with or without yeast and all variations of red, orange and green fluorescence were measured and served as a control. In the assay for phagocytosis and burst 2500 neutrophils were counted, in the killing assay 5000 fungi. The results for phagocytosis, burst and killing were obtained through gating techniques and expressed in percentage of neutrophils which have phagocytozed (have shown oxidative burst) and percentage of dead/phagocytozed yeast as described in detail elsewhere [18]. The exact ratio R of yeast (H0) to neutrophils (N0) was determined in the FCS vs FL1 plot analysis and FSC vs FL2, respectively. Phagocytosis P is given as the fraction of phagocyting neutrophils Npvs all neutrophils N0 in per cent, P = Np/N0, similarly for burst with Nb instead of Np for the neutrophils exhibiting burst, B = Nb/N0. Killing, k, is defined as the number of killed fungi Hk per all fungi H0 in per cent, k = Hk/H0. Finally, we determine the number of phagocytozed fungi Hp in percent of all fungi, hp = Hp/H0. Methods of evaluation To account for the dependence of the neutrophil functions X on the ratio R (stimulus to cell) we have constructed a model which predicts Xt(R) for a specific incubation time t according to the formula log(X(R)) = log(aR [1 – exp(–b/R)]), where a and b are parameters to be determined by a least square fit through the data. The model is applicable to the measured phagocytosis (X = P) and killing (X = K) data with a typical χ2 between 0.1 and 1 and a correlation coefficient r > 0.9. In the limit of high fungal concentration we have from the formula X(R → ∞) ≈ ab = const. In the case of phagocytosis, one knows that P saturates at 100%. Therefore, b = 100/a, leaving only one fitting parameter a. It has the meaning a = Np/H0 for low fungal concentration R → 0, i.e. phagocyting neutrophils Np per supplied fungi H0. This follows from X(R → 0) ≈ aR. Figure 1 illustrates the evaluation with the group of healthy volunteers showing Pt for different incubation times t. Fig. 1. Open in new tabDownload slide Phagocytosis Pt of the healthy controls as a function of the ratio R, fungi/neutrophils, for different incubation times t of 0 min (diamond, dashed-dotted) 6 min (+, dotted), 10 min (x, dashed) and 30 min (square, solid). The lines are fits according to our model; see text. Fig. 1. Open in new tabDownload slide Phagocytosis Pt of the healthy controls as a function of the ratio R, fungi/neutrophils, for different incubation times t of 0 min (diamond, dashed-dotted) 6 min (+, dotted), 10 min (x, dashed) and 30 min (square, solid). The lines are fits according to our model; see text. The same formula describes the killing function X = K = kR, where k = Hk/H0, the killed fungi Hk as percentage of all fungi H0. Here, we have two fitting parameters a and b, since killing does not need to saturate for large R, reflecting the fact that one phagocyte may kill more than one fungus. Again, we extract the killing function in the limit of small ratio R where we obtain from K(R → 0) ≈ aR that k = a for R → 0.Figure 2 shows the killing function for the group of healthy volunteers at different incubation times t. Note that for t = 0 the killing function K = kR does not saturate for large R but follows a straight line. This means that the killing k itself is independent of the ratio R which clearly points to the fact that this (residual) number of measured dead fungi is not due to neutrophil activity which can only set in for t > 0 and is R-dependent. Fig. 2. Open in new tabDownload slide Killing Kt of the healthy controls as a function of the ratio R, fungi/neutrophils, for different incubation times t of 0 min (diamonds, dashed-dotted) 15 min (+, dotted), 30 min (x, dashed), 60 min (triangles, long-dashed) and 120 min (squares, solid). The lines are fits according to our model; see text. Fig. 2. Open in new tabDownload slide Killing Kt of the healthy controls as a function of the ratio R, fungi/neutrophils, for different incubation times t of 0 min (diamonds, dashed-dotted) 15 min (+, dotted), 30 min (x, dashed), 60 min (triangles, long-dashed) and 120 min (squares, solid). The lines are fits according to our model; see text. The respective fitting values with their standard error for killing (a = kt ± SEt) and phagocytosis (a = pt ± SEt) at different incubation times t are the means which we subject to further statistical analysis. More precisely, we apply the unpaired Student’s t-test to see if differences in the neutrophil functions for different groups of probands as described in the introduction are significant. One might think that we could apply a paired t-test, e.g. to the same patient before and after dialysis. However, this is not possible as the ratio R can only be determined precisely a posteriori (see above) resulting in different R within one ‘pair’ of data. From the ratio hp/np one can extract the number of phagocytozed fungi per phagocyting neutrophils noting that hp/np = Hp/Np/R. Fitting hp/np = c/R yields with the fitting parameter c the desired ratio c = Hp/Np, if hp/np is proportional to 1/R. This is indeed the case, as Figure 4 shows. Fig. 4. Open in new tabDownload slide The ratio of phagocytozed fungi (as a fraction of all fungi) per phagocyting neutrophils (as a fraction of all neutrophils) in per cent, hp/np, for healthy controls (triangles), pre-dialysis (circles) and post-dialysis (squares) patients. Each curve is fitted with the function hp/np = cR and the fitting parameter c using a solid, dashed and dotted line, respectively. The fits are indistinguishable to the scale of the figure. Fig. 4. Open in new tabDownload slide The ratio of phagocytozed fungi (as a fraction of all fungi) per phagocyting neutrophils (as a fraction of all neutrophils) in per cent, hp/np, for healthy controls (triangles), pre-dialysis (circles) and post-dialysis (squares) patients. Each curve is fitted with the function hp/np = cR and the fitting parameter c using a solid, dashed and dotted line, respectively. The fits are indistinguishable to the scale of the figure. Results We will present and discuss in detail results on phagocytosis and antimicrobial killing. Our measurements of ROS active neutrophils have not revealed statistically significant differences between the different groups of patients, i.e. before and after dialysis, and with respect to healthy controls. This is illustrated with a typical result in Figure 3. For this reason we refrain from presenting further details of ROS here. Fig. 3. Open in new tabDownload slide ROS results B = Nb/N0 of burst active phagocytes (Nb) as a percentage of all phagocytes (N0) after 30 min incubation time for healthy controls (triangles, bold-solid), HD patients with DM (open circles, dashed) and without DM (filled circles, solid) before dialysis, and HD patients after dialysis (open/filled squares, dotted/dashed-dotted). No significant differences have been found between the different groups concerning the means a, which have been obtained from the simple fitting function B = aR/(1 + R/b)2 with fitting parameters a and b. Fig. 3. Open in new tabDownload slide ROS results B = Nb/N0 of burst active phagocytes (Nb) as a percentage of all phagocytes (N0) after 30 min incubation time for healthy controls (triangles, bold-solid), HD patients with DM (open circles, dashed) and without DM (filled circles, solid) before dialysis, and HD patients after dialysis (open/filled squares, dotted/dashed-dotted). No significant differences have been found between the different groups concerning the means a, which have been obtained from the simple fitting function B = aR/(1 + R/b)2 with fitting parameters a and b. Bearing in mind the different mean age of controls (35.5 ± 14.3 years) and patients (60.8 ± 15.6 years) we first investigated the influence of age on phagocytosis and killing. To this end we formed two subgroups with 15 individuals each from the controls with a mean age of 23.1 ± 2.2 and 53.9 ± 9.0 years, respectively. For these subgroups, phagocytosis after 30 min incubation time (p30) does agree within the standard error, exhibiting values of 55.19 ± 4.1 and 53.52 ± 3.9%, respectively. The same holds for killing after 120 min (k120) with 57.08 ± 4.0 and 60.63 ± 3.0%, respectively. Hence, age does not influence neutrophil functions significantly. Consequently, for the results to follow the samples of the entire control group is compared with samples from all patients. We note in passing that this result is not in contradiction to Wenisch et al. [19], where reduced phagocytosis was found for healthy donors of high age compared with two groups of lower age. Our samples contain only a few contributors of this high age group. From the fit of the curves in Figure 4 we obtain for the healthy controls 1.81 ± 0.07 for Hp/Np after 60 min incubation time. The corresponding values for patients’ pre- and post-dialysis samples are 1.84 ± 0.09 and 1.84 ± 0.08, respectively, with a correlation of the fit r > 0.92 in all three cases. This means that on average roughly two fungi per active neutrophil are phagocytozed in all samples without statistically significant differences between those of patients and healthy controls. In Figure 5 we contrast phagocytosis of healthy controls with that of dialysis patients. In general pt(healthy) is slightly larger than pt(patients). However, the difference is not statistically significant. Further discrimination revealed the tendency that pre-dialysis phagocytosis is slightly higher than post-dialysis phagocytosis. Separating the data according to patients with and without DM shows that this effect is probably due to patients with DM whose pre-dialysis phagocytosis was clearly higher than their post-dialysis phagocytosis while we could not observe this trend in the patients without diabetes. However, the statistics for these small subgroups is too poor to indicate more than tendencies. Overall, we see from our data that dialysis does not alter phagocytosis significantly. Fig. 5. Open in new tabDownload slide Mean phagocytosis pt of the healthy controls (triangles) and all HD patients (circles) for different incubation times. The SE is indicated, lines are to guide the eye. Fig. 5. Open in new tabDownload slide Mean phagocytosis pt of the healthy controls (triangles) and all HD patients (circles) for different incubation times. The SE is indicated, lines are to guide the eye. The neutrophil killing activities represent the major outcome of this work. Results are presented in steps of increasing discrimination between different groups of patients. Figure 6a shows kt for the group of healthy probands and the group of patients suffering from uraemia. The healthy probands have significantly (P < 0.001) better killing for all incubation times. Fig. 6. Open in new tabDownload slide (a) The mean killing kt of the healthy controls (triangles) and all HD patients (circles) for different incubation times. Both curves differ significantly (P < 0.001) for all measured incubation times t > 0. (b) The mean killing kt of the healthy controls (triangles) and HD patients before (circles) and after (squares) dialysis for different incubation times. Pre-dialysis data differs significantly (P < 0.001) for all measured incubation times t from the control data, post-dialysis data differs significantly (P < 0.05) from pre-dialysis data and the control values at t = 15, 30 min. The SE is indicated, lines are to guide the eye. Fig. 6. Open in new tabDownload slide (a) The mean killing kt of the healthy controls (triangles) and all HD patients (circles) for different incubation times. Both curves differ significantly (P < 0.001) for all measured incubation times t > 0. (b) The mean killing kt of the healthy controls (triangles) and HD patients before (circles) and after (squares) dialysis for different incubation times. Pre-dialysis data differs significantly (P < 0.001) for all measured incubation times t from the control data, post-dialysis data differs significantly (P < 0.05) from pre-dialysis data and the control values at t = 15, 30 min. The SE is indicated, lines are to guide the eye. Figure 6b answers the question, if dialysis improves the killing function. To this end we have split the data of the patients into those before and after dialysis. Pre-dialysis data are significantly lower (P < 0.001) than those from healthy probands for all incubation times. Post-dialysis data differs significantly (P < 0.05) from pre-dialysis data and those of healthy probands for t = 15, 30 min. Hence, dialysis improves the killing. Discriminating further between patients with and without DM we find that dialysis is more effective for patients with DM than for patients suffering not from DM, whereas the latter ones maintain a better killing function before, as well as after dialysis. However, these differences are in general too small to become significant with our data set. Only the killing function k120 for patients with and without DM before dialysis is significantly lower compared with that of the healthy controls. The significance level is P < 0.05 for patients without DM and P < 0.001 for patients with DM. The killing data are summarized in Table 1. Table 1. Killing kt in per cent ± SE at different incubation times in samples from different groups of probandsa Time (min) . 0 . 15 . 30 . 60 . 120 . Healthy 4.0 ± 0.4 15.7 ± 0.9 31.6 ± 1.6 47.1 ± 1.9 59.8 ± 2.2 All patients 3.8 ± 0.3 12.2 ± 0.5 24.3 ± 1.0 39.7 ± 1.5 49.4 ± 1.9 Pre-dialysis 3.3 ± 0.3 10.2 ± 0.7 22.0 ± 1.1 38.7 ± 1.8 46.9 ± 2.3 Post-dialysis 4.4 ± 0.5 12.3 ± 0.9 26.6 ± 1.7 41.0 ± 2.6 53.6 ± 3.3 Pre-dialysis no DM 3.8 ± 0.6 11.2 ± 1.1 23.3 ± 1.6 42.3 ± 2.8 50.7 ± 3.2 Pre-dialysis with DM 3.1 ± 0.3 10.4 ± 1.3 21.4 ± 1.7 38.7 ± 2.9 45.3 ± 3.7 Post-dialysis no DM 3.8 ± 0.6 11.6 ± 1.2 26.6 ± 2.9 42.3 ± 3.9 55.7 ± 5.1 Post-dialysis with DM 5.0 ± 0.7 12.6 ± 1.4 26.4 ± 2.3 39.8 ± 3.8 52.8 ± 4.9 Time (min) . 0 . 15 . 30 . 60 . 120 . Healthy 4.0 ± 0.4 15.7 ± 0.9 31.6 ± 1.6 47.1 ± 1.9 59.8 ± 2.2 All patients 3.8 ± 0.3 12.2 ± 0.5 24.3 ± 1.0 39.7 ± 1.5 49.4 ± 1.9 Pre-dialysis 3.3 ± 0.3 10.2 ± 0.7 22.0 ± 1.1 38.7 ± 1.8 46.9 ± 2.3 Post-dialysis 4.4 ± 0.5 12.3 ± 0.9 26.6 ± 1.7 41.0 ± 2.6 53.6 ± 3.3 Pre-dialysis no DM 3.8 ± 0.6 11.2 ± 1.1 23.3 ± 1.6 42.3 ± 2.8 50.7 ± 3.2 Pre-dialysis with DM 3.1 ± 0.3 10.4 ± 1.3 21.4 ± 1.7 38.7 ± 2.9 45.3 ± 3.7 Post-dialysis no DM 3.8 ± 0.6 11.6 ± 1.2 26.6 ± 2.9 42.3 ± 3.9 55.7 ± 5.1 Post-dialysis with DM 5.0 ± 0.7 12.6 ± 1.4 26.4 ± 2.3 39.8 ± 3.8 52.8 ± 4.9 aHealthy volunteers (n = 117), patients before and after dialysis (n = 100), the subgroup of patients without DM before dialysis (n = 26), patients without DM after dialysis (n = 26), patients with DM before dialysis (n = 24), patients with DM after dialysis (n = 24). Open in new tab Table 1. Killing kt in per cent ± SE at different incubation times in samples from different groups of probandsa Time (min) . 0 . 15 . 30 . 60 . 120 . Healthy 4.0 ± 0.4 15.7 ± 0.9 31.6 ± 1.6 47.1 ± 1.9 59.8 ± 2.2 All patients 3.8 ± 0.3 12.2 ± 0.5 24.3 ± 1.0 39.7 ± 1.5 49.4 ± 1.9 Pre-dialysis 3.3 ± 0.3 10.2 ± 0.7 22.0 ± 1.1 38.7 ± 1.8 46.9 ± 2.3 Post-dialysis 4.4 ± 0.5 12.3 ± 0.9 26.6 ± 1.7 41.0 ± 2.6 53.6 ± 3.3 Pre-dialysis no DM 3.8 ± 0.6 11.2 ± 1.1 23.3 ± 1.6 42.3 ± 2.8 50.7 ± 3.2 Pre-dialysis with DM 3.1 ± 0.3 10.4 ± 1.3 21.4 ± 1.7 38.7 ± 2.9 45.3 ± 3.7 Post-dialysis no DM 3.8 ± 0.6 11.6 ± 1.2 26.6 ± 2.9 42.3 ± 3.9 55.7 ± 5.1 Post-dialysis with DM 5.0 ± 0.7 12.6 ± 1.4 26.4 ± 2.3 39.8 ± 3.8 52.8 ± 4.9 Time (min) . 0 . 15 . 30 . 60 . 120 . Healthy 4.0 ± 0.4 15.7 ± 0.9 31.6 ± 1.6 47.1 ± 1.9 59.8 ± 2.2 All patients 3.8 ± 0.3 12.2 ± 0.5 24.3 ± 1.0 39.7 ± 1.5 49.4 ± 1.9 Pre-dialysis 3.3 ± 0.3 10.2 ± 0.7 22.0 ± 1.1 38.7 ± 1.8 46.9 ± 2.3 Post-dialysis 4.4 ± 0.5 12.3 ± 0.9 26.6 ± 1.7 41.0 ± 2.6 53.6 ± 3.3 Pre-dialysis no DM 3.8 ± 0.6 11.2 ± 1.1 23.3 ± 1.6 42.3 ± 2.8 50.7 ± 3.2 Pre-dialysis with DM 3.1 ± 0.3 10.4 ± 1.3 21.4 ± 1.7 38.7 ± 2.9 45.3 ± 3.7 Post-dialysis no DM 3.8 ± 0.6 11.6 ± 1.2 26.6 ± 2.9 42.3 ± 3.9 55.7 ± 5.1 Post-dialysis with DM 5.0 ± 0.7 12.6 ± 1.4 26.4 ± 2.3 39.8 ± 3.8 52.8 ± 4.9 aHealthy volunteers (n = 117), patients before and after dialysis (n = 100), the subgroup of patients without DM before dialysis (n = 26), patients without DM after dialysis (n = 26), patients with DM before dialysis (n = 24), patients with DM after dialysis (n = 24). Open in new tab Discussion Summarizing the results for phagocytosis, we have observed that pt does not differ significantly between HD patients and healthy controls (see Figure 5). Also, dialysis does not have a significant influence on this behaviour. In the present study the patients’ phagocytic activity of neutrophils after 30 min incubation time was lower than in healthy controls but this difference was not statistically significant (Figure 5). This result is in line with a study from 1995, which stated that in general the phagocyting activity is not impaired by uraemia [4.] This is in disagreement with earlier results reporting a reduced phagocytic activity in uraemic patients although phagocytosis was only measured indirectly in these experiments [10,11]. Concerning the effect of dialysis the authors of [10] pointed out a defect of phagocytosis in HD patients before dialysis, eventually increasing during dialysis dependent on the dialysis membranes used (also see Vanholder et al. [11]). The influence of dialysis on phagocytosis seems to be particularly strong if complement-activating membranes are used [10]. On the other hand recent studies report neither pre- nor post-dialysis phagocytosis being different from healthy controls [12,13] in agreement with our findings. Nevertheless, this is not necessarily a contradiction to [10,11] as in these latter studies mostly bio-compatible membranes have been used. Another possible resolution of the conflict between unaltered or reduced phagocytosis in HD patient could be the time span at the onset of dialysis. Vanholder et al. [2] found a significant defect in neutrophil functions mainly in the first 200 days after onset of dialysis. In our study (and also in that by Gastaldello et al. [12]) the patients have been on dialysis for >200 days. Finally, DM seems to reduce phagocytosis in our patients’ granulocytes with an additional difference before and after dialysis. However, statistics with the present sample size is not good enough to render the differences significant. For the neutrophil killing function we found that HD patients do have a significantly lower antimicrobial killing activity. This is true for patients who do or do not suffer from DM. Dialysis improves the killing slightly (10–15%). Yet, post-dialysis killing still remains significantly below the killing of healthy probands. Our results of Figure 6b indicate that factors that lead to reduced killing can be partially removed by dialysis. Moreover, there is a tendency that dialysis is more effective on DM patients than on those without DM. Indeed, there are candidates with such specific properties, namely glucose-modified proteins. One of the granulocyte inhibitory proteins has been shown to inhibit the glucose uptake and the neutrophil oxidative metabolism [3]. Interestingly, this protein is homologous to β2-microblobulin and showed behaviour identical to β2-microglobulin modified by advanced glycation end products (AGEs) [3]. β2-Microglobulin, however, can be removed effectively by high-flux synthetic dialysis membranes. Hence, β2-microglobulin modified by AGEs or other glucose-modified proteins have exactly the properties which could explain Figure 6b as well as the larger effect of dialysis on DM patients. Our results also clearly show that some factors must exist which cannot be removed by dialysis since the killing capability of neutrophils remains consistently lower for HD patients even after dialysis compared with healthy controls. This is in line with the results by Klein et al. [5], who found that only transplantation but not dialysis can correct altered neutrophil functions of patients on dialysis. Cohen et al. [20] recently found that the plasma level of free Ig-light chains is elevated in uraemic pre-dialysis patients as well as in patients with chronic renal failure. Importantly, the free light chain level could not be reduced to normal values by dialysis [20]. It has been shown previously that free Ig-light chains inhibit glucose uptake of neutrophils. Granulocyte inhibitory protein I, a protein with strong homology to IG-light chains, has been isolated from the haemofiltate of HD patients and it has been shown to diminish the intracellular killing of Staphylococcus aureus by neutrophils [3]. Hence, free Ig-light chains possibly play a role in the reduced antimicrobial killing capability. The normal phagocytic activity in combination with a reduced killing capability of neutrophils in HD patients leads us to the conclusion that the functional impairment of uraemic neutrophils is mainly a result of their reduced capability to kill microorganisms intracellularly. Although we did not find a decreased percentage of ROS active neutrophils in HD patients, it is still possible that each neutrophil produces fewer ROS than in healthy controls as we have not determined the intensity of ROS. Published studies give conflicting information about ROS in HD patients, from reduced over unchanged to enhanced [4,6–9,10,12,13]. Bearing in mind that the antimicrobial efficiency of human neutrophils depends not only on the generation of ROS but also on the release of enzymatic or antimicrobial protein content in the granules it might be useful to conduct further studies along these lines. Further support for the role of non-oxidative mechanisms comes from the postulate that neutrophils of uraemic patients have a relative defect in specific granule functions [21]. Summarizing briefly, we have found significantly lower neutrophil antimicrobial killing in HD patients compared with healthy controls. Patients with DM exhibit an even lower killing. The simultaneously measured phagocytosis remained unchanged. Quite generally, comparing the results regarding neutrophil functions between different studies is difficult due to the diverse methods and the different conditions of the experiments. Nevertheless, it is possible to compare the different studies with respect to the relative trends in the neutrophil functions. With our experiments we also have made an effort towards the possibility of quantitative comparisons in the future. We use a whole-blood method, minimizing external sources of error and we have determined the exact cell to stimulus ratio. As we have shown the neutrophil activity depends quite sensitively on this parameter [17], which should be given and determined accurately to make quantitative comparisons possible. We thank Antje Ruppelt and Sigrid Gäbler for excellent technical assistance. This study was supported by the Deutsche Forschungsgemeinschaft (grant AN 327/2-1). Conflict of interest statement. None declared. References 1 Cohen G, Haag-Weber M, Hörl WH. Immune dysfunction in uremia. Kidney Int 1997 ; 52 [Suppl 62] : S79 –S82 2 Vanholder R, Van Biesen W, Ringoir S. Contributing factors to the inhibition of phagocytosis in hemodialyzed patients. Kidney Int 1993 ; 44 : 208 –214 3 Haag-Weber M, Cohen G, Hörl WH. Clinical significance of granulocyte-inhibiting proteins. Nephrol Dial Transplant 2000 ; 15 [Suppl 1] : 15 –16 4 Ward RA, McLeish KR. Polymorphonuclear leukocyte oxidative burst is enhanced in patients with chronic renal insufficiency. J Am Soc Nephrol 1995 ; 5 : 1697 –1702 5 Klein JB, McLeish KR, Ward RA. Transplantation, not dialysis corrects azotemia-dependent priming of the neutrophil oxidative burst. Am J Kidney Dis 1999 ; 33 : 483 –491 6 Nguyen AT, Lethias C, Zingraff J et al. Hemodialysis membrane-induced activation of phagocyte oxidative metabolism detected in vivo and in vitro within microamounts of whole blood. Kidney Int 1985 ; 28 : 158 –167 7 Hirabayashi Y, Kobayashi T, Nishikawa A et al. Oxidative metabolism and phagocytosis of polymorphonuclear leukocytes in patients with chronical renal failure. Nephron 1988 ; 49 : 305 –312 8 Lewis SL, Van Epps DE. Neutrophil and monocyte alterations in chronic dialysis patients. Am J Kidney Dis 1987 ; 9 : 381 –395 9 Porter CJ, Burden RP, Morgan AG et al. Impaired bacterial killing and hydrogen peroxide production by polymorphonuclear neutrophils in end stage renal failure. Nephron 1997 ; 77 : 479 –481 10 Vanholder R, Ringoir S, Dhondt A, Hakim R. Phagocytosis in uremic and hemodialysis patients: a prospective and cross sectional study. Kidney Int 1991 ; 39 : 320 –327 11 Vanholder R, De Smet R, Jacobs V et al. Uraemic toxic retention solutes depress polymorphonuclear response to phagocytosis. Nephrol Dial Transplant 1994 ; 9 : 1271 –1278 12 Gastaldello K, Husson C, Wens R et al. Role of complement and platelet-activating factor in the stimulation of phagocytosis and reactive oxygen species production during hemodialysis. Nephrol Dial Transplant 2000 ; 15 : 1638 –1646 13 Iida T, Umezawa K, Tanaka K et al. Polymorphonuclear cells in chronic hemodialysis patients have intact phagocytotic and impaired bactericidal activities. Nephron 1997 ; 75 : 41 –47 14 Coli L, Tumietto F, De Pascalis A et al. Effects of dialysis membrane nature on intradilytic phagocytizing activity. J Artif Org 1999 ; 22 : 74 –80 15 Moore MA, Kaplan DS, Picciolo GL et al. Effect of cellulose acetate materials on the oxidative burst of human neutrophils. J Biomed Mater Res 2001 ; 55 : 257 –265 16 Bassoe C-F, Smith I, Sornes S et al. Concurrent measurement of antigen- and antibody-dependent oxidative burst and phagocytosis in monocytes and neutrophils. Methods 2000 ; 21 : 203 –219 17 Anding K, Rost JM, Jacobs E, Daschner FD. Flow cytometric measurements of neutrophil functions: the dependence on the stimulus to cell ratio. FEMS Immunol Med Microbiol 2003 ; 35 : 147 –152 18 Salih HR, Husfeld L, Adam D. Simultaneous cytofluorometric measurment of phagocytosis, burst production and killing of human phagocytes using Candida albicans and Staphylococcus aureus as target organisms. Clin Microbiol Infect 2000 ; 6 : 251 –258 19 Wenisch C, Patruta S, Daxboek F et al. Effect of age on human neutrophil function. J Leukoc Biol 2000 ; 67 : 40 –45 20 Cohen G, Rudnicki M, Schmaldienst S, Hörl WH. Effect of dialysis on serum/plasma levels of free immunoglobulin light chains in end-stage renal disease patients. Nephrol Dial Transplant 2002 ; 17 : 879 –883 21 Deicher R, Exner M, Cohen G et al. Neutrophil β2-microglobulin and lactoferrin content in renal failure patients. Am J Kidney Dis 2000 ; 35 : 1117 –1126 Author notes 1Institute of Medical Microbiology and 2Department of Nephrology, Technical University Dresden, Dresden, 3Max Planck-Institute for the Physics of Complex Systems, Dresden and 4Max Planck-Institute of Infection Biology, Berlin, Germany European Renal Association–European Dialysis and Transplant Association
TI - The influence of uraemia and haemodialysis on neutrophil phagocytosis and antimicrobial killing
JO - Nephrology Dialysis Transplantation
DO - 10.1093/ndt/gfg330
DA - 2003-10-01
UR - https://www.deepdyve.com/lp/oxford-university-press/the-influence-of-uraemia-and-haemodialysis-on-neutrophil-phagocytosis-bHR0RlvUDQ
SP - 2067
EP - 2073
VL - 18
IS - 10
DP - DeepDyve
ER -