Acta Med Scand 203: 67-70, 1978
The Effect of Hemodialysis on Neutrophil Chemotactic Responsiveness Bengt BjorkstCn,' S. Michael Mauer, Elaine L. Mills and Paul G. Quie From the Departments of Pediatrics, University of Minnesota Medical School, Mayo, Minneapolis, Minnesota, USA
ABSTRACT. After several hemodialysis treatments, neutrophil chemotactic responsiveness is commonly depressed. The median chemotactic index of 34 patients was 21, compared with 47 for 21 controls tested simultaneously. The depressed chemotactic responsiveness was not restored to normal when leukocytes were washed and resuspended in normal plasma, neither did plasma from patients with depressed chemotaxis affect control neutrophils.
Recently we demonstrated that severely depressed neutrophil chemotactic responsiveness may follow repeated hemodialysis treatment ( 5 ) and it was suggested that this phenomenon could, in part, explain the increased susceptibility to serious bacterial infections noted in uremic patients (3,9). We have extended the studies of neutrophil function in patients on hemodialysis to a larger group, and plasma factors and leukocytes from patients have been studied separately in an effort to explain the depressed chemotactic responsiveness. We found that this responsiveness was not restored when the cells were resuspended in control plasma and that plasma from patients with depressed chemotaxis did not affect control neutrophils. These results indicate that hemodialysis-induced, defective chemotaxis is not caused by lack of a necessary plasma factor or by the presence of a plasma factor that inhibits chemotaxis. I Present address: Department of Pediatrics, University of UmeSi, S-90185 UmeSi, Sweden.
STUDY POPULATION AND METHODS Patients and controls. All 34 patients were in renal failure for at least 2 months and had received at least 8
hemodialyses at the University of Minnesota Hospitals. Renal failure was defined as a blood urea nitrogen higher than 40 mg/lOO mi and a serum creatinine higher than 2.0 mg/lOO ml. Forty-seven healthy adult volunteers served as controls. Pooled human serum (PHS) was obtained from three healthy volunteers. The serum was frozen within two hours of drawing and stored at -20°C prior to use. Samples were only frozen and thawed over. The study was approved by the Committee of the Use of Human Subjects in Research at the University of Minnesota. Chemotaxis. The method has been described in detail elsewhere (7).Briefly, leukocytes were separated by gravity sedimentation, counted and diluted with Hank's BSS containing 30 mM HEPES (pH 7.4) to a final concentration of 2 - 4 10" ~ polymorphonuclear leukocytes (PMN) per ml. Cells from 0.4 cm3of this dilution were deposited onto a 5 micron millipore filter. The filters were placed in a modified Boyden chamber with the supernatant of an overnight culture of E. coli or zymosan activated serum (7) as cytotaxins. After 3 hours of incubation, the chemotaxis filters were removed and stained with hematoxylin. The number of PMN that had migrated completely through the filter in ten random high power fields (using a 5 x 5 mm reticule) were counted and divided by the number of cells ( x 10") delivered to the starting side of the filter, this giving the chemotactic index (CI) To study the effect of plasma and leukocytes from the patients separately, leukocytes from patients or controls were washed twice in Hank's BSS and then resuspended in plasma or PHS to a final concentration of 2-4x1O6 PMN/ml. The plasma samples were either tested within six hours of drawing the blood or stored at -20°C prior to use. Hemodialysis technique. Patients were treated with parallel flow cuprophane membrane dialyzers of the Lindea series (Gambro, Lund, Sweden). A Milton-Ray Acta Med Scand 203
B . BjorkstPn et d.
68
Chemotactic
Chemotact i c
Index
Index
>120
? 120
i 100
.
80
120 I
100
.
80
60 8
40
.
I
x
I
20
60
40
: 20
Hemodialysis patients
n
=34
Controls n=21
Pre d ia I y s i s
Postdialysis
Fig. 1. Chemotactic indices of patients undergoing repeated hemodialysis and healthy controls (individual and median values).
Fig. 2. Chemotactic index of neutrophils drawn prior to and after 4 hours’ hemodialysis from 17 patients undergoing repeated hemodialy sis.
dialysis machine was used for mixing of the concentrate with deionized water and for delivery of the dialysate to the dialyzer. Ascorbic acid, 2-4 g, was added to the dialysate concentrate in order to prevent the method- and sulfhemoglobinemia caused by chloramines present in the deionized water (8). Statistical methods. Student’s t-test for paired observations was employed.
to generate chemotactic factors when incubated with zymosan (Table 11). In three experiments we compared the chemotactic responsiveness of patient leukocytes incubated in autologous plasma and in PHS. PHS did not restore the depressed chemotactic responsiveness. The possibility that the dialyzer membranes could either activate plasma inhibitors to chemotaxis, or remove factors necessary for a normal chemotactic responsiveness was further studied by passing normal PHS and serum chelated with 10 mM EDTA, or 10 mM Mg++ EGTA through the dialyzer or by incubating these sera with dialysis
RESULTS
The median of CIS was lower in patients on hemodialysis than in the healthy controls (Fig. I). However, as shown in the figure, there was a great overlap between the two groups and several patients had CIS well within the normal range. An individual hernodialysis treatment usually did not influence hemotactic responsiveness. In fact, there was no significant difference between neutrophils from the same patients prior to, and after 4 hours’ hemodialysis (Fig. 2). Plasma drawn from patients with depressed chemotactic responsiveness before or after four hours of hemodialysis did not significantly depress chemotactic responsiveness of neutrophils from controls (Table I). Plasma samples drawn prior to and after 4 hours of dialysis did not differ significantly from pooled serum with respect to capacity Acfa Med Scand 203
Table I. Effect of plasma from patients with defective chemotaxis on the chemotactic responsiveness of normal polymorphonuclear leukocytes (PMN) Results are shown as mean chemotactic index (CI) of I 1 experiments in which cells were suspended in normal pooled human serum (PHS) or plasma from patients prior to or after 4 hours’ dialysis PMN suspended in PHS Predialysis plasma Postdialysis plasma
CI (mean? S.E.M.) 57L14.5 505 9.9
42510.1
Effect of hemodialysis on chemotaxis
69
dialysis membranes severely impaired random mobility of PMN and they suggested a serum factor to be responsible for this. However, we found that Results are expressed as mean chemotactic index (CI) of plasma from dialysis patients did not significantly normal polymorphonuclear leukocytes in 8 experiments had no effect on the responsiveness of normal using 10%plasma incubated with zymosan as cytotaxin neutrophils to a chemotactic stimulus and that c1 plasma from dialysis patients did not significantly Source of cytotaxin (mean_+S.E.M.) inhibit chemotaxis of normal PMN. Thus our findings do not support the presence of inhibitory Pooled human serum 27f 6.8 factors to chemotaxis in dialysis plasma. Also, we Predialysis plasma 33+ 10.3 Postdialysis plasma 29+ 4.9 were unable to restore chemotactic responsiveness of patient PMN by incubating them in normal serum. This indicates that the low CI in the patients was caused by a cellular defect and not by lack of a membranes in vitro. Normal neutrophils incubated humoral factor. in these sera did not differ from cells incubated in A direct effect of the dialysis membrane on the control serum with regard to chemotactic responPMN is possible since granulocytes obtained by siveness. adherence to nylon fibers show morphological and biochemical evidence of degranulation (10). HowDlSCU SSlON ever, since we found depressed chemotactic reIn an earlier study we demonstrated that patients sponsiveness in samples drawn up to 72 hours after with chronic renal failure not on hemodialysis had the previous hemodialysis, i.e. in PMN that had not mildly depressed neutrophil chemotactic respon- been present in the circulation when the patient was siveness (3, thus confirming the work of Baum dialyzed, a direct effect of dialysis membranes on et al. (1). Further this work demonstrated that the PMN in the circulation is unlikely. During hemodialysis there is an initial neuchemotaxis was markedly depressed after eight or more hemodialysis treatments and that impaired tropenia, due to sequestration of peripheral neuchemotactic responsiveness could develop after trophils in the lung, and a subsequent rebound several dialyses in uremic patients whose initial CIS neutrophilia due to release of these sequestered were normal (5). Salant et al. (11) reported that neutrophils as well as release of neutrophils from patients had abnormal CIS, i.e. less than 20. We the bone marrow (2). Repeated hemodialysis may creased neutrophil chemotaxis of a seventy equal to reduce bone marrow neutrophil reserves and induce or more severe than that of patients on regular morphological changes in granulocytopoietic bone hemodialysis. In this study we confirm, on a larger marrow cells, including cytoplasmatic vacuolizanumber of patients, our previous findings. How- tion and lysis of chromatin (4). Apparently repeated ever, as shown in Fig. l , only about 50% of our hemodialysis interferes with normal granulocyte patients had abnormal CIS, i.e. less than 20. We maturation in the bone marrow. If the depressed cannot explain the individual variations in our pa- chemotactic responsiveness of neutrophils is a contients or the difference between the results of, on sequence of a disturbed maturation, this would exthe one hand, Salant et al. (11) and, on the other, of plain why several hemodialyses were required to inthe present study and that of Baum et al. (1). duce a depressed chemotactic responsiveness and No significant differences in CI were seen in our why the chemotactic index did not vary signifipatients when we compared neutrophils obtained cantly with an individual hemodialysis. Our studies immediately before dialysis with those obtained af- would support the idea of an effect on bone marrow ter 4 hours of hemodialysis treatment. This con- cells rather than the presence of a serum factor firms previously published findings (1) and supports affecting circulating neutrophils. our earlier suggestion that the hemodialysisThe studies show that repeated hemodialysis inassociated depression in CI develops slowly and duces changes in neutrophil locomotion. Further requires several hemodialysis treatments to become studies are needed to precisely characterize the manifest (5). mechanisms for the changes in neutrophil function Recently Henderson et al. (6) reported that and to define the clinical consequences. Table 11. Generation of cytotaxins in plasma from patients with abnormal chemotactic activity prior to and after 4 hours of dialysis
~
Acfa Med Scond 203
70
B . Bjorksten et (11.
REFERENCES I . Baum, J., Cestero, R. V. M. & Freeman, R. B.: Chemotaxis of the polymorphonuclear leucocyte and delayed hypersensitivity in uremia. Kidney Int (Suppl) 2: S 147, 1975. 2. Brubaker, L. & Nolph, K.: Mechanisms of recovery from neutropenia induced by hemodialysis. Blood 30:623, 1971. 3. Bryan, F. A., Jr: National dialysis registry report: 5th Ann. Contractor's Conference of the Artificial Kidney Program of the National Institutes of Arthritis and Metabolic Institute. O.H.E.W. Publication (NIH) 72-749, p. 204, 1971. 4. Buscarini, L.: La reserve de granulocytes medullaires chez des patients soumis a I'hemodialyse periodique. Acta Haematol 56: 7 , 1976. 5. Greene, W. H., Mauer, S. M., Ray, C. R. & Quie, P. G.: The effect of hernodialysis on neutrophil chemotactic responsiveness. J Lab Clin Med 88: 971, 1976. 6 . Henderson, L. W., Miller, M. E., Hamilton, R. W. & Norman, M. E.: Hemodialysis leukopenia and polymorph random mobility-a possible correlation. J Lab Clin Med 85: 191, 1975.
Acta Med Scand 203
7. Hill, H. R., Sauls, H. S . , Dettloff, J. L. & Quie, P. G.: Impaired leukotactic responsiveness in patients with juvenile diabetes mellitus. Clin Immunol Immunopathol2: 395, 1974. 8. Kjellstrand, C. M., Eaton, T. W., Yawata, Y., Swofford, H., Koplin, C. F., Buselmeier, T. J., von Hartitzch, B. & Jacob, H. S.: Hemolysis in dialysed patients caused by chloramines. Nephron 13: 426, 1974. 9. Kleinknecht, D., Jungers, P., Chanard, J., Barbanel, C. & Ganeval, D.: Uremic and non-uremic complication in acute renal failure. Evaluation of early and frequent dialysis on prognosis. Kidney Int 1: 190, 1972. 10. Klock, J . C. & Bainton, D. F.: Degranulation and abnormal bactericidal function of granulocytes produced by reversible adhesion to nylon wool. Blood 48: 149, 1976. 1 1 . Salant, D. J., Glover, A.-M., Anderson, R., Meyers, A. M., Rabkin, R., Myburgh, J . A. & Rabson, A. R.: Depressed neutrophil chemotaxis in patients with chronic renal failure and after renal transplantation. J Lab Clin Med 88: 536, 1976.
The Effect of Hemodialysis on Neutrophil Chemotactic Responsiveness Bengt BjorkstCn,' S. Michael Mauer, Elaine L. Mills and Paul G. Quie From the Departments of Pediatrics, University of Minnesota Medical School, Mayo, Minneapolis, Minnesota, USA
ABSTRACT. After several hemodialysis treatments, neutrophil chemotactic responsiveness is commonly depressed. The median chemotactic index of 34 patients was 21, compared with 47 for 21 controls tested simultaneously. The depressed chemotactic responsiveness was not restored to normal when leukocytes were washed and resuspended in normal plasma, neither did plasma from patients with depressed chemotaxis affect control neutrophils.
Recently we demonstrated that severely depressed neutrophil chemotactic responsiveness may follow repeated hemodialysis treatment ( 5 ) and it was suggested that this phenomenon could, in part, explain the increased susceptibility to serious bacterial infections noted in uremic patients (3,9). We have extended the studies of neutrophil function in patients on hemodialysis to a larger group, and plasma factors and leukocytes from patients have been studied separately in an effort to explain the depressed chemotactic responsiveness. We found that this responsiveness was not restored when the cells were resuspended in control plasma and that plasma from patients with depressed chemotaxis did not affect control neutrophils. These results indicate that hemodialysis-induced, defective chemotaxis is not caused by lack of a necessary plasma factor or by the presence of a plasma factor that inhibits chemotaxis. I Present address: Department of Pediatrics, University of UmeSi, S-90185 UmeSi, Sweden.
STUDY POPULATION AND METHODS Patients and controls. All 34 patients were in renal failure for at least 2 months and had received at least 8
hemodialyses at the University of Minnesota Hospitals. Renal failure was defined as a blood urea nitrogen higher than 40 mg/lOO mi and a serum creatinine higher than 2.0 mg/lOO ml. Forty-seven healthy adult volunteers served as controls. Pooled human serum (PHS) was obtained from three healthy volunteers. The serum was frozen within two hours of drawing and stored at -20°C prior to use. Samples were only frozen and thawed over. The study was approved by the Committee of the Use of Human Subjects in Research at the University of Minnesota. Chemotaxis. The method has been described in detail elsewhere (7).Briefly, leukocytes were separated by gravity sedimentation, counted and diluted with Hank's BSS containing 30 mM HEPES (pH 7.4) to a final concentration of 2 - 4 10" ~ polymorphonuclear leukocytes (PMN) per ml. Cells from 0.4 cm3of this dilution were deposited onto a 5 micron millipore filter. The filters were placed in a modified Boyden chamber with the supernatant of an overnight culture of E. coli or zymosan activated serum (7) as cytotaxins. After 3 hours of incubation, the chemotaxis filters were removed and stained with hematoxylin. The number of PMN that had migrated completely through the filter in ten random high power fields (using a 5 x 5 mm reticule) were counted and divided by the number of cells ( x 10") delivered to the starting side of the filter, this giving the chemotactic index (CI) To study the effect of plasma and leukocytes from the patients separately, leukocytes from patients or controls were washed twice in Hank's BSS and then resuspended in plasma or PHS to a final concentration of 2-4x1O6 PMN/ml. The plasma samples were either tested within six hours of drawing the blood or stored at -20°C prior to use. Hemodialysis technique. Patients were treated with parallel flow cuprophane membrane dialyzers of the Lindea series (Gambro, Lund, Sweden). A Milton-Ray Acta Med Scand 203
B . BjorkstPn et d.
68
Chemotactic
Chemotact i c
Index
Index
>120
? 120
i 100
.
80
120 I
100
.
80
60 8
40
.
I
x
I
20
60
40
: 20
Hemodialysis patients
n
=34
Controls n=21
Pre d ia I y s i s
Postdialysis
Fig. 1. Chemotactic indices of patients undergoing repeated hemodialysis and healthy controls (individual and median values).
Fig. 2. Chemotactic index of neutrophils drawn prior to and after 4 hours’ hemodialysis from 17 patients undergoing repeated hemodialy sis.
dialysis machine was used for mixing of the concentrate with deionized water and for delivery of the dialysate to the dialyzer. Ascorbic acid, 2-4 g, was added to the dialysate concentrate in order to prevent the method- and sulfhemoglobinemia caused by chloramines present in the deionized water (8). Statistical methods. Student’s t-test for paired observations was employed.
to generate chemotactic factors when incubated with zymosan (Table 11). In three experiments we compared the chemotactic responsiveness of patient leukocytes incubated in autologous plasma and in PHS. PHS did not restore the depressed chemotactic responsiveness. The possibility that the dialyzer membranes could either activate plasma inhibitors to chemotaxis, or remove factors necessary for a normal chemotactic responsiveness was further studied by passing normal PHS and serum chelated with 10 mM EDTA, or 10 mM Mg++ EGTA through the dialyzer or by incubating these sera with dialysis
RESULTS
The median of CIS was lower in patients on hemodialysis than in the healthy controls (Fig. I). However, as shown in the figure, there was a great overlap between the two groups and several patients had CIS well within the normal range. An individual hernodialysis treatment usually did not influence hemotactic responsiveness. In fact, there was no significant difference between neutrophils from the same patients prior to, and after 4 hours’ hemodialysis (Fig. 2). Plasma drawn from patients with depressed chemotactic responsiveness before or after four hours of hemodialysis did not significantly depress chemotactic responsiveness of neutrophils from controls (Table I). Plasma samples drawn prior to and after 4 hours of dialysis did not differ significantly from pooled serum with respect to capacity Acfa Med Scand 203
Table I. Effect of plasma from patients with defective chemotaxis on the chemotactic responsiveness of normal polymorphonuclear leukocytes (PMN) Results are shown as mean chemotactic index (CI) of I 1 experiments in which cells were suspended in normal pooled human serum (PHS) or plasma from patients prior to or after 4 hours’ dialysis PMN suspended in PHS Predialysis plasma Postdialysis plasma
CI (mean? S.E.M.) 57L14.5 505 9.9
42510.1
Effect of hemodialysis on chemotaxis
69
dialysis membranes severely impaired random mobility of PMN and they suggested a serum factor to be responsible for this. However, we found that Results are expressed as mean chemotactic index (CI) of plasma from dialysis patients did not significantly normal polymorphonuclear leukocytes in 8 experiments had no effect on the responsiveness of normal using 10%plasma incubated with zymosan as cytotaxin neutrophils to a chemotactic stimulus and that c1 plasma from dialysis patients did not significantly Source of cytotaxin (mean_+S.E.M.) inhibit chemotaxis of normal PMN. Thus our findings do not support the presence of inhibitory Pooled human serum 27f 6.8 factors to chemotaxis in dialysis plasma. Also, we Predialysis plasma 33+ 10.3 Postdialysis plasma 29+ 4.9 were unable to restore chemotactic responsiveness of patient PMN by incubating them in normal serum. This indicates that the low CI in the patients was caused by a cellular defect and not by lack of a membranes in vitro. Normal neutrophils incubated humoral factor. in these sera did not differ from cells incubated in A direct effect of the dialysis membrane on the control serum with regard to chemotactic responPMN is possible since granulocytes obtained by siveness. adherence to nylon fibers show morphological and biochemical evidence of degranulation (10). HowDlSCU SSlON ever, since we found depressed chemotactic reIn an earlier study we demonstrated that patients sponsiveness in samples drawn up to 72 hours after with chronic renal failure not on hemodialysis had the previous hemodialysis, i.e. in PMN that had not mildly depressed neutrophil chemotactic respon- been present in the circulation when the patient was siveness (3, thus confirming the work of Baum dialyzed, a direct effect of dialysis membranes on et al. (1). Further this work demonstrated that the PMN in the circulation is unlikely. During hemodialysis there is an initial neuchemotaxis was markedly depressed after eight or more hemodialysis treatments and that impaired tropenia, due to sequestration of peripheral neuchemotactic responsiveness could develop after trophils in the lung, and a subsequent rebound several dialyses in uremic patients whose initial CIS neutrophilia due to release of these sequestered were normal (5). Salant et al. (11) reported that neutrophils as well as release of neutrophils from patients had abnormal CIS, i.e. less than 20. We the bone marrow (2). Repeated hemodialysis may creased neutrophil chemotaxis of a seventy equal to reduce bone marrow neutrophil reserves and induce or more severe than that of patients on regular morphological changes in granulocytopoietic bone hemodialysis. In this study we confirm, on a larger marrow cells, including cytoplasmatic vacuolizanumber of patients, our previous findings. How- tion and lysis of chromatin (4). Apparently repeated ever, as shown in Fig. l , only about 50% of our hemodialysis interferes with normal granulocyte patients had abnormal CIS, i.e. less than 20. We maturation in the bone marrow. If the depressed cannot explain the individual variations in our pa- chemotactic responsiveness of neutrophils is a contients or the difference between the results of, on sequence of a disturbed maturation, this would exthe one hand, Salant et al. (11) and, on the other, of plain why several hemodialyses were required to inthe present study and that of Baum et al. (1). duce a depressed chemotactic responsiveness and No significant differences in CI were seen in our why the chemotactic index did not vary signifipatients when we compared neutrophils obtained cantly with an individual hemodialysis. Our studies immediately before dialysis with those obtained af- would support the idea of an effect on bone marrow ter 4 hours of hemodialysis treatment. This con- cells rather than the presence of a serum factor firms previously published findings (1) and supports affecting circulating neutrophils. our earlier suggestion that the hemodialysisThe studies show that repeated hemodialysis inassociated depression in CI develops slowly and duces changes in neutrophil locomotion. Further requires several hemodialysis treatments to become studies are needed to precisely characterize the manifest (5). mechanisms for the changes in neutrophil function Recently Henderson et al. (6) reported that and to define the clinical consequences. Table 11. Generation of cytotaxins in plasma from patients with abnormal chemotactic activity prior to and after 4 hours of dialysis
~
Acfa Med Scond 203
70
B . Bjorksten et (11.
REFERENCES I . Baum, J., Cestero, R. V. M. & Freeman, R. B.: Chemotaxis of the polymorphonuclear leucocyte and delayed hypersensitivity in uremia. Kidney Int (Suppl) 2: S 147, 1975. 2. Brubaker, L. & Nolph, K.: Mechanisms of recovery from neutropenia induced by hemodialysis. Blood 30:623, 1971. 3. Bryan, F. A., Jr: National dialysis registry report: 5th Ann. Contractor's Conference of the Artificial Kidney Program of the National Institutes of Arthritis and Metabolic Institute. O.H.E.W. Publication (NIH) 72-749, p. 204, 1971. 4. Buscarini, L.: La reserve de granulocytes medullaires chez des patients soumis a I'hemodialyse periodique. Acta Haematol 56: 7 , 1976. 5. Greene, W. H., Mauer, S. M., Ray, C. R. & Quie, P. G.: The effect of hernodialysis on neutrophil chemotactic responsiveness. J Lab Clin Med 88: 971, 1976. 6 . Henderson, L. W., Miller, M. E., Hamilton, R. W. & Norman, M. E.: Hemodialysis leukopenia and polymorph random mobility-a possible correlation. J Lab Clin Med 85: 191, 1975.
Acta Med Scand 203
7. Hill, H. R., Sauls, H. S . , Dettloff, J. L. & Quie, P. G.: Impaired leukotactic responsiveness in patients with juvenile diabetes mellitus. Clin Immunol Immunopathol2: 395, 1974. 8. Kjellstrand, C. M., Eaton, T. W., Yawata, Y., Swofford, H., Koplin, C. F., Buselmeier, T. J., von Hartitzch, B. & Jacob, H. S.: Hemolysis in dialysed patients caused by chloramines. Nephron 13: 426, 1974. 9. Kleinknecht, D., Jungers, P., Chanard, J., Barbanel, C. & Ganeval, D.: Uremic and non-uremic complication in acute renal failure. Evaluation of early and frequent dialysis on prognosis. Kidney Int 1: 190, 1972. 10. Klock, J . C. & Bainton, D. F.: Degranulation and abnormal bactericidal function of granulocytes produced by reversible adhesion to nylon wool. Blood 48: 149, 1976. 1 1 . Salant, D. J., Glover, A.-M., Anderson, R., Meyers, A. M., Rabkin, R., Myburgh, J . A. & Rabson, A. R.: Depressed neutrophil chemotaxis in patients with chronic renal failure and after renal transplantation. J Lab Clin Med 88: 536, 1976.
