Kidney International, Vol. 32 (1987), pp.

8488

Bloodmembrane interaction in hemodialysis leads to increased

cytokine production
ANTON LUGER, JOSEF KOVARIK, HANSKRISTER STUMMVOLL, AGATHA URBANSKA,
and THOMAS A. LUGER

Department of Medicine II and Department of Dermatology II, University of Vienna, Ludwig BolizmannInstitute for Dermatovenerological
Serodiagnosis, Laboratory for Ceilbiology, Vienna, Austria

Bloodmembrane interaction in hemodialysis leads to increased of the acute phase response, promotes catabolism of structural
cytokine production. Recently much interest has been focused on the protein matrix and regulates the febrile response [6]. One
role of immunoregulatory cytokines such as interleukin 1 (IL 1) and
interleukin 2 (IL 2) during the pathogenesis of immunological as well as
important immunological function of IL I is its enhancing effect
inflammatory diseases. Therefore peripheral blood mononuclear cells on interleukin 2 (IL 2) production by T lymphocytes. IL 2,
(PBMC) of eight patients undergoing hemodialysis (HD) were tested for formerly called T cell growth factor, provides an essential signal
IL 1 and IL 2 production. Before starting HD, cytokine production by for the transition of activated T cells from G to the S phase of
PBMC in culture was not altered in comparison to normal healthy the cell cycle [7], but also may activate B lymphocytes, killer
controls, however, a significant increase of IL I and IL 2 production
was observed within the first ND hour which lasted throughout the end cells and enhance the T cell production of other cytokines, such
of ND. Moreover direct effects of cellulose membranes on PBMC as interferon or B cell stimulatory factors (B SF) [810].
cytokine production as well as serum IL I levels have been investi- Since altered production of cytokines such as IL 1 and IL 2
gated. Serum IL I levels were already elevated before onset of HD and could be responsible for many events and laboratory findings in
increased further during HD. The discrepancy between PBMC IL 1
production and serum IL 1 levels may be due to the diminished HD [II], the capacity of mononuclear cells to release these
excretion in patients with endstage renal disease. Since addition of factors during HD was tested.
dialysis membrane particles enhanced monocytes to produce more IL 1
as well as lymphocytes to release more IL 2, a direct stimulatory Methods
membrane effect is postulated. The increased release of immunoreg-
ulatory cytokines may account for some of the pathologic findings Patients
observed during hemodialysis.
Eight patients (pts) with chronic renal failure under intermit-
tent ND (3 X S hr/week) using cellulosic membranes (C-DAK,
Cordis Dow, Miami, Florida, USA; membrane surface 1.8 or
Plasma contact with cellophane membrane has been shown to 2.5 m2) and acetate baths were studied. Data of patients are
result in complement activation and leukostasis in the pulmo- shown in Table 1. Blood was drawn from the arterial site of the
nary vasculature U] leading to the well known granulocytopenia arteriovenous fistula at 0, 30, 60 and 300 minutes. The control
and hypoxemia [21 in the initial phase of dialysis. Furthermore, group consisted of 20 healthy volunteers. Informed consent was
at the same time a significant drop in fibronectin [3] as well as a obtained from all subjects. All data are given as mean SE and
rise in temperature later during HD [4] have been described. In compared by multiple Student's i-test with the Bonferroni
a recent report using cuprophane membranes 15] an increased correction. Differences were considered to be significant when
T4/T8 ratio has been shown to parallel the initial neutropenia, P was less than 0.05.
which was mainly caused by a decrease of suppressor T8 cells,
and in contrast to the transient nature of the other mentioned Materials
phenomena, persisted throughout the duration of hemodialysis Peripheral blood mononuclear cells (PBMC) were prepared
(ND) for six hours. from peripheral blood obtained from healthy individuals or
In the last years growing interest also has been focused on patients at various times during HD by gradient centrifugation
immunoregulatory proteins called lymphokines or cytokines. on Ficoll Paque (Pharmacia, Upsala, Sweden). The cells were
Upon stimulation monocytes/macrophages secrete interleukin) incubated on plastic petri dishes coated with fetal calf serum
(IL 1), a nonspecific immunomodulating factor which activates (FCS, Flow Laboratories, Scotland UK) in RPMI 1640 medium
a variety of cells involved in inflammation, such as lympho- (Grand Island Biological Company, Grand Island, New York,
cytes, granulocytes and fibroblasts, and in vivo acts as mediator USA) supplemented with 10% FCS and antibiotics. The non-
adherent cells were harvested, adjusted to 1 X 106 cells/mi in
RPMI 1640 medium containing 1% FCS. Adherent cells were
Received for publication April 12, 1986 harvested from the plastic petri dishes with the aid of a rubber
and in revised form October 10, 1986 policeman and vigorous pipetting, and adjusted to 1 x 106
1987 by the International Society of Nephrology cells/ml in RPMI 1640 medium containing 1% FCS. Adherent

84
 Cytokines in ESRD 85

Table 1. Clinical characteristics of the patients on chronic HD previously. Briefly, the samples were tested for their ability to
Duration maintain the growth of an IL 2 dependent murine T-cell line
of HD (CTLL). Cells were adjusted to a density of 1 x io cells/mi in
Patient Age Sex Diagnosisa months Therapyb RPMI 1640 containing 5% FCS. A total of 1 x i04 cells per well
1 49 m DN 20 1,2 of a 96 well plate were incubated at 37C in 5% CO2 containing
2 49 m CGN 96 1 atmosphere for 48 hours with various concentrations of sample.
3 58 m PKD 17 1 Cells were pulsed for the final 16 hours with 3H-TdR and
4 45 m CGN 8 1
harvested as described. Thus counts per minute at various
5 43 f DN 25 1,3
dilutions of test samples were compared to an IL 2 standard and
6 20 m CGN 3 I
7 25 f CGN 5 1 transformed into units/mi as described previously [141.
8 42 m DN 24 1,3
a DN, diabetic nephropathy; CGN, chronic glomerulonephritis;
Determination of serum IL I activity by HPLC
PKD, polycystic kidney disease
b , aluminiumhidroxide + water soluble vitamins +
Highpressure liquid chromatography (HPLC) was per-
1,25 dihydro- formed with a system supplied from LKB (Bromma, Sweden)
cholecaiciferol; 2, oral hypoglycemics; 3, insulin using two 2150 pumps, a 2152 system controller, a Rheodyne
7010 injector and a Uvicon 720 LC variable wave length
detector. All manipulations were done at room temperature.
cells consisted of 95% macrophages and were not different in Two hundred tl serum were subjected to a size exclusion
patients and controls. column (BioSil, TSK 125, 300 x 7.5 mm and a BioSil TSK
Guard Column 7.5 x 7.5 mm, Bio Rad Laboratories, Rich-
Preparation and assays of IL 1 mond, California, USA). Elution was carried out with phos-
Adherent cells were cultured in RPM! 1640 with or without phate buffered saline (PBS), pH 7.2 at a flow rate of 0.8
addition of 50 g/ml Silica (Sigma Chemical Company, St. mi/minute [15]. Fractions containing IL 1 activity between
Louis, Missouri, USA) or 100 g/ml lipopolysaccharide (LPS, molecular weight 50 kD and 1 kD [161 were collected as
W.E. Coli, Difco, Detroit, Michigan, USA), In some experi- previously determined. The pooled fractions were filtered ster-
ments dialysis membranes were mechanically broken up and ilized and tested for IL 1 activity as described above. In order
layered on the bottom of culture plates before cells were added. to test for IL 1 inhibiting activity pooled serum fractions were
After incubation for 48 hours at 37C in a 5% CO2 atmosphere additionally tested in the presence of 50 U/mI IL 1 (partial-
the supernatants were harvested and assayed for IL 1 activity. lypurified pooled activity of HPLC gel filtration). In some
The enhancement of mitogen stimulated C3H/He (Institute for experiments a monoclonal antibody (Ak 17) against the biolog-
Laboratory Animals, Himberg, Austria) mouse thymocyte pro- ical active site of IL 1 [17] was tested for its effect on serum IL
liferation was used for the detection of IL 1 activity as previ- 1. Anti-IL 1 IgG, 0.25 jig, usually inhibited 70% of the biological
ously described [121. In brief, thymocytes in suspension culture activities of 0.01 ng recombinant human IL lcs or IL 1/3 [18].
were incubated for three days in microculture plates (1.5 X 10
cells/mi) in the presence of 1 jig/ml Concanavalin A and various Results
dilutions of samples. In order to quantitate the degree of DNA
synthesis, cultures were pulsed with 3H-TdR (0.4 MCi/well) for Peripheral blood mononuclear cells (PBMC) IL 1 production
the final 16 hours of incubation and harvested as described. In IL 1 production by peripheral blood mononuclearcells from
order to have a more reliable index of IL 1 activity in each patients undergoing hemodialysis was evaluated. Culture super-
experiment, an IL 1 standard preparation was titrated along natants of adherent cells from patients before starting HD
with the unknown samples and units of IL 1 activity were contained normal levels of IL 1 activity. During the course of
calculated as described previously [131, using the following HD a significant increase of macrophage IL 1 production was
equation: observed in all patients (Fig. 1). Usually a biphasic pattern of
increased IL 1 production was demonstrated with peak levels
reciprocal titer of test sample at within the first hour and at the end of HD. Addition of Silica,
30% of maximal cpm of standard = which is known to increase IL 1 production, further stimulated
100 x U/mi
reciprocal titer of 30% of maximal adherent cells of patients to produce more IL 1 during HD (Fig.
cpm of standard 1). Since the cultures of adherent cells may be contaminated
with some IL 2 producing lymphocytes, and IL 2 like IL I also
One unit (0.01 ng) recombinant IL 1 (Genzyme, Boston, Mas- enhances the proliferation of thymocytes, the supernatants
sachusetts, USA) was equivalent to 10 U/mi of the IL 1 additionally were tested for their capacity to induce the prolif-
standard preparation. eration of IL 2 dependent CTLL. However, no detectable
levels of IL 2 activity were present in the adherent cell
Preparation and assay for IL 2 supernatants, indicating that the activity measured in the
Nonadherent cells were cultured in RPMI 1640 in the pres- thymocyte assay is IL 1.
ence of 10 g/m1 Concanavalin A (Con A, Calbiochem, La
Jolla, California, USA) and 10 ng/ml phorbol myristate acetate Serum IL 1 activity
(PMA, Sigma, St. Louis, Missouri, USA) at 37C in a 5% CO2 It previously has been shown that serum IL 1 usually elutes
containing atmosphere. After 48 hours incubation supernatants between molecular weight 50 kD and I kD [161. Therefore gel
were harvested and tested for IL 2 activity as described filtration fractions within this molecular weight range were
86 Luger et al

100
1000

-J
50
E

5001I1I
a
Cl)

NCO 0 30 60 300
Time, mm NCO 0 60 300
Fig. 1. Basal (U) and silica stimulated() PBMC IL 1 production in
healthy controls (NCO) and in patients with ESRD before and during
Time, mm

HD (0, 30, 60 and 300 mm). Unstimulated IL I production is signifi- Fig. 2. Serum IL 1 activity in healthy controls (NCO) and hemodialysis
cantly elevated at 30 and 300 minutes (P < 0.05); silica stimulated IL I patients at 0, 60 and 300 minutes. Basal levels are significantly elevated
release only at 300 minutes. in patients (P < 0.05) and further increased during HD (P < 0.05).

Table 2. Inhibition of serurn IL 1 activity by rnonoclonal anti-IL 1

lgG

Patient
Serum IL
U/mi
l 0.5
anti-IL I IgG, jsg/ml
0.25 0.12
600
I 12 88b 74b 70b
2 32 75b 63b 31b
3 122 43b 32b NDb
a i U/rnl recombinant IL 1 is equivalent to 10 U/mI of the IL I
standard preparation used in our laboratory c'I
b % inhibition compared to serum IL I activity of each patient
300

pooled and tested for IL 1 activity. Moreover, a monoclonal

antibody which is known to inhibit the biological activities of
both IL la and IL 1(3 as well as to bind both species of IL 1 [18]
was tested for its effect on serum IL 1. The antibody signifi-
cantly blocked serum IL 1 (50-1 kD) mediated thymocytepro-
lileration in a dose dependent manner (Table 2), indicating that
ririF'
NCO 0 30

Time, mm
11 60 300

serum IL 1 and rIL 1 a and rIL 1/3 share a similar site which is Fig. 3. Basal (U) and C + P stimulated () T-cell IL 2 release in
responsible for the biological activity. healthy control (NCO) and in patients during HD at 0, 30, 60 and 300
In addition to PBMC IL 1 production, serum IL 1 activity of minutes, Basal and stimulated IL 2 release is significantly increased at
30 minutes (P < 0.05).
four patients during the course of HD was tested. A significant
increase of already elevated, basal circulating IL 1 activity was
noticed in all patients (Fig. 2). Elevated serum IL 1 levels were
observed within the first hour after starting HD and lasted of HD (Fig. 3). However, only after 30 minutes the levels were
throughout the end. In addition to IL 1 sera of HD patients significantly different from basal levels. Additional stimulation
usually contained an inhibitor of IL 1 activity which eluted of T-cells in vitro with Con A and PMA resulted in a further
between molecular weight 150 kD and 60 kD (data not shown). significant increase of IL 2 secretion at any stage of HD.
However, a similar IL 1 inhibitor is also found in the sera of However, stimulated IL 2 release was increased significantly
healthy control persons [16]. only 30 minutes after starting HD when compared to pre-HD
levels.
Peripheral T-lymphocytes derived IL 2 production
T-lymphocytes from HD patients were tested for their in vitro In vitro membrane effect on PBMC IL I and IL 2 release
IL 2 production. Basal T-cell IL 2 secretion before initiating HD To evaluate a postulated direct effect of the cellulose mem-
was within the range observed in normal control persons. branes on PBMC lymphokine production, monocytes or lym-
Shortly after connecting patients with endstage renal failure to phocytes obtained from normal healthy volunteers were incu-
the hemodialyzer a significant rise in IL 2 synthesis was bated in the presence of cellulose membrane particles.
observed, and IL 2 remained elevated throughout the duration Monocyte IL 1 production was significantly elevated when
 Cytokines in ESRD 87
A penia and leukostasis may be due to down regulation of C5a
receptor expression [22]. Since Ca which is generated through-
out the hemodialysis also stimulates macrophages expressing
receptors for C5a [231 to release IL 1 in vitro [24], generation of
C5a may further contribute to the increased IL 1 production
400 during hemodialysis. Moreover, increased serum IL 1 levels
during the course of dialysis may account for the normalization
-J of leukopenia in later stages of hemodialysis, since in vivo
200
application of IL 1 has been shown to cause a direct release of
neutrophils from the bone marrow [251.
In addition to the enhanced macrophage IL 1 production,
cultured lymphocytes of patients with renal insufficiency in
vitro released significantly more IL 2 during HD when com-
CO Si LPS Memb CO c + P C Memb pared to the normal basal values. Addition of potent stimulants
Fig. 4. Direct effects of dialysis membrane particles on IL 1 and JL 2 such as Con A and PMA induced a further increase of IL 2
release. Abbreviations are: Co, unstimulated; Si, silica; LPS, lipopoly- release. However, only at 30 minutes was unstimulated and
sacchande; C, conA; P,PMA; Memb, cytokine release in the presence
of membrane particles. Membrane particles lead to a significantly stimulated lymphocyte IL 2 production significantly higher than
increased IL I and IL 2 release (P < 0.05) not different from values before HD. Lymphocyte supernatants obtained at later HD
when commonly used stimuli (that is, Si, LPS, C + P and Cresp.) were stages may contain more IL 2 inhibiting mediators leading to IL
added to the cells in vitro. 2 levels within the normal range. As is the case with IL 1,
contact of T lymphocytes with membrane particles under
experimental conditions stimulated cells to release increased
membranes were added to the cultures (Fig. 4A). Similarly levels of IL 2, suggesting a direct effect of membrane material
lymphocytes incubated with cellulose membrane particles pro- on T-cells. Again a similar effect was observed when cellulose
duced significantly more IL 2 activity in comparison to un- or cuprophane were tested. Moreover IL 1 which is well known
treated cells (Fig. 4B). Therefore contact of mononuclear cells to increase IL 2 production by activated T-lymphocytes [261
to dialysis membranes during HD as well as under experimental may account for additional activation of T-cell IL 2 production.
conditions results in an increased release of lymphokines. In culture when lymphocytes were incubated in the presence of
membrane particles IL 1 may be released from a low percentage
Discussion (5%) of contaminating monocytes. However, IL I has been
The present study documents normal basal IL 1 production shown to stimulate only activated lymphocytes to produce
more IL 2 [26]. Thus membranes either directly stimulate
by cultured monocytes in patients with chronic renal failure as
well as a significant elevation of serum IL 1 levels. IL I is lymphocyte IL 2 production or like mitogen they activate
mainly excreted through the kidneys and thus the increased lymphocytes which subsequently in response to IL I release
serum IL 1 levels together with normal PBMC IL I production more IL 2.
might only reflect the renal impairment. However, onset of Although bacterial endotoxins or complement fragments may
dialysis resulted in a significant increase of monocyte IL 1 cause enhanced cytokine production, these findings demon-
production and serum IL 1 levels with maximal levels after 30 strate that direct interaction of mononuclear cells with cellulose
minutes and five hours when patients usually were discon- membranes during hemodialysis and under experimental con-
nected. Thus monocyte dialysis membrane interactions as well dition leads to an increased release of immunoregulatory medi-
as agents generated during dialysis or bacterial contamination ators. The multifaceted biological effects of these factors [6]
may account for monocyte activation, which results in an may explain some of the pathological findings observed during
increased IL 1 production. Since phagocytosis and adherence hemodialysis such as fever, increased levels of acute phase
of certain particles are potent stimulators of macrophage IL 1 proteins, leucocytosis and PGE2 production. In addition
production [19, 201 the contact of monocytes with dialysis cytokines may be responsible for increased catabolic processes
membranes or uptake of inert silicone particles generated decreased levels of cations, such as iron and zinc and increased
during dialysis [21] may induce IL 1 production. However, fibrosis.
incubation of monocytes from hemodialysis patients in the
presence of silica resulted in a further increase of basal IL 1 Acknowledgment
production, suggesting that the way of monocyte activation We thank Mrs. M. Bednar for her secretarial assistance.
during HD may differ from the in vitro effect of Silica, that is,
phagocytosis and adherence. Furthermore, direct exposure of References
monocytes to cellulose membrane particles stimulated cells to
1. CRADDOCK PR, FEHR J, DALMASSO AP, BRIGHAM KL, JACOB HS:
produce more IL 1, indicating a direct interaction of both Hemodialysis leukopenia. Pulmonary vascular leukostasis resulting
membrane materials with monocytes. Similar findings were from complement activation by dialyzer cellophane membrane. J
obtained when monocytes in culture were stimulated with C/in Invest 59:879888, 1977
cuprophane membranes (data not shown). However cupro- 2. KAPLOW LS, GOFFIENT JA: Profound neutropenia during the early
phane membranes usually are not used in our HD center. phase of hemodialysis. JAMA 203:11351137, 1968
3. SCHWARZ HP, GRAF H, LUGER A, KOVARIK J, STUMMVOLL HK:
Complement activation has been shown to occur early during Dialysis hypoxemia: The role of fibronectin and its pathophysiolog-
hemodialysis. However, the transient nature of granulocyto- ical implication. Contrib Nephrol 37:107110, 1984
88 Luger et a!
4. Mooioit Q, PIZZARELLI F, SiscA S: Blood temperature and
vascular stability during hemodialysis and hemofiltration. Trans
AM Soc Artif intern Organs 28:523527, 1982
5. DRATWA M, COLLART F, MASRCARTLEMONE F, DEVETTER L,
THIELEMANS C, WENS R, DUCHATEAU J: Hemodialysis induced
acute changes in T lymphocytes subsets. (abstract) Eur J Gun
Invest 15:A32, 1985
6. DINARELLO CA: Interleukin-1. Rev Infect Dis 6:5195, 1984
7. STADLER BM, DOUGHERTY SF, FARRAR JJ, OPI'ENI-IEIM JJ: Rela-
tionship of cell cycle to recovery of IL 2 activity from human
mononuclear cells, human and mouse T cell lines. J Immunol
127:19361940, 1981
8. KASAFIARA T, HOOKS JJ, DOUGHERTY SF, OPPENHEIM JJ: Inter-
leukin 2-mediated immune Interferon (IFN-y) production by human
T cells and T cell subsets. J Immunol 130:17841789, 1983
9. DOMZIG W, STADLER BM, HERBERMAN RB: Interleukin 2 depen-
dence of human natural killer (NK) cell activity. J Immunol
130:19701973, 1983
10. ZUBLER RH, LOWENTHAL JW, ERARD F, HASHIMOTO N, DEvos
R, MACDONALD R: Activated B cells express receptors for and
proliferate in response to, pure interleukin 2. J Exp Med 160:
11701183, 1984
11. DINARELLO CA: The biology of interleukin- I and its relevance to
hemodialysis. Blood Purif 1:197224, 1983
12. LUGER TA, STADLER BM, KATZ SI, OPPENHEIM JJ: Epidermal cell
(Keratinocyte) derived thymocyte activating factor (ETAF). J
Immunol 127:14931498, 1980
13. LUGER TA, STADLER BM, LUGER BM, MATHIESON BJ, MAGE M,
SCHMIDT JA, OPPENHEIM JJ: Murine epidermal cellderived
thymocyteactivating factor resembles murine Interleukin 1. J
Immunol 128:21472152, 1982
14. STADLER BM, BERENSTEIN EH, SIRAGANIAN RP, OPPENHEIM JJ:
Monoclonal antibody against Interleukin 2 (IL 2). I. Purification of
IL 2 for the production of monoclonal antibodies. J Immunol
128:16201624, 1982
15. KOCK A, LUGER TA: Purification of human Interleukin 1 by high
performance liquid chromatography. J Chromatogr 296:293300,
1984
16. LUGER A, Git H, SCHWARZ HP, STUMMVOLL HK, LUGER TA:
Decreased serum interleukin 1 activity and monocyte IL 1 produc-
tion in patients with fatal sepsis. Critical Care med 14:458461,
1986
17. KOCK A, DANNER M, STADLER BM, LUGER TA: Characterization
of a monoclonal antibody directed against the biologically active
site of human Interleukin 1. J Exp Med 163:463468, 1986
18. LUGER TA, DANNER M, KOCK A: Monoclonal anti-IL I is directed
against a common site of human IL la and IL lf.3. Immunobiology
172:346356, 1986
19. VOGEL SN, ENGLISH KE, O'BRIEN AD: Silica enhancement of
murine endotoxin sensitivity. Infect Immun 38:681685, 1982
20. GERY I, DAVIES P, DERR J, KRETT N, BARRANGER JA: Relation-
ship between production and release of lymphocyte activating
factor (interleukin 1) by murine macrophages. I. Effects of various
agents. Cell Immunol 64:293303, 1981
21. BERLYNE 0, DUDEK E, ADLER AJ, RUBIN JE, SEIDMAN M: Silican
metabolism: The basic facts in renal failure. Kidney mt 28:175177,
1985
22. SKUBITZ KM, CRADDOCK PR: Reversal of hemodialysis
granulocytopenia and pulmonary leukostasis: A clinical manifesta-
tion of selective downregulation of granulocyte responses to Ca
desarg. J Clin invest 67:13831391, 1981
23. CHENOWETH DE, GOODMAN MG, WEIGLE WO: Demonstration of
a specific receptor for human C5a anaphylotoxin on murine mac-
rophages. J Exp Med 156:6878, 1982
25. GOODMAN MG, CHENOWETH DE, WEIGLE WO: Induction of
interleukin- I secretion and enhancement of humoral immunity by
binding of human C5a to macrophage surface C5a. J Exp Med
156:912917, 1982
25. KAMPSCHMIDT RF, UPCHURCH HF: Neutrophil release after injec-
tions of endotoxin or leukocytic endogenous mediator into rats. J
Reticuloendothel Soc 28:191201, 1980
26. LUGER TA, SMOLEN JS, CHUSED TM, STEINBERG AD, O-
PENHEIM JJ: Human lymphocytes with either OKT4 and OKT8
phenotype produce interleukin 2 in culture. J Clin Invest 70:
470473, 1982