- Email: [email protected]
Intraoperative Course and Prognostic Significance of Endotoxin, Tumor Necrosis Factor-alpha and Interleukin-6 in Liver Transplant Recipients
Immunobiol., vol. 182, pp. 425-439 (1991) I. Surgical University Clinic, University School of Medicine, University of Vienna, Vienna, Austria
Intraoperative Course and Prognostic Significance of Endotoxin, Tumor Necrosis Factor-alpha and Interleukin-6 in Liver Transplant Recipients GERHARD HAMILTON, MARIA PRETTENHOFER, ANNA ZOMMER, SUSANNE HOFBAUER, PETER GOTZINGER, FRANZ XAVER GNANT, and REINHOLD FOGGER Received August 13, 1990 . Accepted in Revised Form February 13, 1991
Abstract Early events in reaction of the host immune system to an allograft were studied by intraoperative measurements of endotoxin (ET), tumor necrosis factor-alpha (TNF-a) and interleukin-6 (IL-6) in human liver transplantation. Blood samples were collected prior to operation, before clamping of the liver vessels, at the beginning and end of the anhepatic phase, and 5 and 30 min after recirculation. Diagnosis of rejection and infection in the graft recipients was established by assessment of spontaneous blastogenesis and T-lymphocyte subpopulations in addition to clinical symptoms and results from biopsies and blood chemistry. Of the 27 unmatched liver transplantations included in this study, 8 patients had infections in the first two postoperative weeks and 6 patients showed rejection of their grafts under antithymocyte globulin immunoprophylaxis. Endotoxin was transiently elevated in plasma in the anhepatic phase (2-fold in comparison to normal values) as expected for an accumulation of intestinederived endotoxin during clamping of liver vessels, but no correlation was found with TNF-a levels and specific post-transplant complications. All patients with rejections had high plasma levels of TNF-a immediately after recirculation (mean value 240 pg TNF-a/ml), in contrast to low TNF-a levels in graft recipients without complications or infections. These results indicate that the initiation of rejection in liver transplantation is associated with increased plasma concentrations of TNF-a. The measured TNF-a concentrations are adequate to promote the binding of lymphocytes to allograft endothelial tissue and/or to induce expression of MHC antigens in the graft. Subsequent viral or bacterial infections were preceded by high intraoperative plasma concentrations of interleukin-6 (mean value 1400 pg lL-6/ml). The correlations of rejection with high intraoperative TNF-a levels and of infection with those of lL-6 are statistically significant in Wilcoxon tests for the direct measurements and in Fisher's exact tests for positive test values, with limits of 90 pg/ml for TNF-a and 800 pg/ml for lL-6.
Introduction
Acute rejections of allografted tissues are the result of a complex cascade of interactions in the host immune system, involving many cell types, their Abbreviations: CMV = cytomegalovirus; ET = endotoxin; LTX = liver transplantation; PMBC = peripheral mononuclear blood cells; SB = spontaneous blastogenesis; TNF = tumor necrosis factor.
426 . G.
HAMILTON
et al.
subpopulations and the action of cytokines (1, 2). The primary immune reaction is initiated by antigen-presenting dendritic cells of the allograft and the MHC class I and class II bearing graft cells are subsequently destroyed by the action of T -helper/inducer and in the later stages also by Tsuppressor/cytotoxic lymphocytes. Graft survival depends on the antigenic disparity of the host and recipient, the degree of host sensitization, the degree of surface expression of MHC antigens in the graft and mode and quantity of immunosuppressive therapy applied. The sequence of steps involved in graft rejection and the differential roles of the various cytokines (interleukins, interferons, tumor necrosis factors) is not yet completely resolved (3). In this study we have investigated early events in the host immune response to allografted tissue in liver transplantation. Due to limited storage time, most of the liver grafts are transplanted at a single center across one or more MHC disparities, resulting in frequent and usually reversible rejections. The intraoperative changes in the plasma concentrations of endotoxin, tumor necrosis factor-a and IL-6 were measured in human liver transplantation and correlated with the immediate postoperative complications. Endotoxin was measured because clamping of the liver vessels in the anhepatic phase has been found to be associated with increased levels of endotoxin in animal models (4) and its function as a potent inducer of other cytokines may contribute to graft rejection (5). TNF-a, which is induced directly by endotoxin and IL-6, which is induced in turn by TNF-a (6), was measured here to investigate the immunological response to endotoxin in LTX.
Patients
27 orthotopic liver transplantations (LTX; 2 retransplantations) in 26 patients carried out at our institution between June 1989 and February 1990 were included in this study. The mean age of the patients (19 males, 7 females) was 46 ± 12 years (range: 21-61 years) and indication for LTX was tumor in 4, hepatitis in 4, toxic liver failure in 4, primary biliary cirrhosis in 2 and alcoholic liver cirrhosis in 9 cases. Of these patients, 19 are still alive with a survival time of 5-11 months after transplantation. With exception of one patient, who was transplanted in hepatic coma, the recipients were compensated at the time of surgery. All recipient-donor pairs were matched for blood group but not for their major histocompatibility antigens (2-3 HLA mismatches) and the donors received blood transfusions prior to transplantation. Immunosuppressive therapy was initiated with methylprednisolone at the end of the operation, followed by an anti-thymocyte globulin (ATG; Bio-Merieux, France) preparation at the second day post-transplant and finally after 8-14 days by cyclosporine/prednisolone maintenance therapy.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation . 427
Diagnosis of infection was based on bacteriological cultures from blood samples, urine, bile or ascites for bacterial infections and on virus isolation cultures from blood, urine and sputum and serology for viral infections. Rejections were diagnosed by clinical signs, increases in blood concentrations of bilirubin, liver enzymes, histological examination of liver biopsies (grade I-IV) and demonstration of donor-specific mixed lymphocyte reactivity, which correlates with increased spontaneous blastogenesis and concomitant increases in T -helper/inducer lymphocytes. During transplantation, blood samples (3-5 ml) were drawn into heparinized tubes at the following time points: (A) immediately before operation, (B) before clamping of liver vessels, (C) at the begin of the anhepatic phase, (D) at the end of the anhepatic phase, (E) 5 min after recirculation and (F) 30 min after recirculation. Blood was collected from A. radialis (A) and from V. portae (B-F). Duration of the anhepatic phase was approximately 75 min, ranging from 45-160 min. The blood samples were stored at 4°C, centrifuged and the plasma fraction used in all determinations.
Materials and Methods Measurement of endotoxin Endotoxin was determined using a commercial Limulus amoebocyte lysate test system (Coatest Endotoxin; Kabi Diagnostica, Nykoepping, Sweden; E. coli 0111 :B4 endotoxin standard) and a sample pretreatment method involving ultrafiltration and phenolic extraction according to BERGER et al. with modifications (7). In short, 0.5 ml plasma (stored frozen at -20°C) was centrifuged in Centrisart I (Sartorius, G6ttingen, Germany) molecular weight separation tubes (molecular weight cutoff: 20 kD; centrifugation for 30 min at 2000xg). The fraction > 20 kD was made up to a volume of 1 ml with sterile, pyrogen-free water and proteins precipitated by addition of 2 ml water-saturated phenol (10 min, 68°C). After cooling to 4°e, the aqueous phase was collected and the phenolic phase again extracted with 2 ml of water. The aqueous phases were pooled, centrifuged at 1500xg for 30 min and directly assayed in the LAL test in appropriate dilutions (1:200 to 1:400; test range: 0.1-1.2 EU/ml; 1 endotoxin unit equivalent to 12 ng endotoxin/ml) without further purification. Determinations were done in duplicate in micro titer plates and false positive samples excluded from analysis. Cleavage of the S2423 substrate was assayed at OD415 nm and the endotoxin concentrations of the samples were calculated using the standard samples of the Coatest kit. Measurement of tumor necrosis factor-alpha
TNF-a was determined with an immunoradiometric assay kit consisting of precoated capture and tracer monoclonal antibodies (IRMA, Medgenix diagnostics; Brussels, Belgium). The test was carried out according to the manufacturers instructions using plasma samples in duplicate (200 [,II each). The detection limit of this method is 5 pg TNF-a/ml and the test range was 5-5000 pg/ml. Normal values were 5 ± 3.5 pg TNF-a/ml in healthy controls (mean ± SD; n=16). Measurement of interleukin-6
IL-6 was determined with an ELISA-kit (Quantikine IL-6; Research and Diagnostics System, Minneapolis, MN, USA) with a precoated monoclonal catcher antibody and an
428 . G. HAMILTON et al. enzyme-linked polyclonal antibody preparation for detection of bound IL-6. The detection limit is 5 pg IL-6/ml and the test range 5-2000 pg/ml. No IL-6 could be detected in plasma samples from healthy control persons.
Measurement of spontaneous proliferation and subpopulations of peripheral mononuclear blood cells (PMBC) For the isolation of PMBC, heparinized blood samples (5 ml) were diluted with phosphatebuffered saline (3.5 ml) and layered on top of 4.5 ml Ficoll-Paque (Pharmacia, Uppsala, Sweden) density gradient centrifugation medium. Samples were centrifuged at 200 x g for 30 min and the interphase cells collected and washed with RPMI-1640/Hepes medium supplemented with 10 % fetal bovine serum (Flow, Irvine, Scotland), 4 mM glutamine and 75 [,Igi ml gentamicin. PMBCs were counted and suspended at a cell concentration of 8 x 106 Iml in medium. For the determination of unstimulated thymidine uptake 100 [,II aliquots of the purified PMBC were transferred in duplicate to the wells of microtiter plates and labelled for 18 h with 37 kBq 3H-thymidine (New England Nuclear, MA, USA) under tissue culture conditions. At the end of the incubation the cells were collected with a cell harvester (Skatron, Lier, Norway) and incorporated radioactivity determined by beta-counting. Normal values were 970 ± 410 cpm unstimulated thymidine uptake for PMBC from 19 normal healthy subjects and 2150 ± 1230 cpm unstimulated thymidine uptake for PMBC from 64 liver transplant recipients with stable graft function 3-84 months after transplantation. Increased unstimulated thymidine uptakes (> 8000 cpm) were found to be associated with clinical complications (infection or rejection). For the determination of lymphocyte subpopulations, 50 [,II aliquots of the PMBC suspensions in medium were labelled with appropriate FITC- or PE-conjugated monoclonal antibodies and after washing analyzed using a FACS-analyzer (Becton-Dickinson, Mountain View, CA, USA). Data from 5000 PMBCs were accumulated for each analysis. The percentage of lymphocyte subpopulations was measured in relation to gated lymphocytes (volume-scatter gate). Increases of > 10 % T-helper/inducer lymphocytes within 3 days in the presence of increased SB were found to be associated with immunological activationslrejections.
Mixed lymphocyte cultures Purified PMBC were resuspended in RPMI-1640 medium supplemented with 10 % human AB serum (Biochrome, Berlin, Germany), 4 mM glutamine and 75 [,Ig/ml gentamicin and added to donor spleen cells (4 or 2 x 10 5 PMBC and 5 x 104 spleen cells/well). Donor cells were isolated from spleen by removal of the capsule, mincing of the tissue and pressing of the slices through a stainless steel mesh (RPMI-1640/10 % h-AB serum, 10 [,Ig/ml DNAse type IV, Sigma, St. Louis, MO, USA). Erythrocytes were removed by hypotonic lysis and the irradiated spleen cells (4000 Rad) were frozen in aliquots (medium supplemented with 20 % human AB serum and 10 % dimethyl sulfoxide) and stored at -70°C. For controls spleen cells or PMBC from patients were cultured alone. The flat-bottomed microtiter plates were incubated for 4 days and after labelling for further 18 h with tritiated thymidine (37 kBq/well) the cells were harvested and incorporated radioactivity counted. The MLC index was calculated as quotient of the thymidine incorporation for the MLC culture and the thymidine incorporation in the PMBC control. MLC indices of > 2 were found only in patients with clinical and histological signs of rejections.
Statistical analysis The prognostic significance of the different parameters for complications following LTX was evaluated using Fisher's 2-tailed exact test for positive test results with thresholds of 90 pgl ml for TNF-a and 800 pg/ml for IL-6 respectively and the Wilcoxon test for the absolute values of endotoxin and the cytokines. Differences in mean cytokine plasma concentrations between different LTX groups were analyzed by t-tests.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation . 429
Results
Diagnosis of complications in graft recipients Immunologic diagnosis of complications in these liver graft recipients was achieved by daily assessment of unstimulated thymidine uptake (or spontaneous blastogenesis; SB) into isolated PMBCs (Fig. 1). Normal values of SB in healthy individuals were < 1500 cpm/8 x 105 cells/18 h (n = 19). Liver graft recipients without complications, as shown by absence of adverse clinical signs, normal bilirubin and normal liver-specific enzyme plasma concentrations, showed a minor increase at day 6 after transplantation. Infections as well as rejections were found to be associated with high rates of SB significantly elevated between days 5-9 after operation (p = 0.05 in comparison to the group without complications). In the rejection group the SB dropped in response to treatment with high doses of steroids. It should be noted that infection here includes all types of viral and bacterial infections. Differential diagnosis of rejection versus infection Daily measurements of SB and flow cytometric analysis of T cell subpopulations were carried out for the blood s\lmples from liver graft reci-
S B
•. -.-. -e-
-o
.---e
*
1\
INFECT.]IO~ //*
l~--*
.-- -- ~------.
M
,.. 8 ><
:IE 0..
U
I
I
I
\
I
*
·~~::~,l /.-.-r//I'\.L.J-. 1
5
10
day s
Figure 1. Mean unstimulated thymidine uptake (spontaneous blastogenesis; counts per minute x 103/8 x 105 PMBC/18 h; ± SEM) into peripheral mononuclear blood cells (PMBC) for liver graft recipients without complications (_._._; n = 13) and either infection (---; n = 8) or rejection (-; n = 6) for the first ten postoperative days (SEM < 30 % for all mean values; the mean values for the thymidine uptakes for days 5 and 6 for patients with rejections and for days 5-9 for patients with infections are statistically significantly different from those of the patients without complications; p < 0.05).
430 . G. HAMILTON et al.
pients and the proliferative activity (SB) and the percentage of T -helper/ inducer cells among lymphocytes were compared with the course of donorspecific MLC activity. The results of these tests are shown for a patient without complications (Fig. 2; LTX 130), for a patient with irreversible rejection (Fig. 2; LTX 118) and a patient with infection (Fig. 2; LTX 126;
30
LTX 130
40
20
4 15
2
.----.--;_.---_.--- ..- ---.- --
c:::
o
... ...
CO
30
o
0.
o 40 trc CJ
(,)
c:::
Q)
c:::
"E >....c
><
L TX 118
~
20
I
15
I
.... .....
" ""
.....
I
I
I
1\
\
\
4 \
c:::
\
I
2 (,)
..J
"
L TX 126
~40
4
><
E CJ
"0
30
M
0.
Q)
20
1S.~/
~.
\
\
2
\ \
\
\
1
2
3
4
5
6
7
8
day 5 Figure 2. Course of unstimulated thymidine uptake (SD < 15 %; -), percentage of T-helper/ inducer lymphocytes (SD < 11 %; vertical bars) and donor-specific mixed lymphocyte culture activity (SD < 15 %; ---; MLC index: cpm thymidine incorporation in MLC cultures/cpm thymidine incorporation in unstimulated controls) for a liver transplant recipient without complications (L TX 130), for a recipient with rejection (grade II-III irreversible rejection; rejection classification for biopsies ranging from O-IV; LTX 118) and for a recipient with infection (cholangitis; LTX 126).
B
15 15 nd g 4 13 4 10.2 ±6
A
10 21 20 40 11 4 17.7 ±13
15.1 ±29
±
9.4 8.4
13.4 ±21
48 170 214 14 15 5 78"h ±91
26 14 270 nd 13 5 66 ± 115
8.5 ± 5.9
D
C
20.7 ±23
64 14 202 40 71 9 67',h ± 71
E
TNF-a (pg/ml)' Intraoperative time pointb
41.6 ±56.6
± 110
293 270 412 196 125 130 238,,·,,·h
F d7/I d6/I-II d9/I-II d91lI-III d81II d81lI
Rejection day/gradeC
d5 d4 d7 d7 d6 d7
Increase T-helperd
d6 d6 d8 d7 d8 d8
Increase SBc
d7/4.5 d8/2.7 d6/3.6 nd g d8/3.2 d8/2.1
MLC-day/ indexf
SD < 12 % for duplicate determinations. b A: immediately before operation, B: before clamping of the liver vessels, C: at the begin of the anhepatic phase, D: at the end of the anhepatic phase, E: 5 min after recirculation, and F: 30 min after recirculation. histological grading of liver biopsies (I-IV). d increase in SB > 8000 cpm within 2 successive days. increase in relative percentage (lymphocytes) of T-helperlinducer lymphocytes> 10 % within 3 successive days. MLC index: thymidine incorporation in donor-specific MLC/thymidine incorporation in controls. g nd = not determined. h "p < 0.05; ".". P < 0.001.
mean SD
LTX without complications (n= 13)
118 140 141 145 146 149 mean SD
LTX
Table 1. Intraoperative plasma concentrations of TNF-a for the individual patients with rejections versus the group of patients without complications
tTl
'""
...
l:l
~. o
~
'"0
~
~
..,'"
:;::.
t-<
0"-
~
l:l 0-
l:l"
'".,.,
.,;n
~
p. '"
8
o
l:l 0-
432 . G. HAMILTON et al. Table 2. Intraoperative plasma concentrations of interleukin-6 for the individual patients with infections versus the group of patients without complications Interleukin-6 (pg/ml)' Intraoperative time point b
LTX
B
C
D
62 73 2050 350 675 93 55 725 510 ±680
522 150 1695 2850 2037 489 112 557 1052 ± 1011
LTX without complications (n= 13) mean 56.4 76 23.4 SD ± 71 ±30 ±72
211 ±270
A 82 126 127 132 133 138 142 150 mean SD
a
b
C
0 65 10 28 0 43 0 0 18.2 ±25
0 76 150 nd 12 75 37 79 61 ± 51
E
Infection day/org.
F
1165 1785 240 890 1720 1790 2805 995 2265 1595 480 825 1650 390 1775 975 1255*c 1413"c ±934 ±430
207 ±280
9CMV 7 Serratia m. 5 Klebsiella pn. 9 Acinetobact. an. 14 CMV, HSV 12 Serratia m. 6 Staph. epid. 12 CMV
404 ±360
SD < 13 % for duplicate determinations. A: immediately before operation, B: before clamping of the liver vessels, C: at the begin of the anhepatic phase, D: at the end of the anhepatic phase, E: 5 min after recirculation, and F: 30 min after recirculation. "= p < 0.002.
cholangitis). Rejection is characterized by parallel increases in SB, percentage of CD4+ positive T-Iymphocytes and MLC activity, whereas in infection the increase in SB occurs in the presence of constant or decreasing levels of helper/inducer T cells and low MLC activity. T -suppressor! inducer lymphocytes were constant in both cases « 10 % in LTX 118 and < 5 % in LTX 126). These findings were confirmed in a series of 34 liver graft recipients, in which biopsy-proven rejections and positive MLCactivities were found to correlate with increases in SB (> 8000 cpm within 2 days) and parallel increases in T -helper!inducer lymphocytes (increase > 10 % of the lymphocytes within 3 days; manuscript submitted). For the present study these tests have been used to detect the immunological response to infection and to assign the patients with grade I-rejections to the rejection group (patient 118) or, in the absence of proliferative activity / T -helper!inducer increases and donor-specific MLC-reactivity, to the group of patients without complications (2 patients). No simultaneous increases in SB and T -helper!inducer lymphocytes and reactivity in donorspecific MLC cultures (MLC index> 2.0) were detected in the group of patients without complications.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 433
Intraoperative determination of endotoxin, TNF-a and IL-6 The individual determinations for TNF-a and lL-6 at different time points during liver transplantation are shown in Table 1 for the patients with rejections and in Table 2 for the patients with infections respectively. The mean values of determinations for the group of patients without complications are shown for comparison. The histological and immunological criteria for the diagnosis of rejections are included in Table 1. One patient (llS) with rejection grade I showed concomitant increases in SB and percentage of T -helper/inducer lymphocytes and high donor-specific MLC-reactivity and was included in the rejection group. Two other patients with grade I -rejections exhibited no donor-specific MLC-reactivity and no increases in SB and T -helper/inducer subpopulation. They received no anti-rejection therapy and were assigned to the group of patients without complications. During recirculation, TNF-a was significantly increased for the patients with rejections and lL-6 for the patient group with infections. Mean normal endotoxin value (± SD) for a group of ambulant patients without complications was S.S ± 4 EU. In the course of liver transplantation a transient increase of endotoxin over normal values was observed in the anhepatic phase (Fig. 3). Due to large individual variations, no significant differences were found between the symptom-free group and the liver recipient groups with infections and rejections respectively. The ET plasma concentrations showed no correlation with TNF-a levels measured 30 min after recirculation (correlation coefficient: r2=0.01). For TNF-a, small increases were found in the recipient groups without complications (F; 30 min after recirculation; 42 pg/ml) and infections (F: 60 pg/ml) and a very large increase in the rejection group exceeding 200 pg TNF-a/ml plasma (Fig. 3). Although the values for TNF-a were higher for the rejection group throughout the operation, the difference to the other recipient groups is statistically significant only for the blood samples drawn after recirculation (D-F). TNF-a values for the first postoperative days showed no correlation with infection or rejection (data not shown). The course of IL-6 plasma concentrations during liver transplantation exhibited an increase to 410 pg/ml and 630 pg/ml for the patients without complications and with rejection respectively, whereas the patients, who subsequently developed infections had significantly higher plasma lL-6 levels for the recirculation period.
Prognostic significance of intraoperative endotoxin, TNF-a and IL-6 Correlation of the different parameters with complications after LTX was calculated using the test according to Wilcoxon for all values from A to E and the two-tailed Fisher's exact test for positive and negative test results. For the latter statistical test limits were set to 20 EU endotoxin, 90 pg/ml TNF-a and SOO pg/ml for lL-6. Significant correlations were detected for rejection and TNF-a/30 min after recirculation (p < 0.01) and for infection
434 . G. HAMILTON et al.
.-.-. -e40
INFECTION
0--'"
ET
- - - REJECTION
--t ;:) 20 w
200
TNF
--t
~100
1000
/r--r
I L-6
/ I
*
ABO
E
/
~500
/--r
*
CI
a.
b 10 od
F
sample
Figure 3. Mean values (± SEM) of intraoperative determinations of endotoxin (ET; in endotoxin units/ml), tumor necrosis factor-a (TNF) and interleukin-6 (IL-6) in liver transplant recipients without postoperative complications (_._._; n= 13), and either infection (---; n=8) or rejection (-; n=6). Blood samples were taken before operation (A), before clamping of the liver vessels (B), at the begin and end of the anhepatic phase (C, D) and 5 and 30 min after recirculation (E, F). Statistically significant differences are indicated by an asterisk.
and IL-6/5 min after recirculation (p < 0.05) and IL-6/30 min after recirculation (p < 0.01) in the Wilcoxon test and also for rejection and positive TNF-a values (p < 0.001) and for infection and positive IL-6 values (p < 0.02) in Fisher's exact test. Discussion
Knowledge of the sequence of immunological signals resulting in rejection of solid allografts is not complete. Interleukin-1, IFN-y and TNF-a have been described as early mediators of T -lymphocyte triggering, inducing activation and growth of T - and B-lymphocytes and ultimately inflammation and graft destruction (1, 2). Increased TNF-a levels (cut off 100 pgl ml; range 50-800 pg/ml) have been observed in kidney transplantation
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 435
during allograft rejection (8) and recently human liver graft and experimental cardiac allograft rejections were found to be associated with increased plasma levels of TNF-a (mean values 940 pg/ml; 37 rejections) at the time of the rejection episode (9, 10). Purified TNF-a was reported to accelerate graft rejection in the experimental rat model and antibodies to TNF-a or TNF-~ either alone or in combination, prolonged cardiac graft survival in the experimental animal model without any other immunosuppression, supporting the important role of TNF in allograft rejection (10). In an attempt to investigate the first reactions of the host immune system to solid organ transplants, we have studied the intraoperative release of TNF-a in human liver transplantation. Transplantation of this organ is different from grafting of other organs in several aspects: (a) donor and recipients in most case have one or more HLA mismatches, resulting in a high frequency of usually reversible rejections, (b) clamping of the portal vein without veno-venous bypass, as performed at our center, leads to a release of intestine-derived endotoxin, which is a potent inducer of cytokines and (c) endotoxin accumulates during the anhepatic phase, since it is not eliminated by Kupffer cells of the liver. The role of endotoxin in TNF-a production and induction of rejection in LTX has not been documented. Furthermore, IL-6, which is produced in response to TNF-a and triggers the production of acute phase proteins and possibly also the proliferation of B lymphocytes in vivo, was also determined in this study (11-13). The disadvantage of the liver transplantation model is the high frequency of complications caused by primary organ malfunction, viral and bacterial infections, drug side effects and recurrent disease and the difficulties to establish a correct differential diagnosis of these complications. As shown here, the determination of spontaneous blastogenesis in peripheral blood of the graft recipients in combination with measurements of the T -helper! inducer lymphocyte subpopulation and donor-specific MLC activity can be used to confirm the histological diagnosis of rejection. Determination of cytokines or soluble cytokine receptors and neopterine are alternative nonspecific methods to detect activation of the immune system in graft recipients (14, 15). The different complications can be distinguished by correlation of the unstimulated proliferative activity with changes in the T -helper! inducer lymphocyte subpopulations as proved by measurements of the donor-specific MLC activity. For grade I rejections (lymphocyte infiltration), the histological grading of the biopsies alone correlated poorly with the further clinical course of these patients. In some patients the lymphocytic infiltrate resulted in graft rejection, whereas in other patients the infiltration resolved without rejection therapy. Although the histological grading of liver biopsies is the golden standard of diagnosis of rejection in liver recipients, in the present study the histological diagnosis of grade I rejection was regarded as true rejection if confirmed tw demonstration of donor-specific MLC reactivity and demonstration of proliferative activity/ T -helper!inducer increases. By means of these immunological tests in combination with clinical signs, blood chemistry, biopsies and serological
436 . G. HAMILTON et al.
and bacteriological tests the liver graft recipients were assigned to groups without complications, infections and rejections with high reliability. The immunological tests revealed infection in cases in which the infectious agents could not be isolated or identified despite fever and response to therapies. Endotoxin, tumor necrosis factor-a and interleukin-6 were measured from blood samples collected prior to, and at different time points during, human liver transplantation (n = 27). For determination of ET, plasma components, which could interfere with the Limulus test were removed by ultrafiltration/phenol treatment (7), and TNF-a and IL-6 were assayed with highly sensitive commercial test kits. Endotoxin was elevated moderately in the plasma of liver graft recipients at the end of the anhepatic phase, but there was no correlation to subsequent infections or rejections. Individual differences in cold ischemia time of the graft, duration of the anhepatic phase and other factors might be responsible for differences in ET concentrations. These results are in contrast to a previous study of endotoxemia in human liver transplantation, which reported endotoxin to represent a prognostic factor for LTX (16). However, this study differs from the present work by the inclusion of patients with primary non-function of their graft (10 %), who had high endotoxin blood levels, and the use of a test system which includes interfering blood substances (7). A highly significant increase in TNF-a (over 200 pg/ml plasma) was detected for graft recipients, who subsequently rejected their allografts, in comparison to recipients without complications and infections. Although not significant, the patients in the group with rejections tended to have elevated TNF-a levels prior to allograft exposure. These preoperatively elevated TNF-a levels are not the result of a common underlying disease since these patients had various indications for LTX. In agreement with other work (9) we have not detected elevated TNF-a levels in samples from patients with hepatitis immediately prior to LTX. Our results show no correlation of endotoxin plasma concentrations at the end of the anhepatic phase with TNF-a levels after recirculation in human LTX, supporting the hypothesis that TNF-a is induced by another mechanism, possibly via interleukin-2 (17), in the early phase of reaction to the allograft. The induction of TNF-a immediately after contact with the allograft can contribute to graft rejection by promoting the binding of lymphocytes to graft endothelial tissue (18) or induction of MHC alloantigens in the graft (19). In addition to release by stimulated monocytes in bacterial infections (20), TNF-a is also produced by lymphocytes (21) and participates directly in T cell activation and T cell proliferation (22) and also in the final steps of tissue injury and graft destruction (23). The intraoperative analysis of TNF-a plasma concentrations allowed us to measure the release of TNF-a in response to allograft exposure prior to the initiation of the immunosuppressive therapy, which alters production and release of cytokines. Immunosuppressive therapy for the recipients was started at the end of the operation with steroids and immunoprophylaxis
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 437
for the first post-transplant time was provided by ATG treatment beginning at the second post-transplant day. The TNF-a plasma concentrations under this immunosuppressive regimen showed no correlation with subsequent complications since release of TNF-a is suppressed by steroids (24) and induced by ATG- or OKT-3-mediated lysis of T-Iymphocytes independent of allograft rejection (25, 26). Interleukin-6 has been characterized as B-Iymphocyte differentiation factor, T cell costimulatory factor and as inducer of the acute phase response in hepatocytes (11-13). The intraoperative course of IL-6 was found to correlate with viral and bacterial infectious complications in the liver recipients. The mechanism of IL-6 induction in this patient group is not clear, but the increase became obvious at the beginning of anhepatic phase, indicating that infections may exist prior to transplantation or occur during the anhepatic phase of the liver transplantation. The finding that the production of IL-6 is induced by viruses may help to explain the high levels of IL-6 observed in the liver graft recipients in this study, since these patients show a high incidence of cytomegalovirus (CMV) and other viral infections (27). In summary, early rejections of liver grafts by a group of recipients under ATG immunoprophylaxis have been found to correlate significantly with elevated intraoperative plasma concentrations of TNF-a immediately after allograft exposure. In addition to its direct participation in triggering of T -lymphocytes in the process of allograft rejection, TNF-a may enhance the immunogenicity of the allograft, most likely by induction of MHC antigens and promoting of leukocyte adherence, resulting in a high frequency of subsequent rejections. The results also demonstrate that endotoxin accumulation during the anhepatic phase of liver transplantation is of minor importance for the induction of TNF-a and allograft rejection. Intraoperative elevations in the plasma concentrations of IL-6 were found to be a prognostic factor for early infections in the liver graft recipients.
References 1. SIBLEY, R. K. 1989. Pathology and immunopathology of solid organ graft rejection. Transplant. Proc. 21: 14. 2. TILNEY, N. L. and J. W. KUPIEC-WEGLINSKI. 1989. Advances in the understanding of rejection mechanisms. Transplant. Proc. 21: 10. 3. HALLORAN, P. F., S. M. COCKFIELD, and J. M. ADRENAS. 1989. The mediators of inflammation (IL-1, IFN-y and TNF) and their relevance to rejection. Transplant. Proc. 21: 26. 4. FILIPPONI, F., J. VAN DE STADT, P. IcARD, A. MICHEL, and D. HOUSSIN. 1988. Prevention of death following one-hour occlusion of the portal vein in the rat. Eur. Surg. Res. 20: 39. 5. BEUTLER, B. and A. CERAMI. 1987. Cachectin: more than a tumor necrosis factor. N. Eng!. J. Med. 316: 379. 6. BROUCKAERT, P., D. R. SPRIGGS, G. DEMETRI, D. W. KUFE, and W. FIERS. 1989. Circulating interleukin 6 during continuous infusion of tumor necrosis factor and interferon. J. Exp. Med. 169: 2257.
438 . G. HAMILTON et al. 7. BERGER, D., E. MARZINZIG, M. MARZINZIG, and H. G. BEGER. Quantitative endotoxin determination in blood - chromogenic modification of the Limulus amoebocyte lysate test. Eur. Surg. Res. 20: 128. 8. MAURY, C. P. J. and A. M. TEPPO. 1987. Raised serum levels of cachectin/tumor necrosis factor in renal allograft rejection. J. Exp. Med. 166: 1132. 9. IMAGAWA, D. K., J. M. MILLIS, K. M. OLTHOFF, L. J. DERUS, D. CHIA, L. R. SUGICH, M. OZAWA, R. A. DEMPSEY, Y. IWAKI, P. J. LEVY, P. I. TERASAKI, and R. W. BUSUTTIL. 1990. The role of tumor necrosis factor in allograft rejection. Transplantation 50: 219. 10. IMAGAWA, D. K., J. M. MILLIS, K. M. OLTHOFF, P. SEU, R. A. DEMPSEY, J. HART, P. I. TERASAKI, E. M. W ASEF, and R. W. BUSUTTIL. 1990. The role of tumor necrosis factor in allograft rejection. Transplantation 50: 189. 11. HELFGOTT, D. c., S. B. TATTER, U. SANTHANAM, R. H. CLARICK, N. BHARDWAJ, L. T. MAY, and P. B. SEHGAL. 1989. Multiple forms of interferon-B2/IL-6 in plasma and body fluids during acute bacterial infection. J. Immunol. 142: 948. 12. HEINRICH, P. c., J. V. CASTELL, and T. ANDUS. 1990. Interleukin-6 and the acute phase response. Biochem. J. 265: 621. 13. WAAGE, A., T. ESPEVIK, and A. HALSTENSEN. 1988. Tumor necrosis factor, interleukin-1 and interleukin-6 in human septic shock. Scand. J. Immunol. 28: 267. 14. PERKINS, J. D., D. L. NELSON, J. RAKE LA, P. M. GRAMBSCH, and R. A. F. KROM. 1989. Soluble IL-2 receptor level as an indicator of liver allograft rejection. Transplantation 47: 77. 15. TILG, H., W. VOGEL, W. E. AULITZKY, D. SCHOENITZER, R. MARGREITER, O. DIETZE, G. JUDMAIER, H. WACHTER, and C. HUBER. 1989. Neopterin excretion after liver transplantation and its value in differential diagnosis of complications. Transplantation 48: 594. 16. YOKOYAMA, I., S. TODO, T. MIYATA, R. SELBY, A. G. TZAKIS, and T. E. STARZL. 1989. Endotoxemia and human liver transplantation. Transplant. Proc. 21: 3833. 17. ECONOMOU, J. S., W. H. McBRIDE, and R. ESSNER. 1989. Tumor necrosis factor production by interleukin-2 activated macrophages in vitro and in vivo. Immunology 67:
514. 18. CAVENDER, D., Y. SAEGUSA, and M. ZIFF. 1987. Stimulation of endothelial cell binding of lymphocytes by tumor necrosis factor. J. Immunol. 139: 1855. 19. COLLINS, T., L. A. LAPIERRE, W. FIERS, J. L. STROMINGER, and J. S. POBER. 1986. Recombinant tumor necrosis factor increases mRNA-levels and surface expression of HLA-A,B antigens in vascular endothelial cells and dermal fibroblasts in vitro. Proc. Nat!. Acad. Sci. USA 83: 446. 20. DEBETS,J. M. H., R. KAMPMEIJER, M. P. M. H. VANDERLINDEN, W. A. BUURMAN, and C. F. VAN DER LINDEN. 1989. Plasma tumor necrosis factor and mortality in critically ill septic patients. Crit. Care Med. 17: 489. 21. SUNG, S. S. J., J. M. BJORNDAHL, C. Y. WANG, H. T. KAo, and S. M. Fu. 1988. Production of tumor necrosis factor/cachectin by human T cell lines and peripheral blood T lymphocytes stimulated by phorbol miristate acetate and anti-CD3 antibody. J. Exp. Med. 167: 937. 22. SCHEURICH, P., B. THOMA, U. UCER, and K. PFIZENMAIER. 1987. Immunregulatory activity of recombinant tumor necrosis factor (TNF)-alpha: induction of TNF receptors of human T cells and TNF-alpha mediated enhancement of T cell responses. J. Immunol. 138: 1786. 23. DECKER, T., M.-L. LOHMANN-MATTHES, and G. E. GIFFORD. 1987. Cell-associated tumor necrosis factor (TNF) as a killing mechanism of activated cytotoxic macrophages. J. Immunol. 138: 957. 24. DUPONT, E., L. SCHANDENE, C. DENYS, A. CRUSIAUX, and J. WYBRAN. 1989. Assessment of production of tumor necrosis factor-alpha under the influence of immunosuppressive drugs. Transplant. Proc. 21: 70. 25. DEBETS, J. M. H., K. M. L. LEUNISSEN, H. J. VAN HOOFF, C. J. VAN DER LINDEN, and W. A. BUURMAN. 1989. Evidence of involvement of tumor necrosis factor in adverse reactions during treatment of kidney allograft rejection with antithymocyte globulin. Transplantation 47: 487.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 439 26. ABRAMOWICZ, D., L. SCHANDENE, M. GOLDMAN, A. CRUSIAUX, P. VEREESTRAETEN, L. DE PAUW, J. WYBRAN, P. KINNAERT, E. DUPONT, and C. TOUSSAINT. 1989. Release of tumor necrosis factor, interleukin-2 and gamma-interferon in serum after injection of OKT3 monoclonal antibody in kidney transplant recipients. Transplantation 47: 606. 27. HAMILTON, G., T. POPOV, F. MUEHLBACHER, and F. PIZA. 1989. Early detection of viral infection in liver graft recipients by immune monitoring. Transplant. Proc. 21: 2266. Dr. GERHARD HAMILTON, 1. Surgical University Clinic, Alserstr. 4, A-I090 Vienna, Austria
Intraoperative Course and Prognostic Significance of Endotoxin, Tumor Necrosis Factor-alpha and Interleukin-6 in Liver Transplant Recipients GERHARD HAMILTON, MARIA PRETTENHOFER, ANNA ZOMMER, SUSANNE HOFBAUER, PETER GOTZINGER, FRANZ XAVER GNANT, and REINHOLD FOGGER Received August 13, 1990 . Accepted in Revised Form February 13, 1991
Abstract Early events in reaction of the host immune system to an allograft were studied by intraoperative measurements of endotoxin (ET), tumor necrosis factor-alpha (TNF-a) and interleukin-6 (IL-6) in human liver transplantation. Blood samples were collected prior to operation, before clamping of the liver vessels, at the beginning and end of the anhepatic phase, and 5 and 30 min after recirculation. Diagnosis of rejection and infection in the graft recipients was established by assessment of spontaneous blastogenesis and T-lymphocyte subpopulations in addition to clinical symptoms and results from biopsies and blood chemistry. Of the 27 unmatched liver transplantations included in this study, 8 patients had infections in the first two postoperative weeks and 6 patients showed rejection of their grafts under antithymocyte globulin immunoprophylaxis. Endotoxin was transiently elevated in plasma in the anhepatic phase (2-fold in comparison to normal values) as expected for an accumulation of intestinederived endotoxin during clamping of liver vessels, but no correlation was found with TNF-a levels and specific post-transplant complications. All patients with rejections had high plasma levels of TNF-a immediately after recirculation (mean value 240 pg TNF-a/ml), in contrast to low TNF-a levels in graft recipients without complications or infections. These results indicate that the initiation of rejection in liver transplantation is associated with increased plasma concentrations of TNF-a. The measured TNF-a concentrations are adequate to promote the binding of lymphocytes to allograft endothelial tissue and/or to induce expression of MHC antigens in the graft. Subsequent viral or bacterial infections were preceded by high intraoperative plasma concentrations of interleukin-6 (mean value 1400 pg lL-6/ml). The correlations of rejection with high intraoperative TNF-a levels and of infection with those of lL-6 are statistically significant in Wilcoxon tests for the direct measurements and in Fisher's exact tests for positive test values, with limits of 90 pg/ml for TNF-a and 800 pg/ml for lL-6.
Introduction
Acute rejections of allografted tissues are the result of a complex cascade of interactions in the host immune system, involving many cell types, their Abbreviations: CMV = cytomegalovirus; ET = endotoxin; LTX = liver transplantation; PMBC = peripheral mononuclear blood cells; SB = spontaneous blastogenesis; TNF = tumor necrosis factor.
426 . G.
HAMILTON
et al.
subpopulations and the action of cytokines (1, 2). The primary immune reaction is initiated by antigen-presenting dendritic cells of the allograft and the MHC class I and class II bearing graft cells are subsequently destroyed by the action of T -helper/inducer and in the later stages also by Tsuppressor/cytotoxic lymphocytes. Graft survival depends on the antigenic disparity of the host and recipient, the degree of host sensitization, the degree of surface expression of MHC antigens in the graft and mode and quantity of immunosuppressive therapy applied. The sequence of steps involved in graft rejection and the differential roles of the various cytokines (interleukins, interferons, tumor necrosis factors) is not yet completely resolved (3). In this study we have investigated early events in the host immune response to allografted tissue in liver transplantation. Due to limited storage time, most of the liver grafts are transplanted at a single center across one or more MHC disparities, resulting in frequent and usually reversible rejections. The intraoperative changes in the plasma concentrations of endotoxin, tumor necrosis factor-a and IL-6 were measured in human liver transplantation and correlated with the immediate postoperative complications. Endotoxin was measured because clamping of the liver vessels in the anhepatic phase has been found to be associated with increased levels of endotoxin in animal models (4) and its function as a potent inducer of other cytokines may contribute to graft rejection (5). TNF-a, which is induced directly by endotoxin and IL-6, which is induced in turn by TNF-a (6), was measured here to investigate the immunological response to endotoxin in LTX.
Patients
27 orthotopic liver transplantations (LTX; 2 retransplantations) in 26 patients carried out at our institution between June 1989 and February 1990 were included in this study. The mean age of the patients (19 males, 7 females) was 46 ± 12 years (range: 21-61 years) and indication for LTX was tumor in 4, hepatitis in 4, toxic liver failure in 4, primary biliary cirrhosis in 2 and alcoholic liver cirrhosis in 9 cases. Of these patients, 19 are still alive with a survival time of 5-11 months after transplantation. With exception of one patient, who was transplanted in hepatic coma, the recipients were compensated at the time of surgery. All recipient-donor pairs were matched for blood group but not for their major histocompatibility antigens (2-3 HLA mismatches) and the donors received blood transfusions prior to transplantation. Immunosuppressive therapy was initiated with methylprednisolone at the end of the operation, followed by an anti-thymocyte globulin (ATG; Bio-Merieux, France) preparation at the second day post-transplant and finally after 8-14 days by cyclosporine/prednisolone maintenance therapy.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation . 427
Diagnosis of infection was based on bacteriological cultures from blood samples, urine, bile or ascites for bacterial infections and on virus isolation cultures from blood, urine and sputum and serology for viral infections. Rejections were diagnosed by clinical signs, increases in blood concentrations of bilirubin, liver enzymes, histological examination of liver biopsies (grade I-IV) and demonstration of donor-specific mixed lymphocyte reactivity, which correlates with increased spontaneous blastogenesis and concomitant increases in T -helper/inducer lymphocytes. During transplantation, blood samples (3-5 ml) were drawn into heparinized tubes at the following time points: (A) immediately before operation, (B) before clamping of liver vessels, (C) at the begin of the anhepatic phase, (D) at the end of the anhepatic phase, (E) 5 min after recirculation and (F) 30 min after recirculation. Blood was collected from A. radialis (A) and from V. portae (B-F). Duration of the anhepatic phase was approximately 75 min, ranging from 45-160 min. The blood samples were stored at 4°C, centrifuged and the plasma fraction used in all determinations.
Materials and Methods Measurement of endotoxin Endotoxin was determined using a commercial Limulus amoebocyte lysate test system (Coatest Endotoxin; Kabi Diagnostica, Nykoepping, Sweden; E. coli 0111 :B4 endotoxin standard) and a sample pretreatment method involving ultrafiltration and phenolic extraction according to BERGER et al. with modifications (7). In short, 0.5 ml plasma (stored frozen at -20°C) was centrifuged in Centrisart I (Sartorius, G6ttingen, Germany) molecular weight separation tubes (molecular weight cutoff: 20 kD; centrifugation for 30 min at 2000xg). The fraction > 20 kD was made up to a volume of 1 ml with sterile, pyrogen-free water and proteins precipitated by addition of 2 ml water-saturated phenol (10 min, 68°C). After cooling to 4°e, the aqueous phase was collected and the phenolic phase again extracted with 2 ml of water. The aqueous phases were pooled, centrifuged at 1500xg for 30 min and directly assayed in the LAL test in appropriate dilutions (1:200 to 1:400; test range: 0.1-1.2 EU/ml; 1 endotoxin unit equivalent to 12 ng endotoxin/ml) without further purification. Determinations were done in duplicate in micro titer plates and false positive samples excluded from analysis. Cleavage of the S2423 substrate was assayed at OD415 nm and the endotoxin concentrations of the samples were calculated using the standard samples of the Coatest kit. Measurement of tumor necrosis factor-alpha
TNF-a was determined with an immunoradiometric assay kit consisting of precoated capture and tracer monoclonal antibodies (IRMA, Medgenix diagnostics; Brussels, Belgium). The test was carried out according to the manufacturers instructions using plasma samples in duplicate (200 [,II each). The detection limit of this method is 5 pg TNF-a/ml and the test range was 5-5000 pg/ml. Normal values were 5 ± 3.5 pg TNF-a/ml in healthy controls (mean ± SD; n=16). Measurement of interleukin-6
IL-6 was determined with an ELISA-kit (Quantikine IL-6; Research and Diagnostics System, Minneapolis, MN, USA) with a precoated monoclonal catcher antibody and an
428 . G. HAMILTON et al. enzyme-linked polyclonal antibody preparation for detection of bound IL-6. The detection limit is 5 pg IL-6/ml and the test range 5-2000 pg/ml. No IL-6 could be detected in plasma samples from healthy control persons.
Measurement of spontaneous proliferation and subpopulations of peripheral mononuclear blood cells (PMBC) For the isolation of PMBC, heparinized blood samples (5 ml) were diluted with phosphatebuffered saline (3.5 ml) and layered on top of 4.5 ml Ficoll-Paque (Pharmacia, Uppsala, Sweden) density gradient centrifugation medium. Samples were centrifuged at 200 x g for 30 min and the interphase cells collected and washed with RPMI-1640/Hepes medium supplemented with 10 % fetal bovine serum (Flow, Irvine, Scotland), 4 mM glutamine and 75 [,Igi ml gentamicin. PMBCs were counted and suspended at a cell concentration of 8 x 106 Iml in medium. For the determination of unstimulated thymidine uptake 100 [,II aliquots of the purified PMBC were transferred in duplicate to the wells of microtiter plates and labelled for 18 h with 37 kBq 3H-thymidine (New England Nuclear, MA, USA) under tissue culture conditions. At the end of the incubation the cells were collected with a cell harvester (Skatron, Lier, Norway) and incorporated radioactivity determined by beta-counting. Normal values were 970 ± 410 cpm unstimulated thymidine uptake for PMBC from 19 normal healthy subjects and 2150 ± 1230 cpm unstimulated thymidine uptake for PMBC from 64 liver transplant recipients with stable graft function 3-84 months after transplantation. Increased unstimulated thymidine uptakes (> 8000 cpm) were found to be associated with clinical complications (infection or rejection). For the determination of lymphocyte subpopulations, 50 [,II aliquots of the PMBC suspensions in medium were labelled with appropriate FITC- or PE-conjugated monoclonal antibodies and after washing analyzed using a FACS-analyzer (Becton-Dickinson, Mountain View, CA, USA). Data from 5000 PMBCs were accumulated for each analysis. The percentage of lymphocyte subpopulations was measured in relation to gated lymphocytes (volume-scatter gate). Increases of > 10 % T-helper/inducer lymphocytes within 3 days in the presence of increased SB were found to be associated with immunological activationslrejections.
Mixed lymphocyte cultures Purified PMBC were resuspended in RPMI-1640 medium supplemented with 10 % human AB serum (Biochrome, Berlin, Germany), 4 mM glutamine and 75 [,Ig/ml gentamicin and added to donor spleen cells (4 or 2 x 10 5 PMBC and 5 x 104 spleen cells/well). Donor cells were isolated from spleen by removal of the capsule, mincing of the tissue and pressing of the slices through a stainless steel mesh (RPMI-1640/10 % h-AB serum, 10 [,Ig/ml DNAse type IV, Sigma, St. Louis, MO, USA). Erythrocytes were removed by hypotonic lysis and the irradiated spleen cells (4000 Rad) were frozen in aliquots (medium supplemented with 20 % human AB serum and 10 % dimethyl sulfoxide) and stored at -70°C. For controls spleen cells or PMBC from patients were cultured alone. The flat-bottomed microtiter plates were incubated for 4 days and after labelling for further 18 h with tritiated thymidine (37 kBq/well) the cells were harvested and incorporated radioactivity counted. The MLC index was calculated as quotient of the thymidine incorporation for the MLC culture and the thymidine incorporation in the PMBC control. MLC indices of > 2 were found only in patients with clinical and histological signs of rejections.
Statistical analysis The prognostic significance of the different parameters for complications following LTX was evaluated using Fisher's 2-tailed exact test for positive test results with thresholds of 90 pgl ml for TNF-a and 800 pg/ml for IL-6 respectively and the Wilcoxon test for the absolute values of endotoxin and the cytokines. Differences in mean cytokine plasma concentrations between different LTX groups were analyzed by t-tests.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation . 429
Results
Diagnosis of complications in graft recipients Immunologic diagnosis of complications in these liver graft recipients was achieved by daily assessment of unstimulated thymidine uptake (or spontaneous blastogenesis; SB) into isolated PMBCs (Fig. 1). Normal values of SB in healthy individuals were < 1500 cpm/8 x 105 cells/18 h (n = 19). Liver graft recipients without complications, as shown by absence of adverse clinical signs, normal bilirubin and normal liver-specific enzyme plasma concentrations, showed a minor increase at day 6 after transplantation. Infections as well as rejections were found to be associated with high rates of SB significantly elevated between days 5-9 after operation (p = 0.05 in comparison to the group without complications). In the rejection group the SB dropped in response to treatment with high doses of steroids. It should be noted that infection here includes all types of viral and bacterial infections. Differential diagnosis of rejection versus infection Daily measurements of SB and flow cytometric analysis of T cell subpopulations were carried out for the blood s\lmples from liver graft reci-
S B
•. -.-. -e-
-o
.---e
*
1\
INFECT.]IO~ //*
l~--*
.-- -- ~------.
M
,.. 8 ><
:IE 0..
U
I
I
I
\
I
*
·~~::~,l /.-.-r//I'\.L.J-. 1
5
10
day s
Figure 1. Mean unstimulated thymidine uptake (spontaneous blastogenesis; counts per minute x 103/8 x 105 PMBC/18 h; ± SEM) into peripheral mononuclear blood cells (PMBC) for liver graft recipients without complications (_._._; n = 13) and either infection (---; n = 8) or rejection (-; n = 6) for the first ten postoperative days (SEM < 30 % for all mean values; the mean values for the thymidine uptakes for days 5 and 6 for patients with rejections and for days 5-9 for patients with infections are statistically significantly different from those of the patients without complications; p < 0.05).
430 . G. HAMILTON et al.
pients and the proliferative activity (SB) and the percentage of T -helper/ inducer cells among lymphocytes were compared with the course of donorspecific MLC activity. The results of these tests are shown for a patient without complications (Fig. 2; LTX 130), for a patient with irreversible rejection (Fig. 2; LTX 118) and a patient with infection (Fig. 2; LTX 126;
30
LTX 130
40
20
4 15
2
.----.--;_.---_.--- ..- ---.- --
c:::
o
... ...
CO
30
o
0.
o 40 trc CJ
(,)
c:::
Q)
c:::
"E >....c
><
L TX 118
~
20
I
15
I
.... .....
" ""
.....
I
I
I
1\
\
\
4 \
c:::
\
I
2 (,)
..J
"
L TX 126
~40
4
><
E CJ
"0
30
M
0.
Q)
20
1S.~/
~.
\
\
2
\ \
\
\
1
2
3
4
5
6
7
8
day 5 Figure 2. Course of unstimulated thymidine uptake (SD < 15 %; -), percentage of T-helper/ inducer lymphocytes (SD < 11 %; vertical bars) and donor-specific mixed lymphocyte culture activity (SD < 15 %; ---; MLC index: cpm thymidine incorporation in MLC cultures/cpm thymidine incorporation in unstimulated controls) for a liver transplant recipient without complications (L TX 130), for a recipient with rejection (grade II-III irreversible rejection; rejection classification for biopsies ranging from O-IV; LTX 118) and for a recipient with infection (cholangitis; LTX 126).
B
15 15 nd g 4 13 4 10.2 ±6
A
10 21 20 40 11 4 17.7 ±13
15.1 ±29
±
9.4 8.4
13.4 ±21
48 170 214 14 15 5 78"h ±91
26 14 270 nd 13 5 66 ± 115
8.5 ± 5.9
D
C
20.7 ±23
64 14 202 40 71 9 67',h ± 71
E
TNF-a (pg/ml)' Intraoperative time pointb
41.6 ±56.6
± 110
293 270 412 196 125 130 238,,·,,·h
F d7/I d6/I-II d9/I-II d91lI-III d81II d81lI
Rejection day/gradeC
d5 d4 d7 d7 d6 d7
Increase T-helperd
d6 d6 d8 d7 d8 d8
Increase SBc
d7/4.5 d8/2.7 d6/3.6 nd g d8/3.2 d8/2.1
MLC-day/ indexf
SD < 12 % for duplicate determinations. b A: immediately before operation, B: before clamping of the liver vessels, C: at the begin of the anhepatic phase, D: at the end of the anhepatic phase, E: 5 min after recirculation, and F: 30 min after recirculation. histological grading of liver biopsies (I-IV). d increase in SB > 8000 cpm within 2 successive days. increase in relative percentage (lymphocytes) of T-helperlinducer lymphocytes> 10 % within 3 successive days. MLC index: thymidine incorporation in donor-specific MLC/thymidine incorporation in controls. g nd = not determined. h "p < 0.05; ".". P < 0.001.
mean SD
LTX without complications (n= 13)
118 140 141 145 146 149 mean SD
LTX
Table 1. Intraoperative plasma concentrations of TNF-a for the individual patients with rejections versus the group of patients without complications
tTl
'""
...
l:l
~. o
~
'"0
~
~
..,'"
:;::.
t-<
0"-
~
l:l 0-
l:l"
'".,.,
.,;n
~
p. '"
8
o
l:l 0-
432 . G. HAMILTON et al. Table 2. Intraoperative plasma concentrations of interleukin-6 for the individual patients with infections versus the group of patients without complications Interleukin-6 (pg/ml)' Intraoperative time point b
LTX
B
C
D
62 73 2050 350 675 93 55 725 510 ±680
522 150 1695 2850 2037 489 112 557 1052 ± 1011
LTX without complications (n= 13) mean 56.4 76 23.4 SD ± 71 ±30 ±72
211 ±270
A 82 126 127 132 133 138 142 150 mean SD
a
b
C
0 65 10 28 0 43 0 0 18.2 ±25
0 76 150 nd 12 75 37 79 61 ± 51
E
Infection day/org.
F
1165 1785 240 890 1720 1790 2805 995 2265 1595 480 825 1650 390 1775 975 1255*c 1413"c ±934 ±430
207 ±280
9CMV 7 Serratia m. 5 Klebsiella pn. 9 Acinetobact. an. 14 CMV, HSV 12 Serratia m. 6 Staph. epid. 12 CMV
404 ±360
SD < 13 % for duplicate determinations. A: immediately before operation, B: before clamping of the liver vessels, C: at the begin of the anhepatic phase, D: at the end of the anhepatic phase, E: 5 min after recirculation, and F: 30 min after recirculation. "= p < 0.002.
cholangitis). Rejection is characterized by parallel increases in SB, percentage of CD4+ positive T-Iymphocytes and MLC activity, whereas in infection the increase in SB occurs in the presence of constant or decreasing levels of helper/inducer T cells and low MLC activity. T -suppressor! inducer lymphocytes were constant in both cases « 10 % in LTX 118 and < 5 % in LTX 126). These findings were confirmed in a series of 34 liver graft recipients, in which biopsy-proven rejections and positive MLCactivities were found to correlate with increases in SB (> 8000 cpm within 2 days) and parallel increases in T -helper!inducer lymphocytes (increase > 10 % of the lymphocytes within 3 days; manuscript submitted). For the present study these tests have been used to detect the immunological response to infection and to assign the patients with grade I-rejections to the rejection group (patient 118) or, in the absence of proliferative activity / T -helper!inducer increases and donor-specific MLC-reactivity, to the group of patients without complications (2 patients). No simultaneous increases in SB and T -helper!inducer lymphocytes and reactivity in donorspecific MLC cultures (MLC index> 2.0) were detected in the group of patients without complications.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 433
Intraoperative determination of endotoxin, TNF-a and IL-6 The individual determinations for TNF-a and lL-6 at different time points during liver transplantation are shown in Table 1 for the patients with rejections and in Table 2 for the patients with infections respectively. The mean values of determinations for the group of patients without complications are shown for comparison. The histological and immunological criteria for the diagnosis of rejections are included in Table 1. One patient (llS) with rejection grade I showed concomitant increases in SB and percentage of T -helper/inducer lymphocytes and high donor-specific MLC-reactivity and was included in the rejection group. Two other patients with grade I -rejections exhibited no donor-specific MLC-reactivity and no increases in SB and T -helper/inducer subpopulation. They received no anti-rejection therapy and were assigned to the group of patients without complications. During recirculation, TNF-a was significantly increased for the patients with rejections and lL-6 for the patient group with infections. Mean normal endotoxin value (± SD) for a group of ambulant patients without complications was S.S ± 4 EU. In the course of liver transplantation a transient increase of endotoxin over normal values was observed in the anhepatic phase (Fig. 3). Due to large individual variations, no significant differences were found between the symptom-free group and the liver recipient groups with infections and rejections respectively. The ET plasma concentrations showed no correlation with TNF-a levels measured 30 min after recirculation (correlation coefficient: r2=0.01). For TNF-a, small increases were found in the recipient groups without complications (F; 30 min after recirculation; 42 pg/ml) and infections (F: 60 pg/ml) and a very large increase in the rejection group exceeding 200 pg TNF-a/ml plasma (Fig. 3). Although the values for TNF-a were higher for the rejection group throughout the operation, the difference to the other recipient groups is statistically significant only for the blood samples drawn after recirculation (D-F). TNF-a values for the first postoperative days showed no correlation with infection or rejection (data not shown). The course of IL-6 plasma concentrations during liver transplantation exhibited an increase to 410 pg/ml and 630 pg/ml for the patients without complications and with rejection respectively, whereas the patients, who subsequently developed infections had significantly higher plasma lL-6 levels for the recirculation period.
Prognostic significance of intraoperative endotoxin, TNF-a and IL-6 Correlation of the different parameters with complications after LTX was calculated using the test according to Wilcoxon for all values from A to E and the two-tailed Fisher's exact test for positive and negative test results. For the latter statistical test limits were set to 20 EU endotoxin, 90 pg/ml TNF-a and SOO pg/ml for lL-6. Significant correlations were detected for rejection and TNF-a/30 min after recirculation (p < 0.01) and for infection
434 . G. HAMILTON et al.
.-.-. -e40
INFECTION
0--'"
ET
- - - REJECTION
--t ;:) 20 w
200
TNF
--t
~100
1000
/r--r
I L-6
/ I
*
ABO
E
/
~500
/--r
*
CI
a.
b 10 od
F
sample
Figure 3. Mean values (± SEM) of intraoperative determinations of endotoxin (ET; in endotoxin units/ml), tumor necrosis factor-a (TNF) and interleukin-6 (IL-6) in liver transplant recipients without postoperative complications (_._._; n= 13), and either infection (---; n=8) or rejection (-; n=6). Blood samples were taken before operation (A), before clamping of the liver vessels (B), at the begin and end of the anhepatic phase (C, D) and 5 and 30 min after recirculation (E, F). Statistically significant differences are indicated by an asterisk.
and IL-6/5 min after recirculation (p < 0.05) and IL-6/30 min after recirculation (p < 0.01) in the Wilcoxon test and also for rejection and positive TNF-a values (p < 0.001) and for infection and positive IL-6 values (p < 0.02) in Fisher's exact test. Discussion
Knowledge of the sequence of immunological signals resulting in rejection of solid allografts is not complete. Interleukin-1, IFN-y and TNF-a have been described as early mediators of T -lymphocyte triggering, inducing activation and growth of T - and B-lymphocytes and ultimately inflammation and graft destruction (1, 2). Increased TNF-a levels (cut off 100 pgl ml; range 50-800 pg/ml) have been observed in kidney transplantation
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 435
during allograft rejection (8) and recently human liver graft and experimental cardiac allograft rejections were found to be associated with increased plasma levels of TNF-a (mean values 940 pg/ml; 37 rejections) at the time of the rejection episode (9, 10). Purified TNF-a was reported to accelerate graft rejection in the experimental rat model and antibodies to TNF-a or TNF-~ either alone or in combination, prolonged cardiac graft survival in the experimental animal model without any other immunosuppression, supporting the important role of TNF in allograft rejection (10). In an attempt to investigate the first reactions of the host immune system to solid organ transplants, we have studied the intraoperative release of TNF-a in human liver transplantation. Transplantation of this organ is different from grafting of other organs in several aspects: (a) donor and recipients in most case have one or more HLA mismatches, resulting in a high frequency of usually reversible rejections, (b) clamping of the portal vein without veno-venous bypass, as performed at our center, leads to a release of intestine-derived endotoxin, which is a potent inducer of cytokines and (c) endotoxin accumulates during the anhepatic phase, since it is not eliminated by Kupffer cells of the liver. The role of endotoxin in TNF-a production and induction of rejection in LTX has not been documented. Furthermore, IL-6, which is produced in response to TNF-a and triggers the production of acute phase proteins and possibly also the proliferation of B lymphocytes in vivo, was also determined in this study (11-13). The disadvantage of the liver transplantation model is the high frequency of complications caused by primary organ malfunction, viral and bacterial infections, drug side effects and recurrent disease and the difficulties to establish a correct differential diagnosis of these complications. As shown here, the determination of spontaneous blastogenesis in peripheral blood of the graft recipients in combination with measurements of the T -helper! inducer lymphocyte subpopulation and donor-specific MLC activity can be used to confirm the histological diagnosis of rejection. Determination of cytokines or soluble cytokine receptors and neopterine are alternative nonspecific methods to detect activation of the immune system in graft recipients (14, 15). The different complications can be distinguished by correlation of the unstimulated proliferative activity with changes in the T -helper! inducer lymphocyte subpopulations as proved by measurements of the donor-specific MLC activity. For grade I rejections (lymphocyte infiltration), the histological grading of the biopsies alone correlated poorly with the further clinical course of these patients. In some patients the lymphocytic infiltrate resulted in graft rejection, whereas in other patients the infiltration resolved without rejection therapy. Although the histological grading of liver biopsies is the golden standard of diagnosis of rejection in liver recipients, in the present study the histological diagnosis of grade I rejection was regarded as true rejection if confirmed tw demonstration of donor-specific MLC reactivity and demonstration of proliferative activity/ T -helper!inducer increases. By means of these immunological tests in combination with clinical signs, blood chemistry, biopsies and serological
436 . G. HAMILTON et al.
and bacteriological tests the liver graft recipients were assigned to groups without complications, infections and rejections with high reliability. The immunological tests revealed infection in cases in which the infectious agents could not be isolated or identified despite fever and response to therapies. Endotoxin, tumor necrosis factor-a and interleukin-6 were measured from blood samples collected prior to, and at different time points during, human liver transplantation (n = 27). For determination of ET, plasma components, which could interfere with the Limulus test were removed by ultrafiltration/phenol treatment (7), and TNF-a and IL-6 were assayed with highly sensitive commercial test kits. Endotoxin was elevated moderately in the plasma of liver graft recipients at the end of the anhepatic phase, but there was no correlation to subsequent infections or rejections. Individual differences in cold ischemia time of the graft, duration of the anhepatic phase and other factors might be responsible for differences in ET concentrations. These results are in contrast to a previous study of endotoxemia in human liver transplantation, which reported endotoxin to represent a prognostic factor for LTX (16). However, this study differs from the present work by the inclusion of patients with primary non-function of their graft (10 %), who had high endotoxin blood levels, and the use of a test system which includes interfering blood substances (7). A highly significant increase in TNF-a (over 200 pg/ml plasma) was detected for graft recipients, who subsequently rejected their allografts, in comparison to recipients without complications and infections. Although not significant, the patients in the group with rejections tended to have elevated TNF-a levels prior to allograft exposure. These preoperatively elevated TNF-a levels are not the result of a common underlying disease since these patients had various indications for LTX. In agreement with other work (9) we have not detected elevated TNF-a levels in samples from patients with hepatitis immediately prior to LTX. Our results show no correlation of endotoxin plasma concentrations at the end of the anhepatic phase with TNF-a levels after recirculation in human LTX, supporting the hypothesis that TNF-a is induced by another mechanism, possibly via interleukin-2 (17), in the early phase of reaction to the allograft. The induction of TNF-a immediately after contact with the allograft can contribute to graft rejection by promoting the binding of lymphocytes to graft endothelial tissue (18) or induction of MHC alloantigens in the graft (19). In addition to release by stimulated monocytes in bacterial infections (20), TNF-a is also produced by lymphocytes (21) and participates directly in T cell activation and T cell proliferation (22) and also in the final steps of tissue injury and graft destruction (23). The intraoperative analysis of TNF-a plasma concentrations allowed us to measure the release of TNF-a in response to allograft exposure prior to the initiation of the immunosuppressive therapy, which alters production and release of cytokines. Immunosuppressive therapy for the recipients was started at the end of the operation with steroids and immunoprophylaxis
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 437
for the first post-transplant time was provided by ATG treatment beginning at the second post-transplant day. The TNF-a plasma concentrations under this immunosuppressive regimen showed no correlation with subsequent complications since release of TNF-a is suppressed by steroids (24) and induced by ATG- or OKT-3-mediated lysis of T-Iymphocytes independent of allograft rejection (25, 26). Interleukin-6 has been characterized as B-Iymphocyte differentiation factor, T cell costimulatory factor and as inducer of the acute phase response in hepatocytes (11-13). The intraoperative course of IL-6 was found to correlate with viral and bacterial infectious complications in the liver recipients. The mechanism of IL-6 induction in this patient group is not clear, but the increase became obvious at the beginning of anhepatic phase, indicating that infections may exist prior to transplantation or occur during the anhepatic phase of the liver transplantation. The finding that the production of IL-6 is induced by viruses may help to explain the high levels of IL-6 observed in the liver graft recipients in this study, since these patients show a high incidence of cytomegalovirus (CMV) and other viral infections (27). In summary, early rejections of liver grafts by a group of recipients under ATG immunoprophylaxis have been found to correlate significantly with elevated intraoperative plasma concentrations of TNF-a immediately after allograft exposure. In addition to its direct participation in triggering of T -lymphocytes in the process of allograft rejection, TNF-a may enhance the immunogenicity of the allograft, most likely by induction of MHC antigens and promoting of leukocyte adherence, resulting in a high frequency of subsequent rejections. The results also demonstrate that endotoxin accumulation during the anhepatic phase of liver transplantation is of minor importance for the induction of TNF-a and allograft rejection. Intraoperative elevations in the plasma concentrations of IL-6 were found to be a prognostic factor for early infections in the liver graft recipients.
References 1. SIBLEY, R. K. 1989. Pathology and immunopathology of solid organ graft rejection. Transplant. Proc. 21: 14. 2. TILNEY, N. L. and J. W. KUPIEC-WEGLINSKI. 1989. Advances in the understanding of rejection mechanisms. Transplant. Proc. 21: 10. 3. HALLORAN, P. F., S. M. COCKFIELD, and J. M. ADRENAS. 1989. The mediators of inflammation (IL-1, IFN-y and TNF) and their relevance to rejection. Transplant. Proc. 21: 26. 4. FILIPPONI, F., J. VAN DE STADT, P. IcARD, A. MICHEL, and D. HOUSSIN. 1988. Prevention of death following one-hour occlusion of the portal vein in the rat. Eur. Surg. Res. 20: 39. 5. BEUTLER, B. and A. CERAMI. 1987. Cachectin: more than a tumor necrosis factor. N. Eng!. J. Med. 316: 379. 6. BROUCKAERT, P., D. R. SPRIGGS, G. DEMETRI, D. W. KUFE, and W. FIERS. 1989. Circulating interleukin 6 during continuous infusion of tumor necrosis factor and interferon. J. Exp. Med. 169: 2257.
438 . G. HAMILTON et al. 7. BERGER, D., E. MARZINZIG, M. MARZINZIG, and H. G. BEGER. Quantitative endotoxin determination in blood - chromogenic modification of the Limulus amoebocyte lysate test. Eur. Surg. Res. 20: 128. 8. MAURY, C. P. J. and A. M. TEPPO. 1987. Raised serum levels of cachectin/tumor necrosis factor in renal allograft rejection. J. Exp. Med. 166: 1132. 9. IMAGAWA, D. K., J. M. MILLIS, K. M. OLTHOFF, L. J. DERUS, D. CHIA, L. R. SUGICH, M. OZAWA, R. A. DEMPSEY, Y. IWAKI, P. J. LEVY, P. I. TERASAKI, and R. W. BUSUTTIL. 1990. The role of tumor necrosis factor in allograft rejection. Transplantation 50: 219. 10. IMAGAWA, D. K., J. M. MILLIS, K. M. OLTHOFF, P. SEU, R. A. DEMPSEY, J. HART, P. I. TERASAKI, E. M. W ASEF, and R. W. BUSUTTIL. 1990. The role of tumor necrosis factor in allograft rejection. Transplantation 50: 189. 11. HELFGOTT, D. c., S. B. TATTER, U. SANTHANAM, R. H. CLARICK, N. BHARDWAJ, L. T. MAY, and P. B. SEHGAL. 1989. Multiple forms of interferon-B2/IL-6 in plasma and body fluids during acute bacterial infection. J. Immunol. 142: 948. 12. HEINRICH, P. c., J. V. CASTELL, and T. ANDUS. 1990. Interleukin-6 and the acute phase response. Biochem. J. 265: 621. 13. WAAGE, A., T. ESPEVIK, and A. HALSTENSEN. 1988. Tumor necrosis factor, interleukin-1 and interleukin-6 in human septic shock. Scand. J. Immunol. 28: 267. 14. PERKINS, J. D., D. L. NELSON, J. RAKE LA, P. M. GRAMBSCH, and R. A. F. KROM. 1989. Soluble IL-2 receptor level as an indicator of liver allograft rejection. Transplantation 47: 77. 15. TILG, H., W. VOGEL, W. E. AULITZKY, D. SCHOENITZER, R. MARGREITER, O. DIETZE, G. JUDMAIER, H. WACHTER, and C. HUBER. 1989. Neopterin excretion after liver transplantation and its value in differential diagnosis of complications. Transplantation 48: 594. 16. YOKOYAMA, I., S. TODO, T. MIYATA, R. SELBY, A. G. TZAKIS, and T. E. STARZL. 1989. Endotoxemia and human liver transplantation. Transplant. Proc. 21: 3833. 17. ECONOMOU, J. S., W. H. McBRIDE, and R. ESSNER. 1989. Tumor necrosis factor production by interleukin-2 activated macrophages in vitro and in vivo. Immunology 67:
514. 18. CAVENDER, D., Y. SAEGUSA, and M. ZIFF. 1987. Stimulation of endothelial cell binding of lymphocytes by tumor necrosis factor. J. Immunol. 139: 1855. 19. COLLINS, T., L. A. LAPIERRE, W. FIERS, J. L. STROMINGER, and J. S. POBER. 1986. Recombinant tumor necrosis factor increases mRNA-levels and surface expression of HLA-A,B antigens in vascular endothelial cells and dermal fibroblasts in vitro. Proc. Nat!. Acad. Sci. USA 83: 446. 20. DEBETS,J. M. H., R. KAMPMEIJER, M. P. M. H. VANDERLINDEN, W. A. BUURMAN, and C. F. VAN DER LINDEN. 1989. Plasma tumor necrosis factor and mortality in critically ill septic patients. Crit. Care Med. 17: 489. 21. SUNG, S. S. J., J. M. BJORNDAHL, C. Y. WANG, H. T. KAo, and S. M. Fu. 1988. Production of tumor necrosis factor/cachectin by human T cell lines and peripheral blood T lymphocytes stimulated by phorbol miristate acetate and anti-CD3 antibody. J. Exp. Med. 167: 937. 22. SCHEURICH, P., B. THOMA, U. UCER, and K. PFIZENMAIER. 1987. Immunregulatory activity of recombinant tumor necrosis factor (TNF)-alpha: induction of TNF receptors of human T cells and TNF-alpha mediated enhancement of T cell responses. J. Immunol. 138: 1786. 23. DECKER, T., M.-L. LOHMANN-MATTHES, and G. E. GIFFORD. 1987. Cell-associated tumor necrosis factor (TNF) as a killing mechanism of activated cytotoxic macrophages. J. Immunol. 138: 957. 24. DUPONT, E., L. SCHANDENE, C. DENYS, A. CRUSIAUX, and J. WYBRAN. 1989. Assessment of production of tumor necrosis factor-alpha under the influence of immunosuppressive drugs. Transplant. Proc. 21: 70. 25. DEBETS, J. M. H., K. M. L. LEUNISSEN, H. J. VAN HOOFF, C. J. VAN DER LINDEN, and W. A. BUURMAN. 1989. Evidence of involvement of tumor necrosis factor in adverse reactions during treatment of kidney allograft rejection with antithymocyte globulin. Transplantation 47: 487.
Endotoxin, TNF-alpha and IL-6 Liver Transplantation· 439 26. ABRAMOWICZ, D., L. SCHANDENE, M. GOLDMAN, A. CRUSIAUX, P. VEREESTRAETEN, L. DE PAUW, J. WYBRAN, P. KINNAERT, E. DUPONT, and C. TOUSSAINT. 1989. Release of tumor necrosis factor, interleukin-2 and gamma-interferon in serum after injection of OKT3 monoclonal antibody in kidney transplant recipients. Transplantation 47: 606. 27. HAMILTON, G., T. POPOV, F. MUEHLBACHER, and F. PIZA. 1989. Early detection of viral infection in liver graft recipients by immune monitoring. Transplant. Proc. 21: 2266. Dr. GERHARD HAMILTON, 1. Surgical University Clinic, Alserstr. 4, A-I090 Vienna, Austria