British journal o/Haematologg. 1992. 81, 23-26
Lymphocytes from multi-transfused patients exhibit cytotoxicity against autologous cells EDWARD R . K A M I N S K I , J I L L M. €lows,*JOHN M. GOLDMAN*A N D J . RICHARDBATCHELOR Departments of Immunology and *Haernatology, Royal Postgraduate Medical School, Hammersmith Hospital, London Received 1 Octobcr 1991; accepted for publication 17 December 1991
Summary. We previously demonstrated that multitransfused patients with severe aplastic anaemia (SAA) exhibit high numbers of alloreactive cytotoxic T lymphocyte precursors directed against their HLA identical siblings. In this study a group of patients who had received multiple blood transfusions for SAA. other haematological diseases or acute blood loss were tested for autocytotoxicity and the results compared with those of untransfused controls. These controls consisted of normal individuals, patients with chronic myeloid leukaemia (CMI,) or untransfused patients with SAA. There was a signiticantly higher degree of autocytotoxicity in multitrans-
fused patients. than in the untransfused controls, including untransfused patients with SAA ( P = 0.0001).These results suggest that blood transfusion is responsible for inducing autoreactivity. In one patient, in whom both alloreactive anti-non-MHC and autoreactive cytotoxic T lymphocytes (CTL)had been detected, it was demonstrated that there was no crossreactivity between the alloreactive and autoreactive CTL responses. Inhibition studies using monoclonal antibodies revealed the effector cells to be T lymphocytes and the restricting determinants to be both HLA class I and I1 molecules.
In previous studies we have determined the frequencies of alloreactive cytotoxic T lymphocyte precursors (CTL-p) in various HIA matched and mismatched responder/stimulator combinations by means of limiting dilution analysis (Kaminski et nl. 1988. 1989). The assay variables were established such that no autologous or third party reactivity was demonstrable in normal individuals (Kaminski et ul, 1991). I n a more recent study we demonstrated that in multitransfused patients with severe aplastic anaemia (SAA), high numbers of CT1,-p could be detected against their HLA identical siblings. and that these were probably directed at non H I A antigens (Kaminski et ul. 1990). However, in this latter group of patients we also discovered that autoreactivity was often detectable during the CTL-p analysis. We therefore established an assay for detecting autoreactive CTL. based on our 1,DA system, and used it to study the effect of blood transfusion on autocytotoxicity.
varices (n =4). sickle cell anaemia (n = 2). myelodysplastic syndrome (n = 2 ) or /3-thalassaemia ( n = 1). The patients with SAA or other haematological disorders had all received more than 100 units of blood, and the four patients with acute gastrointestinal bleeds more than 10 units of blood. The control group consisted of 16 untransfused individuals, comprising healthy volunteers ( n = 5 ) . patients with chronic myeloid leukaemia ( n = 5) and patients with severe aplastic anaemia ( n = 6 ) . Cells. Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized venous blood by separation on Ficoll-Paque (Pharmacia). The cells were then washed in KPMI 1 6 4 0 (Flow) and cryopreserved in 10% dimethyl sulphoxide (BDH), 40% fetal calf serum (Seralab) and 50% RPMI until required. Culture medium. This consisted of RPMI 1640 supplemented with 2 mM L-glutamine. 100 IU/ml penicillin, 100 pg/ml streptomycin, and 10% pooled human AB serum (Oxford Blood Transfusion Centre). Autologous mixed lymphocyte culture. The PBMCs were thawed, and 5 x lo4responder PBMCs plated out with equal numbers of irradiated (30 Gy) autologous PBMCs in 96-well, round-bottom micro-titre plates (Flow). Each experiment consisted of 50 replicates. As a control. irradiated autologous cells were cultured without responders ( 1 0 replicates). The cultures were incubated in 10%humidified C02in air at 3 7OC
MATEKIALS A N D METHODS Study group. This consisted of 2 3 multitransfused patients who were diagnosed as severe aplastic anaemia ( n = 14). acute blood loss due to peptic ulceration or oesophageal
Correspondence: Dr E. R . Kaminski, Division of Immunology, Ilepartment of C'rllular and Molecular Sciences. St George's Hospital Medical School. Cranmer Terrace, London SW 17 ORE.
23
24
Edward R. Kaminski et a1
for 7 d. On day 3 the cultures were fed with culture medium containing recombinant interleukin-2 (rIL-2) (Boehringer Mannheim), to give a final concentration of 5 U/ml r - 2 per well. Targetsfor cytotoxicity assay. These were prepared on day 0 by incubating autologous PBMCs with 2 pg/ml of phytohaemagglutinin (PHA) (Wellcome). Every 2-3 d the lymphoblasts were washed, split and fed with rIL-2 (10 U/ml). Cytotoxicity assay. On day 7 of culture, the target ‘blasts’ were washed, counted and labelled with 5’Cr (Amersham), 7.4 MBq/107 cells for 1 h. The target cells (1x lo4 per well) were added to the culture plates and incubated for 4 h. Spontaneous and total 5’Cr release were determined by adding the targets to 100 p1 of medium and 100 pI of 1% Triton X. respectively. At the end of the incubation, 100 p1 of supernatant was removed and the amount of 5’Cr release measured in a gamma counter. The assay wells were scored as positive iftheir jlCr release was greater than the mean plus 3 SD of the control 51Crrelease. The result was expressed as percentage CTL activity and calculated as follows:
”/o CTL activity =
(number of positive wells) x 100 (total number of wells)
Specificitystudies. In one case responder cells were cultured with either autologous or allogeneic PBMCs from an HLA identical sibling. On the day of assay, split well analysis was performed and the cultures tested for cytotoxicity against both autologous and allogeneic targets. Monoclonal antibody (MoAB) inhibition studies. The following MoAB were used to block cytotoxicity assays: Anti-HLA class I (PA2.6) (Andrew McMichael. Oxford); Anti-HLA class I1 (NFK2) (Sue Fuggle. Oxford); Anti-Pan T cell receptor a l p (T Cell Sciences). Statistiral analysis. The Mann-Whitney U test was used to analyse the experimental data.
with underlying aplastic anaemia, blood from a group of 1 6 untransfused individuals, including six patients with SAA. five patients with chronic myeloid leukaemia, and five normal individuals, was assayed. The auto CTL values ranged from 0%to 11%(Fig 1).Autocytotoxicity was in the same low range in the untransfused SAA group as in the other untransfused individuals. Comparison of autocytotoxicity in transfused and untransfused patients The data from the above results was compared in Fig 1. Autocytotoxicity was significantly higher (P=0.0001) in transfused than in untransfused controls (Mann-Whitney U test). Specijcity of allo- and autoreactive CTL responses One of the multitransfused SAA patients (R). in whom a high auto CTL activity was demonstrated had previously been shown to have a high alloreactive CTL-p frequency of 1:2 4 x 10’ against a n HLA identical sibling (S) (Kaminski et al, 1990). An experiment was set up in which the responder (R) cells were primed, either with allogeneic stimulator cells (S)or autologous cells, and. by means of split well analysis, tested for cytotoxicity against either stimulator (S) targets or autologous targets (Fig 2). There was no crossreactivity between the allo- and autoreactive CTL responses.
lo)80
RESULTS Autorytotoxicity in patients with SAA A group of 14 multitransfused patients with SAA were tested for autocytotoxicity. The auto CTL values ranged from 0%to 90% positive wells (Fig 1).All patients had been multitransfused ( > 100 donor units) and there was no correlation between auto CTL activity and whether or not the patient was transfusion dependent at the time of testing. Autocytotoxicity in multitransfused patients with other diseases In order to test whether the autocytotoxicity was induced by blood transfusion in patients with conditions other than SAA, five multitransfused patients with other haematological disorders (sickle cell anaemia. myelodysplastic syndrome or P-thalassaemia) and four patients transfused for acute blood loss, were assayed. The auto CTL values ranged from 10%to 66% positive wells (Fig 1) similar to those found in the transfused SAA patients. Autocytotoxicity in untransfused individuals In order to test further whether the autocytotoxicity was induced by blood transfusion alone or was also associated
I
t
*.
11-14 -transfused&+
n=5
n=4
0 0
Q
O
n-5
n-5
6
O
n-6
untransfused d
Fig 1, Comparison of autocytotoxicity in transfused and untransfused patients. SAA=severe aplastic anaemia, HAEM =other hae-
matological disorder (sickle cell anaemia, myelodysplastic syndrome and /I-thalassaemia). BLEED=acute blood loss (peptic ulceration, oesophagealvarices),NORMAL = normal individuals,CML = chronic myeloid leukaemia. Autocytotoxicity is significantly higher in transfused than in untransfused patients (P=O.OOOI). Mann-Whitney U test.
Blood Transfusion-induced Autocytotoxicity
R+S vs S’
R+S vs R’
R+R vs R’
R+R vs S’
Fig 2. Determination of specificity of allo- and autocytotoxic responses. Responder (R)cells were either primed with allogeneic stimulator ( S ) cells or autologous (K)cells and tested for cytotoxicity against allogeneic or autologous targets (*).
NIL
TCR
HLA I
HLA II
Fig 3. Monoclonal antibody inhibition studies. Autocytotoxicity was assayed in the absence (NIL) and presence of MoAbs to the clip T cell receptor ( ‘ K K ) . HLA class I (HLA I ) and HLA class II (HLA 11) molecules.
Mnrindoriul antibody inhibition studies One of the multitransfused SAA patients was found to have an autocytotoxic response of 3 7%. This autocytotoxicity was inhibited by an anti-ccip T cell receptor MoAb (100%). and both by an anti-class I(40x)and class I1 (65%) MoAb (Fig 3 ) .
DISC 11S S I O N In this study we have demonstrated a significantly higher degree of autocytotoxicity in multitransfused patients with SAA than in the untransfused controls ( P = 0.0001). Multitransfused patients with other diseases also gave high auto CTL activity, whilst untransfused patients with SAA gave low CTL activity in the same range as other untransfused individuals. This suggests that blood transfusion and not aplastic anaemia per SP is responsible for inducing autoreactivity. The possibility that this autocytotoxicity is in fact an alloreactive response against persisting transfusion de-
25
rived lymphocytes is unlikely since many of the ‘high responders’ had not been transfused for more than a year prior to the study. In one patient with a high auto CTL activity, a high alloreactive CTL-p frequency against a n HLA identical sibling had been previously demonstrated (Kaminski et a/, 1990). In order to dissect the alloreactive and autoreactive CTL responses a split-well cytotoxic assay was performed. This clearly demonstrated a lack of crossreactivity between the allo- and autoreactive CTL responses, suggesting that distinct cell populations are involved. Autocytotoxicity in SAA patients has been described previously by Warren et a/ ( I 980), who concluded that this was related to the aplasia and not an effect ofthe transfusion, This interpretation conflicts with the tindings presented here, where the low reactivity of the untransfused SAA patients suggests that transfusion is the cause of the autocytotoxicity. Consistent with this hypothesis was our finding of autoreactive cytotoxic cells in patients transfused for acute blood loss. The coexistence of autocytotoxicity with alloreactivity to HLA identical siblings was also described by Warren et a1 (1980). Similar results were found in a study after bone marrow transplantation, where non-overlapping populations of alloreactive and autoreactive CTLs were demonstrated (Hurme & Sihvola. 1985). Autocytotoxic cells have also been detected in patients with graft-versus-host disease following bone marrow transplantation (Parkman et a / . 1980: Santos, 1989). in patients with tuberculosis (Del Gallo et a / . 1990). in rheumatoid synovial fluid (Sugiyama et al, 1989) and in patients with vitiligo (Vignale et a/, 1988). The effector cells in our study appear to be cytotoxic T lymphocytes restricted both by HLA class I and I1 molecules. The target cells are likely to be T lymphocytes in view of the fact that PHA blasts were used in the assay. However, the antigen specificity of the autoreactive CTLs is unknown. One postulate is that the patient’s CTLs are recognizing foreign peptides, derived from transfused cells, situated in the MHC cleft of autologous cells. This may give rise to autocytotoxic responses. which in some cases may be due to CD8+ cells and in others to CD4+ cells, depending on whether the peptide is situated in a class I or class I1 MHC cleft. A second explanation is that these are responses of a regulatory or suppressor population which is inhibiting the ‘in vivo’ alloreactive response to HLA or minor histocompatibility antigens. Evidence for suppression of alloreactive T lymphocytes by anti-idiotypic cytolytic T cells in rat and murine graft-versus-host reactions was provided by Kimura & Wilson (1984) and Kosmatopoulos et a/ (1987). The mechanisms of anti-idiotypic suppression have been discussed by Batchelor et a1 (1989) and Wilson ( 1989). An hypothesis encompassing both the above explanations has recently been put forward by van Rood & Claas (1990) to explain the well-documented effect of pretransplant blood transfusions (PTBT) improving survival of renal allografts. They propose that CD4 + cells recognize allo-MHC peptides, of a specificity shared by the PTBT and allograft donor, derived from allogeneic cells and presented by autologous T cells. In this case the MHC molecules could be class I1 DR
26
Edward R. Kaminski et a1
molecules, which have been upregulated following activation of the autologous T lymphocytes by allogeneic stimulation. Such cells have been described (De Koster et al, 1989: Ottenhof& Mutis. 1990).The autoreactive CD4 cells then may downregulate the activated alloreactive cells. Although the functional significance of the autoreactive cytotoxic T cells remains to be determined, the following points should be made. (i) The presence of cellular autoreactivity, like autoreactive antibodies. does not necessarily lead to tissue damage. One reason for this may be that the lectin-activated target cells used in this assay have increased expression of HLA class I and class I1 antigens and adhesion molecules making them more sensitive to lysis, whereas in vivo the situation is quite distinct. (ii) The findings demonstrate that autocytotoxic cells can be generated in subjects who are not suffering from autoimmune disease. (iii) In CTLp analysis for bone marrow transplantation, autoreactive CTLs may exist and contribute to the overall CTLp frequency, thus emphasizing the need to include appropriate controls. In conclusion, we have demonstrated autocytotoxicity in multitransfused patients with SAA and other diseases and conclude that this is a n effect of transfusion and not associated with aplastic anaemia per se. Furthermore, we have demonstrated that the allo-non-MHC reactivity previously demonstrated in the same group of patients is not cross reactive with this autoreactivity. which suggests that different populations of cells are responsible for the two phenomena. The effector population appears to consist of cytotoxic T lymphocytes restricted both by HLA class I and I1 antigens. The number of patients in this preliminary study is small and further investigations will be necessary to confirm and extend our findings.
+
ACKNOWLEDGMENTS
We are grateful to Professor E. C. Gordon-Smith for supplying one of the patient samples and to Dr S. Fuggle for generously donating monoclonal antibodies. REFERENCES Batchelor. J.R.. Lombardi, G. & Lechler. R.I. (1989) Speculationson the specificity of suppression. Immunology Today, 10, 37-40. De Koster, H.S.. Anderson. D.C. & Terrnijtelen. A. (1989) T cells sensitized to synthetic HLA-DR3 peptide give evidence of continuous presentation of denatured HLA-DR3 molecules by HLA-DP. journal of Experimental Medicine. 1 6 9 , 1 19 1-1 196. Del Gallo. F.. Lombardi. G.. Piccolella.E.. Gilardini Montani. M.S.. Del Porto. P.. Pugliese. 0..Antonelli. G. & Colizzi, V. (1 990) Increased
autoreactive T cell frequencyin tuberculosis patients. International Archives of Allergy and Applied Immunology, 9 1 , 36-42. Hurme, M. & Sihvola, M. (1985)In vivo activation of autoreactive cytotoxic T lymphocytes after bone marrow transplantation: evidence for the prethymic specificity of T cells. journal of Immunology, 1 3 5 , 1108-1 112. Kaminski, E.. Sharrock. C., Hows, I., Ritter, M., Arthur, C.. Mackinnon, S. & Batchelor, J. (1988) Frequency analysis of cytotoxic T lymphocyte precursors-possible relevance to HLA-matched unrelated donor bone marrow transplantation. Bone Marrow Transplantation. 3, 149-1 55. Kaminski, E., Hows. J.. Man, S., Brookes. P., Mackinnon. S . , Hughes, T.. Avakian, 0..Goldman. J.M. & Batchelor.J.R. ( 1 989) Prediction of graft-versus-host disease by frequency analysis of cytotoxic T cells after unrelated donor bone marrow transplantation. Transplantation, 48. 608-61 3. Kaminski, E.R.. Hows. J.M.. Goldman, J.M. & Batchelor. J.R. (1990) Pretransfused patients with severe aplastic anaemia exhibit high numbers of cytotoxic T lymphocyteprecursors probably directed at non HLA antigens. British journal of Haematology. 76, 40 1-405. Kaminski, E.. Hows. J., Goldman. J. & Batchelor. R. ( 1 99 1 ) Optimising a limiting dilution culture system for quantitating frequencies of alloreactive cytotoxic T lymphocyte precursors. Cellular lmmunology. 1 3 7 , 88-95. Kimura, H. & Wilson, D.B. ( 1984) Anti-idiotypic cytotoxic T cells in rats with graft-versus-host disease. Nature. 308, 46 3-464. Kosmatopoulos, K..Algara, D.C. & Orbach-Arbouys.S. ( 1 987) Antireceptor anti-MHC cytotoxic T lymphocytes: their role in the resistance to graft versus host reaction. journal of Immunology, 1 3 8 , 1038-1041. Ottenhoff. T.H. & Mutis. T. ( 1 990) Specifickilling of cytotoxic T cells and antigen-presenting cells by CD4+ cytotoxic T cell clones. journal of Experimental Medicine, 1 7 1 , 201 1-2024. Parkman. R., Rappeport. J. & Rosen. F. ( 1 980) Human graft-versushost disease. journal of Investigative Dermatology, 74. 2 76-279. Santos. G.W. (1989) Syngeneic or autologous graft-versus-host disease. International journal of Cell Cloning, 7, 92-99. Sugiyama, E., Suzuki, H. & Yano, S. (1989) Autocytotoxic cells in rheumatoid synovial fluids. IournalofRheumatology. 16,276-279. Van Rood, J.J. & Claas, F.H.J.( 1 990) The influence of allogeneic cells on the human T and B cell repertoire. Science, 248, 1388-1 393. Vignale, R., Paciel. J. & Calandria. L. (1988)Immunological changes in common vitiligo. Medicina Cutanea Ibero-Latino-Americana. 16, 34 3-347. Warren. R.P.. Storb, R.. Thomas. E.D.. Su. P.J., Mickelson. E.M. & Weiden. P.L. ( 1 980) Autoimmune and alloimmune phenomena in patients with aplastic anemia: cytotoxicity against autologous lymphocytes and lymphocytes from HLA identical siblings. Blood. 5 6 , 683-689. Wilson, D.B. (1989) Idiotypic regulation of T cells in graft-versushost disease and autoimmunity.Immunological Reviews, 107. 1 59177.
Lymphocytes from multi-transfused patients exhibit cytotoxicity against autologous cells EDWARD R . K A M I N S K I , J I L L M. €lows,*JOHN M. GOLDMAN*A N D J . RICHARDBATCHELOR Departments of Immunology and *Haernatology, Royal Postgraduate Medical School, Hammersmith Hospital, London Received 1 Octobcr 1991; accepted for publication 17 December 1991
Summary. We previously demonstrated that multitransfused patients with severe aplastic anaemia (SAA) exhibit high numbers of alloreactive cytotoxic T lymphocyte precursors directed against their HLA identical siblings. In this study a group of patients who had received multiple blood transfusions for SAA. other haematological diseases or acute blood loss were tested for autocytotoxicity and the results compared with those of untransfused controls. These controls consisted of normal individuals, patients with chronic myeloid leukaemia (CMI,) or untransfused patients with SAA. There was a signiticantly higher degree of autocytotoxicity in multitrans-
fused patients. than in the untransfused controls, including untransfused patients with SAA ( P = 0.0001).These results suggest that blood transfusion is responsible for inducing autoreactivity. In one patient, in whom both alloreactive anti-non-MHC and autoreactive cytotoxic T lymphocytes (CTL)had been detected, it was demonstrated that there was no crossreactivity between the alloreactive and autoreactive CTL responses. Inhibition studies using monoclonal antibodies revealed the effector cells to be T lymphocytes and the restricting determinants to be both HLA class I and I1 molecules.
In previous studies we have determined the frequencies of alloreactive cytotoxic T lymphocyte precursors (CTL-p) in various HIA matched and mismatched responder/stimulator combinations by means of limiting dilution analysis (Kaminski et nl. 1988. 1989). The assay variables were established such that no autologous or third party reactivity was demonstrable in normal individuals (Kaminski et ul, 1991). I n a more recent study we demonstrated that in multitransfused patients with severe aplastic anaemia (SAA), high numbers of CT1,-p could be detected against their HLA identical siblings. and that these were probably directed at non H I A antigens (Kaminski et ul. 1990). However, in this latter group of patients we also discovered that autoreactivity was often detectable during the CTL-p analysis. We therefore established an assay for detecting autoreactive CTL. based on our 1,DA system, and used it to study the effect of blood transfusion on autocytotoxicity.
varices (n =4). sickle cell anaemia (n = 2). myelodysplastic syndrome (n = 2 ) or /3-thalassaemia ( n = 1). The patients with SAA or other haematological disorders had all received more than 100 units of blood, and the four patients with acute gastrointestinal bleeds more than 10 units of blood. The control group consisted of 16 untransfused individuals, comprising healthy volunteers ( n = 5 ) . patients with chronic myeloid leukaemia ( n = 5) and patients with severe aplastic anaemia ( n = 6 ) . Cells. Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized venous blood by separation on Ficoll-Paque (Pharmacia). The cells were then washed in KPMI 1 6 4 0 (Flow) and cryopreserved in 10% dimethyl sulphoxide (BDH), 40% fetal calf serum (Seralab) and 50% RPMI until required. Culture medium. This consisted of RPMI 1640 supplemented with 2 mM L-glutamine. 100 IU/ml penicillin, 100 pg/ml streptomycin, and 10% pooled human AB serum (Oxford Blood Transfusion Centre). Autologous mixed lymphocyte culture. The PBMCs were thawed, and 5 x lo4responder PBMCs plated out with equal numbers of irradiated (30 Gy) autologous PBMCs in 96-well, round-bottom micro-titre plates (Flow). Each experiment consisted of 50 replicates. As a control. irradiated autologous cells were cultured without responders ( 1 0 replicates). The cultures were incubated in 10%humidified C02in air at 3 7OC
MATEKIALS A N D METHODS Study group. This consisted of 2 3 multitransfused patients who were diagnosed as severe aplastic anaemia ( n = 14). acute blood loss due to peptic ulceration or oesophageal
Correspondence: Dr E. R . Kaminski, Division of Immunology, Ilepartment of C'rllular and Molecular Sciences. St George's Hospital Medical School. Cranmer Terrace, London SW 17 ORE.
23
24
Edward R. Kaminski et a1
for 7 d. On day 3 the cultures were fed with culture medium containing recombinant interleukin-2 (rIL-2) (Boehringer Mannheim), to give a final concentration of 5 U/ml r - 2 per well. Targetsfor cytotoxicity assay. These were prepared on day 0 by incubating autologous PBMCs with 2 pg/ml of phytohaemagglutinin (PHA) (Wellcome). Every 2-3 d the lymphoblasts were washed, split and fed with rIL-2 (10 U/ml). Cytotoxicity assay. On day 7 of culture, the target ‘blasts’ were washed, counted and labelled with 5’Cr (Amersham), 7.4 MBq/107 cells for 1 h. The target cells (1x lo4 per well) were added to the culture plates and incubated for 4 h. Spontaneous and total 5’Cr release were determined by adding the targets to 100 p1 of medium and 100 pI of 1% Triton X. respectively. At the end of the incubation, 100 p1 of supernatant was removed and the amount of 5’Cr release measured in a gamma counter. The assay wells were scored as positive iftheir jlCr release was greater than the mean plus 3 SD of the control 51Crrelease. The result was expressed as percentage CTL activity and calculated as follows:
”/o CTL activity =
(number of positive wells) x 100 (total number of wells)
Specificitystudies. In one case responder cells were cultured with either autologous or allogeneic PBMCs from an HLA identical sibling. On the day of assay, split well analysis was performed and the cultures tested for cytotoxicity against both autologous and allogeneic targets. Monoclonal antibody (MoAB) inhibition studies. The following MoAB were used to block cytotoxicity assays: Anti-HLA class I (PA2.6) (Andrew McMichael. Oxford); Anti-HLA class I1 (NFK2) (Sue Fuggle. Oxford); Anti-Pan T cell receptor a l p (T Cell Sciences). Statistiral analysis. The Mann-Whitney U test was used to analyse the experimental data.
with underlying aplastic anaemia, blood from a group of 1 6 untransfused individuals, including six patients with SAA. five patients with chronic myeloid leukaemia, and five normal individuals, was assayed. The auto CTL values ranged from 0%to 11%(Fig 1).Autocytotoxicity was in the same low range in the untransfused SAA group as in the other untransfused individuals. Comparison of autocytotoxicity in transfused and untransfused patients The data from the above results was compared in Fig 1. Autocytotoxicity was significantly higher (P=0.0001) in transfused than in untransfused controls (Mann-Whitney U test). Specijcity of allo- and autoreactive CTL responses One of the multitransfused SAA patients (R). in whom a high auto CTL activity was demonstrated had previously been shown to have a high alloreactive CTL-p frequency of 1:2 4 x 10’ against a n HLA identical sibling (S) (Kaminski et al, 1990). An experiment was set up in which the responder (R) cells were primed, either with allogeneic stimulator cells (S)or autologous cells, and. by means of split well analysis, tested for cytotoxicity against either stimulator (S) targets or autologous targets (Fig 2). There was no crossreactivity between the allo- and autoreactive CTL responses.
lo)80
RESULTS Autorytotoxicity in patients with SAA A group of 14 multitransfused patients with SAA were tested for autocytotoxicity. The auto CTL values ranged from 0%to 90% positive wells (Fig 1).All patients had been multitransfused ( > 100 donor units) and there was no correlation between auto CTL activity and whether or not the patient was transfusion dependent at the time of testing. Autocytotoxicity in multitransfused patients with other diseases In order to test whether the autocytotoxicity was induced by blood transfusion in patients with conditions other than SAA, five multitransfused patients with other haematological disorders (sickle cell anaemia. myelodysplastic syndrome or P-thalassaemia) and four patients transfused for acute blood loss, were assayed. The auto CTL values ranged from 10%to 66% positive wells (Fig 1) similar to those found in the transfused SAA patients. Autocytotoxicity in untransfused individuals In order to test further whether the autocytotoxicity was induced by blood transfusion alone or was also associated
I
t
*.
11-14 -transfused&+
n=5
n=4
0 0
Q
O
n-5
n-5
6
O
n-6
untransfused d
Fig 1, Comparison of autocytotoxicity in transfused and untransfused patients. SAA=severe aplastic anaemia, HAEM =other hae-
matological disorder (sickle cell anaemia, myelodysplastic syndrome and /I-thalassaemia). BLEED=acute blood loss (peptic ulceration, oesophagealvarices),NORMAL = normal individuals,CML = chronic myeloid leukaemia. Autocytotoxicity is significantly higher in transfused than in untransfused patients (P=O.OOOI). Mann-Whitney U test.
Blood Transfusion-induced Autocytotoxicity
R+S vs S’
R+S vs R’
R+R vs R’
R+R vs S’
Fig 2. Determination of specificity of allo- and autocytotoxic responses. Responder (R)cells were either primed with allogeneic stimulator ( S ) cells or autologous (K)cells and tested for cytotoxicity against allogeneic or autologous targets (*).
NIL
TCR
HLA I
HLA II
Fig 3. Monoclonal antibody inhibition studies. Autocytotoxicity was assayed in the absence (NIL) and presence of MoAbs to the clip T cell receptor ( ‘ K K ) . HLA class I (HLA I ) and HLA class II (HLA 11) molecules.
Mnrindoriul antibody inhibition studies One of the multitransfused SAA patients was found to have an autocytotoxic response of 3 7%. This autocytotoxicity was inhibited by an anti-ccip T cell receptor MoAb (100%). and both by an anti-class I(40x)and class I1 (65%) MoAb (Fig 3 ) .
DISC 11S S I O N In this study we have demonstrated a significantly higher degree of autocytotoxicity in multitransfused patients with SAA than in the untransfused controls ( P = 0.0001). Multitransfused patients with other diseases also gave high auto CTL activity, whilst untransfused patients with SAA gave low CTL activity in the same range as other untransfused individuals. This suggests that blood transfusion and not aplastic anaemia per SP is responsible for inducing autoreactivity. The possibility that this autocytotoxicity is in fact an alloreactive response against persisting transfusion de-
25
rived lymphocytes is unlikely since many of the ‘high responders’ had not been transfused for more than a year prior to the study. In one patient with a high auto CTL activity, a high alloreactive CTL-p frequency against a n HLA identical sibling had been previously demonstrated (Kaminski et a/, 1990). In order to dissect the alloreactive and autoreactive CTL responses a split-well cytotoxic assay was performed. This clearly demonstrated a lack of crossreactivity between the allo- and autoreactive CTL responses, suggesting that distinct cell populations are involved. Autocytotoxicity in SAA patients has been described previously by Warren et a/ ( I 980), who concluded that this was related to the aplasia and not an effect ofthe transfusion, This interpretation conflicts with the tindings presented here, where the low reactivity of the untransfused SAA patients suggests that transfusion is the cause of the autocytotoxicity. Consistent with this hypothesis was our finding of autoreactive cytotoxic cells in patients transfused for acute blood loss. The coexistence of autocytotoxicity with alloreactivity to HLA identical siblings was also described by Warren et a1 (1980). Similar results were found in a study after bone marrow transplantation, where non-overlapping populations of alloreactive and autoreactive CTLs were demonstrated (Hurme & Sihvola. 1985). Autocytotoxic cells have also been detected in patients with graft-versus-host disease following bone marrow transplantation (Parkman et a / . 1980: Santos, 1989). in patients with tuberculosis (Del Gallo et a / . 1990). in rheumatoid synovial fluid (Sugiyama et al, 1989) and in patients with vitiligo (Vignale et a/, 1988). The effector cells in our study appear to be cytotoxic T lymphocytes restricted both by HLA class I and I1 molecules. The target cells are likely to be T lymphocytes in view of the fact that PHA blasts were used in the assay. However, the antigen specificity of the autoreactive CTLs is unknown. One postulate is that the patient’s CTLs are recognizing foreign peptides, derived from transfused cells, situated in the MHC cleft of autologous cells. This may give rise to autocytotoxic responses. which in some cases may be due to CD8+ cells and in others to CD4+ cells, depending on whether the peptide is situated in a class I or class I1 MHC cleft. A second explanation is that these are responses of a regulatory or suppressor population which is inhibiting the ‘in vivo’ alloreactive response to HLA or minor histocompatibility antigens. Evidence for suppression of alloreactive T lymphocytes by anti-idiotypic cytolytic T cells in rat and murine graft-versus-host reactions was provided by Kimura & Wilson (1984) and Kosmatopoulos et a/ (1987). The mechanisms of anti-idiotypic suppression have been discussed by Batchelor et a1 (1989) and Wilson ( 1989). An hypothesis encompassing both the above explanations has recently been put forward by van Rood & Claas (1990) to explain the well-documented effect of pretransplant blood transfusions (PTBT) improving survival of renal allografts. They propose that CD4 + cells recognize allo-MHC peptides, of a specificity shared by the PTBT and allograft donor, derived from allogeneic cells and presented by autologous T cells. In this case the MHC molecules could be class I1 DR
26
Edward R. Kaminski et a1
molecules, which have been upregulated following activation of the autologous T lymphocytes by allogeneic stimulation. Such cells have been described (De Koster et al, 1989: Ottenhof& Mutis. 1990).The autoreactive CD4 cells then may downregulate the activated alloreactive cells. Although the functional significance of the autoreactive cytotoxic T cells remains to be determined, the following points should be made. (i) The presence of cellular autoreactivity, like autoreactive antibodies. does not necessarily lead to tissue damage. One reason for this may be that the lectin-activated target cells used in this assay have increased expression of HLA class I and class I1 antigens and adhesion molecules making them more sensitive to lysis, whereas in vivo the situation is quite distinct. (ii) The findings demonstrate that autocytotoxic cells can be generated in subjects who are not suffering from autoimmune disease. (iii) In CTLp analysis for bone marrow transplantation, autoreactive CTLs may exist and contribute to the overall CTLp frequency, thus emphasizing the need to include appropriate controls. In conclusion, we have demonstrated autocytotoxicity in multitransfused patients with SAA and other diseases and conclude that this is a n effect of transfusion and not associated with aplastic anaemia per se. Furthermore, we have demonstrated that the allo-non-MHC reactivity previously demonstrated in the same group of patients is not cross reactive with this autoreactivity. which suggests that different populations of cells are responsible for the two phenomena. The effector population appears to consist of cytotoxic T lymphocytes restricted both by HLA class I and I1 antigens. The number of patients in this preliminary study is small and further investigations will be necessary to confirm and extend our findings.
+
ACKNOWLEDGMENTS
We are grateful to Professor E. C. Gordon-Smith for supplying one of the patient samples and to Dr S. Fuggle for generously donating monoclonal antibodies. REFERENCES Batchelor. J.R.. Lombardi, G. & Lechler. R.I. (1989) Speculationson the specificity of suppression. Immunology Today, 10, 37-40. De Koster, H.S.. Anderson. D.C. & Terrnijtelen. A. (1989) T cells sensitized to synthetic HLA-DR3 peptide give evidence of continuous presentation of denatured HLA-DR3 molecules by HLA-DP. journal of Experimental Medicine. 1 6 9 , 1 19 1-1 196. Del Gallo. F.. Lombardi. G.. Piccolella.E.. Gilardini Montani. M.S.. Del Porto. P.. Pugliese. 0..Antonelli. G. & Colizzi, V. (1 990) Increased
autoreactive T cell frequencyin tuberculosis patients. International Archives of Allergy and Applied Immunology, 9 1 , 36-42. Hurme, M. & Sihvola, M. (1985)In vivo activation of autoreactive cytotoxic T lymphocytes after bone marrow transplantation: evidence for the prethymic specificity of T cells. journal of Immunology, 1 3 5 , 1108-1 112. Kaminski, E.. Sharrock. C., Hows, I., Ritter, M., Arthur, C.. Mackinnon, S. & Batchelor, J. (1988) Frequency analysis of cytotoxic T lymphocyte precursors-possible relevance to HLA-matched unrelated donor bone marrow transplantation. Bone Marrow Transplantation. 3, 149-1 55. Kaminski, E., Hows. J.. Man, S., Brookes. P., Mackinnon. S . , Hughes, T.. Avakian, 0..Goldman. J.M. & Batchelor.J.R. ( 1 989) Prediction of graft-versus-host disease by frequency analysis of cytotoxic T cells after unrelated donor bone marrow transplantation. Transplantation, 48. 608-61 3. Kaminski, E.R.. Hows. J.M.. Goldman, J.M. & Batchelor. J.R. (1990) Pretransfused patients with severe aplastic anaemia exhibit high numbers of cytotoxic T lymphocyteprecursors probably directed at non HLA antigens. British journal of Haematology. 76, 40 1-405. Kaminski, E.. Hows. J., Goldman. J. & Batchelor. R. ( 1 99 1 ) Optimising a limiting dilution culture system for quantitating frequencies of alloreactive cytotoxic T lymphocyte precursors. Cellular lmmunology. 1 3 7 , 88-95. Kimura, H. & Wilson, D.B. ( 1984) Anti-idiotypic cytotoxic T cells in rats with graft-versus-host disease. Nature. 308, 46 3-464. Kosmatopoulos, K..Algara, D.C. & Orbach-Arbouys.S. ( 1 987) Antireceptor anti-MHC cytotoxic T lymphocytes: their role in the resistance to graft versus host reaction. journal of Immunology, 1 3 8 , 1038-1041. Ottenhoff. T.H. & Mutis. T. ( 1 990) Specifickilling of cytotoxic T cells and antigen-presenting cells by CD4+ cytotoxic T cell clones. journal of Experimental Medicine, 1 7 1 , 201 1-2024. Parkman. R., Rappeport. J. & Rosen. F. ( 1 980) Human graft-versushost disease. journal of Investigative Dermatology, 74. 2 76-279. Santos. G.W. (1989) Syngeneic or autologous graft-versus-host disease. International journal of Cell Cloning, 7, 92-99. Sugiyama, E., Suzuki, H. & Yano, S. (1989) Autocytotoxic cells in rheumatoid synovial fluids. IournalofRheumatology. 16,276-279. Van Rood, J.J. & Claas, F.H.J.( 1 990) The influence of allogeneic cells on the human T and B cell repertoire. Science, 248, 1388-1 393. Vignale, R., Paciel. J. & Calandria. L. (1988)Immunological changes in common vitiligo. Medicina Cutanea Ibero-Latino-Americana. 16, 34 3-347. Warren. R.P.. Storb, R.. Thomas. E.D.. Su. P.J., Mickelson. E.M. & Weiden. P.L. ( 1 980) Autoimmune and alloimmune phenomena in patients with aplastic anemia: cytotoxicity against autologous lymphocytes and lymphocytes from HLA identical siblings. Blood. 5 6 , 683-689. Wilson, D.B. (1989) Idiotypic regulation of T cells in graft-versushost disease and autoimmunity.Immunological Reviews, 107. 1 59177.
