Immunobiology TCR cytotoxic T lymphocytes expressing the ... - Nature

11 downloads 0 Views 92KB Size Report
John Radcliff Hospital, Oxford, UK). EBV-LCL were gen- erated from PBMC ..... 15 Mingari MC, Vitale C, Cambiaggi A et al. Cytotoxic T lym- phocytes displaying ...
Bone Marrow Transplantation (2001) 27, 1087–1093  2001 Nature Publishing Group All rights reserved 0268–3369/01 $15.00 www.nature.com/bmt

Immunobiology TCR␥␦ cytotoxic T lymphocytes expressing the killer cell-inhibitory receptor p58.2 (CD158b) selectively lyse acute myeloid leukemia cells H Dolstra1, H Fredrix1, A van der Meer2, T de Witte1, C Figdor3 and E van de Wiel-van Kemenade1 Departments of 1Hematology, 2Bloodtransfusion and Transplantation Immunology and 3Tumor Immunology, University Medical Center St Radboud, Nijmegen, The Netherlands

Summary: Cytotoxic T lymphocytes (CTL) are thought to play an important role in the graft-versus-leukemia (GVL) response. Unfortunately, GVL reactivity is often associated with life-threatening graft-versus-host disease (GVHD). Characterization of CTL that selectively attack leukemic cells but not normal cells may lead to the development of adjuvant immunotherapy that separates GVL from GVHD. Here, we describe TCR␥␦ (V␥9/V␦1) CTL, isolated from the peripheral blood of an AML patient after stem cell transplantation (SCT), that very efficiently lysed freshly isolated acute myeloid leukemia (AML) cells and AML cell lines. Interestingly, HLA-matched non-malignant hematopoietic cells were not killed. We revealed that the killer cell-inhibitory receptor (KIR) p58.2 (CD158b) specific for group 2 HLA-C molecules negatively regulates the cytotoxic effector function displayed by these TCR␥␦ CTL. First, an antibody against HLA-C enhances lysis of nonmalignant cells. Secondly, stable transfection of HLACw*0304 into the class I-negative cell line 721.221 inhibited lysis. Finally, engagement of p58.2 by antibodies immobilized on Fc␥R-expressing murine P815 cells inhibits CD3- and TCR␥␦-directed lysis. Compared to non-malignant hematopoietic cells, AML cells express much lower levels of MHC class I molecules making them susceptible to lysis by p58.2+ TCR␥␦ CTL. Such KIR-regulated CTL reactivity may have a role in the GVL response without affecting normal tissues of the host and leading to GVHD. Bone Marrow Transplantation (2001) 27, 1087–1093. Keywords: stem cell transplantation; graft-versus-leukemia; ␥␦ T cell receptor; cytotoxic T lymphocytes; killer cell-inhibitory receptor; CD158b

Correspondence: Dr H Dolstra, Department of Hematology, University Hospital Nijmegen, Geert Grooteplein 8, PO Box 9101, 6500 HB Nijmegen, The Netherlands Received 3 January 2001; accepted 12 March 2001

Allogeneic stem cell transplantation (SCT) is an effective treatment for patients with leukemia since it causes an immune-mediated elimination of residual leukemic cells, termed graft-versus-host (GVL) reactivity.1,2 Unfortunately, the occurrence of graft-versus-host disease (GVHD) is a frequent and life-threatening complication.1,2 T cellmediated immunity is clearly involved in both the graftversus-host (GVH) and GVL response since T cell depletion of the allogeneic stem cell graft effectively reduces the occurrence of GVHD, but also results in an increased risk of recurrent leukemia.2 Moreover, infusion of T cells from the original stem cell donor to patients with leukemic relapse after allogeneic stem cell transplantation (SCT) induces complete remissions in the majority of patients with chronic myeloid leukemia (CML) and in about 30% of patients with acute myeloid leukemia (AML).3 However, most of the responding patients develop life-threatening GVHD. Antigen recognition by T cells is mediated by the T cell receptor (TCR) that exists in two basic heterodimeric forms, the ␣␤ and ␥␦ TCR. TCR␣␤ cytotoxic T lymphocytes (CTL) recognizing peptide antigens presented by MHC molecules on host leukemic cells have been isolated from SCT recipients and HLA-identical stem cell donors. Some of these TCR␣␤ CTL clones recognize antigens selectively expressed by leukemic cells that may induce only GVL reactivity against the tumor without affecting the normal cells of the host leading to GVHD.4–7 In addition, non-MHC-restricted TCR␥␦ CTL recognizing non-peptide ligands have also been shown to exert specific lysis of tumor cells.8–11 It has been shown that the selective lysis of tumor cells by TCR␥␦ CTL is regulated by killer cellinhibitory receptors (KIRs) expressed at their cell surface.12 These KIRs are a family of Ig-like molecules specifically binding to polymorphic determinants of MHC class I molecules. KIRs with two Ig domains bind to HLA-C allotypes. The KIR p58.1 (CD158a) recognizes an epitope shared by group 1 HLA-C allotypes (Cw2, 4, 5 and 6), whereas p58.2 (CD158b) recognizes a shared epitope of group 2 HLA-C allotypes (Cw1, 3, 7 and 8). A KIR with three Ig domains, p70, recognizes an epitope shared by HLA-Bw alleles and a homodimer of receptors with three Ig domains, p140, recognizes HLA-A3 and HLA-A11.13 The interaction between these KIRs on CTL and the appropriate MHC class

Selective CTL reactivity against leukemia H Dolstra et al

1088

I allotype on potential target cells inhibits TCR-mediated functions including cytotoxic activity and lymphokine production.14–19 In our search for cytotoxic effector cells that selectively recognize leukemic but not normal cells, we have investigated the specificity of T cell cultures generated from PBL of SCT recipients after in vitro stimulation with leukemic cells. Here, we report that a TCR␥␦ CTL clone expressing the p58.2 KIR kills AML cells very efficiently, but does not kill PHA-stimulated T cells and EBV-transformed B cells from the same patients. We show that protection of non-malignant hematopoietic cells from lysis by this TCR␥␦ CTL is mediated by p58.2 which specifically binds to group 2 HLA-C allotypes. We found that AML cells and AML cell lines express a much lower level of MHC class I molecules when compared to non-malignant cells and are, therefore, susceptible to lysis by p58.2+ TCR␥␦ CTL. Material and methods Antibodies and immunofluorescence analysis The following MoAb were used for immunofluorescence analysis, for cell sorting or for inhibition of cytotoxicity: TS2/18 (anti-CD2), SPV-T3b (anti-CD3), RIV6 (antiCD4), WT82 (anti-CD8), Leu19 (anti-CD56), 11F2 (antiTCR-␥␦), IMMU360 (anti-TCR-V␥9), ␦TCS1 (anti-TCRV␦1), IMMU389 (anti-TCR-V␦2), BMA031 (anti-TCR␣␤), W6/32 (anti-HLA-class I), Q5/13 (anti-HLA-DR/DP), B1.23.2 (anti-HLA-B/C), 289HA (anti-HLA-A1), BB7.2 (anti-HLA-A2), EB6 (anti-p58.1 or 50.1, CD158a), GL183 (anti-p58.2 or 50.2, CD158b), DX9 (anti-p70). Immunofluorescence was performed by the indirect method. FITCconjugated goat F(ab⬘)2 anti-mouse IgG and IgM (Tago Immunologicals, Camarillo, CA, USA) was used for staining followed by analysis by an Epics XL flowcytometer (Coulter Electronics, Hialeah, FL, USA). T and NK cell cultures T cells were isolated from the PBL of patient VE (a 53year-old male with AML) 4 months after HLA-identical SCT by cell sorting (Epics Elite flow cytometer, Coulter Electronics). A CTL line was established by stimulating T cells (5 × 105/ml) with irradiated leukemia cells from the patient (106/ml) in IMDM (Gibco BRL, Paisley, UK) supplemented with 10% pooled human serum. On day 7 and day 14, T cell cultures were restimulated with irradiated leukemia cells from the patient (106/ml) and 100 U/ml IL2 (Glaxo, Geneva, Switzerland) was added. From day 21 on, TCR␥␦ CTL cultures were expanded by weekly restimulation with an irradiated EBV-transformed lymphoblastoid cell line (EBV-LCL) of the patient pre-SCT (106/ml), irradiated Daudi cells (105/ml), and 100 U/ml IL2. p58.2 expressing NK cells were isolated by flow cytometry using MoAb GL183 and CD56. CD56+p58.2+ NK cells were expanded by stimulation with irradiated allogeneic PBL, EBV-LCL JY, 1 ␮g/ml PHA and 100 U/ml IL-2. After 14 days, NK cell cultures were used in cytotoxicity assays.

Bone Marrow Transplantation

Target cells Tumor cells were collected from AML patients at diagnosis. Cell lines Lama and Kasumi were derived from patients with AML.20,21 The stable transfectants 721.221.Cw*0304 and 721.221.Cw*0401 were kindly provided by Dr V Braud (Institute of Molecular Medicine, John Radcliff Hospital, Oxford, UK). EBV-LCL were generated from PBMC by transformation with EBV of the EBV-shedding B95–8 cell line and 0.1 ␮g/ml CsA. All cell lines were cultured in IMDM supplemented with 10% FCS. T cell blasts were generated by stimulating PBMC with 4 ␮g/ml PHA in IMDM plus 10% pooled human serum for 3 days. T cell blasts were washed and further cultured with 100 U/ml IL-2 for 3 days. Chromium release assay Specific lysis was determined by chromium release assays. Briefly, 106 target cells were incubated with 100 ␮Ci 51Cr (Amersham, Buckinghamshire, UK) for 1 h. Labeled target cells were mixed in V-bottom microtiter plates (103/well) with various numbers of effector cells in 150 ␮l IMDM with 10% FCS. After 4 h of incubation at 37°C, 100 ␮l supernatant of each sample was collected and radioactivity was determined by a gamma counter. HLA-Cw subtyping HLA-Cw typing was performed by PCR using sequencespecific primers (PCR-SSP). Primers were designed according to Bunce et al22 and extended with additional primers to increase the resolution. The PCR was modified to run under the conditions of the 12th IHW protocols for class I ARMS-PCR. The Cw*0707 and Cw*0709 alleles which possess the sequence for the Asn77, Lys80 p58 KIR motif are not detected by the primers specific for Cw*07 allotypes. The Cw*1507 allele containing the sequence for the Ser77, Asn80 p58 KIR motif is not detected by the primers specific for Cw*15 allotypes.

Results Isolation and characterization of a TCR␥␦ CTL selectively recognizing AML To examine whether CTL with selective anti-leukemic reactivity are present within the T cell repertoire of SCT recipients, we tested cytotoxic activity of T cell cultures generated from PBL by stimulation with irradiated leukemic cells of the patient. One of the T cell cultures obtained showed significant cytotoxicity against the patient’s AML cells (49% specific lysis; E:T ratio 10:1). Phenotypic analysis of this bulk culture revealed that 40% of the cells expressed the ␥␦ TCR, whereas the other 60% of the cells were CD4+TCR␣␤+. The TCR␥␦-expressing cells were purified by flow cytometry and expanded in the presence of irradiated EBV-transformed B cells of patient origin and the MHC class I-negative B cell line Daudi. The resulting T cell culture, named CTL VE, is positive for

Selective CTL reactivity against leukemia H Dolstra et al

a 100

Specific lysis (%)

1089

b

Target cell: AML VE

AML VE AML VH AML HB Lama Kasumi

75

medium anti-CD3 anti-TCRgd

50

anti-CD4 anti-CD8

25

0

anti-HLA-class I anti-HLA-DR/DP 0.3

1

3

10

0

E:T ratio

20

40

60

80

Specific lysis (%)

Figure 1 Specific cytotoxicity of TCR␥␦ CTL VE. (a) Cytotoxicity against AML cells of three patients and AML cell lines Lama and Kasumi. (b) Inhibition of cytotoxicity against AML cells of patient VE. Blocking studies were performed using purified MoAb (10 ␮g/ml) which was present throughout the assay. The effector:target cell ratio was 10:1.

CD2, CD3 and TCR␥␦, but does not express TCR␣␤, CD4 and CD8 (data not shown). Next, we tested the capacity of TCR␥␦ CTL VE to recognize freshly isolated leukemic cells and cell lines derived from AML patients. We observed that AML cells obtained from patient VE at diagnosis were very efficiently lysed by the CTL (Figure 1a). Moreover, AML cells isolated from two other patients and the AML cell lines Lama and Kasumi were also efficiently killed by CTL VE (Figure 1a). Lysis of AML cells was inhibited by anti-CD3 and anti-TCR␥␦ MoAb, but MoAb directed against CD4, CD8, HLA class I and HLA class II were ineffective (Figure 1b). These data indicate that CTL VE recognizes its target ligand through the TCR␥␦CD3 complex which results in strong cytolytic activity against freshly isolated myeloid leukemia cells and leukemic cell lines. The ␥␦ TCR of CTL VE is composed of V␥9 and V␦1 chains, as determined by staining with MoAb specific for these TCR V gene segments (data not shown). In the peripheral blood of healthy individuals the dominant TCR␥␦+ subset co-expresses V␥9 and V␦2 chains, and this population constitutes 70–90% of all TCR␥␦+ T cells.23 The Patient VE

100

TCR␥␦+ subset expressing the V␦1 chain paired with V␥9 or other V␥ chains is less frequent in the peripheral blood of healthy individuals. PBL of patient VE at 4 months after SCT, from which CTL VE is isolated, contained 5% TCR␥␦+ T cells of which 82% expressed the V␦1 gene segment and 30% the V␥9 gene segment (data not shown). Interestingly, this V␥9+/V␦1+ T cell subset represented 30% of all TCR␥␦+ T cells and 1.5% of all PBL. These results indicate that the V␥9V␦1 TCR␥␦+ subset showing strong anti-leukemic cytotoxicity in vitro is preferentially overrepresented in the peripheral blood of patient VE 4 months after SCT. Selective inhibition of cytotoxic activity of TCR␥␦ CTL VE To investigate whether TCR␥␦ CTL VE selectively recognizes leukemia cells, we compared its anti-leukemic reactivity with cytolytic activity mediated against HLAmatched PHA-stimulated T cells and EBV-transformed B cells derived from the AML patients. Interestingly, TCR␥␦ CTL VE did not lyse these non-malignant hematopoietic Patient VH

Patient HB

AML cells

Specific lysis (%)

PHA T blasts 75

EBV-LCL

50

25

0

0.3

1

3

10

0.3

1

3

10

0.3

1

3

10

E:T ratio Figure 2 Specific cytotoxicity of TCR␥␦ CTL VE against AML cells, PHA-stimulated T cell blasts, and EBV-transformed B cells derived from three leukemia patients. Bone Marrow Transplantation

Selective CTL reactivity against leukemia H Dolstra et al

0 0.1

TCR gd

1000

87

0 0.1

p58.1 (CD158a)

1000

Count

134

Count

98

Count

137

Count

1090

0 0.1

p58.2 (CD158b)

1000

0 0.1

p70

1000

Figure 3 Cell surface expression of killer cell inhibitory receptors of TCR␥␦ CTL VE.

cells (Figure 2). Recently, it has been shown that TCR␥␦ CTL express KIRs that regulate the cytotoxic effector function of these cells. Therefore, we analyzed CTL VE for expression of MHC class I-specific KIRs. CTL VE expressed p58.2, but p58.1 and p70 were not present (Figure 3). The p58.2 receptor binds to HLA-C allotypes that have Ser and Asn residues at positions 77 and 80 (ie group 2 HLA-C allotypes).12 HLA-C allelic typing of the AML patients and cell lines revealed that they all were positive for at least one HLA-C subtype with a Ser at position 77 and an Asn at position 80 (Table 1). High expression of these HLA-C subtypes at the cell surface of target cells may result in inhibition of cytotoxicity upon binding with p58.2. Therefore, we determined the level of expression of MHC class I molecules on AML cells and cell lines as well as on non-malignant target cells (Table 2). The results clearly show that freshly isolated AML cells have a significantly lower expression of MHC class I molecules when compared to PHA-stimulated T cells or EBVtransformed B cells. The AML cell lines Lama and Kasumi almost completely lack MHC class I expression. These results suggest that TCR␥␦ CTL VE exerts strong cytolytic activity against AML cells by lack of inhibition due to low expression of group 2 HLA-C molecules, whereas non-

malignant hematopoietic cells are protected from lysis due to high expression. To determine whether AML cells are generally more susceptible to non-MHC-restricted lysis than non-malignant hematopoietic cells, we tested these cells for lysis by two p58.2-expressing NK cell lines generated from PBL of two healthy donors (donor HD and FM). NK cell lines generated from both donors are CD3−, CD4−, CD8−, CD2+, CD56+, and express p58.2 KIR but do not express p58.1 and p70 (data not shown). NK cell line HD efficiently kills the MHC class I-negative AML cell lines Lama and Kasumi, and exerts significant lysis of freshly isolated AML cells but not of PHA-stimulated T cells generated from the same patients (Figure 4). This reactivity pattern is similar to that displayed by the TCR␥␦ CTL VE. NK cell line FM also efficiently kills the MHC class I-negative AML cell lines Lama and Kasumi, but freshly isolated AML cells are resistant to lysis by this NK cell line (Figure 4). As shown in Table 2, freshly isolated AML cells of patients VH and HB do not completely lack expression of all MHC class I alleles, in contrast to the AML cell lines Lama and Kasumi which are almost completely MHC class I negative. Therefore, the presence of inhibitory receptors other than p58.2 may be responsible for the partial or complete inhibition of

Table 1 HLA-C subtype and p58 KIR-related dimorphism of patients and cell lines

Table 2 get cells

HLA-Cw allelesa

p58 KIR-related dimorphism codon 77 and 80

VE

Cw*0304 Cw*1203

Ser77, Asn80 Ser77, Asn80

VH

Cw*0303 Cw*15

Ser77, Asn80 Asn77, Lys80

HB

Cw*07 Cw*15

Ser77, Asn80 Asn77, Lys80

Lama

Cw*0501 Cw*1203

Asn77, Lys80 Ser77, Asn80

Kasumi

Cw*0303 Cw*08

Ser77, Asn80 Ser77, Asn80

Patient

a

HLA-Cw typing was performed by PCR using sequence-specific primers, as described in Materials and methods. The Cw*0707 and Cw*0709 allotypes which possess the sequence for the Asn77, Lys80 p58 KIR motif are not detected by the primers specific for Cw*07 allotypes. The Cw* 1507 allotype containing the sequence for the Ser77, Asn80 p58 KIR motif is not detected by the primers specific for Cw*15 allotypes. All known Cw*08 allotypes possess the sequence for the Ser77, Asn80 p58 KIR motif.

Bone Marrow Transplantation

Patient

Cell surface expression of MHC class I by the different tar-

Cell surface expression ofa

Cell type

Total HLA class I

HLA-A1

HLA-A2 HLA-B/C

VE

AML PHA T blasts EBV-LCL

119 340 370

52 114 144

0 0 0

132 325 322

VH

AML PHA T blasts EBV-LCL

53 361 289

0 0 0

12 188 156

42 343 185

HB

AML PHA T blasts EBV-LCL

119 414 327

22 114 102

0 0 0

50 334 273

Lama

AML

10

0

2

5

Kasumi

AML

2

0

0

2

a Relative fluorescence intensity in arbitrary units. HLA class I types of patients and cell lines are: patient VE (HLA-A1, A31, B38, B51); patient VH (HLA-A2, A24, B51, B62); patient HB (HLA-A1, B8, B51); Lama (HLA-A2, A10, B12, B18); Kasumi (HLA-A10, A26).

Selective CTL reactivity against leukemia H Dolstra et al

TCR gd CTL VE

Target cells:

NK line HD

1091

NK line FM

AML VH AML HB Lama Kasumi PHA T blasts VH PHA T blasts HB 0

20

40

60

0

20

40

60

0

20

40

60

Specific lysis (%) Figure 4 Specific cytotoxicity of p58.2+ NK cell lines HD and FM against AML cells and PHA-stimulated T cell blasts derived from two leukemia patients The effector:target cell ratio was 10:1.

lysis of freshly isolated AML cells by the NK cell lines HD and FM, respectively. Inhibition of cytolytic activity of TCR␥␦ CTL VE by p58.2 engagement Next, we studied the involvement of p58.2 in inhibition of lysis by TCR␥␦ CTL VE in more detail. First, we tested the effect of anti-HLA-B/C MoAb on the lysis of nonmalignant hematopoietic cells. Addition of this antibody during the chromium release assay significantly restored lysis of PHA-stimulated T cells (Figure 5a). Secondly, CTL VE significantly lysed the MHC class I-negative cell line 721.221 either non-transfected or stably transfected with the HLA-Cw*0401 allotype (Asn77, Lys80 motif), whereas HLA-Cw*0304-transfected 721.221 cells (Ser77, Asn80 motif) are resistant to lysis (Figure 5b). Finally, we determined whether triggering of p58.2 influences TCR/CD3directed killing by CTL VE. Binding of anti-CD3 or antiTCR␥␦ MoAb to Fc␥ receptors on P815 cells induces killing of these target cells by CTL VE (Figure 6). Anti-p58.2 MoAb bound to the Fc␥ receptors of P815 target cells inhibits both CD3 and TCR␥␦ MoAb-directed killing, whereas anti-p58.1 MoAb has no effect (Figure 6). These

60

a

60

medium

b

Specific lysis (%)

anti-HLA B/C 40

721.221

results clearly demonstrate that non-malignant hematopoietic cells do express the antigen recognized by the ␥␦ TCR of CTL VE, but engagement of p58.2 by group 2 HLA-C molecules highly expressed on these target cells results in protection from lysis. Discussion CTL reactivity is thought to be important in the elimination of residual leukemic cells after allogeneic SCT. Unfortunately, the occurrence of life-threatening GVHD is strongly associated with this anti-leukemic CTL reactivity. TCR␣␤ CTL have been isolated from the peripheral blood of SCT recipients that recognize antigens selectively expressed by leukemic cells.4–7 Such leukemia-reactive CTL can exert specific anti-leukemic cytotoxicity while they do not attack normal cells and lead to GVHD. The present report describes a leukemia-specific CTL expressing a V␥9/V␦1 TCR. This CTL clone efficiently kills AML cells, whereas HLA-matched PHA-stimulated T cells and EBV-transformed B cells generated from the same leukemia patients are not lysed. We revealed that the leukemia-restricted cytotoxicity displayed by this TCR␥␦ CTL is regulated by the p58.2 KIR. Engagement of p58.2 results in an inhibitory signal that predominates over TCR␥␦CD3 complexmediated activation of the CTL. Therefore, PHA-stimulated T cells and EBV-transformed B cells that express high lev-

721.221.Cw3

anti-HLA DR/DP

40

721.221.Cw4

Target cell: Fcg R+ p815 medium

20

20

0

0

anti-CD3 anti-CD3 + anti-p58.2 anti-CD3 + anti-p58.1 anti-TCRgd

0.3

1

3

10

0.3

1

3

10

E:T ratio Figure 5 Specific cytotoxicity of p58.2+ TCR␥␦ CTL VE. (a) Specific lysis was tested against PHA-stimulated T cell blasts in the absence or presence of anti-HLA-B/C and anti-HLA-DR/DP. Both MoAb are of the IgG1 isotype and were used at a final concentration of 10 ␮g/ml. One representative experiment out of three is shown. (b) Specific lysis against the MHC class I-deficient cell line 721.221, either non-transfected or stably transfected with HLA-Cw*0304 (Ser77, Asn80 p58 KIR motif) or HLA-Cw*0401 (Asn77, Lys80 p58 KIR motif).

anti-TCRgd + anti-p58.2 anti-TCRgd + anti-p58.1 0

20

40

60

80

Specific lysis (%) Figure 6 Inhibition of anti-CD3 and anti-TCR␥␦ MoAb-directed killing by Fc␥ receptor-immobilized anti-p58.2 MoAb on murine P815 cells. Specific cytotoxicity against the murine P815 cell line was assayed after binding of the indicated MoAb to Fc␥ receptors expressed by these cells. Antibodies were used at a final concentration of 10 ␮g/ml. The effector:target cell ratio was 10:1. One representative experiment out of two is shown. Bone Marrow Transplantation

Selective CTL reactivity against leukemia H Dolstra et al

1092

els of group 2 HLA-C molecules trigger the p58.2 receptor resulting in inhibition of lysis, whereas low levels of HLAC molecules as expressed on AML cells do not. Consequently, AML cells are very efficiently lysed by the p58.2+ TCR␥␦ CTL. These findings indicate that cytolytic activity of TCR␥␦+ T cells recognizing broadly distributed antigens can be controlled by the cell surface expression of KIRs, which confirm and extend results of other investigators.17–19 Tumor cell lysis by either TCR␥␦+ T cells and NK cells is clearly dependent on decreased levels of MHC class I molecules on tumor cells.24,25 Lack or down-regulation of MHC class I expression on leukemic cells has been reported.26 We observed that the level of MHC class I expression of freshly isolated AML cells and AML cell lines was much lower than the expression on PHA-stimulated T cells and EBV-transformed B cells. This can result in the loss of protection from non-MHC-restricted lysis by TCR␥␦+ T cells and NK cells. We showed that AML cells and AML cell lines are susceptible to non-MHC-restricted lysis by p58.2+ TCR␥␦ CTL as well as p58.2+ NK cells. Albi et al27 have reported that an increased number of p58.2+ TCR␣␤+CD8+ T cells displaying non-MHC-restricted lysis in vitro was detected in SCT recipients of HLAmismatched bone marrow. These KIR expressing TCR␣␤+ T cells lysed patients’ leukemic cells, whereas non-malignant cells were protected from lysis. Also, it has been reported that p58.2 (CD158b)+ T cells increase after HLAmatched allogeneic SCT.28 Cambiagi et al29 showed that transgenic expression of the CD158b KIR prevents in vivo rejection of an H-2-mismatched mice bone marrow graft. Together, these data suggest that CD158b+ T cells and NK cells displaying non-MHC-restricted lysis may have some role in selective anti-leukemic reactivity following SCT, while normal cells are protected from lysis thereby reducing GVHD. Leukemic relapse after allogeneic SCT is a serious problem. A strategy to augment GVL reactivity without causing GVHD may be to generate CTL with specific anti-leukemic reactivity from allogeneic donors in vitro and adoptively transfer these CTL to BMT recipients with relapsed leukemia. Such CTL directed against leukemia-specific antigens will not exert GVHD. Alternatively, CTL recognizing antigens with a broad tissue distribution and that express KIRs may also selectively exert GVL reactivity. The expression of KIRs on in vitro generated anti-leukemic CTL may be induced by certain cytokines or by transferring KIR genes into CTL employing retroviral vectors that can safely be administered to SCT recipients with relapse.30–32 Acknowledgements We thank Dr J Melenhorst, Dr A Bakker and Dr P Coulie for generously providing MoAb and Dr V Braud for kindly providing the 721.221 Cw*0304 and Cw*0401 stable transfectants. This work was supported by grants from the Dutch Cancer Society (KUN 97–1508).

References 1 Butturini A, Bortin MM, Gale RP. Graft-versus-leukemia following bone marrow transplantation. Bone Marrow Transplant 1987; 2: 233–242. Bone Marrow Transplantation

2 Horowitz MM, Gale RP, Sondel PM et al. Graft-versus-leukemia reactions after bone marrow transplantation. Blood 1990; 75: 555–562. 3 Kolb HJ, Schattenberg A, Goldman JM et al. Graft-versusleukemia effect of donor lymphocyte transfusions in marrow grafted patients. Blood 1995; 86: 2041–2050. 4 Sosman JA, Oetel KR, Smith SD et al. Specific recognition of human leukemic cells by allogeneic T cells: II. Evidence for HLA-D restricted determinants by leukemic cells that are crossreactive with determinants by unrelated nonleukemic cells. Blood 1990; 75: 2005–2016. 5 Van Lochem E, De Gast B, Goulmy E. In vitro separation of host graft-versus-host and graft-versus-leukemia cytotoxic T cell activities. Bone Marrow Transplant 1992; 10: 181–183. 6 Faber LM, van Luxemburg-Heijs SAP, Willemze R, Falkenburg JHF. Generation of leukemia-reactive cytotoxic T lymphocyte clones from the HLA-identical bone marrow donor of a patient with leukemia. J Exp Med 1992; 176: 1283–1289. 7 Dolstra H, Fredrix H, Preijers F et al. Recognition of a B cell leukemia-associated minor histocompatibility antigen by CTL. J Immunol 1997; 158: 560–565. 8 Zocchi MR, Ferrarini M, Rugarli C. Selective lysis of the autologous tumor by ␦TCS1+ ␥␦ tumor infiltrating lymphocytes from human lung carcinoma. Eur J Immunol 1990; 20: 2685–2689. 9 Zocchi MR, Ferrarini M, Migone N, Casorati G. T-cell receptor V␦ gene usage by tumour reactive ␥␦ T lymphocytes infiltrating human lung cancer. Immunology 1994; 81: 234–239. 10 Choudhary A, Davodeau F, Moreau A et al. Selective lysis of autologous tumor cells by recurrent ␥␦ tumor-infiltrating lymphocytes from renal carcinoma. J Immunol 1995; 154: 3932–3940. 11 Maeurer MJ, Martin D, Walter W et al. Human intestinal V␦1+ lymphocytes recognize tumor cells of epithelial origin. J Exp Med 1996; 183: 1681–1696. 12 Moretta A, Biassoni R, Bottino C et al. Major histocompatibility complex class I-specific receptors on human natural killer and T lymphocytes. Immunol Rev 1997; 155: 105–177. 13 Pende D, Biassoni R, Cantoni C et al. The natural killer cell receptor specific for HLA-A allotypes: a novel member of the p58/p70 family of inhibitory receptors that is expressed as a 140-kD disulphide-linked dimer. J Exp Med 1996; 184: 505–518. 14 Ferrini S, Cambiaggi A, Meazza R et al. T cell clones expressing the natural killer-cell related p58 receptor molecule display heterogeneity in phenotypic properties and p58 function. Eur J Immunol 1994; 24: 2294–2298. 15 Mingari MC, Vitale C, Cambiaggi A et al. Cytotoxic T lymphocytes displaying natural killer (NK)-like activity: expression of NK-related functional receptors for HLA class I molecules (p58 and CD94) and inhibitory effect on the TCRmediated target cell lysis or lymphokine production. Int Immunol 1995; 7: 697–703. 16 Phillips JH, Gumperz JE, Parham P, Lanier LL. Superantigendependent, cell-mediated cytotoxicity inhibited by MHC class I receptors on T lymphocytes. Science 1995; 268: 403–405. 17 D’Andrea A, Chang C, Philips JH, Lanier LL. Regulation of T cell lymphokine production by killer cell inhibitory receptor recognition of self HLA class I alleles. J Exp Med 1996; 184: 789–794. 18 Halary F, Peyrat MA, Chanpagne E et al. Control of selfreactive cytotoxic T lymphocytes expressing ␥␦ T cell receptors by natural killer inhibitory receptors. Eur J Immunol 1997; 27: 2812–2821. 19 Nakajima H, Moore YI, Takiguchi M. Functional expression of the NKB1 killer cell inhibitory receptors on a ␥␦ CTL clone. Immunogenetics 1997; 45: 226–228.

Selective CTL reactivity against leukemia H Dolstra et al

20 Seigneurin D, Champelovier P, Mouchiroud G et al. Human chronic myeloid leukemic cell line with positive Philadelphia chromosome exhibits megakaryocytic and erythroid characteristics. Exp Hematol 1987; 15: 822–832. 21 Asou H, Tashiro S, Hamamoto K et al. Establishment of a human acute myeloid leukemia cell line (Kasumi-1) with 8;21 chromosome translocation. Blood 1991; 77: 2031–2036. 22 Bunce M, Welsh KI. Rapid DNA typing for HLA-C using sequence-specific primers (PCR-SSP): identification of serological and non-serologically defined HLA-C alleles including several new alleles. Transplantation 1994; 43: 7–17. 23 Porcelli S, Brenner MB, Band H. Biology of the human gamma delta T-cell receptor. Immunol Rev 1991; 120: 137– 183. 24 Bakker AB, Phillips JH, Figdor CG, Lanier LL. Killer cell inhibitory receptors for MHC class I molecules regulate lysis of melanoma cells mediated by NK cells, gamma delta T cells, and antigen-specific CTL. J Immunol 1998; 160: 5239–5245. 25 Pende D, Accame L, Pareti L et al. The susceptibility to natural killer cell-mediated lysis of HLA class I-positive melanomas reflects the expression of insufficient amounts of different HLA class I alleles. Eur J Immunol 1998; 28: 2384–2394. 26 Dermime S, Mavroudis D, Jiang YZ et al. Immune escape from a graft-versus-leukemia effect may play a role in the relapse of myeloid leukemias following allogeneic bone marrow transplantation. Bone Marrow Transplant 1997; 19: 989–999.

27 Albi N, Ruggeri L, Aversa F et al. Natural killer (NK)-cell function and antileukemic activity of a large population of CD3+/CD8+ T cells expressing NK receptors for major histocompatibility complex class I after ‘three-loci’ HLA-incompatible bone marrow transplantation. Blood 1996; 87: 3993–4000. 28 Tanaka J, Mori A, Ohta S, Yamamoto Y et al. Expression of HLA-C-specific natural killer cell receptors (CD158a and CD158b) on peripheral blood mononuclear cells after allogeneic bone marrow transplantation. Br J Haematol 2000; 108: 778–783. 29 Cambiaggi A, Verthuy C, Naquet P et al. Natural killer cell acceptance of H-2 mismatch bone marrow grafts in transgenic mice expressing HLA-Cw3 specific killer cell inhibitory receptor (CD158b). Proc Natl Acad Sci USA 1997; 94: 8088–8092. 30 Mingari MC, Ponte M, Bertone S et al. HLA class I-specific inhibitory receptors in human T lymphocytes: interleukin 15induced expression of CD94/NKG2A in superantigen- or alloantigen-activated CD8+ T cells. Proc Natl Acad Sci USA 1998; 95: 1172–1177. 31 Mingari MC, Moretta A, Moretta L. Regulation of KIR expression in human T cells: a safety mechanism that may impair protective T-cell responses. Immunol Today 1998; 19: 153–157. 32 Bonini C, Ferrari G, Verzeletti S et al. HSV-TK gene transfer into donor lymphocytes for control of allogeneic graft-versusleukemia. Science 1997; 276: 1719–1724.

1093

Bone Marrow Transplantation