Leukemia (2002) 16, 2259–2266  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Analysis of donor NK and T cells infused in patients undergoing MHC-matched allogeneic hematopoietic transplantation V Pascal1,2, C Brunet2, V Pradel4, X Thirion4, P Andre6, C Faucher3, J Sampol2, F Dignat-George2, DBlaise 3, E Vivier1,2,5 and C Chabannon3

1Centre d’Immunologie INSERM/CNRS de Marseille-Luminy, Marseille, France; 2Laboratoire d’He´matologie, Hoˆpital de la Conception, Marseille, France; 3Institut Paoli-Calmettes, Centre Re´gional de Lutte Contre le Cancer Provence-Alpes-Coˆte d’Azur, Marseille, France; 4De´partement d’Information Me´dicale, Hoˆpital de Sainte Marguerite, Marseille, France; 5Institut Universitaire de France; and 6Innate Pharma, Marseille, France

We retrospectively analyzed the percentages and absolute transplantation, injection of ex vivo activated and expanded numbers of T cells, natural killer (NK) cells and NK cell subsets ␥␦ T cells has been demonstrated to facilitate engraftment.10 in cryopreserved samples of either bone marrow or blood non- T cell-depleted allogeneic MHC-matched hematopoietic grafts. A role of NK cells in transplantation was initially described Using flow cytometry, we found higher numbers of NK cells in in ‘hybrid resistance’, in which F1 hybrid recipients reject par- aphereses than in bone marrow collections. We further investi- ental bone marrow grafts, while accepting solid tissues.11,12 gated the distribution of NK cell subsets, defined by the cell More recently, in a murine model of MHC-incompatible allo- surface expression of MHC class I-specific receptors, in these geneic bone marrow transplantation, Murphy et al showed allogeneic grafts. The distribution of NK cell subsets from the that infusion of in vitro IL-2 activated NK cells could prevent two different origins were similar, with the exception of the CD158a/h+ NK cell subset, whose size appeared to be smaller the development of graft-versus-host disease (GVHD), exert an in bone marrow. The search for relations between the numbers anti-tumoral effect against a co-transplanted tumor cell line, of infused cells and post-transplantation events demonstrated and increase survival in recipient mice.13–15 The molecular that increasing numbers of infused T cells but not NK cells are basis for these effects mediated by NK cells in transplantation related with decreased overall survival. Our study highlights are being unraveled. Indeed, NK cell activity is controlled by the toxicity of infused T cells but not NK cells in allogeneic a dynamic balance between engagement of activating recep- MHC-matched hematopoietic grafts. These data pave the way tors and MHC class I-specific inhibitory NK receptors (NKR) for further trials to investigate the effect of NK cell infusion in 16–18 MHC-matched allogeneic transplantation, and in particular expressed on their surface. In humans, MHC class I-spe- whether ex vivo NK cell expansion and activation may enhance cific NKR belong to two distinct groups of molecules: killer the anti-tumoral effect of the procedure and decrease its cell immunoglobulin (Ig)-like receptors (KIR) are members of a morbidity. multigenic and multiallelic family of Ig-like molecules, while Leukemia (2002) 16, 2259–2266. doi:10.1038/sj.leu.2402670 CD94/NKG2 heterodimers belong to the C-type lectin recep- Keywords: NK cells; T cells; MHC-matched allogeneic transplan- tation; cell therapy tor family. KIR primarily act as receptors for classical MHC class Ia molecules, whereas CD94/NKG2 heterodimers recog- nize the non-classical MHC class Ib molecules, HLA-E.16,18 NKR act as inhibitory receptors when they possess an immu- Introduction noreceptor tyrosine-based inhibitory motif (ITIM) in their intra- cytoplasmic domain. Similarly to NKR, human NK cell activ- Allogeneic cell transplantation has now been evaluated for ating receptors belong to the Ig-like superfamily for the natural three decades,1 in a variety of diseases, including malignant cytotoxicity receptors (NCR: NKp30, NKp44, NKp46), or to and non-malignant disorders. Most of the therapeutic efficacy the lectin-like family (NKG2D). NCR associate with immuno- of allogeneic cell transplantation results from the recognition receptor tyrosine-based activating motif (ITAM) polypeptides of host tumor cells by donor-derived immune effectors, ie the (CD3␨, FcR␥, KARAP/DAP12), whereas NKG2D interact with graft-versus-leukemia (GVL) effect, initially demonstrated in the DAP10 transducing polypeptide.19 NCR and NKG2D patients with leukemias or other types of blood malig- interact with ligands that are upregulated upon cellular stress nancies.2–4 Recently, additional evidence has been reported, and tumor transformation (eg MICA for NKG2D),20,21 or virus suggesting that a similar anti-tumor effect can participate in infection (eg viral hemagglutinins for NKp46).22 Although the cure of poor-prognostic neoplasias of epithelial origin, NCR represent the major receptors for NK cell cytotoxicity, such as breast5 or renal cell carcinomas.6 Precise identifi- other molecules such as CD2, CD69, CD44, CD38 and cation of cellular effectors that are responsible for the GVL CD244 also enhance NK cell killing.19 effect remains to be established. However, T cells play an The importance of NK cells and inhibitory NKR engagement important role in GVL, as demonstrated by the increased inci- by MHC class I molecules has been demonstrated in a number dence of relapse following T cell depletion of the graft in of mouse and human models of MHC-mismatched allogeneic vitro,7 T cell anergisation in vitro,8 or increased immuno- transplantation. Indeed, the rejection of parental bone marrow suppression in vivo.9 A number of arguments suggest that cell graft by a F1 hybrid can be abrogated by the transgenic populations other than ␣␤ T cells can also mediate some of expression of HLA-Cw3 and the cognate KIR (CD158b) in the the clinical events frequently observed after allogeneic trans- parental donor and F1 recipient, respectively.23 Along this plantation, including the GVL effect. Along these lines, in a line, the treatment of leukemia-bearing mice with F(abЈ) frag- murine model of MHC-incompatible allogeneic bone marrow 2 ments of monoclonal antibody (mAb) against Ly49C and I, as well as adoptive transfer of NK cells treated ex vivo with the same F(abЈ) fragments, resulted in significant increases in Correspondence: E Vivier, Centre d’Immunologie INSERM/CNRS de 2 Marseille-Luminy, Case 906, Parc Scientifique de Luminy, 13288 mouse survival, demonstrating that blockade of inhibitory Marseille cedex 09, France; Fax: (33) +491269430 MHC class I-specific NKR enhances anti-tumor activity in Received 11 March 2002; accepted 31 May 2002 vivo.24 In both humans and mice, the presence of KIR/MHC T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2260 class I incompatibility (KIR epitope incompatibility) between Table 1 Donor and patient characteristics donors and recipients of HLA-mismatched allogeneic trans- plantation, ie when the donor KIR repertoire fails to recognize Number 27 the HLA epitopes of the recipient, was recently demonstrated Age in years: median (range) Donors 41 (13–62) to be associated with anti-leukemic activity of donor-derived 25–27 Recipients 40 (16–60) NK cell clones. However, in humans, most allogeneic Sex match (donor/recipient) hematopoietic transplantations are performed from pheno- M/M 8 identical or geno-identical donors. Yet, in such situations of M/F 7 MHC-matched transplantation, a role for NK cells in GVL has F/F 4 been also documented. Indeed, IL-2-activated NK cells F/M 8 28 Diagnosis infused in syngeneic mouse exert a significant GVL effect. Aplastic anemia 2 Moreover, it has recently been demonstrated, in vivo in the Acute lymphoblastic leukemia 7 mouse, that NK cells can reject tumor cell lines expressing Acute myelogenous leukemia 8 syngeneic MHC class I molecules.29,30 Chronic myelogeneous leukemia 2 In this study, we analyzed the NK cell and T cell compart- Myelodysplastic syndrome 2 ments present in hematopoietic bone marrow grafts and aph- Hodgkin’s disease 1 Non-Hodgkin’s lymphoma 4 ereses which were infused in adult patients undergoing MHC- Non-hematopoietic disease 1 matched and non-T cell-depleted allogeneic transplantation. Conditioning regimen Myeloablative 13 Non-myeloablative 14 Methods Graft Bone marrow 18 Aphereses 9 Patients and donors Engraftment Bone marrow 18/18 Twenty-seven patients received a MHC-matched non-T cell- Aphereses 9/9 depleted allogeneic transplant, according to standard eligi- Acute GVHD bility criteria and current protocols at Institut Paoli-Calmettes Grade 0—I (IPC). Informed consent was obtained from donors and recipi- Bone marrow 11/18 Aphereses 3/9 ents, prior to inclusion in protocols. Grafts were from geno- Grade II—IV identical donors for 24 patients, from a syngeneic donor for Bone marrow 7/18 one patient, and from pheno-identical donors for two patients. Aphereses 6/9 Serological typing was used for MHC class I molecules, and Chronic GVHD molecular typing was used for MHC class II molecules. Bone marrow 2/14 Patients were reinfused either with allogeneic bone marrow Aphereses 3/6 (n = 18), or with allogeneic peripheral blood progenitor cells (aphereses, n = 9). In the latter situation, donors were mobil- ized with rhG-CSF prior to apheresis, as previously described.31,32 Conditioning regimens and GVHDprophylaxis were as per institutional protocols, depending on diagnosis Table 2 mAb used in this study and disease status. Donor and patient characteristics and major clinical events are described in Table 1. mAb NKR specificity

CD designation Others designations Graft characterization EB6a CD158a KIR2DL1 p58.1 For all donor/patient pairs, cells were harvested and processed CD158h KIR2DS1 p50.1 according to standard procedures at the IPC Cell Therapy GL183a CD158b1 KIR2DL2 p58.2 Facility. Freshly collected grafts were analyzed for their con- CD158b2 KIR2DL3 p58.3 tent in nucleated cells, hematopoietic progenitors (CD34+ CD158j KIR2DS2 p50.2 cells) and lymphoid subpopulations (CD3+ cells) by flow cyto- ZIN276a CD158e1 KIR3DL1 p70 metry, using standard procedures.33 For regulatory reasons, CD158e2 KIR3DS1 — HP3B1 CD94 — — and as per institutional protocols, 1 ml samples of the cell Z199 CD159a NKG2A — suspension were cryopreserved in 2 ml cryotubes, and kept in the vapor phase of liquid nitrogen. The study described aAnti-KIR mAb do not discriminate inhibitory from activating below was performed on cells recovered from these thawed isoforms.34,50 samples. It was first validated by comparing CD3+ cell counts obtained before and after the freeze–thaw process. These values were not statistically different, and further analyses were conducted with thawed graft samples. (IgG1, GL183), anti-CD158e1/e2 (IgG1, ZIN276), anti-CD94 (IgG2a, HP3B1), anti-CD159a (IgG2b, Z199),34 FITC-conju- Antibodies gated anti-CD3 (IgG1, UCHT1), PC5-conjugated anti-CD16 (IgG1, 3G8) and CD56 (IgG1, NKH.1). These mAb and their The following mouse mAb were used in this study (Table 2): isotype-matched negative control mAb were purchased from PE-conjugated anti-CD158a/h (IgG1, EB6), anti-CD158b1/b2/j Beckman Coulter Immunotech (Marseille, France).

Leukemia T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2261 Flow cytometric analysis of lymphocyte populations grafts, the incidence and severity of chronic GVHDwere increased, so that benefits and disadvantages appear to be bal- Cryopreserved graft samples were analyzed with three-color anced and the role of allogeneic aphereses as an alternative flow cytometry to determine the percentages of total T cells, to bone marrow has not been clearly established yet. In this total NK cells, as well as NK cell subsets in the grafts. Briefly, study, apheresis and bone marrow allogeneic grafts were thawed hematopoietic cells were incubated with the indicated characterized for their content in hematopoietic progenitors PE-labeled anti-NKR, FITC-labeled anti-CD3 and the PC5- (CD34+ cells), as well as in T lymphocytes and NK cells, using labeled anti-CD16 and anti-CD56 mAb, in PBS supplemented flow cytometric analysis. Absolute cell numbers were further with 1% BSA. The pelleted cells were analyzed with an EPICS expressed per kilogram of recipient weight (cell doses). As XL II cytometer (Beckman-Coulter, Margency, France). The shown in Table 3, grafts of bone marrow or blood origin differ analysis was focused on lymphocytes, identified by their for- significantly by their contents in hematopoietic progenitors, ward and right angle scatter features. At least 10 000 events T lymphocytes and NK cells, an observation consistent with were collected in the lymphocyte gate and analyzed. NK cells previous studies.32,37 Indeed, all cell doses infused within aph- were defined as CD3−CD16+CD56+ cells. Five distinct NK cell ereses were significantly higher than those within bone mar- subsets were defined as NK cells expressing at least one of the row samples (Table 3). In particular, patients infused with aph- following NKR: CD94, CD159a, CD158a/h, CD158b1/b2/j or eresis received on average 20-fold more NK cells than patients CD158e1/e2. Absolute numbers of infused cells per kg were infused with bone marrow transplants. computed from the total numbers of viable mononuclear cells, the percentages of lymphocytes in the graft sample, and the recipient weight (cell dose). Phenotypic analysis of NK cell subsets in hematopoietic transplants Statistical analyses

Graphic representation of the NKR cell surface expression on NK cells from bone marrow as well as from apheresis are + NK cells obtained from aphereses and bone marrow trans- mostly CD16 CD56dim NK cells (Figure 1a and b). However, plants, as well as the descriptive statistics of these cell com- as previously described in peripheral blood NK cells, two dis- partments were obtained by using Prism Graphpad software tinct subsets can be identified in graft samples, according to (San Diego, CA, USA). Comparison of distributions was done their CD56 surface expression, ie CD56dim and CD56bright NK with a Mann–Whitney U test. cells. These phenotypic correlates suggest that NK cells The relation between the infused doses of NK and T cells, infused within allogeneic transplantation harbor a heterogen- and post-transplantation clinical events, ie the development eity similar to that observed in peripheral blood of healthy of acute GVHDand overall survival, was studied using multi- donors.38 variate analysis. We used a proportional hazard model (Cox We further investigated the NKR repertoire expressed on NK model) to analyze the relation between these variables and cells obtained from apheresis and bone marrow, using three- the post-transplantation survival. Logistic regression was used color flow cytometry (Figures 2 and 3). This analysis shows to study the relation with acute GVHD. All data were com- that bone marrow collections and aphereses contain NK cells puted using SPSS for Windows software (SPSS, Chicago, IL, that express the CD94/NKG2A-B heterodimers and KIR mol- USA). Multivariate analyses included CD34+ cell dose, log (T ecules. In both types of graft, KIR molecules were expressed + cell dose) and log (NK cell dose), or CD34+ cell dose, log (T only on CD16 CD56dim NK cells (data not shown), as pre- cell dose) and log (NK subset cell dose). A logarithmic trans- viously described for unmanipulated peripheral blood lym- formation was used for T, NK and NK cell subset doses, phocytes.39 The distribution of NK cell subsets expressing because of the log normal distribution of these variables. As CD94/NKG2A-B heterodimers and KIR molecules did not dif- all cell doses were significantly different depending on the fer significantly between either type of graft, with the excep- + nature of the graft (bone marrow vs aphereses), we added tion of the CD158a/h NK cell subset whose size was smaller these data as an indicator variable. The role of T cells on sur- in bone marrow. vival was further investigated with Kaplan–Meier analysis: two groups of patients were arbitrarily clustered through the median value of the T cell dose and overall survival of the two groups was compared using a log-rank test. Table 3 Hematopoietic cell doses infused in allogeneic recipients

Study group Hematopoietic cell doses (106/kg)a Results CD34+ cellb CD3+ T cellb NK cellb NK cells in apheresis and bone marrow hematopoietic transplants Bone marrow n = 18 2 ± 0.2 28 ± 33± 0.7 (0.8–4.1) (12–56) (0.6–13) Availability of clinical grade and recombinant hematopoietic Aphereses n = 97± 2.3 296 ± 47 54 ± 7 growth factors has led to the evaluation of apheresis as an (1.1–24.9) (103–505) (17–90) alternative to bone marrow in hematopoietic transplantation. a ± Several reports have recently been published, demonstrating Results are expressed as mean s.e.m.; ranges are indicated in parentheses. that similarly to previous observations in the autologous set- bStatistically significant differences (P Ͻ 0.001) between values ting, the use of allogeneic blood cells in place of bone marrow obtained from bone marrow grafts and apheresis. Apheresis cells resulted in faster neutrophil and platelet recovery.32,35–37 samples contained 13 ± 2% NK cells and bone marrow grafts con- However, because of the higher immune cell content in blood tained 6 ± 0.5% NK cells.

Leukemia T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2262

Figure 1 Phenotypic analysis of frozen graft samples. Representative phenotypic characterization of NK cells as CD3−CD16+CD56+ lympho- cytes in apheresis (a) and bone marrow (b) frozen allogeneic samples. The numbers indicate the percentages of cells in each quadrant; expression of CD16 and CD56 on NK (CD3−) and T (CD3+) cells in apheresis samples (a) appears to fall within ‘normal ranges’ measured on 49 peripheral blood samples from normal volunteers (data not shown).

Incidence of NK and T cell doses on post- includes both recipients of aphereses and recipients of bone transplantation events marrow cells.

Using multivariate analysis, we investigated the relation between NK and T cell doses, and the occurrence of post- Discussion transplantation events (Table 4). The relation between NK cell doses (total NK cells or NK cell subsets) or T cell doses and The present analysis documents the effects of donor T and acute GVHDwas not demonstrated in this study. The role of NK cells present in hematopoietic grafts, in the clinical events NK cells in the development of chronic GVHDcould not be which follow allogeneic MHC-matched transplantation. statistically explored because only five of the patients who Despite the limited number of donor/recipient pairs studied, were alive 100 days after transplantation developed chronic our data show that infusion of increasing doses of donor T GVHD. cells is deleterious for post-transplantation survival. The selec- While the relation between T cell depletion and the inci- tion of distinct T cell clones in each individual, and the dif- dence of the GVHD/GVL effects is well established,7,40 no firm ficulty of predicting TCR cross-reactivity both contribute to the relation between T cell dose and survival has been reported toxicity associated with adoptive transfer of polyclonal T cells yet. Despite the small number of donor/recipient pairs, we in humans, even in a MHC-matched setting. In contrast, the demonstrated a statistically significant relationship between manipulation of non-conventional lymphocytes, and in parti- the infused T cell dose and patient survival: patients receiving cular NK cells, seems attractive for adoptive transfer. Indeed, high doses of T cells (Ͼ 0.3 × 108 T cells/kg) exhibit a NK cells are much less diverse in their repertoire of activating decreased survival, as shown with Cox multivariate (Table 4) cell surface receptors than T cells for the TCR repertoire and and Kaplan–Meier analyses (Figure 4). This observation does NK cell reactivity is tightly controlled by the recognition of not reflect the origin of allogeneic cells, and thus a potential stress-induced cell surface molecules (ie MICA), as well as by bias in patient selection, as the group with high T cell doses the absence of recognition of self-MHC class I molecules.

Leukemia T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2263

Figure 2 Phenotypic analysis of NK cell subsets present in allogeneic hematopoietic grafts. A representative frozen allogeneic apheresis sample was stained for three-color flow cytometry as described in Methods. Samples were gated on CD3−CD16+CD56+ lymphocytes. NK cell subsets are defined as NK cells expressing at least one of the following NKR: CD94, CD159a, CD158a/h, CD158b1/b2/j or CD158e1/e2. The number in the upper left quadrant indicates the percentage of NK cell subset.

These properties predict an absence of toxicity of NK cell mouse NK cells.41–42 It is thus possible that shaping of the NKR infusions against normal donor tissues in an allogeneic MHC- repertoire is also sequential in humans, with CD158a/h being matched setting. expressed later in the course of NK cell maturation. Neverthe- In the present retrospective study of 27 MHC-matched allo- less, the present phenotypic characterization supports the fact geneic cell transplantations, no relation could be established that NK cells present in hematopoietic grafts, and thus infused between NK cell doses and patient survival or acute GVHD. in patients during the course of allogeneic transplantation, are This observation suggests that infusions of low doses equipped with multiple NKR, and are therefore capable of (Ͻ 0.9 × 108/kg) of bone marrow or apheresis NK cells are not interactions with MHC class I molecules. toxic. Yet, this result does not preclude that donor NK cells Experimental settings used to demonstrate a role for NK may play a role in post-transplantation events as the number cells in murine models of transplantation differ from thera- of patients included in this study may be too limited to dem- peutic human transplantation protocols: (1) murine models onstrate such a relation. The peripheral blood CD16+CD56dim that demonstrate a positive role for NK cells in survival, NK cell subset (Ϸ 90% NK cells) has been described as exert- GVHDand GVL, use IL-2-activated NK cells, and (2) NK cell ing potent cytolytic function in vitro, with a low proliferation doses are approximately 10- to 20-fold higher than NK cell and cytokine production potential. In contrast, the CD16dim/- doses infused in this study (1 × 107 to 2 × 107 IL-2-activated CD56bright NK cell subset (Ϸ 10% NK cells), exerts weaker NK cells were infused in mice, a cell dose approximately cytolytic activity, proliferates more vigorously in response to equivalent to 5 × 108 to 1 × 109 NK cells/kg in humans).13–15 cytokines such as IL-2 and secretes inflammatory cytokines.38 Thus, it remains to be determined whether high doses Our finding that NK cells infused in this study comprise both (Ͼ108/kg) of activated human NK cells may similarly improve CD16+CD56dim and CD16dim/-CD56bright subsets (Figure 1) clinical outcome of human allogeneic transplantations. implies that NK cells infused within allogeneic bone marrow Our study suggests that given the difficulty of controling and aphereses transplantation contain both cytotoxic and donor T cell toxicity, transplantation biologists now have tools cytokine producing NK cells. However, the analysis of the dis- to explore the role of non-conventional lymphocytes involved tribution of NKR+ on NK cells demonstrates a difference of the in innate immunity, particularly the NK cell populations, in distribution pattern of the CD158a/h+ NK cell subset between human allogeneic transplantation, and to establish whether both types of graft. The significance of the lower proportion NK cells may play a positive role in MHC-matched transplan- of the CD158a/h+ NK cell subset in bone marrow remains tation, similar to that described in the haplo-identical set- unclear. An orderly acquisition of NKR has been described on ting.28 While in a MHC-matched setting, donor NK cells

Leukemia T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2264

Figure 4 Post-transplantation survival. The Kaplan–Meier curves represent overall survival for recipients of non-T cell-depleted allo- Figure 3 Distribution of NK cell subsets present in allogeneic geneic transplantation from a MHC-matched donor, who received a hematopoietic grafts. Percentages (mean ± s.e.m.) of NK dose of T cells above (dotted line) or below (solid line) the median (CD3−CD16+CD56+) cells that express indicated NKR, as measured in value for this parameter (0.3 × 108 T cells/kg). Survival for the two cryopreserved and thawed cell samples, obtained from (a) allogeneic groups was compared using a log-rank test. aphereses, (b) allogeneic bone marrow. *Statistically significant differ- ence (P = 0.017) between values obtained from bone marrow grafts and aphereses. secretion. Second, activation of donor NK cells may also occur when the peptides presented by MHC class I molecules expressed on host tumor cells alter appropriate interactions Table 4 Role of the infused NK and T cells in post-transplan- with NKR. Indeed, the nature of the peptides exposed in the tation events groove of molecules is critical for the affinity of KIR/MHC class I interactions44,45. Finally, the cell surface expression of Variables Clinical events ligands for NCR/NKG2Dcan be up-regulated on host tumor (Log (cell dose)) cells, leading to the selective attack by donor NK cells.46,47 Overall survival Acute GVHD Along these lines, it has recently been demonstrated that (Cox model) (Logistic P value regression) mouse NK cells could mediate the rejection of tumor cell lines P value such as EL4 (thymoma) and B16-BL6 (melanoma) when they are transfected with NKG2Dligands, eg RAE-1 and H60. 29,30 Total NK cells 0.323 0.382 These data demonstrate that NK cells can reject tumor cells NK cell subsets: expressing these ligands, despite their expression of MHC CD158a/h+ 0.186 0.223 class I molecules, thus demonstrating that NK cell anti-tumo- + CD158b1/b2/j 0.439 0.658 ral response can also occur in the HLA-matched setting. CD158e1/e2+ 0.916 0.776 + Clinical trials have already demonstrated that the infusion CD94 0.779 0.874 CD159a+ 0.560 0.759 of autologous IL-2-activated adherent NK cells is feasible in 48,49 Total T cells 0.014* 0.972 humans and well tolerated. It is thus tempting to design innovative therapeutic strategies that aim at amplifying and Multivariate analyses including CD34+ cell dose, log (T cell dose) activating NK cells in vivo or in vitro to improve the anti- and log (NK cell or NK cell subset dose), were used to investigate tumor effect of allogeneic cell transplantation while minimiz- the relationship between the infused T and NK cells and post-trans- ing their toxicities. plantation events, ie overall survival and the development of acute GVHD. P values are indicated and P Ͻ 0.05 was considered as a threshold for a statistically significant relation (*). Acknowledgements should be antagonized by recognition of host MHC class I molecules, a number of factors can alter these interactions, The authors thank Pascale Paul (CIML) for helpful discussion, thus providing an incentive to study the phenotype and func- the Hematology Laboratory (Hopital de la Conception) and tion of the donor NK cell compartment. First, reactivity of the Cell Therapy Facility (Institut Paoli-Calmettes) for support donor NK cells against host tumor cells may be the result of during the conduct of this work. This work was supported in a downregulation of MHC class I-specific molecules, as part by institutional grants from INSERM, CNRS, Ministe`re de described in certain types of cancers.43 In such circumstances, l’enseignement supe´rieur et de la Recherche, specific grants inhibitory signals delivered via NKR engagement are not from Ligue Nationale contre le Cancer (Equipe labellise´eLa impaired, leading to NK cell cytolytic activity and cytokine Ligue), from Association pour la Recherche contre le Cancer

Leukemia T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2265 (EV, VP) and by Institut Paoli-Calmettes. The authors also 18 Long EO. Regulation of immune responses through inhibitory thank the Beckman-Coulter Company for its contribution to receptors. Ann Rev Immunol 1999; 17: 875–904. the conduct of this work. 19 Moretta A, Bottino C, Vitale M, Pende D, Cantoni C, Mingari MC, Biassoni R, Moretta L. Activating receptors and coreceptors involved in human natural killer cell-mediated cytolysis. Ann Rev Immunol 2001; 19: 197–223. References 20 Cosman D, Mullberg J, Sutherland CL, Chin W, Armitage R, Fanslow W, Kubin M, Chalupny NJ. ULBPs, novel MHC class I- related molecules, bind to CMV glycoprotein UL16 and stimulate 1 Thomas ED, Storb R, Clift RA, Fefer A, Johnson L, Neiman PE, NK cytotoxicity through the NKG2Dreceptor. Immunity 2001; 14: Lerner KG, Glucksberg H, Buckner CD. Bone-marrow transplan- 123–133. tation. N Engl J Med 1975; 292: 832–843 (part I) and 895-902 21 Pende D, Cantoni C, Rivera P, Vitale M, Castriconi R, Marcenaro (part II). S, Nanni M, Biassoni R, Bottino C, Moretta A, Moretta L. Role of 2 Sullivan KM, Weiden PL, Storb R, Witherspoon RP, Fefer A, Fisher NKG2Din tumor cell lysis mediated by human NK cells: L, Buckner CD, Anasetti C, Appelbaum FR, Badger C, Beatty P, cooperation with natural cytotoxicity receptors and capability of Bensinger W, Berenson R, Bigelow C, Cheever MA, Clift R, Deeg recognizing tumors of nonepithelial origin. Eur J Immunol 2001; HJ, Doney K, Greenberg P, Hansen JA, Hill R, Loughran T, Martin 31: 1076–1086. P, Neiman P, Petersen FB, Sanders J, Singer J, Stewart P, Thomas 22 Mandelboim O, Lieberman N, Lev M, Paul L, Arnon TI, Bushkin ED. Influence of acute and chronic graft-versus-host disease on Y, Davis DM, Strominger JL, Yewdell JW, Porgador A. Recognition relapse and survival after bone marrow transplantation from HLA- of haemagglutinins on virus-infected cells by NKp46 activates lysis identical siblings as treatment of acute and chronic leukemia. by human NK cells. Nature 2001; 409: 1055–1060. Blood 1989; 73: 1720–1728. 23 Cambiaggi A, Verthuy C, Naquet P, Romagne F, Ferrier P, Biassoni 3 Sullivan KM, Storb R, Buckner CD, Fefer A, Fisher L, Weiden PL, R, Moretta A, Moretta L, Vivier E. Natural killer cell acceptance Witherspoon RP, Appelbaum FR, Banaji M, Hansen J, Martin P, of H-2 mismatch bone marrow grafts in transgenic mice express- Sanders JE, Singer J, Thomas ED. Graft-versus-host disease as ing HLA-Cw3 specific killer cell inhibitory receptor. Proc Natl adoptive immunotherapy in patients with advanced hematologic Acad Sci USA 1997; 94: 8088–8092. neoplasms. N Engl J Med 1989; 320: 828–834. 24 Koh CY, Blazar BR, George T, Welniak LA, Capitini CM, Raziud- 4 Tricot G, Vesole DH, Jagannath S, Hilton J, Munshi N, Barlogie din A, Murphy WJ, Bennett M. Augmentation of antitumor effects B. Graft-versus-myeloma effect: proof of principle. Blood 1996; by NK cell inhibitory receptor blockade in vitro and in vivo. Blood 87: 1196–1198. 2001; 97: 3132–3137. 5 Eibl B, Schwaighofer H, Nachbaur D, Marth C, Gachter A, Knapp 25 Ruggeri L, Capanni M, Casucci M, Volpi I, Tosti A, Perruccio K, R, Bock G, Gassner C, Schiller L, Petersen F, Niederwieser D. Evi- Urbani E, Negrin RS, Martelli MF, Velardi A. Role of natural killer dence for a graft-versus-tumor effect in a patient treated with mar- cell alloreactivity in HLA-mismatched hematopoietic stem cell row ablative chemotherapy and allogeneic bone marrow trans- transplantation. Blood 1999; 94: 333–339. plantation for breast cancer. Blood 1996; 88: 1501–1508. 26 Ruggeri L, Capanni M, Urbani E, Perruccio K, Shlomchik WD, 6 Childs R, Chernoff A, Contentin N, Bahceci E, Schrump D, Leit- Tosti A, Posati S, Rogaia D, Frassoni F, Aversa F, Martelli MF, man S, Read EJ, Tisdale J, Dunbar C, Linehan WM, Young NS, Barrett AJ. Regression of metastatic renal-cell carcinoma after non- Velardi A. Effectiveness of donor natural killer cell alloreactivity myeloablative allogeneic peripheral-blood stem-cell transplan- in mismatched hematopoietic transplants. Science 2002; 295: tation. N Engl J Med 2000; 343: 750–758. 2097–2100. 7 Kernan NA. T-cell depletion for the prevention of graft-versus- 27 Karre K. Immunology. A perfect mismatch. Science 2002; 295: host-disease. In: Thomas ED, Blume KG, Forman SJ (eds). Hemato- 2029–2031. poietic Stem Cell Transplantation. Blackwell Science: Oxford, 28 Zeis M, Uharek L, Glass B, Steinmann J, Dreger P, Gassmann W, 1999, pp 186–196. Schmitz N. Allogeneic MHC-mismatched activated natural killer 8 Guinan EC, Boussiotis VA, Neuberg D, Brennan LL, Hirano N, cells administered after bone marrow transplantation provide a Nadler LM, Gribben JG. Transplantation of anergic histoincompat- strong graft-versus-leukaemia effect in mice. Br J Haematol 1997; ible bone marrow allografts. N Engl J Med 1999; 340: 1704–1714. 96: 757–761. 9 Blaise D, Olive D, Michallet M, Marit G, Leblond V, Maraninchi 29 Cerwenka A, Baron JL, Lanier LL. Ectopic expression of retinoic D. Impairment of leukaemia-free survival by addition of interleu- acid early inducible-1 gene (RAE-1) permits natural killer cell- kin-2-receptor antibody to standard graft-versus-host prophylaxis. mediated rejection of a MHC class I-bearing tumor in vivo. Proc Lancet 1995; 345: 1144–1146. Natl Acad Sci USA 2001; 98: 11521–11526. 10 Drobyski WR, Majewski D, Hanson G. Graft-facilitating doses of 30 Diefenbach A, Jensen ER, Jamieson AM, Raulet DH. Rae1 and ex vivo activated gammadelta T cells do not cause lethal murine H60 ligands of the NKG2Dreceptor stimulate tumour immunity. graft-vs.-host disease. Biol Blood Marrow Transplant 1999; 5: Nature 2001; 413: 165–171. 222–230. 31 Blaise D, Jourdan E, Michallet M, Jouet JP, Boiron JM, Michel G, 11 Trinchieri G. Biology of natural killer cells. Adv Immunol 1989; Faucher C, Fegueux N, Schuller MP, Badri N, Chabannon C, Mar- 47: 187–376. aninchi D. Mobilisation of healthy donors with lenograstim and 12 Bennett M. Biology and genetics of hybrid resistance. Adv Immu- transplantation of HLA-genoidentical blood progenitors in 54 nol 1987; 41: 333–445. patients with hematological malignancies: a pilot study. Bone 13 Murphy WJ, Longo DL. The potential role of NK cells in the separ- Marrow Transplant 1998; 22: 1153–1158. ation of graft-versus-tumor effects from graft-versus-host disease 32 Blaise D, Kuentz M, Fortanier C, Bourhis JH, Milpied N, Sutton after allogeneic bone marrow transplantation. Immunol Rev 1997; L, Jouet JP, Attal M, Bordigoni P, Cahn JY, Boiron JM, Schuller 157: 167–176. MP, Moatti JP, Michallet M. Randomized trial of bone marrow 14 Asai O, Longo DL, Tian ZG, Hornung RL, Taub DD, Ruscetti FW, versus lenograstim-primed blood cell allogeneic transplantation in Murphy WJ. Suppression of graft-versus-host disease and amplifi- patients with early-stage leukemia: a report from the Societe Fran- cation of graft-versus-tumor effects by activated natural killer cells caise de Greffe de Moelle. J Clin Oncol 2000; 18: 537–546. after allogeneic bone marrow transplantation. J Clin Invest 1998; 33 Olivero S, Alario T, Ladaique P, Haccoun M, Viens P, Blaise D, 101: 1835–1842. Chabannon C. CD34+ cell enumeration in peripheral blood and 15 Koh CY, Welniak LA, Murphy WJ. Adoptive cellular immunother- apheresis samples, using two laboratory diagnostic kits or an insti- apy: NK cells and bone marrow transplantation. Histol Histo- tutional protocol. Bone Marrow Transplant 1999; 23: 387–394. pathol 2000; 15: 1201–1210. 34 Andre P, Biassoni R, Colonna M, Cosman D, Lanier LL, Long EO, 16 Moretta A, Bottino C, Vitale M, Pende D, Biassoni R, Mingari MC, Lopez-Botet M, Moretta A, Moretta L, Parham P, Trowsdale J, Viv- Moretta L. Receptors for HLA class-I molecules in human natural ier E, Wagtmann N, Wilson MJ. New nomenclature for MHC killer cells. Ann Rev Immunol 1996; 14: 619–648. receptors. Nat Immunol 2001; 2: 661. 17 Lanier L. On guard-activating NK cell receptors. Nat Immunol 35 Schmitz N, Bacigalupo A, Hasenclever D, Nagler A, Gluckman 2001; 2: 23–27. E, Clark P, Bourquelot P, Greinix H, Frickhofen N, Ringden O,

Leukemia T cells and NK cell subsets in human allogeneic hematopoietic grafts V Pascal et al 2266 Zander A, Apperley JF, Gorin C, Borkett K, Schwab G, Goebel M, 43 Marincola FM, Jaffee EM, Hicklin DJ, Ferrone S. Escape of human Russell NH, Gratwohl A. Allogeneic bone marrow transplantation solid tumors from T-cell recognition: molecular mechanisms and vs filgrastim-mobilised peripheral blood progenitor cell transplan- functional significance. Adv Immunol 2000; 74: 181–273. tation in patients with early leukaemia: first results of a randomised 44 Maenaka K, Juji T, Nakayama T, Wyer JR, Gao GF, Maenaka T, multicentre trial of the European Group for Blood and Marrow Zaccai NR, Kikuchi A, Yabe T, Tokunaga K, Tadokoro K, Stuart Transplantation. Bone Marrow Transplant 1998; 21: 995–1003. DI, Jones EY, Van Der Merwe PA. Killer cell Immunoglobulin 36 Vigorito AC, Azevedo WM, Marques JF, Azevedo AM, Eid KA, receptors and T cell receptors bind peptide- major histocompat- Aranha FJ, Lorand-Metze I, Oliveira GB, Correa ME, Reis AR, Mir- ibility complex class I with distinct Thermodynamic and kinetic anda EC, de Souza CA. A randomised, prospective comparison properties. J Biol Chem 1999; 274: 28329–28334. of allogeneic bone marrow and peripheral blood progenitor cell 45 Liberatore C, Capanni M, Albi N, Volpi I, Urbani E, Ruggeri L, transplantation in the treatment of haematological malignancies. Mencarelli A, Grignani F, Velardi A. Natural killer cell-mediated Bone Marrow Transplant 1998; 22: 1145–1151. lysis of autologous cells modified by gene therapy. J Exp Med 37 Powles R, Mehta J, Kulkarni S, Treleaven J, Millar B, Marsden J, 1999; 189: 1855–1862. Shepherd V, Rowland A, Sirohi B, Tait D, Horton C, Long S, 46 Groh V, Rhinehart R, Secrist H, Bauer S, Grabstein KH, Spies T. Singhal S. Allogeneic blood and bone-marrow stem-cell transplan- Broad tumor-associated expression and recognition by tumor- tation in haematological malignant diseases: a randomised trial. derived gammadelta T cells of MICA and MICB. Proc Natl Acad Lancet 2000; 355: 1231–1237. Sci USA 1999; 96: 6879–6884. 38 Cooper MA, Fehniger TA, Caligiuri MA. The biology of human 47 Vivier E, Tomasello E, Paul P. Lymphocyte activation via NKG2D: natural killer-cell subsets. Trends Immunol 2001; 22: 633–640. towards a new paradigm in immune recognition? Curr Opin 39 Andre P, Spertini O, Guia S, Rihet P, Dignat-George F, Brailly H, Immunol 2002; 14: 306–311. Sampol J, Anderson PJ, Vivier E. Modification of P-selectin glyco- 48 Lister J, Rybka WB, Donnenberg AD, deMagalhaes-Silverman M, protein ligand-1 with a natural killer cell-restricted sulfated lactos- Pincus SM, Bloom EJ, Elder EM, Ball ED, Whiteside TL. Autologous amine creates an alternate ligand for L-selectin. Proc Natl Acad peripheral blood stem cell transplantation and adoptive immuno- Sci USA 2000; 97: 3400–3405. therapy with activated natural killer cells in the immediate post- 40 Urbano-Ispizua A, Rozman C, Pimentel P, Solano C, de La Rubia transplant period. Clin Cancer Res 1995; 1: 607–614. J, Brunet S, Perez-Oteiza J, Ferra C, Zuazu J, Caballero D, Carval- 49 deMagalhaes-Silverman M, Donnenberg A, Lembersky B, Elder E, hais A, Diez JL, Espigado I, Martinez C, Campilho F, Sanz MA, Lister J, Rybka W, Whiteside T, Ball E. Posttransplant adoptive Sierra J, Garcia-Conde J, Montserrat E. The number of donor immunotherapy with activated natural killer cells in patients with CD3(+) cells is the most important factor for graft failure after allo- metastatic breast cancer. J Immunother 2000; 23: 154–160. geneic transplantation of CD34(+) selected cells from peripheral 50 Pascal V, Vivier E, Andre´ P. CD158 (killer immunoglobulin-like blood from HLA-identical siblings. Blood 2001; 97: 383–387. receptors family) report. In: Mason D, Andre´ P, Bensussan A, 41 Williams NS, Kubota A, Bennett M, Kumar V, Takei F. Clonal Buckley C, Civin C, Clark E, de Hass M, Goyert S, Hadam M, Hart analysis of NK cell development from bone marrow progenitors D, Horejsi V, Jones Y, Meuer S, Morrissey J, Schwartz-Albiez R, in vitro: orderly acquisition of receptor gene expression. Eur J Shaw S, Simmons D, Turni L, Uguccioni M, van der Schoot M, Immunol 2000; 30: 2074–2082. Vivier E, Zola H (eds). Leucocyte Typing VII. White Cell Differen- 42 Roth C, Carlyle JR, Takizawa H, Raulet DH. Clonal acquisition of tiation Antigens. Oxford University Press: Oxford, 2002, pp inhibitory Ly49 receptors on developing NK cells is successively 412–413. restricted and regulated by stromal class I MHC. Immunity 2000; 13: 143–153.

Leukemia