after bone marrow transplantation for chronic ...

2 downloads 0 Views 1MB Size Report
M Hauch, MV Gazzola, T Small, C Bordignon, L Barnett, I Cunningham, H Castro- Malaspinia, RJ after bone ... Hugo Castro-Malaspinia, Richard J. O'Reilly, and Carolyn A. Keever ... Robert J. Kleberg and Helen C. Kleberg Foundation, the Lila.
From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only.

1990 75: 2250-2262

Anti-leukemia potential of interleukin-2 activated natural killer cells after bone marrow transplantation for chronic myelogenous leukemia M Hauch, MV Gazzola, T Small, C Bordignon, L Barnett, I Cunningham, H Castro- Malaspinia, RJ O'Reilly and CA Keever

Information about reproducing this article in parts or in its entirety may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtml

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2011 by The American Society of Hematology; all rights reserved.

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only.

Anti-Leukemia Potential of Interleukin-2 Activated Natural Killer Cells After Bone Marrow Transplantation for Chronic Myelogenous Leukemia By Michael Hauch, Maria V. Gazzola, Trudy Small, Claudio Bordignon, Lorna Barnett, Isabel Cunningham, Hugo Castro-Malaspinia, Richard J. O’Reilly, and Carolyn A. Keever The anti-leukemia potential of natural killer (NK) cells has been evaluated in 40 patients transplanted for chronic myelogenous leukemia (CML) t o determine whether differences in NK cell function were correlated with subsequent leukemic relapse. Cells from patients and their donors were tested in “Cr release assays against fully allogeneic CML targets and against cultured K562 targets; cells from 26 patients were tested against host-derived CML targets that were cryopreserved before transplantation. Cultured CML targets (K562) were highly susceptible t o lysis by freshly isolated peripheral blood lympocytes (PBL) and to a greater degree by PBL cultured in medium containing interleukin-2 (IL-2) in all assays performed. In contrast, noncultured CML targets were lysed only by IL-2-activated cells from a subset of patients. When present, lytic activity t o CML targets was detectable as early as 3 weeks after bone marrow transplantation, and remained positive throughout the posttransplant period. Optimal lytic activity developed within the first week of culture and required 2250 U/mL of IL-2 in the culture medium. Lytic activity t o

fully allogeneic and host-derived CML targets appeared to be mediated by CD16’ and CD56’ cells but not by CD3’ cells. Lysis of allogeneic CML targets was variable, but patients could be divided into t w o groups: those with and those without lytic activity t o the majority of targets tested. The basis for the differences in lytic activity could not be ascribed t o target susceptibility t o lysis, the proportion of NK cells in the cultures, or t o the phenotype of the NK cell subsets in the cultures. When tested in parallel, the lytic activity of donor and recipient cultures against hostderived CML targets was highly correlated, suggesting that there may be inherent differences in the ability of NK cells t o recognize CML targets. The risk of relapse for patients who failed t o generate lytic activity against host-derived CML targets was significantly increased over that for patients with lytic activity against host leukemia. These data indicate that posttransplant immunotherapy with IL-2 designed to activate NK cells will likely augment the graft-versus-leukemia potential of the graft. 0 1990 by The American Society of Hematology.

0

studies the presence of acute and/or chronic graft-versushost disease (GVHD)’-4also has been found to be associated with a decreased incidence of relapse in patients transplanted for leukemia. Consonant with this observation have been studies reporting an increase in leukemic relapse in recipients of marrow from which mature T cells have been removed for GVHD prevention.’-’ However, the GVHD and GVL potential of marrow grafts can be separated in animal models,’0.” and augmentation of GVHD by abbreviations in GVHD prophylaxis or by increasing the dose of immunocompetent T cells administered with the graft has resulted in increased mortality without a reduction in the incidence of leukemia relapse in a recent human trial.” Several possible mechanisms have been proposed to explain the GVL e f f e ~ tOne . ~ of these possibilities involves the activation of nonspecific effector systems such as natural killer (NK) cells during GVHD. There is a substantial body of evidence that supports the role of N K cells in defense against leukemia. In murine models, an association between N K activity and susceptibility to leukemias has been ~ h o w n ’ ~ . ‘ ~ and resistance to leukemia can be transferred with cloned N K cells.’’ In humans there appears to be an increased incidence of hematopoietic malignancies in individuals with low N K activity,l6 and impaired N K functions are commonly seen in patients with leukemia and preleukemic Because high numbers of circulating cells with NK phenotype and full functional activity have been demonstrated early following all types of BMT,’*-*’ we have concentrated our efforts to determine whether these N K cells can function as anti-leukemia effectors during the posttransplant period. For our initial studies we have examined peripheral blood mononuclear (PBMC) obtained a t intervals after transplant from a series of patients transplanted for chronic myelogenous leukemia (CML). We have compared the lytic activities of resting and interleukin-2 (IL-2)-activated patient cells

N E RATIONALE for using bone marrow transplantation (BMT) as a treatment for leukemia is that this approach permits the administration of total body irradiation (TBI) and chemotherapeutic drugs designed to eliminate leukemic cells at doses that would otherwise induce a permanent marrow aplasia. The administration of donor marrow is required to restore hematopoiesis and immune function. Evidence also suggests that the efficacy of an allogeneic BMT as a treatment for leukemia is not strictly limited to the conditioning regimen, but may also involve the activity of immunocompetent cells transplanted with the marrow that eliminates residual host leukemic cells. The evidence for a such a graft-versus-leukemia (GVL) effect in humans is strongly suggested by the reduced incidence of relapse reported after HLA-matched marrow grafts in comparison with grafts from identical twins.’ In a number of

From the Bone Marrow Transplantation Service, Memorial Sloan-Kettering Cancer Center, New York, NY. Submitted July 29, 1989; accepted February 2, 1990. Supported in part by Grant No. CA23766 from the National Institutes of Health, the Zelda Weintraub Cancer Foundation, the Robert J. Kleberg and Helen C. Kleberg Foundation, the Lila Acheson De Witt Wallace Foundation, the Charles A . Dana Foundation, the Vicent Astor Fund, the Andrew S. Gaflney Foundation, and the Yankee-Frito Lay Challenge Fund. M.V.G. is supported by a grant from the Italian Association for Cancer Research (AIRC). Address reprint requests to Carolyn A . Keever. PhD. Division of Bone Marrow Transplantation. Ohio State University, Doan Hall, 410 WlOth Avenue, Columbus, OH43210. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.section 1734 solely to indicate this fact. 0 1990 by The American Society of Hematology. 0006-4971/90/751I-0002$3.00/0 2250

Blood, Vol 7 5 , No 1 1 (June 11, 1990: pP 2250-2262

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. 2251

ANTILEUKEMIA ACTIVITY POST BMT

with cells from the bone marrow donors against host-derived and/or allogeneic CML targets. We have examined the kinetics of activation, the target cell specificity, and the phenotype of the anti-leukemia effectors. We have further assessed whether the potential for anti-leukemia activity might be associated with posttransplant CML relapse. MATERIALS AND METHODS

Patients. All of the 40 patients included in this study were transplanted for Philadelphia chromosome positive (Ph'+) CML from HLA identical donors. The median age of the patients was 29 with a range of 3 to 45 years, and consisted of 24 males and 16 females. Thirty-two patients were transplanted for CML in first chronic phase, four patients were in second chronic phase, and three patients were transplanted in accelerated phase (as defined by additional chromosomal abnormalities). One patient was transplanted with greater than 30% marrow myeloblasts after a period during which treatment with Idarubicin (Erbamont Inc, Adria Laboratories, Dublin, OH) and cytosine arabinoside resulted in disappearance of the Ph' chromosome, and is thus considered to have been in blast crises. All patients had previously received hydroxyurea; four patients had received a-interferon, and four patients received chemotherapy with either Idarubicin and cytosine arabinoside (three cases) or Adriamycin, Vincistine, and prednisone (one case) before conditioning for BMT. Pretransplant conditioning consisted of hyperfractionated TBI (1,375 cGy to 1,500 cGy) followed by cyclophosphamide (60 mg/kg x 2). Thirty-five of the patients were studied after receiving marrow depleted of T cells by treatment with soybean agglutinin and sheep erythrocytes (SBA-E-BM) as previously described.26Donor peripheral blood was added back to the marrow in five of these cases to achieve a T-cell dose of 1.0 to 1.5 x 105/kg (SBA-E- + T). Two patients received marrow that was partially depleted of T cells by treatment with SBA (SBA-) alone, and three patients received conventional, unseparated marrow grafts. Four grafts were secondary transplants after graft failure or leukemia relapse. Recipients of conventional marrow grafts received short course methotrexate (MTX) with cyclosporine posttransplant as GVHD pro phyla xi^.'^ Recipients of T-cell depleted BMT received no posttransplant GVHD prophylaxis, and one recipient of partially T-cell depleted marrow received cyclosporine alone. No patient developed chronic GVHD, and only one patient developed acute, fatal GVHD (grade 111) after transplantation of SBA-E-BM, to which donor PBMC had been added back. One patient experienced late graft failure, without detection of Ph'+ cells before death at week 13. Clinical relapse occurred in nine patients, and was defined as the reappearance in the marrow of metaphases containing the Ph' chromosome and the development of hematologic changes characteristic of the underlying disease. Isolated appearance of Ph' in the marrow has occurred in four patients. These patient characteristics are summarized in Table 1 (see Table 7 for presentation in detail). These studies were performed with informed consent following protocol approval by the Institutional Review Board. Preparation of leukemic and normal targets. Before transplantation, mononuclear leukemic cells were isolated from patient peripheral blood or spleen on gradients of Ficoll-Hypaque (Lymphoprep, Nyegaard & Co, Oslo, Norway). Peripheral blood from one patient who relapsed after transplant in the blast crises of CML was obtained for use as targets only. For some samples, further fractionation was performed by incubation on nylon wool columns for 60 minutes at 37OC, and separation of the eluted cells on discontinuous step gradients of five different densities of Percoll (Pharmacia, Uppsala, Sweden) ranging from 40% to 60% Percoll. In two cases, PBMC were depleted of mature cells by immune rosetting

Table 1. Patient Characteristics

No. Patients Type of transplant SBA-E-BMT SBA-E- plus peripheralT cells SBAConventional BMT Disease Status 1st Chronic phase 2nd Chronic phase Accelerated phase Blastic CML Age Median Range Sex M F Clinical relapse Cytogenetic relapset

32 4 3 1 29 3-45 24 16

9 5

"Number of secondary transplants within each transplant type. tMinor populations of host-derivedPh" marrow metaphases without evidence of peripheral disease.

with chromium chloride treated human erythrocytes as described." The PBMC were treated with a combination of monoclonal antibodies (MoAbs) (in ascites form), all of which were adjusted to a final dilution of 1:200. The antibodies used included: B-53.4 (CD4) and B-36.1 (CD5) (T lymphocytes), B-73.1 (CD16) (NK cells), B13.4.1 (monocytes and mature myeloid cells), and OKMl (CDll) (monocytes, NK cells, T-suppressor cells, granulocytes) (kindly provided by Dr G. Trinchieri, Wistar Institute, Philadelphia, PA). Differentials of Wright-Giemsa stained cytocentrifuge preparations of the cells were performed before and/or after cryopreservation. Only those preparations containing 275% myeloid cells were used as targets. The target from the blast crises CML patient was fractionated on a Percoll gradient and consisted of greater than 98% blasts by morphology. SBA-E-BM cells were obtained from normal bone marrow donors and were cryopreserved on the day of transplant. Leukemic cells and SBA-E-BM were cryopreserved in aliquots of approximately 5 x IO6 cells/mL in a freezing solution of 90% fetal bovine serum (FBS) and 10% dimethylsulfoxide (DMSO) and were stored in liquid nitrogen. PBMC from marrow donors, patients after transplantation, and normal controls were separated on FicollHypaque gradients and were cryopreserved in a freezing solution of 20% FBS, 10% DMSO, and 70% RPMI-1640. N K and lymphokine activated killer cell (LAK). Heparinized peripheral blood was separated on Ficoll-Hypaque gradients and was depleted of monocytes by plastic adherence before assay of NK lytic function. LAK assays were performed on freshly isolated or cryopreserved PBMC after 3 to 7 days of culture at 37OC in a humidified atmosphere of 5% CO,. The cells were cultured at a concentration of 1 x lo6 cells/mL in RPMI 1640 medium supplemented with 25 mmol/L HEPES, 100 U/mL penicillin, 100 pg/mL streptomycin, 2 mmol/L L-glutamine, 10% heat-inactivated pooled human serum (M.A. Bioproducts, Walkersville, MD), with human recombinant IL-2 (kindly provided by the Cetus Corp, Emeryville, CA). NK and LAK lytic activites were measured in standard 4 hour 5'Cr release assays at multiple effector to target cell ratios as previously described." The K562 cell line, which was established from the pleural effusion of a patient in CML blast crises, was used as target for fresh NK cell activity and was included in each assay of LAK activity. K562 targets were labeled in a total volume of 0.5 mL RPMI-10% FBS for 1 to 2 hours at 37OC with 125 pCi of Sodium

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. 2252

HAUCH ET AL

"Chromate, specific activity 5 mCi/mL) (New England Nuclear, Boston, MA). CML and PBMC targets were thawed on the day of testing and were labeled with 250 pCi of Sodium Chromate for 2 hours at 37OC. The targets were washed and adjusted for testing as previously described.25 The results are expressed as the percent specific lysis at a given E:T ratio calculated from the formula: CPM Test - CPM Spontaneous x 100 %Specific Lysis = CPM Maximum - CPM Spontaneous or as lytic units per lo6 cells, with 1 lytic unit defined as the number of cells required to lyse 33.3% of 5,000 target cells.29 Immunofluorescence analysis. Two-color immunofluorescence analysis was performed by standard techniques, using fresh PBMC or cultured LAK cells. A pan leukocyte antigen, HLe (CD45)FITC, served as a control to gate out residual red blood cells. Each two-color fluorescence study included a double negative control (MsIgG-FITC/MsIgG-PE), a positive FITC/negative PE control (HL e-FITC/MsIgG-PE), and a negative FITC/positive PE control (MsIgG-FICT/Leu4-PE). The Moabs used to generate the data reported in this study were purchased from Becton-Dickinson (Mountain View, CA) and included Leu4 (CD3), Leu2 (CD8), Leu16 (CD20), Leu19 (CD56), and Leul 1 (CD16). Immunofluorescence samples were analyzed on a FACScan flow cytometer (BectonDickinson) or alternatively on an Epic C flow cytometer (Coulter Immunology, Hialeah, FL). Antibody plus complement depletions. IL-2activated cells were treated with mouse MoAbs plus low toxicity baby rabbit complement (Pel Freez, Rogers, AR) or with complement alone as previously de~cribed.~'OKT3 (anti-CD3) was used as culture supernatant from the hybridoma cell line obtained from the American Type Culture Collection and was kindly provided by Dr Robert Evans (Rosewell Park Memorial Institute, Buffalo, NY). Leul l b ascites were a kind gift of Steve Brown (University of Kentucky, Lexington). NKHI, ascites were a generous gift from Dr J. Ritz (Dana Faber Cancer Insitute, Boston, MA), and OKT8 was purchased from Orthro Diagnostics (Raritan, NJ). The antibodies were used at a previously determined optimal dilution in a volume of 0.1 mL/106 treated cells. OKT3 was used undiluted, OKT8 was diluted 1:20, Leul l b was used at a dilution of 150, and NKHI, was used at a dilution of 1:125. After treatment of effector cells, all preparations were adjusted to the concentration of the complement control before assay. In preliminary experiments, direct immunofluorescence analysis of cells depleted in this manner showed 1% to 3% contamination with cells bearing the relevant antigen. RESULTS

Recovery of N K cell numbers and function after transplantation for CML. W e have previously reported t h e early regeneration of NK cells and NK and L A K activity in leukemic patients after T-cell depleted and conventional BMT.2S,30 Figure 1 shows t h a t NK cell recovery is also very rapid in the subgroup of patients transplanted for CML. T h e proportion of peripheral CD56' N K cells is elevated in most patients through t h e first year posttranspiant and is highest during the first 6 months (Fig 1, panel A). Two-color immunofluorescence analysis showed t h a t nearly all of t h e C D 5 6 + cells coexpress CD16, a n d t h a t there were few C D 3 / C D 5 6 coexpressing cells (less than 2%) (not shown). These N K cells a r e fully functional as measured by lytic activity t o K562 targets. Normal t o high activity to K562 was maintained throughout t h e posttransplant period (Fig 1, panel B). T h e activity of t h e N K cells could be markedly enhanced by 3 t o 7 days of culture in medium containing

100

AI

0

+ 60

(D v)

a

40

0

t

0

8

-

B

C

::

i(

!

0 0 0

1

.1

I

I

I

r 0

40

m m

0

0

=

ooo

QI

8

cn

0

. I

0

Q

8

0

0

20

m

0

s

c8 0

am I

0

4

0

0

0 Normal PBMC

CML PBMC

CML Enriched

Normal BM

Fig 2. Lytic activity against noncultured targets from normal controls and CML patients was tested after 3 to 7 days of culture in medium containing 250 U/mL IL-2. The data represent the highest activity of effectors from two or more individuals. Normal PBMC (N = 16) and PBMC from CML patients (N = 22) were separated on gradients of Ficoll-Hypaque: CML-enriched samples (N = 14) were enriched for early myeloid cells on discontinuous gradients of Percoll or by immune rosetting as described in Material and Methods. Normal bone marrow (N = 5 ) was enriched for immature cells and depleted of residual CD2’ cells by treatment with soybean agglutininand sheep erythrocytes as described in Material and Methods. All targets had been cryopreserved and were thawed and labeled with ”Cr on the day of testing. The data are expressed as the percent specific lysis a t an effector to target cell ratio of 100 to 1.

Thereafter, subsequent cultures were supplemented with 250 U/mL IL-2. Our previous data had also indicated that optimal LAK activity was generated after 3 to 4 days of culture of PBMC from patients less than 5 months after transplant, but required a longer culture period (5 to 7 days) for normal controls and patients greater than 5 months p~sttransplant.~’ It has been suggested that the kinetics of LAK activation may be dependent on the type of target used as well as the

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. 2254

HAUCH ET AL

concentration of IL-2 in the cultures.36 Therefore, we varied the day of assay to determine a t what time anti-leukemia activity developed and the degree to which this activity could be maintained in cultures that were supplemented with fresh medium and kept a t optimal cell concentrations (1 to 2 x lo6 cells/mL). Anti-leukemia activity was detectable as early as day 3 of culture and could be maintained for as long as 3 weeks. The activity did not appear to be consistently enhanced by culture for more than 1 week (not shown). For subsequent experiments, assays were performed after 3 to 7 days of culture. Lytic activity to CML targets with time after transplantation. The results shown in Fig 1 show that although the proportion of N K cells within the lymphoid population of C M L patients declines with time after transplant, N K and LAK lytic activity to K562 targets remains at a high level. We assessed whether lytic activity to noncultured C M L targets varied with time after transplant. The data in Table 2 are from multiple LAK assays of PBMC from six representative patients toward allogeneic C M L targets and C M L Table 2. Lytic Activity to CML Targets Over Time in

Individual Patients Patient No.

68 1

669

684

776

760

806

Mo

0.8 1.4 3.1 4.1 15.2 21.0 0.7 1.2 6.2 17.3 0.8 3.0 4.8 6.7 9.2 20.3 23.7 0.7 1.2 3.0 6.8 1.5 4.4 6.2 7.6 12.1 1.4 3.7 6.2

% Lysis HostCML.

% Lysis AlldMLt

Lytic Units K562

25.9 37.6 19.2

NT NT

1,094 597

NT

NT

NT NT NT

38.1 48.7 17.7

14.3 17.5 17.3

16.6

NT

24.2

20.1 35.2 30.3 26.7 15.3 NT NT

0.0 0.0 0.0 0.0 NT

5.7 0.0 5.4 0.0 0.1 4.6 2.0

NT NT NT

21.6 27.1 34.4/69.3$ NT

47.5$ 53.8$ NT

6.6 9.9 17.0 5.1 7.2 18.5 16.2 10.1 9.0 30.3 54.0

313 537 82 1,000 432 22 1,656 145 160 345 622 71 83 145 505 722 553 2,952 432 209 712 657 2,300 85 75 73

Abbreviation: NT, not tested. *Host-derived CML targets were cryopreserved before BMT. tDate from a single allogeneic CML target is shown for each patient unless otherwise indicated. The same allogeneic target was not used for every patient. $The allogeneic CML target tested at months 6.7, 20.3, and 23.7 differs from that used in earlier testings. Results from both targets are shown for month 6.7.

targets of host origin. Cells from three of these patients had lytic activity to host-derived C M L that was present throughout the posttransplant course (0.8 to 9 months), while IL-2-activated cells from three additional patients failed to lyse host-derived C M L targets at all times after transplant (0.8 to 12 months). The IL-2-activated cells from all six patients were able to lyse allogeneic C M L targets. There was no clear association of lytic activity toward K562 with lytic activity to C M L targets in these or in other experiments not shown. Comparison of lytic activity by effectors from patients and their bone marrow donors. The data in Table 3 are from 15 experiments in which IL-2-activated cells from the patients and their HLA identical bone marrow donors (cyropreserved a t the time of transplant) were tested in parallel. The lytic activity of donor and patient cultures to host C M L targets was very similar (excluding donorrecipient pair no. 806, r = 3 2 ) . Five donor-recipient pairs had lytic activity toward the host-derived C M L targets whereas there was no LAK activity to host targets by seven donor-recipient pairs. The donor-recipient activities were discordant in only two instances. The failure of donor or patient effectors to lyse host C M L targets was not likely due to the insensitivity of these targets to lysis, since targets that were not lysed by patient or donor effectors could be lysed by effectors from at least one other unrelated patient in every instance tested ( N = 5). Although there were fewer parallel studies performed, the data suggest a positive correlation between patient and donor lytic activity to specific allogeneic C M L targets (r = .83, N = 6). In contrast, there was no significant correlation between lytic activity of IL-2activated cultures from donors and patients to K562 targets, nor was there a correlation of K562 lytic activity with activity to noncultured C M L targets. Phenotype of anti-leukemia effector. Lysis of tumor targets by IL-2-activated PBMC has been shown to be primarily mediated by CD3- cells bearing CD16 and/or CD56 in both normal individuals and cancer patient^.^^.^^ However, CD3+ cells, which do or do not express CD16 or CD56, can also contribute to these a ~ t i v i t i e s . We ~ ~ . have ~~ previously demonstrated that LAK activity to cultured tumor targets during the posttransplant period is primarily mediated by cells with an N K phenotype (CD3-, CD16+, CD56+).’j To confirm that the LAK activity against noncultured C M L targets was due to IL-2-activiated N K cells and not T lymphocytes, we performed depletion experiments using MoAbs plus complement. The results in Table 4 show that lysis of either host-derived or fully allogeneic C M L targets was only minimally affected by removal of CD3’ or CD8+ cells, whereas removal of CD16+ cells, CD56+ cells, or both cell types consistently reduced, or eliminated the lysis of these targets. The phenotype of the anti-CML effector was the same in patients tested early ( 5 2 months) or from 1 to 5 years after transplant. A similar pattern of depletion was seen when K562 cells were used as targets, although in cultures with a high degree of lytic activity, cytotoxicity was not completely eliminated by the treatments. These data indicate that the lytically active anti-leukemia effector does have an N K phenotype and not a T-cell phenotype, and that

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. ANTI-LEUKEMIA ACTIVITY POST 8MT

2255

Table 3. Lysis of CML Targets by LAK Cells From Patients and Their Marrow Donors Targets. Host-CML

K562

UPN

684 734 68 1 546 669 70 1 742 646 776 806 859 634 633 624 830

Mos

0.8 1.2 0.8 23.6 1.2 6.9 3.8 3.5 1.o 3.8 3.2 8.6 9.5 8.7 1.o

Patient

Donor

69 97 502 140 48 142 49 64 35 195 168 25 118 184 85

145 133 439 252 432 194 378 70 43 74 177 25 159 88 164

AIIo-CML

Patient

Donor

Patient

Donor

26.8 24.8 22.8 17.7 17.3 14.4 6.5 7.2 5.4 4.6 3.6 0.5 2.8 0.0 NT

21.3 20.0 10.9 3.0 15.4 16.7 7.0t 5.3 5.3 41.4 7.9 0.0 4.3 0.0 NT

17.9 NT 48.7 40.3 NT NT 11.2 NT NT 30.3 NT NT NT NT 8.1

7.9 NT 30.3 23.7 NT NT 9.0 NT NT 35.6 NT NT NT NT 5.9

LAK activity was tested after 3 to 7 days of culture. PBMC from marrow donors were cryopreservedat the time of transplant. PBMC from the patients were obtained at the indicatedweeks after transplant. Assays of donor and patient lytic activity were performed on the same date. Abbreviation: NT, not tested. ‘Lysis of K562 targets is expressed as lytic units/lO’ effectors. Lysis of host-derived and fully allogeneic CML targets is expressed as percent specific lysis. tThe lytic activities of fully allogeneic LAK cells against targets from patients nos. 742, 646, 776, 859, and 634 were 30.3%. 35.9%. 41.0%. 36.6%. and 14.3%, respectively. Fully allogeneic LAK effectors were not tested against targets from patients nos. 633 and 624.

NK cells with anti-leukemia potential can be found throughout the posttransplant period. Activity of LAK efectors from patients toward multiple CML targets. Clones exhibiting major histocompatibility complex nonrestricted killing often display a spectrum of target cell recognition that may not be detectable in the bulk populations from which they are derived.m42We assessed whether the IL-2-activated cells from individual patients would preferentially lyse certain CML targets. The data in Table 5 show the reactivity of effectors from nine patients who, over the course of this study, had been tested to targets

from five or more fully allogeneic CML patients, including target cells isolated from the patient in CML blast crises (greater than 98% myeloblasts), target 791. Effectors from each patient were able to lyse one or more CML targets, and conversely, IL-2-activated cells from each of the patients failed to lyse one or more CML targets. In every instance, targets that were not lysed by effectors from the patients shown in the table were lysed by effectors from other patients or controls. IL-2-activated cells from four of the patients tested (nos. 753,701,790,806)had low or no lytic activity to the majority of targets tested, while LAK effectors from five

Table 4. Phenotype of Anti-Leukemia Effectors MoAb

Target.

Control

OKT3

OKT8

Leu1l b

NKH1

68 1 1.6 mos 852

Host-CML K562 Allo-CMLt

29.5 32 35.3

20.0 26 30.3

2.1 mos 675 16.1 mos 592 36.4 mos 532 46.0 mos 489 5 1.3 mos

K562 Allo-CML K562 Allo-CML K562 Allo-CML K562 AIIo-CML K562

30 27.7 88 13.0 118 14.9 130 26.7 72

43 22.0 88 15.7 114 23.3 26 1 28.7 60

NT NT 32.2 36 24.0 112 14.3 30 1 18.2 198 33.2 157

5.9 2 NT NT 12.8 9 8.6 39 2.7 54 15.6 41

NT NT NT NT 12.5 9 0.1 23 7.9 56 10.3 50

UPN

Leu11

+ NKHl

NT NT 8.0 0.0 NT NT 0.0 51 NT NT NT NT

Effectors were treated with the indicated MoAb followed by rabbit complement or with complement alone (Control) after 7 days of culture in medium containing IL-2, as described in Materials and Methods. PBMC for culture were obtained at the indicated months after transplantation. Abbreviation: NT, not tested. ‘Data for lysis of CML targets are expressed as percentage specific lysis at the highest E:T ratio tested (100: 1 or 5 0 1). Data for lysis of K562 targets are expressed as lytic units/ 1Os effectors. tThe fully allogeneic CML targets used in all experiments were isolated from a single patient in blast crises CML.

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. 2256

HAUCH ET AL Table 5. Activity of LAK Effectors From Individual Patients Against Multiple CML Targets Effectors(UPN)

Targets.

68 1

791-P 786-P 859-P 1000-P 646-P 689-F 4 16-P 634-F 681-F 1001-P 1002-P 734-F 742-P 7 19-F 753-F 70 1-F 537-F 684-F 831-P 675-F 674-F 703-F 806-P

48.7 0.8 14.7 15.8 30.1 64.0 14.3 10.5 37.6t

684

69.3 7.7 8.4 5.8 29.0 22.7

20.6

852

55.7 25.5 6.4 1.6

753

31.6 5.9

12.2

7.5 14.3

21.4 28.2 10.4 2.9

0.0 3.0 6.8

1o.ot 15.2

580

13.9 60.9

675

70 1

790

806

44.4

11.5

9.0 1.4 9.7 12.9

7.2 30.8

25.5 2.9

5.3 0.0 2.4 0.0 26.1

62.9

37.6 35.9 15.6

2.3 14.4t 6.5 8.2

35.2t 43.6

3.9

23.6t 41.0 39.7 4.6t

LAK assays were performed at multiple intervals after transplant. The data represent the highest degree to which the indicated target was lysed in any experiment. Data are expressed as the percent specific lysis at an effector to target cell ratio of 100: 1. *Targets were separated from peripheral blood OT spleen by fractionation over gradients of Percoll (P) or Ficoll-Hypaque (F) as described in Materials and Methods. Target 79 1-P was isolated from a patient who relapsed after BMT in the blast crisis of CML. tAutologous host-CML target.

patients (nos. 681, 684, 852, 580, 675) had lytic activity (210%) to most targets. Although we have not ruled out in every case the potential contribution of in vivo activated HLA allospecific cytotoxic T lymphocytes (CTL) to the activities seen, the data in Table 4 indicate that lytic activity to fully allogeneic C M L targets can be mediated by N K cells. These data suggest that there may be preferential target cell recognition by LAK effectors from a single individual, but also show that the LAK effectors from some patients have an overall greater capacity to lyse C M L targets than those from other patients. Immunojuorescence analysis of IL-2-activated cells. Cells with N K and LAK activity can be divided into several subpopulations based on the expression of CD3 or CD8 in combination with either CD16, CD56, or both antigens.43 The possibility was considered that there may be differences in N K cell mediated lytic activity to C M L targets based on the number of N K cells or on the relative proportion of N K cell subpopulations in the cultures. We performed two-color immunofluorescence analysis of cells from eight patients with and without anti-CML activity before and/or after 4 to 7 days of culture. The IL-2-activated cells in these experiments were tested against homogeneous targets of myeloblasts isolated from a blast crisis C M L patient. In most of the five patients whose cells were phenotyped before and after culture, CD16+ cells and, to a greater extent, CD56+ cells were selectively expanded. The CD16+ cells nearly all coexpressed CD56 both before and after culture. However, after culture as many as 50% of the CD56+ cells were CD16-. There were marked differences in the proportion of

cells with N K phenotype in the cultures; however, these differences did not appear to correlate with the degree of anti-CML activity. Indeed, the culture that consisted of almost exclusively CD56' cells (patient no. 675) failed to lyse the C M L target, while the highest specific lysis was by a culture (patient no. 852) that contained only 16% CD56+ cells. We also did not detect a correlation between the proportion of CD3+CD56+,CD3+CD16+,or CD8+CD56+ cells and anti-leukemia activity (Table 6). Correlation of anti-CML activity with clinical features. We analyzed the data to determine whether LAK activity to noncultured C M L targets or cultured C M L targets (K562) was correlated with leukemic relapse. Of the 40 patients studied, nine patients subsequently relapsed with their leukemia (3 to 59 months after transplant), and five additional patients were found to have minor populations of hostderived Ph'+ cells in their marrow but had no clinical evidence of leukemia in the periphery. For this analysis, 210% lysis of host targets or of any allogeneic C M L target a t any time after transplant was considered positive for anti-leukemia activity. We found no correlation between clinical relapse and the degree of lytic activity to K562 targets ( N = 40) or the capacity to lyse fully allogeneic C M L targets ( N = 37). However, when lysis of host C M L targets ( N = 26) was examined there was a significant correlation with clinical relapse. IL-2-activated cells from 8 of the 9 patients who clinically relapsed had been tested against host C M L targets; cells from seven of these patients did not lyse host CML. In contrast, IL-2-activated cells from only 4 of 18 patients who have not experienced

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. 2257

ANTI-LEUKEMIA ACTIVITY POST BMT

Table 6. Phenotype of PBL Before and After Culture in Medium Containing IL-2

Posttransplant

Day

UPN

Mo

Tested

852

2.5

0 5 0 7 0 7 0 7 0 7 7 7 7

859

4.1

675

2.5

866

12.4

753

20.7

867 865 790

1.6 3.9 15.7

Lytic Activity"

CML

K562

25.5

68

0.0

9

7.1

154

4.4

64

5.5 15.6 11.4 0.0

170 270 1,000 158

CD3

CD8

CD20

CD16

CD56

CD16/ CD56

CD31 CD56

CD3/ CD16

CD81 CD56

83.7 57.3 58.2 59.0 11.3 1.5 64.5 76.2 49.0 23.9 13.1 18.5 28.7

64.3 69.0 55.9 68.2 21.1 38.2 40.0 75.0 28.0 29.2 42.4 49.0 43.6

36.7 NT 25.3 1.3 20.5 0.4 14.2 2.0 19.0 1.9 0.1 NT NT

6.0 17.0 10.2 16.6 46.0 65.1 14.5 13.3 25.0 42.9 46.0 39.4 19.4

3.4 15.7 13.1 31.6 61.4 96.6 14.8 22.8 23.8 67.9 84.9 70.6 38.2

1.4 10.0 8.4 13.7 45.6 73.6 8.8 8.6 19.0 36.8 41.4 35.3 17.4

1.9 5.0 1.0 2.7 2.3 1.0 2.4 10.4 0.7 2.9 3.3 1.9 0.8

1.3 6.0 0.2 0.7 0.9 0.3 1.3 2.3 1.7 2.0 0.8 0.4 0.5

NT 5.0 3.4 8.1 19.0 39.0 4.5 11.3 7.2 12.0 30.1 34.2 24.8

Cells were phenotyped by two-color immunofluorescencebefore and/or after culture for the indicated number of days in medium containing 250 U/mL IL-2. Data are expressed as the percentage of cells expressing the indicated antigen or combination of antigens. Abbreviation: NT. not tested. "Lytic activity was tested in 5'Cr release assays after culture. Lysis of fully allogeneic CML targets is expressed as percent specific lysis at an effector to target cell ratio of 100: 1. CML targets were from patient no. 79 1 in blast crisis CML. Lysis of K562 targets is expressed as lytic units/lOs effector cells.

clinical relapse failed to lyse host C M L targets (chi2 = 9.7, P = .0002) (Fig 3). Kaplan-Meier estimates of the probability of relapse for patients who did nor did not lyse host C M L targets are shown in Fig 4.The probability of clinical relapse for patients with cytolytic activity to host targets was 7% at 39 months after transplant compared with 72% for those who failed to lyse this target (P = .002). Classification into the category of patients without lytic activity did not simply reflect fewer tests to host targets in relapsed patients. Relapsed patients were tested 2.9 + 2.0 times (range 1 to 6 ) to host targets with consistent negative results, which is nearly identical to the mean number of tests performed in nonrelapsed patients, 3.1 + 2.1 (range 1 to 9) (Table 7). All four patients with minor populations of Ph'+ marrow metaphases but no evidence of peripheral disease had lytic activity to host-derived C M L targets. Not all patients with Ph' marrow metaphases experience clinical relapse; how+

301

20

1

ever, a significant proportion subsequently d0.44-46If all patients with cytogenetic relapse are considered as relapsed, the probability of relapse estimates becomes 38% for those with lytic activity and 67% for those without lytic activity to host C M L targets (P= 0.45) (not shown). A number of other factors must also be considered when evaluating C M L relapse post-BMT, including patient age, interval from diagnosis to transplantation, the type of transplant (T-cell depleted v T-cell replete), and most significantly, the disease Our study involves a status a t the time of tran~plantation.~-~.~' relatively small number of patients who are heterogenous for these characteristics. Listed in Table 7 are the clinical and experimental data for each of these patients. Because of the small numbers, statistically valid multivariate analysis could not be performed. However, it would appear from the data that as expected, 2 of the 3 patients with advanced disease (acclerated-phase CML, blastic CML) who were tested to host targets fell into the group of patients who experienced clinical relapse. If these patients (unique patient nos. [UPNs] Lytic Activity to Host Target (92.8%)

JI

C 0 0

4" 01 = 1

.-C

No Lytic Activity to Host Target (28.3%)

Q

:

c

Relapsed

5

20

$

0

0

Fig 3. The lytic activity of IL-2-activated patient PBMC against host-derived CML targets was tested after 3 to 7 days of culture. The patients are divided into two groups, those who subsequent to the testing relapsed with their disease (Relapsed). and those who have shown no evidence of peripheral disease to date (Not Relapsed). Data are expressed as the percent specific lysis at an effector to target cell ratio of 100 to 1.

9

18

27

36

45

54

Months After Transplantation Fig 4. Kaplan-Meier estimates of the probability of remaining in remission for patients with (N = 15) or without (N = 11) lytic activity to host-derived CML targets when tested during the posttransplant period. Tick marks indicate the last follow-up evaluation. P = .W2.

From bloodjournal.hematologylibrary.org by guest on July 10, 2011. For personal use only. HAUCH ET AL

2258

Table 7. Summary of Clinical and Experimental Patient Data Host

Current

UPN

Age

Disease.

lntervalt

Trans Typef

624 646 689 719 742 760 806 633 634 776 859 530 546 559 580 655 669 675 68 1 684 70 1 734 753 804 815 830 416 489 517 532 56 1 592 595 667 674 790 852 865 867 878

42 32 27

Acc CP Acc 2nd CP CP 2nd CP CP CP CP CP CP CP CP CP CP CP CP CP 2nd CP CP CP CP CP Blastic CP CP CP CP Acc CP 2nd CP CP CP CP CP CP CP CP CP CP

9.7 25.9 9.3 7.5 32.1 23.2 15.4 7.6 40.9 15.6 25.7 20.9 8.2 10.5 12.8 11.5 12.0 12.6 25.9 59.7 48.1 10.8 7.4 57.9 15.6 8.4 13.6 10.4 48.2 28.3 6.5 5.8 31.0 17.6 67.2 5.3 5.1 9.1 31.3 17.7

SBA-ES8A-ESBA-EConv Conv:3 SBA-ESBA-ESBA-E-:2 SBA-ESBA-ES8A-ES8A-ES8A-ES8A-ES8A-ESBA-:2 SBA-EF+T S8A- E- +T SBA-E- +T SBA-ESBA-ES8A-ESBA-ESBA-ESBASBA-ESBA-ES8A-ESBA-ESBA-EConv:2 S8A-ES8A-ES8A-E- +T SBA-EF+T S8A-ES8A-ES8A-ES8A-ES8A-E-

3 32 28 29 29 46 34 35 26 32 32 24 31 31 38 14 18 40 22 15 31 38 19 22 19 29 33 16 17 32 32 19 27 17 12 45 35

Status5

Re1 Re1 Re1 Re1 Re1 Re1 Re1

EXP A&W A&W A&W cyto cyto A&W A&W A&W A&W

EXP A&W Cyto Cyto A&W A&W A&W Re1 EXP Re1 A&W A&W A&W A&W A&W A&W EXP EXP A&W Cyto A&W EXP

A&W

Mod

8.7 7.6 2.6 9.2 20.2 12.6 17.2 18.8 42.0 24.0 11.8 25.7152.7 17.4132.0 49.5 47.2 38.9 37.6 16.6 36.4 23.8135.7 2.5133.6 29.0 27.0 20.3 8.7 6.1 59.1 57.8 54.1 52.5 10.0 45.3 45.4 3.2 3.3 22.4 4.4112.3 12.3 3.1 10.2

Target1

Allogeneic Test Mo.#

Taraetv

6.0 3.5-6.1 1.2-2.0 1 .O-4.3 0.9-4.7 1.5-12.1 1.O-6.1 9.5-12.9 8.6-14.7 0.7-6.8 2.2-3.2 20.0-42.5 23.6 14.3-25.0 19.8-3 1.7 1.5 0.7-6.2 0.9-3.5 0.8-3.1 0.8-9.2 0.8-8.3" 0.6-12.4 0.6-6.4 8.7- 13.3 3.4 0.9-1.3

Abbreviation: NT, not tested. 'Disease status at time of transplant: CP, first chronic phase: 2nd CP, second chronic phase: Acc, accelerated phase; Blastic, blastic phase. tTime from diagnosis of CML to initial BMT. $Type of last BMT: number of transplants. §Current status of the patient: Rel, clinical relapse: Cyto, cytogenetic relapse only: Exp, expired: A&W, alive and well. IIMonth of current status. For patients with cytogeneic relapse, first figure represents month Ph' chromosome was first detected: second figure represent current follow-up time. ncytotoxicity to CML targets. First figure represent number of assays in which cytolytic activity to CML target was 2 10%: second figure represents number of separate assays in which the targetk) was tested. #Test month for host CML targets. The first figure represents the months after transplant the target was first tested: the second figure represents the month after transplant the last assay was performed. ''The negative assay for patient 70 1 was at 0.8 months after transplant.

624,689,804) are excluded from the analysis, the probability of relapse for patients with and without lytic activity to host targets becomes 8 and 65% respectively (P = .01) (not shown). DISCUSSION

The anti-leukemic effects of an unmanipulated marrow allograft may be based on any or all of a series of factors,

including the intensity of the cytoreductive regimen used:' the type and duration of drugs administered to prevent or treat GVHD,'