A Functional Analysis of Hematopoietic Growth Factor Production from Class I and Class 11 Human Alloreactive T Cell Clones

Transcription

1 Original Paper International Journal of Cell Cloning 6: (1988) A Functional Analysis of Hematopoietic Growth Factor Production from Class I and Class 11 Human Alloreactive T Cell Clones finneth E Mangan", Adriana Zeevib, Rene Duquesnoyb, Richard K. Shadduct, Alan M. Gewirtz" 'Department of Medicine, Temple University School of Medicine, Philadelphia, Pennsylvania, USA; bdepartment of Pathology and Central Blood Bank, University of Pittsburgh, Pittsburgh, Pennsylvania, USA; 'Department of Medicine, Montefiore Hospital, University of Pittsburgh, Pittsburgh, Pennsylvania, USA Key Words. T cells Hematopoiesis Growth factors Abstract. Class I and Class II human alloreactive T cell clones or their conditioned media were mixed with progenitor cell-enriched null cells to assess their ability to stimulate human hematopietic progenitor cell (HPC) growth. Optimal release of erythroid, myeloid or megakaryocyte colony-stimulating factors occurred after 72 hours and required contact of cloned T cells with irradiated stimulator cells expressing the appropriate major histocompatibility complex (MHC) determinants recognized by the T cells. Individual clones were quite heterogeneous in their capacity to release hematopoeitic growth factors. Clones that produced optimal levels of factors that stimulated granulocyte-macrophage colony growth did not always produce equivalent amounts of factors that stimulated erythroid colony growth and vice versa when tested against identical target cells. Class II clones released nearly twice as much interleukin 3 activity as Class I clones. Class II clones that lacked cell-mediated lympholytic (CML) activity against B or T lymphoblastoid targets were consistent stimulators of HPC growth. In contrast, Class I or Class II clones that contained CML activity either poorly stimulated or inhibited HPC growth. These CML-positive clones produced greater amounts of gamma interferon. Our findings may have important implications for HPC growth following allogeneic mismatched bone marrow transplantation. Correspondence: Kenneth F. Mangan, M.D., Department of Medicine, Temple University School of Medicine, Bone M m Transplantation Program, 3400 N. Broad Street, Philadelphia, PA 19140, USA. Received May 23, 1988; provisionally accepted June 21, 1988; accepted for publication July 18, /88/$ oalphamed Press

2 Hematopiesis and Alloreactive T Cell Clones 325 Introduction It has been well established that T lymphocytes can elaborate factors that support or suppress hematopoietic progenitor cell (HPC) growth [l]. Most of these early observations were made employing lectin-stimulated T cells or T cells derived from malignant T cell lines [2-71. Other studies employed blood T cells of unknown antigen specificity [%lo]. Very little information is available concerning the release of HPC growth factors from antigen-specific clones [U]. Since T cells communicate with other cells in a major histocompatibility complex (MHC) restricted fashion [12], we have examined the release of hematopoietic growth factors from a series of human alloreactive clones triggered by specific alloantigens. We believe these studies may provide important and relevant clues regarding the regulation of hematopoiesis following a matched or mismatched allogeneic bone marrow transplant. In this report, we present a functional analysis of HPC growth factor production from a series of 12 Class I or Class II alloreactive T cell clones. The results reveal that individual clones are quite heterogeneous in their capacity to generate various HPC growth hctors. Striking differences in growth factor production were observed between T cell clones of various classes and cytotoxic phenotype. In addition, hematopoietic growth factor production could not be predicted by cell surface phenotype alone. Materials and Methods HLA Typing All blood and marrow donors gave informed consent for participation in these studies as approved by the Institutional Review Boards. All donors were human leukocyte antigen (HLA) typed: typing for antigens A, B and C was performed using the National Institutes of Health microlymphocytotoxicity technique with 144 alloantisera. Serologic typing of HLA-D region markers HLA-DR, DQ, DRw52 or DRw53 was performed by prolonged incubation in microlymphocytotoxicity tests with alloreactive sera and rabbit complement using B lymphocyte-enriched preparations as described [W]. Genemtion of Alloreactive T Cell Clones Peripheral blood mononuclear cells were isolated from heparin anticoagulated blood by Ficoll-Hypaque density gradient centrifugation. Alloreactive clones were generated according to previously described methods [14, 151. Briefly, 3-day unidirectional, mixed lymphocyte reaction- (MLR) activated cells were seeded in a soft agar culture medium. After 5-8 days, long-term cultures were generated from individual colonies in the presence of an interleukin 2 (IL-2) source and irradiated (4,000 rads) feeder cells derived from cells

3 Mangan/Zeevi/Duquesnoy/Shadduck/Gewirtz 326 initiating the primary MLR (original stimulators). IL-2 was derived from phytohemagglutinin- (PHA) activated lymphocytes as described [14, 151. The cloned cells were maintained through supplementation with IL-2 every 3 days and with 2.5 x lo5 irradiated feeder cells every 6 days. Specificiry Testing of Alloreactive T Cell Clones in Secondary Proliferation Assays The secondary proliferation of alloreactive clones was assessed in 3-day primed lymphocyte testing assays (PLT) as described [ In brief, 5 x lo3 cloned cells were incubated with 5 x lo4 irradiated peripheral blood lymphocytes (PBL) as stimulators. For each assay, a panel of HLA-typed stimulators was used in order to determine the HLA specificity of the alloreactive clones. Proliferation was assessed by 'H-thymidine uptake (1 pcilculture; Research Products International Corporation, Elk Grove Village, IL), which was added during the final 20 h of incubation. Each culture was harvested in a liquid scintillation counter. Cell Mediated Lympholytic (CML) Assay The cytolytic activity of alloreactive T cell clones was assessed as previously described [16]. Epstein-Barr virus transformed B lymphoblastoid cell lines (Human Genetic Mutant Cell Repository, Camden, NJ) were targets for Class I- and Class 11-specific CML cells. PHA-transformed T lymphoblasts were also targets for Class I cells. Cell suspensions of 1 x 106 cells and 0.1 ml of tissue culture media were labeled with 0.1 ml of Na2 51Cr04 (specific activity Cilg; New England Nuclear, Boston, MA) for 90 min at 37 C and washed twice prior to resuspension in tissue culture media. In each CML assay, effector cells were mixed with SICr-labeled target cells at a ratio of lo:]. Following incubation at 3PC for 4 h, the plates were centrifuged at 200 g for 5 min, and the supernatants were sampled and counted in a gamma counter. The percentage specific 51Cr release was calculated using the following formula: % = experimental 51Cr release - spontaneous 51Cr release Total 510 release - spontaneous 5ICr release Spontaneous release was determined by incubating target cells in medium alone. Total release was determined by incubating target cells with 5% triton X (Fisher Scientific Co., Chicago, IL). Isolation of Progenitor Cell-Enriched Null Cells Null cells enriched in erythroid burst-forming units (BFU-e) or granulocyte/macrophage colony-forming units (CFU-gm) were prepared from starter populations of blood mononuclear cells as previously described [9]. In brief, blood mononuclear cells obtained by Ficoll-Hypaque centrifugation were depleted of adherent monocytes by incubation in fetal calf serum (FCS; Hyclone, Logan, UT) coated in plastic Petri dishes (Fisher Scientific, Pittsburgh, PA) for 1 to 2 h, followed by removal of B cells by panning with anti-immunoglobulin-coated Petri dishes. The non-adherent B-depleted cells were then rosetted with 2-aminoethylisothiouronium bromide-treated sheep erythrocytes (Rockland, Inc., Cochranville, PA) and subjected to a second Ficoll-Hypaque density gradient centrifugation. Progenitor-enriched null cells that remained at the interface contained less than 5% B or T cells as judged by immunofluorescence with pan-t cell (OKTII) or pan-b cell (Leul2) monoclonal antibodies. They contained less than 2% monoqtes as judged by morphology

4 Hematopiesis and Alloreactive T Cell Clones 321 or staining with alpha napthyl acetate esterase. Immunofluorescent analysis of cloned T cells and target null cells was performed employing the Orthospectrum I11 flow cytometer (Westwood, MA). All cells (> 95% viable) were washed 3 times in alpha minimum essential medium prior to use in cultures. BFU-e Assay Day 14 blood BFU-e were assayed in a methylcellulose erythroid culture system [17]. In brief, 2 x lo5 target null cells were mixed with 2 x lo4 to 5 x lo4 cloned T cells in 0.8% methylcellulose (Fisher) containing 1 x mercaptoethanol (Fisher), 1 % deionized bovine serum (Sigma Chemical Co., St. Louis, MO) albumin and 30% horse serum (Hyclone). Each dish containined 1.0 IU/ml of human urinary erythropoietin (specific activity, 500 IUlmg of protein; Alpha Corporation, Los Angeles, CA). Conditioned media derived from the MO hairy cell line [7] (courtesy of Dr. David Golde, UCLA, Los Angeles, CA) was used as a source of burst-promoting activity (BPA) and was added to control cultures only at a final concentration of 10%. These cultures were used as a positive control for burst-promoting activity (BPA). In some studies 2 X lo5 cloned T cells were pre-incubated with 1 x lo6 irradiated (2,000 rads) PBL derived from the original stimulator or from unrelated stimulator cells not expressing the relevant HLA determinant. The cells were incubated for 3 h in a 5% C02 humidified atmosphere before addition to null cell targets. In other studies, irradiated stimulator PBL were added directly to culture dishes containing cloned T cells and null cells. Triplicate cultures per variable were incubated for 14 days at 37 C in a 5% C02 humidified atmosphere then stained in situ with benzidine. Benzidine-positive colonies containing greater than 50 cells per aggregate were scored as BFU-e. CFU-gm Assay Committed myeloid progenitors (CFU-gm) were assayed from null cell targets in the bilayer agar gel system as previously described [18]. Control cultures contained a standard source of granulocytemacrophage colony-stimulating factor (GM-CSF) derived from the DET fibroblastic cell line in a 0.5% agar underlayer [19]. GM-CSF was omitted in all other cultures. Underlayers contained cloned T cells with or without irradiated original stimulator or unrelated stimulator PBL (4 x lo5). Overlayers contained 1 X lo5 to 2 x lo5 null target cells in 0.3% agar. After days in a 7.5% C02 atmosphere at 3TC, colonies of greater than 50 cells each were numerated in situ employing an inverted microscope. CFU-meg Assay Megakaryocyte colonies were grown in plasma clot cultures as previously described [20]. Target cells were light-density bone marrow mononuclear cells plated at a density of 5 x 105 cells/dish. Control cultures contained a reference source of megakaryocyte colonystimulating factor (Meg-CSF) derived from 10% (vlv) aplastic anemia serum. Megakaryocyte colonies were enumerated using a rabbit anti-human platelet glycoprotein anti-serum and an indirect immunofluorescence assay as described [20]. A cluster of three or more intensely fluorescent cells was scored as one colony. The cultures were carried out in a 5% C02 atmosphere for a period of 12 days. Preparation of T Clone-Conditioned Media Cloned T cells (1 x lo5) were cultured alone or in the presence of 1 X lo5 irradiated

5 ManganlZeevilDuquesnoy/Shadduck/Gewirtz 328 original stimulator PBL for h at 37 C in a 5% C02 atmosphere in RPMI 1640 medium (GIBCO Laboratories, Grand Island, NY) containing 10% human serum. In other experiments, irradiated B lymphoblastoid cells expressing appropriate HLA antigens were used as stimulators. At the end of the incubation time, conditioned media were assayed for the presence of BPA, GM-CSF or Meg-CSF at 10-15% concentrations. Activity was assessed employing target null cells or bone marrow cells prepared as described above. Controls consisted of media alone or peripheral blood stimulators alone. As controls for any contributions of growth factors provided by the original stimulator cells (O.S.) alone or the medium alone, data was expressed as follows: Colonies % Reference Control = [loo%][xi where x = [#colonies (clone + O.S.)] - [#colonies (O.S.)] [#colonies (reference BPA, GM-CSF, Meg-CSF)] - [#colonies (media)] Interleukin 2 Assay Interleukin 2 (IL-2) was assessed in 72-hour samples as described [16]. Alloreactive IL-2-dependent T cell clones (1 x LO4) were incubated with media alone, control or 100 pl of the samples to be tested for IL-2 activity. After 3 days of incubation, the amount of proliferation was assessed by 'H-thymidine incorporation. Gamma-Interferon (IFNy) Assay IFNy levels were assessed in media conditioned by T cell clones employing the Centocor radioimmunoassay kit (Centocor Inc., Malvern, PA) exactly as described by the manufacturer. Interleukin 3 Assay Interleukin 3 (IL-3) was detected as described by using an IL-3-sensitive mouse cell line (DA-I; a gift of Dr. James Ihle, Frederick Cancer Research Institute, Frederick, MD) [21]. In brief, 100 pl of test supernatant was added to 5 X lo4 DA-1 cellslwell, incubated for 10 h at 37"C, 5% C02 and then pulsed with 1 pci of 3H-thymidine. The cultures were harvested 4 h later by a Skatron Auto Harvester (Sterling, VA) and counted on a Packard scintillation counter (Sterling, VA). Control cultures were tested in triplicate with recombinant human IL-3 (Genzyme, Boston, MA). Stutistics A comparison of cohorts was made employing the Student's t test. Results Specificity Testing and CML Activity of T Cell Clones Three separate MLR combinations (responder cells and irradiated original stimulator cells) were employed to generate 12 separate alloreactive T cell clones. The HLA phenotypes of the responder-stimulator combinations are shown in Table I. Each clone demonstrated significant secondary proliferative responses (using the PLT) toward its original stimulator (Table 11). The PLT specificity of

6 ~~~ Hematopiesis and Alloreactive T Cell Clones 329 Table I. HLA phenotypes of MLR combinations used to generate T cell clones Responder-Stimulator Cell Clones Combinations HLA Phenotypes A B DR DQ DRw # 1-2 HA 2,32 7,52 2,- 1,- 52,- AZ' 3,24 7,35 6,7 1,2 52,- # 3-6 DB 2,3 44,62 4,- 3,- 53,- DT" 3,25 18,44 2,7 1,2 52,53 # 7-12 AZ 3,24 7,35 6,7 1,2 52,- RS' 2,- 57,62 6,7 1,2 52,53 *Indicates donor source of original irradiated stimulator cells. The hyphen (-) indicates the absence of a phenotype. lslble II. Characteristics of alloreactive T cell clones Secondary proliferative responseb CML' T cell # Clone Class PLT' (CPM X 10') (%) phenotyped 1 HA-6 II DR (-1 T4 2 HA-18 II DR T4 3 DB-7 II DR (-1 T4 4 DB-22 II DR (-1 T4 5 DB DRw (-1 T4 6 DB-29 II DQwl 29.5 (-1 T4 7 AZ DR (-1 T4 8 AZ-18 I B (-) T8 9 AZ-20 I B T8/T4 10 AZ-39 I B T4 11 AZ-40 I A T4 12 AZ-43 n DR T4 "Specificity of clone was determined by employing a panel of HLA-typed donor lymphocytes as stimulators. b5 X 103 cloned T cells were incubated with 5 X lo4 irradiated original stimulator cells and thymidine uptake was measured. Data indicates count per minute (CPM) employing original stimulators. 'CML activity against B lymphoblastoid cells or PHA-transformed T lymphoblasts as described in Methods. (-) indicates no CML activity detected. dt cell phenotype as determined by flow cytometry employing OKT4 (helperhducer) or OKT8 (suppressorkytotoxic) T cell monoclonal antibodies. Clones were > 99% positive. Clone #9 expressed both T8 and T4 antigens (>90% positive).

7 Mangan/Zeevi/Duquesnoy/Shadduck/Gewirtz 330 each clone was determined by employing a panel of HLA-typed donor lymphocytes as stimulators, previously described above. As summarized in Table 11, the PLT specificity of the 12 clones was associated with a variety of Class I and Class I1 alloantigens. Each clone was also tested for its ability to lyse Epstein-Barr virus-transformed B lymphoblastoid cell lines (CML activity). Clones 2, 9, 10, 11 and 12 all displayed CML activity in a T r release assay. These clones were termed killer clones. Flow cytometric analysis indicated that most clones expressed the OKT4 helpedinducer T cell antigen. However, clone 8 was OKT8+ and clone 9 equally expressed ( > 90%) both antigens (Table II). Release of Hematopoietic Growth Factors from T Cell Clones is Triggered by Contact with Stimulator Cells Expressing Appropriate HLA Antigens Mitogen-stimulated or unstimulated blood T cells may enhance the growth of erythroid or myeloid progentior cells from autologous or allogeneic null cells. However, the physiologic significance of these observations in the context of a specific immune response is uncertain. To examine the specificity of this T cell function and the role of alloantigens in triggering T cell release of hematopoietic growth factors, pure populations of cloned T cells were co-cultured with BFU-e enriched target null cells in the absence of an exogenous burst-promoting source. Null cells were obtained from the same donor source (autologous responder) as the clones. The data were expressed as the % reference control containing MO cell BPA. Typical experiments in the BFU-e cultures showing the effects of two T cell clones (#3 and #6) with PLT specificities for the Class I1 antigens DR7 and DQwl, respectively, are given in Figure 1. Addition of 5 X lo4 T cells from clones (Panel C) or 1 x lo5 irradiated (2,000 rads) blood lymphocytes (Panel B) from original stimulators to 1 x los null cells resulted in little or no release of BPA when compared to cultures containing null cells alone (Panel A). In contrast, addition of clones to original stimulator PBL expressing DR7 or DQwl antigens triggered release of a BPA equivalent to twice reference control (Panel D). PBL which did not express DR7 or DQwl determinants, however, failed to trigger BPA release (Panels E and F), i.e., the release of BPA was HLA restricted. Irradiation of T cells prior to co-incubation with irradiated original stimulators abolished the stimulatory effects suggesting that growth factors were being released from T cells rather than the irradiated stimulator cells (Panel G). To further exclude the possibility that BPA was released from the irradiated stimulator cells, T cell clones were stimulated by irradiated pure B lymphoblastoid cells expressing appropriate HLA antigens, and their conditioned media were assayed for BPA at a final concentration of 15% in culture. As shown in Table 111,

8 Hematopiesis and Alloreactive T Cell Clones 331 Fig. 1. Effect of cloned T cells (5 X lo5) from a DR7 or DQwl-specific clone on the growth of BFU-e from 1 X lo5 autologous null cells cultured in the presence of 1 X lo5 original or 1 x 105 non-reactive irradiated stimulator cells. Only original stimulators expressed HLA antigens DR7 and DQwl. Data represents mean f SD of triplicate cultures expressed as % reference control containing exogenous BPA source from MO cellconditioned media. Exogenous BPA was omitted from cultures shown in panels A-G. Sixteen BFU-e proliferated in 1 X lo5 null cells containing reference BPA. BPA was released into cultures when cloned T cells were exposed to original stimulators expressing appropriate. HLA antigens (panel D) but not from stimulators expressing irrelevant HLA antigens (DR4, DQw3, panel F). Little or no BPA was released from null cells alone (panel A), clones alone (panel C) or irradiated stimulators alone (panels B and E). Irradiation of T clone (2,000 rads) before culture with original stimulators eliminated BPA release (panel G). release of BPA required contact of B lymphoblastoid cells with the T cell clone (BPA was 87% of MO cell BPA control). In contrast, little or no BPA was released in media conditioned by either T cell clones alone or by the irradiated B lymphoblastoid cells alone. Variable Production of Soluble Hematopoietic Progenitor Cell Growth Factor@) from Class I and Class II Alloreactive T Cell Clones Since soluble hematopoietic growth factors were released when T clones were triggered by irradiated stimulator cells expressing appropriate HLA antigens, con-

9 Mangan/Zeevi/Duquesnoy/Shadduck/Gewirtz 332 able IJI. Release of burst-promoting activity into conditioned media from a DQwl-specific clone stimulated by irradiated DQwl-specific B cells Cell source for conditioned media BFU-e/l X lo5 null cells' % MO cell control None 9 f 4 MO cell control 30 f 5 Cloned T cells 9 f 4 irradiated B lymphoblastoid cells 10 f 4 Cloned T cells and irradiated B lymphoblastoid cells 26 f 3b "Conditioned media was tested at a final concentration of 15% against progenitor cellenriched null cells. Values indicate mean f SD for quadruplicate plates. % c.01, versus media control 87 ditioned media from various stimulated clones were tested for their ability to support myeloid or erythroid colony growth. Initial studies were carried out to determine the kinetics of growth factor production from stimulated T cell clones. Peak release of BFU-e- and CFU-gm-enhancing factors occurred at 72 h (data not shown). Little or no constitutive release of growth factors was observed in media conditioned by T clones alone or by irradiated stimulator cells alone. Assays for IL-2 employing IL-2-dependent T clones revealed that IL-2 levels had fallen to nearly background levels in 72-hour-conditioned media samples, excluding the possibility that this factor accounted for stimulation of myeloid or erythroid colonies (Table IV). The stimulatory (or inhibitory) effects of T cell-conditioned media from 12 separate clones on BFU-e and CFU-gm growth is shown in Figure 2. Conditioned media were tested in triplicate or quadruplicate dishes at a final concentration of 15% against null cell targets derived from a single donor. As controls for the potential contribution of original irradiated stimulator cells alone or the effects of the FCS, control-conditioned media was prepared and tested separately. Colony counts were adjusted as described in Materials and Methods to eliminate contributions of growth factors from stimulator cells or FCS. Eight Class Il clones and four Class I clones were tested in BFU-e and CFU-gm assays. Five clones showed killer activity toward B or T lymphoblastoid targets (Table II). As shown in Figure 2, there was marked variability between clones and within individual clones in their ability to support either erythroid or myeloid colony growth. Clones, such as clone number one which was Class I1 DR6-specific, was a strong erythroid

10 Hematopiesis and Alloreactive T Cell Clones 333 'Igble IV. Detection of IL-2 in T cell clone-conditioned media Cell source for conditioned mediaa Incubation times 24 Hoursb 72 Hours HA-18 clone HA-18 clone + S* DB-29 clone DB-29 clone + S* (CPM) , ,509 (CPM) 458 1, 'conditioned media from clones #2 (HA18) and #6 (DB29) as prepared in the presence or absence of original stimulators (S*) for 24 or 72 h incubation times. 100 pl of conditioned media was then added to 1 X lo4 IL-2-dependent T cell clones and 3H-thymidine uptake was measured. %ata represents means of triplicate cultures expressed as counts per minute (CPM). enhancer, but did not stimulate myeloid colony growth as well. In contrast, clone number 6, which was a Class I1 DQwl clone, was a potent myeloid colony stirnulator but a weak erythroid colony stimulator. Conditioned media derived from clone 5 also enhanced CFU-meg growth seven-fold in comparison to an aplastic anemia serum control (data not shown). As a group, Class I1 clones were stronger enhancers of erythroid and myeloid colony growth than Class I clones (p <.05). Additional studies were carried out to determine whether release of an erythroid BPA was due to release of rnultipotential growth factor IL-3 (Table V). Clones 1 and 3, which were excellent supporters of BFU-e (Fig. 2), also released the highest levels of IL-3 as assessed by assays employing the IL-3-sensitive line, DA-I. Clone 5 was the strongest producer of IL-3, but caused only modest stimulation in the BFU-e assay, suggesting that IL-3 was not completely responsible for this activity. Clones 2, 6, 8, 9, 10 and 12, which all showed minimal activity (less than the reference BPA control in the BFU-e assay, Fig. 2), also produced very little IL-3 activity in the DA-1 assay (159% f 47%, media control). In comparison to Class I clones, Class Il clones released nearly twice as much IL-3 (266% f 102% control vs. 142% f 65% control, p <.05, Table V). Clones 1 and 6 were also exposed to the mitogen PHA for 72 h, and IL-3 production was cornpared to IL-3 production following alloantigen stimulation. No significant differences were observed in IL-3 output under these conditions. Comparison of CML-Positive Clones with CML-Negative Clones Most Class I clones dislayed CML killer activity, whereas only two of eight

11 ManganlZeevilDuquesnoy/ShadducklGewirtz 334 Fig. 2. The effects of conditioned media prepared from Class I or Class II alloreactive T cell clones on growth of BFU-e or CFU-gm from 1 X 105 nu11 cells. Clones were exposed to original stimulators for 72 h, and media was tested at a final concentration of 15% (v/v). No exogenous source of BPA or GM-CSF was present in test cultures. Data represents mean f 1 SD of triplicate plates expressed as A reference controls containing exogenous sources of BPA or GM-CSF as described in Materials and Methods. Reference control plates corresponding to 100% (line) contained an average of 16 BFU-ell X los null cells and 76 CFU-gdl X lo5 null cells. The PLT specificities, class and cell-mediated lympholytic activity of each clone is shown. Class I1 clones were CML-positive. As a group, in comparison to conditioned media from CML-negative clones, conditioned media from CML-positive clones were either poor stimulators or inhibitors of BFU-e or CFU-gm growth (p <.05, Fig. 3). Clones seven and twelve afforded a unique opportunity to compare the possible influence of CML activity and HPC growth-enhancing function in clones which recognized identical HLA antigens. Both clones were Class II DR6-specific, however, clone 7 was CML-negative and clone 12 was CML-positive. As shown in Figure 4, conditioned media from the DR6 non-killer clone enhanced erythroid, myeloid and megakaryocyte colony growth (in comparison to reference controls), whereas the DR6 killer clone was inhibitory in all three progenitor cell assays.

12 ~~~ Hematopiesis and Alloreactive T Cell Clones 335 'Lgble V. 1L-3 production by human alloreactive clones # Clone Class PLT 3H-thymidine uptake" 1 HA-6 II DR6 329' (311)b 2 HA-18 II DR DB -7 II DR7 356' 5 DB DRw52 388' 6 DB -29 I1 DQwl 169 (200)b 12 A DR6 199' 8 A2-18 I B62 218' 10 A2-39 I B A2-20 I B '100 p1 of 72-hour supernatant from alloantigen-stimulated clones were added to 5 x lo4 IL-3-dependent DA-1 cells, incubated overnight and pulsed with 1 pci 3H-thymidine. Results are expressed as percent media controls. Each supernatant was tested in triplicate. Controls yielded a mean of 12,195 CPM. Addition of 20, 100 and 200 IU human recombinant IL-3 yielded 20,270, 23,247 and 32,483 CPM, respectively. qalues in parentheses derived from T clones stimulated with PHA for 72 h. 'p <.05 vs. media control IFNy levels, as measured by radioimmunoassay, were 15 IU/ml from the DR6 killer clone and only 2 IU/ml from the DR6 non-killer clone. In other studies, clone 10, which was also CML-positive, released 17 IU/ml; whereas, CMLnegative clone I released only 4 IU of IFNy per ml. Discussion In the present in vitro model of hematopoiesis, we have shown that essentially pure populations of well-characterized alloreactive-cloned T cells generated growth stimulatory and/or inhibitory factor@) for erythroid, myeloid or megakaryocyte progenitor cells when triggred by irradiated stimulators (PBL or B lymphoblastoid cells) expressing appropriate HLA antigens recognized by the clone. In contrast, irradiated stimulator cells that did not express HLA antigens recognized by the clone failed to induce a release of hematopoietic progenitor cell (HPC) growth factors. Antigen-specific cloned T cells were reported to release granulocyte and eosinophilic progenitor cell CSF in an HLA-restricted fashion when triggered by appropriate antigens in the presence of appropriate antigen-presenting cells [ll]. The present studies extend these observations to erythroid and megakaryocyte progenitor cells and further suggest that Class I and Class I1 alloantigens on accessory

13 Mangan/Zeevi/Duquesnoy/Shadduck/Gewirtz 336 Fig. 3. Comparison of the effects of conditioned media prepared from Class I and II killer (CML+) and non-killer (CML-) clones on BFU-e and CFU-gm growth from null cells. Data represents mean f 1 SEM of 7 CML- clones and 5 CML+ clones expressed as % reference controls. Conditioned media for each clone was tested in triplicate or quadruplicate as described in Materials and Methods. CML+ clones provided little or no growthenhancing effect for BFU-e or CFU-gm compared to CML- clones (p <.05). cells may act as triggers for the release of HPC growth factors [22]. While we cannot totally exclude a contributon of growth factor production from irradiated stimulator cells in these experiments, the majority of evidence suggests that the T cell clones were the source of growth factors, rather than the irradiated stimulator cells, since irradiation of T cell clones prior to co-culture with appropriate stimulator cells abolished their growth factor-producing capacity. Furthermore, pure populations of B lymphoblastoid cells which were employed as stimulators did not release growth factors into conditioned media (Table 111). Media conditioned by irradiated stimulator cells supplied little or no support of HPC growth. Finally, in the conditioned media experiments, data was expressed

14 Hematopiesis and Alloreactive T Cell Clones 337 Fig. 4. Comparison of the effects of a conditioned media prepared from a DR6 killer (CML+) and DR6 non-killer (CML-) clone or BFU-e, CFU-gm and CFU-meg growth. Conditioned media were tested at 15% (v/v) for CFU-gm and BFU-e and 10% (v/v) for CFU-meg. Data represents mean f 1 SD for triplicate or quadruplicate plates expressed as 46 reference controls as described in Materials and Methods. Sixteen BFU-e/l X 105 null cells, 76 CFU-gm/l X lo5 null cells and 26 CFU-rnegb x lo5 marrow mononuclear cells were present in reference controls. as percent reference control to minimize any potential contribution from irradiated stimulator cells. Earlier studies suggested that surface phenotype, i.e., OKT4+ or OK=+, might predict the growth stimulatory or inhibitory effects of T cells in progenitor cell assays [ The present studies, however, if confirmed by additional studies with OKT8+ clones, suggest that the growth stimulatory or inhibitory effects of pure populations of T cells on progenitor cells may not be reliable predictions based on surface phenotypic differences alone. For example, clones 8 and 9 which expressed the OKT8' antigen could release modest amounts of erythroid BPA and granulocyte CSF into conditioned media. On the other hand, growth stimulatory effects of cloned T cells in myeloid and erythroid colony assays appear to corre-

15 ManganIZeevilDuquesnoylShadducklGew irtz 338 late well with the presence of CML activity of the T cell clones. This effect was observed whether the clone was Class I or Class 11. The lack of stimulation by these clones may be due to their failure to generate optimal amounts of CSF or to the release of competing inhibitory lymphokines such as IFNy (Fig. 4). Both CML-positive clones tested released 5-7 times as much IFNy as that observed using two CML-negative clones. Virtually all Class I and Class 11 OKT4' or OKT8' cloned T cells released a factor(s) which supported erythroid or myeloid progenitor cell growth or both. We noted, however, a marked heterogeneity in the capacity of individual clones to release BPA or granulocyte CSF when tested under identical conditions. Thus, each clone appeared to have its own unique functional profile. Similar observations have been reported employing murine alloreactive T cell clones [29]. This heterogeneity was in part due to the variable release of the multipotential growth factor IL-3 (Table V). However, discordant stimulation of BFU-e and CFU-gm growth (clone 6) was also noted. This observation suggests that more than one IL-3 growth factor and GM-CSF may be triggered by alloantigens simultaneously. The variable effects of clones were not likely due to release of IL-2, since this lymphokine was nearly depleted in the 72-hour supernatants tested ('hble IV). However, competing inhibitory lymphokines such as IFNy, tumor necrosis factor or others may also have contributed to some of these differences as noted above. Our observations may have important implications for growth of progenitor cells in the setting of allogeneic mismatched bone marrow transplants. These studies suggest that expansion of some alloreactive T cell clones may have more favorable effects on progenitor cell growth than others. In particular, clones with CML activity would appear to inhibit progenitor cell growth and may be involved in graft rejection or suppresive mechanisms during graft-versus-host disease. Understanding the factors which lead to the proliferation of CML-positive T clones may prove important in understanding disturbed hematopoiesis following a specific immune response or following mismatched bone marrow transplantation. Acknowledgments We are grateful for the technical assistance of Mark Mullaney, Donna Hodge. Loretta Divechiu and the secretarial assistance of Kathy Donnelly. Supported in part by Grants AI21410, AI18923, AI23303, (2, and ( from The National Institutes of Health. Dr. Gewirtz is the recipient of a Research Career Develop ment Award from the National Institutes of Health (CAOl324).

16 Hematopiesis and Alloreactive T Cell Clones 339 References Mangan KF. Immune disregulation of hematopoiesis. Ann Rev Med 1987;38: Banisadre M, Ash RC, Ascensao JL, Kay NE, Zanjani ED. Suppression of erythropoiesis by mitogen activated T lymphocytes in vitro. In: Baum SJ, Ledney GD, Khan A, eds. Experimental Hematology Today. Basel: S. Karger, 1981: Zuckerman KS, Haak M. Mitogen induced stimulation and suppression of erythroid burst promoting activity production of human mononuclear cells. Br J Haematol 1983 $5: Talley R, Rinehart J, Balcerzak SP. Autologous and allogeneic suppressor lymphocyte inhibition of human erythroid colony formng unit proliferation. Exp Hematol 1982;10: Sakamaki H, Hamaguch H, Fomsama S, Shishido H. Inhibitory effect of PWMstimulated OKT4+ subsets on erythro-, granulo-, and megakaryocytopoiesis in vim. Br J Haematol 1985;61: Hamburger A. Enhancement of human erythroid progenitor cll growth by media conditioned by a human T-lymphocyte line. Blood 1980;56: Golde DW, Bersch N, Quan SG, Lusis A. Production of erythroid potentiating activity by a human T-lymphoblast cell line. Proc Natl Acad Sci USA 1980;1: Nathan D, Chess L, Hillman D, et al. Human erythroid burst forming unit: T cell requirement for proliferation in vitro. J Exp Med 1986; , Mangan KF, DesForges J. The role of T lymphocytes and monocytes in the regulation of human erythropoietic peripheral blood burst forming units. Exp Hematol 1980;8: Torok-Storb B, Martin P, Hansen J. Regulation of in vitro erythropoiesis by normal T cells: two T cell subsets with opposing function. Blood 1981;58: Griffin J, Mever S, Schlossman S, Reinherz E. T cell regulation of myelopoiesis: analysis at a clonal level. J Immunol 1984;133: Schwartz R. T lymphocyte recognition of antigen in association with gene products of the major histocompatibility complex. AM Rev Immunol 1985;3: Zeevi A, Scheffel C, Annen K, et al. Association of PLT specificity of alloreactive lymphocyte clones with HLA-DR, MB and MT determinants. Immunogenetics 1980 ;16: Zeevi A, Chiu K, Duquesnoy R. Functional differences between cells from MLRcolonies grown in soft agar cultures. J Immunol 1980;l25: Duquesnoy R, Zeevi A. Immunogenetic analysis of the HLA complex with alloreactive T cell clones. Human Immunology 1983;8: Zeevi A, Duquesnoy R. Specificity of alloactivated human T lymphocyte clones in secondary proliferation, cell mediated lympholysis and interleukin-2 release. J Immunogenetics 1985;E : Ogawa M. Methylcellulose clonal cell culture of murine and human erythropoietic precursors, In: Murphy MJ, ed. In Vitro Aspects of Erythropoiesis. New York: Springer-Verlag, 1978:248. Shadduck R, Winkelstein A, Zeigler Z. Aplastic anemia following infectious mononucleosis: possible immune etiology. Exp Hematol l979;7:

17 ManganlZeevilDuquesnoyIShadducklGew irtz Inoue S, Ottenbreit M. Heterogeneity of human colony forming cells. Blood 1978;51: Gewirtz A, Xu WY, Mangan KF. Role of natural killer cells in comparison with T lymphocytes and monocytes in the regulation of normal human megakaryocytopoiesis in vitro. J Immunol 1987;139: Ythier A, Abbud-Filho M, William JM, et al. Interleukin 2-dependent release of interleukin 3 activity by T4+ human T cell clones. Proc Natl Acad Sci USA 1985; 82 : Torok-Storb B, Hansen J. Modulation of in vitro BFU, growth by normal Ia-positive T cells is restricted by HLA-DR. Nature 1982;298: Mangan KF, Chikkappa G, Bieler L, Scharfman W, Parkinson D. Regulation of human blood erythroid burst forming unit (BFU,) proliferation by T lymphocyte subpopulations defined by Fc receptors and monoclonal antibodies. Blood 1982;59: Haq A, Rinehart J, Balcerzak S. T cell subset modulation of blood erythroid burst forming unit proliferation. J Lab Clin Med 1983;101: Wisniewski D, Strife A, Wachter M, Clarkson B. Regulation of human peripheral blood burst-forming unit growth by T lymphocytes and T lymphocyte subpopulations defined by OKT4 and OKT8 monoclonal antibodies. Blood 1985;65: Chikkappa G, Phillips F! Regulation of normal human blood neurophilic, macrophagic and eosinophilic committed stem cell proliferation by autologous blood T lymphocyte subsets. Blood 1984;63: Hesketh P, Sullivan R, Valeri R, McCarrall L. Production of granulocyte-monocyte colony stimulating activity by isolated human T lymphocyte subpopulations. Blood 1984;63: Bagby G, Rigas V, Bennett R, Vanderbard A, Garewal H. Interaction of lactoferrin, monocytes and T lymphocytes subset regulation of steady-state granulopoiesis in vitro. J Clin Invest 1981;68: Sanderson C, Strath M, Warren D, Ogawa A, Kirkwood T. The production of lymphokines by primary alloreactive T clones: a coordinate analysis of 233 clones in seven lymphokine assays. Immunology 1980;56: