Bone Marrow Transplantation, (1999) 23, 1013–1017  1999 Stockton Press All rights reserved 0268–3369/99 $12.00 http://www.stockton-press.co.uk/bmt Expression of activation antigen CD134 (OX40) has no predictive value for the occurrence or response to therapy of acute graft-versus-host disease in partial T cell-depleted bone marrow transplantation

APA Gadisseur1, JW Gratama2, C Lamers2, JWJ van Esser1, RLH Bolhuis2 and JJ Cornelissen1

Departments of 1Hematology and 2Immunology, Daniel den Hoed Cancer Center, Rotterdam, The Netherlands

Summary: defined as an outpouring of endogenous resulting in an attack of various tissues by antigen-spe- CD134 (OX40) is a member of the tumor necrosis factor cific and non- specific effector leukocyte populations.3,4 family which is expressed by activated T lymphocytes. Acute GVHD is preferably diagnosed on the basis of CD134 expression on T cells was monitored during the clinical signs together with biochemical data (liver first 35 days post-transplant in 14 patients, receiving enzymes, bilirubin).5 Other ways of monitoring and pre- either an HLA-identical sibling bone marrow transplant dicting acute GVHD have been proposed, such as the (BMT), a matched unrelated transplant (MUD-BMT) or repopulation pattern of the T lymphocytes.6 The rate of an autologous peripheral blood progenitor cell trans- CD8+ T cell repopulation and the expression of T cell plant (PBPCT). The sibling and unrelated grafts were + activation markers such as CD25 have been shown to pre- partially depleted of T cells. CD134 expression on CD4 dict GVHD in some patient cohorts.7,8 T cells peaked between 7 and 14 days after BMT, with + CD134 (OX40) is a 50-kDa type I transmembrane pro- a mean peak value of 45% of CD4 cells (range 26–70%) tein that is expressed on activated CD4+ T lymphocytes over all three patient groups. The observed pattern of + 9 + + and to a lesser extent on CD8 T lymphocytes. It is a CD4 CD134 expression, an increase during the first 2 member of the tumor necrosis factor (TNF)/nerve growth weeks post-BMT followed by a gradual decline towards factor (NGF) family. CD134 is rarely expressed by lym- values of 15–40%, was similar in all groups. No differ- phocytes under normal conditions and, if it is, mainly by ence in the kinetics of CD134 expression by CD4+ T cells T lymphocytes residing in the lymph nodes. The ligand was observed between the patients that did or did not (L) for CD134 is a gp34 membrane bound , struc- develop graft-versus-host disease (GVHD), nor did the turally similar to other ligands in the TNF receptor family clinical effect of any treatment given for GVHD corre- late with alterations in CD134 expression by CD4+ T with a still incompletely documented tissue distribution + + (activated B cells, vascular endothelial cells, heart, skel- cells. Absolute CD4 ,CD134 T cell numbers showed a 10 more rapid increment after autologous PBPCT than etal muscle, testis). The CD134/CD134L (gp34) system after sibling or MUD transplants. We conclude that is involved in inducing T cell-dependent expression of CD134+ by CD4+ T lymphocytes cannot responses, differentiation of activated B cells into immun- serve as a surrogate marker for allo-reactivity. CD134+ oglobin producing plasma cells, and ‘direct’ adhesion of 11–13 expression may reflect lymphocyte regeneration, rather activated T cells to vascular endothelial cells. Studies than alloreactivity. in animal models have shown enhanced expression of Keywords: graft-versus-host disease; T cell; CD134 anti- CD134 by lymphocytes in the blood after bone marrow gen; bone marrow transplantation; reconstitution transplantation as early as 5 days post-BMT and peaking at 12 days post-BMT.14 As the increment in CD134 expression ran parallel to the development of GVHD in the rat model, it was suggested that CD134+ T lympho- Graft-versus-host disease (GVHD) is still a major and cytes may serve as a possible target for the therapy of potentially lethal complication of allogeneic bone marrow acute GVHD. Preliminary data in partially mismatched transplantation (BMT).1 It has been shown that GVHD is human bone marrow transplants indicated that CD134 initiated by donor-derived, alloreactive cytotoxic T lym- + expression, while not predictive of acute GVHD, may be phocytes (CTLs, CD8 ) and helper T lymphocytes 15 + predictive of the response to anti-GVHD therapy. The (CD4 ), which is then further amplified by elaboration of current study was set out to investigate in human recipi- 2,3 cytokines. The ‘ storm’ concept of GVHD is ents of HLA-matched allogeneic T cell-depleted marrow grafts whether the kinetics of repopulation of CD4+ and CD8+ lymphocytes and their CD134 expression would be Correspondence: Dr JJ Cornelissen, Department of Hematology, Daniel den Hoed Cancer Center, Groene Hilledijk 301, 3075 EA Rotterdam, predictive of clinical acute GVHD and/or its response to The Netherlands immunosuppressive therapy. Received 21 October 1998; accepted 3 December 1998 CD134 expression does not predict GVHD APA Gadisseur et al 1014 Patients and methods depletion was performed either by sheep E-rosetting or CD34+ selection using the CellPro device (CellPro, Patients Wezembeek-Oppem, Belgium). Acute GVHD was graded according to standard criteria.5 CD134 expression by T cells was followed in 14 patients with hematological malignancies, who received either an autologous or allogeneic stem cell graft. Three patients Flow cytometry received an autologous peripheral blood stem cell trans- plant (auto-PBSCT). Seven patients received an allogeneic Heparin anticoagulated whole blood samples were stained HLA-identical sibling stem cell transplantation. Stem cells directly using fluorescein isothiocyanate (FITC) and phyco- were harvested from the bone marrow in six patients (allo- erythrin (PE) conjugated monoclonal (mAbs). BMT) and from the peripheral blood in one patient (allo- The following mAb mixtures were used: CD4 (anti- PBSCT). Four patients received a graft from an HLA-A, Leu3ab/FITC) + CD134 (OX40/PE), CD8 (anti- B, DRB1 identical unrelated donor (MUD). The mean age Leu2a/FITC) + CD134 (OX40/PE), and CD4/FITC or of the patients was 43 years (range 22–65), eight were CD8/FITC + mIgG1/PE as isotype controls. All mAbs were female, six male. Indications for the BMT were non-Hodg- obtained from Becton Dickinson (BDIS, San Jose, CA, kin lymphoma (NHL, n = 4), acute myeloid leukemia USA). In short, 200 ␮l aliquots of whole blood were incu- (AML, n = 3), chronic myeloid leukemia (CML, n = 2), bated with mAb mixtures for 20 min at room temperature myelodysplastic syndrome (MDS, n = 2), aplastic anemia in the dark. After lysing the erythrocytes using FACS (AA, n = 2) and chronic lymphocytic leukemia (CLL, n = lysing buffer (BDIS), the remaining leukocytes were 1) (Table 1). Blood samples were taken at day 0, +7, +10, washed once and resuspended in PBS containing 1% para- +14, +21, +28 and +35 after transplantation, and if possible, formaldehyde and 0.04 ␮g/ml LDS-751 (Exciton, Dayton, at longer intervals thereafter. OH, USA). During flow cytometric data acquisition a live gate was set on nucleated cells (LDS-751+), and sufficient nucleated cells were acquired so as to contain at least 1000 Transplant procedures CD4+ or CD8+ lymphocytes. During list mode data analysis Different conditioning regimens were used depending on the CD4+ or CD8+ lymphocytes were selected on the basis diagnosis and donor type (related or unrelated). Eleven of their low sideward scatter signals in combination with patients (11/14) received myelo-ablative conditioning by their CD4 or CD8 positivity. The expression of CD134 by cyclophosphamide and total body irradiation (cyclo/TBI). the CD4+ or CD8+ lymphocytes was analyzed by placing Two non-Hodgkin’s lymphoma patients were treated with the marker to discriminate between positive and negative BCNU, etoposide (VP16), cytosine arabinoside and cyclo- signals at the foot of the negative peak in the isotype con- phosphamide (BEAC regimen). One patient (1/14) was trol histogram. With this marker setting the CD134-stained conditioned with busulfan and melphalan. Patients with an cells were analyzed and from the %CD134+ lymphocytes unrelated donor received additional lymphocyte depletion (ie fraction of CD4+ or CD8+ lymphocytes) the % lympho- by anti-thymocyte globulin (ATG-Merieux, Pasteur Mer- cytes exceeding the marker setting on each isotype control ieux, Lyon, France) before BMT. All allogeneic grafts were was subtracted. The detection limit was 0.5% of CD4+ or administered after in vitro partial T cell depletion. T cell CD8+ lymphocytes expressing the CD134 antigen.

Table 1 Patient characteristics

UPN Sex Disease Stem cell Conditioning GVHD Steroidsa transplantation regimen maximum grade

379 F NHL auto-PBSCT BEAC — — 381 F NHL auto-PBSCT BEAC — — 394 F CML auto-PBSCT Bu/Melp — — 380 F AML allo-BMT Cyclo/TBI — — 382 F CLL/NHL allo-BMT Cyclo/TBI I (skin) + 385 M ALL allo-PBSCT Cyclo/TBI II (skin) + 395 M MDS-RA allo-BMT Cyclo/TBI III (GI) ++ 396 M CLL allo-BMT Cyclo/TBI I (skin) topical 397 M NHL allo-BMT Cyclo/TBI II (skin) + 398 F MDS-RA allo-BMT Cyclo/TBI — — 386 F AML MUD-BMT Cyclo/TBI/ATG I (skin) topical 387 F CML MUD-BMT Cyclo/TBI/ATG II (skin) ++ 388 M AA-PNH MUD-BMT Cyclo/TBI/ATG III (skin, GI) ++ 392 F AML MUD-BMT Cyclo/TBI/ATG I (skin) topical

NHL = non-Hodgkin lymphoma; CML = chronic myeloid leukemia; AML = acute myeloid leukemia; CLL = chronic lymphocytic leukemia; MDS-RA = myelodysplastic syndrome, type refractory anemia; AA-PNH = aplastic anemia-paroxysmal nocturnal hemoglobinuria; BEAC = BCNU, etoposide, arabinoside-C, cyclophosphamide; Bu/Melp = busulfan/melphalan; Cyclo = cyclophosphamide; TBI = total body irradiation; ATG = anti-thymocyte globulin; GI = gastro-intestinal tract. aSteroids were administered topically or prednisolone intravenously (+: 1 mg/kg/day; ++:2× 1 mg/kg/day). CD134 expression does not predict GVHD APA Gadisseur et al 1015 Assessment of absolute CD4+CD134+ and CD8+CD134+ UPN 380 UPN 382 counts 80 a UPN 385 70 UPN 395 Absolute CD4+ or CD8+ lymphocyte counts were obtained UPN 396 + + 60 UPN 397 UPN 398 by multiplying the proportion of CD4 or CD8 lympho- 50 cytes (fraction of nucleated cells) with the absolute leuko- 40 cyte counts as assessed using a Technicon H1 automated 30 cell counter (Technicon, Tarrytown, NY, USA). Blood 20 samples from apparently healthy donors served as a refer- + 10 ence source for CD134 expression by the CD4 and 0 + CD8 lymphocytes. 5 10152025303540 Days after BMT UNP 386 80 b UPN 387 Results 70 UPN 388

+ UPN 392 60 Expression of CD134 by CD4+ and CD8+ lymphocytes 50 40 + CD134 expression on CD4 T lymphocytes of healthy 30 donors was invariably low with a mean percentage of 8.1% 20 + CD134% of CD4 (range: 2.3–14.7; s.d., 3.2). The CD134 expression by CD8 10 T lymphocytes of healthy donors was undetectable (ie 0 Ͻ0.5% of CD8+ lymphocytes). In contrast, the expression 5 10152025303540 of CD134 by CD4+ T lymphocytes of BMT recipients was Days after BMT strongly increased, ranging between 5 and 70%. CD134 expression by the CD8+ lymphocytes of BMT recipients 80 c 70 remained low and in many cases below the detection limit UPN 379 Ͻ + 60 UPN 381 (ie 0.5% of the CD8 lymphocytes). The expression of UPN 394 CD134 became detectable on the CD8+ subset (mean 4.6%, 50 range 0.6–25%) in 55/86 (64%) of the samples. The results 40 of CD134 expression on CD4+ lymphocytes were evaluated 30 further in the three patient groups. 20 10 0 HLA-matched sibling transplants: The expression of 5 10152025303540 + CD134 as a percentage of the CD4 cells showed peak Days after BMT levels up to 58% (mean 40.4%, range 26–58%) in this Figure 1 CD134 expression (%) by CD4+ T cells in recipients of (a) group of patients (Figure 1a). Peak values were reached an allogeneic HLA-matched sibling stem cell transplantation (SCT), in between day +7 and +21. Five out of seven patients recipients of an HLA-A, B, DRB1 matched unrelated SCT (b), and in developed GVHD. Maximum grade and time of onset of recipients of an autologous SCT (c). GVHD are presented in Table 2. Four patients experienced skin GVHD and one patient developed gastro-intestinal GVHD. No correlation between the levels or the pattern in HLA-matched unrelated transplants: Patients who received time of CD134 expression on CD4+ lymphocytes and the an unrelated donor graft showed similar levels of CD134 severity or clinical course of GVHD could be determined. expression as patients who received a sibling graft (mean Two patients showed a steep rise before GVHD could be 49.75%, range 5–70%). Peak values were reached between observed clinically, while two other patients showed a mod- days +7to+14 after transplantation. Two patients showed erate increase before the onset of GVHD. However, both an early rise of CD134+CD4+ cells followed by a decline patients without GVHD also showed steep rises of CD134+ (Figure 1b). All four patients showed some evidence of expression as a percentage of CD4+ lymphocytes. Although GVHD during the study period (Table 2). Three patients the percentage of CD134+CD4+ expression showed a experienced acute GVHD limited to the skin (grade I to II) decline after initial peaking there was no evident correlation and one patient showed extensive GVHD grade III of the between the initiation of systemic steroids for GVHD and skin, gastro-intestinal tract and liver. No correlation was that decline. All patients with GVHD responded to steroid observed between the severity and course of GVHD and therapy. The absolute CD134+CD4+ values remained rela- the levels and pattern in time of CD134 expression as a tively low with levels ranging between 1–10/␮l and peak percentage of CD4+ lymphocytes. The highest levels were values between 20–35/␮l (Figure 2a). Absolute observed in patient UPN 387 showing acute GVHD grade CD134+CD4+ values showed a clear increase from day 0 II while patient UPN 388 with grade III GVHD showed a onwards. One patient with GVHD grade III (UPN 395) low and decreasing CD134+CD4+ level when the picture showed a steep rise just before GVHD was diagnosed but of GVHD worsened. Absolute CD134+CD4+ numbers rose showed a further rise after therapy with systemic steroids around day +14 with peak levels 10–18/␮l (Figure 2b). The had been initiated, upon which these symptoms quickly levels of CD134+CD4+ T cells showed an initial peak but resolved. became much lower thereafter. CD134 expression does not predict GVHD APA Gadisseur et al 1016 Table 2 CD134 expression and acute GVHD

UPN BMT GVHD CD134 as % CD4+ Absolute CD134+CD4+

Onset Maximum Peak Peak Peak Peak (day) grade day level day level (/␮l)

379 auto-PBSCT — — +29 30% +29 113 381 auto-PBSCT — — +9 45% +14 68 394 auto-PBSCT — — +7 63% +14 152 380 allo-sib-BMT — — +14 53% +14 3 382 allo-sib-BMT +26 II +21 36% +10 17 385 allo-sib-BMT +21 II +10 26% +31 20 395 allo-sib-BMT +21 III +21 58% +35 32 396 allo-sib-BMT +21 I +21 28% +28 18 397 allo-sib-BMT +14 II +7 47% +35 15 398 allo-sib-BMT — — +14 35% +21 6 386 MUD-BMT +36 I +48 39% +38 18 387 MUD-BMT +11 II +11 70% +14 11 388 MUD-BMT +12 III +14 55% +14 17 392 MUD-BMT +28 I +36 26% +36 3

Autologous transplants: These control patients were treated with high-dose chemotherapy and autologous peri- pheral blood stem cell transplantation. Expression of UPN 380 + UPN 382 CD134 as a percentage of CD4 cells was high. Peak values 1000 a UPN 385 (mean 46%, range 30–63%) were detected around 7 to 10 UPN 395 days after transplantation, thereafter levels declined (Figure UPN 396 + + 100 UPN 397 1c). The absolute CD134 CD4 levels also peaked rela- UPN 398 tively early (ie day +15) and declined subsequently (Figure ␮ 10 2c). Peak values ranged between 70 and 160/ l. None of the patients experienced severe infections. The highest + 1 value for CD134 expression as a percentage of CD4 lym- 510152025303540 phocytes was detected in patient UPN 394 who recovered Days after BMT 0.1 quickly after transplantation and experienced no compli- cations.

1000 b UPN 386 UPN 387 Discussion 3 UPN 388 100 UPN 392

/mm CD134 (OX40) is a member of the NGFR/TNFR super- + family of receptors expressed on activated T cells, which, 10 CD4

+ upon cross-binding, results in T cell proliferation, cytokine secretion, differentiation of B cells, and adhesion of acti- 1 9–13 CD134 510152025303540 vated T cells to vascular endothelial cells. Experimental Days after BMT studies in animals have suggested an association between 0.1 CD134 expression on peripheral blood T lymphocytes and onset and course of acute GVHD.14 No such correlation UPN 379 was observed in the present study. The appearance of 1000 UPN 381 + + c UPN 394 CD134 CD4 T lymphocytes was not associated with the onset of acute GVHD in patients receiving a bone marrow 100 graft from either an HLA genotypically matched sibling donor or a HLA-A, B, DR matched unrelated donor. Fur- 10 thermore, a decline of CD134+CD4+ T lymphocytes was not associated with GVHD responding to therapy, neither 1 was the persistence of GVHD associated with a continued 510152025303540 presence of CD134+CD4+ T lymphocytes. Most strikingly, Days after BMT + + 0.1 high levels of CD134 CD4 T lymphocytes were detected Figure 2 Absolute numbers of CD134+ CD4+ T cells in recipients of an after autologous transplantation and relative high values allogeneic sibling SCT (a), recipients of a MUD-SCT (b) and in recipients were also observed in patients who received an allogeneic of an autologous SCT (c). graft without any GVHD. Our observations do not support CD134 expression does not predict GVHD APA Gadisseur et al 1017 the view that CD134 expression is predictive of GVHD or 5 Glucksberg H, Storb R, Fefer A et al. Clinical manifestations is a useful tool for monitoring response to GVHD. In of graft-versus-host disease in human recipients of marrow addition, earlier suggestions to develop a therapeutic from HLA-matched sibling donors. Transplantation 1974; 18: approach aimed at the elimination of CD134+CD4+ 295–304. lymphocytes are not supported by our findings. 6 Gratama J, Wuersch A, Nissen C et al. Influence of graft- Prevention of GVHD following allogeneic BMT can be versus-host disease prophylaxis on early T lymphocyte performed by anti-T cell therapy combining cyclosporine repopulation in allogeneic bone marrow transplantation. Exp Hematol 1986; 14: 133–137. and methotrexate and alternatively, by T cell depletion 7 Gratama J, Naipal A, Oljans P et al. T lymphocyte repopu- (TCD) of the graft, which is often combined with cyclospo- lation and differentiation after marrow transplantation. Early 1,16–18 rine. Depletion of virtually all T cells from the graft shifts in the ration between T4+ and T8+ T lymphocytes corre- is the most effective way of GVH prevention. However, it late with the occurrence of acute graft-versus-host disease. is associated with an increased risk of graft failure and an Blood 1984; 63: 1416–1423. increased relapse rate. Therefore, partial instead of com- 8 Miyamoto T, Akashi K, Hayashi S et al. Serum concentration plete TCD has been developed in order to prevent severe of the soluble interleukin-2 receptor for monitoring acute GVHD, while preserving an anti-leukemic effect and avoid- graft-versus-host disease. Bone Marrow Transplant 1996; 17: ing the risk of graft failure.19–21 It has been shown that the 185–190. outcome of allo-BMT following partial TCD is associated 9 Durkop H, Lanza U, Himmelreich P, Stein H. Expression of with a low incidence of GVHD and a relapse rate for the human OX40 antigen in normal and neoplastic tissues. Br patients with acute leukemia, which seems comparable to J Haematol 1995; 91: 927–931. non-TCD BMT.21 However, a major disadvantage of either 10 Baum P, Gayle R, Ramsdell F et al. Molecular character- complete or partial TCD is a delayed immune recovery, ization of murine and human OX40/OX40 ligand systems: identification of a human OX40 ligand as the HTLV-1 regu- which may be associated with opportunistic infections. The + + + lated protein gp34. EMBO J 1994; 13: 3992–4001. low numbers of CD4 T cells and the CD134 CD4 subset 11 Imura A, Hori T, Imada K et al. The human Ox40/gp34 sys- observed in the present study in patients who received an tem directly mediates adhesion of activated T cells to vascular allogeneic graft may reflect a slow immune recovery. In endothelial cells. J Exp Med 1996; 183: 2185–2195. addition, the absence of any correlation between GVH and 12 Stuber E, Strober W. The T cell–B cell interaction via OX40- + + the presence of CD134 CD4 T cells may be due to T cell OX40L is necessary for the T cell-dependent humoral immune depletion of the graft. It does not exclude the possibility response. J Exp Med 1996; 183: 979–989. that the pattern of CD134+CD4+ T cells in patients receiv- 13 Higashimura N, Takasawa N, Tanaka Y et al. Induction of ing an unmanipulated graft may be associated with onset OX40, a receptor of gp34, on T cells by trans-acting transcrip- and course of GVHD. tional activator, Tax, of human T cell leukemia virus type I. Enhanced expression of CD134 of CD4+ T cells is con- Jpn J Cancer Res 1996; 87: 227–231. sidered to reflect activation of T cells as part of an inflam- 14 Tittle T, Weinberg A, Steinkeler C, Maziarz R. Expression of matory reaction. Opportunistic infections, as well as GVH, the T cell activation antigen OX40 identifies alloreactive T may induce inflammation and induction of CD134+CD4+ T cells in acute graft-versus-host disease. Blood 1997; 89: cells following allogeneic BMT. In the present study, high 4652–4658. + + 15 Lamb L, Abhyankar S, O’ Neal W et al. Expression of CD134 numbers of CD134 CD4 T cells were observed following (OX40) on T cells during the first 100 days following T cell- autologous transplantation in patients, who did not experi- depleted bone marrow transplantation from partially mis- ence any infection at that time. It suggests that CD134+ + matched related donors: correlation with incidence and sever- expression by CD4 cells comes early in lymphocyte ity on graft-versus-host disease. Blood 1996; 90: 542a. repopulation and may be a marker for T cell repopulation 16 Poynton CH. T cell depletion in bone marrow transplantation. in general. Further experimental and clinical studies are Bone Marrow Transplant 1988; 3: 265–269. needed to define the role of CD134+/CD4+ T cells in lym- 17 Marmont A, Horowitz MM, Gale RP et al. T cell depletion phocyte repopulation and lymphocyte development follow- of HLA-identical transplants in leukemia. Blood 1991; 78: ing stem cell transplantation. 2120–2124. 18 Van Bekkum D. Graft-versus-host disease. Biological mech- anisms and clinical practice. In: van Bekkum DW, Lo¨wenberg References B (eds). Bone Marrow Transplant. Marcel Dekker: New York, 1985, pp 147–153. 1 Lazarus H, Vogelsang G, Rowe J. Prevention and treatment 19 Lo¨wenberg B, Wagemaker G, Van Bekkum DW et al. Graft- of acute graft-versus-host disease: the old and the new. A versus-host disease following transplantation of ‘one log’ vs report from the Eastern Cooperative Oncology Group ‘two log’ T lymphocyte-depleted bone marrow from HLA- (ECOG). Bone Marrow Transplant 1997; 19: 577–600. identical donors. Bone Marrow Transplant 1986; 1: 133–137. 2 Theobald M. Allorecognition and graft-versus-host disease. 20 Verdonck LF, De Gast GC, Van Heugten HG, Dekker A. A Bone Marrow Transplant 1995; 15: 489–498. fixed low number of T cells in HLA-identical allogeneic bone 3 Antin J, Ferrara J. Cytokine dysregulation and acute graft- marrow transplantation. Blood 1990; 75: 776–780. versus-host disease. Blood 1992; 80: 2964–2968. 21 Verdonck LF, Dekker AW, De Gast GC et al. Allogeneic bone 4 Ferrara J. Cytokine dysregulation as a mechanism of graft- marrow transplantation with fixed low number of T cells in versus-host disease. Curr Opin Immun 1993; 3: 794–799. the marrow graft. Blood 1994; 83: 3090–3093.