- Email: [email protected]
Alloreactive T cell clonotype recruitment in a mixed lymphocyte reaction: Implications for graft engineering
Experimental Hematology 34 (2006) 788–795
Alloreactive T cell clonotype recruitment in a mixed lymphocyte reaction: Implications for graft engineering Phillip Scheinberga, David A. Priceb, David R. Ambrozakc, A. John Barretta, and Daniel C. Douekb a
Hematology Branch, National Heart, Lung, and Blood Institute, Bethesda, Md., USA; bHuman Immunology Section; cImmunology Laboratory, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Md., USA (Received 20 December 2005; revised 22 February 2006; accepted 2 March 2006)
Objective. The selective elimination of alloreactive T cells from donor stem cell grafts prior to hematopoietic stem cell transplantation (HSCT) is an important goal in the prevention of graft-vs-host disease (GVHD). However, in HLA-identical donor-recipient pairs, it has proven difficult to identify alloreactive T cells using in vitro systems pretransplant due, in part, to their low frequency and a lack of methodological standardization. To better understand the alloresponse between HLA-identical related pairs, we characterized the alloreactive T cells generated in a mixed lymphocyte reaction (MLR) assay system. Methods. HSCT donor peripheral blood mononuclear cells (responder) were labeled with carboxyfluorescein diacetate, succinimidyl ester (CFSE) dye and cocultured with irradiated HSCT recipient cells (stimulator) in a one-way MLR. Alloreactive T cells were sorted by upregulation of activation markers (CD25 in most cases) and the responding clonotypes were defined by sequencing the complementarity region 3 (CDR3) of the T cell receptor b-chain. Results. We show that the recruitment of alloreactive CD4+ T cells is highly variable. Oligoclonal CD4+ T-cell expansions in repeated MLRs performed in the same donor-recipient pair showed inconsistent recruitment of clonotypes. The recruitment of alloreactive CD8+ T cells was more consistent in repeated assays, with the same clonotypes identified in the same donorrecipient pair performed under different conditions. Conclusion. Taken together, our data show that even in culture conditions constrained to eliminate background proliferation, stochastic events and low precursor frequencies preclude reproducible elicitation of immunodominant T cell clonotypes with the potential to cause GVHD. Ó 2006 International Society for Experimental Hematology. Published by Elsevier Inc.
The development of ex vivo coculture assays between donor and recipient pairs undergoing hematopoietic stem cell transplantation (HSCT) has been used to demonstrate alloimmune T cell responses in vitro that correspond to events in vivo postgraft. Since its first introduction by Bach et al., the mixed lymphocyte reaction (MLR) has been used to identify alloreactivity between donor-recipient pairs in organ and stem cell transplant recipients [1]. However, the lack of standardization prevents its routine widespread use. Many variable parameters must be validated formally in order to ensure the reproducibility of this approach; these include the nature of the stimulator cells that will best recruit alloreactive T cells, the techniques
Offprint requests to: Phillip Scheinberg, M.D., Hematology Branch, NHLBI, 10 Center Drive, Building 10 CRC, Rm 3-5140, MSC 1202, Bethesda, MD 20892-1202; E-mail: [email protected]
employed to identify responding cells, and the optimal culture conditions to enable maximal resolution of alloreactive T cells. In mismatched MLRs, immunodominant alloresponses to major histocompatibility antigens (MaHA) can be identified, as has been illustrated in a pretransplant HLA-DRmismatched MLR, where a single CD4D T cell clonotype identified in vitro accounted for almost 100% of the circulating CD4D T cells 100 days postgraft, and correlated with emergence of graft-vs-host disease (GVHD) [2]. In the HLA-matched setting, minor histocompatibility antigens (MiHA) are perceived as the main stimulus for donor alloreactive T-cell activation and expansion in vivo, resulting in GVHD and graft-vs-leukemia (GVL) effects [3,4]. Although the antigenic disparity between HLA-identical siblings involves hundreds of different MiHA, only a few immunodominant antigens are likely to trigger an antihost T-cell response [5]. Furthermore, MLRs between
0301-472X/06 $–see front matter. Copyright Ó 2006 International Society for Experimental Hematology. Published by Elsevier Inc. doi: 10.1016/j.exphem.2006.03.001
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
HLA-identical HSCT related pairs usually elicit either in a minimal or no response, despite the fact that such donor lymphocytes can induce lethal graft-vs-host reactions in the transplanted recipient. The identification of alloreactive T cells in the matched MLR setting is hampered by their very low precursor frequency, estimated to be less than 0.1% [6,7]. Therefore, an assay with a high signal-to-noise ratio, with minimal or no background, is critical for the appropriate characterization of alloreactive clonotypes. Another limiting factor in the recruitment of alloreactive T cells in the matched MLR is inadequate antigen presentation in vitro. For this reason, modified MLRs have been used in which cytokines are added to the culture or stimulators are pretreated to augment their antigen presenting capacity [8]. The problem with these approaches is that background also increases, which may limit the specific identification of alloreactive T cells. In the present study, we developed modifications that increase the sensitivity of the standard MLR to enhance reliable identification of alloreactive T cell clonotypes. This was achieved primarily by manipulations that reduced nonspecific background stimulation and the augmentation of antigen presentation based on the use of different stimulator cell populations in the same donor-recipient pair. Alloreactive T cells were isolated by flow cytometry based on the upregulation of a T-cell activation marker (CD25 in most experiments) and characterized by sequencing the complementarity-determining region 3 (CDR3) of the T cell receptor (TCR) b-chain.
Material and methods Patients All subjects tested gave written permission for their blood to be used for research under Institutional Review Board–approved National Heart, Lung, and Blood Institute stem cell allotransplantation protocols. Isolation of human peripheral blood mononuclear cells (PBMCs) A leukapheresis was collected from donor and recipient prior to transplantation. Peripheral blood mononuclear cells (PBMCs) were prepared by density-gradient centrifugation over Ficoll-Paque (MP Biomedicals, Aurora, OH, USA); washed in culture medium (RPMI-1640; Rockville, MD, USA), 10% heat-inactivated human AB serum (Gemini Bioproducts, Woodland, CA, USA), and 65 ug/mL gentamicin (Biosource, Rockville, MD, USA); and then cryopreserved in liquid nitrogen. Generation of monocyte-derived dendritic cells (DCs) PBMCs were resuspended in culture medium at a final concentration of 3 to 5 3 106 cells/mL and incubated in a standard tissue culture flask for 2 hours at 37 C in a 5% CO2-containing atmosphere. Nonadherent cells were removed by vigorous pipetting and placed in a 50-mL Falcon (supernatant). The remaining adherent cells were cultured in medium (containing 1% heat-inactivated human AB serum) supplemented with 200 ng/mL recombinant human
789
GM-CSF (R&D Systems, Minneapolis, MN, USA) and 4 ng/mL recombinant human interleukin IL-4 (R&D Systems). Fresh cytokines were added every 2 to 3 days. For the maturation of DCs, culture medium was replaced on day 6, and 1100 U/mL recombinant human TNF-a (R&D Systems) was added for 24 hours [9]. Generation of activated B cells The cells in the supernatant were washed in culture medium, and B cells were positively separated by using the MIDI-magnetic cell sorting (MACS) protocol provided by the manufacturer (Miltenyi Biotec, Auburn, CA, USA). Briefly, after centrifugation, the cell pellet was resuspended in 80 uL buffer (phosphate-buffered saline [PBS], 0.5% heat inactivated human AB serum); 20 uL anti-CD19 microbeads/107 cells were then added and the cells were incubated with the beads for 15 minutes at 4 C. Cells were then washed twice, resuspended in 500 uL buffer, and magnetically separated with an LSD column. After rinsing the column, the fraction with the magnetically labeled cells (CD19D) in the plunger was flushed and counted with a hemocytometer. Positively selected B cells were cultured for 6 days in culture medium (1-2 3 106/ mL) in a tissue culture flask and synthetic CpG oligodeoxynucleotide (CpG; TCG TCG TTT TGT CGT TTT GTC GTT; Coley Pharmaceutical Group, Ottawa, Canada) was added to a final concentration of 10 ug/mL [10]. Generation of activated T cells A 25-cm2 tissue culture flask was prepared for polyclonal T-cell stimulation in the solid phase by addition of 1 ug anti-CD3 monoclonal antibody (Muromonab-CD3 Ortho Biotech Products, Raritan, NJ, USA) diluted in 2 mL PBS, followed by overnight incubation at 37 C in a 5% CO2-containing atmosphere. The antibody suspension was removed the following day from the coated flask which was rinsed vigorously with PBS. The unlabeled cell fraction from above (effluent) was resuspended in medium (1-2 3 106 cells/mL) supplemented with 100 U/mL IL-2 (Peprotech, Rocky Hill, NJ, USA) and cultured in the anti-CD3 monoclonal antibody treated tissue culture flask for 6 days at 37 C in a 5% CO2-containing atmosphere [11]. The culture medium was changed as needed. Carboxyfluorescein diacetate, succinimidyl ester (CFSE) labeling Responder PBMCs (1-2 3 106/mL) were labeled with 0.25 uM CFSE (Molecular Probes, Carlsbad, CA, USA) in serum-free PBS for 7 minutes at 37 C. Filtered human AB serum was then added to stop the reaction, and the cells were washed with RPMI-1640. The supernatant was decanted after centrifugation and the cell pellet was resuspended in 1 mL filtered human AB serum before washing again with RPMI-1640. Mixed lymphocyte reaction assay culture Initial allogeneic MLRs were performed by culturing 20 to 50 3 106 responder (HSCT graft donor) PBMCs with irradiated (2500 rads) stimulator (HSCT graft recipient) PBMCs at a 1:1 ratio for 6 days in culture medium. Later, allogeneic MLRs were performed by culturing 10 to 20 3 106 responder cells (HSCT graft donor) with irradiated (2500 rads) DCs, B or T cells (stimulator cells derived from the HSCT recipient as described above), for 3 days. The responder-to-stimulator ratio was 8:1 for DCs (DCMLR), 4:1 for B cells (B-MLR), and 1:1 for T cells (T-MLR). An autologous MLR under identical conditions was performed
790
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
in parallel with each allogeneic MLR. All MLRs were performed on samples from sibling pairs matched at 6/6 HLA loci (HLA-A, B, and DR), unless otherwise indicated. Depletion of cell fractions according to the expression of surface activation markers Responder cells were depleted of the CD25D fraction one day after CFSE labeling, and before the MLR. PBMCs were labeled with anti-CD25 magnetic microbeads (Miltenyi Biotec, Auburn, CA, USA) according to protocol provided by the manufacturer. The magnetically labeled CD25D cells were retained on the MSD column; cells in the effluent (CD252) were counted and used in the MLR as responders. Removal of CD25D, CD69D, and CD71D cells was performed in one HLA-matched MLR. Responder PBMCs were stained with anti-CD25, anti-CD69, and anti-CD71 phycoerythrin (PE)-conjugated mAbs (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA), washed, and bound to anti-PE microbeads (20 mL/107 cells; Miltenyi Biotec, Auburn, CA, USA) according to protocol provided by the manufacturer. The cells were then washed again and subjected to magnetic separation. Cells expressing one or more of the selection antigens (CD25DCD69DCD71D) were magnetically labeled and retained on the MSD column; cells in the effluent were counted and used in the MLR as responders. Flow cytometry The expression of surface antigens on treated DCs, B and T cells were analyzed by flow cytometry. Conjugated mAbs used for phenotypic characterization of the various cell populations were as follows: 1) treated DCs: CD1a-PE and CD83-fluorescein isothiocyanate (FITC); 2) B cells: CD19-FITC, CD80-PE, and CD86-allophycocyanin (APC; Becton Dickinson Pharmingen, San Diego, CA, USA); and 3) T cells: CD3-FITC and CD25-PE. The purity of activated T and B cells was greater than 95%. All T cells used as stimulators were activated as evidenced by CD25 upregulation. B cells following culture with CpG had significant upregulation of CD80 and CD86 molecules compared to baseline. The purity of mature DCs was greater than 80% (data not shown). The expression of T-cell activation markers on the CFSElabeled responder cells was analyzed by six-color flow cytometry. The following directly conjugated mAbs were used: CD25-PE; CD11c-, CD14-, CD16-, CD19 PE-Cy5 (dump channel); CD4 PE-Cy7; and CD8 APC-Cy7. In one matched MLR, a combination of T-cell activation markers was used in the APC channel: CD25, CD69, and CD71 APC. 7-AAD was used as a viability marker for all sorts. Flow cytometric cell sorting All sorts were performed with a modified FACS DIVA (BD Pharmingen). Instrument set-up was performed according to the manufacturer’s instructions. All sorts were performed at 25 lb/in2. Instrument compensation was performed with antibody capture beads (BD Pharmingen) stained singly with individual mAbs used in the test samples. Sort gates were set such that only donor-derived (CFSE-labeled) activated cells were included. In the initial 6-day MLRs, in which PBMCs were used as stimulators, CD4DCFSEDhi and CD4DCFSEDlo were sorted separately. Later, in the three-day MLRs, CD4DCFSEDhi and CD8DCFSEDhi cells that had upregulated activation marker(s) were included in the sorting gate. All cells were sorted into RNAlater (Ambion, Austin, TX, USA) and frozen at 280 C.
Clonotypic assay Specific T cell clonotypes in the sorted populations were identified using a strand-switch–anchored reverse transcriptase polymerase chain reaction (RT-PCR) that amplifies all expressed TCRB gene products without bias [12,13]. Briefly, mRNA was extracted from the sorted antigen-activated T cells (Oligotex kit, Qiagen, Valencia, CA, USA). Amplification of TCRB VDJ junctions was performed by using a modified version of the SMART procedure (switching mechanism at the 50 end of RNA transcript) to generate cDNA and a template-switch–anchored PCR with a 30 C-region primer. The PCR product was ligated into the pGEMT Easy vector (Promega, Madison, WI, USA) and used to transform Escherichia coli. Individual colonies were selected, amplified by PCR with M13 primers, and then sequenced to obtain TCRB CDR3 sequences corresponding to all sorted T cells that had been activated in the MLR.
Results HLA-matched MLR with PBMC stimulators and unmanipulated responders PBMCs were initially used as the stimulating cell population. The ratio of donor (responder) to patient (stimulator) cells was 1:1 in these assays. Cultures were incubated for 6 days prior to washing and antibody labeling. Plots were gated on CD4D and CD8D T cells, and CD25DCFSEDlo and CD25DCFSEDhi populations were sorted separately for RNA extraction (Fig. 1). As the responding CD8D T cell populations were small (!20 cells), further study
Figure 1. The top panels are gated on CD4D T cells; the bottom panels are gated on CD8D T cells. Control cultures are shown on the left in each case; experimental MLRs are shown on the right. A 0.27% CD4DCD25DCFSEDlo and 0.8% CD4DCD25DCFSEDhi response above background is shown; cells were sorted directly into RNAlater for clonotype analysis. The clonotypes detected in each of these sorted populations are shown in Table 1.
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
791
Table 1. Replicate MLRs from same donor-recipient pair with unmanipulated responding cells No. of distinct clonotypes
CDR3 region of clonotype(s) with frequency O10%
Freq. (%)
Repertoire
Sorted cells #1 CD4DCD25DCFSEDhi
6
18
32 18 11 14 11 14
oligoclonal
CD4DCD25DCFSEDlo
CASSPWSRQQRELRNEQYF CASSPRGDTQYFG CSARTPGLAGGETQYF CSAPKGGLPLPYNEQFF CASSFSGRNNEQFF CASRITANNEQFF
Sorted cells #2 CD4DCD25DCFSEDhi CD4DCD25DCFSEDlo
2 36
CSAREGATTYEQYF CASRITANNEQFFz
98 3z
oligoclonal polyclonal
MLR from same donor-recipient pair
polyclonal
MLR, mixed lymphocyte reaction; CDR3, complementarity-determining region 3. z This clonotype was identified previously but was present only once in the repeat MLR.
of these cells was precluded. The responding CD4D T cells segregated into two populations based on the degree of CFSE fluorescence: bright (CFSEDhi) and dim (CFSEDlo). Clonotyping of the sorted CD4DCD25DCFSEDhi and CD4DCD25DCFSEDlo T cell populations generally revealed polyclonal TCRB sequences with no shared clonotypes between populations. There was also no identity between clonotypes in replicate MLR cultures performed several weeks later using the same PBMC combinations (Table 1). On only one occasion in a total of 16 pairs studied and over 1000 sequences analyzed was the same CDR3 sequence (CASRITANNEQFF) identified in the CD4DCD25D CFSEDlo population in duplicate MLRs (Table 1). HLA-matched MLR with PBMC stimulators and CD25-depleted responders To minimize background in the clonotype analysis, donor (responder) cells were depleted of CD25-expressing populations after CFSE labeling but prior to the MLR. The MLRs were then cultured for 3 days before washing and antibody labeling. Plots were gated on CD4D and CD8D T cells, and CD25DCFSEDhi cells were sorted for RNA extraction (Fig. 2). Again, the low magnitude of the responding CD8D T cell populations (!20 cells) precluded clonotypic analysis. The CD4D T cell clonotypes identified in duplicate MLRs were different; one MLR yielded only one clonotype, while the other yielded 45 clonotypes with no clear hierarchy. The same experiment was repeated with the same donor-recipient pair; once more, one MLR yielded only one clonotype that differed from the single clonotype identified in the previous MLR with this PBMC combination, while the other yielded 18 different sequences with 3 dominant clonotypes (Table 2). HLA-DR-mismatched MLR with B cell, T cell, and DC stimulators and CD25-depleted responders To minimize the heterogeneity of the stimulator cell populations, monocyte-derived DCs, activated B cells, and activated T cells were generated as described in the Materials
and Methods. The following donor (responder) to patient (stimulator) ratios, based on pilot optimization experiments, were used: DCs, 8:1; activated B cells, 4:1; and activated T cells, 1:1. The MLRs were incubated for 3 days before washing and antibody labeling. Plots were gated on CD4D and CD8D T cells, and CD25DCFSEDhi populations were sorted separately for RNA extraction (Fig. 3). The clonotypes derived from the CD4D T cell sort in each of the T-, B-, and DC-MLRs were polyclonal and distinct from each other (Table 3). There were 10 clonotypes in the T-MLR that had a frequency of greater than 5% each; in the B- and DC-MLR, there were no clonotypes that had a frequency greater than 5%. The CD8D sorted T cells were more oligoclonal, with approximately five different clonotypes in each MLR. One CD8D T cell clonotype (TCRB CDR3 sequence CSARDLAAGGLGETQYF) was present in all three MLRs, while two further clonotypes
Figure 2. Three-day MLRs in an HLA-matched donor-recipient pair performed in parallel. Responder (recipient) cells were CD25 depleted after CFSE labeling prior to the MLR. A small response above background was observed in the CD4D T cell population (upper panels); the constituent clonotypes are shown in Table 2. No CD8D T cell response was observed. Autologous control cultures in each case are shown in the left panels.
792
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
Table 2. Replicate MLRs from the same donor-recipient pair with CD25-depleted responding cells MLR from same donor-recipient pair performed in parallel (MLR 1 and 2)
CDR3 region of clonotype(s) with frequency O10%
No. of distinct clonotypes
Freq. (%)
Repertoire
Sorted cells–MLR1 & 2 1: CD4DCD25DCFSEDhi 2: CD4DCD25DCFSEDhi
1 45
CSAGTSGGTFTKQYF None
100
monoclonal polyclonal
Sorted cells–MLR 1 & 2 (repeat) 1: CD4DCD25DCFSEDhi 2: CD4DCD25DCFSEDhi
1 18
CASSQGDRSVGTQYF CASSWKLTGDRYNEQFF CASRLGRLNTEAFF CASSFLAGGPGEQFF
100 17 15 15
monoclonal polyclonal
MLR, mixed lymphocyte reaction; CDR3, complementarity-determining region 3.
(CASSLAPVNYGYTF and CASSPDGYGYTF) were present in the B- and DC-MLRs, but not the T-MLR (Table 3). HLA-matched MLR with B cell, T cell, and DC stimulators and CD25-depleted responders The same experiment was performed with samples from HLA-matched siblings; in this case, responder PBMCs were depleted of CD25D, CD69D, and CD71D cells prior to the MLR, and a combination of activation markers was used in the APC channel as described in the Materials and Methods. CD69 and CD71 were included to enhance the identification of alloreactive T cells in the matched MLR [14–16]. With the exception of the B-MLR, the clonotypes from the responding CD4D T cell populations were polyclonal and distinct from each other (Table 4). In the sorted CD8D T cell populations, the response was more oligoclonal (Table 4). One clonotype (TCRB CDR3
Figure 3. Mixed lymphocyte reaction with HLA-DR-mismatched sibling PBMCs and prior CD25 depletion of responder cells. Three different stimulators were used: T cells (top row), B cells (middle row), and dendritic cells (DCs; bottom row). The response is greater with DCs and lesser with T cells; in all cases, both CD4D and CD8D responding T cell populations were sorted for clonotypic analysis (Table 3).
CASSPAQHPPRERDTQYF) was present in the T- and B-MLR, but not in the DC-MLR.
Discussion The antigens that drive the T cell response in the HLAmatched allogeneic setting are for the most part unknown, but include broadly expressed minor histocompatibility antigens (e.g., H-Y, HA-8), tissue-restricted minor histocompatibility antigens (e.g., HB-1), and antigens that are either overexpressed or aberrantly expressed by malignant cells (e.g., PR-1, WT-1) [17–19]. Furthermore, the recruitment and characterization of alloreactive T cells between HLA-identical siblings has long been a challenge because such T cells are present at very low precursor frequencies. Here, we report methodological inconsistencies that complicate the reliable identification of alloreactive T cell clonotypes in current HLA-matched donor-recipient MLR protocols and describe manipulations that illuminate the underlying reasons for these limitations. Initially, we used a one-way MLR with irradiated PBMCs as stimulators; cells that had upregulated CD25 after 6 days were isolated by flow cytometric sorting as described previously [20,21]. With this approach, alloreactive CD4D T cell clonotypes have been identified pretransplant and tracked in vivo post-HSCT by qPCR; expansions have been associated with acute GVHD [2]. In our experiments, dividing and nondividing cells were analyzed separately. Consistent with previous reports, CD8D T cell responses were not observed under these conditions [22]. Our analysis was therefore generally restricted to CD4D T cells. By sorting CD4DCD25DCFSEDlo T cells, we expected to identify oligoclonal populations that were responding in an alloreactive manner to mHAg. In most cases, polyclonal T cell populations were elicited. When oligoclonal populations were observed, the same clonotype was only rarely detected in replicate or parallel MLRs (Fig. 1; Table 1). Possible explanations for this inconsistency include: 1) nonspecific stimulation of T cells in the MLRs resulting in a ‘‘background’’ that would become relatively more
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
793
Table 3. Clonotypes from sibling HLA-DR-mismatched MLR using T cells, B cells, and DC stimulators Stimulators and sorted cell populations
No. of distinct clonotypes
CD4D sorted cells ‘‘T’’: CD4DCD25DCFSEDhi
39
‘‘B’’: CD4DCD25DCFSEDhi ‘‘DC’’: CD4DCD25DCFSEDhi
24 32
CD8D sorted cells ‘‘T’’: CD8DCD25DCFSEDhi
7
‘‘B’’: CD8DCD25DCFSEDhi
5
‘‘DC’’: CD8DCD25DCFSEDhi
3
CDR3 region of clonotype(s) with frequency O5%
CASSLRIGNTEAFF CASSQVLTGGGSPLHF CASSSNYQWETQYF CASSYSPLTGELFF CASTKVGTGLYEQYF CASRLADTQYF CASRTPGTSPLHF CSPTPGGRETQYG CASNQGLAGPQETQYF CASSLRDSDYGYTF None None CASSYYGGRWTGELFF CSARDLAAGGLGETQYFz CASSQATGRNYGYTF CSARDLAAGGLGETQYFz CASSLAPVNYGYTF* CASSPDGYGYTF* CSARDLAAGGLGETQYFz CASSLAPVNYGYTF* CASSPDGYGYTF*
Freq. (%)
Repertoire
8 8 8 5 5 5 5 5 5 5
polyclonal
polyclonal polyclonal 42 17 13 58 17 17 66 25 8
oligoclonal
oligoclonal
oligoclonal
MLR, mixed lymphocyte reaction; CDR3, complementarity-determining region 3; ‘‘T,’’ T-MLR; ‘‘B,’’ B-MLR; ‘‘DC,’’ DC-MLR. z clonotype recruited in T-, B-, and DC-MLR. *clonotype recruited in the B- and DC-MLR.
substantial in the presence of a small alloresponse; and 2) variable allostimulation of PBMCs. In order to determine whether the failure to consistently elicit alloreactive clonotypes consistently was due to the intrusion of background
noise from nonspecifically activated cells, we depleted CD25D cells from the responder population prior to the MLR. This maneuver did not improve the selectivity of the clones induced in MLR (Fig. 2; Table 2).
Table 4. HLA-matched MLR with T cells, B cells, and DC stimulators Stimulators and sorted cell populations
No. of distinct clonotypes
CD4D sorted cells ‘‘T’’: CD4DCD25DCFSEDhi
13
‘‘B’’: CD4DCD25DCFSEDhi ‘‘DC’’: CD4DCD25DCFSEDhi
3 19
CD8D sorted cells ‘‘T’’: CD8DCD25DCFSEDhi
6
‘‘B’’: CD8DCD25DCFSEDhi
4
‘‘DC’’: CD8DCD25DCFSEDhi
4
z
Same clonotype identified in T- and B-MLR.
CDR3 region of clonotype(s) with frequency $10%
Freq. (%)
Repertoire
CSAGGQGAYEQYF CAISAGQGEGYGYTF CAWSISGLETQYF CASSLAGAWDTEAF CASSRVGLPSDEQY CASSTNPRGNQPQH CASSSPRDRTNEQY
18 18 18 88 14 11 11
polyclonal
CASSPNYSNQPQHF CASRRMTGGYDYEQYF CASSLSQGLEQPQHF CASFPGPWLAKNIQYF CASSPAQHPPRERDTQYFz CSARELAGASSYEQY CASSVNRQGGIYEQY CASSPAQHPPRERDTQYFz CASSRRSSYNEQF CASSFGLEGNSPLH
42 23 13 10 10 46 32 19 59 23
oligoclonal polyclonal
oligoclonal
oligoclonal
oligoclonal
794
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
In an attempt to optimize allorecruitment and maintain a more homogeneous stimulator cell population, monocyte-derived DCs, activated T cells, and activated B cells were generated and cocultured with responder cells. The inconsistent recruitment of CD4D T cell clonotypes persisted. However, the CD8D T cell repertoire was more oligoclonal, and some clonotypes were present in MLRs performed with different stimulators. This was observed in both an HLADR-mismatched (Fig. 3; Table 3) and in an HLA-matched MLR (Table 4). As CD25 upregulation was infrequent in samples from HLA-identical siblings, CD69 and CD71 were added as activation markers; this allowed for the identification and sorting of both CD4D and CD8D T cell subsets. How should the finding of wide diversity and nonreproducibility in T cell clonotypes that are elicited in HLAmatched MLRs be interpreted? First, the diversity of CD4D T cell clonotypes elicited within a few days of culture is consistent with the known degeneracy of CD4 T-cell recognition [23]. Thus, the oligoclonal T cells elicited could have recognized a single peptide-MHC complex on the stimulator cells. In contrast, the degree of cross-reactivity in CD8D T cell populations might be lower [23]; this could explain the greater degree of consistency in the recruitment of alloreactive CD8D T cells observed in our system, although further experiments are required to confirm this. Nevertheless, for the CD4D T cell populations in particular, the nonreproducibility between duplicate and replicate cultures implies stochastic selection of individual clonotypes in these cultures. It is somewhat harder to explain the failure to elicit consistent oligoclonal responses in many cultures. This could be interpreted as either a broad base of weak responses to multiple antigens or background noise in the absence of a strong allostimulation. When compared to a more homogeneous stimulator population, such as EBV lymphoblastoid cell lines, allostimulation by PBMCs may be less consistent [24]. Second, antigens may not be presented properly to alloreactive T cells in culture by resting PBMCs. While an alloreactive T cell is likely to respond to antigens on the surface of monocytes, the same may not occur in the case of resting B or T lymphocytes. The more consistent and predictable results obtained with the HLA-mismatched pairs, within which antigenic disparity is presumably stronger, suggests that at least in this in vitro system the alloresponse in HLAmatched pairs represents a broad base of very weak alloresponses to multiple antigens. It is possible that only in prolonged culture or in vivo can a limited number of alloreactive clonotypes emerge to exert functional T cell responses causing, for example, GVHD, GVL, or rejection reactions. Methods for removing GVHD-causing T cells from the graft are being developed that deplete selected T cell subsets (‘‘selective allodepletion’’) recruited in an ex vivo MLR [15,25]. Selective allodepletion of alloreactive donor
T cells in vitro has been shown to preserve antiviral and antileukemic responses [24,26]. In addition, children undergoing HLA-haploidentical HSCT for congenital hematopoietic disorders given selectively depleted T cells did not develop severe GVHD [27]. We have reported similar promising results using selectively depleted T cell transplants in HLA-matched recipients [28]. The failure of the MLR approach used here to identify the alloresponse at the level of individual T cell clonotypes is disappointing. The implications of our observations, however, are important for graft engineering and the selective depletion of alloreactive T cells ex vivo prior to HLA-matched HSCT. While it remains possible that the removal of donor T cells activated by the recipient will be sufficient to reduce or prevent GVHD, the current technology is incapable of defining the basis of the alloreaction. Although the HTLP frequency assay has been used to grade the alloresponse and monitor efficiency of selective depletion approaches, standardization is lacking and reproducibility has been poor; therefore, there are currently no robust predictive assays for GVHD in the HLA-matched setting [28]. Acknowledgments This research was supported by the Intramural Research Program of the NIH; National Heart, Lung, and Blood Institute; and National Institute of Allergy and Infectious Diseases. David A. Price is a Medical Research Council (UK) Clinician Scientist.
References 1. Bach FH, Voynow NK. One-way stimulation in mixed leukocyte cultures. Science. 1966;153:545. 2. Michalek J, Collins RH, Hill BJ, Brenchley JM, Douek DC. Identification and monitoring of graft-versus-host specific T-cell clone in stem cell transplantation. Lancet. 2003;361:1183. 3. Chao NJ. Minors come of age: Minor histocompatibility antigens and graft-versus-host disease. Biol Blood Marrow Transplant. 2004;10:215. 4. Falkenburg JH, Marijt WA, Heemskerk MH, Willemze R. Minor histocompatibility antigens as targets of graft-versus-leukemia reactions. Curr Opin Hematol. 2002;9:497. 5. Perreault C, Roy DC, Fortin C. Immunodominant minor histocompatibility antigens: the major ones. Immunol Today. 1998;19:69. 6. Sherman LA, Chattopadhyay S. The molecular basis of allorecognition. Annu Rev Immunol. 1993;11:385. 7. Wang XN, Taylor PR, Skinner R, et al. T-cell frequency analysis does not predict the incidence of graft-versus-host disease in HLA-matched sibling bone marrow transplantation. Transplantation. 2000;70:488. 8. Bishara A, Brautbar C, Nagler A, et al. Prediction by a modified mixed leukocyte reaction assay of graft-versus-host disease and graft rejection after allogeneic bone marrow transplantation. Transplantation. 1994;57:1474. 9. Nguyen XD, Eichler H, Dugrillon A, Piechaczek C, Braun M, Kluter H. Flow cytometric analysis of T cell proliferation in a mixed lymphocyte reaction with dendritic cells. J Immunol Methods. 2003;275:57. 10. Krieg AM. Antitumor applications of stimulating toll-like receptor 9 with CpG oligodeoxynucleotides. Curr Oncol Rep. 2004;6:88. 11. Solomon SR, Tran T, Carter CS, et al. Optimized clinical-scale culture conditions for ex vivo selective depletion of host-reactive donor lymphocytes: a strategy for GvHD prophylaxis in allogeneic PBSC transplantation. Cytotherapy. 2002;4:395.
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795 12. Douek DC, Betts MR, Brenchley JM, et al. A novel approach to the analysis of specificity, clonality, and frequency of HIV-specific T cell responses reveals a potential mechanism for control of viral escape. J Immunol. 2002;168:3099. 13. Price DA, West SM, Betts MR, et al. T cell receptor recognition motifs govern immune escape patterns in acute SIV infection. Immunity. 2004;21:793. 14. Fehse B, Frerk O, Goldmann M, Bulduk M, Zander AR. Efficient depletion of alloreactive donor T lymphocytes based on expression of two activation-induced antigens (CD25 and CD69). Br J Haematol. 2000;109:644. 15. Koh MB, Prentice HG, Lowdell MW. Selective removal of alloreactive cells from haematopoietic stem cell grafts: graft engineering for GVHD prophylaxis. Bone Marrow Transplant. 1999;23:1071. 16. van Dijk AM, Kessler FL, Stadhouders-Keet SA, Verdonck LF, de Gast GC, Otten HG. Selective depletion of major and minor histocompatibility antigen reactive T cells: towards prevention of acute graftversus-host disease. Br J Haematol. 1999;107:169. 17. Goulmy E, Schipper R, Pool J, et al. Mismatches of minor histocompatibility antigens between HLA-identical donors and recipients and the development of graft-versus-host disease after bone marrow transplantation. N Engl J Med. 1996;334:281. 18. Mutis T, Gillespie G, Schrama E, Falkenburg JH, Moss P, Goulmy E. Tetrameric HLA class I–minor histocompatibility antigen peptide complexes demonstrate minor histocompatibility antigen–specific cytotoxic T lymphocytes in patients with graft-versus-host disease. Nat Med. 1999;5:839. 19. Dolstra H, Fredrix H, Maas F, et al. A human minor histocompatibility antigen specific for B cell acute lymphoblastic leukemia. J Exp Med. 1999;189:301.
795
20. Cavazzana-Calvo M, Fromont C, Le Deist F, et al. Specific elimination of alloreactive T cells by an anti-interleukin-2 receptor B chain–specific immunotoxin. Transplantation. 1990;50:1. 21. Montagna D, Yvon E, Calcaterra V, et al. Depletion of alloreactive T cells by a specific anti-interleukin-2 receptor p55 chain immunotoxin does not impair in vitro antileukemia and antiviral activity. Blood. 1999;93:3550. 22. Martins SL, St John LS, Champlin RE, et al. Functional assessment and specific depletion of alloreactive human T cells using flow cytometry. 2004;104:3429. 23. Wilson DB, Wilson DH, Schroder K, et al. Specificity and degeneracy of T cells. Mol Immunol. 2004;40:1047. 24. Amrolia PJ, Muccioli-Casadei G, Yvon E, et al. Selective depletion of donor alloreactive T cells without loss of antiviral or antileukemic responses. Blood. 2003;102:2292. 25. Szabolcs P, Park KD, Marti L, et al. Superior depletion of alloreactive T cells from peripheral blood stem cell and umbilical cord blood grafts by the combined use of trimetrexate and interleukin-2 immunotoxin. Biol Blood Marrow Transplant. 2004;10:772. 26. Davies JK, Koh MB, Lowdell MW. Antiviral immunity and T-regulatory cell function are retained after selective alloreactive T-cell depletion in both the HLA-identical and HLA-mismatched settings. Biol Blood Marrow Transplant. 2004;10:259. 27. Andre-Schmutz I, Le Deist F, Hacein-Bey-Abina S, et al. Immune reconstitution without graft-versus-host disease after haemopoietic stem-cell transplantation: a phase 1/2 study. Lancet. 2002;360:130. 28. Solomon SR, Mielke S, Savani BN, et al. Selective depletion of alloreactive donor lymphocytes: a novel method to reduce the severity of graft-versus-host disease in older patients undergoing matched sibling donor stem cell transplantation. Blood. 2005;106:1123.
Alloreactive T cell clonotype recruitment in a mixed lymphocyte reaction: Implications for graft engineering Phillip Scheinberga, David A. Priceb, David R. Ambrozakc, A. John Barretta, and Daniel C. Douekb a
Hematology Branch, National Heart, Lung, and Blood Institute, Bethesda, Md., USA; bHuman Immunology Section; cImmunology Laboratory, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Md., USA (Received 20 December 2005; revised 22 February 2006; accepted 2 March 2006)
Objective. The selective elimination of alloreactive T cells from donor stem cell grafts prior to hematopoietic stem cell transplantation (HSCT) is an important goal in the prevention of graft-vs-host disease (GVHD). However, in HLA-identical donor-recipient pairs, it has proven difficult to identify alloreactive T cells using in vitro systems pretransplant due, in part, to their low frequency and a lack of methodological standardization. To better understand the alloresponse between HLA-identical related pairs, we characterized the alloreactive T cells generated in a mixed lymphocyte reaction (MLR) assay system. Methods. HSCT donor peripheral blood mononuclear cells (responder) were labeled with carboxyfluorescein diacetate, succinimidyl ester (CFSE) dye and cocultured with irradiated HSCT recipient cells (stimulator) in a one-way MLR. Alloreactive T cells were sorted by upregulation of activation markers (CD25 in most cases) and the responding clonotypes were defined by sequencing the complementarity region 3 (CDR3) of the T cell receptor b-chain. Results. We show that the recruitment of alloreactive CD4+ T cells is highly variable. Oligoclonal CD4+ T-cell expansions in repeated MLRs performed in the same donor-recipient pair showed inconsistent recruitment of clonotypes. The recruitment of alloreactive CD8+ T cells was more consistent in repeated assays, with the same clonotypes identified in the same donorrecipient pair performed under different conditions. Conclusion. Taken together, our data show that even in culture conditions constrained to eliminate background proliferation, stochastic events and low precursor frequencies preclude reproducible elicitation of immunodominant T cell clonotypes with the potential to cause GVHD. Ó 2006 International Society for Experimental Hematology. Published by Elsevier Inc.
The development of ex vivo coculture assays between donor and recipient pairs undergoing hematopoietic stem cell transplantation (HSCT) has been used to demonstrate alloimmune T cell responses in vitro that correspond to events in vivo postgraft. Since its first introduction by Bach et al., the mixed lymphocyte reaction (MLR) has been used to identify alloreactivity between donor-recipient pairs in organ and stem cell transplant recipients [1]. However, the lack of standardization prevents its routine widespread use. Many variable parameters must be validated formally in order to ensure the reproducibility of this approach; these include the nature of the stimulator cells that will best recruit alloreactive T cells, the techniques
Offprint requests to: Phillip Scheinberg, M.D., Hematology Branch, NHLBI, 10 Center Drive, Building 10 CRC, Rm 3-5140, MSC 1202, Bethesda, MD 20892-1202; E-mail: [email protected]
employed to identify responding cells, and the optimal culture conditions to enable maximal resolution of alloreactive T cells. In mismatched MLRs, immunodominant alloresponses to major histocompatibility antigens (MaHA) can be identified, as has been illustrated in a pretransplant HLA-DRmismatched MLR, where a single CD4D T cell clonotype identified in vitro accounted for almost 100% of the circulating CD4D T cells 100 days postgraft, and correlated with emergence of graft-vs-host disease (GVHD) [2]. In the HLA-matched setting, minor histocompatibility antigens (MiHA) are perceived as the main stimulus for donor alloreactive T-cell activation and expansion in vivo, resulting in GVHD and graft-vs-leukemia (GVL) effects [3,4]. Although the antigenic disparity between HLA-identical siblings involves hundreds of different MiHA, only a few immunodominant antigens are likely to trigger an antihost T-cell response [5]. Furthermore, MLRs between
0301-472X/06 $–see front matter. Copyright Ó 2006 International Society for Experimental Hematology. Published by Elsevier Inc. doi: 10.1016/j.exphem.2006.03.001
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
HLA-identical HSCT related pairs usually elicit either in a minimal or no response, despite the fact that such donor lymphocytes can induce lethal graft-vs-host reactions in the transplanted recipient. The identification of alloreactive T cells in the matched MLR setting is hampered by their very low precursor frequency, estimated to be less than 0.1% [6,7]. Therefore, an assay with a high signal-to-noise ratio, with minimal or no background, is critical for the appropriate characterization of alloreactive clonotypes. Another limiting factor in the recruitment of alloreactive T cells in the matched MLR is inadequate antigen presentation in vitro. For this reason, modified MLRs have been used in which cytokines are added to the culture or stimulators are pretreated to augment their antigen presenting capacity [8]. The problem with these approaches is that background also increases, which may limit the specific identification of alloreactive T cells. In the present study, we developed modifications that increase the sensitivity of the standard MLR to enhance reliable identification of alloreactive T cell clonotypes. This was achieved primarily by manipulations that reduced nonspecific background stimulation and the augmentation of antigen presentation based on the use of different stimulator cell populations in the same donor-recipient pair. Alloreactive T cells were isolated by flow cytometry based on the upregulation of a T-cell activation marker (CD25 in most experiments) and characterized by sequencing the complementarity-determining region 3 (CDR3) of the T cell receptor (TCR) b-chain.
Material and methods Patients All subjects tested gave written permission for their blood to be used for research under Institutional Review Board–approved National Heart, Lung, and Blood Institute stem cell allotransplantation protocols. Isolation of human peripheral blood mononuclear cells (PBMCs) A leukapheresis was collected from donor and recipient prior to transplantation. Peripheral blood mononuclear cells (PBMCs) were prepared by density-gradient centrifugation over Ficoll-Paque (MP Biomedicals, Aurora, OH, USA); washed in culture medium (RPMI-1640; Rockville, MD, USA), 10% heat-inactivated human AB serum (Gemini Bioproducts, Woodland, CA, USA), and 65 ug/mL gentamicin (Biosource, Rockville, MD, USA); and then cryopreserved in liquid nitrogen. Generation of monocyte-derived dendritic cells (DCs) PBMCs were resuspended in culture medium at a final concentration of 3 to 5 3 106 cells/mL and incubated in a standard tissue culture flask for 2 hours at 37 C in a 5% CO2-containing atmosphere. Nonadherent cells were removed by vigorous pipetting and placed in a 50-mL Falcon (supernatant). The remaining adherent cells were cultured in medium (containing 1% heat-inactivated human AB serum) supplemented with 200 ng/mL recombinant human
789
GM-CSF (R&D Systems, Minneapolis, MN, USA) and 4 ng/mL recombinant human interleukin IL-4 (R&D Systems). Fresh cytokines were added every 2 to 3 days. For the maturation of DCs, culture medium was replaced on day 6, and 1100 U/mL recombinant human TNF-a (R&D Systems) was added for 24 hours [9]. Generation of activated B cells The cells in the supernatant were washed in culture medium, and B cells were positively separated by using the MIDI-magnetic cell sorting (MACS) protocol provided by the manufacturer (Miltenyi Biotec, Auburn, CA, USA). Briefly, after centrifugation, the cell pellet was resuspended in 80 uL buffer (phosphate-buffered saline [PBS], 0.5% heat inactivated human AB serum); 20 uL anti-CD19 microbeads/107 cells were then added and the cells were incubated with the beads for 15 minutes at 4 C. Cells were then washed twice, resuspended in 500 uL buffer, and magnetically separated with an LSD column. After rinsing the column, the fraction with the magnetically labeled cells (CD19D) in the plunger was flushed and counted with a hemocytometer. Positively selected B cells were cultured for 6 days in culture medium (1-2 3 106/ mL) in a tissue culture flask and synthetic CpG oligodeoxynucleotide (CpG; TCG TCG TTT TGT CGT TTT GTC GTT; Coley Pharmaceutical Group, Ottawa, Canada) was added to a final concentration of 10 ug/mL [10]. Generation of activated T cells A 25-cm2 tissue culture flask was prepared for polyclonal T-cell stimulation in the solid phase by addition of 1 ug anti-CD3 monoclonal antibody (Muromonab-CD3 Ortho Biotech Products, Raritan, NJ, USA) diluted in 2 mL PBS, followed by overnight incubation at 37 C in a 5% CO2-containing atmosphere. The antibody suspension was removed the following day from the coated flask which was rinsed vigorously with PBS. The unlabeled cell fraction from above (effluent) was resuspended in medium (1-2 3 106 cells/mL) supplemented with 100 U/mL IL-2 (Peprotech, Rocky Hill, NJ, USA) and cultured in the anti-CD3 monoclonal antibody treated tissue culture flask for 6 days at 37 C in a 5% CO2-containing atmosphere [11]. The culture medium was changed as needed. Carboxyfluorescein diacetate, succinimidyl ester (CFSE) labeling Responder PBMCs (1-2 3 106/mL) were labeled with 0.25 uM CFSE (Molecular Probes, Carlsbad, CA, USA) in serum-free PBS for 7 minutes at 37 C. Filtered human AB serum was then added to stop the reaction, and the cells were washed with RPMI-1640. The supernatant was decanted after centrifugation and the cell pellet was resuspended in 1 mL filtered human AB serum before washing again with RPMI-1640. Mixed lymphocyte reaction assay culture Initial allogeneic MLRs were performed by culturing 20 to 50 3 106 responder (HSCT graft donor) PBMCs with irradiated (2500 rads) stimulator (HSCT graft recipient) PBMCs at a 1:1 ratio for 6 days in culture medium. Later, allogeneic MLRs were performed by culturing 10 to 20 3 106 responder cells (HSCT graft donor) with irradiated (2500 rads) DCs, B or T cells (stimulator cells derived from the HSCT recipient as described above), for 3 days. The responder-to-stimulator ratio was 8:1 for DCs (DCMLR), 4:1 for B cells (B-MLR), and 1:1 for T cells (T-MLR). An autologous MLR under identical conditions was performed
790
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
in parallel with each allogeneic MLR. All MLRs were performed on samples from sibling pairs matched at 6/6 HLA loci (HLA-A, B, and DR), unless otherwise indicated. Depletion of cell fractions according to the expression of surface activation markers Responder cells were depleted of the CD25D fraction one day after CFSE labeling, and before the MLR. PBMCs were labeled with anti-CD25 magnetic microbeads (Miltenyi Biotec, Auburn, CA, USA) according to protocol provided by the manufacturer. The magnetically labeled CD25D cells were retained on the MSD column; cells in the effluent (CD252) were counted and used in the MLR as responders. Removal of CD25D, CD69D, and CD71D cells was performed in one HLA-matched MLR. Responder PBMCs were stained with anti-CD25, anti-CD69, and anti-CD71 phycoerythrin (PE)-conjugated mAbs (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA), washed, and bound to anti-PE microbeads (20 mL/107 cells; Miltenyi Biotec, Auburn, CA, USA) according to protocol provided by the manufacturer. The cells were then washed again and subjected to magnetic separation. Cells expressing one or more of the selection antigens (CD25DCD69DCD71D) were magnetically labeled and retained on the MSD column; cells in the effluent were counted and used in the MLR as responders. Flow cytometry The expression of surface antigens on treated DCs, B and T cells were analyzed by flow cytometry. Conjugated mAbs used for phenotypic characterization of the various cell populations were as follows: 1) treated DCs: CD1a-PE and CD83-fluorescein isothiocyanate (FITC); 2) B cells: CD19-FITC, CD80-PE, and CD86-allophycocyanin (APC; Becton Dickinson Pharmingen, San Diego, CA, USA); and 3) T cells: CD3-FITC and CD25-PE. The purity of activated T and B cells was greater than 95%. All T cells used as stimulators were activated as evidenced by CD25 upregulation. B cells following culture with CpG had significant upregulation of CD80 and CD86 molecules compared to baseline. The purity of mature DCs was greater than 80% (data not shown). The expression of T-cell activation markers on the CFSElabeled responder cells was analyzed by six-color flow cytometry. The following directly conjugated mAbs were used: CD25-PE; CD11c-, CD14-, CD16-, CD19 PE-Cy5 (dump channel); CD4 PE-Cy7; and CD8 APC-Cy7. In one matched MLR, a combination of T-cell activation markers was used in the APC channel: CD25, CD69, and CD71 APC. 7-AAD was used as a viability marker for all sorts. Flow cytometric cell sorting All sorts were performed with a modified FACS DIVA (BD Pharmingen). Instrument set-up was performed according to the manufacturer’s instructions. All sorts were performed at 25 lb/in2. Instrument compensation was performed with antibody capture beads (BD Pharmingen) stained singly with individual mAbs used in the test samples. Sort gates were set such that only donor-derived (CFSE-labeled) activated cells were included. In the initial 6-day MLRs, in which PBMCs were used as stimulators, CD4DCFSEDhi and CD4DCFSEDlo were sorted separately. Later, in the three-day MLRs, CD4DCFSEDhi and CD8DCFSEDhi cells that had upregulated activation marker(s) were included in the sorting gate. All cells were sorted into RNAlater (Ambion, Austin, TX, USA) and frozen at 280 C.
Clonotypic assay Specific T cell clonotypes in the sorted populations were identified using a strand-switch–anchored reverse transcriptase polymerase chain reaction (RT-PCR) that amplifies all expressed TCRB gene products without bias [12,13]. Briefly, mRNA was extracted from the sorted antigen-activated T cells (Oligotex kit, Qiagen, Valencia, CA, USA). Amplification of TCRB VDJ junctions was performed by using a modified version of the SMART procedure (switching mechanism at the 50 end of RNA transcript) to generate cDNA and a template-switch–anchored PCR with a 30 C-region primer. The PCR product was ligated into the pGEMT Easy vector (Promega, Madison, WI, USA) and used to transform Escherichia coli. Individual colonies were selected, amplified by PCR with M13 primers, and then sequenced to obtain TCRB CDR3 sequences corresponding to all sorted T cells that had been activated in the MLR.
Results HLA-matched MLR with PBMC stimulators and unmanipulated responders PBMCs were initially used as the stimulating cell population. The ratio of donor (responder) to patient (stimulator) cells was 1:1 in these assays. Cultures were incubated for 6 days prior to washing and antibody labeling. Plots were gated on CD4D and CD8D T cells, and CD25DCFSEDlo and CD25DCFSEDhi populations were sorted separately for RNA extraction (Fig. 1). As the responding CD8D T cell populations were small (!20 cells), further study
Figure 1. The top panels are gated on CD4D T cells; the bottom panels are gated on CD8D T cells. Control cultures are shown on the left in each case; experimental MLRs are shown on the right. A 0.27% CD4DCD25DCFSEDlo and 0.8% CD4DCD25DCFSEDhi response above background is shown; cells were sorted directly into RNAlater for clonotype analysis. The clonotypes detected in each of these sorted populations are shown in Table 1.
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
791
Table 1. Replicate MLRs from same donor-recipient pair with unmanipulated responding cells No. of distinct clonotypes
CDR3 region of clonotype(s) with frequency O10%
Freq. (%)
Repertoire
Sorted cells #1 CD4DCD25DCFSEDhi
6
18
32 18 11 14 11 14
oligoclonal
CD4DCD25DCFSEDlo
CASSPWSRQQRELRNEQYF CASSPRGDTQYFG CSARTPGLAGGETQYF CSAPKGGLPLPYNEQFF CASSFSGRNNEQFF CASRITANNEQFF
Sorted cells #2 CD4DCD25DCFSEDhi CD4DCD25DCFSEDlo
2 36
CSAREGATTYEQYF CASRITANNEQFFz
98 3z
oligoclonal polyclonal
MLR from same donor-recipient pair
polyclonal
MLR, mixed lymphocyte reaction; CDR3, complementarity-determining region 3. z This clonotype was identified previously but was present only once in the repeat MLR.
of these cells was precluded. The responding CD4D T cells segregated into two populations based on the degree of CFSE fluorescence: bright (CFSEDhi) and dim (CFSEDlo). Clonotyping of the sorted CD4DCD25DCFSEDhi and CD4DCD25DCFSEDlo T cell populations generally revealed polyclonal TCRB sequences with no shared clonotypes between populations. There was also no identity between clonotypes in replicate MLR cultures performed several weeks later using the same PBMC combinations (Table 1). On only one occasion in a total of 16 pairs studied and over 1000 sequences analyzed was the same CDR3 sequence (CASRITANNEQFF) identified in the CD4DCD25D CFSEDlo population in duplicate MLRs (Table 1). HLA-matched MLR with PBMC stimulators and CD25-depleted responders To minimize background in the clonotype analysis, donor (responder) cells were depleted of CD25-expressing populations after CFSE labeling but prior to the MLR. The MLRs were then cultured for 3 days before washing and antibody labeling. Plots were gated on CD4D and CD8D T cells, and CD25DCFSEDhi cells were sorted for RNA extraction (Fig. 2). Again, the low magnitude of the responding CD8D T cell populations (!20 cells) precluded clonotypic analysis. The CD4D T cell clonotypes identified in duplicate MLRs were different; one MLR yielded only one clonotype, while the other yielded 45 clonotypes with no clear hierarchy. The same experiment was repeated with the same donor-recipient pair; once more, one MLR yielded only one clonotype that differed from the single clonotype identified in the previous MLR with this PBMC combination, while the other yielded 18 different sequences with 3 dominant clonotypes (Table 2). HLA-DR-mismatched MLR with B cell, T cell, and DC stimulators and CD25-depleted responders To minimize the heterogeneity of the stimulator cell populations, monocyte-derived DCs, activated B cells, and activated T cells were generated as described in the Materials
and Methods. The following donor (responder) to patient (stimulator) ratios, based on pilot optimization experiments, were used: DCs, 8:1; activated B cells, 4:1; and activated T cells, 1:1. The MLRs were incubated for 3 days before washing and antibody labeling. Plots were gated on CD4D and CD8D T cells, and CD25DCFSEDhi populations were sorted separately for RNA extraction (Fig. 3). The clonotypes derived from the CD4D T cell sort in each of the T-, B-, and DC-MLRs were polyclonal and distinct from each other (Table 3). There were 10 clonotypes in the T-MLR that had a frequency of greater than 5% each; in the B- and DC-MLR, there were no clonotypes that had a frequency greater than 5%. The CD8D sorted T cells were more oligoclonal, with approximately five different clonotypes in each MLR. One CD8D T cell clonotype (TCRB CDR3 sequence CSARDLAAGGLGETQYF) was present in all three MLRs, while two further clonotypes
Figure 2. Three-day MLRs in an HLA-matched donor-recipient pair performed in parallel. Responder (recipient) cells were CD25 depleted after CFSE labeling prior to the MLR. A small response above background was observed in the CD4D T cell population (upper panels); the constituent clonotypes are shown in Table 2. No CD8D T cell response was observed. Autologous control cultures in each case are shown in the left panels.
792
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
Table 2. Replicate MLRs from the same donor-recipient pair with CD25-depleted responding cells MLR from same donor-recipient pair performed in parallel (MLR 1 and 2)
CDR3 region of clonotype(s) with frequency O10%
No. of distinct clonotypes
Freq. (%)
Repertoire
Sorted cells–MLR1 & 2 1: CD4DCD25DCFSEDhi 2: CD4DCD25DCFSEDhi
1 45
CSAGTSGGTFTKQYF None
100
monoclonal polyclonal
Sorted cells–MLR 1 & 2 (repeat) 1: CD4DCD25DCFSEDhi 2: CD4DCD25DCFSEDhi
1 18
CASSQGDRSVGTQYF CASSWKLTGDRYNEQFF CASRLGRLNTEAFF CASSFLAGGPGEQFF
100 17 15 15
monoclonal polyclonal
MLR, mixed lymphocyte reaction; CDR3, complementarity-determining region 3.
(CASSLAPVNYGYTF and CASSPDGYGYTF) were present in the B- and DC-MLRs, but not the T-MLR (Table 3). HLA-matched MLR with B cell, T cell, and DC stimulators and CD25-depleted responders The same experiment was performed with samples from HLA-matched siblings; in this case, responder PBMCs were depleted of CD25D, CD69D, and CD71D cells prior to the MLR, and a combination of activation markers was used in the APC channel as described in the Materials and Methods. CD69 and CD71 were included to enhance the identification of alloreactive T cells in the matched MLR [14–16]. With the exception of the B-MLR, the clonotypes from the responding CD4D T cell populations were polyclonal and distinct from each other (Table 4). In the sorted CD8D T cell populations, the response was more oligoclonal (Table 4). One clonotype (TCRB CDR3
Figure 3. Mixed lymphocyte reaction with HLA-DR-mismatched sibling PBMCs and prior CD25 depletion of responder cells. Three different stimulators were used: T cells (top row), B cells (middle row), and dendritic cells (DCs; bottom row). The response is greater with DCs and lesser with T cells; in all cases, both CD4D and CD8D responding T cell populations were sorted for clonotypic analysis (Table 3).
CASSPAQHPPRERDTQYF) was present in the T- and B-MLR, but not in the DC-MLR.
Discussion The antigens that drive the T cell response in the HLAmatched allogeneic setting are for the most part unknown, but include broadly expressed minor histocompatibility antigens (e.g., H-Y, HA-8), tissue-restricted minor histocompatibility antigens (e.g., HB-1), and antigens that are either overexpressed or aberrantly expressed by malignant cells (e.g., PR-1, WT-1) [17–19]. Furthermore, the recruitment and characterization of alloreactive T cells between HLA-identical siblings has long been a challenge because such T cells are present at very low precursor frequencies. Here, we report methodological inconsistencies that complicate the reliable identification of alloreactive T cell clonotypes in current HLA-matched donor-recipient MLR protocols and describe manipulations that illuminate the underlying reasons for these limitations. Initially, we used a one-way MLR with irradiated PBMCs as stimulators; cells that had upregulated CD25 after 6 days were isolated by flow cytometric sorting as described previously [20,21]. With this approach, alloreactive CD4D T cell clonotypes have been identified pretransplant and tracked in vivo post-HSCT by qPCR; expansions have been associated with acute GVHD [2]. In our experiments, dividing and nondividing cells were analyzed separately. Consistent with previous reports, CD8D T cell responses were not observed under these conditions [22]. Our analysis was therefore generally restricted to CD4D T cells. By sorting CD4DCD25DCFSEDlo T cells, we expected to identify oligoclonal populations that were responding in an alloreactive manner to mHAg. In most cases, polyclonal T cell populations were elicited. When oligoclonal populations were observed, the same clonotype was only rarely detected in replicate or parallel MLRs (Fig. 1; Table 1). Possible explanations for this inconsistency include: 1) nonspecific stimulation of T cells in the MLRs resulting in a ‘‘background’’ that would become relatively more
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
793
Table 3. Clonotypes from sibling HLA-DR-mismatched MLR using T cells, B cells, and DC stimulators Stimulators and sorted cell populations
No. of distinct clonotypes
CD4D sorted cells ‘‘T’’: CD4DCD25DCFSEDhi
39
‘‘B’’: CD4DCD25DCFSEDhi ‘‘DC’’: CD4DCD25DCFSEDhi
24 32
CD8D sorted cells ‘‘T’’: CD8DCD25DCFSEDhi
7
‘‘B’’: CD8DCD25DCFSEDhi
5
‘‘DC’’: CD8DCD25DCFSEDhi
3
CDR3 region of clonotype(s) with frequency O5%
CASSLRIGNTEAFF CASSQVLTGGGSPLHF CASSSNYQWETQYF CASSYSPLTGELFF CASTKVGTGLYEQYF CASRLADTQYF CASRTPGTSPLHF CSPTPGGRETQYG CASNQGLAGPQETQYF CASSLRDSDYGYTF None None CASSYYGGRWTGELFF CSARDLAAGGLGETQYFz CASSQATGRNYGYTF CSARDLAAGGLGETQYFz CASSLAPVNYGYTF* CASSPDGYGYTF* CSARDLAAGGLGETQYFz CASSLAPVNYGYTF* CASSPDGYGYTF*
Freq. (%)
Repertoire
8 8 8 5 5 5 5 5 5 5
polyclonal
polyclonal polyclonal 42 17 13 58 17 17 66 25 8
oligoclonal
oligoclonal
oligoclonal
MLR, mixed lymphocyte reaction; CDR3, complementarity-determining region 3; ‘‘T,’’ T-MLR; ‘‘B,’’ B-MLR; ‘‘DC,’’ DC-MLR. z clonotype recruited in T-, B-, and DC-MLR. *clonotype recruited in the B- and DC-MLR.
substantial in the presence of a small alloresponse; and 2) variable allostimulation of PBMCs. In order to determine whether the failure to consistently elicit alloreactive clonotypes consistently was due to the intrusion of background
noise from nonspecifically activated cells, we depleted CD25D cells from the responder population prior to the MLR. This maneuver did not improve the selectivity of the clones induced in MLR (Fig. 2; Table 2).
Table 4. HLA-matched MLR with T cells, B cells, and DC stimulators Stimulators and sorted cell populations
No. of distinct clonotypes
CD4D sorted cells ‘‘T’’: CD4DCD25DCFSEDhi
13
‘‘B’’: CD4DCD25DCFSEDhi ‘‘DC’’: CD4DCD25DCFSEDhi
3 19
CD8D sorted cells ‘‘T’’: CD8DCD25DCFSEDhi
6
‘‘B’’: CD8DCD25DCFSEDhi
4
‘‘DC’’: CD8DCD25DCFSEDhi
4
z
Same clonotype identified in T- and B-MLR.
CDR3 region of clonotype(s) with frequency $10%
Freq. (%)
Repertoire
CSAGGQGAYEQYF CAISAGQGEGYGYTF CAWSISGLETQYF CASSLAGAWDTEAF CASSRVGLPSDEQY CASSTNPRGNQPQH CASSSPRDRTNEQY
18 18 18 88 14 11 11
polyclonal
CASSPNYSNQPQHF CASRRMTGGYDYEQYF CASSLSQGLEQPQHF CASFPGPWLAKNIQYF CASSPAQHPPRERDTQYFz CSARELAGASSYEQY CASSVNRQGGIYEQY CASSPAQHPPRERDTQYFz CASSRRSSYNEQF CASSFGLEGNSPLH
42 23 13 10 10 46 32 19 59 23
oligoclonal polyclonal
oligoclonal
oligoclonal
oligoclonal
794
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795
In an attempt to optimize allorecruitment and maintain a more homogeneous stimulator cell population, monocyte-derived DCs, activated T cells, and activated B cells were generated and cocultured with responder cells. The inconsistent recruitment of CD4D T cell clonotypes persisted. However, the CD8D T cell repertoire was more oligoclonal, and some clonotypes were present in MLRs performed with different stimulators. This was observed in both an HLADR-mismatched (Fig. 3; Table 3) and in an HLA-matched MLR (Table 4). As CD25 upregulation was infrequent in samples from HLA-identical siblings, CD69 and CD71 were added as activation markers; this allowed for the identification and sorting of both CD4D and CD8D T cell subsets. How should the finding of wide diversity and nonreproducibility in T cell clonotypes that are elicited in HLAmatched MLRs be interpreted? First, the diversity of CD4D T cell clonotypes elicited within a few days of culture is consistent with the known degeneracy of CD4 T-cell recognition [23]. Thus, the oligoclonal T cells elicited could have recognized a single peptide-MHC complex on the stimulator cells. In contrast, the degree of cross-reactivity in CD8D T cell populations might be lower [23]; this could explain the greater degree of consistency in the recruitment of alloreactive CD8D T cells observed in our system, although further experiments are required to confirm this. Nevertheless, for the CD4D T cell populations in particular, the nonreproducibility between duplicate and replicate cultures implies stochastic selection of individual clonotypes in these cultures. It is somewhat harder to explain the failure to elicit consistent oligoclonal responses in many cultures. This could be interpreted as either a broad base of weak responses to multiple antigens or background noise in the absence of a strong allostimulation. When compared to a more homogeneous stimulator population, such as EBV lymphoblastoid cell lines, allostimulation by PBMCs may be less consistent [24]. Second, antigens may not be presented properly to alloreactive T cells in culture by resting PBMCs. While an alloreactive T cell is likely to respond to antigens on the surface of monocytes, the same may not occur in the case of resting B or T lymphocytes. The more consistent and predictable results obtained with the HLA-mismatched pairs, within which antigenic disparity is presumably stronger, suggests that at least in this in vitro system the alloresponse in HLAmatched pairs represents a broad base of very weak alloresponses to multiple antigens. It is possible that only in prolonged culture or in vivo can a limited number of alloreactive clonotypes emerge to exert functional T cell responses causing, for example, GVHD, GVL, or rejection reactions. Methods for removing GVHD-causing T cells from the graft are being developed that deplete selected T cell subsets (‘‘selective allodepletion’’) recruited in an ex vivo MLR [15,25]. Selective allodepletion of alloreactive donor
T cells in vitro has been shown to preserve antiviral and antileukemic responses [24,26]. In addition, children undergoing HLA-haploidentical HSCT for congenital hematopoietic disorders given selectively depleted T cells did not develop severe GVHD [27]. We have reported similar promising results using selectively depleted T cell transplants in HLA-matched recipients [28]. The failure of the MLR approach used here to identify the alloresponse at the level of individual T cell clonotypes is disappointing. The implications of our observations, however, are important for graft engineering and the selective depletion of alloreactive T cells ex vivo prior to HLA-matched HSCT. While it remains possible that the removal of donor T cells activated by the recipient will be sufficient to reduce or prevent GVHD, the current technology is incapable of defining the basis of the alloreaction. Although the HTLP frequency assay has been used to grade the alloresponse and monitor efficiency of selective depletion approaches, standardization is lacking and reproducibility has been poor; therefore, there are currently no robust predictive assays for GVHD in the HLA-matched setting [28]. Acknowledgments This research was supported by the Intramural Research Program of the NIH; National Heart, Lung, and Blood Institute; and National Institute of Allergy and Infectious Diseases. David A. Price is a Medical Research Council (UK) Clinician Scientist.
References 1. Bach FH, Voynow NK. One-way stimulation in mixed leukocyte cultures. Science. 1966;153:545. 2. Michalek J, Collins RH, Hill BJ, Brenchley JM, Douek DC. Identification and monitoring of graft-versus-host specific T-cell clone in stem cell transplantation. Lancet. 2003;361:1183. 3. Chao NJ. Minors come of age: Minor histocompatibility antigens and graft-versus-host disease. Biol Blood Marrow Transplant. 2004;10:215. 4. Falkenburg JH, Marijt WA, Heemskerk MH, Willemze R. Minor histocompatibility antigens as targets of graft-versus-leukemia reactions. Curr Opin Hematol. 2002;9:497. 5. Perreault C, Roy DC, Fortin C. Immunodominant minor histocompatibility antigens: the major ones. Immunol Today. 1998;19:69. 6. Sherman LA, Chattopadhyay S. The molecular basis of allorecognition. Annu Rev Immunol. 1993;11:385. 7. Wang XN, Taylor PR, Skinner R, et al. T-cell frequency analysis does not predict the incidence of graft-versus-host disease in HLA-matched sibling bone marrow transplantation. Transplantation. 2000;70:488. 8. Bishara A, Brautbar C, Nagler A, et al. Prediction by a modified mixed leukocyte reaction assay of graft-versus-host disease and graft rejection after allogeneic bone marrow transplantation. Transplantation. 1994;57:1474. 9. Nguyen XD, Eichler H, Dugrillon A, Piechaczek C, Braun M, Kluter H. Flow cytometric analysis of T cell proliferation in a mixed lymphocyte reaction with dendritic cells. J Immunol Methods. 2003;275:57. 10. Krieg AM. Antitumor applications of stimulating toll-like receptor 9 with CpG oligodeoxynucleotides. Curr Oncol Rep. 2004;6:88. 11. Solomon SR, Tran T, Carter CS, et al. Optimized clinical-scale culture conditions for ex vivo selective depletion of host-reactive donor lymphocytes: a strategy for GvHD prophylaxis in allogeneic PBSC transplantation. Cytotherapy. 2002;4:395.
P. Scheinberg et al./ Experimental Hematology 34 (2006) 788–795 12. Douek DC, Betts MR, Brenchley JM, et al. A novel approach to the analysis of specificity, clonality, and frequency of HIV-specific T cell responses reveals a potential mechanism for control of viral escape. J Immunol. 2002;168:3099. 13. Price DA, West SM, Betts MR, et al. T cell receptor recognition motifs govern immune escape patterns in acute SIV infection. Immunity. 2004;21:793. 14. Fehse B, Frerk O, Goldmann M, Bulduk M, Zander AR. Efficient depletion of alloreactive donor T lymphocytes based on expression of two activation-induced antigens (CD25 and CD69). Br J Haematol. 2000;109:644. 15. Koh MB, Prentice HG, Lowdell MW. Selective removal of alloreactive cells from haematopoietic stem cell grafts: graft engineering for GVHD prophylaxis. Bone Marrow Transplant. 1999;23:1071. 16. van Dijk AM, Kessler FL, Stadhouders-Keet SA, Verdonck LF, de Gast GC, Otten HG. Selective depletion of major and minor histocompatibility antigen reactive T cells: towards prevention of acute graftversus-host disease. Br J Haematol. 1999;107:169. 17. Goulmy E, Schipper R, Pool J, et al. Mismatches of minor histocompatibility antigens between HLA-identical donors and recipients and the development of graft-versus-host disease after bone marrow transplantation. N Engl J Med. 1996;334:281. 18. Mutis T, Gillespie G, Schrama E, Falkenburg JH, Moss P, Goulmy E. Tetrameric HLA class I–minor histocompatibility antigen peptide complexes demonstrate minor histocompatibility antigen–specific cytotoxic T lymphocytes in patients with graft-versus-host disease. Nat Med. 1999;5:839. 19. Dolstra H, Fredrix H, Maas F, et al. A human minor histocompatibility antigen specific for B cell acute lymphoblastic leukemia. J Exp Med. 1999;189:301.
795
20. Cavazzana-Calvo M, Fromont C, Le Deist F, et al. Specific elimination of alloreactive T cells by an anti-interleukin-2 receptor B chain–specific immunotoxin. Transplantation. 1990;50:1. 21. Montagna D, Yvon E, Calcaterra V, et al. Depletion of alloreactive T cells by a specific anti-interleukin-2 receptor p55 chain immunotoxin does not impair in vitro antileukemia and antiviral activity. Blood. 1999;93:3550. 22. Martins SL, St John LS, Champlin RE, et al. Functional assessment and specific depletion of alloreactive human T cells using flow cytometry. 2004;104:3429. 23. Wilson DB, Wilson DH, Schroder K, et al. Specificity and degeneracy of T cells. Mol Immunol. 2004;40:1047. 24. Amrolia PJ, Muccioli-Casadei G, Yvon E, et al. Selective depletion of donor alloreactive T cells without loss of antiviral or antileukemic responses. Blood. 2003;102:2292. 25. Szabolcs P, Park KD, Marti L, et al. Superior depletion of alloreactive T cells from peripheral blood stem cell and umbilical cord blood grafts by the combined use of trimetrexate and interleukin-2 immunotoxin. Biol Blood Marrow Transplant. 2004;10:772. 26. Davies JK, Koh MB, Lowdell MW. Antiviral immunity and T-regulatory cell function are retained after selective alloreactive T-cell depletion in both the HLA-identical and HLA-mismatched settings. Biol Blood Marrow Transplant. 2004;10:259. 27. Andre-Schmutz I, Le Deist F, Hacein-Bey-Abina S, et al. Immune reconstitution without graft-versus-host disease after haemopoietic stem-cell transplantation: a phase 1/2 study. Lancet. 2002;360:130. 28. Solomon SR, Mielke S, Savani BN, et al. Selective depletion of alloreactive donor lymphocytes: a novel method to reduce the severity of graft-versus-host disease in older patients undergoing matched sibling donor stem cell transplantation. Blood. 2005;106:1123.