Identification of a Novel UTY-Encoded Minor Histocompatibility Antigen

Authors


Dr B. K. Mortensen, MD, Department of Hematology, HCT Unit, Section 4041, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen Ø, Denmark. E-mail: bokokmortensen@hotmail.com; and A. Stryhn, PhD, Laboratory of Experimental Immunology, Faculty of Health Sciences, University of Copenhagen, Blegdamsvej 3, 2200 Copenhagen N, Denmark. E-mail: astryhn@sund.ku.dk

Abstract

Minor histocompatibility antigens (mHags) encoded by the Y-chromosome (H-Y-mHags) are known to play a pivotal role in allogeneic haematopoietic cell transplantation (HCT) involving female donors and male recipients. We present a new H-Y-mHag, YYNAFHWAI (UTY139–147), encoded by the UTY gene and presented by HLA-A*24:02. Briefly, short peptide stretches encompassing multiple putative H-Y-mHags were designed using a bioinformatics predictor of peptide-HLA binding, NetMHCpan. These peptides were used to screen for peptide-specific HLA-restricted T cell responses in peripheral blood mononuclear cells obtained post-HCT from male recipients of female donor grafts. In one of these recipients, a CD8+ T cell response was observed against a peptide stretch encoded by the UTY gene. Another bioinformatics tool, HLArestrictor, was used to identify the optimal peptide and HLA-restriction element. Using peptide/HLA tetramers, the specificity of the CD8+ T cell response was successfully validated as being HLA-A*24:02-restricted and directed against the male UTY139–147 peptide. Functional analysis of these T cells demonstrated male UTY139–147 peptide-specific cytokine secretion (IFNγ, TNFα and MIP-1β) and cytotoxic degranulation (CD107a). In contrast, no responses were seen when the T cells were stimulated with patient tumour cells alone. CD8+ T cells specific for this new H-Y-mHag were found in three of five HLA-A*24:02-positive male recipients of female donor HCT grafts available for this study.

Introduction

In the setting of allogeneic haematopoietic cell transplantation, minor histocompatibility antigens (mHags) are known to play an important role in generating the immune responses leading to graft-versus-leukaemia (GVL) effects and graft-versus-host-disease (GVHD) [1, 2]. mHags are results of polymorphisms in the recipients genome encoding proteins that can be recognized by donor T cells [1]. More than 30 mHags are now characterized [3]. Approximately one-third of these are encoded by the Y-chromosome (H-Y-mHags) [4–9] and two of them by the UTY gene [10, 11]. Existence of many more mHags seems reasonable considering the large number of genetic polymorphisms, and the fact that most mHags reported so far have been restricted by a few common HLA alleles leaving many more to be examined. Characterization of more mHags is required, if patients in the future should benefit from treatment based on T cells specific for mHags. mHag identification methods classically start by having an mHag reactive clone and subsequently search for its peptide specificity and the HLA restriction using techniques such as cDNA-expressional cloning [10–13], high-performance liquid chromatography, genetic linkage analysis [4–6, 14–18] or whole genome association (WGA) scanning [19]. Recently, techniques based on bioinformatics have been introduced to reverse this process by predicting new mHags [9, 20], and then search for the corresponding T cell responses. Here, we present a novel H-Y-mHag and a method to detect new mHags based on such a reverse immunology approach.

Materials and methods

Human subjects.  Blood was obtained from patients before and longitudinally after non-myeloablative conditioning haematopoietic cell transplantation (NMC-HCT) at the Hematopoietic Cell Transplantation Unit, Department of Hematology, Rigshospitalet, Copenhagen. Blood was also obtained from donors and healthy volunteers. All provided written informed consent in accordance with the declaration of Helsinki. Peripheral blood mononuclear cells (PBMCs) were stored at minus 150 °C after lymphoprep separation.

HLA typing.  Patients and donors were high-resolution typed at HLA-A and HLA-B loci using sequence-based typing.

Prediction of potential mHag epitopes.  In this study, we report the results for one patient, UPN (unique patient number) 694, but in parallel, H-Y-mHags were likewise predicted for 31 other male recipients of female donor HCT grafts. Twenty-seven genes or gene families are present on the Y-chromosome. Eighteen genes (TGIF2LY, PCDH11Y, SRY, RPS4Y1, ZFY, AMELY, TBL1Y, PRKY, USP9Y, DBY, UTY, TMSB4Y, NLGN4Y, Cyorf15A, Cyorf15B, SMCY, EIF1AY and RPS4Y2) were screened for potential H-Y-mHags. The remaining nine genes are ampliconic gene families and these were excluded because their expression is restricted to the testis [21]. The Ensembl database (http://www.ensembl.org) was used to obtain the protein products encoded by these 18 genes. Potential mHags were predicted by the NetMHCpan method (http://www.cbs.dtu.dk/services/NetMHCpan/) [22] using the 15 most common HLA-A and HLA-B alleles of the patients and considering all 8-mer to 11-mer peptides within the selected proteins. A peptide was considered to be a potential epitope, if the predicted affinity was below 500 nm. To reduce the number of peptides for testing, we removed all predicted mHags, which were also encoded by the X-chromosome, and all submers (i.e. shorter peptides), which were contained within a longer selected peptide. The reason for the latter was that the submers can be generated from the longer versions by peptide cleavage during in vitro cellular immune response assays. Finally, only the top 30 strongest predicted binders were included for each of the 15 HLA alleles. Predicted H-Y-mHags, which were selected owing to restriction to alleles of another male patient, but predicted to be cross-presented by an HLA molecule of patient UPN694, were also included in this study. In total, a panel of 324 peptides were synthesized and 141 peptides included epitopes predicted to bind to the HLA-A and HLA-B molecules of patient UPN694.

Prediction of the optimal epitope and HLA restriction.  HLArestrictor (http://www.cbs.dtu.dk/services/HLArestrictor/) was used to identify the optimal peptide and HLA-restriction element as previously described [23].

HLA molecules, peptides and stability assays.  HLA class I (HLA-I) molecules were produced recombinantly as previously described [24]. The peptides were synthesized by standard 9-fluorenylmethyloxycarbonyl chemistry, and purified by reversed-phase high-performance liquid chromatography (at least 80%, usually > 95% purity), and validated by mass spectrometry (Shafer-N, Copenhagen, Denmark).

The stabilities of peptide-HLA-I complexes were determined by a scintillation proximity assay as previously described (SPA) [25].

Generation, maturation and peptide pulsing of dendritic cells.  Dendritic cells (DCs) were generated by purification with magnetic CD14 beads according to the manufacturer’s instructions (Miltenyi, Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). CD14+ cells were cultured in 6 or 12 wells for 11 days in X-vivo 15, supplemented with 5% heat-inactivated human AB serum at 37 °C, 5% CO2 in humidified air. Granulocyte monocyte colony-stimulating factor (GM-CSF), final concentration 250 U/ml and IL-4, final concentration 500 U/ml, was added. Half of the media was changed every 3rd day and new cytokines were added. Forty-eight hours before use, the dendritic cells were matured by incubation with tumour necrosis factor-α (TNFα, 10 ng/ml), IL-1β (10 ng/ml), IL-6 (100 U/ml) and prostaglandin E2 (1 μg/ml). DCs were pulsed with peptides by incubation with the peptide, or the peptide mixture, at a final concentration of 2 μm each for 90 min at 37 °C, washed and irradiated (2000 rad).

In vitro stimulation of T cells.  Peripheral blood mononuclear cells from 2, 9 and 11 years post-HCT were used since only a limited amount of blood could be drawn at each time point and repeated analysis was needed for validation. PBMCs were stimulated with single peptides or peptides mixtures (final concentration 2 μm) in wells with 107 cells/cm2/ml. After 18–24 h incubation, the cells were washed, IL-2 (final concentration 50 U/ml) was added and the cells were expanded for another 10 days. Half of the media was changed every 2nd to 3rd day and new IL-2 was added. IL-15 (final concentration 15 ng/ml) was added from day 5. To increase the number of specific T cells, the cells were eventually harvested and stimulated with irradiated, peptide-pulsed DCs at T cell/DC ratio 7:1 and expanded for another 9 days with IL-2 and IL-15.

Intracellular cytokine staining with flow cytometry.  The cell cultures were harvested, washed and incubated with or without peptide or peptide mixtures, final concentration 2 μm for 4 h at 37 °C in 96 U wells. Brefeldin A (Sigma-Aldrich Danmark A/S, Brøndby, Denmark) was present for the last 3 h of incubation. EDTA final concentration 1.40 mm was added and the cells were incubated at room temperature (RT). Afterwards permeabilizing agent (Perm 2; Becton Dickinson a/s, Albertslund, Denmark) was added and the cells were incubated at RT, then washed and stained with different fluorochrome mixtures, i.e. anti-CD3-APCCy7, anti-CD4-PerCp, anti-CD8-APC, anti-CD69-PE, anti-TNF-α-FITC, anti-INF-γ-FITC (Biolegend; Becton Dickinson). The cells were then washed, fixed in 1% formaldehyde and analysed with flow cytometry on LSRII (Becton Dickinson) using diva software (Beckton Dickinson BD a/s, Albertslund, Denmark).

Tetramer staining.  Major Histocompatibility Complex (MHC) tetramers were produced as previously described using streptavidin conjugated with either PE or APC [24]. Cells were resuspended in 22 μl tetramer (concentration 269 nm). In some cases, double tetramer staining was performed by adding 22 μl of the tetramer complex labelled with PE and 22 μl of the same tetramer complex labelled with APC. After incubation at RT, the cells were washed and stained with anti-CD3-pacific blue, anti-CD4-APCCy7, anti-CD8-PerCP, anti-HLA-DR-FITC (Biolegend; Becton Dickinson). The cells were then washed, fixed in 1% formaldehyde and analysed with flow cytometry on LSRII (Becton Dickinson) using diva software.

Cytotoxicity assay.  To demonstrate cytotoxicity of the T cells specific for the mHag, DCs were pulsed with peptide, irradiated and incubated with the expanded T cell culture and anti-CD107a-PE (Biolegend), 2 μl/105 cells for 4 h at 37 °C, DC/T cell ratio 1:10. Brefeldin A was present the last 3 h. The cells were incubated with tetramers conjugated with APC, permeabilized and stained with anti-CD3-pacific blue, anti-CD4-APC-Cy7, anti-CD8-PerCP, anti-TNF-α-FITC and anti-INF-γ-FITC as described above, fixed in 1% formaldehyde and analysed with flow cytometry on LSRII using diva software.

Results

Characteristics of patient and donor

A male patient, UPN694, received non-myeloablative conditioning haematopoietic cell transplantation (NMC-HCT) for relapsed follicular lymphoma with his sister as donor, according to a protocol described earlier [26, 27]. He was not in remission at HCT and lymphoma cells could be demonstrated in the blood by flow cytometry. The lymphoma cells had a characteristic phenotype for follicular lymphoma cells: positive for CD10, CD19 and CD20, negative for CD5 and CD23, and lambda light chain restricted (data not shown). The NMC-HCT was complicated by tumour lysis syndrome, acute GVHD overall grade III and extensive chronic GVHD, but the patient was brought into remission and is still alive almost 12 years after HCT; he is off immunosuppression and in good condition. Patient and donor were HLA-identical siblings with HLA-A and HLA-B alleles: HLA-A*03:01, HLA-A*24:02, HLA-B*07:02, HLA-B*35:08.

Screening for cytokine responses

As described in materials and methods, a panel of 324 peptides was designed for H-Y-mHag screening of our cohort of male recipients of female HCT and 141 of these peptides contained epitopes predicted to bind to the HLA molecules of patient UPN694. The peptides were organized into a matrix with 12 rows and 12 columns, altogether generating 24 different mixtures (Fig. 1A) where each peptide was present in one row mixture and one column mixture only. Two-year post-HCT PBMCs were split into 24 different wells, and in vitro stimulated with each of the 24 peptide mixtures. Each culture was harvested, split in two and analysed for cytokine response by intracellular cytokine staining (ICS) either in the absence of peptide (background response) or presence of peptide mixtures (Fig. 1B). One column mixture (column 4) and one row mixture (row C) gave a significant CD8+ T cell response (Fig. 1C). The peptide shared between the two stimulatory mixtures was the 11-mer YFYYNAFHWAI (UTY137–147). Indeed, reanalysis of the PBMCs showed that the UTY137–147 peptide was responsible for the detected CD8+ T cell cytokine response. No significant CD4+ responses were observed (data not shown).

Figure 1.

 Peripheral blood mononuclear cells (PBMCs) from patient UPN694 2 years post-HCT were screened for recognition of 141 predicted HY minor peptides. (A) The peptides were pooled in 24 mixtures [12 row mixtures (mix A–L) and 12 column mixtures (mix 1–12)] organized by a 12 × 12 matrix. PBMCs were in vitro stimulated for 11 days with each of the 24 peptide mixtures. Each culture was split in two and analysed by intracellular cytokine staining (ICS) for response to the relevant peptide mixture versus no peptide (background response). (B) An example of the ICS analysis of T cells in vitro stimulated with peptide mixture 4, split and analysed in the presence or absence of peptide mixture 4. CD8+ T cells were gated, and CD69 versus cytokine (TNF-α and IFN-γ) stainings were plotted. The upper right quadrant gives the frequency of CD8+ cytokine-producing T cells, (C) the frequency of CD8+ cytokine-producing T cells for all 24 peptide mixtures. Black bars represent the background response (no peptide) and grey bars the specific response (peptide mix).

Determining the optimal peptide epitope and HLA restriction

The HLArestrictor tool was used to screen all possible 8–11-mers within the UTY137–147 peptide for potential binding to the patients HLA-A and HLA-B molecules. Only HLA-A*24:02 was predicted to bind UTY137–147 and some of its submers. These submers were synthesized, and peptide-HLA stability assays were performed for all available patient HLA-A and HLA-B molecules. The interaction between the 9-mer YYNAFHWAI (UTY139–147) and HLA-A*24:02 was the most stable of those examined (half-life 35 h at 37°, Table 1). HLA-A*03:01, HLA-B*07:02 and HLA-B*35:08 were not predicted to bind to UTY137–147 or any of its submers. Experimentally, no stable interactions with HLA-A*03:01 or HLA-B*07:02 (HLA-B*35:08 was not available) could be observed (data not shown). For control purposes, the X-chromosomal homologue of male UTY139–147, the UTX gene encoded female peptide, HYNAFQWAI (UTX142–150), and two male-female chimeric peptides, HYNAFHWAI and YYNAFQWAI, were also synthesized and analysed for HLA-A*24:02 stability (Table 1). The female UTX142–150 bound to HLA-A*24:02 with a half-life of 6.5 h at 37 °C, whereas the two male-female chimeric peptides interacted with half-lives of 25–27 h (Table 1).

Table 1.   Affinity and stability of peptide-HLA complexes.
PeptideAmino-acid SequenceHLA-A*24:02
Predicted binding (KD in nm)Measured stability (T1/2 in hours)
  1. Predicted affinity and measured stability of the indicated peptide-HLA-A*24:02 complexes.

  2. aMale-female chimeric peptides.

UTY137–147YFYYNAFHWAI180.3
UTY137–145YFYYNAFHW74.2
UTY138–145FYYNAFHW60.8
UTY138–146FYYNAFHWA3780.5
UTY138–147FYYNAFHWAI71.3
UTY139–146YYNAFHWA2042.3
UTY139–147YYNAFHWAI535.4
UTX142–150HYNAFQWAI366.5
 HYNAFHWAIa1824.8
 YYNAFQWAIa726.9

The YFYYNAFHW (UTY137–145), FYYNAFHWAI (UTY138–147) and YYNAFHWAI (UTY139–147), as well as the UTX142–150 peptide and the two male/female chimeric peptides, all supported peptide-HLA-A*24:02 complex formation and tetramer production. PBMCs 9 years post-HCT were in vitro stimulated with a mixture of the 11-mer UTY137–147 and all submers predicted to bind to HLA-A*24:02, and subsequently stained with each of the tetramers. The HLA-A*24:02/UTY139–147 tetramer stained 0.9% of the CD8-positive cells (Fig. 2C). Neither the HLA-A*24:02/UTY139–147 tetramer nor the HLA-A*24:02/UTY138–147 tetramer showed any staining (Fig. 2A,B). As negative controls, irrelevant HLA-A*24:02 tetramers were generated with the HLA-A*24:02-restricted CMV epitope, IE1248–256, AYAQKIFKI (note, the patient was CMV-negative before HCT) or with the female UTX142–150 peptide (or the male/female chimeric peptides described above). None of these negative control HLA-A*24:02 tetramers showed any staining of the CD8+ T cells (Fig. 2D,E, and data not shown). Thus, tetramer staining demonstrated that UTY139–147 is the optimal peptide epitope and validated HLA-A*24:02 as the restricting HLA molecule. To further demonstrate that the 9-mer UTY139–147 is the optimal epitope, in vitro stimulated T cells were analysed in a dose response of the 9-mer UTY139–147 versus the 11-mer UTY137–147. Cytokine responses could be observed when stimulating with as little as 5 nm of the 9-mer UTY139–147 whereas 500 nm of the 11-mer UTY137–147 was needed to obtain the same magnitude of response (Fig. 3).

Figure 2.

 Tetramer staining of in vitro stimulated patient UPN694 peripheral blood mononuclear cells (PBMCs) obtained 9 years post-HCT. PBMCs were in vitro stimulated for 11 days with a mixture of the 11-mer peptide UTY137–147 and its predicted strong HLA-A*24:02 binding submers (UTY137–145, UTY138–145, UTY138–146, UTY138–147, UTY139–146, UTY139–147). HLA-A*24:02 tetramers could be generated with three of these peptides (UTY137–145, UTY138–147, UTY139–147) (A–C), the X-chromosomal homologue peptide (C) and an irrelevant tetramer (D). CD8+ cells were gated and the plots show tetramer staining with the indicated tetramers.

Figure 3.

 Titration of the peptides YYNAFHWAI (UTY139–147) and YFYYNAFHWAI (UTY137–147). Peripheral blood mononuclear cells from patient UPN694 11 years post-HCT were in vitro stimulated with UTY137–147 for 11 days, UTY137–147 pulsed DCs were added and the cells were cultured for another 9 days. The cells were analysed by intracellular cytokine staining for recognition of a titration of the 11-mer (UTY137–147) and the optimal 9-mer (UTY139–147). CD8+ cells were gated and the frequency of cytokine (TNF-α and IFN-γ)-producing cells was determined as described in Fig. 1. The frequencies of cytokine-producing CD8+ T cells (background subtracted) are plotted against decreasing peptide concentrations (nm).

Development of UTY139–147-specific T cells after HCT

To verify that UTY139–147-specific donor T cells had been generated in vivo in the patient, we examined PBMCs from the donor prior to the HCT, as well as patient PBMCs obtained 47 days and 9 years post-HCT, for ex vivo staining with the HLA-A*24:02/UTY139–147 tetramer. Double tetramer staining was used to enhance identification of the tetramer-positive populations. Briefly, CD8+ cells were gated, and only CD8+ cells that stained for both HLA-A*24:02/UTY139–147 PE and HLA-A*24:02/UTY139–147 APC were considered positive. For PBMCs obtained ex vivo 47 days post-HCT, we found that 0.1% of all CD8+ cells were specific for HLA-A*24:02/UTY139–147 (Fig. 4B). In contrast, no HLA-A*24:02/UTY139–147-specific T cells could be detected ex vivo in PBMCs from the donor before HCT, nor in PBMCs from the patient 9 years post-HCT (Fig. 4A,C). Even after in vitro stimulation with the peptide UTY139–147, there were no detectable UTY139–147-specific T cells in the donor PBMCs (Fig. 4D). In contrast, UTY139–147-specific T cells were readily detectable in both the in vitro stimulated PBMCs day 47 post-HCT (10.6%; Fig. 4E) and 9 years post-HCT (0.6%; Fig. 4F). The fact that UTY139–147-specific T cells could be detected in the patient post-HCT, but not in the donor, strongly suggests that the UTY139–147-specific T cells have been primed in the patient within the first 47 days post-HCT. Most importantly, the ex vivo detection of UTY139–147-specific T cells at 47 days post-HCT excludes the possibility that the UTY139–147-specific T cells observed are the result of an in vitro priming of naïve CD8+ T cells.

Figure 4.

 HLA-A*24:02/UTY139–147-tetramer analysis of Peripheral blood mononuclear cells (PBMCs) from patient UPN694 and his donor. PBMCs from the donor (A and D), from the patient at day 47 post-HCT (B and E) and from the patient 9 years post-HCT (C and F). In the top panels (A–C), the PBMCs were analysed ex vivo. PBMCs are stained with HLA-A*24:02/UTY139–147 tetramers labelled with either PE or APC. CD8+ T cells were gated and the plots show PE versus APC labelled tetramers. In the bottom panels (D–F), the donor and patient PBMCs were in vitro stimulated with the 9-mer UTY139–147 for 11 days and then stained with HLA-A*24:02/UTY139–147 PE labelled tetramer. The CD8+ T cells were gated and the plots show HLA-A*24:02/UTY139–147 tetramer stained CD8+ T cells.

Functionality of UTY139–147-specific T cells

To determine the cytokine profile of the UTY139147-specific T cells, 9 years post-HCT PBMCs were in vitro stimulated with the peptide UTY139147. To increase the number of UTY139147-specific T cells, UTY139–147 pulsed HLA-A*24:02-positive DCs were added and the cells were cultured for another 9 days. When restimulated with the peptide and analysed by ICS, T cells produced IFN-γ, TNF-α and MIP-1β (Fig. 5).

Figure 5.

 Cytokine production of the UTY139–147-specific CD8+ T cells. Peripheral blood mononuclear cells (PBMCs) from patient UPN694 9 years post-HCT were in vitro stimulated with the peptide UTY139–147 for 11 days. UTY139–147 pulsed DCs were added and the cells were cultured for another 9 days. The cells were harvested and their capacity to produce cytokines was analysed by intracellular cytokine staining in the absence (A and C) or presence of peptide UTY139–147 (B and D). Upper panels (A and B) show TNF-α versus IFN-γ responding CD8+ T cells, bottom panels (C and D) show cytokine (TNF-α and IFN-γ) versus MIP-1β responding CD8+ T cells.

To demonstrate that UTY139147-specific T cells exerted cytotoxic activity, PBMCs from 11 years post-HCT were in vitro stimulated and analysed using DCs from a female volunteer sharing only the allele HLA-A*24:02 with the patient and DCs from a HLA-A*24:02-negative volunteer. The UTY139147-specific T cells were positive for CD107a, a cytotoxic degranulation marker, when stimulated with HLA-A*24:02-positive DCs pulsed with UTY139–147, but not when stimulated with unpulsed DCs, with DCs pulsed with the female variant UTX142–150 (Fig. 6A–C), nor with HLA-A*24:02-negative DCs pulsed with UTY139–147 (data not shown). The UTY139–147-specific T cells also produced the cytokines TNF-α and IFN-γ when stimulated with HLA-A*24:02-positive DCs loaded with UTY139–147 (data not shown). UTY139–147-specific T cells were also stimulated with HLA-A*24:02-positive male DCs. Cytotoxic degranulation and cytokine production were also in this case only observed, if the DCs were pulsed exogenous with UTY139–147 (data not shown). A leucapheresis product obtained from the patient before the HCT and containing 73% follicular lymphoma cells was available. UTY139–147-specific T cells co-cultured with these tumour cells showed neither degranulation (Fig. 6D) nor cytokine production (data not shown), unless the tumour cells were pulsed with exogenous UTY139–147 peptide (Fig. 6E). This suggests that the thawed tumour cells did not express, process and/or present the UTY139–147 epitope in a manner that was stimulatory to the in vitro stimulated T cells.

Figure 6.

 Cytotoxicity of the UTY139–147-specific CD8 T cells. Peripheral blood mononuclear cells from patient UPN694 11 years after HCT were in vitro stimulated with the peptide UTY139–147 for 11 days. UTY139–147 pulsed DCs were added and the cells were cultured for another 9 days. The cells were harvested and their cytotoxic response was determined by a 3-h stimulation with various target cells that had been pulsed (or not pulsed) with UTY139–147, and then analysed by flow cytometry for expression of the degranulation marker CD107a. Plots show gated CD8+ cells labelled with the HLA-A*24:02/UTY139–147 tetramer and stained for CD107a after co-culture with the following target cells: HLA-A*24:02-positive DCs not pulsed with peptide (A), HLA-A*24:02-positive DCs pulsed with UTY139–147 (B), HLA-A*24:02-positive DCs pulsed with UTX142–150 (C), patient lymphoma cells (73% pure) not pulsed with peptide (D) and patient lymphoma cells pulsed with UTY139–147 (E).

UTY139–147-specific T cells in other HLA-A*24:02-positive male recipients of female grafts

Post-HCT PBMCs were available from four other HLA-A*24:02-positive male recipients of female donor grafts, from two HLA-A*24:02-positive female recipients of female donor grafts, from two HLA-A*24:02-positive male recipient of male donor grafts and from one HLA-A*24:02-positive female recipient of a male donor graft. All donor/recipient pairs were either HLA-identical siblings or matched at 10 of 10 HLA alleles. Characteristics of these recipients are shown in Table 2. PBMCs obtained after HCT were in vitro stimulated with the peptide UTY139–147 and stained with the HLA-A*24:02/UTY139–147 tetramer. UTY139–147-specific T cells were detected in three of five HLA-A*24:02-positive male recipients of female donor grafts (Fig. 7A). In contrast, no UTY139–147-specific T cells could be detected post-HCT in the two female recipients of female donor grafts, in the two male recipient of male donor grafts or in the female recipient of a male donor graft (Fig. 7B).

Table 2.   Characteristics of HLA-A*24:02 donor/recipient pairs.
UPNDiseaseYears post-HCTPatient sexDonor sexUTY139–147-specific T cells
  1. CLL, Chronic lymphocytic leukaemia; AML, Acute myeloid leukaemia; FL, Follicular lymphoma; MM, Multiple myeloma; NHL, Non-Hodgkins lymphoma; MDS, Myelodysplastic syndrome.

766CLL2MFYes
804AML9MFYes
694FL9MFYes
816CLL1MFNo
733MM1MFNo
628NHL, MDS11FFNo
1104CLL4FFNo
925MDS4MMNo
1194CLL3MMNo
988AML2FMNo
Figure 7.

 Tetramer staining with HLA-A*24:02/UTY139–147 tetramer of Peripheral blood mononuclear cells (PBMCs) post-HCT from HLA-A*24:02-positive patients after in vitro stimulation with peptide UTY139–147 for 11 days. Plots show gated CD8 cells. All donor/recipient pairs were either HLA-identical siblings, or matched at 10 of 10 alleles. UPN, donor and patient sex information is indicated above the plots (F/M: female donor/male recipient, F/F: female donor/female recipient, M/M: male donor/male recipient, M/F: male donor/female recipient). Disease and years post-HCT the PBMCs were obtained as indicated in Table 2.

Discussion

We have used a reverse immunology strategy to identify novel H-Y-mHags. Using the NetMHCpan method, short peptide stretches encompassing multiple predicted candidate epitopes were designed and synthesized. PBMCs obtained post-HCT from male patients of female HCT grafts were screened for T cell responses directed against these peptides. The 11-mer UTY137–147 encoded by the UTY gene induced a strong response in one HCT recipient. The HLArestrictor was then used to predict the epitope present within the corresponding peptide and the appropriate HLA-restriction element. The corresponding tetramer was generated and used to validate the T cells as being directed against the male UTY139–147 peptide presented by HLA-A*24:02. UTY139–147-specific T cells were detected in three of five HLA-A*24:02-positive male recipients of female donor grafts. The observed frequency is similar to frequencies of other H-Y mHag epitopes found by Ofran et al. [9]. Importantly, UTY139–147-specific T cells were detected ex vivo at day 47 post-HCT. They were also detected after UTY139–147 stimulation in vitro of PBMCs obtained at day 47 post-HCT and 9 years post-HCT, but not of donor PBMCs obtained prior to the HCT. Thus, priming of the naïve CD8+ T cells must have occurred in vivo and among the donor T cells after the HCT. That our in vitro stimulation conditions are not leading to in vitro priming is also supported by our finding of UTY139–147-specific T cells solely in HLA-A*24:02-positive male recipients of female grafts.

In principle, a mHag could either have an ubiquitous or a more cell- or tissue-restricted, expression profile. HCT is routinely used against malignancies of the haematopoietic system, where the engrafted immune system replaces the host immune system. In this case, mHags expressed solely by the host haematopoietic tissue could be attacked by the donor immune system leading to GVL effects without harming the host. In contrast, donor immune responses directed at mHags, which are expressed in non-haematopoietic tissues (e.g. ubiquitously), could lead to detrimental GVHD. Thus, whether a mHag contributes to beneficial GVL effects, or a harmful GVHD, is in part determined by the expression profile of the mHag in question. The existence of two other H-Y-mHags in the UTY gene (the B60-restricted UTY82–91 peptide RESEEESVSL, and the B8-restricted UTY566–573 peptide LPHNHTDL) has previously been reported [10, 11]. Although the UTY mRNA has been found to be ubiquitously expressed [10, 11], the Human Protein Atlas (http://www.proteinatlas.org) reports UTY protein expression to be restricted to haematopoietic tissue. In agreement, Warren et al. reported that the B8-restricted UTY566–573 only shows lytic activity in vitro against certain haematopoietic cell lines (EBV-LCLs, PHA-stimulated T cell blasts and primary leukaemic blasts), and not against fibroblasts [10, 11]. The latter would suggest that the antigenically important expression profile is restricted to haematopoietic tissue.

Whether the mHags expressed by the UTY gene can be used as targets for T cell therapy is still an open question. In the case of the proband, we noted that it was not possible to demonstrate T cell reactivity against the patients follicular lymphoma cells unless we pulsed the cells with exogenous peptide. That the latter is possible demonstrates that the cells successfully express HLA-A*24:02. The former suggests that the epitope is not being generated in the follicular lymphoma cells, at least not under the in vitro culture conditions used here. This does not exclude the possibility that the follicular lymphoma cells in situ express the epitope. For one of the tetramer-positive recipients of female HCT graft that this study was subsequently extended to, UPN766, we have previously reported the induction of tumour-specific CD107a+CD8+ T cells when stimulating post-HCT CD8+ T cells with pre-HCT CLL (chronic lymphocytic leukaemia) cells [28].

If on the other hand the UTY gene is expressed ubiquitously, then the mHag identified here could contribute to GVHD. It is noteworthy that the patient developed acute intestinal GVHD overall grade III at day 48 after HCT, at the time when a high frequency of the UTY139–147-specific T cells was detected ex vivo. In the two other patients (UPN766 and UPN804), UTY139–147-specific T cells were detected in samples obtained approximately 2 and 9 years post-HCT, respectively. UPN766 developed acute GVHD overall grade II at day 31 after HCT and both developed extensive chronic GVHD. Overall, this mHag could be involved in GVHD responses as well as in GVL effects.

Owing to the enormous polymorphism of the HLA system, each beneficial mHag will only be of value for patients with the proper HLA restriction elements. Therefore, the availability of many and differently restricted mHags are important factors in translating these epitopes into novel and personalized immunotherapies. The discovery of this new mHag demonstrates the feasibility of a bioinformatics, ‘reverse immunology’-driven approach to mHag discovery. With this method, the labour-intensive and often unsuccessful process of finding the HLA restriction and the peptide epitope for an isolated T cell clone in the entire human proteome can be avoided. The mode of detection could lead to improvements in the future identification of H-Y mHags, and eventually also of autosomally expressed mHags. For example, known haematopoietically expressed proteins could be selected and candidate peptides could be found based on prediction of strong binding peptides in association with non-synonymous nucleotide polymorphisms. This could lead to identification of mHags relevant as targets for T cell therapy.

In conclusion, we have identified a novel mHag, UTY139–147 (YYNAFHWAI) presented by HLA-A*24:02, which is frequently recognized in HLA-A*24:02-positive male recipients of female HCT grafts. This suggests that HLA-A*24:02/UTY139–147 represents an immunodominant H-Y mHag epitope.

Acknowledgment

We thank the medical, nursing and laboratory staff of the participating departments for their contribution to this study. We also thank Mrs Anne Bjørlig for excellent technical assistance. This study was supported with grants from the Danish Cancer Society, the Lundbeck Foundation and the Danish Research Council for Technology and Production Sciences.

Ancillary