Dr Y Akatsuka, Division of Immunology, Aichi Cancer Centre Research Institute, 1-1 Kanokoden, Chikusa-ku, Nagoya 464-8681, Japan. E-mail: firstname.lastname@example.org
Minor histocompatibility antigens (mHags) play crucial roles in the induction of graft versus host disease (GVHD) and/or graft versus leukaemia (GVL) effects following human leucocyte antigen (HLA)-identical haematopoietic stem cell transplantation (HSCT). Using HLA-A*3101- and -A*3303-restricted cytotoxic T lymphocyte (CTL) clones generated from different post-HSCT recipients, we identified two novel mHag epitopes encoded by the leader sequence of cathepsin H (CTSH) isoform a. The nonameric sequence ATLPLLCAR was defined as an HLA-A*3101-restricted epitope (CTSHR/A31), while a decameric peptide featuring a one N-terminal amino acid extension, WATLPLLCAR, was presented by HLA-A*3303 (CTSHR/A33). The immunogenicity of both epitopes was totally dependent on the polymorphic C-terminal arginine residue and substitution with glycine completely abolished binding to the corresponding HLA molecules. Thus, the immunogenicity of this mHag is exerted by differential HLA binding capacity. CTSH is relatively ubiquitously expressed at protein levels, thus it may be involved in GVHD and anti-leukaemic/tumour responses. Interestingly, however, CTL clones predominantly lysed targets of haematopoietic cell origin, which could not be explained in terms of the immunoproteasome system. Although the mechanisms involved in the differential susceptibility remain to be determined, these data suggest that CTSH-encoded mHags could be targets for GVL effects.
Minor histocompatibility antigens (mHags) are major histocompatibility complex (MHC)-bound peptides derived from cellular proteins that are encoded by polymorphic genes, mostly due to single nucleotide polymorphism (SNP) (Goulmy et al, 1996; Simpson & Roopenian, 1997). Disparities in some mHags with allogeneic haematopoietic stem cell transplantation (HSCT) have been shown to be associated with graft versus host disease (GVHD) (Goulmy et al, 1996; Tseng et al, 1999; Akatsuka et al, 2003a), graft rejection (Marijt et al, 1995) and/or graft versus leukaemia (GVL) effects (Bonnet et al, 1999; Marijt et al, 2003). Although methods for identification of MHC-bound antigens have improved, only a limited number of mHags have been reported to date (Bleakley & Riddell, 2004; Spierings et al, 2004). Thus, identification of novel mHags should facilitate further understanding of the mechanisms involved in generating mHags and also of the pathogenesis of GVHD, which may lead to development of effective immunotherapy.
This study identified the mHag epitopes recognised by two HLA-A*3101 and -A*3303-restricted cytotoxic T lymphocyte (CTL) clones, which were encoded by the same non-synonymous SNP in the cathepsin H (CTSH) gene. The polymorphic amino acid (aa) responsible for antigenicity was an arginine (Arg) in place of glycine (Gly) at the C terminus of the identified epitopes. The latter substitution resulted in more profound loss of peptide-HLA binding than previously demonstrated with HA-1/A2 (den Haan et al, 1998) and HA-2 (Pierce et al, 2001). CTSH has been shown to be expressed relatively ubiquitously, but unexpectedly, CTL clones specific for CTSH-encoded mHag lysed haematopoietic cells, including primary acute myeloid leukaemia (AML) cells obtained from patients, but not some cytokine-treated non-haematopoietic cells, even though high levels of CTSH protein expression and immunoproteasome components comparable with those of haematopoietic cells were found. The selective susceptibility of haematopoietic cells suggests that CTSH should be further evaluated as a potential target to augment GVL effects.
Materials and methods
Patients, cell collection and cultures
Cytotoxic T lymphocyte clones, clinical characteristics of the two patients from whom the clones were obtained with unique patient numbers (UPNs) 027 and 028, are summarised in Table I. This study was approved by the Institutional Review Board of the Aichi Cancer Centre according to the Declaration of Helsinki. All blood or tissue samples were collected after written informed consent was obtained. Epstein–Barr virus-transformed B lymphocyte cell lines (B-LCL) were derived from donors, recipients, and normal volunteers and maintained in RPMI 1640 medium (Sigma-Aldrich, St Louis, MO, USA) supplemented with 10% fetal calf serum (IBL, Takasaki, Japan). Individual patients’ primary fibroblast lines established pre-HSCT from skin and bone marrow, normal renal epithelial cells and renal cell carcinoma cell lines established from restriction HLA-matched unrelated individuals and 293T cells were grown in Iscove's-modified Dulbecco's medium (Sigma-Aldrich) supplemented with 10% fetal calf serum. B-LCLs from Centre d'Etude du Polymorphism Humain (CEPH) families (provided by Dr P. Martin, Human Immunogenetics Program, Fred Hutchinson Cancer Research Center, Seattle, WA, USA) were transduced with HLA-A*3101 or -A*3303 cDNA-encoding retroviral vectors as described previously (Akatsuka et al, 2003b).
Table I. Characteristics of patients and cytotoxic T lymphocyte clones.
Generation and cytotoxicity analysis of CTL lines and clones
Cytotoxic T lymphocyte lines were generated from post-HSCT peripheral blood mononuclear cells (PBMC) (c. 1 × 106) by primary stimulation with irradiated (33 Gy) pre-HSCT recipient PBMC (c. 1 × 106) followed by weekly restimulation with irradiated (33 Gy) recipient activated B cells (2 × 106, see below) twice. Interleukin (IL)-2 (20 U/ml; Chiron, Emeryville, CA, USA) was added on days 1 and 5 after the second and third stimulation. CTL clones were generated by limiting dilution as previously reported (Akatsuka et al, 2002). The CTLs were expanded as previously described (Walter et al, 1995) and frozen until use. All cultures were performed in RPMI 1640 medium supplemented with 9% pooled human serum. The cytotoxic activity of the CTL lines and clones was evaluated by standard 51Cr release assays. Some target cells were treated with cytokines as indicated for 48 h. Percentage-specific lysis was calculated as follows: ((experimental cpm − spontaneous cpm)/(maximum cpm − spontaneous cpm)) × 100.
For antibody blocking experiments, cells were incubated for 30 min with predetermined concentrations of monoclonal antibodies W6/32 (anti-HLA class I) or HDR-1 (anti-HLA-DR; provided by K. Ito, Kurume University, Fukuoka, Japan) before mixing with CTL clones. The antigen specificity of tumour cell lysis was further determined in a cold target inhibition assay by adding unlabelled cold target cells (indicated) to labelled target cells (hot targets) at serial ratios of cold-to-hot target cells. The percentage of inhibition was calculated as follows: ((% specific lysis without cold target − % specific lysis with cold target)/(% specific lysis without cold target)) × 100.
Construction of a cDNA library and expression screening
A cDNA library was constructed using the SuperScript Plasmid System (Invitrogen, Carlsbad, CA, USA). A sample (2 μg) of messenger RNA isolated from B-LCL recognised by CTL clones with a FastTrack 2·0 kit (Invitrogen) was converted into cDNA using an oligo-dT primer containing a NotI site in its 3′-end and SuperScript II reverse transcriptase (Invitrogen). cDNA was ligated to SalI adaptors, and then digested with NotI, size fractionated by column chromatography and ligated into the SalI and NotI cut pCMVSPORT6 vector (Invitrogen). Recombinant plasmids were transformed into Escherichia coli DH10B by electroporation, and clones were selected with ampicillin. The library contained 1·5 × 106 cDNA clones with an average insert size of approximately 2500 bp. cDNA pools, each consisting of c. 120 clones were expanded for 24 h in 96-deep well plates, and plasmid DNA was extracted with a QIAprep 96 Turbo Miniprep kit (Qiagen, Valencia, CA, USA).
The 293T cells (n = 20 000), stably transduced with HLA-A*3303 or -A*3101 were plated in 96-well flat-bottomed plates, cultured overnight at 37°C, then transfected with 0·12 μg of plasmid containing a pool of the cDNA library using Trans IT-293 (Mirus Technologies, Madison, WI, USA). CTL clones (10 000 cells/well) were added to each well 20 h after transfection. After overnight incubation in the presence of 10 U/ml IL-2 at 37°C, 50 μl of supernatant was collected and interferon (IFN)-γ was measured by enzyme-linked immunosorbent assay (ELISA) with 3,3′,5,5′-tetramethylbenzidine substrate (Sigma-Aldrich).
Genotyping of CTSH polymorphisms
Genomic DNA was isolated from each B-LCL with a DNA blood kit (Qiagen) and amplified by polymerase chain reaction (PCR). The primer sequences used to amplify CTSH exon 1 were as follows: sense, 5′-GAACTAGAGCTGGGGAGTTA-3′ antisense, 5′-CCCGCCTATAATGCAGTTTA-3′. PCR products were purified and directly sequenced with the same primer and a BigDye Terminator kit (ver. 3·1; Applied Biosystems, Foster City, CA, USA) using an ABI PRISM 3100 (Applied Biosystems).
HLA peptide-binding assay
We used a quantitative ELISA-based assay capable of measuring the affinity of the interaction between peptide and HLA as described previously, with some modifications (Sylvester-Hvid et al, 2002). In brief, purified recombinant HLA molecules in 8 mol/l of urea, 10 mmol/l of EDTA, 25 mmol/l of 2-(N-morpholino)ethanesulfonic acid (MES) and 0·1 mmol/l of dithiothreitol were diluted to 4 μg/ml in refolding buffer containing 400 mmol/l of Arg, 100 mmol/l Tris pH 8·0, 2 mmol/l of EDTA, 5 mmol/l of reduced glutathione, 0·5 mmol/l of oxidised glutathione, 0·2 mmol/l of phenyl methyl sulphonyl fluoride (all from Sigma-Aldrich), and 2 μmol/l purified β2-microglobulin (β2m) on ice. Ten-fold dilutions of each peptide were made with 100% dimethyl sulphoxide in 96-well round-bottomed polypropylene plates, then 1 μl of individual aliquots were transferred into new plates and 99 μl of the above HLA-β2m mixture was added to each well (i.e. 100-fold dilution for each peptide solution). The plates were incubated on a shaker at 4°C for 48–72 h. One day before ELISA analysis, 96-well ELISA plates (Costar, Cambridge, MA, USA) were coated with 50 μl/well W6/32 monoclonal antibody (mAb) (10 μg/ml) in 50 mmol/l of carbonate–bicarbonate buffer, pH 9·6 (Sigma, St. Louis, MO, USA) and kept overnight at 4°C. After washing three times with washing buffer containing 0·05% Tween-20 (Sigma) in phosphate-buffered saline (PBS), the wells were blocked for 1 h. Just prior to the ELISA analysis, the reaction volume was diluted 10 times by PBS at 4°C, and 50 μl/well of aliquots were transferred in duplicate to the W6/32 mAb-coated plates. The plates were incubated for 2 h at room temperature, and then washed six times. To detect properly refolded complexes, incubation was for 2 h at room temperature with 100 μl/well of horseradish peroxidase (HRP)-conjugated anti-human β2m mAb (1:1000 dilution; Dako, Copenhagen, Denmark), followed by washing as above. Finally, colour development was performed by ELISA as above.
Reverse transcription PCR analysis of components for antigen processing
A reverse transcription PCR assay was used to examine the expression of LMP2, LMP7, PA28α, TAP1, TAP2 and β-actin (as an internal control) mRNA. Total RNA was extracted using an RNeasy Mini Kit (Qiagen), and cDNA was synthesised by standard methods. Specific primers used were prepared as reported previously (Ito et al, 2006). The PCR products were separated in 2·5% agarose and visualised with ethidium bromide staining.
Cells were fixed and permeabilised with Cytofix/Cytoperm (BD Biosciences, San Diego, CA, USA), incubated for 60 min with goat anti-cathepsin H monoclonal and mouse anti-HLA-DM antibodies (both from Santa Cruz Biotechnology, Santa Cruz, CA, USA). Then cells were reacted with appropriate fluorescence-labelled secondary antibodies for 60 min. Finally, stained cells were washed and cytocentrifuged, and analysed by laser scanning confocal microscopy using Radiance 2100 K-3 (Bio-Rad, Hercules, CA, USA).
Tetramer construction and flow cytometric analysis
HLA-A*3101 or -A*3303 tetramers incorporating the identified peptides were produced as described previously (Altman et al, 1996). For staining, PBMC or T-cell lines were incubated with the tetramer at a concentration of 20 μg/ml at room temperature for 15 min followed by fluorescein isothiocyanate-conjugated anti-CD3 (Becton-Dickinson, San Diego, CA, USA) and Tricolor anti-CD8 mAb (Caltag, Burlingame, CA, USA) on ice for 15 min. Cells were analysed with a FACSCalibur flow cytometer and CellQuest software (Becton-Dickinson).
Characterisation of CTL clones
Cytotoxicity assays using a panel of B-LCLs showed that the CD8+ CTL clone 2A10, derived from UPN 027, was HLA-A*3303-restricted, while 1A8, from UPN 028, was HLA-A*3101 restricted (data not shown, and Table 1). Each lysed its respective recipient-derived B-LCLs and phytohaemagglutinin (PHA)-stimulated T-cell blasts, but not donor-derived B-LCLs or natural killer-sensitive K562 cells (Fig 1A and B). Specific lysis of respective recipient B-LCL was inhibited by anti-HLA class I mAb, but not by mAb against anti-HLA-DR, indicating recognition of HLA class I-restricted mHags (Fig 2C and D). No cytotoxicity was observed against the recipient's dermal- and bone marrow-derived fibroblasts even after treatment with IFN-γ and tumour necrosis factor-α (Fig 1E and F). UPN 028 was also HLA-A*3303 positive and his B-LCL were found to be lysed by 2A10 as well. Primary leukaemic blasts from UPN 028 were tested for recognition by the CTL clones. Both 2A10 and 1A8 lysed the leukaemic blasts (Fig 1E and F), indicating that the mHag was expressed on myeloid leukaemic cells. The U937 cell line (Sundstrom & Nilsson, 1976) (subsequently confirmed to be mHag-positive by genotyping) was lysed by both CTL clones only when the respective HLA allele was transduced, confirming HLA-restricted recognition by these CTL clones (Fig 1E and F). Finally, mHag-specific recognition of the U937 cell line was confirmed by cold target inhibition assays. Killing of HLA-transduced U937 was inhibited by the addition of respective recipient B-LCLs or donor LCLs pulsed with the cognate mHag identified in the subsequent study (see below) (Fig 1G and H), while addition of unpulsed donor B-LCLs showed no inhibition. These data suggest that the mHags recognised by 2A10 and 1A8 clones were presented by the U937 cell line. The preferential lysis of haematopoietic cells by these CTLs prompted us to identify the gene(s) encoding the mHags as potential therapeutic targets.
Identification of the gene encoding the mHags as cathepsin H
Two-point linkage analysis was performed to identify the gene encoding the mHags recognised by 2A10 and 1A8, as reported previously (Akatsuka et al, 2003b). Surprisingly, the two CTL clones showed identical lytic patterns against CEPH B-LCLs in five families tested (Fig 2A and B, and data not shown), suggesting that the polymorphic genes controlling the expression of the two mHags were located on a narrow chromosomal region or, alternatively, that a single gene encoded both mHags, as seen with the BCL2A1 mHags (Akatsuka et al, 2003b). However, because the mapped region contains many genes whose characteristics have not yet been fully elucidated, we could not identify candidate gene(s) by this in silico approach.
A cDNA library was generated from UPN 027 B-LCL and used for expression cloning. In the initial screening, one of 96 plasmid pools induced IFN-γ production by 2A10, which was then subcloned into individual cDNA clones and rescreened (Fig 3A). A single cDNA, referred to as B9, was identified that stimulated 2A10 CTL to produce IFN-γ, and recognition of 293T cells required transfection of both HLA-A*3303 and B9 cDNA (Fig 3B, left panel). As the preliminary linkage analysis results showed that the genetic region controlling the mHag recognised by 1A8 was found to be in the same location as that for 2A10 (4·1-Mbp distance around chromosome 15q24–q25, data not shown), we examined whether CTL clone 1A8 might also recognise the B9 cDNA product. Indeed, CTL 1A8 recognised 293T cells expressing B9 cDNA and HLA-A*3101, suggesting that 2A10 and 1A8 CTL clones recognised two different epitopes of the same gene product (Fig 3B, right panel). Subsequently, the cDNA insert of B9 was sequenced and a search of the BLAST database (http://www.ncbi.nlm.nih.gov/BLAST/Blast.cgi) revealed that this cDNA was nearly identical to transcriptional variant 1 encoding isoform a of CTSH (GenBank accession no. NM_004390) (Fig 3C), which existed within the region identified by linkage analysis (data not shown).
Identification of HLA-A*3303 and -A*3101-restricted epitopes on CTSH
We next sought to identify the polymorphism(s) in CTSH responsible for antigenicity. Splice variant 1 of CTSH mRNA contains three reported non-synonymous coding SNPs (Fig 3C). It was found that the presence of the adenine (translation, Arg) but not guanine (translation, Gly) at nucleotide (nt) position 126 (aa position 11) was associated with recognition by both 2A10 and 1A8 CTL clones. The observed frequency of the adenine-containing allele among 21 Japanese individuals was 0·11.
We then searched for candidate epitope sequences spanning this SNP using on-line algorithms, such as BIMAS (http://bimas.dcrt.nih.gov/molbio/hla_bind) (Parker et al, 1994), and identified nonameric ATLPLLCAR or decameric WATLPLLCAR as candidate epitopes. Both HLA-A*3303 and -A*3101 have Arg or Lys as preferred C-terminal anchor residues, whereas Gly is not a preferred residue (Falk et al, 1994; Parker et al, 1994; Takiguchi et al, 2000). Minigene experiments demonstrated that CTL clone 2A10 recognised 293T/A*3303 expressing WATLPLLCAR but not that expressing nonameric peptide, while 1A8 recognised 293T/A*3101 expressing both ATLPLLCAR and WATLPLLCAR (Fig 3D). Neither CTL clone recognised 293T cells transfected with minigenes encoding the Gly allele.
To determine whether the identified epitopes could sensitise donor HLA-A*3303 or -A*3101-positive target cells to lysis by 2A10 and 1A8, epitope reconstitution assays were performed. As predicted by minigene experiments, 2A10 recognised only the decameric WATLPLLCAR with half-maximal lysis at 2 nmol/l (Fig 4A), although peptide-HLA-binding assays suggested that both nonameric and decameric peptides could bind to HLA-A*3303 equally (Fig 4B). Thus, peptide WATLPLLCAR represents the HLA-A*3303-restricted mHag epitope recognised by 2A10 (designated as CTSHR/A33). 1A8 recognised the nonameric peptide ATLPLLCAR in the context of HLA-A*3101, with half-maximal lysis at 20 nmol/l, and also the decameric peptide WATLPLLCAR, but at a 100-fold higher concentration (Fig 4C). However, in this case, the HLA-binding assay demonstrated that almost only the nonameric peptide could bind to HLA-A*3101 (Fig 4D). These data demonstrate that the nonameric peptide ATLPLLCAR represents the HLA-A*3101-restricted 1A8 epitope (designated as CTSHR/A31).
Finally, A31/CTSHR or A33/CTSHR tetramers, which were prepared in our laboratory, was able to detect growing populations of CD8+ T cells in T-cell lines generated from UPN 028s post-HSCT PBMC by two different stimulators, recipient pre-HSCT PBMC and mHag peptide-pulsed donor PBMC, indicating that CTSHR mHags were indeed immunogenic and sensitised precursor T cells were generated in the recipients. However, tetramer-positive cells were not detectable as a distinguishable cluster in the unstimulated CD8+ fraction of any of the post-HSCT samples tested (data not shown), implying that the precursor frequency CTSHR-specific T cells in unstimulated PBMCs was <10−3 (data not shown).
Lysosomal localisation of CTSH does not correlate with susceptibility to lysis
Because mHags CTSHR/A31 and CTSHR/A33 are generated only from the isoform a of CTSH that is anticipated to localise to lysosomes, confocal microscopy was used as this approach would make it possible to address the question of whether preferential accumulation of CTSH isoform a in lysosomes correlated with the differential susceptibility to CTL of individual target cells. Haematopoietic cells, including not only monocytes, PHA-blasts, B-LCLs but also most primary leukaemia cells were positive for CTSH, but co-localisation of CTSH with lysosome-resident HLA-DM was not evident except for normal monocytes and AML FAB M4 and M5 subtypes; four of four M4 and three of three M5 primary leukaemic cells were strongly positive, but B-ALL cells from one patient were only weakly stained (Fig 5 and data not shown). Contrary to our expectations, dermal fibroblasts, normal renal epithelial cells, renal carcinoma cell lines and keratinocytes were diffusely stained (Fig 5 and data not shown), suggesting that not only CTSH variants 1 encoding the isoform a, but also variant 2 are expressed in most kinds of cells.
No apparent correlation of preferential CTL cytotoxicity against haematopoietic cells in proportion to CTSH protein expression
Cytokine-pretreated dermal and bone marrow fibroblasts, which expressed relatively low levels of CTSH protein, were not lysed by the CTLs (Figs 1A,B and 5). We then investigated whether or not dermal fibroblasts from UPN 028, unrelated renal normal epithelial cells and carcinoma cell lines (possessing HLA-A*3101 and CTSHR allele confirmed by genotyping) expressing high levels CTSH protein (Fig 5) were susceptible to lysis. Surprisingly, these non-haematopoietic cell lines were not lysed by the CTLs even after cytokine pretreatment (Fig 6A). These cells were, however, not resistant to cytolysis as they became susceptible when cognate CTSHR/A31peptide was pulsed exogenously (Fig 6B).
Activated haematopoietic cells are known to express immunoproteasomes containing LMP2, LMP7 and/or PA28α and to alter proteasome cleavage specificity (Kloetzel, 2001), which thus may account for preferential generation and presentation of the CTSH mHags. However, except for keratinocytes, cytokine-pretreated non-haematopoietic cells expressed these immunoproteasome components at levels comparable with those of haematopoietic cells tested (Fig 6C).
Minor histocompatibility antigens are known to play an important role as alloantigens inducing GVHD and/or GVL effects after HLA-identical allogeneic HSCT. Thus, it is still necessary to identify novel mHags and to study the potential clinical significance and mechanisms involved in the generation of antigenicity. This study demonstrated that splice variant 1 of CTSH mRNA transcribed from the CTSH gene, located on chromosome 15q24–q25, encodes two novel mHag epitopes restricted by two different HLA-A alleles but belonging to the same HLA-A3 supertype. Nonameric peptide ATLPLLCAR (CTSHR/A31) was presented by HLA-A*3101, while decameric peptide with one N-terminal aa extension, WATLPLLCAR (CTSHR/A33), was presented by HLA-A*3303. The non-synonymous coding SNP located on CTSH exon 1 composed the C-terminal anchor motif of both epitopes. The epitope reconstitution assay and peptide–HLA-binding assay indicated that peptides with Gly instead of Arg at the C terminus, which is the sole anchor position for both HLA-A*3101 and -A*3303 (Falk et al, 1994), were unable to bind to these HLA molecules at all, suggesting that lack of HLA-binding peptide in donors homozygous for CTSHG may be attributable to allo-responses against recipients carrying CTSHR allele following allo-HSCT.
To date, various mechanisms involved in mHag generation have been reported as follows: peptide binding to MHC observed in HA-1/A2 (den Haan et al, 1998) and HA-2 (Pierce et al, 2001); proteosomal cleavage in HA-3 (Spierings et al, 2003); peptide transport in HA-8 (Brickner et al, 2001); recognition of MHC-peptide complex by cognate T cells in SMCY/B7 (Wang et al, 1995), DFFRY/A1 (Pierce et al, 1999), HB-1 (Dolstra et al, 2002) and HA-1/B60 (Mommaas et al, 2002); and differential protein expression in UGT2B17 (Murata et al, 2003), PANE1 (Brickner et al, 2006) and LRH-1 (Rijke et al, 2005). Among those, our mHags were considered to be generated by differential peptide binding to MHC, and the substitution of Gly for Arg at the C-terminus, resulting in the complete loss of binding, was sufficient to account for the difference in recognition of CTSH/A31+ or CTSH/A33+ cells from their negative counterparts. The differential binding mechanism is similar to the cases of HA-1/A2 and HA-2, but seems to be more significant because there are only 12- to 15-fold differences in peptide binding in HA-1/A2 and HA-2, respectively, although binding affinity to HLA-A2 was evaluated by competition-based assays (den Haan et al, 1998).
Initially, to identify the gene encoding the mHag recognised by CTL 2A10, two-point linkage analysis (Akatsuka et al, 2003b; Rijke et al, 2005) was conducted. Coincidentally, not only 2A10 but also 1A8 CTL clones showed an identical lytic pattern, suggesting that the polymorphic genes controlling the expression of the two mHags were located on a narrow chromosomal region or, alternatively, that a single gene encoded both mHags, as seen with the BCL2A1 mHags (Akatsuka et al, 2003a,b). However, the mapped region was found to contain many genes whose characteristics had not yet been fully elucidated, so we did not further pursue candidate gene(s) by this in silico approach. Rijke et al (2005) successfully applied quantitative PCR to identify the P2X5 gene by comparing the expression pattern of candidate genes in three cell types (B-LCL, monocytes and fibroblasts) with the lytic pattern of these cells by their CTL. Considering the fact that we were indeed very close to the target gene, CTSH, by linkage analysis, extensive quantitative PCR analyses tailored for the candidate genes may have identified of CTSH as the gene encoding mHag recognised by our CTL clones.
The primary AML cells of UPN 028 were recognised by the CTL clones. Thus we speculate that CTSH may function as a target of a GVL response. Indeed, the two patients with high-risk AML from whom CTSH-specific CTLs were isolated continue to be disease-free after more than two and a half years. However, one concern is that CTSH is expressed not only in haematopoietic cells (Greiner et al, 2003) but also some epithelial cells including type II pneumocytes (Brasch et al, 2002), suggesting that it may also be a target for GVHD and targeting this molecule with CTLs might carry the risk of toxicity. For example, mild bronchiolitis obliterans organising pneumonia observed in patient UPN 027 might be associated with chronic GVHD (Afessa et al, 2001), and it is conceivable that immune responses to CTSH expressed in type II pneumocytes could be a factor. Unfortunately, we could not exclude this possibility because bronchoalveolar lavage, which might contain T cells that could be tested with CTSHR tetramers, was not performed.
It was unexpected, but quite encouraging, that our CTSHR-specific CTLs demonstrated lytic activity against haematopoietic cells including leukaemic cells but not fibroblasts or renal normal epithelial and carcinoma cells, although the latter non-haematopoietic cells did express CTSHR proteins. It has been shown that the splicing variant 1 of CTSH encoding the mHag epitopes results in a longer protein (isoform a) which is likely to be localised to lysosomes, and variant 2 produces a shorter protein (isoform b) which is more likely to be a secreted protein (Waghray et al, 2002). Thus, to find differences between CTL-sensitive and -resistant target cells, we studied the expression of CTSH mRNA, intracellular distribution of CTSH protein, and also the expression of genes involved in antigen processing and presentation. However, no clear evidence to explain the differential recognition of target cells with high CTSH expression by the CTLs was obtained. In line with this, it is of note that an HLA-B8-restricted line specific for the mHag encoded by UTY was reported to lyse only haematopoietic cells, irrespective of the relatively ubiquitous expression of the gene in male cells (Warren et al, 2000). Because our mHags were derived from the leader sequence of CTSH isoform a precursor, more active translation in haematopoietic cells which accompany the production of excised leader peptide, compared with stable mature CTSH detectable by the antibody used in this study, may account for the differential susceptibility to the CTLs. In any event, determination of mechanisms involved in differential antigen expression between haematopoietic and non-haematopoietic cells is a high priority in future studies.
Finally, the restriction molecules of our CTLs, HLA-A*3303 and -A*3101, are known to belong to the HLA-A3 supertype (Falk et al, 1994; Sette & Sidney, 1999), which is found in 46% of Japanese and 38% of Caucasians (Sette & Sidney, 1999). Accordingly, the A3 supertype now includes A*0301, A*1101, as well as A*6801. HLA alleles in the A3 supertype use a similar anchor motif of Arg or Lys at their C-termini. Thus, it may be worthwhile to examine whether nonameric CTSHR/A31 or decameric CTSHR/A33 peptides would also bind to HLA molecules belonging to A*0301 and A*6801, because these HLA alleles are common in Caucasians and may create an opportunity to evaluate the significance of CTSHR-encoded mHags. The results obtained in this study suggest that CTSH may be used as a target for GVL responses after allogeneic HSCT if potential GVHD induction can be manipulated by suitable techniques, such as suicide gene introduction into CTLs. Animal models to determine the susceptibility of CTSH-positive leukaemic progenitor cells and other tissue cells to CTSHR-specific CTLs in vivo are now being developed using immune-deficient mouse systems.
The authors thank Drs S. Riddell, P. Martin (Fred Hutchinson Cancer Research Center) and W. Ho for critically reading the manuscript, and Yoshinori Ito, Ayako Demachi-Okamura, Michiyo Nakayama, Yasue Matsudaira, Keiko Nishida, and Hiromi Tamaki for their expert technical assistance. This study was supported in part by grants-in-aid for Scientific Research (C) (no. 17591025) and Scientific Research on Priority Areas (B01) (no. 17016089), from the Ministry of Education, Culture, Science, Sports, and Technology, Japan; Research on Human Genome, Tissue Engineering Food Biotechnology (to Y.A.) and the Second and Third Team Comprehensive 10-year Strategy for Cancer Control (no. 30), from the Ministry of Health, Labour, and Welfare, Japan; and grants from Daiko Foundation (to Y.A.) and Japan Leukaemia Research Fund (to Y.A.).