|Home | About | Journals | Submit | Contact Us | Français|
Minor histocompatibility antigens (mHAgs) constitute the targets of the graft-versus-leukemia response after HLA-identical allogeneic stem cell transplantation. Here, we have used genetic linkage analysis to identify a novel mHAg, designated lymphoid-restricted histocompatibility antigen–1 (LRH-1), which is encoded by the P2X5 gene and elicited an allogeneic CTL response in a patient with chronic myeloid leukemia after donor lymphocyte infusion. We demonstrate that immunogenicity for LRH-1 is due to differential protein expression in recipient and donor cells as a consequence of a homozygous frameshift polymorphism in the donor. Tetramer analysis showed that emergence of LRH-1–specific CD8+ cytotoxic T cells in peripheral blood and bone marrow correlated with complete remission of chronic myeloid leukemia. Furthermore, the restricted expression of LRH-1 in hematopoietic cells including leukemic CD34+ progenitor cells provides evidence of a role for LRH-1–specific CD8+ cytotoxic T cells in selective graft-versus-leukemia reactivity in the absence of severe graft-versus-host disease. These findings illustrate that the P2X5-encoded mHAg LRH-1 could be an attractive target for specific immunotherapy to treat hematological malignancies recurring after allogeneic stem cell transplantation.
HLA-identical allogeneic hematopoietic stem cell transplantation (SCT) is a curative therapy for various forms of hematological malignancies (1). It has become the treatment of choice for patients with chronic myeloid leukemia (CML) and is currently the most effective consolidation treatment for patients with acute leukemia (2, 3). The therapeutic efficacy is attributed to the graft-versus-leukemia (GVL) reaction, an immune response in which donor-derived CTLs eliminate malignant cells of the recipient (4, 5). Minor histocompatibility antigens (mHAgs) are the target antigens of the GVL response, and expansion of mHAg-specific CTLs has been shown to precede clinical remission in relapsed CML patients treated with donor lymphocyte infusions (DLIs) (6, 7). However, alloreactive CTL responses induced upon DLI generally lack tumor specificity and are often accompanied by graft-versus-host disease (GVHD). Selective GVL responses in the absence of GVHD may result from mHAgs with expression limited to hematopoietic cells such as HA-1 (8), HA-2 (9, 10), and BCL2A1 (11) or to lineage-specific hematopoietic cells such as HB-1 (12, 13). These hematopoietic cell–restricted antigens represent potential targets for specific immunotherapy after allogeneic SCT (14–16).
mHAgs are HLA-bound peptides derived from cellular proteins that differ in amino acid sequence between donor and recipient due to genetic polymorphisms (4, 17–19). Thus far, approximately 15 human mHAgs have been identified through mass spectrometry of HLA-eluted peptides (8, 9, 20–24), cDNA library cloning (13, 25), or genetic linkage analysis (11). Immunogenicity of most of the molecularly characterized human mHAgs results from single nucleotide polymorphisms (SNPs) in homologous donor and recipient genes causing amino acid substitutions within the epitope (8–11, 13, 20–24). In 1 instance, an mHAg encoded by the UGT2B17 gene resulted from differential protein expression in donor and recipient cells as a consequence of a homozygous gene deletion in the donor (25).
Here, we report the identification of a novel hematopoietic cell–restricted mHAg, designated lymphoid-restricted histocompatibility antigen–1 (LRH-1), which elicited an HLA-B7–restricted CTL response in a CML patient treated with DLI. Using genetic linkage analysis and gene profiling, we have identified P2X5 as the polymorphic gene encoding the LRH-1 antigenic peptide. Our data provide what we believe to be the first evidence that mHAg disparity can result from differential protein expression due to a homozygous frameshift polymorphism. Furthermore, we have analyzed the kinetics of the LRH-1–specific CTL response in the CML patient and show that emergence of LRH-1/B7 tetramer+ T cells is associated with GVL reactivity following DLI. Finally, our observation that LRH-1 is selectively expressed within the hematopoietic system including leukemic CD34+ progenitor cells offers a novel target antigen for the development of immunotherapy for the treatment of relapsed leukemia.
In 1996, we transplanted a marrow graft into a 35-year-old woman (UPN389) with Philadelphia chromosome–positive (Ph+) CML from her HLA-identical brother (A24, A29, B7, B44, Cw7, Cw16) after receiving informed consent. Partial T cell depletion was performed using counterflow centrifugation, and 0.8 × 106 CD3+ cells/kg body weight were given within the graft (26). The patient suffered from acute GVHD grade III (skin and liver), which was successfully treated with the addition of corticosteroids. At 6 months after SCT, FISH analysis of a BM aspirate showed 7.5% Ph+ cells. Cyclosporin A (CsA) and corticosteroids were gradually withdrawn within 2 months without reactivation of GVHD. However, 1 month after cessation of immunosuppressive treatment, Ph+ cells increased to 85% in the BM. Furthermore, these cells had additional chromosomal abnormalities with del(13) and t(4;11), and the patient had progressed to accelerated phase. The patient was given lymphocytes from the original BM donor (DLI-1; 0.1 × 108 CD3+ cells/kg body weight) but did not respond. After a second and higher dose of DLI (DLI-2; 0.7 × 108 CD3+ cells/kg body weight), the patient developed GVHD of the skin and pancytopenia. For GVHD of the skin, she was treated with CsA and mycophenolate mofetil for different intervals. Six months after DLI-2, BM cells were of donor origin. Since the patient had become and remained pancytopenic, she was given a T cell–depleted marrow booster from her donor. Nine months after DLI-2 and 2 months after the stem cell booster, the patient was in complete cytogenetic and molecular remission.
To isolate mHAg-specific CD8+ T cells involved in the GVL response in CML patient UPN389, we stimulated PBMCs obtained 3 months after DLI-2 with CML cells obtained at hematological relapse. After multiple stimulations, TCR repertoire analysis revealed that more than 90% of CD8+ T cells expressed a clonal TCR-BV21.3 chain (data not shown). This CD8+ CTL clone, termed RP1, mediated specific lysis against the EBV–lymphoblastoid cell line (EBV-LCL) of the recipient, but not against donor EBV-LCL and NK-sensitive K562 cells (Figure (Figure1A).1A). Specific lysis of recipient EBV-LCL was substantially inhibited by anti–HLA class I and anti–HLA-B/C Abs, but not by Ab against anti–HLA class II, indicating recognition of an HLA class I–restricted mHAg (data not shown). Testing of EBV-LCLs from unrelated individuals sharing expression of HLA-B7 with the recipient and EBV-LCL from an HLA class I–mismatched individual that were transduced with HLA-B*0702 revealed that CTL RP1 recognizes an mHAg presented by HLA-B7 (Figure (Figure11B).
Next we investigated whether the mHAg recognized by CTL RP1 was selectively expressed by hematopoietic cells. CTL RP1 significantly lysed cells of lymphoid origin such as EBV-LCLs, CD40-activated B cells, and phytohemagglutinin-stimulated (PHA-stimulated) T cells, whereas monocytes and monocyte-derived DCs were not recognized (Figure (Figure1,1, C and D). Furthermore, no cytotoxicity of CTL RP1 was observed against TNF-α– and IFN-γ–pretreated BM-derived fibroblasts (Figure (Figure1C).1C). By contrast, all target cell types were significantly recognized by the allo–HLA-B7–specific CTL KOR18, indicating their susceptibility to CTL-mediated recognition (data not shown). Together these data indicate that CTL RP1 recognizes an HLA-B7–restricted mHAg that is expressed by hematopoietic cells of the lymphoid lineage. Based on this CTL recognition pattern, we provisionally designated this novel mHAg as LRH-1.
We performed genetic linkage analysis to map the chromosomal region containing the LRH-1–encoding gene. EBV-LCLs from parental individuals of 13 Centre d’Etude du Polymorphism Humain (CEPH) families (provided by P. Martin, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA) were transduced with HLA-B*0702 and tested for CTL RP1 recognition. Six CEPH families (i.e., 1331, 1332, 1347, 1362, 1413, and 102), of which either one or both parents were positive, were selected for further study. CTL RP1 recognition of HLA-B*0702–transduced EBV-LCLs from all individuals of these families yielded informative LRH-1 segregation patterns for genome-wide linkage analysis (Figure (Figure2A).2A). All HLA-B*0702–transduced EBV-LCLs were significantly recognized by the allo–HLA-B7–specific CTL KOR18, indicating that the HLA-restriction molecule was properly expressed (data not shown). A cluster of markers on chromosome 17p13.2 was significantly linked (i.e., log odds ratio [lod] scores > 3.0) to the locus encoding LRH-1 (Figure (Figure2B).2B). The maximum lod score obtained was 13.77 at θ = 0.001 for marker D17S1810. A 7.5-cM critical region of no recombination was defined between the flanking markers D17S1845 (7.19 cM) and D17S938 (14.69 cM) on the Marshfield genetic map. This region corresponds to a distance of 3.5 Mb according to the UCSC Genome Bioinformatics database (http://genome.ucsc.edu/; ref. 27).
To narrow down the genetic region that includes the LRH-1 locus, we performed haplotype analysis using markers between D17S831 and D17S796. This analysis identified informative recombinant haplotypes in individual 13182 of family 102, individual 4889 of family 1332, and individual 8564 of family 1362 (Table (Table1).1). The recombinant paternal haplotype inherited by individual 13182 in family 102 indicated that the gene encoding LRH-1 is located telomeric to marker D17S1876. In an analogous fashion, the recombinant maternal haplotype inherited by individual 4889 in family 1332 and the recombinant paternal haplotype inherited by individual 8564 in family 1362 implied that the LRH-1–coding gene lies centromeric to marker D17S1845. Taken together, these data indicated that the gene encoding LRH-1 is located between marker D17S1845 and D17S1876, a region of 1.64 Mb on chromosome 17p13.2.
In the region deduced from linkage analysis, we identified 30 known genes using the NCBI (http://www.ncbi.nlm.nih.gov/) and UCSC Genome Bioinformatics databases. Eleven of these genes represented members of the olfactory receptor gene family and were excluded from further analysis. For the remaining 19 candidate genes, we developed real-time quantitative PCR and determined the expression level of each candidate gene in LRH-1–positive (EBV-LCL and Raji) and LRH-1–negative cell types (monocytes and fibroblasts). Expression profiles of CTNS, P2X5, and ATP2A3 were most closely associated with the pattern of recognition by CTL RP1 (Table (Table2).2). However, the expression level of CTNS and ATP2A3 was only 2- to 3-fold lower in monocytes compared with EBV-LCL, whereas P2X5 mRNA was expressed at a 20-fold-lower level. Furthermore, P2X5 mRNA was not detected in fibroblasts. These data suggested that the P2X5 gene was the most likely candidate gene to encode LRH-1.
To confirm these results, we constructed expression plasmids encoding the P2X5 isoforms 1 and 2 using cDNA isolated from the homozygous LRH-1–positive CEPH individual 1331-8234. P2X5 isoform 2 lacks exon 3 (i.e., 24 aas) and 1 codon in exon 7. CTL RP1 efficiently recognized 293T-HLA-B*0702 cells transfected with P2X5 isoform 1, whereas isoform 2 was not recognized (Figure (Figure3A).3A). Both P2X5 isoforms cloned from the homozygous LRH-1–negative donor EBV-LCLs did not stimulate IFN-γ release by CTL RP1 (Figure (Figure3A).3A). Furthermore, 293T-HLA-B*0702 cells transfected with a deletion construct encoding exons 1–5 of P2X5 isoform 1 from CEPH individual 1331-8234 were efficiently recognized by CTL RP1 (Figure (Figure3A).3A). Collectively, these data demonstrate that the LRH-1 antigenic epitope is derived from the 24-aa sequence encoded by exon 3 of P2X5 isoform 1.
We searched the human genome database for previously identified polymorphisms in exon 3 of P2X5 and found a cytosine deletion polymorphism in the region encoding aas 110 and 111 of P2X5 isoform 1 (Figure (Figure3B).3B). To determine whether this polymorphism could be involved in recognition by CTL RP1, we sequenced and compared the P2X5 exon 3 sequence (72 bp in length) from the homozygous LRH-1–negative donor EBV-LCL and the homozygous LRH-1–positive EBV-LCL of CEPH individual 1331–8234. Interestingly, 1 cytosine nucleotide in the donor P2X5 exon 3 sequence was indeed deleted between bp 329 and 333 of the open reading frame, resulting in a frameshift at aa position 112 and a putative truncated P2X5 protein of 146 aas (Figure (Figure3B).3B). These results demonstrate that a single nucleotide deletion polymorphism in the P2X5 sequence controls expression of the LRH-1 antigenic epitope.
As a result of the deletion polymorphism in exon 3, the P2X5 isoform 1 of the recipient contains a disparate NQRQNVCAE sequence compared with that of the donor (Figure (Figure3B).3B). To identify the antigenic epitope, we analyzed this disparate sequence flanked at both sides by 9 naturally occurring aas for HLA-B7–binding peptides using BIMAS prediction software. This resulted in the identification of the nonamer peptide TPNQRQNVC and the decamer peptide TPNQRQNVCA, which both have a predicted BIMAS binding score of 2.0. We synthesized these peptides and tested their ability to stimulate IFN-γ release by CTL RP1 upon loading on donor EBV-LCL at various concentrations. Both peptides were efficiently recognized by CTL RP1 (Figure (Figure3C).3C). However, half-maximal IFN-γ release for the 9-mer peptide was obtained at pulsing EBV-LCL with approximately 150 nM, whereas the 10-mer peptide showed half-maximal IFN-γ release at a concentration of approximately 350 nM, suggesting higher avidity of CTL RP1 for the 9-mer epitope. Furthermore, HLA-B7 tetramers constructed with the 9-mer peptide efficiently stained CTL RP1, whereas HLA-B7 tetramers constructed with the 10-mer peptide were negative (data not shown). To further define the minimal epitope, we tested truncated versions of the 9-mer peptide. A peptide with a deletion of the Thr aa residue at the NH2 terminus failed to sensitize donor EBV-LCLs even at very high peptide concentrations, indicating that the Thr is the N-terminal end of the naturally processed peptide. Deletion of the Cys aa residue at the COOH terminus also resulted in a significant decrease in recognition (Figure (Figure3C).3C). Taken together, these data indicated that the 9-mer TPNQRQNVC peptide is the naturally presented epitope at the cell surface by HLA-B7 molecules that is recognized by CTL RP1.
To detect LRH-1–specific CD8+ T cells in vivo, we constructed an HLA-B7 tetramer using the 9-mer peptide TPNQRQNVC. Before DLI, tetramer+ cells could not be detected (≤0.01%) in the peripheral blood of CML patient UPN389 (data not shown), whereas LRH-1–specific CD8+ T cells were clearly detectable after DLI-2 (Figure (Figure4).4). The PBMC sample that was used for isolation of CTL RP1 contained 0.27% of tetramer+ cells among CD8+ T cells. This percentage increased rapidly in the peripheral blood to a peak level of 1.6% at 6 months after cessation of CsA treatment (Figure (Figure4).4). At that time, 1.4% LRH-1–specific CD8+ T cells could also be detected in the BM (data not shown). These in vivo–expanded LRH-1–specific CD8+ T cells displayed the CD45RA+CCR7–CD27– cytotoxic effector phenotype, indicative of an ongoing GVL reaction (data not shown). The number of circulating tetramer+ cells declined to 0.25% at 12 months but persisted at a low frequency (0.05–0.10%) (Figure (Figure4).4). Specific detection of LRH-1–specific CTLs by the tetramers was verified after in vitro stimulation of PBMC samples with peptide-pulsed donor EBV-LCLs. After 2 rounds of stimulation, tetramer+ CD8+ T cells markedly increased up to 6.8% for month 3, 13.9% for month 6, 4.2% for month 12, and 2.0% for month 38 (Figure (Figure44).
Correlation of the emergence of LRH-1–specific CTLs with clinical outcome of DLI-2 revealed that the peak of tetramer+ cells was followed immediately by pancytopenia and a dramatic decrease in Bcr-Abl in the peripheral blood without an increase of GVHD (Figure (Figure5,5, A and B). Thereafter, the patient remained in complete cytogenetic remission for more than 2 years despite readministration of CsA to treat chronic skin GVHD. However, 4 years after DLI-2, PCR for Bcr-Abl had become substantially positive (Figure (Figure5B).5B). The patient was given DLI-3 at a relatively low dose of 1 × 106 CD3+ cells/kg body weight in order to avoid severe GVHD. The patient did not develop GVHD, and her tumor transformed into CML in accelerated phase with almost all BM cells from recipient origin and with del(13) and t(4;11) as found during the first relapse. She was treated with imatinib mesylate (Glivec), and despite DLI-4 (0.1 × 108 CD3+ cells/kg body weight) and DLI-5 (0.6 × 108 CD3+ cells/kg body weight), she entered myeloid blastic phase and died without GVHD. During this second relapse and concomitant treatment, LRH-1–specific CD8+ T cells were present but did not expand (Figure (Figure5,5, A and B).
To more comprehensively define which cell types and tissues could be targeted by LRH-1–specific CTLs, we developed a real-time quantitative PCR that detects P2X5 transcripts encoding the LRH-1 epitope. The mean P2X5 RNA expression level compared with the reference cell line was 2.68 in EBV-LCLs, 0.28 in monocytes, and 0.10 in fibroblasts (Figure (Figure6A).6A). Based on our results that LRH-1–specific CTL recognizes EBV-LCLs but not monocytes and fibroblasts, we arbitrarily set the cutoff level at 0.4 to discriminate between LRH-1–positive and –negative cell types. CD34+CD38– and CD34+CD38+ cells isolated from myeloid leukemias, including CML in blast crisis and acute myeloid leukemia (AML), showed significant P2X5 RNA expression levels, as well as CD34+ progenitor subsets isolated from normal BM and G-CSF–mobilized blood (Figure (Figure6A).6A). Interestingly, leukemic CD34+CD38– cells appear to have a significantly higher P2X5 RNA level compared with normal CD34+CD38– cells (P = 0.02; ANOVA; Figure Figure6A).6A). CD34+CD38– cells obtained at first relapse from CML patient UPN389 had a relatively low P2X5 RNA expression of 0.61, whereas CD34+CD38+ cells had a higher expression level of 2.67. A slightly lower P2X5 expression level of 1.31 was found in CD34+CD38+ cells obtained at second relapse (data not shown). We were unable to isolate enough CD34+CD38– cells (<5% of the total CD34+ population) from the second relapse sample. These data suggest that leukemic CD34+ progenitor subsets could be potentially targeted in vivo by LRH-1–specific CTLs.
Freshly isolated as well as cultured cells of the lymphoid lineage including B cells, EBV-LCLs, T cells, PHA–T cell blasts, and NK cells showed significant levels of P2X5 RNA expression (Figure (Figure6B).6B). In contrast, mature cells of the myeloid lineage including monocytes and DCs and nonhematopoietic cells including fibroblasts, keratinocytes, melanocytes, human umbilical vein endothelial cells (HUVECs), and proximal tubular epithelial cells (PTECs) had a very low level of or completely lacked P2X5 transcription. Analysis of P2X5 expression in normal tissues showed high (>1.0) transcription levels in lymphoid tissues such as spleen, tonsils, thymus, and BM (Table (Table3).3). No significant (<0.4) P2X5 expression could be detected in most adult normal tissues, except for low levels in neuronal and muscle tissue. Fetal brain showed a relatively high P2X5 expression level (2.80), whereas expression levels in the adult brain (0.43) and cerebellum (0.70) were much lower. Importantly, no P2X5 expression was detected in the GVHD target organs skin, liver, colon, and small intestine (Table (Table3).3). These data demonstrate that P2X5 is selectively expressed in leukemic and normal CD34+ progenitor cells and hematopoietic cells of the lymphoid lineage.
mHAgs play a pivotal role in inducing T cell responses mediating GVL reactivity after HLA-identical allogeneic SCT (4, 5). Molecular definition of mHAg peptides is crucial for the development of immunotherapeutic protocols to either selectively boost the GVL reaction or to inhibit GVHD. In this study, we have identified a novel hematopoietic cell–restricted mHAg, designated LRH-1, which is presented by HLA-B*0702 and encoded by the P2X5 purinergic receptor gene. CTLs specific for LRH-1 (RP1 clone) were isolated from a CML patient that was successfully treated with high-dose DLI for relapse in accelerated phase after allogeneic SCT. Here, we show a direct association between in vivo expansion of LRH-1–specific CTLs and the disappearance of Bcr-Abl–positive CML tumor cells following DLI. These findings indicate a causative role for LRH-1 in GVL reactivity. Although the patient achieved a complete cytogenetic remission after high-dose DLI, low levels of Bcr-Abl transcripts remained detectable (Figure (Figure5).5). Therefore, it remains to be investigated whether targeting of LRH-1 in patients with CML might be able to eradicate the most undifferentiated leukemic stem cells. GVHD was not increased by the emergence of LRH-1–specific CTLs despite interruption of CsA treatment. However, it is difficult to conclude that LRH-1–specific CTLs had no role in GVHD, since the patient required readministration of CsA for the treatment of chronic skin GVHD.
The level of expression of mHAgs may determine whether it is involved in GVL and/or GVHD (5). Previous studies by RNA dot blot analysis showed that P2X5 is selectively expressed in brain and lymphoid tissues (28). Here, we have used sensitive real-time quantitative PCR to determine relative expression levels of the P2X5 splice variant 1 encoding mHAg LRH-1. Our results confirm that P2X5 is highly expressed in various cell types of lymphoid origin and lymphoid tissues. Interestingly, we observed that the P2X5 gene is significantly expressed in leukemic CD34+ subpopulations from most CML as well as AML patients. We were unable, however, to demonstrate in vitro CTL RP1 recognition of CD34+ cells of the CML patient from whom the LRH-1–specific CTL was isolated (data not shown). This may be explained by the relatively low P2X5 expression level (0.61) in the leukemic CD34+CD38– progenitor subset of this CML patient, which represented approximately 50% of the total CD34+ population (data not shown). But, P2X5 expression in vivo could be higher, resulting in increased sensitivity to CTL recognition. The efficiency of killing of leukemic CD34+ progenitor cell subsets by LRH-1–specific CTLs is currently under investigation using sensitive in vitro progenitor cell cytotoxicity assays and could be studied in vivo in the NOD/SCID leukemia model (14, 29).
Importantly, P2X5 is not expressed in prominent GVHD tissues such as skin, liver, colon, and small intestine. Furthermore, mature cells of myeloid origin and various nonhematopoietic cell types were negative for P2X5. In nonlymphoid tissues, we detected only high P2X5 expression levels in fetal brain, whereas in adult tissues, we observed low expression in brain, cerebellum, and skeletal muscle, which could not be explained by the presence of hematopoietic cells (Table (Table3).3). This low level of P2X5 expression could be caused by a small subset of cells present in these tissues (30, 31). But based on the fact that both brain and skeletal muscle are tissues that are not very likely to be infiltrated with lymphocytes without significant inflammation (32), it seemed unlikely that LRH-1–specific CTLs damage these cells in vivo. This is supported by our finding that emergence of LRH-1–specific CTLs was not accompanied by significant toxicity to these organs.
Most previously described human mHAgs result from disparities in protein sequences between donor and recipient due to nonsynonymous SNPs in their encoding genes. The resulting aa substitutions may affect either peptide processing by the proteasome (10), transporter associated with antigen processing translocation of the peptide into the ER (23), binding of the peptide to MHC (8), or recognition of the MHC-peptide complex by mHAg-specific T cells (20, 22, 33). In the case of an mHAg encoded by the UGT2B17 gene, disparity resulted from differential protein expression due to a homozygous gene deletion in the donor, although the genetic basis for this gene deletion remains to be elucidated (25). Here, we demonstrate for the first time to our knowledge that mHAgs can arise from frameshift polymorphisms that are disparate between recipients and their HLA-identical transplant donors.
P2X5 is a member of the P2X purinergic receptor family consisting of at least 7 proteins, P2X1 to P2X7, which are capable of forming ATP-gated cation channels by homo- or heteromultimerization (34). Transcription of the human P2X5 gene results in expression of 2 major isoforms as a result of differential splicing (28). Splice variant 1 encodes a long isoform consisting of 422 aas, whereas translation of splice variant 2 results in a shorter isoform lacking exon 3 (397 aas). Our results indicate that a donor-specific, homozygous frameshift polymorphism in exon 3 of splice variant 1 leads to the formation of a truncated aberrant version of the long P2X5 isoform. Hence, donor-derived T lymphocytes are not tolerant of the wild-type exon 3 of the recipient. The high genotype frequency (i.e., 46%; unpublished observations) of healthy individuals that are homozygous for the cytosine deletion in exon 3 indicates that deficiency for the long P2X5 isoform is common without any significant phenotypic characteristics. The short P2X5 isoform as well as redundancy among P2X family members may compensate for the deficiency of the long P2X5 isoform.
Hematopoietic cell–restricted mHAgs are suitable targets for the application of immunotherapy to prevent or to treat recurrences of hematopoietic malignancies after allogeneic SCT. Based on its hematopoietic-specific expression, the P2X5-encoded mHAg LRH-1 could be an attractive candidate. P2X5 is highly expressed in a broad range of lymphoid malignancies including T and B cell acute lymphoblastic leukemia, chronic lymphocytic leukemia, various types of B cell lymphoma, and multiple myeloma (unpublished observations). Furthermore, CD34+CD38– and CD34+CD38+ myeloid leukemic progenitor cells were found to have enriched levels of P2X5 transcripts compared with mature nonleukemic monocytes (P < 0.05; ANOVA; Figure Figure6A),6A), suggesting that LRH-1–specific CTLs may preferentially target immature myeloid leukemia cells. Therefore, vaccination with the LRH-1 peptide would be an attractive approach for induction of GVL reactivity. Recently we developed a genotyping assay for the LRH-1 mHAg, and of the 45 HLA-B7–positive SCT donors and recipients analyzed so far, 6 patients (~13%) positive for LRH-1 mHAg received a transplant from an LRH-1–negative donor (unpublished observations). These patients would be eligible for LRH-1 peptide–based vaccination. Studies are underway to determine whether LRH-1–specific CD8+ T cells can be detected or induced ex vivo in these LRH-1 disparate transplant recipients.
In conclusion, we describe a novel hematopoietic cell–restricted mHAg, designated LRH-1, that can function as a GVL-specific antigen for hematological malignancies. The significance of mHAg LRH-1 for cellular immunotherapy of relapsed leukemia is demonstrated by the direct correlation between expansion of LRH-1–specific CTLs and disappearance of CML, and the expression of its encoding gene in leukemic CD34+ progenitor cells. In addition, we describe a novel mechanism responsible for antigenicity of mHAg LRH-1. Our data demonstrate for the first time to our knowledge that differential protein expression as a consequence of a homozygous frameshift polymorphism is the basis for mHAg disparity. This novel mechanism may apply to other mHAgs involved in alloresponses after HLA-identical allogeneic SCT.
CD8+ CTL clone RP1 was isolated from PBMCs obtained 3 months after DLI-2 by weekly stimulation with CML cells obtained at hematological relapse in Iscove’s modified Dulbecco’s medium (IMDM) (Invitrogen Corp.) supplemented with 10% human serum (HS) (PAA Laboratories). Relapse CML cells that were used as stimulator cells contained 5% CD34+ blasts, 3% CD19+ B cells, 29% CD3+ T cells, and 60% CD13+CD33+ myeloid cells. Myeloid cells were completely of recipient origin, and lymphocytes were mixed chimeric. After initial stimulation, CTL RP1 (0.5 × 106) was cultured in IMDM/10% HS containing irradiated (80 Gy) recipient EBV-LCLs (0.5 × 106), irradiated (60 Gy) allogeneic PBMCs (0.5 × 106) from 2 donors, 100 IU/ml IL-2 (Eurocetus), and 1 μg/ml PHA-M (Boerhinger) in a total volume of 2 ml. The HLA-B7–alloreactive CTL KOR18 was kindly provided by E. Goulmy (Leiden University Medical Center). CTL KOR18 (2 × 106) was cultured in IMDM/10% HS containing irradiated HLA-B7–positive EBV-LCLs (5 × 106), irradiated allogeneic PBMCs (5 × 106) from 2 donors, 100 IU/ml IL-2, and 1 μg/ml PHA-M in a total volume of 15 ml. CTL cultures were used in cytotoxicity and stimulation assays either 7 days after stimulation or 1 day after thawing a frozen aliquot. All cell lines and primary cells were cultured in IMDM/10% FCS. EBV-LCLs from the CEPH families were provided by P. Martin (Fred Hutchinson Cancer Research Center, Seattle, Washington, USA) or purchased from the Coriell Institute for Medical Research. B cells, T cells, and monocytes were isolated from PBMCs using anti-CD19, anti-CD3, and anti-CD14 immunomagnetic beads (Miltenyi Biotec), respectively. B cell blasts were generated by culturing CD19+ B cells in IMDM/10% HS with 1 μg/ml CD40L (gift from Immunex Corp.) and 100 U/ml IL-4 (Schering-Plough) for 5 days. T cell blasts were generated by stimulating CD3+ T cells in IMDM/10% HS with 20 μg/ml PHA-M for 3 days. Thereafter, PHA-activated T cells were washed and further cultured with 100 IU/ml IL-2 for 2 additional days. Immature DCs were generated by culturing CD14+ monocytes in IMDM/5% HS with 800 U/ml GM-CSF (Schering-Plough) and 500 U/ml IL-4 for 6 days. Immature DCs were maturated by culturing for 3 additional days in IMDM/5% HS with 800 U/ml GM-CSF, 500 U/ml IL-4, 20 ng/ml TNF-α (Endogen), and 10 μg/ml Prostin E2 (Pfizer). Fibroblasts were isolated from BM aspirates of healthy stem cell donors. BM was resuspended in 20 ml IMDM/20% FCS and incubated overnight in 75-cm2 tissue culture flasks. The nonadherent fraction was removed, and fibroblasts were further cultured in IMDM/20% FCS to passage 3 before analysis.
Chromium release assays were performed to assess cytolytic activity of CTLs as described previously (12). Fibroblasts were pretreated with 10 ng/ml TNF-α and 100 U/ml IFN-γ (Boehringer) for 2 days before 51Cr labeling. The following Abs were used for inhibition of CTL reactivity: SPV-T3b (anti-CD3), W6/32 (anti–HLA class I), B1.23.2 (anti–HLA-B/C), and SPVL-3 (anti–HLA-DQ).
HLA-B*0702 cDNA (provided by E. Goulmy) was subcloned in the LZRS-IRES-EGFP vector. The resulting LZRS-HLA-B*0702-IRES-EGFP vector was used to generate a stable producer cell line. Retroviral transduction was performed using non–tissue culture–treated 35-mm2 dishes (BD) coated with 10 μg/ml RetroNectin (Takara Bio Inc.). Target cells (106) were resuspended in 2 ml virus supernatant and transferred to RetroNectin-coated dishes. After 24 hours of incubation, cells were collected and incubated with fresh virus supernatant. Finally, transduced cells were cultured for 5 additional days before use in CTL stimulation assays as described previously (13). Release of IFN-γ was determined by ELISA (Endogen; Pierce Biotechnology Inc.).
Mapping of the LRH-1 locus was accomplished in a subset of CEPH reference families. Genotypes were retrieved from the CEPH database, version 9.0 (http://www.cephb.fr/cephdb/; refs. 35, 36). Two-point linkage was performed using the MLINK subroutine of the FASTLINK (version 4.0) program (http://www.cs.rice.edu/~schaffer/fastlink.html) as reported previously (11, 36, 37). An autosomal dominant, single gene model was assumed for the LRH-1 trait. The population frequency of the encoding allele was estimated at 0.5 and the penetrance of the LRH-1 phenotype was set at 0.9. Population allele frequencies for the genetic markers were used when present in the CEPH database; otherwise equal allele frequencies were assumed. Resulting lod scores were examined for regions of highly significant linkage, and the identification of the LRH-1 locus was interpreted from the Marshfield genetic map (38).
Total RNA was isolated by the TRIzol method (Invitrogen Corp.) and treated with DNase. First-strand cDNA was prepared using oligo-dT primer and Powerscript RT (BD Biosciences — Clontech). Gene-specific primer sets were designed at the 3′ end of each candidate gene. One microliter of cDNA was amplified in a 50-μl reaction mixture containing 1.25 U AmpliTaq Gold (Applied Biosystems), 300 nM of each primer, 250 μM of each dNTP, 5 mM MgCl2, and 1× SYBR green PCR buffer (Applied Biosystems). The porphobilinogen deaminase (Pbgd) gene was used to normalize expression of each candidate gene. The following primers and TaqMan probe were used for Pbgd real-time PCR: Pbgd-F 5′-GGCAATGCGGCTGCAA-3′, Pbgd-R 5′-GGGTACCCACGCGAATCAC-3′, and Pbgd-probe 5′-(VIC)-CTCATCTTTGGGCTGTTTTCTTCCGCC-(TAMRA)-3′. Pbgd-PCR was performed as described above using 300 nM TaqMan probe. PCR amplification was performed using an ABI Prism 7700 (Applied Biosystems) with the following PCR conditions: enzyme activation for 10 minutes at 95°C, followed by 45 cycles of 95°C for 15 seconds and 60°C for 1 minute. The cycle threshold (Ct) was determined by software analysis (Sequence Detection System, version 1.9.1; Applied Biosystems). The expression of candidate genes was calculated using EBV-LCLs of the recipient as calibrator with the following formula: [2–(ΔΔCt)] × 100 (39).
Constructs containing the full-length or a truncated form of the P2X5 transcript variants were generated by PCR. cDNA was prepared from a homozygous LRH-1–positive and LRH-1–negative EBV-LCLs and amplified using specific primers. The following primers were used: P2X5-FLF 5′-CATGGGGCAGGCGGG-3′ and P2X5-FLR 5′-AACGTAAGCAGAGGCAATTCACGT-3′. A truncated form containing exons 1–5 of P2X5 transcript variant 1 was generated with the P2X5-FLF forward primer and reverse primer P2X5-TRR 5′-GGGGAAACGGATGTGGTTCT-3′. PCR products were purified using the QIAGEN gel extraction kit and cloned into vector pCR3 (Invitrogen Corp.) using the unidirectional TA cloning kit (Invitrogen Corp.). Plasmids (400 ng) were used for transfection into 3.5 × 104 293T-HLA-B*0702 cells plated in poly-D-lysine–coated 96-well flat-bottom wells (BD) using Lipofectamine (Invitrogen Corp.). Transfected cells were cultured for 2 additional days and used in CTL stimulation assays.
HLA-B7–binding peptides were predicted using BIMAS software (http://bimas.dcrt.nih.gov/molbio/hla_bind/; ref. 40). Peptides were synthesized by solid-phase strategies on an automated multiple peptide synthesizer (AMS 422; ABIMED). Peptides were purified by reverse-phase HPLC and lyophilized. Finally, peptides were dissolved in DMSO, diluted in IMDM to a concentration of 1 mM, and stored at –20°C before use. In peptide recognition assays, target cells were preincubated with various concentrations of peptide for 30 minutes in a volume of 100 μl prior to the addition of CTLs.
PE-labeled LRH-1/HLA-B7 tetramers using peptide TPNQRQNVC were produced as described previously (41). LRH-1 tetramers were validated by staining against CTL RP1. For staining, PBMC or CTL cell lines were incubated with 20 μg/ml LRH-1 tetramer for 30 minutes at room temperature. After washing with PBS/0.5% BSA, cells were labeled with the appropriate concentration anti-CD8 (LT8-FITC; ProImmune), anti-CD3 (UCHT1-Cy7; Beckman Coulter), anti-CD45 (J33-ECD; Beckman Coulter), and 1 μg/ml 7-amino-actinomycin D (7AAD) at room temperature for 15 minutes. After washing, cells were resuspended in PBS/0.5% BSA and analyzed using the Coulter FC500 flow cytometer (Beckman Coulter).
Bcr-Abl PCR analysis was performed on an ABI Prism 7700 system (Applied Biosystems) as described previously (42). In brief, total RNA from cells was extracted using TRIzol (Invitrogen Corp.). Reverse transcription was performed on 1 μg of total RNA using random hexamers and Mo-MuLV RT (Invitrogen Corp). Expression of the Pbgd gene was used to normalize Bcr-Abl expression. This normalized Bcr-Abl expression in patient samples was related to a standard curve obtained from K562 cells diluted into normal BM cells.
PBMCs were isolated from normal donors using Ficoll-Paque (Amersham Biosciences) and used for isolation of different cell subsets. PBMCs were labeled with the appropriate concentration of FITC-, PE-, and/or Cy5-conjugated Abs. Labeled cells were sorted into CD19+ B cells, CD3+ T cells, CD3–CD16/56+ NK cells, and CD14+ monocytes using the Coulter EPICS Elite flow cytometer (Beckman Coulter). Normal and leukemic CD34+CD38– and CD34+CD38+ cells were sorted from cell suspensions after labeling with CD34-FITC and CD38-PE Abs (BD). Immature DCs, mature DCs, and fibroblasts were cultured as described above. Keratinocytes were provided by S. Ebeling (University Medical Center Utrecht, Utrecht, The Netherlands). PTECs, HUVECs, and melanocytes were provided by E. Spierings (Leiden University Medical Center). Total RNA from each of these cell types was isolated using the Zymo RNA isolation kit II (Zymo Research Corp.). RNA of normal tissues was obtained from the RNA Master Panel II (BD Biosciences — Clontech). First-strand cDNA was prepared from total RNA using oligo-dT, random hexamers, and Mo-MuLV RT (Invitrogen Corp.). In parallel to P2X5 expression, we used CD45 expression to detect the presence of hematopoietic cells in normal tissues. The Pbgd housekeeping gene was used to normalize P2X5 and CD45 expression. One microliter of cDNA was amplified in a 50-μl reaction mixture containing 1.25 U AmpliTaq Gold (Applied Biosystems), 300 nM gene-specific forward and reverse primer, 150–300 mM gene-specific TaqMan probe (150 nM for P2X5, 200 nM for CD45, and 300 nM for Pbgd), 250 μM of each dNTP, 5 mM MgCl2, and 1× TaqMan PCR buffer (Applied Biosystems). The following gene-specific primers and TaqMan probes were used: P2X5: P2X5-F 5′-TCCTGGCGTACCTGGTCGT-3′, P2X5-R 5′-CTTCATTCTCAGCACAGACGTTC-3′, and P2X5-probe 5′-(TET)-TGGGTGTTCCTGATAAAGAAGGGTTACCA-(TAMRA)-3′; and CD45: CD45-F 5′-TTAGAAATGAGTCGCATAAGAATTGC-3′, CD45-R 5′-CTCCAGGATAGTCTCCATTGTGAAA-3′, and CD45-probe 5′-(TET)-ATTTCCGTGTAAAAGATC-(TAMRA)-3′. PCR conditions were as follows: enzyme activation for 10 minutes at 95°C, followed by 45 cycles of 95°C for 15 seconds and 60°C for 1 minute. P2X5 and CD45 mRNA expression was quantified by determining calibration functions using JVM-2 as reference cell line. The level of P2X5 and CD45 expression of test samples was calculated relative to the P2X5 and CD45 expression in the JVM-2 cell line. P2X5 mRNA expression levels of different cell types were compared using ANOVA and Duncan post-hoc comparison. P < 0.05 was considered significant.
We thank Hanny Fredrix, Arie Pennings (now deceased), and Gertie Vierwinden for technical assistance. We thank Ingrid Overes for CD45 PCR analysis and Shahla Abdollahi for help in LRH-1 phenotyping of LCLs. We thank Els Goulmy and Eric Spierings for providing the HLA-B*0702 cDNA, HLA-B7–alloreactive CTL KOR18, and RNA from nonmalignant cell types. We acknowledge our Molecular Diagnostics unit for performing Bcr-Abl PCR analysis. We thank Saskia Ebeling for providing RNA from keratinocytes. We thank Elke de Grouw for providing RNA of CD34+ progenitor subsets. We acknowledge Joop Jansen and Gosse Adema for critically reading the manuscript. We thank Fred Falkenburg (Department of Hematology, Leiden University Medical Center) for advice and discussion. This work was supported by grants from the Dutch Cancer Society (KUN 2000-2294), the Bekales Foundation, the W.M. Keck Foundation, and the George S. and Delores Doré Eccles Foundation.
See the related Commentary beginning on page 3397.
Nonstandard abbreviations used: CEPH, Centre d’Etude du Polymorphism Humain; CML, chronic myeloid leukemia; CsA, cyclosporin A; DLI, donor lymphocyte infusion; GVHD, graft-versus-host disease; GVL, graft-versus-leukemia; IMDM, Iscove’s modified Dulbecco’s medium; LCL, lymphoblastoid cell line; lod, log odds ratio; LRH-1, lymphoid-restricted histocompatibility antigen–1; mHAg, minor histocompatibility antigen; PHA, phytohemagglutinin; SCT, stem cell transplantation.Conflict of interest: The authors have declared that no conflict of interest exists.