Toward targeting B cell cancers with CD4+ CTLs: identification of a CD19-encoded minor histocompatibility antigen using a novel genome-wide analysis

Abstract

Some minor histocompatibility antigens (mHags) are expressed exclusively on patient hematopoietic and malignant cells, and this unique set of antigens enables specific targeting of hematological malignancies after human histocompatability leucocyte antigen (HLA)-matched allogeneic stem cell transplantation (allo-SCT). We report the first hematopoietic mHag presented by HLA class II (HLA-DQA1*05/B1*02) molecules to CD4(+) T cells. This antigen is encoded by a single-nucleotide polymorphism (SNP) in the B cell lineage-specific CD19 gene, which is an important target antigen for immunotherapy of most B cell malignancies. The CD19(L)-encoded antigen was identified using a novel and powerful genetic strategy in which zygosity-genotype correlation scanning was used as the key step for fine mapping the genetic locus defined by pairwise linkage analysis. This strategy was also applicable for genome-wide identification of a wide range of mHags. CD19(L)-specific CD4(+) T cells provided antigen-specific help for maturation of dendritic cells and for expansion of CD8(+) mHag-specific T cells. They also lysed CD19(L)-positive malignant cells, illustrating the potential therapeutic advantages of targeting this novel CD19(L)-derived HLA class II-restricted mHag. The currently available immunotherapy strategies enable the exploitation of these therapeutic effects within and beyond allo-SCT settings.

Figure 1.

Five-step identification of the SNP encoding for the mHag recognized by clone 21. (A) mHag phenotypes of CEPH families (indicated with Utah database ID numbers) were determined using methods described in the Materials and methods section. CEPH family 1362 is depicted as an example (male, square; female, circle; mHag⁺, black; mHag⁻, white; undetermined, gray). Phenotypes of families 1331, 1408, and 1416 are given in Fig. S1 (available at http://www.jem.org/cgi/content/full/jem.20080713/DC1). (B) Genome-wide pairwise two-point linkage analysis using the mHag phenotypes from families 1331, 1362, and 1408. Multiple significant lod scores >3 (at a recombination fraction of θ = 0.001) were identified on chromosome 16 in the depicted region. (C) Narrowing of the mHag locus using haplotype data from family 1416. As depicted, the mHag⁺ children 1189 and 2387 inherited the dark gray recombinant haplotype from the father, who is also mHag⁺. Thus the mHag locus was narrowed to the 16.8-cM region, which is defined by the shared part of the paternal allele of children 1189 and 2387. (D) Zygosity-genotype correlation analysis for fine mapping the mHag locus. The r² values in the y axis represent the correlation between the mHag zygosity of 15 CEPH individuals (Table S1) with the genotypes for 4146 HapMap SNPs in this region. Each bar represents a single SNP. Two SNPs (rs7184597 and rs3924376) with 100% correlation (r² = 1) are indicated. (E) The location of rs7184597 and rs3924376 (both light gray) in the intronic regions of the RABEP2 gene, which is neighbored by the CD19 gene. Also indicated are the nine nonsynonymous or transcription/translating-altering SNPs in these two genes. The mHag phenotypes and zygosities (Pheno./Zyg.) and the SNP genotypes for five informative CEPH individuals, the SC donor, and the SC recipient are depicted (−, deletion of the base pair). Only rs2904880 matched exactly with the phenotypes.

Figure 2.

CD19^L encodes for the mHag recognized by clone 21. (A) The CD19 gene with rs2904880 in the third exon encoding a valine (V) to leucine (L) substitution at position 174. (B) IFN-γ response of clone 21 to mHag⁻ donor (Do) EBV-LCLs transduced either with an empty vector (mock) or with the CD19^L-encoding vector. Response to mHag⁺ recipient (Rt) EBV-LCLs (LCLs) is depicted as positive control. The mean and SEM of three experiments are depicted. (C and D) IFN-γ response of clone 21 toward serial concentrations of 15-mer peptides derived either from CD19^L (▪) or from CD19^V (○; C) or toward CD19^L-derived overlapping 15-mer peptides (D). Donor EBV-LCLs were used as APCs. The core sequence recognized by clone 21 is highlighted in gray. Error bars represent the SEM of triplicate cultures. (E) In vivo presence of CD19^L-specific clone 21. Genomic DNA isolated from patient PBMCs at the indicated days after allo-SCT was used to amplify the TCR of clone 21. Vβ16-specific PCR was used as positive control. Genomic DNA from clone 21 (cl.21) and third-party PBMCs (irrel.) were used as positive and negative controls, respectively. Also indicated are the severity of acute and chronic GvHD, the serum M protein levels, and the immunosuppressive prednisone treatment during the monitoring period.

Figure 3.

Genome-wide mapping of the CD19^L-mHag locus by zygosity-genotype correlation analysis. The r² values on the y axis represent the correlation between the mHag zygosities of 23 CEPH individuals with the genotypes for all HapMap SNPs (public release 21a). 15 of these CEPH individuals are depicted in Table S1. For eight other individuals that were also phenotyped by clone 21 recognition, the zygosity information was derived from the rs2904880 genotype information in the latest HapMap release (public release 22; unpublished data). Each bar represents a single SNP; only r² values >0.5 are shown. Chromosome 16 was analyzed in more detail (bottom). All r² values are depicted.

Figure 4.

Retrospective genome-wide mapping of the HMSD-mHag locus by zygosity-genotype correlation analysis. The correlation analysis of the HMSD mHag, using zygosity data from 7–14 CEPH trios, is depicted as a representative example. On the y axis, r² values >0.5 are depicted representing the correlation between the HapMap-derived mHag zygosities of the CEPH individuals (Table S2, available at http://www.jem.org/cgi/content/full/jem.20080713/DC1) and the indicated number of trios with their genotypes for all HapMap SNPs (public release 23). The single r² = 1 peak after analysis with 11 trios consisted of seven SNPs within the same LD block.

Figure 5.

Retrospective genome-wide mapping of nonsynonymous SNPs with various allele frequencies by zygosity-genotype correlation analysis. Summary of correlation analyses for 149 HapMap SNPs (Table S3, available at http://www.jem.org/cgi/content/full/jem.20080713/DC1). From each SNP one of the alleles was designated to encode for a fictive mHag. For each 5% mHag frequency interval between 5 and 95%, and 10, 5, or 4 mHags were analyzed. The analyses were executed for a dataset without phenotyping errors (A) or with 7–10% false-positive phenotypes (B). The false-positive phenotypes were introduced in the corresponding datasets by randomly changing −/− typings into +/−, avoiding Mendelian segregation errors. The criterion for positive identification is r² = 1 at the genomic locus of the analyzed mHag without any r² = 1 false-positive hits at other genomic loci (A), or the criterion is r² is above the theoretical r² (Fig. S2, available at http://www.jem.org/cgi/content/full/jem.20080713/DC1) at the mHag genomic locus without any false-positive r² above this value at irrelevant genomic loci (B). The tables show the number of mHags analyzed for each frequency indicated below in the figures, as well as the number of successfully identified mHags. The figures show the number of used trios for only the successfully mapped mHags.

Figure 6.

The helper activity of CD19^L-specific clone 21. (A) The LRH-1–specific HLA-B7–restricted CD8⁺ clone cocultured with CD19^L-negative or -positive HLA-B7/DQA1*05/B1*02 EBV-LCLs (LCLs) in the presence of its own epitope. Different dilutions of irradiated clone 21 were added in the cultures and in some conditions were supplemented with the 15-mer CD19^L peptide. The proliferation of the LRH-1–specific CD8⁺ clone after 48 h is depicted. Error bars represent the SEM of triplicate cultures. The proliferation in the presence of CD19^L-positive EBV-LCLs was significantly higher than CD19^L- negative EBV-LCLs (*, P < 0.05). (B and C) Immature DCs generated from HLA-DQ2–matched monocytes loaded with 15-mer CD19^L peptide PEIWEGEPPCLPPRD or irrelevant peptide LPPRDSLNQSLSQDL (irr.pept.; B) or with apoptotic CD19^L-negative EBV-LCLs (apCD19^Lneg-LCL) transduced with CD19^L (apCD19^Lpos-LCL; C) were cultured with clone 21 (cl.21). Apoptosis was induced by incubation of EBV-LCLs with FasL. The CD4⁺ mHag-specific T cell clone 3AB11 (irr.cl.) or CD40L-expressing fibroblasts were used as negative and positive controls, respectively. DC maturation was assessed as described in Materials and methods.

Figure 7.

The effector function of CD19^L-mHag–specific clone 21. (A) Granzyme B production of clone 21 in response to recipient (Rt) or donor (Do) EBV-LCLs (LCLs). (B) The lysis of recipient EBV-LCLs or donor EBV-LCLs by clone 21 in the absence or presence of 15-mer CD19^L-peptide at an effector/target ratio of 50:1. The error bars represent the SEM of duplicate cultures. (C) IFN-γ response of clone 21 toward CD19⁺ malignant cells from 18 B-CLL patients. The mHag genotypes of the patients were determined by partial sequencing of the chromosomal DNA extracted from PBMCs. The mean and SEM are shown for the indicated number of patient samples. The difference between the HLA-matched CD19^L-positive patients and the others was statistically significant (*, P < 0.05). (D) The lysis of HLA-DQB1*02 and CD19^L-positive (n = 3) and CD19^L-negative (n = 2) B-CLL samples by clone 21 at different effector/target ratios. Error bars indicate the SEM of the different B-CLL samples.