Matching for the non-conventional MHC-I MICA gene significantly reduces the incidence of acute and chronic GVHD
R. Carapito (Strasbourg, FR)
The success of unrelated hematopoietic cell transplantations (HCT) relies greatly on Human Leukocyte Antigen (HLA) matching between patient and donor. But even in HLA matched transplantations, adverse clinical outcomes such as graft-versus-host-disease (GvHD) remain alarmingly high. The highly polymorphic MHC class I chain-related gene A (MICA) encodes a stress-induced glycoprotein interacting with NKG2D, an activating receptor expressed on the surface of cytotoxic CD8+ αβ and γδ T lymphocytes and natural killer cells. MICA may act as a transplantation antigen and therefore matching of donors and patients at the MICA locus could have a beneficial effect for patients undergoing unrelated HSCT. Using high-resolution sequence-based genotyping of MICA, we retrospectively analyzed the clinical impact of MICA-mismatches in a large multicentre cohort of 922 unrelated transplantations matched 10/10 for HLA-A, -B, -C, -DRB1, and –DQB1. We further examined severe acute GvHD patients with respect to MICA and NKG2D expression using immunohistochemical staining on intestinal biopsies and flow cytometry on PBMCs, respectively. Among the 922 pairs, 113 (12.3%) were mismatched for MICA. MICA mismatches were significantly associated with an increased incidence of grade III-IV acute GVHD (HR, 1.83; 95% CI, 1.50 to 2.23; P<0.001), chronic GvHD (HR, 1.50; 95% CI, 1.45 to 1.55; P<0.001) and non-relapse mortality (HR, 1.35; 95% CI, 1.24 to 1.46; P<0.001). The increased risk of GvHD was mirrored by a lower risk of relapse (HR, 0.50; 95% CI, 0.43 to 0.59; P<0.001), indicating a possible graft-versus-leukemia effect. In addition, expression of MICA and its ligand NKG2D were shown to be induced in intestinal tissue and circulating CD8+ T cells of acute grades III-IV GvHD patients, respectively. In conclusion, when possible, selecting a MICA-matched donor significantly influences key clinical outcomes of HCT in which a marked reduction of GvHD is paramount. The tight linkage disequilibrium between MICA and HLA-B renders identifying a MICA-matched donor readily feasible in clinical practice.
Examination of the regulation of CXCL10 expression, and CXCL10 DNA sequence variation and disease associations
M. Dorak (Liverpool, GB)
Chemokine proteins have important roles in inflammation and immunity. One member of this family, CXCL10, appears to be pleiotropic with effects in autoimmune disorders, transplantation, infectious diseases and cancer. We examined the regulation of expression of CXCL10 (eQTL effects and microRNAs (niRNAs); co-regulated gene sets), and disease associations of CXCL10 sequence variants, eQTLs/meQTLs for CXCL10 together with related miRNA variants to learn about its pathophysiological roles. The CO-Regulation Database (CORD) identified 182 concordantly co-expressed genes (at the 5-fold threshold), including five additional CXCL genes in the vicinity of CXCL10, and ten MHC genes, including HLA-DRA/DQA1/DPA1, but not their beta-chain coding counterparts. The most significant co-expression pattern was noted in the spleen followed by other immune-related tissues and cells (neutrophils, thymus, B-cells). The gene set enrichment analysis of the co-expressed gene set on the GSEA/MSigDB tool suggested enrichment of genes involved in breast and prostate cancers, immune and inflammatory response, and response to interferon and regulatory T cells (FDR<1E-50). The co-expressed gene set did not have enrichment for common targets of any miRNA. We selected the miRNAs targeting CXCL10, and determined their other target genes using TargetScan. Eighty miRNAs were identified, and the targets of each miRNA were subjected to the same GSEA analysis. The results were not nearly significant as the co-expressed gene set suggesting that miRNAs do not play a major role in the regulation of CXCL10 expression. Examination of disease associations of SNPs from each miRNA gene region in GWAS databases yielded results for autoimmune (RA, AS, CD, Psoriasis), and for MS, T1D, MG, GD and allergy/asthma at P<1E-04. SNPs within CXCL10 did not show any GWAS associations, but SNPs acting as eQTL/meQTL in blood for CXCL10 showed GWAS associations with longevity, aging, IBD and breast cancer (P<1E-04). Although we did not find strong evidence for miRNA-mediated CXCL10 expression, the variants near miRNAs showed stronger genetic associations with inflammatory and immune disorders. We conclude that the role played by CXCL10 is stronger in autoimmunity, inflammation and possibly cancer than in transplantation.
Phased whole-gene characterization of novel KIR3DL2 alleles using a dual redundant sequencing strategy
K. Lang (Dresden, DE)
The human killer-cell immunoglobulin-like receptor (KIR) genes form a complex genomic locus, with extensive gene copy number variation and allelic diversity. Next to the human leucocyte antigen (HLA) genes, KIR genotypes have been demonstrated to affect disease predisposition and transplantation outcome. To facilitate the selection of donors based on KIR genotypes, we extended the routine workflow for typing of newly registered potential stem cell donors by KIR in 2014. Up to now we have KIR genotyped more than 1.5 million samples based on our next-generation sequencing approach. This resulted in the discovery of hundreds of novel alleles. For the submission of novel alleles, full-length phased sequence characterization is desirable to build a comprehensive knowledge database. Therefore, we initiated projects for high quality full-length sequence characterization of novel KIR alleles starting with KIR3DL2. We designed PCR primers flanking the UTR-regions of KIR3DL2 resulting in a 17.5 kb whole-gene amplicon. Two independent long-range PCR products are sequenced on long-read and short-read sequencing platforms. Long-read-based consensus sequences are polished with short, high-fidelity Illumina sequences using the Dual Redundant Reference Sequencing software (DR2S, DKMS Life Science Lab). Analysis of the generated sequencing data revealed high specificity of the assay for KIR3DL2. The long-read sequencing technologies are capable of generating fully phased sequences for KIR3DL2 spanning the entire 17 kb gene. This enables high-quality phase-defined characterization of novel KIR alleles without the requirement for time consuming cloning steps. Here we demonstrate a method for phased whole-gene KIR characterization based on long-read sequences complemented with high quality short read-data. We have started to apply this approach to submit novel and partially defined KIR3DL2 alleles to the IPD-KIR database.
KIR allele polymorphism investigated by Next Generation Sequencing technology
K. Gagne (Nantes, FR)
The impact of natural killer cell alloreactivity on hematopoeitic stem cell transplantation outcome is still debated due to the heterogeneity of graft parameters, HLA class I environment, nature of KIR/KIR ligand genetic combinations studied and KIR+ NK cell repertoire size. KIR genes are known to be polymorphic in terms of gene content, copy number variation and number of alleles. This allele polymorphism may impact both on the phenotype and function of KIR+ NK cells. We speculate that this polymorphism may alter donor KIR+ NK cell phenotype/function thus modulating post-HSCT KIR+ NK cell alloreactivity. Today, the study of KIR allele polymorphism of KIR genes remains difficult to assess by lack of appropriate techniques. We therefore, developed a Next-Generation-Sequencing method on a MiSeq platform to investigate the KIR allele polymorphisms of all KIR genes. Firstly, we used genomic DNA from well-characterized cell lines to ensure the reliability and specificity of our method. Our results highlighted successful long-range KIR gene amplifications for all reference samples using intergenic KIR primers. Secondly, two different bioinformatics pipelines were used for KIR allele typing. The analysis of sequencing data of all reference samples by using BiRD pipeline and Profiler software demonstrated that all KIR genes were completely sequenced with a sufficient read depth and a high percentage of mapping. KIR allele assignment was feasible for all samples and permits to identify 191 previously un-characterized KIR alleles with standard methods. Comparison of high-resolution KIR typing obtained to those published, reported a high rate of concordance for most centromeric and telomeric KIR genes. In particular, Profiler software revealed a high reliability for KIR allele assignment. Overall, our results showed a reliable NGS-KIR method to investigate the broad KIR allelic polymorphism to improve our knowledge on KIR+ NK cell alloreactivity in HSCT and KIR+ NK cell implication in control of viral infections.
The first genome-wide association study of pemphigus foliaceus reveals new variants associated with differential susceptibility
D. Augusto (Curitiba, BR)
Pemphigus foliaceus (PF) is a neglected autoimmune blistering skin disease characterized by auto-antibodies against desmoglein 1. It occurs sporadically all over the world but is endemic in Brazil and a few other countries, including Tunisia, Peru and Colombia. In Brazil, PF represents a local health problem, with the highest incidence (up to 0.83 cases per 10,000) and prevalence (3.2% in the Terena population) ever reported for an autoimmune disease. We performed a genome-wide association study (GWAS) with 235 patients and 5,658 controls, which have been genotyped with an Illumina platform. After standard quality control procedures, 204,967 single nucleotide polymorphisms (SNP) remained. We performed logistic regression analysis using principal components to correct for possible population stratification. As expected on basis of our previous studies, most of the SNPs that reached genome-wide significance are in the major histocompatibility complex (MHC) class II region. However, our results spotted new markers in MHC and non-MHC genes and intergenic regions that were not seen in previous studies (e.g. an intergenic MHC SNP, OR = 5.3, p = 2 x 10-18; a NOTCH4 variant, OR = 4.2, p = 4 x 10-7, a long non-coding RNA variant at chromosome 7, OR = 0.4, p = 10-5). Our findings help to dissect the associations between MHC genotypes and PF susceptibility. In addition, we identified new genes and non-coding elements that might be involved in PF etiology and suggest new pathways for this complex and poorly understood disease.
Allele-level KIR genotyping of almost 200,000 registry samples – towards a new standard profile for potential blood stem-cell donors
G. Schoefl (Dresden, DE)
The human killer-cell immunoglobulin-like receptor (KIR) family of genes is a key regulator of natural killer cell activity. As well as the the human leucocyte antigen (HLA) genes, KIR genotypes have been demonstrated to affect disease predisposition and transplantation outcome. Owing to the complexity of the KIR locus with extensive gene copy number variation and allelic diversity, high-resolution characterization of KIR genotypes has so far been applied only to smaller cohorts. Here we present a high-throughput KIR genotyping approach that delivers both, copy numbers and allelic information for individual KIR genes, using next generation sequencing of PCR-amplified exonic regions, multiplexed across all genes. Specifically, exons 3, 4, 5, 7, 8 and 9 are sequenced and the resulting reads are assigned to reference KIR alleles. After a series of normalization steps to compensate for differences in starting material and amplification biases among genes and alleles, reads obtained from all amplicons are combined to determine copy number and allelic information. After successful validation, we applied this workflow to nearly 200,000 registry donors and despite the focus on throughput and cost efficiency, we achieved allelic resolution for 83% of all typed KIR genes across samples. For these genes, we report allele diversity and the relative abundance of alleles. We demonstrate that high-resolution KIR genotyping is feasible using a very cost-efficient workflow. This approach enables KIR genotyping for population genetics, disease association studies and other applications requiring large cohorts.
As of October 2016 we have been applying this workflow to the analysis of all volunteer samples registering as potential donors for hematopoietic stem cell transplantation at DKMS. High-resolution KIR genotyping results will thus become available to donor centers and may soon be used as an additional selection criterion to improve transplantation outcome.
Macaque BTN3A genes are conserved orthologs of the human γδ T cell modulators
N. de Groot (Rijswijk, NL)
Butyrophilins (BTN) share structural homology with B7 family members at the extracellular domain level but have in contrast to B7 mostly an extracellular B30.2 signalling part domain. BTN3A play a central role in modulation of γδ T cells, which are highly frequent in gut and mucosal tissues. For example, BTN3A1 activates Vγ9Vδ2 T-cells by phosphoantigen binding, and seems to have diverse additional functions. Since BTN3A genes are not present in the mouse, macaques are an important model species to study functional aspects of these, not yet well understood molecules. In the extended HLA region three BTN3A genes are located; BTN3A1, BTN3A2 and BTN3A3, which could also be defined in rhesus macaques. cDNA analyses showed that all three BTN3A genes are transcribed in PBMC’s and colon tissues. In contrast to humans, rhesus monkeys have an additional gene, BTN3A3-like, which, however, appears to a pseudogene. Genomic DNA and cDNA analysis of 32 rhesus macaques belonging to two families proved that all four genes show modest allelic variation, and at least 12 BTN3A haplotypes could be defined. Analyses of these alleles showed their co-segregation with MHC alleles with the exception of two MHC homozygous individuals, which are heterozygous for BTN3A genes. Our results indicate that rhesus macaque BTN3A genes are localized indeed on the extended MHC as in humans. Additionally, BTN3A transcripts of both species cluster together in the phylogenetic tree, and it can therefore be assumed that the genes represent true orthologs. Based on the comparable function of certain subsets of γδ T cells in human and non-human primates and the high level of similarity of the BTN3A molecules shown here, divers experiments can be thought of as proliferation studies with humanized γδ T cells and inhibition experiments with monoclonal antibodies in cancer research.
Unprecedented Killer Immunoglobulin-like Receptor (KIR) gene polymorphisms in rhesus macaques
J. Bruijnesteijn (Rijswijk, NL)
The killer-cell immunoglobulin-like receptors (KIR) are highly polymorphic immune-regulating receptors, which are expressed on NK-cells and subsets of T-cells in humans and non-human primates. The receptors are classified in different lineages (I, II, III and V) based on their structure. The KIR genes display copy number variation and allelic polymorphism. This makes it challenging to identify all different KIR alleles and define complete haplotypes. Multiple studies have described genomic DNA based KIR haplotypes in different species using the conventional sequencing methods (Sanger and 454 sequencing). In humans, this revealed two haplotypic organisations: the mainly inhibitory haplotype A or the more activating haplotype B. In rhesus macaques, a species used in biomedical research as a model for several human diseases, this haplotypic distribution of KIR genes is absent. Additionally, in humans an expansion of lineage III KIR genes is observed, whereas in macaques lineage II KIR show an extreme expansion. In this study, single-molecule real-time (SMRT) sequencing was performed on the Pacific Biosciences’ (PacBio) Sequel to characterize transcribed lineage II KIRs in out-bred rhesus macaque families. Per animal, about 4,000 to 10,000 full-length reads were obtained. To verify that the complete KIR genotype was identified, the samples were run in duplicate with two different primer pairs. The results confirm an extreme plasticity of the KIR system in macaques. A huge gene content variation is observed resulting in different gene configurations, reflected by gene copy variation and polymorphism of both activating and inhibitory KIRs. Next to identifying novel mamu-KIR alleles (N=45) and two previously uncharacterized genes, we were able to define haplotypes based on full-length cDNA sequences. The results suggest that the KIR gene system, at least in macaques, is even more complex than the MHC gene system, which was always considered to be the most complex gene family in the animal kingdom.