Knock-out of multiple classical HLA class II beta chains by one-shot CRISPR-Cas9 mediated genome editing
P. Crivello (Essen, DE)
The class II Human Leukocyte Antigens (HLA-II) HLA-DR, DQ and DP are heterodimers formed by the association of polymorphic alpha and beta chains co-expressed on a variety of immune cell subsets as well as certain tumors. The possibility of modifying any cell type of interest to express only a single HLA-II antigen is desirable for addressing many questions in immunology and transplantation, and could be of potential clinical use in cell therapy. Here we have used CRISPR-Cas9 genome editing to comprehensively knock-out HLA-II beta genes in the B cell line MGAR (HLA-DRB1*15:01, DRB5*01:01, DQB1*06:02, DPB1*04:01). In order to target the different genes simultaneously, we designed a single guide RNA (sgRNA) spanning a conserved region in exon 2 of the HLA-II beta genes. In-silico analysis suggested that neither HLA-II alpha chains nor the processing machinery components DM and DO were targeted. After lentiviral vector transduction encoding the sgRNA, Cas9 and GFP, MGAR cells were sorted for GFP+/HLA-DR- expression and found to be negative also for HLA-DQ and DP, with median fluorescence intensity identical to the isotype control, while expression of HLA class I remained at wild type (WT) levels. HLA-DP expression could be rescued to WT levels by electroporation of the mRNA encoding HLA-DPB1*04:01. This fully restored recognition by HLA-DPB1*04:01-specific alloreactive T cells in CD137 upregulation assays (40.3%, 11.4%, 43.4% T cells responding to MGAR WT, sgRNA-edited, and sgRNA-rescued, respectively). Our results show that CRISPR-Cas9 mediated knock-out of multiple HLA-II beta genes is feasible in B-LCLs, and that expression can be rescued functionally by transfection of a full-length HLA-DP beta chain. This is a potentially attractive new tool for the flexible generation of HLA-II knock-out cells with a variety of possible experimental and clinical applications.
Single Molecule Real-Time (SMRT®) DNA sequencing of HLA genes from 128 International Histocompatibility Workshop cell lines
T. Turner (London, GB)
High-resolution HLA typing has many important clinical applications, including disease association studies and donor selection for transplantation. In 1987, the 10th International Histocompatibility Workshop (IHW) established a panel of 107 B-lymphoblastoid cell lines (B-LCLs), representing diverse ethnicities and HLA haplotypes, which remain some of the best characterised B-LCLs in the immunogenetics field. Pacific Biosciences’ SMRT DNA sequencing generates long reads lengths in isolation, allowing unambiguous high-resolution HLA typing. Here, SMRT DNA sequencing was used to type full-length HLA class I alleles and near full-length HLA class II alleles of 128 B-LCLs, including the 107 IHW panel, with the aim of improving and standardising their HLA typing in the IPD-IMGT/HLA Database. Of 922 allele types assigned, SMRT sequencing matched the existing resolution for 34% of alleles and improved the resolution for 61%. Thirty-seven ambiguous typing results remained for HLA class II alleles where distinguishing polymorphisms are outside our current target region. Thirty discrepancies to previous typing were observed and confirmed with repeat sequencing, including novel intronic polymorphisms (8), differences in zygosity (9), variation outside traditional target regions (8) and potential errors (5). Thirty-five allele types previously unreported in the IPD-IMGT/HLA Database were also determined. With four-field typing we observed clear patterns of linkage disequilibrium between HLA-B and HLA-C alleles, suggesting intronic variation could be used to determine haplotypes. This work has further demonstrated the efficacy and reproducibility of SMRT DNA sequencing to generate allelic level HLA typing, including the discovery of novel polymorphisms in well-characterised B-LCLs. To this end we have improved and maintained the usefulness of these B-LCLs as a resource for the immunogenetics community.
Nanopore sequencing for HLA allele specific RNA expression studies
T. Johansson (Helsinki, FI)
The highly polymorphic human leukocyte antigens (HLA) are crucial in presentation of self, non-self and tumor antigens to T cells, and play a role in autoimmunity, infection responses and transplantation. Expression levels of certain HLA alleles have recently been showm to differ but this has not been studied systematically throughout all HLA genes, cell types and disease conditions. We have developed a targeted RNA-based HLA sequencing method for Oxford Nanopore. Reverse transcription of the whole transcriptome is first performed with template switching where every transcript gets a unique molecular barcode. After full length cDNA amplification the targeted HLA transcripts are further enriched with the 5’ universal primer and HLA gene-specific reverse primers. In this study we performed amplification of HLA-A, -B, -C, -DRA1, -DRB1, DRB3/4/5, -DQA1, -DQB1, -DPA1 and -DPB1 on 50 peripheral blood mononuclear cells (PBMCs). After the enrichment the amplicons are pooled by individual in equal molarities into two HLA gene pools, five loci in each. Nanopore sample barcodes are incorporated by PCR and the sequencing adapters by ligation to both ends of the amplicons. Nanopore 2D sequencing was carried out on Spot-ON flow cells by MinION Mk 1b. During the real-time sequencing the data is uploaded to Metricho
Ultra-high resolution HLA class II typing using the Oxford Nanopore MinION
D. De Santis (Perth, AU)
Single molecule sequencing from Oxford Nanopore MinION offers a major breakthrough for long-read sequencing with great potential for routine diagnostic purposes. The aim of this project is to develop a robust HLA typing assay on the MinION specifically for HLA class II genes namely, HLA-DRB1, -DQA1, -DQB1, -DPA1 and -DPB1. This high-throughput assay would open opportunity for registry and diagnostic laboratories to obtain affordable, reliable and ultra-high-resolution HLA class II typing results.
A panel of 30 samples including most frequent-allele groups and UCLA reference samples was selected for this high-throughput HLA typing assay. The entire full-length HLA-DRB1, -DQA1, -DQB1, -DPA1 and -DPB1 were amplified individually by gene-specific PCR. Amplicons of each sample were pooled and barcoded following the PCR Barcoding and 2D library preparation workflow for MinION. Finally, sequencing was performed on the latest MinION R9.4 flowcell and data was analysed by a combination of HLA assignment software that is restricted to exon analysis. High-quality data (Q-score > 17) was obtained for five HLA class II loci of all the selected samples from multiple MinION sequencing runs. The extracted reads covered the complete region of interest with sufficient read-depth per heterozygous allele. HLA alleles were effectively separated and unambiguous third-field resolution results were obtained. Analysis up to four-field resolution is possible, yet challenging due to insufficient reference data for HLA class II intronic regions. Finally, no allele dropout was identified and 4 discrepancies to previous Sanger results were mainly due to the difference in homopolymer regions of coding region. The assay we have developed on Oxford Nanopore MinION demonstrated that single molecule sequencing can provide accurate and high-quality typing results for HLA class II and suitable for routine diagnostic purposes.
NGS typing results using Oxford Nanopore Sequencing. Can MinION data be reliably used for HLA typing?
E. Rozemuller (Utrecht, NL)
Nanopore sequencing is one of the latest DNA sequencing technologies, in which single molecules are sequenced directly. It offers the potential of read lengths of tens of kilobases. However, the obtained reads are relatively error prone compared to short read technologies. In particular, the length of repetitive sequences such as homopolymers cannot be determined accurately. This hampers the application for accurate HLA typing. Here we will present the typing results for HLA class I samples based on Next Generation Sequencing (NGS) data obtained with the MinION instrument from Oxford Nanopore Technologies (ONT). Using successive iterations of ONT chemistries (7.3, 9, and 9.4) we have mapped the template-complement (2D) reads to the known reference alleles using Burrows-Wheeler Aligner (BWA), which resulted in mappability increasing from 98 to ~99.5% for 2D reads. Additionally, we have used NGSengine® (GenDx) to map the reads to the allele database, increasing the mappability from ~4 to ~70% for 2D reads. The lower mappability with NGSengine is due to the noise present within the data in combination with the thresholds set within NGSengine software. In 80% of the 30 validation samples the best match presented by NGSengine corresponds with the previous typing, however the majority of the samples have exon mismatches. Further inspection of these exon mismatches revealed that the majority involve homopolymers or repetitive sequences. We will discuss the challenges of HLA typing with MinION data, such as the differences in data quality and typing results between 1D (template-only) and 2D data. With the current status of the MinION data quality, the results look promising. Further improvements of the chemistry and/or the base calling algorithm as well as the analysis software are necessary before it can be applied for routine HLA typing.
W/S, X, X2 and Y-boxes sequences of DRB* genes determined by Next Generation Sequencing
M. ALIZADEH (Rennes, FR)
W/S, X, X2 and Y-boxes are conserved cis-acting regulatory elements present within the first 150–300 base pairs upstream of the transcription initiation site of each MHC class II gene and bind to proteins forming the enhanceosome complex. We developed an approach using NGS sequencing from the long range PCR for HLA-DRB1, -DRB3 and -DRB5 typing. All these loci are amplified in a single vial in two different PCRs, one amplifying exon 1 from 5’UTR to intron 1 and the other from intron 2 to 3’ UTR. We developed also software that allows analysis of all loci. We describe here the 5’UTR sequences of these genes, determined by 4 different laboratories on 41 Workshop B lymphoblastoid cell lines: 20 different DRB1 alleles, 4 DRB3 and 2 DRB5 alleles, each present in 1-6 lines. Of those, the 5’UTR has not yet been described for 5 DRB1, 2 DRB3 and the 2 DRB5 alleles tested. Comparing the sequences published (in particular for DRB4) and those obtained we could conclude that these DNA motifs are identical between, all the DRB1*07 and *09 alleles (called DRB1*07 group), all the DRB1*01, *15 and *16 alleles (called DRB1*01 group), all the DRB1*03, *11, *12, *13, *14, *08 alleles (called DRB1*03 group). Only 1 position is different between DRB1*01 and DRB1*03 groups, between DRB5 alleles and DRB1*01 group, 2 between DRB1*04 alleles and DRB1*07 group, 3 between DRB3 alleles and DRB1*03 group respectively. DRB4 and DRB1*04, *07, *09 alleles are more distant. These observations are in line with the evolutionary theory of the “two branches” for DR haplotypes generation, the DR53 haplotype diverging from all the other human DR-haplotypes after the separation of the new and old world monkeys. The DR1 haplotype probably emerged by DRB5 deletion in a DR51 haplotype, and the DRB1*08 gene was created through a gene contraction event in a primordial DR52 haplotype between a DRB1* and a DRB3* gene. These interesting data can be useful for studies on MHC class II expression regulation.
The macaque A2*05 gene encodes for specialist MHC molecules with HLA-B*27-like peptide-binding characteristics
N. de Groot (Rijswijk, NL)
HLA-B*27 is one of the molecules strongly associated with elite control (EC) in HIV-1 infected individuals. EC is also observed in the simian immunodeficiency virus (SIV) infection model in rhesus macaques, an important model to test new HIV-1 vaccine candidates or vaccine components. Rhesus macaques EC are found to express particular MHC class I allotypes, and some, such as Mamu-B*008, have peptide binding characteristics that resemble that of HLA-B*27. The rhesus macaque MHC class I region is expanded. The duplicated A gene, Mamu-A2*05, is abundantly present in macaques, however, the molecules encoded by this gene are characterised by lower cell surface expression and a highly conserved peptide-binding cleft.
We have characterized the peptide-binding motif of Mamu-A2*05:01 by tandem mass spectrometry and elucidated the peptide-binding potential for viral-derived epitopes by a peptide-binding competition assay.
The analyses revealed that Mamu-A2*05 is a specialist MHC class I allotype preferring the basic amino acid arginine at the second position of the peptide, and hydrophobic and polar amino acids at the C-terminal end. These preferences are shared with HLA-B*27 and Mamu-B*008. In contrast, however, Mamu-A2*05 binds preferentially 8-mer peptides. We showed that the lower cell surface expression is due to retention in the endoplasmic reticulum. Peptide-binding studies demonstrated that Mamu-A2*05 is able to bind SIV-epitopes that in the context of Mamu-B*008 evoke a strong CD8+ T-cell response potentially contributing to EC.
In conclusion, Mamu-A2*05 is a specialist MHC class I molecule sharing features with HLA-B*27, and most likely is transported to the cell surface only when the appropriate peptides become available.
Alloreactivity against HLA-DPB1 molecules arises from diverse T-cell receptor repertoires in the permissive and non-permissive context
E. Arrieta Bolaños (Essen, DE)
Permissiveness of HLA-DPB1 mismatches based on the T-cell epitope (TCE) model is now a well-established phenomenon in stem cell transplantation, but its biological basis is still not completely understood. We hypothesized that permissiveness might affect the breadth of the TCR repertoire, reflecting tolerance to shared TCE in the permissive context. To test this hypothesis, we evaluated the TCR repertoires of total CD4+ cells from three self-TCE3/3 individuals responding against a non-permissive TCE1 allele (DPB1*09:01) and a permissive TCE3 allele (DPB1*02:01) expressed on transfected HeLa cells via TCR-Vβ cytometry and immunosequencing. Flow-cytometric analysis revealed responsiveness (CD137+CD4+ T-cells) from all Vβ families against both alleles in all individuals. However, overall fold expansions of specific Vβ correlated with CD137 positivity, and most families responded with higher levels to DPB1*09:01 than to DPB1*02:01. Immunosequencing of TCRB-CDR3 of sorted CD137+CD4+ cells showed that both permissive and non-permissive CD137+ samples had a lower repertoire richness (Daley-Smith estimation) than pre-culture samples. In addition, permissiveness did not affect clonality, with a median of 0.39 [0.36-0.45] and 0.39 [0.30-0.46] against DPB1*02:01 (permissive) and DPB1*09:01 (non-permissive), respectively. There was little overlap between repertoires responding against DPB1*09:01 and DPB1*02:01 within the same individual (median Morisita index: 0.12 [0.005-0.68]), and no overlap against the same allele across individuals. Analysis of the 10 most frequent TCRB-CDR3 sequences revealed almost no sharing between cultures, and comparable cumulative frequencies for DPB1*02:01 (55-67%) and DPB1*09:01 (58-62%). These preliminary results show the uniqueness and heterogeneity of TCR alloresponses and, for the first time, that permissiveness of HLA-DP mismatching does not significantly impact the breadth of the total CD4+ TCR repertoire responding to an allogeneic DP molecule. Future work in a larger cohort and with other HLA-DPB1 alleles in order to further dissect the relationship between TCR repertoire and HLA-DP permissiveness is guaranteed.