Postergroup No. 5/1
Form of presentation:
Internal structure of HLA genes revealed by next-generation sequencing
M. LI (Mountain View, CA, US)
MicroRNAs regulate the expression of genes by binding to the 3'UTR region of these genes. Polymorphism in this region affects the regulatory mechanism of microRNAs. This is particularly the case with HLA-C. The region of the HLA-C gene publicly known and available at the IPD-IMGT/HLA Database does not include the binding site for microRNAs. Next generation sequencing (NGS) allows the analysis of expanded regions of the genome. The microRNA-binding region of the 3’UTR segment of HLA-C was studied using Mia Fora (Immucor) NGS methodology in a large sample set. Seven polymorphic variants were identified in the HLA-C 3’UTR region. Most HLA-C alleles present only one of these genetic variants each, but some of them (C*02:10, C*04:03:01 and C*12:02:02) show multiple variants each. This is an example of the expanded horizons opened by the NGS HLA typing technology that go beyond allele assignment. Other poorly studied regions of the HLA genome are included in this study showing how these regions reveal not only important functional properties of HLA molecules but also phylogenetic cues useful to determine the evolutionary relationship among alleles. These cues from poorly studied low-entropy segments of the HLA genome show the internal structure of HLA genes, which is essential to develop tools to analyse their DNA sequence. Sequence analysis software tools can be enhanced by taking advantage of the internal structure of HLA genes. Specialised knowledge of the HLA environment is one of the keys to the efficiency of these software tools.
Novel HLA-G whole gene amplification approach reveals linkage disequilibrium between extended 3’UTR and HLA-G at the allelic resolution level
J. Drabbels (Leiden, NL)
Several studies have been conducted which assessed the influence of the polymorphic sites present in the regulatory regions of the HLA-G gene on its expression. With regard to the 3 prime untranslated region (3’UTR) of HLA-G, a number of different haplotypes have been defined, including the 14-bp indel region and additional SNPs covering cDNA region 2940-3509. Linkage disequilibrium between the 3’UTR haplotypes and high-resolution HLA-G alleles has been studied by several research groups that are interested in the relevance of HLA-G in clinical transplantation and in pregnancy outcome. However, no HLA-G alleles have yet been submitted to the most recent IPD-IMGT/HLA Database (3.27.0) to provide the fully phased whole HLA-G gene sequence with inclusion of the 3’UTR region. We have developed an HLA-G whole-gene NGS-based typing assay (NGSgo-AmpX HLA-G, GenDx) enabling full phasing over the entire HLA-G region, and allowing the identification of HLA-G alleles at the allelic resolution level including the extended 3’UTR. Here, we present HLA-G typing data obtained for 450 clinical samples, using the NGSgo workflow compatible with Illumina platforms. HLA-G typings were confirmed using the CE-marked HLA-G NGSgo-AmpX assay and NGSengine® software (GenDx). A high, but not absolute, linkage disequilibrium between IMGT-defined HLA-G alleles and the extended 3’UTR region was identified. Full associations using estimated genotype and haplotype frequencies were found between HLA-G*01:01:01:01 and UTR-01, between HLA-G*01:01:01:05 and UTR-04, and between HLA-G*01:04 and UTR-03. In addition, several novel HLA-G alleles were identified.
Quality control in next-generation sequencing HLA Typing
M. LI (Mountain View, CA, US)
The main problem in the implementation of a new truly innovative technology is the lack of proper standards for its validation. Typing results from next-generation sequencing (NGS) cannot be compared to results by current standard methods because these fall short in many ways. In so far as validation of a new technology must rely on external verification, and since there is no current established typing method of comparable coverage and resolution, only a biological system with its own internal logic can meet the demands for proper validation. Reproducibility studies testing the same samples repeatedly do not meet the requirements for analysing alleles in different genetic contexts. Sexual reproduction and family segregation studies are a proper validation. A set of samples of mother-father-child trios is the ideal set to validate NGS HLA typing. These trios present the same alleles in different diploid combinations. A set of 500 such trios was used with Mia Fora (Immucor) revealing the strengths and problems of the methodology and providing the necessary feed-back for tuning up the sequence analysis and allele assignment algorithms. NGS software must have the feed-back control mechanisms to use typing-error data to optimise its analytical tools. In addition, automated sequential control systems in NGS should ideally incorporate the logic of population genetics. One of the main problems was the reference sequence library. Since the IMGT allele sequence library was built in sub-optimal and error-prone conditions it is not surprising to find it unfit to meet the data analysis demands of NGS. These demands include: 1) A realistic catalogue of HLA alleles based on adequate documentation. 2) Allele frequency information in the main broad world populations. 3) Quantitative linkage disequilibrium data specific for each of the main world populations.
Full-length sequencing of a novel MICA allele variant
E. Bauer (Birkenfeld, DE)
The MICA (MHC I chain related gene A) gene is located on chromosome 6 within the HLA region and encodes cell stress–inducible, highly polymorphic cell surface proteins that function as ligands of the activating NKG2D receptor on natural killer (NK) cells. Recent reports describe the importance of MICA matching in the context of hematopoietic stem cell transplantation and therefore prospective typing may facilitate a further improvement in selecting suitable donors. After development of a full-length MICA next generation sequencing (NGS) assay as an integrated module of our NGS HLA typing routine we found a so far undescribed MICA coding sequence variant. This presumably new allele is almost identical to the MICA*008:04 allele except the A to G nucleotide exchange at the position 1087 in exon 6 which leads to an amino acid change in codon 340 from Thr to Ala. Next generation sequencing was performed on a MiSeq instrument with 500 cycles PE V2 chemistry from Illumina. The initially generated MICA full-length amplicons (13 kb) were enzymatically fragmented and adapter ligated according to a standard NEB library preparation protocol. For verification reasons the generated fastq data were analysed with two different sequence analysis tools provided by the software suppliers Omixon (HLA Twin) and GenDx (NGSengine). The subsequent validation of the NGS results was performed by Sanger sequencing all six exons of this sample in forward and reverse directions on a 3730XL DNA Analyzer (Applied Biosystems) followed by an Assign SBT v4.7.1 (Conexio Genomics) sequence analysis. Sanger sequencing confirmed the initially observed nucleotide substitution as novel and so far undescribed MICA sequence variant. Implementing full-length MICA genotyping into HLA routine typing workflows will definitely increase our knowledge about the variability of the MICA locus at reasonable costs without neglecting allelic variants that might be important for donor selection and transplantation outcome.
NGS based HLA typing: comparison of four protocols and corresponding software
P. Neukirchen (Cologne, DE)
High resolution HLA typing results without ambiguities have become increasingly difficult to obtain when applying SBT and PCR-SSP. NGS (next-generation sequencing) based HLA Typing has great potential for generating precise and unambiguous results for multiple samples in a single assay potentially without additional testing to resolve ambiguities. We compared four commercially available reagent kits including their corresponding software (GenDx/NGSgo-Ampx, Illumina/TruSight HLAv2, Immucor/MIAFORA 11 Flex and Omixon/HOLOTYPE HLA). To evaluate the four protocols, we considered the amplification strategy, the target generation, the flexibility of the kits and their CE certification, as well as the need for additional equipment, the practicability of the workflow and the costs. Altogether the same 24 samples (21 clinical and 3EPT) were analyzed using the kits. HLA-A, -B, -C, -DRB1 and -DQB1 were typed. The kits from Illumina/TruSight HLAv2, Immucor/MIAFORA 11 Flex and Omixon/HOLOTYPE also contain amplification primers for HLA-DQA1 and -DPB1. The Immucor/MIAFORA 11 Flex also amplified HLA-DPA1 and DRB3/4/5. Sequences were obtained using the MiSeq Illumina platform and analyzed with the corresponding software NGSengine, Assign truSight, MiaFora and Twin, respectively. The final NGS typing results were compared to known SBT and consensus results of EPT programs and comprised the verification of homozygous results, rare alleles and typing errors. With all four protocols for MiSeq we experienced concordance of the NGS results with the already known typings. Positive practice in the processing (easy to follow work flow, straightforward analysis) was observed. The minimum number of remaining ambiguities depended partly on the alleles detected. Implementation of the NGS workflow promises to be a highly efficient and reliable method in our HLA-typing routine but its cost effectiveness needs to be followed up.
More with less: multiplexed real-time PCR HLA typing using SYBR® Green and multi-peak Tm analysis.
Z. Antovich (South San Francisco, US)
HLA typing plays a critical role in solid organ transplantation. The degree of HLA matching and especially the avoidance of immunogenic antigens in pre-sensitized patients is directly correlated with improved outcomes. As of October 2016, the IPD-IMGT/HLA Database has grown to contain over 15,600 alleles and will certainly continue to expand. Current technologies used to evaluate immune status will also need to expand and broaden their ability to detect allele specific differences that code for immunogenic antigens. Epitope mapping will likely play an important role in this effort and provide a framework to group alleles in a practical way. HLA typing will also need to evolve to split these differences with improved resolution. This study validates a new multiplexing HLA typing method that allows better resolution and flexibility to add additional resolution over time.
The LinkSēq™ typing kit from Linkage Biosciences uses real-time PCR combined with melt curve analysis to define HLA specificity. An evolution of this approach consists of multiplexing several targets into a single reaction and identifying the products by way of distinct melt curve signatures, thereby drastically reducing the number of required reactions. Among the new multiplex assays that were validated, the example presented in this study combines 3 separate reactions targeting HLA-B*15, HLA-DPB1*03 and HLA-DQB1*02 alleles into a single reaction without requiring any protocol changes. Software algorithms were updated to define these new multiplexed results while keeping the same analysis workflow and thus a transparent modification to end-users. This study demonstrates that multiplexing several LinkSēq assays into the same reaction is a practical and reproducible approach and should be applicable across the full spectrum of reactions. In addition to cost and labor advantages, multiplexing represents a leap forward for high-throughput HLA applications where the need for increased resolution is growing.
Validation and routine setting of HLA typing by Next Generation Sequencing using the HOLOTYPE HLA (OMIXON) kits : a multicentric experience
X. Lafarge (Bordeaux, FR)
This study presents the experience of eight French laboratories that utilized the HOLOTYPE HLA kits (OMIXON) for performing HLA typing. This technique necessitates one dedicated PCR per locus. Several kits exist for different numbers of samples (24-96) and loci to type (2, 5, 7 or 11). Two different PCR programs are necessary for class I and II. Amplicon quantification is recommended by the manufacturer. Appropriately diluted amplicons are pooled for each individual and then fragmented, enzymatically, repaired and tagged with indexes. Indexed fragments are pooled and purified with magnetic beads. A selection of the appropriate fragments is performed based on their size by preparative electrophoresis. Afterwards, the library is quantified by qPCR. After appropriate dilution, the library is sequenced on a MiSeq (ILLUMINA). The interpretation of the sequences obtained is done by the HLATwin software using systematically two algorithms. The reliability of the interpreted results is expressed by the calculation of a number of quality criterias. The validation of the kits was performed in all eight laboratories. The sensitivity/specificity was evaluated on 20 EPT samples. The results obtained with the previous methods used in the laboratories (SBT Celera, SSO reverse Luminex One Lambda + PCRSSP OLERUP) were compared on routine samples, including particular samples (new and null alleles). Results revealed excellent concordance with consensual typings obtained in EPT samples, or with results obtained with the other techniques on routine samples. The 4-6 digit resolution is excellent and new alleles are detected. In addition, quantification of the amplicons is not necessary. The only discrepancies observed in the results obtained, were due to new alleles not detected with the previously used technique. They concerned polymorphisms not explored by the previous technique: positions not tested by SSO probes, or exons not routinely sequenced in SBT.
Longitudinal transcriptomic profiling of de novo hepatitis B vaccine recipients
E. Bartholomeus (Edegem, BE)
System vaccinology trials study the influence of a vaccine on gene expression and the resulting characteristics of the immune response by measuring induced antibodies and activated T cells. Not all vaccines are equally effective and there exists a large variability between individual immune responses. In this study we used the hepatitis B vaccine (Engerix-B, GSK) which is in most cases highly immunogenic. However, still 5% of the population are non-responders after the first 3 doses. We recruited 35 individuals without any known previous exposure to the hepatitis B virus or the hepatitis B vaccine. They received the Engerix vaccination on days 0 and 30. For profiling the gene expression changes due to hepatitis B vaccination, stabilized RNA was extracted from blood samples collected on day 0 (before the administration of the first vaccine dose), day 3 and day 7 after dose 1. The RNA samples were prepared using the QuantSeq 3' mRNA-Seq Library Prep (Lexogen GmbH), an alternative for total RNA-seq which copies once the 3’ end of each mRNA strand. The prepared libraries were equimolar pooled and sequenced on an Illumina NextSeq (v2 high output, 150 cycli). We created a bioinformatics pipeline optimized to deal with the 3’mRNA-seq data and processed the resulting reads with DESeq2. When we grouped all data of the principal component analysis, we noted that the variability between individuals was larger than the variation between different time points per individual. Nevertheless, on the individual level we found significant differentially expressed genes (log fold change above 2) in different pathways. The results are consistent with the fact that subunit vaccines, like Engerix-B, tend to have less significant general changes in gene expression than live attenuated vaccines. The next step is to combine the short time gene expression data with the ELISA antibody titres to profile the immune response in order to identify RNA signatures that distinguish responders from non-responders.
Human HLA class II monoclonal antibodies: crucial tools to define antibody epitopes on HLA class II molecules.
C. Kramer (Leiden, NL)
Recent evidence suggests that HLA epitope matching represents an efficient approach to prevent induction of donor-specific antibodies (DSA) after transplantation. For proper epitope matching it is essential to define the most immunogenic epitopes for these DSAs. Human monoclonal antibodies (mAbs) directed against various HLA antigens have been instrumental in defining antibody epitopes on HLA class I molecules. For HLA class II antigens, however, the number of human mAbs unfortunately remains very limited, which explains why the number of antibody verified epitopes on HLA class II molecules is still relatively small. The generation of a larger panel of well-defined HLA class II specific mAbs is therefore required. We decided to set up a system to isolate and produce human mAbs recombinantly. This approach will allow the generation of a well-controlled panel of anti-HLA class II mAbs characterized at the molecular level. A proof of principle experiment was performed with the established human B cell hybridomas WIM8E5 and MUS4H4, specific for HLA-A11 and HLA-A24 respectively. Variable heavy (VH) chain and variable light (VH) chain PCR products of the mAbs were generated and cloned into pcDNA3.3 expression vectors containing the heavy chain constant domain for human IgG1 and either the kappa or lambda light chain constant domain. Subsequently, the mAbs were expressed in Expi293 cells by co-transfection of the VH and VL chain containing vectors. The recombinant human mAbs produced, were tested against luminex single antigen beads and showed similar specificities to the original mAbs. These proof of principle data indicate that it is feasible to clone and produce recombinant human mAbs against HLA. We are currently developing a system to culture flow cytometry-sorted memory B cells that are specific for HLA class II molecules with the aim to generate novel human HLA class II mAbs. These recombinant reagents will then be used for the verification of HLA class II epitopes.