ABSTRACT. T-cell recognition of antigenic peptides presented by HLA molecules at the cell surface is critical for mounting an efficient adaptive immune response during acute viral infection such as COVID-19 caused by SARS-CoV-2. Recent data suggest that the depth of peptide coverage and the breadth of T cells that are able to respond are both important parameters associated with disease outcome. Strong T-cell responses against SARS-CoV-2 have also been reported in unexposed individuals, pointing to a possible role of heterologous immunity. In this study, we performed immunosequencing of the TCR CDR3β region in a large cohort of 116 alloHSCT recipients and their corresponding healthy donors collected prior to the emergence of SARS-CoV-2. We used bioinformatics analyses and a large database of about 150,000 SARS-CoV-2 specific T-cell sequences in order to investigate the composition of the TCR repertoire regarding the presence of SARS-CoV-2 specific clonotypes in unexposed subjects among the more than 3.5 million CDR3β sequences that we retrieved by immunosequencing. We also performed peptide binding predictions based on the reference proteome of the virus and by using the HLA class I high resolution typing data of the 116 patients. We could show that every individual is equipped with a large and diverse repertoire of clonotypes sharing their CDR3β sequence with a SARS-CoV-2 specific T cell. Furthermore, the composition of the anti-SARS-CoV-2 repertoire was very similar among individuals, in healthy donors but also in the context of immune reconstitution in recipients, despite significant differences previously reported when accounting for the whole repertoire or for CMV-specific clonotypes only. In addition, each individual had the potential to cover a diverse repertoire of SARS-CoV-2 derived peptides (i.e., a few thousands strong and weak binders), but, interestingly, some inter-individual differences were observed when only accounting for a strong affinity level of binding.

ABSTRACT. Massively Parallel Sequencing or Next-Generation Sequencing (NGS) allowed an in-depth evaluation of thousands of whole genomes and transcriptomes. While the NGS data analysis for most genes is now trivial, with many well-documented protocols, some genes require special attention to achieve reliable genotyping and expression-level estimations. This is the case for most HLA genes, where sequence similarity and high levels of polymorphism jeopardize genotyping and expression analysis when using conventional tools. Here we present a new algorithm implemented in the hla-mapper package that supports RNA sequencing (RNA-seq) data. This algorithm minimizes read alignment errors, allowing more accurate downstream analysis such as genotyping, haplotyping, detection of alternative transcripts, and expression level estimations. To assess the hla-mapper RNA algorithm performance, we used simulated transcriptomes to compare the results obtained with the aligner STAR and after the hla-mapper optimization. We also demonstrate that, when using conventional tools for sequence alignment, real transcriptomes present the same pattern observed in the simulated ones: underestimation of the expression levels for most HLA genes, overestimation of others, and many genotype errors, particularly for HLA-B, HLA-C, and HLA-DRB1. Alignment errors are particularly intense in alleles that present many nucleotide differences from the reference genomes. These alignment errors are minimized or prevented with hla-mapper, allowing accurate genotyping and expression-level estimations for HLA genes, including allele-specific expression levels. After optimization, the expression levels in real transcriptomes are very different from the original reported by STAR. Therefore, post-processing transcriptome with hla-mapper is essential for reliable HLA genotypes and expression estimates in RNA-seq data.

ABSTRACT. The Human Leukocyte Antigen (HLA) is a major component in tissue compatibility for solid organs and hematopoietic stem-cell transplantation. The differences between donor and recipient HLA alleles, called HLA mismatches, is a major factor for de novo Donor-Specific Antibodies (dnDSA) development. Several methods have been proposed to evaluate the immunogenicity of the graft and thus the risk of developing dnDSA. Among these, amino-acids matching, and eplet matching showed promising results. However, these methods do not consider the allele immunogenicity in its entirety. In this project, we hypothesized structural differences relate to immunogenicity. We developed a score based on predicted 3D structure superposition for all HLA class I (HLA-A, HLA-B and HLA-C) alleles. In practice, we retrieved all HLA amino acid sequences and performed homology modeling on all class I HLA alleles in multi-templates mode using available solved structures from the Protein Data Bank (PDB). We then superimposed modelled structures two by two resulting in a square symetrical matrix of Root Mean Square Distance (RMSD). For a given pair of donor recipient, the calculated score corresponds to the sum of RMSD of each unknown alleles (by the recipient in solid organ and by the donor in HSCT) for each HLA locus. We compared this score to the epitopic mismatch score and found a partial correlation (R²=0.55). Finally, we validated our score on a kidney transplant cohort by performing cox regression on longitudinal data. Our HLA-3Diff score was strongly associated with class I dnDSA occurrence (FDR=5.3e-5, HR=1.89) as well as rejection (FDR=0.01, HR=1.58). In perspective, we plan to perform the same work on HLA class II alleles and refine our score on the basis of available scientific knowledge. Altogether, this new metric could help matching donors and recipients and minimize the risks of rejection and dnDSA occurence especially for non-familial transplantation in a context of graft shortage.

ABSTRACT. Cytomegalovirus (CMV) is a β-herpes virus that is ubiquitous in all human populations, with worldwide prevalence of seropositivity ranging from 60 to >90%, reaching 80% in some parts of the USA. While both acute and chronic infection are benign for the majority of infected individuals, CMV can cause severe effects in immunocompromised individuals, including transplant recipients. Additionally, CMV has the ability to evade immune response causing a persistent asymptomatic infection. Downregulation of HLA class I expression is one of the mechanisms that CMV uses for evading the immune system. HLA mismatch is also associated with CMV reactivation, with high risk of opportunistic infection after transplant. Here, we evaluated the role of HLA variation with CMV seropositivity in 397,672 individuals registered with Be The Match. We examined 5 HLA loci (HLA-A, HLA-B, HLA-C, HLA-DRB1, HLA-DQB1) for association with CMV serostatus, while adjusting for sex, age, ancestry and socioeconomic status. Eleven risk alleles were identified, with C*03:04 and DRB1*04:01 demonstrating strongest association in European ancestry individuals (OR= 1.07; CI97,5%= 1.04 - 1.10; p= 1.23E-08; and OR= 1.06; CI97,5%= 1.04 - 1.09; p= 3.40E-08 respectively). Ten protective alleles identified included HLA-DRB1*01:03 in both Hispanic and European ancestry (OR= 0.52; CI97,5%= 0.44 - 0.61; p= 2.14E-14; and OR= 0.71; CI97,5%= 0.671 - 0.75; p= 1.35E-32 respectively). Interestingly, HLA-DRB1*01:01 was identified as a risk allele (OR= 1.05; CI97,5%= 1.032 - 1.08; p= 2.36E-06) in individuals with European ancestry. HLA-DRB1*01:01 and -DRB1*01:03 have 98.48% homology, differing in only three amino acids residues at positions 67, 70 and 71 in the peptide binding groove. Despite their overall similarity, these differences appear to mediate opposing effects in CMV seropositivity. Our results provide population-scale evidence of the role of HLA in mediating control of this common and important virus.

ABSTRACT. The HLA class I genes encode key molecules for antigen presentation, recognition of self and non-self-antigens, and the modulation of NK cell activity. HLA polymorphisms influence transplant outcomes and the susceptibility to many autoimmune and infectious diseases. The SNP-HLA Reference Consortium (SHLARC) aims to accelerate immunogenetics studies by improving HLA imputation from SNP data. Imputation accuracy depends on a suitable reference panel. However, HLA genes are highly polymorphic, and the available reference panels are mostly European, resulting in poor accuracy when imputing diverse populations. Here, we tested imputation performance for HLA-A, HLA-B, HLA-C, HLA-E, HLA-F, HLA-G, HLA-H, MICA, and MICB genes. We used a bioinformatics pipeline developed to minimize genotype errors for HLA genes, obtaining the genotypes (SNPs and indels) and the HLA alleles for 4,592 samples from 80 different populations, including 1,170 admixed Brazilians. The generated data were used to build reference panels to predict HLA alleles in 192 samples (test set) extracted from the data but not used for the reference. The F1 score was used to evaluate these models. Our models predicted alleles with an F1 score ranging from 0.80 (HLA-B) to 0.86 (HLA-A) for the classical HLA genes, while for other HLA loci and MIC genes the F1 score ranged from 0.92 (HLA-F) to 1 (HLA-G). We detected a poor performance for HLA-E (F1 = 0.70) due to a misclassification of HLA-E*01:11 to the common HLA-E*01:03 allele in only one sample. Overall, imputation provided a fast and simple method for inferring HLA alleles from SNPs, with accurate results for all genes, including those not commonly evaluated in imputation (e.g., MICA). Imputation errors are mostly related to rare and population-specific alleles, such as Native-American alleles, which emphasizes the importance of larger and more diverse reference panels for improving HLA imputation accuracy.

ABSTRACT. A strong link between the COVID-19 severity and the HLA-C*04:01 allele has been replicated in several Caucasian populations including Armenians. The results have led to an idea that HLA-C*04:01 may affect the immune response via three biological mechanisms: i) disruption of the HLA-C mediated protection harnessing natural killer cells (NK); ii) causing NK hyporesponsiveness through KIR2DL1; or iii) over-activation and exhaustion of CTL and NK cells by stimulating functional KIR2DS4. To test those hypotheses, we re-analyzed HLA-genotypes and RNA-sequencing data of Overmyer et al. [Cell Systems 2021; 12:23-40]. An ordinal regression of patients’ status (i.e. nonCOVID vs. COVID-nonICU vs. COVID-ICU) against HLA-C has corroborated the increase in the disease severity with increasing HLA-C*04:01 dosage (P< 0.003). DESeq2 analyses of the transcriptome (16444 loci) within COVID subset mapped 3586 down-regulated and 4031 up-regulated loci to the disease severity at FDR P<0.05. The results of enrichment analyses of those 7617 genes indicated aberrations in processes, such as T cell activation, inflammatory response, positive regulation of both NK-mediated cytotoxicity and interferon-gamma production. However, only 563 down- and 341 up-regulated loci had nominally associated with the HLA-C*04:01 carriage, reflecting its genetic association with severe symptoms. Using GTEx data and rs5010528 as a proxy for HLA-C*04:01 (R2 = 0.97, 1kG EUR cohort), we found that HLA-C*04:01 was associated with multiple tissues (e.g. lung, heart and blood) RNA expressional and splicing changes in >10 protein-coding loci situated close to HLA-C. The ontology analysis of the loci implicated HLA-C*04:01 in altering antigen processing and presentation of endogenous peptide antigen via HLA class I via ER pathway (FDR P<0.0001), protection from NK-mediated cytotoxicity (P<0.004), and innate immune response to other organisms (P<0.009). The work was supported by the Science Committee of RA (grant E17).

ABSTRACT. HLA genes are strongly associated with transplantation outcome, disease onset and severity, while their comprehensive study utilizing state-of-the-art methods represents one of the major goals of immunology. The introduction of Next Generation Sequencing (NGS) can provide the opportunity to investigate non-coding regions of the HLA genes and to associate them with disease prevalence. MicroRNAs (miRs) represent single-stranded non-coding RNAs that can efficiently modulate the HLA phenotype, through sequence-specific interactions with their corresponding mRNAs. This study aimed to investigate the potential association between the 3’ UTR of HLA class I (A, B and C) genes with specific miRs, that can affect the HLA phenotype. Initially, we used the genetic sequence (indexed in IPD-IMGT/HLA database) of the most frequent HLA class I alleles in Greece, as has been previously by our laboratory, to find potential associations with specific miRs. For this purpose, miRbase and miRDB, were used to search potential miRs related to the 3’ UTR genetic sequence of HLA class I alleles. Moreover, the function of the detected miRs was identified, using the available databases such as Target Scan, miR path v3, mirWalk2.0, miRanda and miRNAmap. The results of this study showed the presence of about 149, 175, and 152 associated miRs with the HLA A, B and C alleles, respectively. For each locus allele, approximately over 30 miRs were differentially detected. Specifically, for the A locus, the hsa-mir-7111-3p, hsa-mir-7111-5p, hsa-miR-4436b-3p, hsa-miR-7114-3p, for B locus, the hsa-mir-125a-3p, has-miR-1207-5p and for C locus, the has-miR-103a-5p and has-miR-4646-3p, were found to play a major role in post-transcriptional HLA gene regulation. Based on the above results, the identification of specific miRs may reveal significant data regarding the HLA class I phenotype modulation, thus playing important role in decision-making for transplantation or disease administration.

ABSTRACT. The Pangenome (PGP) and Thousand Genome Project (1KGP) sequence data provide new opportunities for understanding MHC diversity. Dot matrix analysis of 47 fully phased MHC sequences from the PGP revealed a 60kb DNA indel between HLA-W and HLA-J on 27 of the 47 PGP sequences. A homology search of the 60kb indel sequence with sequences maintained by IMGT/HLA identified a sequence in the indel with 100% homology to HLA-Y*01:01 and a second sequence with 88% identity with HLA-P*01:01:01:01. Subsequent phylogenetic analysis placed the second sequence in a broader cluster with alleles of HLA-P and HLA-W. Assuming this to be a novel HLA Class I sequence we named it HLA-OLI Subsequent analysis of 3202 whole genome sequences from 1KGP with Type Stream Visual optimized for genotyping whole genome sequence data revealed that HLA-Y and HLA-OLI (Y/OLI) were present together in 886 of the 3202 samples. Y/OLI was associated with HLA-A*29:01:01, 30:01:01, 33:01:01, 33:03:01 and 34:01:01, confirming HLA-A~HLA-Y associations of previous studies. We could not confirm an association with HLA-A*02:03 and 02:05 but did identify a possible new association with HLA-A*01:01:01:11. Further analysis suggests a segmental replication structure of the alpha block of the MHC with at least 4 homologous segments that include HLA-V and HLA-P, HLA-H and HLA-T, HLA-A and HLA-W and HLA-Y and HLA-OLI. The HLA-H/HLA-T segment is absent on haplotypes with HLA-A*23 and HLA-A*24 and the HLA-Y/HLA-OLI segment is present only on some haplotypes with the HLA-V/HLA-H segment and HLA-A/HLA-W segment being present on all (or most haplotypes) This study of the MHC using “big data” has provided new insights into the structure and evolution of the MHC alpha block. We have identified a new HLA-Class I sequence, the location of HLA-Y and provide evidence that the MHC alpha block is made up of varying numbers of replicated segments containing HLA class I genes and pseudogenes

ABSTRACT. Most human killer cell immunoglobulin-like receptors (KIR) are expressed by Natural Killer (NK) cells and recognize HLA class I molecules as ligands. Uniquely, KIR3DL3 is a conserved but polymorphic inhibitory KIR recognizing a B7 family ligand, HHLA2, and is implicated for immune checkpoint targeting. KIR3DL3 is also the most genetically diverse KIR, with a subset of polymorphic residues having genetic patterns consistent with the action of diversifying natural selection. Because the expression profile and biological function of KIR3DL3 remained elusive, we searched extensively for KIR3DL3 transcripts, revealing expression is highly enriched in γδ and CD8+ T cells rather than NK cells. These KIR3DL3 expressing cells are rare in the blood and thymus, but more common in the lungs and digestive tract. High resolution flow cytometry and single cell transcriptomics showed that peripheral blood KIR3DL3+ T cells have an activated transitional memory phenotype and are hypofunctional. The TCR usage is biased towards genes from early rearranged TCR-α variable segments or Vδ1 chains. Using calcium flux and luciferase reporter assays, we show TCR-mediated stimulation of γδ T cells can be inhibited through KIR3DL3 ligation. Surface plasmon resonance shows KIR3DL3 binds HHLA2 with similar affinity as KIR3DL1 binds Bw4+HLA class I. Whereas we detected no impact of KIR3DL3 polymorphism on ligand binding, variants in the proximal promoter and at residue 86 can reduce expression. In summary, we demonstrate that KIR3DL3 is upregulated in response to unconventional T cell stimulation and that individuals may vary in their ability to express KIR3DL3. These results have implications for the personalized targeting of KIR3DL3/HHLA2 checkpoint inhibition.

ABSTRACT. Naïve T-cells (TN) are considered the main mediators of alloreactivity after hematopoietic cell transplantation (HCT), prompting clinical trials with TN-depleted allografts to prevent graft-versus-host disease (GvHD). Both minor histocompatibility antigens (mHAg) and HLA-DP mismatches can be alloreactive CD4+ T-cell targets. We have demonstrated that immunopeptidome divergence between mismatched HLA-DP, regulated by the peptide editor HLA-DM, determines the strength and T-cell receptor (TCR) diversity of alloresponses. Here, we hypothesized that immunopeptidome divergence could also influence memory T cell (TM) alloreactivity against HLA-DP. TN (CD45RA+CD45RO-) and TM (CD45RA-CD45RO+) cells from healthy individuals were stimulated with HeLa cells expressing single matched (DPB1*04:01) or permissively (DPB1*04:02) or non-permissively (DPB1*09:01) mismatched HLA-DP allotypes, with/without HLA-DM. In the presence of HLA-DM, TN alloresponses (quantified by CD137 upregulation) against matched (14.4±13.9% vs 5.9±7.7%, p=0.0041) and permissively mismatched (16.7±14.3% vs 3.0±3.3%, p=0.005) HLA-DP were significantly stronger than TM responses. In contrast, TM responses against non-permissive mismatches were strong and not significantly different from TN (40.0±18.1% vs 33.4±22.3%, p=0.08). Using immunopeptidomics and comparative whole-exome sequencing, we show that the absence of HLA-DM significantly increases the number of candidate mHAg presented by HLA-DP (7.9±5.1 to 18.9±7.2, p=0.016), concomitantly increasing the strength and TCR α+β diversity of alloresponses to comparable levels in TN and TM (33.9±14.6% vs 36.3±15.7%, p=ns). In conclusion, TM cells contribute substantially to alloresponses against divergent immunopeptidomes from non-permissive HLA-DP mismatches and deregulated mHAg. These findings bear implications for TN depletion strategies to prevent GvHD, and suggest potential new avenues to harness mHAg alloreactivity in cellular therapy.

ABSTRACT. Complications due to infection with or reactivation of human cytomegalovirus (HCMV) remain a clinically challenging problem in immunocompromised patients. Knowledge of viral targets is critical to improve monitoring of high-risk patients and to optimize antiviral T-cell therapy. We aimed to identify naturally presented HLA-A*11:01-restricted HCMV-derived T-cell epitopes from HCMV infected professional antigen presenting cells (APCs) to expand the spectrum of immunogenic targets. Monocyte-derived and genetically-engineered dendritic cells (DCs) were generated to provide a stable platform for soluble (s)HLA-A*11:01 production. After infection of sHLA-A*11:01-secreting cells with wild type HCMV or a mutant lacking known immune evasion molecules (US2-6+11), sHLA-A11-bound peptides were isolated by w6/32 immunoaffinity chromatography followed by mass spectrometric analysis. More than 50 peptides were identified and screened for their viral protein origin and binding strength to HLA-A*11:01 using established databases. The 25 highest scoring candidates were selected for in vitro evaluation of immunogenicity, cytotoxicity, clinical suitability and relevance, and specific T cells for 5 candidates were detected in healthy CMV+ donors by IFN-γ-EliSpot. Their complex stability was demonstrated by in vitro peptide binding assays. Highly proliferative and cytotoxic memory T cells were detected after stimulation with the UL36-derived A11SAL and UL122-derived A11SVS peptides in healthy CMV+ donors and HCMV-infected patients. Surprisingly, A11SAL-specific memory T cells exhibited functional properties at levels comparable to those T cells against the known immunodominant pp65-derived A02NLV peptide. The A11SAL peptide is a new major target of the anti-HCMV immune response. The newly identified HCMV peptides expand the repertoire of immunodominant targets and will improve strategies for identifying high-risk patients, and enhancing therapeutic options with HCMV-specific T cells.

ABSTRACT. Human Leukocyte Antigens (HLA) present a large repertoire of peptides (immunopeptidome) including epitopes relevant for T cell immunity. A fraction of these peptides originate from alternative open reading frames (ORF) embedded in canonical genes or in long non-coding RNA (lncRNA), or interspersed in intergenic regions (cryptic immunopeptidome). This represents an unexplored reservoir of epitopes potentially useful in cellular immunotherapy of cancer. Here we characterized the cryptic immunopeptidome of class II HLA-DP alltoypes in the monocytic leukemia cell line THP-1. Single HLA-DP expressing THP-1 cell lines were generated by CRISPR/Cas9 knockout of endogenous HLA class II genes followed by transduction with lentiviral vectors encoding 3 frequent prototype HLA-DP antigens: DP10, DP401 and DP402. HLA-DP peptides were eluted after immunoaffinity chromatography and analyzed by tandem Mass Spectrometry. Peptide identification was performed using a THP-1 specific protein database built from the 3-frame in silico translation of transcripts in total RNA sequencing data. A total of 2104, 1963 and 2206 peptides were identified in DP10, DP401 and DP402, with 10.0%, 5.4% and 8.1% derived from cryptic ORFs, respectively. Among all 6273 peptides identified, 117 (1.9%) carried amino acids corresponding to public or THP-1 specific genetic variations, i.e. potential minor histocompatibility antigens or neoepitopes. In addition, 569 (9%) cryptic peptides were identified overall, the majority arising from alternative ORFs in canonical genes (83.8%) or in lncRNA (10.7%). One peptide was derived from the lncRNA CCDC26 whose expression levels have been associated with poor prognosis of AML, and 16 peptides were found from alternative ORFs of 15 tumor associated antigens including TP53, GFI1 and ETV6. In conclusion, we show here that a relevant fraction of the HLA-DP immunopeptidome in THP-1 is of cryptic origin and includes potential new targets for T-cell based immunotherapy.

ABSTRACT. HSCT donor-derived virus-specific T cell (VST) therapy is an alternate prophylactic treatment for pre-emptive treatment of viral infections after allogeneic HSCT. However, HSCT-donor derived T cells have the major limitation of requiring a prolonged time for the generation. Therefore, it is necessary to have off-the-shelf, ready-to-use, third party donor derived VSTs. We have generated and characterized in-vitro adenovirus (AdV), cytomegalovirus (CMV), and Epstein Barr Virus (EBV) - specific T cells from third-party donors. These donors (n=145) were selected based on common HLA alleles namely HLA-A*02, A*24, A*11, B*40, B*15 DRB1*07, DRB1*15 in Indian population and, the seropositive status (n=50) against respective target viruses. The VSTs were generated by co-culturing donor-derived PBMCs with donor monocyte-derived dendritic cells pulsed with virus-specific overlapping peptides namely BLZF1/LMP2/EBNA-1 for EBV VST, with pp65 for CMV-VST, Hexon and Penton for AdV-VSTs. During the standardization of protocols we have compared VSTs generated using both In house and commercial available peptides. The results based on characterization were comparable. Immunophenotyping of the VSTs showed prominent CD3+ CD8+ and CD3+CD4+ subset having effector memory CD45 RA- CCR7 - (TEM), terminal effector memory (TEMRA) CD45 RA+ CCR7 - subsets in all the VSTs generated. Also CD3+ CD8+ and CD3+CD4+ subsets had high percentage of TC1 and Th1 cells. We then functionally characterized the VSTs using IFN gamma secretion assay post antigen stimulation, which was found to be in the range of 38 to 45% for EBV, AdV in the total live cell population and 90% positive for CMV VSTs. Our data suggests that with current approach of generation of VST bank, a broad range of HSCT patient’s population can be targeted for VST adoptive immunotherapy. Further toxicity of VSTs will be analyzed in immunocompromised mice models.

ABSTRACT. HLA compatibility is not commonly considered in the development of other allogeneic IECTs, such as CAR-modified cells. However, HLA matching, or knocking out problematic HLAs, could improve the persistence and duration of efficacy of such “HLA-independent” therapies. However, even in situations where finding HLA compatible starting material is usually easy, there is no need to manufacture “HLA redundant” products. Optimally sizing the bank to maximize population coverage while minimizing manufacturing costs is an open problem. To address this aspect of IECT development, we developed an optimal coverage problem, combined with graph algorithms for donor selection, under different, clinically plausible scenarios where the cost of additional donor recruitment and/or the cost of preventing class I or class II HLA expression through gene editing were of interest. We compared the efficiency of different optimization algorithms – a greedy solution, a linear programming (LP), and integer linear programming (ILP) vs. random donor selection over millions of candidate donors. The additional population coverage per donor decreases with the number of donors. All proposed algorithms consistently achieve the optimal coverage with far fewer donors than the random choice. The number of randomly selected donors required to achieve a desired coverage increases with increasing population. However, when optimal donors are selected, the number of donors required may counterintuitively decrease with increasing population size. Gene editing was generally more expensive than recruiting additional donors. When choosing the donors and patients from different populations, the number of random donors required drastically increases, while the number of optimal donors does not change. Random donors fail to cover populations different from their original populations, while a small number of optimal donors from one population can cover a different population.

ABSTRACT. Identification of HLA antibody utilizes bead-based technology, which offers a snapshot of the antibody profile at the time of collection. However, antibody profiles can fluctuate with time, specifically a current serum profile may not detect historical sensitization. Memory B cells in circulation specific to HLA-alloantigen can be present without detection by SAB assay, though may be differentiated into plasma cells and rapidly produce high affinity HLA specific antibodies. Karahan et al published a protocol for stimulation of memory B cells and characterization for the presence of HLA specific IgG. The aim of this project is to reproduce the findings of Karahan et al in the Western Australian cohort. The study subject was selected based on the history of sensitization. Peripheral blood and serum samples were collected from healthy donors consisting of multiparous women, and individuals who had never been exposed to alloantigens. Specifically, B cells from PBMC were incubated with agonists and growth media for 10 days at 37°C. The stimulated cells were characterised by Immunophenotyping assay using flow cytometry. The supernatant was purified for IgG using a protein G affinity purification method. The purified products were then evaluated for HLA IgG specificity. We observed an expansion of CD19+ B cells in 10 days of stimulation and those cells carried a characteristic antibody-secreting-cell phenotype. In immunized individuals, 3/5 were found to have HLA-specific IgG antibody in concentrated supernatants. The HLA specificities detected in the supernatants were absent in the current serum, and accounted by a single mismatch eplet from the sensitizing partner. In all unsensitized individuals, the supernatant was negative for HLA specificity by SAB assay as expected. Therefore, the stimulation assay was shown to be specific to B cell memory, however, further investigation is required to examine the assay sensitivity and producibility.

ABSTRACT. The bio-ethical regulatory framework applicable to immunogenetics is different in various countries, although common principles and supra national texts do exist; it impacts on medical and scientific collaboration between countries. In France a bioethics law exists since 1994; it covers the use of body elements for research and therapy, stem cells, genetic testing, human genomics, transplantation and other domains without impact on immunogenetics. It has been periodically revised (2004, 2011, 2021). After a public consultation in 2018 and reports from bodies such as the National bioethics committee, the State Council and a parliamentary Commission the revised law was discussed, voted in the parliament and published on 2 August 2021. Some decrees are still pending for the application of the law. However, a good picture of the framework for immunogenetics can be drawn and allows to better organise lawful collaborations, be it for clinical care or for research. Of special interest is the analysis of the measures proposed by citizens and what was kept of them in the final text, as an example of societal dialogue with possible impact on regulations. We will present the relevant modifications for the practice of immunogenetics and their possible consequences on collaborative practices in Europe, for transplantation, genetic testing, genomic medicine and stem cell research. The new measures deal with living cross organ donation, intra-family donation from minors or vulnerable persons, for organs and hematopoietic stem cells, the transmission of genetic information concerning a deceased person, for the medical benefit of a family member, the conditions for revealing incidental findings occurring in the course of testing and not related to the aim of the test prescribed, algorithms and massive data management in the context of medicine, new conditions for research on embryos and stem cells of embryonic origin; an entire chapter of the law deals with genomic medicine.

ABSTRACT. Permissive T-cell epitope (TCE) HLA-DP mismatches (mM) are an established criterion for unrelated donor (UD) selection in hematopoietic cell transplantation (HCT). We have recently demonstrated that frequent permissive mM can be stratified according to their immunopeptidome divergence into core vs non-core subsets. Here, we sought to refine the definition of these permissive subsets and explore how their directionality affects their associations with HCT outcome. HLA-DP matching status was scored in 8420 adult patients who received a 1st HCT from 10/10-matched UD for AML, ALL, or MDS from 2005-2020. TCE group 3 (TCE3) permissive pairs (N=3178) were stratified into core and non-core mM, and compared with allele-matched (N=2390) and non-permissively mM (N=2598) pairs. An extended core allele definition including alleles with 0-5 exon-2 amino acid differences respect to the high-frequency core alleles (DPB1*02:01, 04:01, 04:02) was applied. Multivariable models were tested in parallel to the standard TCE model, including graft-versus-host (GvH) and host-versus-graft (HvG) vectors. In the standard TCE model, non-relapse mortality (NRM) risks were significantly increased for non-permissive (HR 1.3 [99% CI: 1.0-1.5]; p=0.0011) but not for permissive (HR 1.1 [0.9-1.3]; p=0.08) mM compared to allele matches. Both permissive and non-permissive mM were associated with increased acute GvH disease (aGvHD) and lower relapse. Among permissive pairs, only GvH non-core mM showed significantly increased risks of aGvHD (HR 1.5 [1.3-1.8]; p<0.0001) and reduced risks of relapse (HR 0.77 [0.63-0.93]; p=0.0005) compared to allele-matched pairs. Importantly, none of the permissive subsets associated with increased NRM, resulting in better relapse-free survival for non-core GvH mM (HR 0.86 [0.74-1.0]; p=0.0121). Our results confirm the clinical relevance of the refined definition of core alleles, and identify GvH non-core permissive mM as a subset associated with improved disease control.

ABSTRACT. We recently isolated a multiple myeloma-reactive T-cell receptor (TCR) that appeared to be restricted for the HLA null allele C*04:09N. HLA-C*04:09N differs from HLA-C*04:01 in a frameshift within its cytoplasmic domain resulting in 32 additional amino acids that prevent cell surface expression. Based on our observations in multiple myeloma, we hypothesized that HLA-C*04:09N can present antigen sufficient for T cell activation in an appropriate microenvironment. We asked experimentally whether HLA-C*04:09N i) is detectable on the cell surface, and ii) can present peptides to T cells. We generated an HLA class I-deficient osteosarcoma cell line (U2OSHLA-KO) and established single cell clone-derived cell lines that stably expressed HLA-C*04:01 or HLA-C*04:09N. In line with previous observations, HLA-C*04:09N expression was detectable intracellularly but not at the cell surface by flow cytometry and conventional fluorescence microscopy. Yet, we asked whether antigen presentation could still be sufficient for T cell activation. We expressed an HLA-C*04:01-restricted TCR specific for the QYDPVAALF (QYD) epitope in human peripheral blood T cells (TQYD) to probe for antigen recognition in different HLA contexts. TQYD were incubated with i) U2OSHLA-KO or ii) U2OS expressing only HLA-C*04:09N, both previously transfected with a QYD-minigene or a different epitope control minigene. Only U2OS cells expressing HLA-C*04:09N and the QYD-minigene activated TQYD as indicated by interferon-γ and granzyme B secretion. T cell activation was blocked by an HLA class I-specific antibody, which suggests that HLA-C*04:09N-T-cell interaction occurred at the cell surface. In summary, HLA-C*04:09N expression is sufficient for antigen presentation and T-cell activation in vitro. We assume that certain in vivo situations will enable efficient antigen presentation on HLA-C*04:09N and suggest consideration of this allele in clinical context such as allogeneic stem cell transplantation.

ABSTRACT. Reconstitution of T cell immunity after allogeneic stem cell transplantation (aHSCT) is essential towards achieving immune protection and is linked to several transplant related complications. We thought to investigate the spatiotemporal dynamics of T – cell receptor (TCR) repertoire in a cohort of 32 aHSCT recipients and their respective donors. We performed high – throughput TCR complementarity-determining region 3 (CDR3) β sequencing as a measure of T cell immunity on serial peripheral DNA samples at 4 time points during the first – year post – aHSCT. Recipient’s TCR diversity was significantly reduced after aHSCT (p < 0.001) reflected by an increased clonality and reduced TCR richness. This reduced diversity sustained during the entire follow – up period and remained stable. Recipient – donor overlap was persistently low post – aHSCT (median morisita index of 0.04), with increasing similarity of recipient’s repertoire during immune recovery. R+/D+, R+/D- CMV configurations were subject to significant higher clonalities than in R-/D+, R-/D- recipients (p < 0.001) and the clonality within the different groups didn’t change throughout the follow-up. Cumulative frequency and the number of anti – CMV clonotypes didn’t differ pre and post – aHSCT and didn’t significantly correlate with clonality in R+/D+, R+/D- CMV configurations. In conclusion, clonal diversity was stably maintained throughout the time course suggesting changes in the repertoire immediately after transplantation. Repertoire reconstitution mainly consists of de novo clonotype generation and reaches statis by 9 to 12 months post – HSCT. CMV specific clonotypes follow homeostatic stability and expansion of clones in CMV positive recipients seems to be rather driven by a bystander effect than antigenic selection.

ABSTRACT. HLA Evolutionary Divergence (HED), a numerical metric calculated from the amino acid variability in the peptide binding pocket of HLA allotypes representing predicted immunopeptidome diversity, has been associated with outcome of hematopoietic cell transplantation (HCT). Here we explored the impact of HED scores as continuous variable (rather than cohort-specific cut-off values) in HLA-matched HCT. 17,525 adult patients with acute leukemia or myelodysplastic/myeloproliferative disease, transplanted from 10/10 HLA-matched related (N=3,682) or unrelated (N=13,843) donors between 2010 and 2019, were included. (Cause-specific) proportional hazards models including common predictors were used to investigate the association of HLA-A,-B,-C,-DRB1, and -DQB1 locus-specific HED scores as continuous variables with a linear effect, with overall survival (OS), relapse-free survival (RFS), non-relapse mortality (NRM), relapse, acute and chronic graft-versus-host disease (GvHD). The distribution of HED sores was similar in related and unrelated transplants (0-15 for HLA-A,-B,-C; 0-20 for HLA-DRB1,-DQB1). No significant associations were found between HED scores at any locus and any outcome endpoint after related donor HCT. In contrast, in the unrelated donor setting, increasing HED scores at HLA-B (HED-B) were associated with improved OS (hazard ratio [HR] 0.99 per unit difference, p=0.004) and RFS (HR 0.99, p<0.001), as well as lower NRM (HR 0.98, P=0.004) and aGvHD (grade 2-4 (HR=0.99, P=0.014). For RFS, we observed a significant interaction between HED-B and B-Leader status (strongest protective HED-B effect for B-Leader MM). In conclusion, we show a weak but consistent protective association between increasing HED-B scores and survival after HLA-matched unrelated but not related HCT for malignant disease, possibly mirroring effects on T-cell reconstitution and/or B-Leader related mechanisms. These findings will be of aid in patient risk stratification after allogeneic HCT.

ABSTRACT. Hematopoietic stem cell transplant (HSCT) is an effective therapeutic procedure for the treatment of patients (R) with onco-haematological disorders. Improved transplant protocols have brought to increasingly satisfactory results in the field of transplant from haploidentical donors (D). Recently, a hypothesis has emerged that, in addition to DSA (donor specific antibodies), RSA may play an important role in transplant outcome, especially in relation to GvHD. To evaluate the role of RSAs, 51 D/R pairs (41 adults and 10 pediatrics) undergoing haplo- HSCT (2016 -2022) were analyzed. The donors were tested for Ab-HLA with Luminex technology, using an MFI =1000 as positivity cut-off. 23 (46%) were Ab-HLA+. 8/23 (35%) had RSA and 15/23 (65%) were RSA-. aGvHD occurred in 37% of Rs (18/48) while cGvHD occurred in 25% (10/40). 100% of RSA+ recipients (8/8) experienced acute or chronic GvHD vs 40% of RSA- recipients (p<0.01). 75% of RSA+ recipients developed aGvHD vs 30% of RSA- recipients (p=0.04, OR=6.69). 67% of RSA+ recipients developed cGvHD vs 18% of RSA- recipients (p=0.025, OR=8.63). Multivariate analyses confirmed the relationship between RSA and chronic and acute GvHD (p=0.008, p=0.01), as already described in few papers. Interestingly, the only low MFI (959) was found in a son vs. father transplant with aGvHD-I while 3 other cases, with MFI: 17772, 24542 and 16227, had aGvHD-III and cGvHD, aGvHD-II, aGvHD-III and cGvHD, respectively. These data seem to indicate that the level of MFI could correlate with incidence and severity of GvHD, although the result needs to be confirmed in larger cohorts. In addition, all Rs were DSA-, suggesting that RSAs are the key players in an immunological milieu in the D/R pair that could lead to the development of GvHD. It is therefore possible to hypothesize that donor immunization may be involved in B-mediated alloreactivity toward the recipient, leading mainly to the development of cGvHD, but also to aGvHD.

ABSTRACT. Hematopoietic stem cell transplantation from an HLA-haploidentical donor (haplo-HSCT) is now commonly used to treat hematological malignancies in the absence of a fully-matched donor. HLA typing is not considered the first criterion for donor selection (absence of donor specific antibody, age, gender and CMV). HED, a measure of HLA allele diversity, has been proposed to predict response to cancer immunotherapy, solid organ transplant rejection and outcome of HLA-matched HSCT. We retrospectively analyzed the impact of HED on survival and relapse in 107 haplo-HSCT recipients, transplanted between 2014 and 2022 in Saint-Louis hospital, Paris. HED was calculated as genetic distance between pairwise HLA-A, -B, -C, -DRB1, -DQB1 and -DPB1 alleles in recipients (HED-R), donors (HED-D) and mismatched haplotypes (HED-MM). Both 1st quartile and median HED values at each locus were used to stratify recipients, donors and MM as “high” or “low” HED. Survival analysis, using Kaplan-Meier curves, was performed for progression-free (PFS) and overall survival (OS). Cumulative incidence of relapse (CIR) and non-relapse mortality (NRM) were calculated using the method of Fine and Gray with their competing risks (death and disease recurrence, respectively). We found impaired OS in recipients with low (<1st quartile) class I HED (log-rank p=0.014). This effect of low class I HED-R was also significant when analyzing death from relapse (fine-gray test p=0.032) and PFS (p=0.036). No impact of HED-D was found on outcomes. Interestingly, OS and PFS were decreased with a high HED-MM (p=0.051 and p=0.022 respectively), while the number of mismatches was not associated with outcomes. Thus, outcome of haplo-HSCT is not only impacted by class I HED-R, as previously shown for matched-unrelated donor HSCT, but also by the level of class I HLA divergence between the unshared haplotypes. These results suggest that HED-MM could be considered as a selection criterion for haploidentical donors.

ABSTRACT. Hematopoietic Stem Cell Transplants (HSCT) are a curative treatment for many hematologic diseases. Such transplants require high resolution HLA-matched donors. Current HLA matching algorithms depend on imputation which in turn relies on a self-identification of race and ethnicity (SIRE) , which presents a challenge for multi-racial/ethnic individuals or those with unknown race/ethnicity. We present a novel algorithm MR GRIMM “Multi-Race Graph IMputation and Matching” that simultaneously imputes the ethnic group and high-resolution HLA genotypes using the SIRE as a Bayesian prior, and propose a novel method to impute typing inconsistent with current haplotypes. The prediction performance of MR GRIMM was measured using a validation dataset of over 170,000 donor-recipient potentially matched pairs from US registry searches. A higher AUC was observed compared to single-race method on all populations. Recall of the race/ethnic group imputation from the HLA was measured by comparing to ethnic self-definition- using the SIRE. Accuracies of 0.74 and 0.55 were obtained for the prediction of 5 rollup and 21 detailed US population groups using the HLA typing respectively. Improved race/ethnic group designation can improve the accuracy of match predictions as well as increase the accuracy of genetic association studies. We anticipate that the operational implementation of this algorithm in the registry search would help reduce disparity in access to HLA-matched HSCT.

ABSTRACT. Brazilian populations carry large components of European, African, and Native American ancestry. This has implications for finding a compatible donor for bone marrow transplants. Here, we investigate HLA polymorphism in a Brazilian population to understand how admixture impacts the chances of finding a compatible donor for transplant. Using whole genome sequencing data of 2629 individuals from the Recipient Epidemiology and Donor Evaluation Study-III Brazil Sickle Cell Disease Cohort (REDS), we applied the ADMIXTURE and GNOMIX approaches to infer each individual's global genetic ancestry, as well as their local ancestry in the MHC region, and the HLA allelic diversity. Because of the sequence similarity between HLA genes and their high variability, we used a specific approach to make the HLA genotyping and allele calls (HLA-mapper software). The global ancestry proportion of the cohort is 47.9% African, 43.6% European, and 6.7% Native American. There are 87 HLA-A predicted proteins, 129 HLA-B, 69 HLA-C, 158 HLA-DRB1, 89 HLA-DQB1, and 63 HLA-DPB1, some not described in IMGT. To understand the impact of admixture on the chances of finding a donor in REDOME, the Brazilian Marrow Donor Registry, we treated REDS as patients looking for a donor in REDOME. The match proportion decreases markedly when we attempt to match over more loci, going from 80% (6/6) to 1.3% (12/12). Considering IBGE self-reported categories, individuals who self-identify as "white" have a higher average proportion of matches to potential donors than those who identify as "mixed" or "black" (e.g. 16% and 10% for 10/10 match). Also, individuals with higher African ancestry have lower chances of finding a compatible donor when compared to individuals with less African ancestry (e.g. 16% and 7% for a 10/10 match). These results, generated using new and accurate HLA allele calls, confirm the finding that the proportion of African ancestry influences the chances of finding a compatible donor in REDOME.

ABSTRACT. During the first and second waves of coronavirus-19 disease, Sardinia had one of the lowest hospitalization and related mortality rates in Europe. However, in contrast with this evidence, the Sardinia population showed a very high frequency of the Neanderthal risk locus variant rs35044562, considered to be a major risk factor for a severe SARS-CoV-2 disease course. We evaluated 358 patients who had tested positive for SARS-CoV-2 and 314 healthy Sardinian controls (Italy). Patients were divided according to WHO classification: 120 patients asymptomatic, 90 pauci-symptomatic, 108 with a moderate disease course and 40 severely ill. The allele frequencies of Neanderthal-derived genetic variants reported as being protective (rs1156361) or causative (rs35044562) for severe illness were calculated in patients and controls. The Thalassemia variant (rs11549407), the HLA haplotypes, the KIR genes, as well as KIRs and their HLA class I ligand combinations were also investigated. The rs35044562 and rs1156361 Neanderthal variants revealed a distribution in Hardy–Weinberg equilibrium (HWE) both in SARS-CoV-2 patients and the control population (X2HWE = 0.82, p = 0.37 and X2HWE = 0.13, p= 0.72, respectively). Our findings reported an increased risk for severe disease in Sardinian patients carrying the rs35044562 high-risk variant [OR 5.32 (95% CI 2.53-12.01), p<0.0001]. Conversely, the protective effect of the HLA-A*02:01,B*18:01,DRB*03:01 three-loci extended haplotype in the Sardinian population was shown to efficiently contrast the high risk of a severe and devastating outcome of the infection predicted for carriers of the Neanderthal locus [OR 15.47 (95% CI 5.8 – 41.0), p<0.0001]. This result suggests that the balance between risk and protective immunogenetic factors plays an important role in the evolution of COVID-19. A better understanding of these mechanisms may well turn out to be the biggest advantage in the race for the development of more efficient drugs and vaccines.

ABSTRACT. While several studies have been conducted on HLA and Covid-19 infection, few investigated the role of HLA polymorphism on vaccination response. We previously analysed the HLA allele and haplotype frequencies in a hundred weak responders (antibody levels <5th percentile) after mRNA anti-Covid vaccine and found some suggestions about specific alleles and one haplotype [Crocchiolo et al., HLA 2022]. We moved on typing individuals enrolled in the same cohort study (“RENAISSANCE” [Pani et al., Mayo Clin Proc 2021]) with antibody titers above the 95th percentile at six months after vaccination, in order to search for any alleles predictive of strong humoral response. Individuals with clinical history of Covid-19 or positive anti-nucleocapside antibodies were excluded. Allelic frequencies were compared with those of weak responders and of general population, taken from the national bone marrow donor registry, IBMDR. N= 123 evaluable individuals presented with >95th percentile antibody titers at six months after BNT162b2 vaccine. One-third of them had >2080 BAU/mL, lowest value was 1261 BAU/mL. Comparison of allelic frequencies with weak responders showed a significant different proportion of individuals carrying A*03:01, A*24:02 and DRB1*16:01 (21.5% vs. 3.6%, 7.7% vs. 14.9% and 3.2% vs. 8.1% respectively). Moreover, when looking at alleles of the ancestral haplotype A3-B35-C4-DR1, we observed frequencies of 4.47%, 0.84% and 0% in the present cohort, IBMDR and weak responders, respectively. After adjusting for age, gender and BMI, the presence of A*03:01 confirmed to be statistically significant (p<0.0001) and was predictive of high antibody titers at six months with an odds ratio of 12.5 with respect to weak antibody levels. Age was the only other significant variable, with an odds ratio of 0.96. Clinical collection data is underway to correlate Covid-19 infection rates in both cohorts, in an attempt to find a definite correlation between HLA and vaccine protection.

ABSTRACT. The Tshepo study is a clinical trial exploring non-nucleoside reverse transcriptase inhibitor (NNRTI)-based antiretroviral therapy (ART) in HIV-positive patients from Botswana. Here, we propose to explore for the first time the impact of host genetic factors on the development of virological failure (VF) among ART-treated adults in Botswana. Participants were genotyped and after quality control and SNP imputation, 798,232 SNPs were tested for association with VF in 46 cases vs. 521 controls. Logistic mixed models were performed and adjusted for gender, age, CD4 levels, virus load, BMI, and ancestry. Ten SNPs located on the first chromosome reached the genome-wide significance threshold (p-value < 6.26e-8). These SNPs overlapped the NEGR1 promoter in a 200kb high linkage disequilibrium region, which suggests they all derived from the same haplotype. Among the significant SNPs, rs145668083 (p-value=2.31e-8, odd ratio or OR= 44.66) is a 5bp deletion directly located in a regulatory region of NEGR1, suggesting a potential effect on its expression. NEGR1 was previously associated with psychological disorders and BMI, but BMI was similar among patients with and without VF (p>0.05). In addition, four SNPs located near the UBE2H promoter exhibited a suggestive negative association with VF (p-value = 2.09e-6, OR=0.005). UBE2H encodes an E2 ubiquitin ligase involved in MHC-I presentation and UBE2H knockdown was previously shown to disrupt early steps of HIV in vitro replication. This study provides compelling evidence that NEGR1 variations are strongly associated with VF development among ART-treated adults in Botswana, independently from BMI. This is the first report to identify a non-HLA gene associated with VF. Non-adherence to ART is often a multidimensional issue involving HIV mutational resistance but also psychosocial factors (including mental illness) that could explain the association between NEGR1 variants with VF in our cohort.

ABSTRACT. Multiple sclerosis (MS) is a multifactorial disorder characterized by the destruction of the myelin sheath and progressive neurodegeneration. The human leukocyte antigen (HLA) allele HLA-DRB1*15:01 has been consistently demonstrated to be the factor exerting the largest genetic contribution to MS susceptibility, along with other classical HLA alleles. However, the complement system has been suspected to play a role in neuroinflammation due to the presence of complement activation products in MS lesions. The Complement 4 (C4) gene is located within the HLA region and has been associated with experimental autoimmune encephalomyelitis in C4-deficient animal models. However, analysis is difficult due to structural and genetic variation within the HLA region. C4 gene products consist of two isotypes (C4A and C4B) and two isoforms, long (C4L) and short (C4S), owing to a H-ERV insertion. We used our novel bioinformatics tool, C4Investigator, to interrogate the role of C4 copy number variation in MS. Analysis of C4 copy number, controlling for sex and HLA-DRB1*15:01, in affected individuals of European (428) and African (199) ancestry against ancestry-matched healthy controls (438 and 124 respectively) found higher C4L copy number (3+) in European ancestry MS patients compared to healthy controls (p = 6.26e-3, OR = 1.51, 95% CI 1.12-2.03). In comparison, affected individuals of African ancestry had lower C4L copy number (0-1) compared to healthy controls (p = 1.18e-2, OR = 0.48, 95% CI 0.27-0.84). Overall, the median C4L copy number among healthy European individuals was 3, while among African individuals the median copy number was 2, suggesting that copy number differentials in one group may have opposing outcomes with respect to disease depending on ancestry. These findings suggest that the proportion of C4S/C4L may influence disease risk in multiple sclerosis, and that this variation modulates risk differently across ancestral groups.

ABSTRACT. Background: HLA-E is a unique, non – classical HLA class 1 molecule. HLA-E modulates both innate and adaptive immunity via interaction with natural killer (NK) cells and T cell receptors. Tumor cells are reported to overexpress HLA-E. HLA-E interacts with CD94/NKG2A receptors on NK cells and prevents NK cell mediated lysis and contributes to the immune escape of tumor cells. Aims and objectives: The aim of the present study was to analyse the HLA-E protein expression in HPV infected cervical carcinoma. The tumor tissue was obtained from Cervical squamous cell carcinoma patients. Clinically normal tissue was obtained from patients undergoing hysterectomy. Extracted DNA was subjected to Human Papillomavirus (HPV) genotyping. Western blotting was performed on the tissue lysates. The protein intensity data was acquired using the image software and normalized with internal control GAPDH. In addition, peripheral blood samples were collected as well and subjected to ELISA to determine the serum HLA-E levels. Result and conclusions: There were 32 cases of Squamous cell carcinoma of cervix and 43 cases of clinically normal cervix for which tissue samples were available. The high-risk HPV was observed in all cases of carcinoma cervix cases enrolled for the study and the subtypes observed were HPV 16 alone in 31(72%) and rest 12 cases had multiple subtypes. HLA-E expression was significantly increased in cervical cancer tissue as compared to normal cervical tissue (p=0.000). The blood samples were available in 43 cases of carcinoma cervix and 55 cases of normal cervix. The soluble HLA-E levels were significantly increased in carcinoma cervix as compared to normal cervix group (P= 0.038). The study provides support for exploring HLA-E based therapy which would plausibly function by inhibiting the immunoregulatory role of HLA-E.

ABSTRACT. We found a higher incidence of myocarditis in young males who had received Pfizer-BioNTech BNT162b2 vaccinations as compared to historical controls and unvaccinated individuals. The analyses focused on risk following the first and second vaccine in adults and adolescents, as well as risk in adults following the third (booster) vaccine. Males, mainly aged 12−30 years, were found to be at higher risk. However, the question remains what causes lead one specific young male, but not another, to develop post-vaccination myocarditis. The human antigen leukocyte antigen (HLA) is known to play an important role in infectious and auto-inflammatory diseases. We hypothesized that differences in HLA alleles could lead to either protection or susceptibility to vaccination-induced myocarditis. On this basis, HLA typing was performed using next-generation sequencing technology for the HLA-A, -B, -C, -DRB1, -DQB1 and –DPB1 loci, in 21 well-characterized patients who developed myocarditis after the second Pfizer BNT162b2 vaccination. The HLA genotypes were compared with high-resolution HLA data of 272 healthy controls from the Hadassah Bone Marrow registry samples, who are representative of HLA frequencies in the Israeli population. Our findings demonstrated that in HLA class II, DRB1*14:01 (19.04% vs. 5.3%, Pcorr=0.028, OR=4.17), HLA-DQB1*05:03 (19.04% vs. 6.06%, Pcorr=0.034, OR=3.64) and DRB1*15:03 (7.14% vs. 0.0%, Pcorr=0.003, OR=41.76) were significantly associated with disease susceptibility. We further discovered susceptibility motifs in the HLA-DR peptide-binding grooves: His60 (Pcorr0.01, OR=3.52) and Arg70 (Pcorr=0.0047, OR=3.43). Our findings suggest that immunogenetic fingerprints in HLA peptide-binding grooves may have changed the binding affinity of different peptides derived from the Pfizer-BioNTech BNT162b2 vaccination, and induced myocarditis.

ABSTRACT. Pediatric Acute-Onset Neuropsychiatric Syndrome (PANS) is a relatively recently-described condition in which children exhibit sudden and inexplicable behavioral changes characteristic of obsessive-compulsive disorder (OCD) and is often accompanied by disordered eating or difficulty completing daily tasks. PANS has been associated with other autoimmune diseases and is observed most often in patients who have experienced a recent, prior infection or inflammation of the upper respiratory tract. The disease course of PANS has been defined as relapsing-remitting; most of the children impacted by PANS show improved symptoms over time and may relapse after an extended period. Given the apparent immunologic underpinnings in PANS pathology, we sought to elucidate the role of HLA in this little-understood disease. Using our custom next-generation sequencing methods, we genotyped 149 PANS cases with European ancestry for 6 HLA loci (HLA-A, -B, -C, -DRB1, -DQB1, and -DPB1) and compared the results to 2026 previously genotyped ancestry-matched healthy controls. Analysis at the allele and amino acid-levels revealed a significant association of disease with the HLA-B locus, with the risk signal focused on positions 80-83. The amino acids at positions 80-83 comprise the Bw4 epitope, which mediates binding to the inhibitory receptor KIR3DL1 on the surface of natural killer (NK) cells. Of the four significant positions, HLA-B position 80-I (Isoleucine), which is known to increase binding affinity to KIR3DL1, displayed the strongest association (p = 2.63E-05, OR = 1.85, 95% CI = 1.36, 2.48). Our results provide strong evidence for a role of HLA-KIR interaction in PANS, suggesting the involvement of NK-mediated immunopathology in this enigmatic pediatric neurological disease.

ABSTRACT. T cells, and especially cytotoxic T cells are at the forefront of the fight against viral infection. The killer cells are able not only to distinguish between self and foreign peptides, but also to engage in the fight to clear the viral infection by eliminating the infected cells. Our Lab is focused on understanding how T cell engage with viral peptide antigens, that are presented by highly polymorphic molecules called Human Leukocyte Antigens (HLA). T cells have receptors on their surface called T cell receptor (TCR) that allow them to recognise the composite surface of the peptide-HLA complex. Using X-ray crystallography we can understand at the atomic level both peptide antigens presentation and TCR recognition, both important to determine the quality of the subsequent immune response. We can then link those structural information with our cellular assay that determine the strength and magnitude of the anti-viral response, providing the basis for peptide modification to reach stronger response or an understanding of viral mutation that led to viral escape.

Our current work compared the T cell response, at the antigen level against 32 single epitope derived from Spike, between COVID-19 recovered and vaccinated donors. We have shown that the booster shot (3rd dose) increase the antigen-specific T cell response, increase the level of T cell cross-reactivity against variant of SARS-CoV-2, but also alter the phenotype of the T cell. Those results are important to future guide vaccination advise and better understand the immune response to SARS-CoV-2 infection.