ABSTRACT Advances in sequencing technology have enabled extensive identification of polymorphisms in the regulatory regions of the HLA‐E gene. However, the regulatory mechanisms of gene expression for most polymorphic sites remain unclear, limiting research on the associations between HLA‐E genetic variation and disease. In this study, we investigated the regulatory potential of two polymorphic sites in HLA‐E: NT‐26 G/T in the promoter (rs76971248) and NT + 3345 T/C in the 3′ UTR (rs376446052) (nucleotide numbering according to the IPD‐IMGT/HLA Database v3.55 released January 2024). Bioinformatics tools were used to predict transcription factors and miRNAs that may bind to these polymorphic regions. Dual‐luciferase reporter assays were performed to validate the interactions between the HLA‐E promoter and transcription factor HOXA5, as well as between the 3′ UTR and miR‐619‐5p. Additionally, TaqMan probe assays were conducted to analyse the frequency of NT‐26 polymorphism in healthy individuals and solid tumour patients. Bioinformatics analysis predicted a stronger binding affinity of HOXA5 to the mutant HLA‐E promoter (NT‐26 T). Dual‐luciferase assays confirmed that the NT‐26 T variant significantly reduced basal promoter activity compared to the wild‐type (NT‐26 G) ( p < 0.0001). Furthermore, HOXA5 binding induced a more pronounced suppression of transcriptional activity in the mutant promoter ( p < 0.01). For the 3′ UTR, miR‐619‐5p significantly inhibited the transcriptional activity of both the wild‐type and mutant (NT + 3345 C) constructs ( p < 0.05), with no differential effect observed between the two variants ( p > 0.05). Genotyping analysis revealed that the frequency of the NT‐26 T allele did not differ significantly between solid tumour patients and healthy controls ( p > 0.05), contrasting with its previously reported lower frequency in haematological malignancies. In conclusion, the HLA‐E regulatory polymorphism NT‐26 functionally modulates gene expression by altering HOXA5‐mediated transcriptional repression, contributing to a disease‐specific association pattern, particularly in haematological malignancies. In contrast, the NT + 3345 polymorphism does not influence miR‐619‐5p‐dependent regulation. Suggesting that this particular 3′ UTR variant may lack a direct functional role in miRNA‐mediated control of HLA‐E expression. These findings elucidate the mechanistic roles of HLA‐E regulatory variants and provide a foundation for understanding their impact on disease susceptibility.
ABSTRACT Identifying HLA amino acid residues that account for an observed specificity (reactivity profile) is critical for serological categorisation at B‐cell epitope level. Human monoclonal antibody, LB_DQB0501_C (hIgG1l), produced by an individual typed DQA1*01:02 , DQA1*03:01 , DQB1*03:02 , DQB1*06:09 against an immuniser typed DQA1*01:01 , DQA1*01:02 , DQB1*05:01 , DQB1*06:09 , reacts strongly to many but not all DQB1*02 , *03 , *05 and *06 single antigen beads. Within each first‐field molecular typing group, some beads were recognised weakly by the antibody or not at all even at 10 μg/mL. Only one prediction (DQB1_130R185T as a pair) was made using AminoAcid3D (AA3D), an algorithm that defines each HLA molecule as a series of overlapping 3‐dimensional clusters and deduces the potential essential region that could account for an observed reactivity profile. Among various scenarios, DQB1_130R185T was predicted only when the weak signals (TMFI about 1000) were included as positive. Both 130R and 185T, about 10 Å apart, were confirmed as essential for binding using engineered variants as the change of either led to complete loss of binding in both cell and bead assays. Additional residues in the vicinity were also tested for their relative contributions. Besides polymorphic residues in DQB1 at positions 135 or likely 182 that could explain the signal differences within each DQB1*02 or DQB1*05 group, three conserved residues (DQB1_132F, _174H and _183P) were also shown to be critical in securing the binding, indicating multiple residues including conserved are needed to maintain a stable interaction. Epitopes distal to the peptide‐binding domains have not been sufficiently described by the conventional serological typing. Even using HLAMatchmaker, 130R185T was not predicted despite both 130R (verified, level A2) and 185T (unverified) are listed separately as an HLA‐DQ eplet in HLA Eplet Registry. This study provides evidence that AA3D is a useful research tool for epitope analysis. Once a predicted specificity determinant is experimentally verified, both polymorphic and conserved positions nearby should be further investigated for their contributions to the epitope. Such knowledge could become the basis of the ongoing efforts in serological categorisation.
ABSTRACT The highly polymorphic yet homologous nature of HLA antigens (Ags) constitutes a milieu of public and private B‐cell epitopes of varying degrees far exceeding the coverage of the existing cross‐reactive groups (CREGs). To support the on‐going efforts in updating serological categorisation and designing the next generation single antigen bead (SAB) panel of higher resolution and broader population coverage, an algorithm named AminoAcid3D (AA3D) has been developed as a research tool for designing SAB panels and predicting which amino acid residues (termed ‘eps’) could be involved in Ag‐antibody (Ab) interactions according to SAB reactivity profiles. AA3D discerns patterns directly based on Ag sequence and structure, allows size flexibility of contact area and accommodates the uncertainties involving molecular modelling, without pre‐filtering and/or pre‐assigning eps based on polymorphism or predicted attributes such as solvent exposure, protruding index, physicochemical properties, and so forth. Therefore, AA3D enables a more comprehensive approach in capturing various scenarios manifested in the SAB assay. After the eps are indicated using AA3D based on experimentally determined reactivity profiles, further filters can then be applied to rank their probability as the specificity determinants and epitope participants. AA3D provides a powerful research tool for systematic evaluation of spatial parameters that predicts the essential residue(s)/region for specificity of an epitope from a single ep to a cluster of increasing size. Data collected and knowledge gained with AA3D contribute to defining HLA serology. The concepts, hypotheses, algorithms, designs, applications, examples and future directions are described and discussed.
ABSTRACT Identifying HLA amino acid residues that account for an observed specificity (reactivity profile) is critical for serological categorisation at the B‐cell epitope level. Human monoclonal antibody, LB_DQB0402_A, produced by an individual typed DQA1*03:03 , DQA1*05:05/*05:11 , DQB1*03:01 , DQB1*03:01 against an immuniser typed DQA1*01:02 , DQA1*05:05 , DQB1*03:01 , DQB1*06:02 , reacts strongly to DQB1*04/05/06 single antigen beads (SABs), for which 55R, present in DQB1*04/05/06 alleles, has been previously proposed to be the target eplet. In this study, the antibody was also observed to react with DQB1*03 SABs albeit with reduced intensities and only at antibody concentrations higher than 1 μg/mL. Additional weak reactivities to DQB1*02 SABs were also observed when the concentration was further increased to 10 μg/mL. Different clusters responsible for the three concentration‐dependent reactivity profiles were predicted by AminoAcid3D algorithm, and two (DQB1_46V and DQB1_55R) were verified using engineered variants (EVs) derived from DQA1*02:01 ‐ DQB1*06:01 expressed on cells and immobilised on beads, with bead assay more sensitive than cell assay. However, EVs targeting the third cluster (DQB1_81H) did not impact binding, suggesting 81H is not involved in binding to DQB1*04/05/06 when strong interacting 46V and 55R are present. On the other hand, 81H was predicted to account for the weak interactions with DQB1*02 antigens when neither 46V nor 55R are present. As to DQB1*03 , only 46V and 81H are present, so the reduced signals could be due to the missing 55R. This antibody appears to pivot toward the stronger interacting residues available on different antigens. Although HLA alloantibodies develop and affinity mature against a specific foreign conformation, reactivity spread is expected if another antigen shares the same essential residue/region for specificity (ERS) and similar constituents of an epitope. However, this study demonstrates that reactivities in an in vitro setting can still occur even to antigens, including self, not carrying the same ERS as the immuniser, albeit of a much weaker intensity only observed above certain high antibody concentrations. While the clinical relevance of such observations remains unknown, this level of differentiation impacts the analysis of antigenicity.
