
HLA associations are among the strongest genetic risk factors in immune-mediated disease, yet they rarely explain incomplete penetrance, tissue localization, or opposing effects of different alleles. We use T cell-mediated drug hypersensitivity and multiple sclerosis as complementary mechanistic anchors. Drug hypersensitivity shows that a defined exposure can create or alter a peptide-HLA surface and may recruit pre-existing class I-restricted memory in some early reactions, while allowing naive-derived responses in others. Patch-confirmed abacavir reactions beginning as early as 36 h, together with responses in drug-unexposed HLA-B*57:01 carriers, are difficult to reconcile with obligatory de novo naive priming in every case; in other drug hypersensitivity reactions the relative contributions of memory- and naive-derived cells to clinical injury remain unresolved. Multiple sclerosis adds a lifelong antigen source and separates HLA-DRB1*15:01-associated class II specificity from class I associations with measured Epstein-Barr virus burden and from candidate tissue execution. Together, these systems separate three HLA functions: specificity and focusing, antigen-source control, and tissue execution; and three temporal stages: imprinting and licensing, tissue recall, and local amplification. For T-cell receptor-mediated injury, we propose interacting requirements involving receptor-HLA compatibility, functional ligand and abundance, natural tissue presentation, and local regulation. Within this general architecture, Keystone Epitope Theory proposes that persistent, human-adapted organisms exert disproportionate influence on durable immune memory. That memory-source hypothesis is independently testable and is not a prerequisite for the broader framework. We distinguish antigenic target recognition from inflammatory amplification and propose a lesion-anchored, receptor-first experimental strategy with explicit disconfirming outcomes. Predictive-value calculations distinguish clinical screening decisions from mechanistic inference and show how measured modifiers can motivate competing, experimentally testable causal models.
Syphilis has long been recognized to induce various autoantibodies, yet the prevalence, specificity, and clinical significance of these serological abnormalities remain poorly characterized due to limited sample sizes, incomplete antibody panels, and a lack of longitudinal verification for autoimmune disease (AID) outcomes. In this large-scale retrospective matched case-control study conducted at Zhongshan Hospital of Xiamen University from 2020 to 2023, we evaluated 14 autoantibodies in 1042 syphilis patients and 2084 age- and sex-matched non-syphilis controls, and tracked AID diagnoses through December 2025. The overall autoantibody positivity rate was significantly higher in syphilis patients than in controls (38.5% vs. 23.7%; adjusted OR = 2.02, 95% CI: 1.72–2.37, P < 0.001), with the strongest associations observed for anti-SSB (aOR = 4.57), anti-SSA (aOR = 3.16), and ANA (aOR = 2.60), findings that remained consistent across all age groups and both sexes. During follow-up, 88 syphilis patients (8.4%) developed 16 distinct AIDs, most commonly Sjögren's disease (26.1%) and rheumatoid arthritis (21.6%), with disease-specific autoantibody profiles mirroring those of primary AIDs. Syphilis-associated autoantibodies are not an incidental serological phenomenon but rather a clinically meaningful signal of confirmed autoimmune disease, highlighting the urgent need for autoimmune evaluation in autoantibody-positive syphilis patients.