Type 1 diabetes results from an autoimmune-mediated destruction of pancreatic β-cells, yet considerable heterogeneity exists regarding disease onset, progression and severity. Emerging evidence suggests that this variability reflects not only differences in immune responses, but also differences in the architecture and immune landscape of the endocrine pancreas. Our recent Steve Morgan Foundation Grand Challenge studies have provided important 'lessons from nature', revealing that the endocrine tissue in the human pancreas exists as a continuum of endocrine objects (EOs), ranging from isolated endocrine cells and small clusters to large islets. These structures differ in their cellular composition, vascularisation, innervation and maturity, creating distinct β-cell microenvironments that evolve throughout life and may shape susceptibility to autoimmune attack. Strikingly, β-cells in small EOs are preferentially lost in type 1 diabetes, suggesting EO size and organisation contribute to β-cell vulnerability. In parallel, we are advancing our understanding of changes in classical and non-classical human leukocyte antigen (HLA-I) expression in type 1 diabetes which can impact β-cell visibility to autoreactive T cells, and have demonstrated that soluble forms of HLA-I can be released from β-cells under inflammatory conditions associated with type 1 diabetes. Taken together, our findings support a model in which disease progression is shaped by interactions between a heterogeneous immune response and a heterogeneous endocrine landscape that can vary across the life course. Improving our understanding of how endocrine architecture, immune niches and HLA-mediated antigen presentation interact could identify new opportunities to protect and expand β-cell populations through targeted therapies in the future.