IntroductionPregnancy triggers physiological adaptations across multiple organ systems to support foetal growth and prepare the mother for delivery1. This includes the pancreatic islets of Langerhans, which house insulin-producing β cells and glucagon-producing α cells that regulate glucose metabolism2,3. Insulin facilitates cellular glucose uptake to lower blood glucose levels, while glucagon increases blood glucose levels by promoting glucose release from the liver2,3. As pregnancy progresses, insulin resistance gradually increases, requiring a compensatory rise in insulin secretion to maintain glucose homoeostasis4,5. Compared to pre-pregnancy, insulin secretion increases by 50% in early gestation and 100% by late gestation6. Pregnancy, thus, represents a state of enhanced islet-cell function. Understanding the mechanisms that enable islet plasticity during pregnancy could inform the development of pharmacotherapies to enhance insulin secretion, offering potential novel strategies for diabetes management.Failure to adequately increase insulin secretion in pregnancy leads to gestational diabetes mellitus (GDM), defined as the first occurrence of hyperglycaemia during pregnancy1. GDM is the most common pregnancy-related condition and is associated with materno-foetal complications, such as: large-for-gestational-age offspring, shoulder dystocia, and neonatal intensive care admission. Furthermore, GDM elevates the future risk of type 2 diabetes for the mother and increases the child’s lifelong risk of type 2 diabetes, obesity, and cardiovascular disease7,8. Elucidating the molecular mechanisms underlying pregnancy-related islet adaptations is crucial for understanding GDM.Little is known about how human islets adapt to increase insulin secretion during pregnancy or why they fail to secrete enough insulin in GDM. Studies mapping transcriptomic profiles of pregnant mouse islets have thus far shaped our understanding of global islet changes during pregnancy9,10,11,12. However, changes in mRNA expression do not always correlate with changes in protein levels13. Only one in-depth proteomic study on islets from mouse pregnancy has been conducted, and, to our knowledge, no similar research has used pancreatic tissue from pregnant women14. Characterising the human islet proteome in pregnancy could uncover mechanisms driving pregnancy-related islet adaptations and functional plasticity.Histological studies in humans show that β-cell area increases by 1.4- to 2.4-fold during pregnancy, but the mechanisms behind this expansion remain unclear15,16. Mouse models suggest this process is regulated by lactogenic hormones, such as prolactin and placental lactogen. Lactogenic hormones bind to the prolactin receptor (PRLR) on β cells, activating signalling pathways that promote β-cell proliferation and insulin secretion. Serotonin synthesis within β cells, stimulated by lactogenic hormones, further promotes β-cell proliferation during pregnancy, via the serotonin 2B (5-HT2B) receptor, which is upregulated during mouse pregnancy and downregulated post-partum1,4,5. Additionally, α-cell mass increases in mouse pregnancy due to increases in cell size and replication17. While these mechanisms are well-documented in mice, their role in human islets remains unclear.The limitations of using mouse models to study human pregnancy have been previously described18,19. Moreover, human and mouse islets have several biological differences. They differ in cellular composition, organisation, architecture, size, number, vascularisation, gene expression signatures, and functional responses to stimuli. Importantly, mice β cells exhibit robust proliferative capacity, while adult human β cells are largely quiescent, with replication rates below 0.1%20,21,22,23. These species differences highlight the importance of cautious interpretation when translating findings from mouse models of pregnancy to human biology. Furthermore, whether β-cell replication is crucial to the increased insulin secretion observed during human pregnancy remains controversial5, underscoring the need for studies using human tissues.Difficulty accessing well-preserved pancreatic tissue presents a challenge to research into islet biology. Pancreatic samples can only be obtained postmortem. However, autopsy-derived samples are often degraded due to rapid autodigestion by pancreatic digestive enzymes from the exocrine pancreas, leading to tissue necrosis, disruption of islet architecture, and destruction of islet proteins24. Additionally, studies on human pregnancy present practical and ethical challenges, further limiting access to pancreatic tissue from pregnant women. The Network for Pancreatic Organ Donors with Diabetes (nPOD) at the University of Florida is the largest biorepository of human pancreatic tissue worldwide. It collects pancreata from heart-beating donors, with a cold ischaemia time of less than 24 hours between organ collection and tissue processing. This rapid preservation process minimises postmortem tissue degradation, providing high-quality samples. Among the limited global sources, nPOD offers well-preserved pancreatic tissue from pregnant women, enabling the study of ex vivo islet architecture and protein expression in pregnancy25,26. Laser-capture microdissection (LCM) is a powerful technique for isolating, from complex tissues, minority cell populations that have been preserved in their native environment, allowing for their precise molecular analyses27,28. This method can be applied to pancreatic tissues to isolate islets for further study.We utilise formalin-fixed paraffin-embedded (FFPE) pancreatic tissue sections from nPOD to isolate pancreatic islets by LCM and by liquid-chromatography mass spectrometry (LC-MS/MS) to characterise their protein expression profiles in pregnant women compared to non-pregnant controls. Additionally, we characterise pregnancy-associated changes in whole islet, as well as α- and β-cell metrics. High-resolution imaging of whole tissue sections was combined with powerful automatic image analysis and quantification, to ensure reproducibility and a lack of bias. We validate antibodies for and examine the expression of PRLRs and 5-HT2B receptors in human α and β cells during pregnancy, while also measuring the abundance of glucagon-like peptide-1 (GLP-1) in α cells in pregnancy. By integrating deep proteomic profiling of LCM-isolated human pancreatic islets with high-resolution whole tissue imaging, we provide the most extensive characterisation, to date, of pregnancy-associated molecular changes that occur in human islets.ResultsSuccessful isolation and deep proteomic profiling of human islets and exocrine tissue using LCM and LC-MS/MSIslets and exocrine tissue were successfully isolated from FFPE pancreas tissue sections from both pregnant and non-pregnant donors using LCM (Table 1 and Supp. Data S1). On average 10 islets per donor were analysed (Supp. Data S1). Isolated islets underwent unbiased proteomic characterisation by LC-MS/MS, followed by quantification of the detected proteins (Fig. 1a, 1b and 1c) (Supp. Data S3).Fig. 1: Identification, isolation and analysis of isolated islets and exocrine tissue.a Schematic illustrating the experimental workflow: pancreatic islets were isolated from formalin-fixed paraffin embedded (FFPE) pancreatic tissue sections. Their proteome was characterised by liquid chromatography-mass spectrometry (LC-MS/MS) and between group comparisons analysed. Data analysis included between-group comparisons using permutation-based testing to correct for multiple comparisons using a false discovery rate (FDR) of < 0.05. No statistical data are presented in this figure. Created in BioRender. Stefana, I. (https://BioRender.com/3j35qyz). b Identification of islets by brightfield and fluorescence (488 nm channel) microscopy. Images of islets at 20x magnification before laser-capture microdissection (LCM) are shown. Scale bar = 200 µm. c Brightfield and fluorescence images of islets post-LCM. Scale bar = 200 µm. Images in (b, c) are representative of 12 biological replicate samples included in the analysis, with 9–12 islets dissected per sample, and similar results observed across all islets within each biological replicate. d Number of proteins detected per sample type following LC-MS/MS. e Principal component analysis plot of the samples. Isolated islets are represented as circles (pregnant cases in pink, controls in green), while exocrine tissue samples are shown as squares (pregnant cases in blue, controls in orange). f Identification of islet proteins, volcano plot comparing protein expression between isolated islets (pregnant and non-pregnant controls) and exocrine tissue (pregnant and non-pregnant controls). Significantly upregulated proteins are shown in purple, and downregulated proteins in red. Selected proteins with a log2 fold change > 3 or