C24-14 The Acute Lung Injury Global Network-Align-for Preclinical Translational Research: An International Collaboration to Identify Phenotypes in Severe Lung Injury Models

Rationale Recent advances in statistical and computational methods have enabled the identification of latent structure within complex biological and clinical data revealing distinct phenotypes of acute respiratory distress syndrome (ARDS) (Sinha AnnuRevMed(2023)74:457-71). Although various animal models for ARDS exist, most currently lack phenotype-specific discrimination. This has led to a significant gap in informing on personalised treatments for phenotype-specific trial. Here we present the newly established ALIGN for Preclinical Translational Research initiative that aims to advance animal models of severe lung injury by incorporating existing and emerging phenotypes, and accurately model clinical lung injury heterogeneity. Methods ALIGN for Preclinical Translational Research is an international collaboration between 15 research institutions across three continents. The network aims to enable translation of human heterogeneity in large animal models will be achieved through: 1)synthesising a constantly expanding large dataset of infectious and non-infectious porcine lung injury models from disparate laboratories; 2) provide shared resources and sufficient sample sizes for investigators to test research hypotheses using complex modelling that was previously unachievable; and 3) provide a regular forum for investigators to present and discuss innovations in large mammal models of lung injury. As preliminary analyses, using data from porcine studies conducted at the Hospital Clinical, Barcelona, Spain, we evaluated the pathogen-specific heterogeneity of host response in three distinct infective models of intra-tracheal installation: streptococcus pneumoniae (N = 40), pseudomonas aeruginosa (N = 83), and methicillin-resistant staphylococcus aureus (N = 18), with similar inoculum doses across experiments. We analysed serum tumour necrosis factor (TNF)-α, interleukin (IL)-1β, IL-6, IL-8, and IL-10, with each biomarker quantified at three timepoints across 24-hour intervals after the initial inoculum. We fitted mixed effects generalised additive models (GAMs) for each biomarker with individualised random intercepts to test for pathogen-specific differences in longitudinal temporal trajectories. Results Biomarker levels diverged across pathogen-specific models at 24 hours for most analytes. For IL-6 and IL-8, these differences diminished by 48-72 hours, whereas TNF-α showed model-specific divergence that persisted through to 72 hours. Mixed-effects GAMs demonstrated that temporal trajectories for each biomarker differed significantly between the three infection models. Notably, individual biomarkers exhibited distinct temporal patterns, suggesting each provided orthogonal temporal information. Conclusions ALIGN for Preclinical Translational Research is a network focused towards rapidly advancing the development of innovative animal models that reflect distinct lung injury phenotypes. Our preliminary data show that circulating biomarkers trajectories in lung injury models of infection have divergent trajectories that are pathogen-specific, recapitulating similar heterogeneity in humans (Sinha NatCommun(2025)22;16(1):7848).