The presence of missing values (MVs) in label-free quantitative proteomics greatly reduces the completeness of data. Imputation has been widely utilized to handle MVs, and selection of the proper method is critical for the accuracy and reliability of imputation. Here we present a comparative study that evaluates the performance of seven popular imputation methods with a large-scale benchmark dataset and an immune cell dataset. Simulated MVs were incorporated into the complete part of each dataset with different combinations of MV rates and missing not at random (MNAR) rates. Normalized root mean square error (NRMSE) was applied to evaluate the accuracy of protein abundances and intergroup protein ratios after imputation. Detection of true positives (TPs) and false altered-protein discovery rate (FADR) between groups were also compared using the benchmark dataset. Furthermore, the accuracy of handling real MVs was assessed by comparing enriched pathways and signature genes of cell activation after imputing the immune cell dataset. We observed that the accuracy of imputation is primarily affected by the MNAR rate rather than the MV rate, and downstream analysis can be largely impacted by the selection of imputation methods. A random forest-based imputation method consistently outperformed other popular methods by achieving the lowest NRMSE, high amount of TPs with the average FADR < 5%, and the best detection of relevant pathways and signature genes, highlighting it as the most suitable method for label-free proteomics.

Glucocorticoids (GCs) are efficacious drugs used for treating many inflammatory diseases, but the dose and duration of administration are limited because of severe side effects. We therefore sought to identify an approach to selectively target GCs to inflamed tissue. Previous work identified that anti–tumor necrosis factor (TNF) antibodies that bind to transmembrane TNF undergo internalization; therefore, an anti-TNF antibody-drug conjugate (ADC) would be mechanistically similar, where lysosomal catabolism could release a GC receptor modulator (GRM) payload to dampen immune cell activity. Consequently, we have generated an anti–TNF-GRM ADC with the aim of inhibiting pro-inflammatory cytokine production from stimulated human immune cells. In an acute mouse model of contact hypersensitivity, a murine surrogate anti–TNF-GRM ADC inhibited inflammatory responses with minimal effect on systemic GC biomarkers. In addition, in a mouse model of collagen-induced arthritis, single-dose administration of the ADC, delivered at disease onset, was able to completely inhibit arthritis for greater than 30 days, whereas an anti-TNF monoclonal antibody only partially inhibited disease. ADC treatment at the peak of disease was also able to attenuate the arthritic phenotype. Clinical data for a human anti–TNF-GRM ADC (ABBV-3373) from a single ascending dose phase 1 study in healthy volunteers demonstrated antibody-like pharmacokinetic profiles and a lack of impact on serum cortisol concentrations at predicted therapeutic doses. These data suggest that an anti–TNF-GRM ADC may provide improved efficacy beyond anti-TNF alone in immune mediated diseases while minimizing systemic side effects associated with standard GC treatment.

Adalimumab but neither etanercept nor certolizumab-pegol has been reported to induce a wound-healing profile in vitro by regulating macrophage differentiation and matrix metalloproteinase expression, which may underlie the differences in efficacy between various TNF-α inhibitors in impaired wound healing in patients with hidradenitis suppurativa, a chronic inflammatory skin disease. To examine and compare the efficacy of various TNF inhibitors in cutaneous wound healing in vivo, a human TNF knock-in Lepr^db/db mouse model was established to model the impaired cutaneous wound healing as seen in hidradenitis suppurativa. The vehicle group exhibited severe impairments in cutaneous wound healing. In contrast, adalimumab significantly accelerated healing, confirmed by both histologic assessment and a unique healing transcriptional profile. Moreover, adalimumab and infliximab showed similar levels of efficacy, but golimumab was less effective, along with etanercept and certolizumab-pegol. In line with histologic assessments, proteomics analyses from healing wounds exposed to various TNF inhibitors revealed distinct and differential wound-healing signatures that may underlie the differential efficacy of these inhibitors in accelerating cutaneous wound healing. Taken together, these data revealed that TNF inhibitors exhibited differential levels of efficacy in accelerating cutaneous wound healing in the impaired wound-healing model in vivo.

Understanding the heterogeneity of Rheumatoid Arthritis (RA) and identifying therapeutic targets remain challenging using traditional bulk transcriptomics alone, as it lacks the spatial and protein-level resolution needed to fully capture disease and tissue complexities. In this study, we applied Laser Capture Microdissection (LCM) coupled with mass spectrometry-based proteomics to analyze histopathological niches of the RA synovium, enabling the identification of protein expression profiles of the diseased synovial lining and sublining microenvironments compared to their healthy counterparts. In this respect, key pathogenetic RA proteins like membrane proteins (TYROBP, AOC3, SLC16A3, TCIRG1, and NCEH1), and extracellular matrix (ECM) proteins (PLOD2, OGN, and LUM) showed different expression patterns in diseased synovium compartments. To enhance our understanding of cellular dynamics within the dissected regions, we further integrated the proteomic dataset with single-cell RNA sequencing (scRNA-seq), and deduced cell type enrichment, including T cells, fibroblasts, NK cells, myeloid cells, B cells, and synovial endothelial cells. By combining high-resolution spatial proteomics and transcriptomic analyses, we provide novel insights into the molecular mechanisms driving RA, and highlight potential protein targets for therapeutic intervention. This integrative approach offers a more comprehensive view of RA synovial pathology, and mitigates the limitations of traditional bulk transcriptomics in target discovery.

Idiopathic pulmonary fibrosis (IPF) is a progressive lung disease characterized by repetitive alveolar injuries with excessive deposition of extracellular matrix (ECM) proteins. A crucial need in understanding IPF pathogenesis is identifying cell types associated with histopathological regions, particularly local fibrosis centers known as fibroblast foci. To address this, we integrated published spatial transcriptomics and single-cell RNA sequencing (scRNA-seq) transcriptomics and adopted the Query method and the Overlap method to determine cell type enrichments in histopathological regions. Distinct fibroblast cell types are highly associated with fibroblast foci, and transitional alveolar type 2 and aberrant KRT5-/KRT17+ (KRT: keratin) epithelial cells are associated with morphologically normal alveoli in human IPF lungs. Furthermore, we employed laser capture microdissection-directed mass spectrometry to profile proteins. By comparing with another published similar dataset, common differentially expressed proteins and enriched pathways related to ECM structure organization and collagen processing were identified in fibroblast foci. Importantly, cell type enrichment results from innovative spatial proteomics and scRNA-seq data integration accord with those from spatial transcriptomics and scRNA-seq data integration, supporting the capability and versatility of the entire approach. In summary, we integrated spatial multi-omics with scRNA-seq data to identify disease-associated cell types and potential targets for novel therapies in IPF intervention. The approach can be further applied to other disease areas characterized by spatial heterogeneity.