Performances of the deep-learning cell detection model in TCGA cohorts (oral cavity, uterine cervix and larynx SCC) and GR cohorts (oral cavity SCC)
Representative video of granulomas made of multiple MGC in oral cavity SCC from GR cohort
Psoriasis is a systemic inflammatory skin disorder with a prevalence of 2% in adults. Up to 30% of affected individuals further develop psoriatic arthritis (PsA), which is characterized by additional joint inflammation. Myeloperoxidase (MPO) is strongly expressed by neutrophils and, to a lesser extent, also by other myeloid cells. MPO converts hydrogen peroxide to secondary reactive oxygen species (ROS) and is thus primarily considered to induce tissue damage. However, recent studies suggest a protective role of MPO in psoriatic diseases. We aimed to investigate the role of MPO in PsA using the mouse model of mannan-induced PsA. MPO-deficient (Mpo-/-) mice showed exacerbated skin inflammation, joint swelling, and bone degradation associated with increased infiltration of neutrophils, classically activated macrophages, and T cells as well as increased inflammatory cytokine expression in the affected tissues. In the absence or blockade of MPO, in vitro neutrophil stimulation resulted in reduced NET formation and enhanced degranulation characterized by increased neutrophil elastase (NE) activity. In addition, in vitro differentiated macrophages from Mpo-/- mice showed increased interleukin (Il)-6 mRNA expression. Altogether, our findings suggest that MPO controls inflammatory responses in PsA, at least in part, by reducing neutrophil degranulation and serine protease release and, putatively, by reducing inflammatory cytokine production by macrophages.
[This corrects the article DOI: 10.3389/fimmu.2016.00557.].
Rheumatoid arthritis (RA) and other inflammatory arthritis are systemic diseases that primarily affect the joints, characterized by synovial inflammation and progressive cartilage and bone degradation. The temporomandibular joint (TMJ) is reported to be involved in over 50% of RA cases, often leading to severe jaw pain and compromised oral function. Despite its prevalence, TMJ involvement is often underestimated, and its cellular and molecular mechanisms remain poorly understood. Due to the unique biological and functional properties of the TMJ, inflammatory pathways observed in other joints such as the well-studied ankle joint may not directly apply to the TMJ. This study aimed to establish a reliable inflammatory arthritis model for investigating TMJ-specific pathomechanisms. The human TNF-α transgenic (hTNFtg) mouse model effectively replicated TMJ pathology seen in arthritic patients, including increased synovial inflammation (p=0.0024) and severe bone loss (p=0.009) as compared to control mice assessed by micro-computed tomography and histomorphometry. These changes were driven by increased osteoclast numbers (p=0.0331) and upregulation of genes associated with bone resorption such as Acp5 (p=0.0003) and Ctsk (p=0.0025). Notably, we observed that the TMJ displays a unique pattern of immune cell infiltration and pro-inflammatory cytokine expression compared to the ankle joint, particularly with respect to T cell recruitment. These findings were further supported by bulk RNA sequencing, which revealed overall increased inflammation in both the ankle joint and TMJ of hTNFtg mice compared to the control group. Interestingly, while the expression of immune cell and pro-inflammatory cytokine-related gene sets was higher in the ankle joint, the TMJ showed increased expression of genes associated with energy consumption and bone resorption-related enzymes. These findings highlight the TMJ as a distinct anatomical site with heightened susceptibility to arthritis-related damage and emphasize the need for greater awareness and targeted research to improve disease management for affected individuals.
Computed tomography (CT) is routinely used for three-dimensional non-invasive imaging. Numerous data-driven image denoising algorithms were proposed to restore image quality in low-dose acquisitions. However, considerably less research investigates methods already intervening in the raw detector data due to limited access to suitable projection data or correct reconstruction algorithms. In this work, we present an endto- end trainable CT reconstruction pipeline that contains denoising operators in both the projection and the image domain and that are optimized simultaneously without requiring ground-truth high-dose CT data [1]. In addition to experiments with shallow convolutional neural networks, we use trainable bilateral filter layers as known denoising operators [2]. These custom filter layers only require gradient-based optimization of four parameters, each with well-defined effect on the filtering operation. Our experiments reveal that including an additional projection denoising operator in the CT reconstruction pipeline improved the overall denoising performance by 82.4–94.1