Spatial transcriptomic characterization of pathologic niches in IPF.

Despite advancements in antifibrotic therapy, idiopathic pulmonary fibrosis (IPF) remains a medical condition with unmet needs. Single-cell RNA sequencing (scRNA-seq) has enhanced our understanding of IPF but lacks the cellular tissue context and gene expression localization that spatial transcriptomics provides. To bridge this gap, we profiled IPF and control patient lung tissue using spatial transcriptomics, integrating the data with an IPF scRNA-seq atlas. We identified three disease-associated niches with unique cellular compositions and localizations. These include a fibrotic niche, consisting of myofibroblasts and aberrant basaloid cells, located around airways and adjacent to an airway macrophage niche in the lumen, containing SPP1+ macrophages. In addition, we identified an immune niche, characterized by distinct lymphoid cell foci in fibrotic tissue, surrounded by remodeled endothelial vessels. This spatial characterization of IPF niches will facilitate the identification of drug targets that disrupt disease-driving niches and aid in the development of disease relevant in vitro models.

A Minimal Approach to Demonstrate Concordance of Digital and Conventional Microscopy in Toxicologic Pathology.

Digitalization of pathology workflows has undergone a rapid evolution and has been widely established in the diagnostic field but remains a challenge in the nonclinical safety context due to lack of regulatory guidance and validation experience for good laboratory practice (GLP) use. One means to demonstrate that digital slides are fit for purpose, that is, provide sufficient quality for pathologists to reach a diagnosis, is conduction of comparison studies, which have been published both, for veterinary and human diagnostic pathology, but not for toxicologic pathology. Here, we present an approach that uses study material from nonclinical safety studies and that allows for the statistical comparison of concordance rates for glass and digital slide evaluation while minimizing time and effort for the involved personnel. Using a benchmark study design, we demonstrate that evaluation of digital slides fits the purpose of nonclinical safety evaluation. These results add to reports of successful workflow validations and support the full adaptation of digital pathology in the regulatory field.

Diverse potential of secretome from natural killer cells and monocyte-derived macrophages in activating stellate cells.

Chronic liver diseases, such as non-alcoholic steatohepatitis (NASH)-induced cirrhosis, are characterized by an increasing accumulation of stressed, damaged, or dying hepatocytes. Hepatocyte damage triggers the activation of resident immune cells, such as Kupffer cells (KC), as well as the recruitment of immune cells from the circulation toward areas of inflammation. After infiltration, monocytes differentiate into monocyte-derived macrophages (MoMF) which are functionally distinct from resident KC. We herein aim to compare the in vitro signatures of polarized macrophages and activated hepatic stellate cells (HSC) with ex vivo-derived disease signatures from human NASH. Furthermore, to shed more light on HSC activation and liver fibrosis progression, we investigate the effects of the secretome from primary human monocytes, macrophages, and NK cells on HSC activation. Interleukin (IL)-4 and IL-13 treatment induced transforming growth factor beta 1 (TGF-ß1) secretion by macrophages. However, the supernatant transfer did not induce HSC activation. Interestingly, PMA-activated macrophages showed strong induction of the fibrosis response genes COL10A1 and CTGF, while the supernatant of IL-4/IL-13-treated monocytes induced the upregulation of COL3A1 in HSC. The supernatant of PMA-activated NK cells had the strongest effect on COL10A1 induction in HSC, while IL-15-stimulated NK cells reduced the expression of COL1A1 and CTGF. These data indicate that other factors, aside from the well-known cytokines and chemokines, might potentially be stronger contributors to the activation of HSCs and induction of a fibrotic response, indicating a more diverse and complex role of monocytes, macrophages, and NK cells in liver fibrosis progression.

Learning image representations for anomaly detection: Application to discovery of histological alterations in drug development.

We present a system for anomaly detection in histopathological images. In histology, normal samples are usually abundant, whereas anomalous (pathological) cases are scarce or not available. Under such settings, one-class classifiers trained on healthy data can detect out-of-distribution anomalous samples. Such approaches combined with pre-trained Convolutional Neural Network (CNN) representations of images were previously employed for anomaly detection (AD). However, pre-trained off-the-shelf CNN representations may not be sensitive to abnormal conditions in tissues, while natural variations of healthy tissue may result in distant representations. To adapt representations to relevant details in healthy tissue we propose training a CNN on an auxiliary task that discriminates healthy tissue of different species, organs, and staining reagents. Almost no additional labeling workload is required, since healthy samples come automatically with aforementioned labels. During training we enforce compact image representations with a center-loss term, which further improves representations for AD. The proposed system outperforms established AD methods on a published dataset of liver anomalies. Moreover, it provided comparable results to conventional methods specifically tailored for quantification of liver anomalies. We show that our approach can be used for toxicity assessment of candidate drugs at early development stages and thereby may reduce expensive late-stage drug attrition.