Domain generalization across tumor types, laboratories, and species - Insights from the 2022 edition of the Mitosis Domain Generalization Challenge

Recognition of mitotic figures in histologic tumor specimens is highly relevant to patient outcome assessment. This task is challenging for algorithms and human experts alike, with deterioration of algorithmic performance under shifts in image representations. Considerable covariate shifts occur when assessment is performed on different tumor types, images are acquired using different digitization devices, or specimens are produced in different laboratories. This observation motivated the inception of the 2022 challenge on MItosis Domain Generalization (MIDOG 2022). The challenge provided annotated histologic tumor images from six different domains and evaluated the algorithmic approaches for mitotic figure detection provided by nine challenge participants on ten independent domains. Ground truth for mitotic figure detection was established in two ways: a three-expert majority vote and an independent, immunohistochemistry-assisted set of labels. This work represents an overview of the challenge tasks, the algorithmic strategies employed by the participants, and potential factors contributing to their success. With an F1 score of 0.764 for the top-performing team, we summarize that domain generalization across various tumor domains is possible with today's deep learning-based recognition pipelines. However, we also found that domain characteristics not present in the training set (feline as new species, spindle cell shape as new morphology and a new scanner) led to small but significant decreases in performance. When assessed against the immunohistochemistry-assisted reference standard, all methods resulted in reduced recall scores, with only minor changes in the order of participants in the ranking.

Tracking cell turnover in human brain using 15N-thymidine imaging mass spectrometry

Microcephaly is often caused by an impairment of the generation of neurons in the brain, a process referred to as neurogenesis. While most neurogenesis in mammals occurs during brain development, it thought to continue to take place through adulthood in selected regions of the mammalian brain, notably the hippocampus. However, the generality of neurogenesis in the adult brain has been controversial. While studies in mice and rats have provided compelling evidence for neurogenesis occurring in the adult rodent hippocampus, the lack of applicability in humans of key methods to demonstrate neurogenesis has led to an intense debate about the existence and, in particular, the magnitude of neurogenesis in the adult human brain. Here, we demonstrate the applicability of a powerful method to address this debate, that is, the in vivo labeling of adult human patients with 15N-thymidine, a non-hazardous form of thymidine, an approach without any clinical harm or ethical concerns. 15N-thymidine incorporation into newly synthesized DNA of specific cells was quantified at the single-cell level with subcellular resolution by Multiple-isotype imaging mass spectrometry (MIMS) of brain tissue resected for medical reasons. Two adult human patients, a glioblastoma patient and a patient with drug-refractory right temporal lobe epilepsy, were infused for 24 h with 15N-thymidine. Detection of 15N-positive leukocyte nuclei in blood samples from these patients confirmed previous findings by others and demonstrated the appropriateness of this approach to search for the generation of new cells in the adult human brain. 15N-positive neural cells were easily identified in the glioblastoma tissue sample, and the range of the 15N signal suggested that cells that underwent S-phase fully or partially during the 24 h in vivo labeling period, as well as cells generated therefrom, were detected. In contrast, within the hippocampus tissue resected from the epilepsy patient, none of the 2,000 dentate gyrus neurons analyzed was positive for 15N-thymidine uptake, consistent with the notion that the rate of neurogenesis in the adult human hippocampus is rather low. Of note, the likelihood of detecting neurogenesis was reduced because of (i) the low number of cells analyzed, (ii) the fact that hippocampal tissue was explored that may have had reduced neurogenesis due to epilepsy, and (iii) the labeling period of 24 h which may have been too short to capture quiescent neural stem cells. Yet, overall, our approach to enrich NeuN-labeled neuronal nuclei by FACS prior to MIMS analysis provides a promising strategy to quantify even low rates of neurogenesis in the adult human hippocampus after in vivo15N-thymidine infusion. From a general point of view and regarding future perspectives, the in vivo labeling of humans with 15N-thymidine followed by MIMS analysis of brain tissue constitutes a novel approach to study mitotically active cells and their progeny in the brain, and thus allows a broad spectrum of studies of brain physiology and pathology, including microcephaly.

The specific DNA methylation landscape in focal cortical dysplasia ILAE type 3D

Focal Cortical Dysplasia (FCD) is a frequent cause of drug-resistant focal epilepsy in children and young adults. The international FCD classifications of 2011 and 2022 have identified several clinico-pathological subtypes, either occurring isolated, i.e., FCD ILAE Type 1 or 2, or in association with a principal cortical lesion, i.e., FCD Type 3. Here, we addressed the DNA methylation signature of a previously described new subtype of FCD 3D occurring in the occipital lobe of very young children and microscopically defined by neuronal cell loss in cortical layer 4. We studied the DNA methylation profile using 850 K BeadChip arrays in a retrospective cohort of 104 patients with FCD 1 A, 2 A, 2B, 3D, TLE without FCD, and 16 postmortem specimens without neurological disorders as controls, operated in China or Germany. DNA was extracted from formalin-fixed paraffin-embedded tissue blocks with microscopically confirmed lesions, and DNA methylation profiles were bioinformatically analyzed with a recently developed deep learning algorithm. Our results revealed a distinct position of FCD 3D in the DNA methylation map of common FCD subtypes, also different from non-FCD epilepsy surgery controls or non-epileptic postmortem controls. Within the FCD 3D cohort, the DNA methylation signature separated three histopathology subtypes, i.e., glial scarring around porencephalic cysts, loss of layer 4, and Rasmussen encephalitis. Differential methylation in FCD 3D with loss of layer 4 mapped explicitly to biological pathways related to neurodegeneration, biogenesis of the extracellular matrix (ECM) components, axon guidance, and regulation of the actin cytoskeleton. Our data suggest that DNA methylation signatures in cortical malformations are not only of diagnostic value but also phenotypically relevant, providing the molecular underpinnings of structural and histopathological features associated with epilepsy. Further studies will be necessary to confirm these results and clarify their functional relevance and epileptogenic potential in these difficult-to-treat children.

A comprehensive multi-domain dataset for mitotic figure detection

The prognostic value of mitotic figures in tumor tissue is well-established for many tumor types and automating this task is of high research interest. However, especially deep learning-based methods face performance deterioration in the presence of domain shifts, which may arise from different tumor types, slide preparation and digitization devices. We introduce the MIDOG++ dataset, an extension of the MIDOG 2021 and 2022 challenge datasets. We provide region of interest images from 503 histological specimens of seven different tumor types with variable morphology with in total labels for 11,937 mitotic figures: breast carcinoma, lung carcinoma, lymphosarcoma, neuroendocrine tumor, cutaneous mast cell tumor, cutaneous melanoma, and (sub)cutaneous soft tissue sarcoma. The specimens were processed in several laboratories utilizing diverse scanners. We evaluated the extent of the domain shift by using state-of-the-art approaches, observing notable differences in single-domain training. In a leave-one-domain-out setting, generalizability improved considerably. This mitotic figure dataset is the first that incorporates a wide domain shift based on different tumor types, laboratories, whole slide image scanners, and species.

Ganglioglioma with adverse clinical outcome and atypical histopathological features were defined by alterations in PTPN11/KRAS/NF1 and other RAS-/MAP-Kinase pathway genes

Exome-wide sequencing studies recently described PTPN11 as a novel brain somatic epilepsy gene. In contrast, germline mutations of PTPN11 are known to cause Noonan syndrome, a multisystem disorder characterized by abnormal facial features, developmental delay, and sporadically, also brain tumors. Herein, we performed a deep phenotype-genotype analysis of a comprehensive series of ganglioglioma (GG) with brain somatic alterations of the PTPN11/KRAS/NF1 genes compared to GG with common MAP-Kinase signaling pathway alterations, i.e., BRAFV600E. Seventy-two GG were submitted to whole exome sequencing and genotyping and 84 low grade epilepsy associated tumors (LEAT) to DNA-methylation analysis. In 28 tumours, both analyses were available from the same sample. Clinical data were retrieved from hospital files including disease onset, age at surgery, brain localization, and seizure outcome. A comprehensive histopathology staining panel was available in all cases. We identified eight GG with PTPN11 alterations, copy number variant (CNV) gains of chromosome 12, and the commonality of additional CNV gains in NF1, KRAS, FGFR4 and RHEB, as well as BRAFV600E alterations. Histopathology revealed an atypical glio-neuronal phenotype with subarachnoidal tumor spread and large, pleomorphic, and multinuclear cellular features. Only three out of eight patients with GG and PTPN11/KRAS/NF1 alterations were free of disabling-seizures 2 years after surgery (38% had Engel I). This was remarkably different from our series of GG with only BRAFV600E mutations (85% had Engel I). Unsupervised cluster analysis of DNA methylation arrays separated these tumours from well-established LEAT categories. Our data point to a subgroup of GG with cellular atypia in glial and neuronal cell components, adverse postsurgical outcome, and genetically characterized by complex alterations in PTPN11 and other RAS-/MAP-Kinase and/or mTOR signaling pathways. These findings need prospective validation in clinical practice as they argue for an adaptation of the WHO grading system in developmental, glio-neuronal tumors associated with early onset focal epilepsy.

Identification of a novel role for matrix metalloproteinase-3 in the modulation of B cell responses in multiple sclerosis

There has been a growing interest in the presence and role of B cell aggregates within the central nervous system of multiple sclerosis patients. However, very little is known about the expression profile of molecules associated with these aggregates and how they might be influencing aggregate development or persistence in the brain. The current study focuses on the effect of matrix metalloproteinase-3, which is associated with B cell aggregates in autopsied multiple sclerosis brain tissue, on B cells. Autopsied brain sections from multiple sclerosis cases and controls were screened for the presence of CD20⁺ B cell aggregates and expression of matrix metalloproteinase-3. Using flow cytometry, enzyme-linked immunosorbent assay and gene array as methods, in vitro studies were conducted using peripheral blood of healthy volunteers to demonstrate the effect of matrix metalloproteinase-3 on B cells. Autopsied brain sections from multiple sclerosis patients containing aggregates of B cells expressed a significantly higher amount of matrix metalloproteinase-3 compared to controls. In vitro experiments demonstrated that matrix metalloproteinase-3 dampened the overall activation status of B cells by downregulating CD69, CD80 and CD86. Furthermore, matrix metalloproteinase-3-treated B cells produced significantly lower amounts of interleukin-6. Gene array data confirmed that matrix metalloproteinase-3 altered the proliferation and survival profiles of B cells. Taken together, out data indicate a role for B cell modulatory properties of matrix metalloproteinase-3.

AI Prediction of Neuropathic Pain after Lumbar Disc Herniation—Machine Learning Reveals Influencing Factors

The treatment options for neuropathic pain caused by lumbar disc herniation have been debated controversially in the literature. Whether surgical or conservative therapy makes more sense in individual cases can hardly be answered. We have investigated whether a machine learning-based prediction of outcome, regarding neuropathic pain development, after lumbar disc herniation treatment is possible. The extensive datasets of 123 consecutive patients were used to predict the development of neuropathic pain, measured by a visual analogue scale (VAS) for leg pain and the Oswestry Disability Index (ODI), at 6 weeks, 6 months and 1 year after treatment of lumbar disc herniation in a machine learning approach. Using a decision tree regressor algorithm, a prediction quality within the limits of the minimum clinically important difference for the VAS and ODI value could be achieved. An analysis of the influencing factors of the algorithm reveals the important role of psychological factors as well as body weight and age with pre-existing conditions for an accurate prediction of neuropathic pain. The machine learning algorithm developed here can enable an assessment of the course of treatment after lumbar disc herniation. The early, comparative individual prediction of a therapy outcome is important to avoid unnecessary surgical therapies as well as insufficient conservative therapies and prevent the chronification of neuropathic pain.

DNA methylation-based classification of malformations of cortical development in the human brain

Malformations of cortical development (MCD) comprise a broad spectrum of structural brain lesions frequently associated with epilepsy. Disease definition and diagnosis remain challenging and are often prone to arbitrary judgment. Molecular classification of histopathological entities may help rationalize the diagnostic process. We present a retrospective, multi-center analysis of genome-wide DNA methylation from human brain specimens obtained from epilepsy surgery using EPIC 850 K BeadChip arrays. A total of 308 samples were included in the study. In the reference cohort, 239 formalin-fixed and paraffin-embedded (FFPE) tissue samples were histopathologically classified as MCD, including 12 major subtype pathologies. They were compared to 15 FFPE samples from surgical non-MCD cortices and 11 FFPE samples from post-mortem non-epilepsy controls. We applied three different statistical approaches to decipher the DNA methylation pattern of histopathological MCD entities, i.e., pairwise comparison, machine learning, and deep learning algorithms. Our deep learning model, which represented a shallow neuronal network, achieved the highest level of accuracy. A test cohort of 43 independent surgical samples from different epilepsy centers was used to test the precision of our DNA methylation-based MCD classifier. All samples from the test cohort were accurately assigned to their disease classes by the algorithm. These data demonstrate DNA methylation-based MCD classification suitability across major histopathological entities amenable to epilepsy surgery and age groups and will help establish an integrated diagnostic classification scheme for epilepsy-associated MCD.

Multilobar unilateral hypoplasia with emphasis on the posterior quadrant and severe epilepsy in children with FCD ILAE Type 1A

OBJECTIVE

Focal cortical dysplasia (FCD) Type 1 and its three subtypes have yet not been fully characterized at the clinical, anatomopathological, and molecular level (International League Against Epilepsy [ILAE] FCD classification from 2011). We aimed to describe the clinical phenotype of patients with histopathologically confirmed FCD1A obtained from a single epilepsy center between 2002 and 2016.

METHODS

Medical records were retrieved from the hospital's archive. Results from electroencephalography (EEG) video recordings, neuroimaging, and histopathology were reevaluated. Magnetic resonance imaging (MRI) post-processing was retrospectively performed in nine patients. DNA methylation studies were carried out from archival surgical brain tissue in 11 patients.

RESULTS

Nineteen children with a histopathological diagnosis of FCD1A were included. The average onset of epilepsy was 0.9 years (range 0.2-10 years). All children had severe cognitive impairment and one third had mild motor deficits, yet fine finger movements were preserved in all patients. All patients had daily seizures, being drug resistant from disease onset. Interictal electroencephalography revealed bilateral multi-regional epileptiform discharges. Interictal status epilepticus was observed in 8 and countless subclinical seizures in 11 patients. Regional continuous irregular slow waves were of higher lateralizing and localizing yield than spikes. Posterior background rhythms were normal in 16 of 19 children. Neuroimaging showed unilateral multilobar hypoplasia and increased T2-FLAIR signals of the white matter in 18 of 19 patients. All children underwent tailored multilobar resections, with seizure freedom achieved in 47% (Engel class I). There was no case with frontal involvement without involvement of the posterior quadrant by MRI and histopathology. DNA methylation profiling distinguished FCD1A samples from all other epilepsy specimens and controls.

SIGNIFICANCE

We identified a cohort of young children with drug resistance from seizure onset, bad EEG with posterior emphasis, lack of any focal neurological deficits but severe cognitive impairment, subtle hypoplasia of the epileptogenic area on MRI, and histopathologically defined and molecularly confirmed by DNA methylation analysis as FCD ILAE Type 1A.

Modulation of Pancreatic Neuroendocrine Neoplastic Cell Fate by Autophagy-Mediated Death

INTRODUCTION

Autophagic cell death in cancer cells can be mediated by inhibition of deacetylases. Although extensive studies have focused on the autophagic process in cancer, little is known about the role of autophagy in degrading cytosolic and nuclear components of pancreatic neuroendocrine neoplastic (pNEN) cells leading to cell death, thus improving the therapy of patients affected by pNEN.

METHODS

2D and 3D human pNEN and pancreatic stellate cells were treated with panobinostat and bafilomycin. Autophagy markers were detected by RT-qPCR, immunofluorescence, and Western blot. Autophagosomes were detected by electron microscopy and their maturation by real-time fluorescence of LC3B stable transfected cells. ChIP was performed at the cAMP responsive element. Immunofluorescence was performed in murine pancreatic tissue.

RESULTS

We observed that pan-deacetylase inhibitor panobinostat treatment causes autophagic cell death in pNEN cells. We also found that although AMPK-α phosphorylation is counterbalanced by phosphorylated AKT, it is not capable to inhibiting autophagic cell death. However, the binding activity of the cAMP responsive element is prompted by panobinostat. Although autophagy inhibition prevented autophagosome synthesis, maturation, and cell death, panobinostat treatment induced the accumulation of mature autophagosomes in the cytosol and the nucleus, leading to disruption of the organelles, cellular digestion, and decay. Observation of autophagosome membrane proteins Beclin1 and LC3B aggregation in murine pancreatic islets indicates that autophagy restoration may also lead to autophagosome aggregation in murine insulinoma cells. A basal low expression of autophagy markers was detectable in patients affected by pNEN, and, interestingly, the expression of these markers was significantly lower in metastatic pNEN.

DISCUSSION/CONCLUSION

Our study highlights that the autophagy functional restoration and prolongation of this catabolic process, mediated by inhibition of deacetylase, is responsible for the reduction of pNEN cells. Prompting of autophagy cell death could be a promising strategy for the therapy of pNEN.

Mosaic trisomy of chromosome 1q in human brain tissue associates with unilateral polymicrogyria, very early-onset focal epilepsy, and severe developmental delay

Polymicrogyria (PMG) is a developmental cortical malformation characterized by an excess of small and frustrane gyration and abnormal cortical lamination. PMG frequently associates with seizures. The molecular pathomechanisms underlying PMG development are not yet understood. About 40 genes have been associated with PMG, and small copy number variations have also been described in selected patients. We recently provided evidence that epilepsy-associated structural brain lesions can be classified based on genomic DNA methylation patterns. Here, we analyzed 26 PMG patients employing array-based DNA methylation profiling on formalin-fixed paraffin-embedded material. A series of 62 well-characterized non-PMG cortical malformations (focal cortical dysplasia type 2a/b and hemimegalencephaly), temporal lobe epilepsy, and non-epilepsy autopsy controls was used as reference cohort. Unsupervised dimensionality reduction and hierarchical cluster analysis of DNA methylation profiles showed that PMG formed a distinct DNA methylation class. Copy number profiling from DNA methylation data identified a uniform duplication spanning the entire long arm of chromosome 1 in 7 out of 26 PMG patients, which was verified by additional fluorescence in situ hybridization analysis. In respective cases, about 50% of nuclei in the center of the PMG lesion were 1q triploid. No chromosomal imbalance was seen in adjacent, architecturally normal-appearing tissue indicating mosaicism. Clinically, PMG 1q patients presented with a unilateral frontal or hemispheric PMG without hemimegalencephaly, a severe form of intractable epilepsy with seizure onset in the first months of life, and severe developmental delay. Our results show that PMG can be classified among other structural brain lesions according to their DNA methylation profile. One subset of PMG with distinct clinical features exhibits a duplication of chromosomal arm 1q.

Deep Learning-Based Quantification of Pulmonary Hemosiderophages in Cytology Slides

Exercise-induced pulmonary hemorrhage (EIPH) is a common condition in sport horses with negative impact on performance. Cytology of bronchoalveolar lavage fluid by use of a scoring system is considered the most sensitive diagnostic method. Macrophages are classified depending on the degree of cytoplasmic hemosiderin content. The current gold standard is manual grading, which is however monotonous and time-consuming. We evaluated state-of-the-art deep learning-based methods for single cell macrophage classification and compared them against the performance of nine cytology experts and evaluated inter- and intra-observer variability. Additionally, we evaluated object detection methods on a novel data set of 17 completely annotated cytology whole slide images (WSI) containing 78,047 hemosiderophages. Our deep learning-based approach reached a concordance of 0.85, partially exceeding human expert concordance (0.68 to 0.86, mean of 0.73, SD of 0.04). Intra-observer variability was high (0.68 to 0.88) and inter-observer concordance was moderate (Fleiss' kappa = 0.67). Our object detection approach has a mean average precision of 0.66 over the five classes from the whole slide gigapixel image and a computation time of below two minutes. To mitigate the high inter- and intra-rater variability, we propose our automated object detection pipeline, enabling accurate, reproducible and quick EIPH scoring in WSI.

Same same but different: A Web-based deep learning application revealed classifying features for the histopathologic distinction of cortical malformations

OBJECTIVE

The microscopic review of hematoxylin-eosin-stained images of focal cortical dysplasia type IIb and cortical tuber of tuberous sclerosis complex remains challenging. Both entities are distinct subtypes of human malformations of cortical development that share histopathological features consisting of neuronal dyslamination with dysmorphic neurons and balloon cells. We trained a convolutional neural network (CNN) to classify both entities and visualize the results. Additionally, we propose a new Web-based deep learning application as proof of concept of how deep learning could enter the pathologic routine.

METHODS

A digital processing pipeline was developed for a series of 56 cases of focal cortical dysplasia type IIb and cortical tuber of tuberous sclerosis complex to obtain 4000 regions of interest and 200 000 subsamples with different zoom and rotation angles to train a neural network. Guided gradient-weighted class activation maps (Guided Grad-CAMs) were generated to visualize morphological features used by the CNN to distinguish both entities.

RESULTS

Our best-performing network achieved 91% accuracy and 0.88 area under the receiver operating characteristic curve at the tile level for an unseen test set. Novel histopathologic patterns were found through the visualized Guided Grad-CAMs. These patterns were assembled into a classification score to augment decision-making in routine histopathology workup. This score was successfully validated by 11 expert neuropathologists and 12 nonexperts, boosting nonexperts to expert level performance.

SIGNIFICANCE

Our newly developed Web application combines the visualization of whole slide images with the possibility of deep learning-aided classification between focal cortical dysplasia IIb and tuberous sclerosis complex. This approach will help to introduce deep learning applications and visualization for the histopathologic diagnosis of rare and difficult-to-classify brain lesions.

Morphological and Immunohistochemical Characterization of Human Intrinsic Gastric Neurons

Our knowledge about human gastric enteric neuron types is even more limited than that of human intestinal types. Here, we immunohistochemically stained wholemounts and sections of gastric specimens obtained from 18 tumor-resected patients. Myenteric wholemounts were labeled for choline acetyl transferase (ChAT), neuronal nitric oxide synthase (NOS), and the human neuronal protein HuC/D (as pan-neuronal marker for quantitative analysis) or alternatively for neurofilament (for morphological evaluation). ChAT-positive neurons outnumbered NOS-positive neurons (56 vs. 27%), and neurons negative for both markers accounted for 17%. Two larger groups of neurons (each between 12 and 14%) costained for ChAT and vasoactive intestinal peptide (VIP) or for NOS and VIP, respectively. Clear morphochemical correlation was found for uniaxonal stubby type I neurons (ChAT+; putative excitatory inter- or motor neurons), for uniaxonal spiny type I neurons (NOS+/VIP+; putative inhibitory motor or interneurons), and for multiaxonal type II neurons (ChAT+; putative afferent neurons; immunostaining of additional wholemounts revealed their coreactivity for somatostatin). Whereas these latter neuron types were already known from the human intestine, the morphology of gastric myenteric neurons coreactive for ChAT and VIP was newly described: they had numerous short, extremely thin dendrites and resembled, together with their cell bodies, a "hairy" head. In our sections, nerve fibers coreactive for ChAT and VIP were commonly found only in the mucosa. We suggest these myenteric ChAT+/VIP+/hairy neurons to be mucosal effector neurons. In contrast to myenteric neurons, the much less common submucosal neurons were not embedded in a continuous plexus and did not display any clear morphochemical phenotypes.

Contribution of LTi and TH17 cells to B cell aggregate formation in the central nervous system in a mouse model of multiple sclerosis

Background

In a subgroup of patients suffering from progressive multiple sclerosis (MS), which is an inflammation-mediated neurodegenerative disease of the central nervous system (CNS), B cell aggregates were discovered within the meninges. Occurrence of these structures was associated with a more severe disease course and cortical histopathology. We have developed the B cell-dependent MP4-induced experimental autoimmune encephalomyelitis (EAE) as a mouse model to mimic this trait of the human disease. The aim of this study was to determine a potential role of lymphoid tissue inducer (LTi) and T_H17 cells in the process of B cell aggregate formation in the MP4 model.

Methods

We performed flow cytometry of cerebellar and splenic tissue of MP4-immunized mice in the acute and chronic stage of the disease to analyze the presence of CD3⁻CD5⁻CD4⁺RORγt⁺ LTi and CD3⁺CD5⁺CD4⁺RORγt⁺ T_H17 cells. Myelin oligodendrocyte glycoprotein (MOG):35–55-induced EAE was used as B cell-independent control model. We further determined the gene expression profile of B cell aggregates using laser capture microdissection, followed by RNA sequencing.

Results

While we were able to detect LTi cells in the embryonic spleen and adult intestine, which served as positive controls, there was no evidence for the existence of such a population in acute or chronic EAE in neither of the two models. Yet, we detected CD3⁻CD5⁻CD4⁻RORγt⁺ innate lymphoid cells (ILCs) and T_H17 cells in the CNS, the latter especially in the chronic stage of MP4-induced EAE. Moreover, we observed a unique gene signature in CNS B cell aggregates compared to draining lymph nodes of MP4-immunized mice and to cerebellum as well as draining lymph nodes of mice with MOG:35–55-induced EAE.

Conclusion

The absence of LTi cells in the cerebellum suggests that other cells might take over the function as an initiator of lymphoid tissue formation in the CNS. Overall, the development of ectopic lymphoid organs is a complex process based on an interplay between several molecules and signals. Here, we propose some potential candidates, which might be involved in the formation of B cell aggregates in the CNS of MP4-immunized mice.

Alarin in different human intestinal epithelial cell types

Alarin (AL), a new member of the galanin family, has been localized in various CNS regions, mainly in rodents. Among other effects, it modulates food intake. Therefore, we analyzed the immunohistochemical distribution pattern of AL in human intestinal epithelia. Cryosections of 12 human bowel samples were immunohistochemically double-stained for AL and α-defensin 5 (αD; first set). Two further sets of sections were quadruple-stained either (second set) for AL, chromogranin (CG), synaptophysin (SY), and somatostatin (SO) or (third set) for AL, CG, Peptide Y (PY), and 5-hydroxytryptamine (5-HT). Slides were digitized and quantitative analysis of co-localization rates was undertaken. Small bowel: most of AL-positive cells (56%) were αD-positive Paneth cells located within the base of the crypts (first set). In the second set, about 27% of AL-labeled cells were co-reactive for SY and CG, likely representing entero-endocrine cells. In the third set, the largest subpopulation of AL-positive cells was not co-reactive for other markers applied (89%); most of them were likely Paneth cells. Large bowel: co-localization of AL with αD was not detected (first set). In the second set, AL was frequently co-localized with the other three markers applied (68%). In the third set, AL was frequently co-localized with 5-HT and CG (31%) as well as with PY and 5-HT (22%). Due to its presence in various enteroendocrine as well as Paneth cells, AL may be involved in different physiological and pathological processes.

Digital pathology imaging and computer-aided diagnostics as a novel tool for standardization of evaluation of aganglionic megacolon (Hirschsprung disease) histopathology

Based on a recently introduced immunohistochemical panel (Bachmann et al. 2015) for aganglionic megacolon (AM), also known as Hirschsprung disease, histopathological diagnosis, we evaluated whether the use of digital pathology and 'machine learning' could help to obtain a reliable diagnosis. Slides were obtained from 31 specimens of 27 patients immunohistochemically stained for MAP2, calretinin, S100β and GLUT1. Slides were digitized by whole slide scanning. We used a Definiens Developer Tissue Studios as software for analysis. We configured necessary parameters in combination with 'machine learning' to identify pathological aberrations. A significant difference between AM- and non-AM-affected tissues was found for calretinin (AM 0.55% vs. non-AM 1.44%) and MAP2 (AM 0.004% vs. non-AM 0.07%) staining measurements and software-based evaluations. In contrast, S100β and GLUT1 staining measurements and software-based evaluations showed no significant differences between AM- and non-AM-affected tissues. However, no difference was found in comparison of suction biopsies with resections. Applying machine learning via an ensemble voting classifier, we achieved an accuracy of 87.5% on the test set. Automated diagnosis of AM by applying digital pathology on immunohistochemical panels was successful for calretinin and MAP2, whereas S100β and GLUT1 were not effective in diagnosis. Our method suggests that software-based approaches are capable of diagnosing AM. Our future challenge will be the improvement of efficiency by reduction of the time-consuming need for large pre-labelled training data. With increasing technical improvement, especially in unsupervised training procedures, this method could be helpful in the future.

Panobinostat mediated cell death: a novel therapeutic approach for osteosarcoma

Osteosarcoma is an aggressive cancer with a poor long term prognosis. Neo-adjuvant poly-chemotherapy followed by surgical resection remains the standard treatment, which is restricted by multi-drug resistance. If first-line therapy fails, disease control and patient survival rate drop dramatically. We aimed to identify alternative apoptotic mechanisms induced by the histone deacetylase inhibitor panobinostat in osteosarcoma cells. Saos-2, MG63 and U2-OS osteosarcoma cell lines, the immortalized human osteoblast line hFOB and the mouse embryo osteoblasts (MC3T3-E1) were treated with panobinostat. Real time viability and FACS confirmed the cytotoxicity of panobinostat. Cell stress/death related factors were analysed by RT-qPCR and western blot. Cell morphology was assessed by electron microscopy. 10 nM panobinostat caused cell viability arrest and death in all osteosarcoma and osteoblast cells. P21 up-regulation was observed in osteosarcoma cells, while over-expression of p73 was restricted to Saos-2 (TP53^-/-). Survivin and Bcl-2 were suppressed by panobinostat. Endoplasmic reticulum (ER) stress markers BiP, CHOP, ATF4 and ATF6 were induced in osteosarcoma cells. The un-spliced Xbp was no further detectable after treatment. Autophagy players Beclin1, Map1LC3B and UVRAG transcripts over-expressed after 6 hours. Protein levels of Beclin1, Map1LC3B and p62 were up-regulated at 72 hours. DRAM1 was stable. Electron micrographs revealed the fragmentation and the disappearance of the ER and the statistically significant increase of autophagosome vesiculation after treatment. Panobinostat showed a synergistic suppression of survival and promotion of cell death in osteosarcoma cells. Panobinostat offers new perspectives for the treatment of osteosarcoma and other malignant bone tumours.

Calbindin D28k-Immunoreactivity in Human Enteric Neurons

Calbindin (CALB) is well established as immunohistochemical marker for intrinsic primary afferent neurons in the guinea pig gut. Its expression by numerous human enteric neurons has been demonstrated but little is known about particular types of neurons immunoreactive for CALB. Here we investigated small and large intestinal wholemount sets of 26 tumor patients in order to evaluate (1) the proportion of CALB⁺ neurons in the total neuron population, (2) the colocalization of CALB with calretinin (CALR), somatostatin (SOM) and vasoactive intestinal peptide (VIP) and (3) the morphology of CALB⁺ neurons. CALB⁺ neurons represented a minority of myenteric neurons (small intestine: 31%; large intestine: 25%) and the majority of submucosal neurons (between 72 and 95%). In the submucosa, most CALB⁺ neurons co-stained for CALR and VIP (between 69 and 80%) or for SOM (between 20 and 3%). In the myenteric plexus, 85% of CALB⁺ neurons did not co-stain with the other markers investigated. An unequivocal correlation between CALB reactivity and neuronal morphology was found for myenteric type III neurons in the small intestine: uniaxonal neurons with long, slender and branched dendrites were generally positive for CALB. Since also other neurons displayed occasional CALB reactivity, this protein is not suited as an exclusive marker for type III neurons.

The enteric nervous system is a potential autoimmune target in multiple sclerosis

Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system (CNS) in young adults that has serious negative socioeconomic effects. In addition to symptoms caused by CNS pathology, the majority of MS patients frequently exhibit gastrointestinal dysfunction, which was previously either explained by the presence of spinal cord lesions or not directly linked to the autoimmune etiology of the disease. Here, we studied the enteric nervous system (ENS) in a B cell- and antibody-dependent mouse model of MS by immunohistochemistry and electron microscopy at different stages of the disease. ENS degeneration was evident prior to the development of CNS lesions and the onset of neurological deficits in mice. The pathology was antibody mediated and caused a significant decrease in gastrointestinal motility, which was associated with ENS gliosis and neuronal loss. We identified autoantibodies against four potential target antigens derived from enteric glia and/or neurons by immunoprecipitation and mass spectrometry. Antibodies against three of the target antigens were also present in the plasma of MS patients as confirmed by ELISA. The analysis of human colon resectates provided evidence of gliosis and ENS degeneration in MS patients compared to non-MS controls. For the first time, this study establishes a pathomechanistic link between the well-established autoimmune attack on the CNS and ENS pathology in MS, which might provide a paradigm shift in our current understanding of the immunopathogenesis of the disease with broad diagnostic and therapeutic implications.

Fibromyalgia syndrome: metabolic and autophagic processes in intermittent cold stress mice

Fibromyalgia is characterized by widespread musculoskeletal pain, fatigue, and depression. The aim was to analyze potential mitochondrial dysfunction or autophagy in mice after exposure to intermittent cold stress (ICS). Muscle and liver specimens were obtained from 36 mice. Lactate dehydrogenase (LDH) activity was measured. Microtubule-associated protein light chain 3 (MAP1LC3B) and glycogen content were determined histologically; muscle ultrastructure by electron microscopy. Mitochondrial- and autophagy-related markers were analyzed by RT-qPCR and Western blotting. ATP level, cytotoxicity, and caspase 3 activity were measured in murine C2C12 myoblasts after ICS exposure. Coenzyme Q10B (COQ10B) transcript was up-regulated in limb muscle of ICS mice, whereas its protein content was stable. Cytochrome C oxidase 4 (COX4I1) and LDH activity increased in limb muscle of male ICS mice. Glycogen content was lower in muscle and liver tissue of male ICS mice. Electron micrographs of ICS mice specimens showed mitochondrial damage and autophagic vesicles. A significant up-regulation of autophagic transcripts of MAP1LC3B and BECLIN 1 (BECN1) was observed. Map1lc3b protein showed an aggregated distribution in ICS mice and SqSTM1/p62 (p62) protein level was stable. Furthermore, ATP level and caspase activity, detected as apoptotic marker, were significantly lowered after ICS exposure in differentiated C2C12 myoblasts. The present study shows that ICS mice are characterized by mitochondrial dysfunction, autophagic processes, and metabolic alterations. Further investigations could dissect autophagy process in the proposed model and link these mechanisms to potential therapeutic options for fibromyalgia.

Autophagy-related cell death by pan-histone deacetylase inhibition in liver cancer

Autophagy is a homeostatic, catabolic degradation process and cell fate essential regulatory mechanism. Protracted autophagy triggers cell death; its aberrant function is responsible for several malignancies. Panobinostat, a potent pan-deacetylase inhibitor, causes endoplasmic reticulum stress-induced cell death. The aim of this study was to investigate the role of autophagy in deacetylase inhibitor-triggered liver cancer cell death.HepG2 (p53wt) and Hep3B (p53 null) liver cancer cell lines were exposed to panobinostat. RT-qPCR and western blot confirmed autophagic factor modulation. Immuno-fluorescence, -precipitation and -histochemistry as well as transmission electron microscopy verified autophagosome formation. The cytotoxicity of panobinostat and autophagy modulators was detected using a real time cell viability assay.Panobinostat induced autophagy-related factor expression and aggregation. Map1LC3B and Beclin1 were significantly over-expressed in HepG2 xenografts in nude mice treated with panobinostat for 4 weeks. Subcellular distribution of Beclin1 increased with the appearance of autophagosomes-like aggregates. Cytosolic loss of p53, in HepG2, and p73, in Hep3B cells, and a corresponding gain of their nuclear level, together with modulation of DRAM1, were observed. Autophagosome aggregation was visible after 6 h of treatment. Treatment of cells stably expressing GFP-RFPtag Map1LC3B resulted in aggregation and a fluorescence switch, thus confirming autophagosome formation and maturation. Tamoxifen, an inducer of autophagy, caused only a block in cell proliferation; but in combination with panobinostat it resulted in cell death.Autophagy triggers cell demise in liver cancer. Its modulation by the combination of tamoxifen and panobinostat could be a new option for palliative treatment of hepatocellular carcinoma.

Neurovascular Interface in Porcine Small Intestine: Specific for Nitrergic rather than Nonnitrergic Neurons

In the 1970s, by using classic histological methods, close topographical relationships between special areas of enteric ganglia and capillaries were shown in the pig. In this study, by application of double and triple immunohistochemistry, we confirmed this neurovascular interface and demonstrated that these zones are mainly confined to nitrergic neurons in the myenteric and the external submucosal plexus. In the upper small intestine of the pig, the respective neurons display type III morphology, i.e. they have long, slender and branched dendrites and a single axon. In another set of experiments, we prepared specimens for electron-microscopical analysis of these zones. Both ganglia and capillaries display continuous basement membranes, the smallest distances between them being 1,000 nm at the myenteric and 300 nm at the external submucosal level. The capillary endothelium was mostly continuous but, at the external submucosal level, scattered fenestrations were observed. This particular neurovascular relationship suggests that nitrergic neurons may require a greater amount of oxygen and/or nutrients. In guinea pig and mouse, previous ischemia/reperfusion experiments showed that nitrergic neurons are selectively damaged. Thus, a preferential blood supply of enteric nitrergic neurons may indicate that these neurons are more vulnerable in ischemia.

Interstitial cells of Cajal: crucial for the development of megacolon in human Chagas' disease?

AIM

Megacolon, chronic dilation of a colonic segment,is accompanied by extensive myenteric neuron loss. However, this fails to explain unequivocally the formation of megacolon. We aimed to study further enteric structures that are directly or indirectly involved in colonic motility.

METHOD

From surgically removed megacolon segments of seven Chagasic patients, three sets of cryosections from oral, megacolonic and anal zones were immunohistochemically quadruple-stained for smooth-muscle actin (SMA), synaptophysin (SYN, for nerve fibres), S100 (glia) and c-Kit (interstitial cells of Cajal, ICCs). Values of area measurements were related to the appropriate muscle layer areas and these proportions were compared with those of seven non-Chagasic control patients.

RESULTS

Whereas nerve and glia profile proportions did not mirror unequivocally the changes of Chagasic colon calibre (nondilation/dilation/nondilation), the proportions of SMA (i.e. muscle tissue density) and c-Kit (i.e. ICC density) did so: they decreased from the oral to the megacolonic segment but increased to the anal zones (muscle tissue density: control 68.3%, oral 54.3%, mega 42.1%, anal 47.6%; ICC-density: control 1.8%, oral 1.1%, mega 0.4, anal 0.8%).

CONCLUSION

Of the parameters evaluated, muscle tissue and ICC densities may be involved in the formation of Chagasic megacolon, although the mechanism of destruction cannot be deduced.

Mucosal layers and related nerve fibres in non-chagasic and chagasic human colon--a quantitative immunohistochemical study

Chagasic megacolon is accompanied by extensive myenteric and, simultaneously, moderate submucosal neuron loss. Here, we examined changes of the innervation pattern of the lamina propria (LP) and muscularis mucosae (MM). Two alternating sets of cryosections were taken from seven non-chagasic colonic and seven chagasic megacolonic specimens (the latter included both the dilated megacolonic and the non-dilated transitional oral and anal zones) and were immunohistochemically triple-stained for smooth-muscle actin (SMA), synaptophysin (SYN) and glial acid protein S100 and, alternatively, for SMA, vasoactive intestinal peptide (VIP) and somatostatin (SOM). Subsequent image analysis and statistical evaluation of nervous tissue profile areas revealed that, in LP, the most extreme differences (i.e. increase in thickness or decrease in nerve, glia and muscle tissue profile area, respectively) compared with control values occurred in the dilated megacolonic zone itself. In contrast, the most extreme differences in the MM were in the anal-to-megacolonic zone (except the profile area of muscle tissue, which was lowest in the megacolonic zone). This parallels our previous results in the external muscle coat. A partial and selective survival of VIP-immunoreactive in contrast to SOM-immunoreactive nerve fibres was observed in both mucosal layers investigated. Thus, VIPergic nerve elements might be crucial for the maintenance of the mucosal barrier. The differential changes of neural tissue parameters in LP and MM might reflect a multifactorial rather than a pure neurogenic development of megacolon in chronic Chagas' disease.

What neurons hide behind calretinin immunoreactivity in the human gut?

Calretinin (CALR) is often used as an immunohistochemical marker for the histopathological diagnosis of human intestinal neuropathies. However, little is known about its distribution pattern with respect to specific human enteric neuron types. Prior studies revealed CALR in both myenteric and submucosal neurons, most of which colabel with choline acetyl transferase (ChAT). Here, we specified the chemical code of CALR-positive neurons in small and large intestinal wholemounts in a series of 28 patients. Besides other markers, we evaluated the labeling pattern of CALR in combination with vasoactive intestinal peptide (VIP). In colonic submucosa, CALR and VIP were almost completely colocalized in about three-quarters of all submucosal neurons. In the small intestinal submucosa, both the colocalization rate of CALR and VIP as well as the proportion of these neurons were lower (about one-third). In the myenteric plexus of both small intestine and colon, CALR amounted to 11 and 10 %, respectively, whereas VIP to 5 and 4 % of the whole neuron population, respectively. Colocalization of both markers was found in only 2 and 3 % of myenteric neurons, respectively. In section specimens, nerve fibers coreactive for CALR and VIP were found in the mucosa but not in the muscle coat. Summarizing the present and earlier results, CALR was found in at least one submucosal and two myenteric neuron populations. Submucosal CALR+/VIP+/ChAT± neurons innervate mucosal structures. Furthermore, CALR immunoreactivity in the myenteric plexus was observed in morphological type II (supposed primary afferent) and spiny type I (supposed inter- or motor-) neurons.

Endoplasmic reticulum stress plays a pivotal role in cell death mediated by the pan-deacetylase inhibitor panobinostat in human hepatocellular cancer cells

Panobinostat, a pan-deacetylase inhibitor, represents a novel therapeutic option for cancer diseases. Besides its ability to block histone deacetylases (HDACs) by promoting histone hyperacetylation, panobinostat interferes with several cell death pathways providing a potential efficacy against tumors. We have previously demonstrated that panobinostat has a potent apoptotic activity in vitro and causes a significant growth delay of hepatocellular carcinoma (HCC) tumor xenografts in nude mice models. Here, we show that treatment with panobinostat is able to induce noncanonical apoptotic cell death in HepG2 and in Hep3B cells, involving the endoplasmic reticulum (ER) stress by up-regulation of the molecular chaperone binding immunoglobulin protein/glucose-regulated protein 78, activation of eukaryotic initiation factor 2α-activating transcription factor 4 (tax-responsive enhancer element B67) and inositol requiring 1α-X-box binding protein 1 factors, strong increase and nuclear translocation of the transcription factor C/EBP homologous protein/growth arrest and DNA damage-inducible gene 153, and involvement of c-Jun N-terminal kinase. These signaling cascades culminate into the activation of the ER-located caspase-4/12 and of executioner caspases, which finally lead to cell demise. Our results clearly show that panobinostat induces an alternative ER stress-mediated cell death pathway in HCC cells, independent of the p53 status.

Selective survival of calretinin- and vasoactive-intestinal-peptide-containing nerve elements in human chagasic submucosa and mucosa

Chronic Chagas' disease is frequently characterized by massive myenteric neuron loss resulting in megacolon with severely and irreversibly disturbed motility. Here, we focused on two submucosal neuron populations, immunoreactive for calretinin (CALR) or somatostatin (SOM), and their respective mucosal nerve fibres in chagasic megacolon. Surgically removed megacolonic segments of seven chagasic patients were compared with seven age- and region-matched non-chagasic control segments. Evaluation included immunohistochemical triple-staining of cryosections for CALR, SOM and peripherin or for CALR and vasoactive intestinal peptide (VIP) and of submucosal whole-mounts for CALR, SOM and the pan-neuronal marker anti-HuC/D. Submucosal neuron counts in chagasic tissue revealed neuron numbers reduced to 51.2 % of control values. In cryosections, nerve fibre area measurements revealed 8.6 % nerve fibre per mucosal area in control segments, but this value decreased to 1.5 % in megacolonic segments. In both evaluations, a disproportionate decrease of SOM-reactive nerve elements was observed. The proportions of SOM-positive neurons related to the total neuron number declined to 2 % (control 10 %) and the proportion of SOM-reactive mucosal nerve fibres related to the whole mucosal area to 0.014 % (control 1.8 %)in chagasic tissue. The second set of cryosections revealed extensive colocalization of CALR with VIP in both surviving submucosal perikarya and mucosal nerve fibres. We suggest that VIP, a neuroprotective and neuroeffectory peptide typically contained in submucosal neurons, allows both the VIP-containing neurons to endure and the patients to survive by maintaining their mucosal barrier, despite the almost complete loss of colonic motility for decades.

Partial, selective survival of nitrergic neurons in chagasic megacolon

One frequent chronic syndrome of Chagas’ disease is megacolon, an irreversible dilation of a colonic segment. Extensive enteric neuron loss in the affected segment is regarded as key factor for deficient motility. Here, we assessed the quantitative balance between cholinergic and nitrergic neurons representing the main limbs of excitatory and inhibitory colonic motor innervation, respectively. From surgically removed megacolonic segments of four patients, each three myenteric wholemounts (from non-dilated oral, megacolonic and non-dilated anal parts) was immunohistochemically triple-stained for choline acetyltransferase, neuronal nitric oxide synthase (NOS) and the panneuronal human neuronal protein Hu C/D. Degenerative changes were most pronounced in the megacolonic and anal regions, e.g. bulked, honeycomb-like ganglia with few neurons which were partly enlarged or atrophic or vacuolated. Neuron counts from each 15 ganglia of 12 megacolonic wholemounts were compared with those of 12 age- and region-matched controls. Extensive neuron loss, mainly in megacolonic and anal wholemounts, was obvious. In all three regions derived from megacolonic samples, the proportion of NOS-positive neurons (control: 55%) was significantly increased: in non-dilated oral parts to 61% (p = 0.003), in megacolonic regions to 72% (p < 0.001) and in non-dilated anal regions to 78% (p < 0.001). We suggest the chronic dilation of megacolonic specimens to be due to the preponderance of the nitrergic, inhibitory input to the intestinal muscle. However, the observed neuronal imbalance was not restricted to the dilated regions: the non-dilated anal parts may be innervated by ascending, cholinergic axons emerging from less affected, more anally located regions.

Cellular plasticity of trans- and dedifferentiation markers in human hepatoma cells in vitro and in vivo

Tumor cells have the capability to trans- and to dedifferentiate, for example by reactivating embryonic development genes and stem cell characteristics. The aim of our study was to show the differential expression of stem- and progenitor cell markers in human hepatocellular carcinoma cell lines (HCC). Different human HCC cell lines (HUH7, HUH7 5-15, HUH7 pcDNA3.1, Hep3B and HepG2) were cultured under standard conditions in vitro or implanted subcutaneously (5x10(6) cells) in male NMRI mice. Specimens were characterized by quantitative real-time PCR, Western blotting, methylation-specific PCR and immunohistochemistry for markers of differentiation (cytokeratins, vimentin), embryonic development or stem cells (PTC, PDX-1, SHH, Thy1, c-kit, CD34, beta-catenin, Ki-67). The investigated HCC cell lines showed different patterns of marker expression allowing to distinguish four distinct groups: the classical cholangiocellular type (Huh-7, Huh-7 pcDNA3.1, Hep3B) with expression of CK7/19, beta-catenin and CD34; a dedifferentiated mesenchymal-proliferative type (Huh-7 5-15) characterized by CK19, Vimentin and Ki-67; a dedifferentiated embryonic-development type (Hep3B implanted in matrigel) with expression of CK19, beta-catenin and PTC and a classical HCC type (HepG2) showing CK18/19 and beta-catenin expression. HCC cell lines showed significantly different expression patterns of differentiation markers in a xenograft model. Furthermore, direct association of some markers was observed. The groups differ from each other in expression patterns, but also show that environmental factors play an important role in the behaviour of cells.