16580 research outputs found
Sort by
Molecular and functional evaluation of myeloid cell programming in a patient-derived colon cancer organoid co-culture
No full textTumor-associated monocytes and macrophages accumulate in colorectal cancer (CRC) and play a critical role in shaping the tumor microenvironment (TME). While they have been linked to both pro- and anti-tumor functions, understanding the cues that instruct the phenotype of individual subsets, as well as their functional impact on the TME remains challenging. In this study, I established co-cultures comprising primary human monocytes and patient-derived organoids (PDOs) from CRC tumors to emulate myeloid cell-carcinoma interactions in vitro. Across diverse PDOs, monocytes acquired a shared phenotype that transcriptionally resembled IL1B-programmed monocytes previously identified in CRC patients. This phenotype emerged independently of the tumors’ mutational profiles or CMS type, and was arising by carcinoma cell contact without the need for additional stromal components. At the mechanistic level, I demonstrated that soluble mediators secreted by PDOs were sufficient to induce the expression of IL1B signature genes such as the chemokines CXCL2, CXCL5, and CXCL7. In parallel, phagocytic uptake of tumor debris reduced the expression of MHC class II molecules on the surface of myeloid cells, suggesting an impairment of antigen-presenting capacity. In addition, this in vitro system allowed the functional assessment of PDO-exposed monocytes demonstrating a compromised capacity to mount an inflammatory response upon TLR stimulation. This uncoupling of gene expression from effector function points to a refractory state of monocytes upon contact with carcinoma cells, potentially contributing to an immune evasive TME where tumor-derived inflammation is maintained at a transcriptional level but fails to translate into effective anti-tumor activity. Together, my PDO–myeloid cell co-cultures offer an elegant model system to study the interplay between human epithelial cancer cells and monocytes, and to advance the understanding of myeloid plasticity and function in CRC. Future studies can leverage this model to identify and validate therapeutic strategies aimed at overcoming monocyte/macrophage dysfunction, for instance by restoring antigen presentation, breaking refractoriness to innate stimuli, or selectively targeting IL1B-driven inflammatory programs. By capturing the personalized interplay between patient-derived tumor cells and primary immune cells, this system bridges experimental modelling and clinical heterogeneity, thereby advancing the development of TAM-based therapies with improved bench-to-bedside translation for CRC patients
Vom Bildnis zum Bit. Deep Learning zur Analyse der Ikonographie antiker Münzen
No full textThe classification of ancient coins remains a central but highly labor-intensive task in numismatics, as it relies on heterogeneous descriptive traditions and the manual comparison of large corpora. Although automated image recognition has been explored for several decades, practical applications have so far been limited by insufficient data standardization and the complexity of interpretative iconographic descriptions. This study investigates the potential and current limits of an AI-based approach to the automated classification of Roman coins.
Focusing on Roman coinage (ca. 320 BCE–491 CE), the research combines Natural Language Processing and computer vision to address both textual and visual sources. Interpretative coin descriptions from digitized GLAM collections were transformed into machine-readable data using a custom Named Entity Recognition system based on a knowledge graph, incorporating alternative spellings, fuzzy matching, and targeted data augmentation. In parallel, coin images were preprocessed and used to train a Convolutional Neural Network to recognize the most frequent iconographic representations.
The results demonstrate that reliable automated recognition of iconographic motifs is achievable for standardized coin images, while also revealing limitations caused by descriptive variance, image quality, and mild overfitting in the recognition of persons and attributes. The study shows that high-quality, interoperable data are a prerequisite for successful automation and that explainability and statistical robustness remain key requirements for practical use.
Overall, the proposed approach provides a scalable foundation for automated coin classification in Roman numismatics and establishes a benchmark for future work. It also highlights perspectives for extending the method to other numismatic corpora, improving data interoperability, and supporting both academic research and heritage documentation through AI-assisted workflows
Monitoring urban greenhouse gas emissions using high-resolution models
No full textUrban regions emit large amounts of CO2, are drivers of national inventory uncertainty, and have a direct impact on the day-to-day life of their inhabitants. Therefore, an independent assessment of urban emissions supports climate action and governance. This work presents a complete pipeline for improving urban emissions estimates. It includes setting up, optimizing, and evaluating atmospheric transport as well as CO2 and CO simulations and estimating
emissions based on aircraft measurements.
We optimize the physics configuration of the WRF atmospheric transport model and show that its performance exceeds previous studies. We use this optimized configuration to generate MACRO-2018, a 1 km resolved dataset of meteorology as well as CO2 and CO concentrations in German metropolitan areas for 2018. This dataset provides two different physics configurations and biogenic models. We evaluate its performance against high-precision CO2 and CO measurements and find mean absolute biases of 5.2 ppm to 6.4 ppm for CO2, exceeding the quality of comparable datasets. For CO, we find mean absolute biases of 25.4 ppb to 28.6 ppb. Finally, based on this dataset we set up a Bayesian inversion framework which assimilates aircraft measurements to optimize the emissions of Berlin and the background CO2 concentration simultaneously. Using this approach, we reduce the emissions estimate uncertainty from a simpler and widely-used mass balance approach by a factor of five
Eine „ergiebige Quelle […], aus der Heimatliebe und Bürgersinn neue Kräfte heben“: Die museale Modellierung von Heimat(en) in der deutsch-belgischen Grenzregion
No full textDer Beitrag skizziert die Geschichte des Stadtgeschichtlichen Museums und des Couven-Museums in Aachen sowie des Eupener Museums vom frühen 20. Jahrhundert bis in die Gegenwart und geht der Frage nach, welche Kontinuitäten und Veränderungen sich in den dortigen Darstellungen von Heimat(en) beobachten lassen. Der Blick auf die musealen Ansätze im deutsch-belgischen Grenzraum ist eingebettet in die im Beitrag vorgestellte Konzeption des Dissertationsprojektes der Autorin, das museale Modellierungen von Heimat(en) in deutschen Grenzregionen untersucht
Spatial and temporal localization of unipolar brush cells in mice and humans using multiplexing immunofluorescence imaging approaches
No full textPediatric brain tumors are the leading cause of cancer-related mortality in children, with medulloblastomas being the most prevalent embryonal type. Among its four molecular
subgroups, Group 4 is the most common yet least understood. Standard treatments (surgery, chemotherapy, radiotherapy) improve survival but cause severe side effects, including
immunosuppression and neurocognitive decline. Current genetically modified mouse models fail to recapitulate the observed heterogeneity of Group 4 tumors, and attempts to generate accurate models using known driver oncogenes, such as PRDM6 and MYCN, have been unsuccessful. This may be due to our limited understanding of the tumor’s origin within the unipolar brush cell (UBC) lineage, particularly the critical time window in which these oncogenes exert their effects. Despite their identification nearly 50 years ago, UBCs remain poorly characterized, with most studies limited to electrophysiology and a handful of molecular
analyses.
To address this gap, I combined immunofluorescence-based imaging with transcriptomic datasets, including my own single-cell RNA sequencing (scRNA-seq), to refine UBC lineage definitions and their relevance to Group 4 tumor modeling. Pseudo-age-matched comparisons between human and mouse cerebella revealed species-specific differences, including earlier UBC progenitor emergence in human fetal development and a higher proportion of early differentiating UBCs. Using an UBC reporter mouse, I enriched and isolated GFP-positive cells from E16.5 and P0 cerebella, confirming all UBC cell states and uncovering unexpected heterogeneity, with some UBCs lacking Lmx1a expression. This finding challenges the assumption that all UBCs follow a uniform lineage and suggests instead an additional lineage
diversity within this population.
To validate these results, I optimized multiplexed Immuno-SABER staining, evolving from single/double stainings to simultaneous visualization of up to 13 proteins in developing mouse and human cerebella. This enabled a refined molecular characterization of the rhombic lip (RL), revealing three distinct compartments: (1) early progenitors in the ventricular zone (RL VZ), (2) fate-determined granule cell (GC)/UBC progenitors in the subventricular zone
(RL SVZ), and (3) proliferative progenitors, potentially of the GC lineage, in the exterior rhombic lip (eRL). Temporal analyses showed that UBC progenitors appeared in the RL from
E16.5 onward but were absent along migratory pathways, where only differentiating UBCs were detected. Mature UBCs first emerged at P7 and became more abundant by P28, with a
surprising degree of protein expression heterogeneity, including populations lacking TBR2, the presumed universal UBC marker.
To further investigate lineage commitment, I applied Immuno-SABER to genetically modified mouse models and human tissue. Lineage tracing of Lmx1a-expressing progenitors confirmed their contribution to UBC development, though LMX1A was not consistently expressed in all UBCs. Functional ablation of TBR2 in Lmx1a+ progenitors disrupted UBC maturation, preventing proper migration and differentiation, with no mature UBCs detected at P28. However, TBR2+ GCs, particularly in the inner external granular layer (EGL), remained unaffected, suggesting distinct TBR2 dependencies within cerebellar lineages. In human fetal cerebellar tissue, TBR2+ developing UBCs exhibited expected protein expression patterns predicted by transcriptomic analyses. However, murine UBCs lacked the anticipated co-expression of ATOH1 and LMX1A with TBR2, highlighting species-specific
transcriptional and translational discrepancies. In patient-derived xenograft (PDX) models, TBR2 was detected in Group 3 tumors but absent in mixed Group 3/4 tumors, raising
questions about how TBR2 expression is regulated during tumor development.
In summary, my work integrates transcriptomic and high-content imaging analyses to refine cerebellar lineage definitions to improve preclinical modeling of Group 4 medulloblastomas in the future. UBC heterogeneity, species-specific differences, and unexpected marker variability
suggest that additional UBC subtypes may exist beyond the traditionally recognized two. Future studies should investigate the roles of TBR2 and LMX1A in UBC lineage specification and tumor heterogeneity, as these factors may be key to unraveling Group 4 medulloblastoma biology and improving disease modeling
Die Bedeutung des Hippocampus beim semantischen Gedächtnisabruf am Beispiel der transienten globalen Amnesie
No full textDie transiente globale Amnesie (TGA) ist ein neurologisches Krankheitsbild mit dem Leitsymptom einer akuten Gedächtnisstörung, die sich nach einigen Stunden spontan zurückbildet. Als klinisches Modell einer Funktionsstörung des Hippocampus stellt sie ein einzigartiges experimentelles Paradigma dar, um die Bedeutung des Hippocampus für Gedächtnisprozesse zu evaluieren. Die vorliegende Arbeit hatte zum Ziel, am Beispiel der TGA die Rolle des Hippocampus beim Abruf semantischer Gedächtnisinhalte zu untersuchen. Im Einzelnen wurde untersucht, ob während einer Hippocampusdysfunktion der Abruf von in den letzten Jahren wiederholt erlernten semantischen Informationen in gewissem Umfang möglich ist, obwohl der episodische Abruf aus dem gleichen Zeitraum kaum gelingt (Ziel 1); ob eine hippocampale Funktionsstörung den semantischen Abruf in Abhängigkeit von den verwendeten Abrufstrategien einschränkt und ob sich die Beeinträchtigung je nach Abrufstrategie in unterschiedlichem Ausmaß manifestiert (Ziel 2); ob eine Hippocampusdysfunktion zu spezifischen Beeinträchtigungen im Abruf semantisch-räumlichen Wissens führt und falls ja, ob diese Beeinträchtigung auf Störungen der visuell-räumlichen Verarbeitung zurückzuführen ist (Ziel 3) und schließlich, ob es einen Zusammenhang zwischen semantischer Abrufleistung und funktionell-anatomischer DWI-Läsionslokalisation gibt (Ziel 4). Insgesamt wurden 20 Patient*innen mit einer TGA in die Studie eingeschlossen. Für die Ziele 1-3 wurden Daten der akuten und postakuten Phase von 17 Patient*innen in die Auswertung einbezogen; zusätzlich wurden 17 gesunde Kontrollpersonen untersucht, die sich hinsichtlich ihrer soziodemographischen Daten nicht von den Patient*innen unterschieden haben. Für Ziel 4 gingen Daten von 15 Patient*innen in die Analyse ein.
Die Ergebnisse der vorliegenden Arbeit zeigen, dass der Abruf bestimmter semantischer Informationen, die in den letzten 5-10 Jahren wiederholt erlernt wurden, trotz eingeschränkter Hippocampusfunktion möglich ist, während episodische Inhalte aus dem gleichen Zeitraum kaum abrufbar sind. Dennoch ist der Hippocampus am semantischen Abruf beteiligt, wenn auch weniger ausgeprägt als beim episodischen Abruf. Dieses Ergebnis unterstützt die Annahme der Multiple Trace Theory (z. B. Nadel & Moscovitch, 1997), dass in den letzten Jahren erworbene Inhalte durch multiples Lernen extrahippocampal gespeichert werden und somit weniger Hippocampusbeteiligung beim Abruf benötigen, wobei eine solche beim semantischen Abruf nicht gänzlich ausgeschlossen wird. Weiterhin kann festgehalten werden, dass eine Dysfunktion des Hippocampus die semantische Wortflüssigkeitsleistung allgemein und differenziell in Abhängigkeit von den verwendeten Abrufstrategien einschränkt. Insbesondere scheint der Hippocampus wesentlich zum semantischen Abruf beizutragen, wenn semantische Inhalte flexibel neu verknüpft werden müssen. Dieser Befund ist mit der Annahme der Relational Memory Theory (Cohen & Eichenbaum, 1993; Eichenbaum & Cohen, 2001) vereinbar, dass eine Beteiligung des Hippocampus unabhängig von der Art der abzurufenden Inhalte (episodisch oder semantisch) erfolgt, sobald Informationen flexibel abgerufen werden müssen. Tendenziell findet sich eine stärkere Beeinträchtigung unter Hippocampusdysfunktion im semantischen Abruf von räumlichen Inhalten im Vergleich zum Abruf von Inhalten ohne räumliche Aktivierung, was jedoch nicht sicher von Defiziten in der visuell-räumlichen Verarbeitung differenziert werden kann. Dieser Befund steht eher im Einklang mit der Cognitive Map Theory (z. B. O'Keefe & Nadel, 1978), die dem Hippocampus eine wichtige Funktion bei allozentrisch-räumlicher Repräsentation zuschreibt. Bezüglich der bildgebenden Befunde kann ein Zusammenhang zwischen semantischen Abrufleistungen im Wortflüssigkeitsparadigma und der Lokalisation von DWI-Läsionen festgestellt werden – DWI-Läsionen im posterioren Hippocampus sind mit größeren Leistungsverlusten in semantischen Wortflüssigkeitsaufgaben assoziiert, die auf einer episodisch-räumlichen Abrufstrategie basieren. Dagegen werden bei anterioren DWI-Läsionen größere Leistungsverluste in semantischen Wortflüssigkeitsaufgaben, die eine Neuverknüpfung semantischen Wissens erfordern, beobachtet. Dieses Ergebnismuster passt zu Konzeptionen, wonach der anteriore Hippocampus Ereignisse grob repräsentiert und Verallgemeinerungen mit neuartigen Assoziationen über Ereignisse hinweg ermöglicht, während der posteriore Hippocampus Ereignisse auf einer feineren und detaillierten zeitlich-räumlichen Skala abbildet. Die Lokalisation der Läsion im Hippocampus scheint somit die semantische Abrufleistung zu beeinflussen.
In der Gesamtschau lässt sich eine Beteiligung des Hippocampus am semantischen Abruf feststellen, die auf einen substanziellen und differenziellen Beitrag des Hippocampus zur relationalen und räumlichen Verarbeitung zurückzuführen ist. Die Beziehung zwischen episodischem und semantischem Gedächtnis ist interdependent, wobei insbesondere das fronto-hippocampale Netzwerk eine wichtige Rolle zu spielen scheint. Diese Arbeit plädiert für eine Modifikation der klassischen Theorien der Hippocampusfunktion mit sich hieraus ergebenden Implikationen für die Diagnostik und Behandlung von Gedächtnisstörungen
Hermann von Helmholtz : Nachruf (1895) und Biographie-Artikel (1906)
No full textDer Verfasser, der von 1865 bis 1871 Helmholtz' Assistent in Heidelberg war, schildert im ersten Beitrag knapp das Leben Hermann von Helmholtz und geht umfassend auf dessen physiologische und physikalische Publikationen ein.
Im zweiten Beitrag widmet er sich ausführlicher der Heidelberger Zeit und benennt zahlreiche Personen aus Helmholtz' Umfeld.
Die Beiträge werden durch Kurzbiographien der dort erwähnten Personen ergänzt
Development of a Novel Framework to Explore Correlations in Multinuclear MR Spectroscopy: Enabling Advanced Quantification of In Vivo 1H Spectra Supported by 31P Data
No full textMagnetic resonance spectroscopic imaging (MRSI) enables non-invasive quantification of metabolite concentrations in vivo. Proton (1H) MRSI is relatively sensitive, but has difficulties discriminating metabolites with similar chemical shifts. In contrast, phosphorus (31P) MRSI shows larger chemical shift dispersion, but is less sensitive. Both techniques provide mutual information that can be exploited by exploring their correlations. In this work, a novel framework for such an exploration was developed with the initial aim to advance 1H MRS(I) quantification using 31P MRS(I) as ground truth, employing sequential multiparametric MR protocols and machine learning analysis.
This framework was applied to create models that robustly discriminate glycerophosphocholine (GPC) and phosphocholine (PC) in 1H spectra for the cases of synthetic spectra
(n_trained = 160,000), acquired spectra of model solutions at B0 = 9.4T (n_trained = 5,760), and MRSI from healthy volunteers at B0 = 7T (n_trained = 1,521). In all cases, the created interpretable, non-linear gradient boosting model outperformed the state-of-the-art method LCModel, e.g., for synthetic spectra, with a mean absolute percentage error about 7 times lower. Feature analysis identified not only spectral features around the total choline resonance as important for the GPC/PC discrimination, but in vivo also spectral features from other metabolites, like glutamine/glutamate, not known before.
In conclusion, the framework proved to enable novel applications, and might ultimately pave the way for approaches overcoming limitations of individual in vivo MRSI methods by combining their strengths
Frustrated Magnetism and Localization in a Long-Range Interacting Spin System
No full textFrustration, bond disorder, and long-range interactions can produce novel, highly nontrivial phases of matter. This thesis investigates a bond-disordered, dipolar-interacting XY spin model, studied through both numerical simulations and experiments on a Rydberg quantum simulator. Three major results are achieved: (i) We analyze energetic-magnetic hysteresis in two disorder configurations, and find pronounced hysteresis in the strongly disordered configuration for low energies. This is a first experimental evidence of energetic-magnetic hysteresis in putative isolated glasses. (ii) We develop an extension of the generalized Kibble-Zurek mechanism for reverse quenches, which allows to characterize putative spin glass quantum phase transitions by global magnetization measurements. We validate this extension numerically exactly for one-dimensional systems, and subsequently experimentally extract a critical exponent consistent with spin glass critical exponents measured in other systems. This is a first tentative experimental evidence of a spin glass phase in an isolated dipolar interacting spin system. (iii) We studied time reversal on a Rydberg quantum simulator and showed numerically that, in bond-disordered power-law interacting models, time-reversal–based protocols reveal a localization mechanism distinct from conventional many-body localization at finite sizes