Artificial intelligence (AI) holds much promise for enabling highly desired imaging diagnostics improvements. One of the most limiting bottlenecks for the development of useful clinical-grade AI models is the lack of training data. One aspect is the large amount of cases needed and another is the necessity of high-quality ground truth annotation. The aim of the project was to establish and describe the construction of a database with substantial amounts of detail-annotated oncology imaging data from pathology and radiology. A specific objective was to be proactive, that is, to support undefined subsequent AI training across a wide range of tasks, such as detection, quantification, segmentation, and classification, which puts particular focus on the quality and generality of the annotations. The main outcome of this project was the database as such, with a collection of labeled image data from breast, ovary, skin, colon, skeleton, and liver. In addition, this effort also served as an exploration of best practices for further scalability of high-quality image collections, and a main contribution of the study was generic lessons learned regarding how to successfully organize efforts to construct medical imaging databases for AI training, summarized as eight guiding principles covering team, process, and execution aspects.

The dataset consists of 101 H&E-stained colon whole slide images (WSI) - 52 abnormal and 49 benign cases. All significant abnormal findings identified are outlined and categorized into 15 types such as hyperplastic polyp, high grade adenocarcinoma and necrosis. Other tissue components such as mucosa, submucosa, as well as the surgical margin are delineated to create a complete histological map. In total, 756 separate annotations have been made to segment the different tissue structures and link them to ontological information.

In the developing Drosophila central nervous system (CNS), neural progenitor (neuroblast [NB]) selection is gated by lateral inhibition, controlled by Notch signaling and proneural genes. However, proneural mutants still generate many NBs, indicating the existence of additional proneural genes. Moreover, recent studies reveal involvement of key epithelial-mesenchymal transition (EMT) genes in NB selection, but the regulatory interplay between Notch signaling and the EMT machinery is unclear. We find that SoxNeuro (SoxB family) and worniu (Snail family) are integrated with the Notch pathway, and constitute the missing proneural genes. Notch signaling, the proneural, SoxNeuro, and worniu genes regulate key EMT genes to orchestrate the NB selection process. Hence, we uncover an expanded lateral inhibition network for NB selection and demonstrate its link to key players in the EMT machinery. The evolutionary conservation of the genes involved suggests that the Notch-SoxB-Snail-EMT network may control neural progenitor selection in many other systems.

During cell cycle progression, the activity of the CycE-Cdk2 complex gates S-phase entry. CycE-Cdk2 is inhibited by CDK inhibitors (CKIs) of the Cip/Kip family, which include the human p21(Cip)(1) and Drosophila Dacapo (Dap) proteins. Both the CycE and Cip/Kip family proteins are under elaborate control via protein degradation, mediated by the Cullin-RING ligase (CRL) family of ubiquitin ligase complexes. The CRL complex SCFFoxw7/Ago targets phosphorylated CycE, whereas p21(Cip)(1) and Dap are targeted by the CRLCdf2 complex, binding to the PIP degron. The role of CRL-mediated degradation of CycE and Cip/Kip proteins during CNS development is not well understood. Here, we analyse the role of ago (Fbxw7)-mediated CycE degradation, and of Dap and p21(Cip)(1) degradation during Drosophila CNS development. We find that ago mutants display over-proliferation, accompanied by elevated CycE expression levels. By contrast, expression of PIP degron mutant Dap and p21(Cip)(1) transgenes inhibit proliferation. However, surprisingly, this is also accompanied by elevated CycE levels. Hence, ago mutation and PIP degron Cip/Kip transgenic expression trigger opposite effects on proliferation, but similar effects on CycE levels.

During embryonic development, a number of genetic cues act to generate neuronal diversity. While intrinsic transcriptional cascades are well-known to control neuronal sub-type cell fate, the target cells can also provide critical input to specific neuronal cell fates. Such signals, denoted retrograde signals, are known to provide critical survival cues for neurons, but have also been found to trigger terminal differentiation of neurons. One salient example of such target-derived instructive signals pertains to the specification of the Drosophila FMRFamide neuropeptide neurons, the Tv4 neurons of the ventral nerve cord. Tv4 neurons receive a BMP signal from their target cells, which acts as the final trigger to activate the FMRFa gene. A recent FMRFa-eGFP genetic screen identified several genes involved in Tv4 specification, two of which encode components of the U5 subunit of the spliceosome: brr2 (l(3) 72Ab) and Prp8. In this study, we focus on the role of RNA processing during target- derived signaling. We found that brr2 and Prp8 play crucial roles in controlling the expression of the FMRFa neuropeptide specifically in six neurons of the VNC (Tv4 neurons). Detailed analysis of brr2 revealed that this control is executed by two independent mechanisms, both of which are required for the activation of the BMP retrograde signaling pathway in Tv4 neurons: (1) Proper axonal pathfinding to the target tissue in order to receive the BMP ligand. (2) Proper RNA splicing of two genes in the BMP pathway: the thickveins (tkv) gene, encoding a BMP receptor subunit, and the Medea gene, encoding a co-Smad. These results reveal involvement of specific RNA processing in diversifying neuronal identity within the central nervous system.

The Notch pathway controls proliferation during development and in adulthood, and is frequently affected in many disorders. However, the genetic sensitivity and multi-layered transcriptional properties of the Notch pathway has made its molecular decoding challenging. Here, we address the complexity of Notch signaling with respect to proliferation, using the developing Drosophila CNS as model. We find that a Notch/Su(H)/E(spl)-HLH cascade specifically controls daughter, but not progenitor proliferation. Additionally, we find that different E(spl)-HLH genes are required in different neuroblast lineages. The Notch/Su(H)/E(spl)-HLH cascade alters daughter proliferation by regulating four key cell cycle factors: Cyclin E, String/Cdc25, E2f and Dacapo (mammalian p21(CIP1)/p27(KIP1)/p57(Kip2)). ChIP and DamID analysis of Su(H) and E(spl)-HLH indicates direct transcriptional regulation of the cell cycle genes, and of the Notch pathway itself. These results point to a multi-level signaling model and may help shed light on the dichotomous proliferative role of Notch signaling in many other systems.

The human central nervous system (CNS) displays the greatest cellular diversity of any organ system, consisting of billions of neurons, of numerous cell sub-types, interconnected in a vast network. Given this enormous complexity, decoding the genetic programs controlling the multistep process of CNS development remains a major challenge. While great progress has been made with respect to understanding sub-type specification, considerably less is known regarding how the generation of the precise number of each sub-type is controlled.

The aim of this thesis was to gain deeper knowledge into the regulatory programs controlling cell specification and proliferation. To address these questions I have studied the Drosophila embryonic CNS as a model system, to thereby be able to investigate the genetic mechanisms at high resolution. Despite the different size and morphology between the Drosophila and the mammalian CNS, the lineages of their progenitors share similarity. Importantly for this thesis, both species progenitors show elaborate variations in their proliferation modes, either giving rise to daughters that can directly differentiate into neurons or glia (type 0), divide once (type I), or multiple times (type II).

The studies launched off with a comprehensive chemical forward genetic screen, for the very last born cell in the well-studied lineage of progenitor NB5-6T: the Ap4 neuron, which expresses the neuropeptide FMRFa. NB5-6T is a powerful model to use, because it undergoes a programmed type I>0 daughter cell proliferation switch. An FMRF-eGFP transgenic reporter was utilized as readout for successful terminal differentiation of Ap4/FMRFa and thereby proper lineage progression of the ∼20 cells generated. The strongest mutants were mapped to genes with both known and novel essential functions e.g., spatial and temporal patterning, cell cycle control, cell specification and chromatin modification. Subsequently, we focused on some of the genes that showed a loss of function phenotype with an excess of lineage cells. We found that Notch is critical for the type I>0 daughter cell proliferation switch in the NB5-6T lineage and globally as well. When addressing the broader relevance of these findings, and to further decipher the Notch pathway, we discovered that selective groups of E(spl) genes is controlling the switch in a close interplay with four key cell cycle factors: Cyclin E, String, E2F and Dacapo, in most if not all embryonic progenitors. The Notch mediation of the switch is likely to be by direct transcriptional regulation. Furthermore, another gene identified in the screen, sequoia, was investigated. The analysis revealed that sequoia is also controlling the daughter cell switch in the CNS, and this partly through context dependent interactions with the Notch pathway.

Taken together, the findings presented in this thesis demonstrate that daughter cell proliferation switches in Drosophila neural lineages are genetically programmed, and that Notch contributes to the triggering of these events. Given that early embryonic processes is frequently shown to be evolutionary conserved, you can speculate that changeable daughter proliferation programs could be applied to mammals, and contribute to a broader understanding of proliferation processes in humans as well.

Neural progenitors typically divide asymmetrically to renew themselves, while producing daughters with more limited potential. In the Drosophila embryonic ventral nerve cord, neuroblasts initially produce daughters that divide once to generate two neurons/glia (type I proliferation mode). Subsequently, many neuroblasts switch to generating daughters that differentiate directly (type 0). This programmed type I>0 switch is controlled by Notch signaling, triggered at a distinct point of lineage progression in each neuroblast. However, how Notch signaling onset is gated was unclear. We recently identified Sequoia (Seq), a C2H2 zinc-finger transcription factor with homology to Drosophila Tramtrack (Ttk) and the positive regulatory domain (PRDM) family, as important for lineage progression. Here, we find that seq mutants fail to execute the type I>0 daughter proliferation switch and also display increased neuroblast proliferation. Genetic interaction studies reveal that seq interacts with the Notch pathway, and seq furthermore affects expression of a Notch pathway reporter. These findings suggest that seq may act as a context-dependent regulator of Notch signaling, and underscore the growing connection between Seq, Ttk, the PRDM family and Notch signaling.

The expression of neuropeptides is often extremely restricted in the nervous system, making them powerful markers for addressing cell specification . In the developing Drosophila ventral nerve cord, only six cells, the Ap4 neurons, of some 10,000 neurons, express the neuropeptide FMRFamide (FMRFa). Each Ap4/FMRFa neuron is the last-born cell generated by an identifiable and well-studied progenitor cell, neuroblast 5-6 (NB5-6T). The restricted expression of FMRFa and the wealth of information regarding its gene regulation and Ap4 neuron specification makes FMRFa a valuable readout for addressing many aspects of neural development, i.e., spatial and temporal patterning cues, cell cycle control, cell specification, axon transport, and retrograde signaling. To this end, we have conducted a forward genetic screen utilizing an Ap4-specific FMRFa-eGFP transgenic reporter as our readout. A total of 9781 EMS-mutated chromosomes were screened for perturbations in FMRFa-eGFP expression, and 611 mutants were identified. Seventy-nine of the strongest mutants were mapped down to the affected gene by deficiency mapping or whole-genome sequencing. We isolated novel alleles for previously known FMRFa regulators, confirming the validity of the screen. In addition, we identified novel essential genes, including several with previously undefined functions in neural development. Our identification of genes affecting most major steps required for successful terminal differentiation of Ap4 neurons provides a comprehensive view of the genetic flow controlling the generation of highly unique neuronal cell types in the developing nervous system.

During central nervous system (CNS) development, progenitors typically divide asymmetrically, renewing themselves while budding off daughter cells with more limited proliferative potential. Variation in daughter cell proliferation has a profound impact on CNS development and evolution, but the underlying mechanisms remain poorly understood. We find that Drosophila embryonic neural progenitors (neuroblasts) undergo a programmed daughter proliferation mode switch, from generating daughters that divide once (type I) to generating neurons directly (type 0). This typelgreater than0 switch is triggered by activation of Dacapo (mammalian p21(CIP1)/p27(KIP1)/p57(Kip2)) expression in neuroblasts. In the thoracic region, Dacapo expression is activated by the temporal cascade (castor) and the Hox gene Antennapedia. In addition, castor, Antennapedia, and the late temporal gene grainyhead act combinatorially to control the precise timing of neuroblast cell-cycle exit by repressing Cyclin E and E2f. This reveals a logical principle underlying progenitor and daughter cell proliferation control in the Drosophila CNS.

During neural lineage progression, differences in daughter cell proliferation can generate different lineage topologies. This is apparent in the Drosophila neuroblast 5-6 lineage (NB5-6T), which undergoes a daughter cell proliferation switch from generating daughter cells that divide once to generating neurons directly. Simultaneously, neural lineages, e.g. NB5-6T, undergo temporal changes in competence, as evidenced by the generation of different neural subtypes at distinct time points. When daughter proliferation is altered against a backdrop of temporal competence changes, it may create an integrative mechanism for simultaneously controlling cell fate and number. Here, we identify two independent pathways, Prospero and Notch, which act in concert to control the different daughter cell proliferation modes in NB5-6T. Altering daughter cell proliferation and temporal progression, individually and simultaneously, results in predictable changes in cell fate and number. This demonstrates that different daughter cell proliferation modes can be integrated with temporal competence changes, and suggests a novel mechanism for coordinately controlling neuronal subtype numbers.

n/a

n/a

A central theme in developmental biology pertains to how the diversity of different cell types is generated. In addition, it is important to understand how the numbers of each cell type are regulated. In the developing Drosophila ventral nerve cord, only six neurons, the Ap4 neurons, express the neuropeptide gene FMRFamide (FMRFa). This is the result of proper lineage development and a cascade of regulatory information leading to final cell specification. In addition to these cascades, FMRFa expression is critically dependent upon a retrogarade TGFβ/BMP signal from the axonal target. Its restricted expression pattern and the wealth of information regarding its gene regulation, makes FMRFa a useful marker for understanding cell specification, as well as axon path finding and retrograde signaling. To identify novel genes acting at any level of neuronal development, including pattern formation, stem cell competence, cell cycle control, cell specification, axon transport and retrograde signaling, we have conducted a single cell resolution, forward genetic screen utilizing an FMRFa-EGFP reporter as our read-out. A total of 9,781 EMS-mutated chromosomes were screened for perturbations in FMRFa-EGFP expression, and 611 mutants were identified. Complementation tests showed that many of the previously known regulators had been isolated, which confirmed the validity of the screen. However, in addition to these known genes, the majority of mutants represent novel genes with previously undefined functions in neural development.