Guided interactive image segmentation using machine learning and color-based image set clustering.

MOTIVATION: Over the last decades, image processing and analysis have become one of the key technologies in systems biology and medicine. The quantification of anatomical structures and dynamic processes in living systems is essential for understanding the complex underlying mechanisms and allows, i.e. the construction of spatio-temporal models that illuminate the interplay between architecture and function. Recently, deep learning significantly improved the performance of traditional image analysis in cases where imaging techniques provide large amounts of data. However, if only a few images are available or qualified annotations are expensive to produce, the applicability of deep learning is still limited. RESULTS: We present a novel approach that combines machine learning-based interactive image segmentation using supervoxels with a clustering method for the automated identification of similarly colored images in large image sets which enables a guided reuse of interactively trained classifiers. Our approach solves the problem of deteriorated segmentation and quantification accuracy when reusing trained classifiers which is due to significant color variability prevalent and often unavoidable in biological and medical images. This increase in efficiency improves the suitability of interactive segmentation for larger image sets, enabling efficient quantification or the rapid generation of training data for deep learning with minimal effort. The presented methods are applicable for almost any image type and represent a useful tool for image analysis tasks in general. AVAILABILITY AND IMPLEMENTATION: The presented methods are implemented in our image processing software TiQuant which is freely available at tiquant.hoehme.com. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Quantitative modeling identifies critical cell mechanics driving bile duct lumen formation.

Biliary ducts collect bile from liver lobules, the smallest functional and anatomical units of liver, and carry it to the gallbladder. Disruptions in this process caused by defective embryonic development, or through ductal reaction in liver disease have a major impact on life quality and survival of patients. A deep understanding of the processes underlying bile duct lumen formation is crucial to identify intervention points to avoid or treat the appearance of defective bile ducts. Several hypotheses have been proposed to characterize the biophysical mechanisms driving initial bile duct lumen formation during embryogenesis. Here, guided by the quantification of morphological features and expression of genes in bile ducts from embryonic mouse liver, we sharpened these hypotheses and collected data to develop a high resolution individual cell-based computational model that enables to test alternative hypotheses in silico. This model permits realistic simulations of tissue and cell mechanics at sub-cellular scale. Our simulations suggest that successful bile duct lumen formation requires a simultaneous contribution of directed cell division of cholangiocytes, local osmotic effects generated by salt excretion in the lumen, and temporally-controlled differentiation of hepatoblasts to cholangiocytes, with apical constriction of cholangiocytes only moderately affecting luminal size.

On the Mechanisms of Biliary Flux.

Since the late 1950s, transport of bile in the liver has been described by the "osmotic concept," according to which bile flows into the canaliculi toward the ducts, countercurrent to the blood flow in the sinusoids. However, because of the small size of canaliculi, it was so far impossible to observe, let alone to quantify this process. Still, "osmotic canalicular flow" was a sufficient and plausible explanation for the clearance characteristics of a wide variety of choleretic compounds excreted in bile. Imaging techniques have now been established that allow direct flux analysis in bile canaliculi of the intact liver in living organisms. In contrast to the prevailing osmotic concept these analyses strongly suggest that the transport of small molecules in canalicular bile is diffusion dominated, while canalicular flow is negligibly small. In contrast, with the same experimental approach, it could be shown that in the interlobular ducts, diffusion is augmented by flow. Thus, bile canaliculi can be compared to a standing water zone that is connected to a river. The seemingly subtle difference between diffusion and flow is of relevance for therapy of a wide range of liver diseases including cholestasis and NAFLD. Here, we incorporated the latest findings on canalicular solute transport, and align them with extant knowledge to present an integrated and explanatory framework of bile flux that will undoubtedly be refined further in the future.

Intravital Dynamic and Correlative Imaging of Mouse Livers Reveals Diffusion-Dominated Canalicular and Flow-Augmented Ductular Bile Flux.

BACKGROUND AND AIMS: Small-molecule flux in tissue microdomains is essential for organ function, but knowledge of this process is scant due to the lack of suitable methods. We developed two independent techniques that allow the quantification of advection (flow) and diffusion in individual bile canaliculi and in interlobular bile ducts of intact livers in living mice, namely fluorescence loss after photoactivation and intravital arbitrary region image correlation spectroscopy. APPROACH AND RESULTS: The results challenge the prevailing "mechano-osmotic" theory of canalicular bile flow. After active transport across hepatocyte membranes, bile acids are transported in the canaliculi primarily by diffusion. Only in the interlobular ducts is diffusion augmented by regulatable advection. Photoactivation of fluorescein bis-(5-carboxymethoxy-2-nitrobenzyl)-ether in entire lobules demonstrated the establishment of diffusive gradients in the bile canalicular network and the sink function of interlobular ducts. In contrast to the bile canalicular network, vectorial transport was detected and quantified in the mesh of interlobular bile ducts. CONCLUSIONS: The liver consists of a diffusion-dominated canalicular domain, where hepatocytes secrete small molecules and generate a concentration gradient and a flow-augmented ductular domain, where regulated water influx creates unidirectional advection that augments the diffusive flux.

Bile Microinfarcts in Cholestasis Are Initiated by Rupture of the Apical Hepatocyte Membrane and Cause Shunting of Bile to Sinusoidal Blood.

Bile duct ligation (BDL) is an experimental procedure that mimics obstructive cholestatic disease. One of the early consequences of BDL in rodents is the appearance of so-called bile infarcts that correspond to Charcot-Gombault necrosis in human cholestasis. The mechanisms causing bile infarcts and their pathophysiological relevance are unclear. Therefore, intravital two photon-based imaging of BDL mice was performed with fluorescent bile salts (BS) and non-BS organic anion analogues. Key findings were followed up by matrix-assisted laser desorption ionization imaging, clinical chemistry, immunostaining, and gene expression analyses. In the acute phase, 1-3 days after BDL, BS concentrations in bile increased and single-cell bile microinfarcts occurred in dispersed hepatocytes throughout the liver caused by the rupture of the apical hepatocyte membrane. This rupture occurred after loss of mitochondrial membrane potential, followed by entry of bile, cell death, and a "domino effect" of further death events of neighboring hepatocytes. Bile infarcts provided a trans-epithelial shunt between bile canaliculi and sinusoids by which bile constituents leaked into blood. In the chronic phase, ≥21 days after BDL, uptake of BS tracers at the sinusoidal hepatocyte membrane was reduced. This contributes to elevated concentrations of BS in blood and decreased concentrations in the biliary tract. Conclusion: Bile microinfarcts occur in the acute phase after BDL in a limited number of dispersed hepatocytes followed by larger infarcts involving neighboring hepatocytes, and they allow leakage of bile from the BS-overloaded biliary tract into blood, thereby protecting the liver from BS toxicity; in the chronic phase after BDL, reduced sinusoidal BS uptake is a dominant protective factor, and the kidney contributes to the elimination of BS until cholemic nephropathy sets in.

Quantitative cell-based model predicts mechanical stress response of growing tumor spheroids over various growth conditions and cell lines.

Model simulations indicate that the response of growing cell populations on mechanical stress follows the same functional relationship and is predictable over different cell lines and growth conditions despite experimental response curves look largely different. We develop a hybrid model strategy in which cells are represented by coarse-grained individual units calibrated with a high resolution cell model and parameterized by measurable biophysical and cell-biological parameters. Cell cycle progression in our model is controlled by volumetric strain, the latter being derived from a bio-mechanical relation between applied pressure and cell compressibility. After parameter calibration from experiments with mouse colon carcinoma cells growing against the resistance of an elastic alginate capsule, the model adequately predicts the growth curve in i) soft and rigid capsules, ii) in different experimental conditions where the mechanical stress is generated by osmosis via a high molecular weight dextran solution, and iii) for other cell types with different growth kinetics from the growth kinetics in absence of external stress. Our model simulation results suggest a generic, even quantitatively same, growth response of cell populations upon externally applied mechanical stress, as it can be quantitatively predicted using the same growth progression function.

Modulating Portal Hemodynamics With Vascular Ring Allows Efficient Regeneration After Partial Hepatectomy in a Porcine Model.

OBJECTIVE: To investigate safety and efficacy of temporary portal hemodynamics modulation with a novel percutaneously adjustable vascular ring (MID-AVR) onto a porcine model of 75% hepatectomy. BACKGROUND: Postoperative liver failure is a leading cause of mortality after major hepatectomy. Portal flow modulation is an increasingly accepted concept to prevent postoperative liver failure. Nonetheless, the current strategies have shortcomings. METHODS: Resection was performed under hemodynamic monitoring in 17 large, white pigs allocated into 2 groups. Eight pigs had ring around the portal vein for 3 days with the aim of reducing changes in hemodynamics due to hepatectomy. Analysis of hemodynamics, laboratory, and histopathological parameters was performed. RESULTS: Percutaneous inflation, deflation, and removal of the MID-AVR were safe. Two (25%) pigs in the MID-AVR group and 4 (45%) controls died before day 3 (P = NS). A moderate increase of portal flow rate per liver mass after resection was associated with better survival (P = 0.017). The portocaval pressure gradient was lower after hepatectomy in the MID-AVR group (P = 0.001). Postoperative serum bilirubin levels were lower in the MID-AVR group (P = 0.007 at day 5). In the MID-AVR group, the Ki67 index was significantly higher on day 3 (P = 0.043) and the architectural derangement was lower (P < 0.05). Morphometric quantification of the bile canaliculi revealed a significantly lower number of intersection branches (P < 0.05) and intersection nodes (P < 0.001) on day 7 compared with the preoperative specimen, in the control group. These differences were not found in the ring group. CONCLUSIONS: MID-AVR is safe for portal hemodynamics modulation. It might improve liver regeneration by protecting liver microarchitecture.

Quantitative analysis of hepatic macro- and microvascular alterations during cirrhogenesis in the rat.

Cirrhosis represents the end-stage of any persistent chronically active liver disease. It is characterized by the complete replacement of normal liver tissue by fibrosis, regenerative nodules, and complete fibrotic vascularized septa. The resulting angioarchitectural distortion contributes to an increasing intrahepatic vascular resistance, impeding liver perfusion and leading to portal hypertension. To date, knowledge on the dynamically evolving pathological changes of the hepatic vasculature during cirrhogenesis remains limited. More specifically, detailed anatomical data on the vascular adaptations during disease development is lacking. To address this need, we studied the 3D architecture of the hepatic vasculature during induction of cirrhogenesis in a rat model. Cirrhosis was chemically induced with thioacetamide (TAA). At predefined time points, the hepatic vasculature was fixed and visualized using a combination of vascular corrosion casting and deep tissue microscopy. Three-dimensional reconstruction and data-fitting enabled cirrhogenic features to extracted at multiple scales, portraying the impact of cirrhosis on the hepatic vasculature. At the macrolevel, we noticed that regenerative nodules severely compressed pliant venous vessels from 12 weeks of TAA intoxication onwards. Especially hepatic veins were highly affected by this compression, with collapsed vessel segments severely reducing perfusion capabilities. At the microlevel, we discovered zone-specific sinusoidal degeneration, with sinusoids located near the surface being more affected than those in the middle of a liver lobe. Our data shed light on and quantify the evolving angioarchitecture during cirrhogenesis. These findings may prove helpful for future targeted invasive interventions.

Model Prediction and Validation of an Order Mechanism Controlling the Spatiotemporal Phenotype of Early Hepatocellular Carcinoma.

Recently, hepatocyte-sinusoid alignment (HSA) has been identified as a mechanism that supports the coordination of hepatocytes during liver regeneration to reestablish a functional micro-architecture (Hoehme et al. in Proc Natl Acad Sci 107(23):10371-10376, 2010). HSA means that hepatocytes preferentially align along the closest micro-vessels. Here, we studied whether this mechanism is still active in early hepatocellular tumors. The same agent-based spatiotemporal model that previously correctly predicted HSA in liver regeneration was further developed to simulate scenarios in early tumor development, when individual initiated hepatocytes gain increased proliferation capacity. The model simulations were performed under conditions of realistic liver micro-architectures obtained from 3D reconstructions of confocal laser scanning micrographs. Interestingly, the established model predicted that initiated hepatocytes at first arrange in elongated patterns. Only when the tumor progresses to cell numbers of approximately 4000, does it adopt spherical structures. This prediction may have relevant consequences, since elongated tumors may reach critical structures faster, such as larger vessels, compared to a spherical tumor of similar cell number. Interestingly, this model prediction was confirmed by analysis of the spatial organization of initiated hepatocytes in a rat liver tumor initiation study using single doses of 250 mg/kg of the genotoxic carcinogen N-nitrosomorpholine (NNM). Indeed, small clusters of GST-P positive cells induced by NNM were elongated, almost columnar, while larger GDT-P positive foci of approximately the size of liver lobuli adopted spherical shapes. From simulations testing numerous possible mechanisms, only HSA could explain the experimentally observed initial deviation from spherical shape. The present study demonstrates that the architecture of small cell clusters of hepatocytes early after initiation is still controlled by physiological mechanisms. However, this coordinating influence is lost when the tumor grows to approximately 4000 cells, leading to further growth in spherical shape. Our findings stress the potential importance of organ micro-architecture in understanding tumor phenotypes.

Inferring Growth Control Mechanisms in Growing Multi-cellular Spheroids of NSCLC Cells from Spatial-Temporal Image Data.

We develop a quantitative single cell-based mathematical model for multi-cellular tumor spheroids (MCTS) of SK-MES-1 cells, a non-small cell lung cancer (NSCLC) cell line, growing under various nutrient conditions: we confront the simulations performed with this model with data on the growth kinetics and spatial labeling patterns for cell proliferation, extracellular matrix (ECM), cell distribution and cell death. We start with a simple model capturing part of the experimental observations. We then show, by performing a sensitivity analysis at each development stage of the model that its complexity needs to be stepwise increased to account for further experimental growth conditions. We thus ultimately arrive at a model that mimics the MCTS growth under multiple conditions to a great extent. Interestingly, the final model, is a minimal model capable of explaining all data simultaneously in the sense, that the number of mechanisms it contains is sufficient to explain the data and missing out any of its mechanisms did not permit fit between all data and the model within physiological parameter ranges. Nevertheless, compared to earlier models it is quite complex i.e., it includes a wide range of mechanisms discussed in biological literature. In this model, the cells lacking oxygen switch from aerobe to anaerobe glycolysis and produce lactate. Too high concentrations of lactate or too low concentrations of ATP promote cell death. Only if the extracellular matrix density overcomes a certain threshold, cells are able to enter the cell cycle. Dying cells produce a diffusive growth inhibitor. Missing out the spatial information would not permit to infer the mechanisms at work. Our findings suggest that this iterative data integration together with intermediate model sensitivity analysis at each model development stage, provide a promising strategy to infer predictive yet minimal (in the above sense) quantitative models of tumor growth, as prospectively of other tissue organization processes. Importantly, calibrating the model with two nutriment-rich growth conditions, the outcome for two nutriment-poor growth conditions could be predicted. As the final model is however quite complex, incorporating many mechanisms, space, time, and stochastic processes, parameter identification is a challenge. This calls for more efficient strategies of imaging and image analysis, as well as of parameter identification in stochastic agent-based simulations.

Adverse outcome pathways: opportunities, limitations and open questions.

Adverse outcome pathways (AOPs) are a recent toxicological construct that connects, in a formalized, transparent and quality-controlled way, mechanistic information to apical endpoints for regulatory purposes. AOP links a molecular initiating event (MIE) to the adverse outcome (AO) via key events (KE), in a way specified by key event relationships (KER). Although this approach to formalize mechanistic toxicological information only started in 2010, over 200 AOPs have already been established. At this stage, new requirements arise, such as the need for harmonization and re-assessment, for continuous updating, as well as for alerting about pitfalls, misuses and limits of applicability. In this review, the history of the AOP concept and its most prominent strengths are discussed, including the advantages of a formalized approach, the systematic collection of weight of evidence, the linkage of mechanisms to apical end points, the examination of the plausibility of epidemiological data, the identification of critical knowledge gaps and the design of mechanistic test methods. To prepare the ground for a broadened and appropriate use of AOPs, some widespread misconceptions are explained. Moreover, potential weaknesses and shortcomings of the current AOP rule set are addressed (1) to facilitate the discussion on its further evolution and (2) to better define appropriate vs. less suitable application areas. Exemplary toxicological studies are presented to discuss the linearity assumptions of AOP, the management of event modifiers and compensatory mechanisms, and whether a separation of toxicodynamics from toxicokinetics including metabolism is possible in the framework of pathway plasticity. Suggestions on how to compromise between different needs of AOP stakeholders have been added. A clear definition of open questions and limitations is provided to encourage further progress in the field.

Model-guided identification of a therapeutic strategy to reduce hyperammonemia in liver diseases.

BACKGROUND & AIMS: Recently, spatial-temporal/metabolic mathematical models have been established that allow the simulation of metabolic processes in tissues. We applied these models to decipher ammonia detoxification mechanisms in the liver. METHODS: An integrated metabolic-spatial-temporal model was used to generate hypotheses of ammonia metabolism. Predicted mechanisms were validated using time-resolved analyses of nitrogen metabolism, activity analyses, immunostaining and gene expression after induction of liver damage in mice. Moreover, blood from the portal vein, liver vein and mixed venous blood was analyzed in a time dependent manner. RESULTS: Modeling revealed an underestimation of ammonia consumption after liver damage when only the currently established mechanisms of ammonia detoxification were simulated. By iterative cycles of modeling and experiments, the reductive amidation of alpha-ketoglutarate (α-KG) via glutamate dehydrogenase (GDH) was identified as the lacking component. GDH is released from damaged hepatocytes into the blood where it consumes ammonia to generate glutamate, thereby providing systemic protection against hyperammonemia. This mechanism was exploited therapeutically in a mouse model of hyperammonemia by injecting GDH together with optimized doses of cofactors. Intravenous injection of GDH (720 U/kg), α-KG (280 mg/kg) and NADPH (180 mg/kg) reduced the elevated blood ammonia concentrations (>200 µM) to levels close to normal within only 15 min. CONCLUSION: If successfully translated to patients the GDH-based therapy might provide a less aggressive therapeutic alternative for patients with severe hyperammonemia.

TiQuant: software for tissue analysis, quantification and surface reconstruction.

MOTIVATION: TiQuant is a modular software tool for efficient quantification of biological tissues based on volume data obtained by biomedical image modalities. It includes a number of versatile image and volume processing chains tailored to the analysis of different tissue types which have been experimentally verified. TiQuant implements a novel method for the reconstruction of three-dimensional surfaces of biological systems, data that often cannot be obtained experimentally but which is of utmost importance for tissue modelling in systems biology. AVAILABILITY AND IMPLEMENTATION: TiQuant is freely available for non-commercial use at msysbio.com/tiquant. Windows, OSX and Linux are supported. CONTACT: hoehme@uni-leipzig.de SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Concomitant expression of far upstream element (FUSE) binding protein (FBP) interacting repressor (FIR) and its splice variants induce migration and invasion of non-small cell lung cancer (NSCLC) cells.

Transcription factors integrate a variety of oncogenic input information, facilitate tumour growth and cell dissemination, and therefore represent promising therapeutic target structures. Because over-expression of DNA-interacting far upstream element binding protein (FBP) supports non-small cell lung cancer (NSCLC) migration, we asked whether its repressor, FBP-interacting repressor (FIR) is functionally inactivated and how FIR might affect NSCLC cell biology. Different FIR splice variants were highly expressed in the majority of NSCLCs, with the highest levels in tumours carrying genomic gains of chromosome 8q24.3, which contained the FIR gene locus. Nuclear FIR expression was significantly enriched at the invasion front of primary NSCLCs, but this did not correlate with tumour cell proliferation. FIR accumulation was associated with worse patient survival and tumour recurrence; in addition, FIR over-expression significantly correlated with lymph node metastasis in squamous cell carcinomas (SCCs). In vitro, we applied newly developed methods and modelling approaches for the quantitative and time-resolved description of the pro-migratory and pro-invasive capacities of SCC cells. siRNA-mediated silencing of all FIR variants significantly reduced the speed and directional movement of tumour cells in all phases of migration. Furthermore, sprouting efficiency and single cell invasiveness were diminished following FIR inhibition. Interestingly, the silencing of FIR isoforms lacking exon 2 (FIR(Δexon2)) alone was sufficient to reduce lateral migration and invasion. In summary, by using scale-spanning data derived from primary human tissues, quantitative cellular analyses and mathematical modelling, we have demonstrated that concomitant over-expression of FIR and its splice variants drives NSCLC migration and dissemination.

Integrated metabolic spatial-temporal model for the prediction of ammonia detoxification during liver damage and regeneration.

UNLABELLED: The impairment of hepatic metabolism due to liver injury has high systemic relevance. However, it is difficult to calculate the impairment of metabolic capacity from a specific pattern of liver damage with conventional techniques. We established an integrated metabolic spatial-temporal model (IM) using hepatic ammonia detoxification as a paradigm. First, a metabolic model (MM) based on mass balancing and mouse liver perfusion data was established to describe ammonia detoxification and its zonation. Next, the MM was combined with a spatial-temporal model simulating liver tissue damage and regeneration after CCl4 intoxication. The resulting IM simulated and visualized whether, where, and to what extent liver damage compromised ammonia detoxification. It allowed us to enter the extent and spatial patterns of liver damage and then calculate the outflow concentrations of ammonia, glutamine, and urea in the hepatic vein. The model was validated through comparisons with (1) published data for isolated, perfused livers with and without CCl4 intoxication and (2) a set of in vivo experiments. Using the experimentally determined portal concentrations of ammonia, the model adequately predicted metabolite concentrations over time in the hepatic vein during toxin-induced liver damage and regeneration in rodents. Further simulations, especially in combination with a simplified model of blood circulation with three ammonia-detoxifying compartments, indicated a yet unidentified process of ammonia consumption during liver regeneration and revealed unexpected concomitant changes in amino acid metabolism in the liver and at extrahepatic sites. CONCLUSION: The IM of hepatic ammonia detoxification considerably improves our understanding of the metabolic impact of liver disease and highlights the importance of integrated modeling approaches on the way toward virtual organisms.

Modeling dynamics of cell-to-cell variability in TRAIL-induced apoptosis explains fractional killing and predicts reversible resistance.

Isogenic cells sensing identical external signals can take markedly different decisions. Such decisions often correlate with pre-existing cell-to-cell differences in protein levels. When not neglected in signal transduction models, these differences are accounted for in a static manner, by assuming randomly distributed initial protein levels. However, this approach ignores the a priori non-trivial interplay between signal transduction and the source of this cell-to-cell variability: temporal fluctuations of protein levels in individual cells, driven by noisy synthesis and degradation. Thus, modeling protein fluctuations, rather than their consequences on the initial population heterogeneity, would set the quantitative analysis of signal transduction on firmer grounds. Adopting this dynamical view on cell-to-cell differences amounts to recast extrinsic variability into intrinsic noise. Here, we propose a generic approach to merge, in a systematic and principled manner, signal transduction models with stochastic protein turnover models. When applied to an established kinetic model of TRAIL-induced apoptosis, our approach markedly increased model prediction capabilities. One obtains a mechanistic explanation of yet-unexplained observations on fractional killing and non-trivial robust predictions of the temporal evolution of cell resistance to TRAIL in HeLa cells. Our results provide an alternative explanation to survival via induction of survival pathways since no TRAIL-induced regulations are needed and suggest that short-lived anti-apoptotic protein Mcl1 exhibit large and rare fluctuations. More generally, our results highlight the importance of accounting for stochastic protein turnover to quantitatively understand signal transduction over extended durations, and imply that fluctuations of short-lived proteins deserve particular attention.

How predictive quantitative modelling of tissue organisation can inform liver disease pathogenesis.

From the more than 100 liver diseases described, many of those with high incidence rates manifest themselves by histopathological changes, such as hepatitis, alcoholic liver disease, fatty liver disease, fibrosis, and, in its later stages, cirrhosis, hepatocellular carcinoma, primary biliary cirrhosis and other disorders. Studies of disease pathogeneses are largely based on integrating -omics data pooled from cells at different locations with spatial information from stained liver structures in animal models. Even though this has led to significant insights, the complexity of interactions as well as the involvement of processes at many different time and length scales constrains the possibility to condense disease processes in illustrations, schemes and tables. The combination of modern imaging modalities with image processing and analysis, and mathematical models opens up a promising new approach towards a quantitative understanding of pathologies and of disease processes. This strategy is discussed for two examples, ammonia metabolism after drug-induced acute liver damage, and the recovery of liver mass as well as architecture during the subsequent regeneration process. This interdisciplinary approach permits integration of biological mechanisms and models of processes contributing to disease progression at various scales into mathematical models. These can be used to perform in silico simulations to promote unravelling the relation between architecture and function as below illustrated for liver regeneration, and bridging from the in vitro situation and animal models to humans. In the near future novel mechanisms will usually not be directly elucidated by modelling. However, models will falsify hypotheses and guide towards the most informative experimental design.

Protocols for staining of bile canalicular and sinusoidal networks of human, mouse and pig livers, three-dimensional reconstruction and quantification of tissue microarchitecture by image processing and analysis.

Histological alterations often constitute a fingerprint of toxicity and diseases. The extent to which these alterations are cause or consequence of compromised organ function, and the underlying mechanisms involved is a matter of intensive research. In particular, liver disease is often associated with altered tissue microarchitecture, which in turn may compromise perfusion and functionality. Research in this field requires the development and orchestration of new techniques into standardized processing pipelines that can be used to reproducibly quantify tissue architecture. Major bottlenecks include the lack of robust staining, and adequate reconstruction and quantification techniques. To bridge this gap, we established protocols employing specific antibody combinations for immunostaining, confocal imaging, three-dimensional reconstruction of approximately 100-µm-thick tissue blocks and quantification of key architectural features. We describe a standard procedure termed 'liver architectural staining' for the simultaneous visualization of bile canaliculi, sinusoidal endothelial cells, glutamine synthetase (GS) for the identification of central veins, and DAPI as a nuclear marker. Additionally, we present a second standard procedure entitled 'S-phase staining', where S-phase-positive and S-phase-negative nuclei (stained with BrdU and DAPI, respectively), sinusoidal endothelial cells and GS are stained. The techniques include three-dimensional reconstruction of the sinusoidal and bile canalicular networks from the same tissue block, and robust capture of position, size and shape of individual hepatocytes, as well as entire lobules from the same tissue specimen. In addition to the protocols, we have also established image analysis software that allows relational and hierarchical quantifications of different liver substructures (e.g. cells and vascular branches) and events (e.g. cell proliferation and death). Typical results acquired for routinely quantified parameters in adult mice (C57Bl6/N) include the hepatocyte volume (5,128.3 ± 837.8 µm(3)) and the fraction of the hepatocyte surface in contact with the neighbouring hepatocytes (67.4 ± 6.7 %), sinusoids (22.1 ± 4.8 %) and bile canaliculi (9.9 ± 3.8 %). Parameters of the sinusoidal network that we also routinely quantify include the radius of the sinusoids (4.8 ± 2.25 µm), the branching angle (32.5 ± 11.2°), the length of intersection branches (23.93 ± 5.9 µm), the number of intersection nodes per mm(3) (120.3 × 103 ± 42.1 × 10(3)), the average length of sinusoidal vessel per mm(3) (5.4 × 10(3) ± 1.4 × 10(3)mm) and the percentage of vessel volume in relation to the whole liver volume (15.3 ± 3.9) (mean ± standard deviation). Moreover, the provided parameters of the bile canalicular network are: length of the first-order branches (7.5 ± 0.6 µm), length of the second-order branches (10.9 ± 1.8 µm), length of the dead-end branches (5.9 ± 0.7 µm), the number of intersection nodes per mm(3) (819.1 × 10(3) ± 180.7 × 10(3)), the number of dead-end branches per mm(3) (409.9 × 10(3) ± 95.6 × 10(3)), the length of the bile canalicular network per mm(3) (9.4 × 10(3) ± 0.7 × 10(3) mm) and the percentage of the bile canalicular volume with respect to the total liver volume (3.4 ± 0.005). A particular strength of our technique is that quantitative parameters of hepatocytes and bile canalicular as well as sinusoidal networks can be extracted from the same tissue block. Reconstructions and quantifications performed as described in the current protocols can be used for quantitative mathematical modelling of the underlying mechanisms. Furthermore, protocols are presented for both human and pig livers. The technique is also applicable for both vibratome blocks and conventional paraffin slices.