Daniel Seeler (1,2,3), Nastasja Grdseloff (2), Claudia Jasmin Rödel (2), Charlotte Kloft (4), Salim Abdelilah-Seyfried (2), Wilhelm Huisinga (1,2)
(1) Institute of Mathematics, University of Potsdam, Germany; (2) Institute of Biochemistry and Biology, University of Potsdam, Germany; (3) Graduate Research Training Program PharMetrX: Pharmacometrics & Computational Disease Modelling, Berlin/Potsdam, Germany; (4) Institute of Pharmacy, Freie Universität Berlin, Germany
Objectives: The geometry of the vasculature is largely influenced by its biophysical environment, which impacts the morphology of endothelial cells (ECs) lining the vessel interior. In response to strong blood flow, ECs elongate and align their cell junctions in parallel to the flow direction. These cell morphological changes impact vessel diameter and thus blood flow dynamics [1]. The result is a feedback cycle where ECs detect biophysical changes and adapt their shapes.
Our objective is to better understand this feedback cycle during vasculogenesis with the long-term aim to comprehend vascular dysregulation in human diseases. As a model system for the developing vasculature, we characterize the dorsal aorta (DA) of zebrafish embryos. Previous geometric models of the DA (i) have either projected isolated ECs onto elliptic vessel cross-sections (CSs) and then unrolled the cylinder to determine EC morphometrics [1], or (ii) modeled the vessel by varying elliptic CSs without integrating EC morphology [2]. Here, we present a novel approach that leverages EC outlines and junctional contacts to reconstruct the 3D vessel geometry without cartographic distortions created by unrolling.
Methods: To obtain EC contours for model development, we used transgenic lines of zebrafish expressing PECAM1-EGFP as a junctional marker. We imaged ECs of the DA in wild-type zebrafish embryos at 48 and 72hours post fertilization (hpf) using 3D confocal microscopy. 3D coordinates on cell junctions were manually annotated per EC in Imaris software (Bitplane Inc.). Our aim was to use this junctional data to first estimate CS shapes of the DA, interpolate between these to reconstruct the vessel and then divide it into its constituting ECs.
We first fitted smoothing splines to enrich recorded EC outline data with interpolated data. Next, per spline equidistant points were extracted for fitting the CSs of the DA. To account for the previously known dorsal-ventral asymmetry of the DA [3] and for the left-right asymmetry found in our data, we developed a CS model consisting of four quarter superellipses. As our data was still sparse per CS, we included neighboring CSs in the fitting process using Gaussian weights decaying with distance. To ensure robust estimation, we chose the width of the weight function adaptively in space. We investigated the impact of manual annotation by comparing repeated cell contour annotations of the same DA. After projecting each EC contour onto the estimated vessel CSs, we computed an adaptive triangulation containing all edges of the projected EC contours.
Results: We were able to fit a smoothing spline to each cell contour with low residual errors. Smoothing allowed us to filter outliers created by the manual annotation process and reduce the overlap of neighboring ECs. Interpolated data from the cell contour splines improved the identifiability of CSs. The estimated CSs had low projection errors.
Both data annotations resulted in low projection errors and CS diameters of similar magnitude, making our geometric model robust to annotation errors. We inferred a decrease of vessel diameter between 48hpf and 72hpf, which is in line with experimental findings [1,4]. Extracting a mesh for each EC enabled us to perform individual cell morphometric analyses. We found that EC areas and perimeters were in the range of previous experimental observations [1]. Between 48hpf and 72hpf ECs elongate and thereby increase the length of cell contact zones, that are oriented parallel to the direction of blood flow.
Conclusions: Our geometric model allows a high-precision analysis of both 3D blood vessel geometry and EC morphology without cartographic distortions. Next, we aim to employ this modeling approach to quantify the effects of mutant genotypes and drug treatments on EC morphometrics in zebrafish embryos. Ultimately, we plan to investigate feedback mechanisms by leveraging the geometric model’s simple parameterization. After changing parameters of the vessel’s CSs, EC morphology can be recomputed.
One promising application of our model are cerebral cavernous malformations, which are dilated, blood-filled lesions primarily found in capillaries and small veins of the brain vasculature. These vascular malformations are prone to bleed and can thus result in hemorrhagic stroke. This disease is characterized by aberrant blood flow sensing leading to pathological EC morphology. Currently, no drug therapy is approved for its treatment [5].
References:
[1] Sugden, W. W.; Meissner, R.; Aegerter-Wilmsen, T.; Tsaryk, R.; Leonard, E. V.; Bussmann, J.; Hamm, M. J.; Herzog, W.; Jin, Y.; Jakobsson, L.; Denz, C. & Siekmann, A. F., Endoglin controls blood vessel diameter through endothelial cell shape changes in response to haemodynamic cues, Nature Cell Biology, 2017, 19, 653-665
[2] Campinho, P.; Lamperti, P.; Boselli, F. & Vermot, J., Three-dimensional microscopy and image analysis methodology for mapping and quantification of nuclear positions in tissues with approximate cylindrical geometry, Philosophical Transactions of the Royal Society of London. Series B, Biological sciences, 2018, 373
[3] Campinho, P.; Lamperti, P.; Boselli, F.; Vilfan, A. & Vermot, J., Blood Flow Limits Endothelial Cell Extrusion in the Zebrafish Dorsal Aorta, Cell Reports, 2020, 31, 107505
[4] Kissa, K.; Herbomel, P. Blood stem cells emerge from aortic endothelium by a novel type of cell transition, Nature, 2010;464(7285):112-5.
[5] Abdelilah-Seyfried, S.; Tournier-Lasserve, E. & Derry, W. B., Blocking Signalopathic Events to Treat Cerebral Cavernous Malformations, Trends in Molecular Medicine, Elsevier BV, 2020, 26, 874-887
Reference: PAGE 29 (2021) Abstr 9820 [www.page-meeting.org/?abstract=9820]
Poster: Drug/Disease Modelling - Other Topics