LiverMap pipeline enables multilobular morphometric 3D imaging and reveals dysregulation of structural and functional zonation within the cirrhotic human liver

Introduction: The liver is considerably complex both structurally and functionally, and its unique microstructure is critical for its proper function and homeostasis. In liver cirrhosis, the liver’s intricately organized microvasculature can become massively rearranged. Yet, non-diseased human liver tissue and cirrhotic microvascular rearrangement has never been visualized in 3D on the multilobular scale in human tissue. Prior to our work, most of the liver’s structure on the cellular level had been elucidated via two-dimensional (2D) traditional histology, which has several limitations. For one, sectioning tissue into thin slices can physically alter the examined structures. Beyond that, the 2D nature of histology can misrepresent actual biological structures: 1) 3D convoluted structures often appear as multiple distinct objects in a 2D cross-section, 2) cross-sections may not accurately capture the distribution of objects within a tissue volume, and 3) 2D slices may simply not contain rare or sparse objects [1]. Thus, we hypothesized that multilobular 3D mapping of multiscale architectural and functional features would enable morphometry of normal human liver and unveil volumetric pathological remodeling in human cirrhosis. The disruption of the intricate architecture in the liver is pathologic, and studying how changes in architecture correlate with disease states may provide avenues for understanding liver disease and lead to possible strategies for treatment. Furthermore, the ability to more accurately model 3D human liver tissue in its normal and disease states may offer new insights for liver tissue engineering and provide 3D blueprints for reconstructing human liver tissue.


Materials and Methods: This work introduces the LiverMap pipeline, which enables 3D multilobular visualization and quantification of structural and functional features within human liver tissue. First, liver samples were excised from human patients undergoing partial hepatectomy or liver transplantation. Samples from oncologic partial hepatectomy were dissected at least two centimeters away from any present tumor, a distance at which tissue is expected to be free from field effect [2, 3]. These liver samples were then sliced to a thickness of 2mm via a vibratome and fixed in paraformaldehyde. After rinsing, samples were punched into 6mm discs with biopsy punches, a size expected to contain approximately four to six human lobules. Each patient sample was biopsy-punched twice to establish two separate sets of non-fibrotic and cirrhotic tissues. One set of tissue samples was then immunostained with the Zonation Panel (cytokeratin 19 [CK19] for bile ducts; glutamine synthetase [GLUL] for pericentral hepatocytes; and CD32b for type II sinusoids). The other set of tissue samples was immunostained with the Architecture Panel (CK19; smooth muscle actin [SMA] for portal veins/central venules; and CD31 for hepatic arterioles). All tissue sections were then optically cleared using our modified, liver-optimized version of the Ce3D protocol [4]. After staining and clearing, the optically transparent tissue samples were imaged at high resolution with a confocal laser scanning microscope (Leica SP8). Lastly, Imaris (Bitplane) and Vesselucida (MBF Biosciences) software were utilized to computationally reconstruct 3D human liver images for enhanced visualization and quantification.


Results, Conclusion, and Discussion: From our multilobular analysis, we found a reduction in expression of GLUL in cirrhotic tissue compared to non-fibrotic tissue, which may indicate decreased ammonia metabolism [5]. We also observed a spatial dysregulation of CD32b expression in cirrhotic tissue, suggesting a shift in the ratio of Type I to Type II liver sinusoidal endothelial cells (LSECs). This observation implies an imbalance in sinusoid homeostasis and likely entails functional repercussions, such as increased lipoprotein scavenging and decreased functions related to blood-derived leukocyte adhesion and antigen clearance [6]. Additionally, we found that cirrhotic tissue had a significantly larger quantity of small CK19+ ductule segments, which may have implications for biliary pressure and flow. We also found that the cirrhotic tissue had a significantly lower quantity of central venules, which may have severe functional implications, such as dysfunctional blood drainage that may contribute to portal vein hypertension. Furthermore, we also observed that CD31+ vessel networks expand in cirrhotic tissue, which suggests a severely dysfunctional endothelial phenotype called capillarization and correlates with loss of sinusoid fenestrations. Together, the significant reductions in central venule network volume and GLUL expression in conjunction with the expansions of CD31+ vessel networks and CK19+ bile duct networks suggest an overall decentralization/proportalization phenotype in human cirrhosis.

In this work, we introduce the LiverMap pipeline and present the first multilobular 3D visualization and volumetric characterization of sparse and convoluted structures within human liver tissue. Here, we surpassed the image size of existing 3D human liver imaging studies by about 5X through imaging a depth equivalent to ~100 traditional histology tissue sections and by roughly 200X the previously captured tissue volume at cellular resolution. With this novel capability, we examine both non-fibrotic and cirrhotic human liver tissue on the multilobular scale which enables the identification of significant architectural abnormalities associated with cirrhosis. Our 3D models suggest a decentralization/proportalization phenotype within human cirrhosis and identify that these structural changes correlate with functional zonation dysregulation. The LiverMap pipeline may enable more accurate modeling and engineering of 3D human liver tissue as well as provide new insights into the histopathology of cirrhosis in the human liver.

References: [1] J. T. C. Liu, A. K. Glaser, K. Bera, L. D. True, N. P. Reder, K. W. Eliceiri and A. Madabhushi, "Harnessing non-destructive 3D pathology," Nature Biomedical Engineering, 2021.

[2] H. Chai and R. E. Brown, "Field Effect in Cancer–An Update," Annals of Clinical & Laboratory Science, vol. 39, no. 4, 2009.

[3] Y. Hoshida, A. Villanueva and J. M. Llovet, "Molecular profiling to predict hepatocellular carcinoma outcome," Expert Review of Gastroenterology & Hepatology, vol. 3, no. 2, 2009.

[4] W. Li, R. N. Germain and M. Y. Gerner, "High-dimensional cell-level analysis of tissues with Ce3D multiplex volume imaging," Nature Protocols, 2019.

[5] Y. Zhou, T. Eid, B. Hassel and N. C. Danbolt, "Novel aspects of glutamine synthetase in ammonia homeostasis," Neurochemistry International, 2020.

[6] O. Strauss, A. Phillips, K. Ruggiero, A. Bartlett and R. Dunbar, "Immunofluorescence identifies distinct subsets of endothelial cells in the human liver," Scientific Reports, vol. 7, 2017.