3-dimensional imaging and mathematical modelling of human pulmonary lymphatics

Abstract

Lymphatics drain fluid, cells, and essential metabolites back to the circulatory system to maintain fluid homoeostasis. The fluid balance is critical for normal tissue function and is often disturbed in lung diseases such as COPD, asthma, and lung cancer. Although 6.5 million people in the UK are undergoing active treatment for these diseases, little is known about how lymphatics influence lung fluid dynamics.This study developed and applied a methodology to assess the anatomical structure and predict the function of the 3D lymphatic network in healthy and diseased peripheral human lung tissue.High-resolution X-ray micro-computed tomography (μCT) provided the 3D structural blueprint of the lung samples down to a voxel size of 5 μm. The pulmonary lymphatic network was subsequently identified and segmented within the same tissue using immunohistochemistry and anti-podoplanin antibodies (D240 and LP21). These micrographs were used to manually segment the lymphatic structures within the μCT dataset alongside the blood vessels, airways and lung parenchyma.A physiologically accurate image-based mathematical model was established using these geometries. Flow in the blood and microlymphatics are described using the Stokes equations and flow through the interstitium using Darcy's law. Starling’s law describes flow across the blood and lymphatic vessel walls, with a uni-directional restraint given to the lymphatic vessel wall to represent the primary valve condition. In addition to changes in input geometry, the model shows sensitivity to pulmonary blood vessel pressure and vessel flow velocities.To validate the methodology, ten formalin-fixed, paraffin-embedded peripheral lung biopsies were used, four controls, four from patients with acute respiratory injury or inflammation and two with severe COPD disease. Inter- and intra-sample lymphatic heterogeneity was assessed by vessel volume, vessel surface area, network branch and junction number, vessel tortuosity and fractal dimension. Two hierarchical lymphatic populations were identified; precollecting lymphatics and microlymphatics. Three microlymphatic subpopulations were also defined based on their structural heterogeneity and anatomical location in the lung: intralobular(IL); subpleural (SP) and bronchovascular(BVB). This suggested pulmonary fluid dynamics may vary in different anatomical locations. In the active disease group, five of the six microlymphatic measures were distinct to both the COPD and control groups (P<0.0001). Whereas, no morphological differences were seen between the control and COPD sample groups, suggesting fluid dynamics in the control and COPD samples would be similar, whilst those in the active disease sample would be different.To test the predictions from the morphometry assessment, a finite element mesh was created for the lymphatics, blood vessels and parenchyma in five random volumes from the IL, SP and BVB regions in the control group, and IL region of the acute disease and COPD samples. Using COMSOL Multiphysics ©, finite element simulation software and the mathematical model, fluid flow within the lung geometries were then simulated and compared. Modelling results predict fluid drainage does vary between anatomical areas of the peripheral lung but in the IL region, the COPD geometries drain fluid less efficiently than the control (0.0015 ml.min1 /mm3 lymphatic volume and 0.0024 ml.min-1 /mm3 lymphatic volume respectively). The model considers both blood vessel and interstitial structural variation and their proximity to the lymphatics in 3D space. This more sophisticated analysis accounts for some of the discrepancies between the morphometry assessment and the modelling predictions. This highlights the benefit of incorporating image-based modelling into biological systems to provide higher-level consideration of structure function relationships of the whole tissue.

More information

Identifiers

ORCID for Stephanie Robinson: orcid.org/0000-0001-5436-2929

ORCID for Tiina Roose: orcid.org/0000-0001-8710-1063

Catalogue record

Export record