Correlative 3D Imaging and Microfluidic Modelling of Human Pulmonary Lymphatics using Immunohistochemistry and High-resolution μCT

Abstract

Lung lymphatics maintain fluid homoeostasis by providing a drainage system that returns fluid, cells and metabolites to the circulatory system. The 3D structure of the human pulmonary lymphatic network is essential to lung function, but it is poorly characterised. Image-based 3D mathematical modelling of pulmonary lymphatic microfluidics has been limited by the lack of accurate and representative image geometries. This is due to the microstructural similarity of the lymphatics to the blood vessel network, the lack of lymphatic-specific biomarkers, the technical limitations associated with image resolution in 3D, and sectioning artefacts present in 2D techniques. We present a method that combines lymphatic specific (D240 antibody) immunohistochemistry (IHC), optimised high-resolution X-ray microfocus computed tomography (μCT) and finite-element mathematical modelling to assess the function of human peripheral lung tissue. The initial results identify lymphatic heterogeneity within and between lung tissue. Lymphatic vessel volume fraction and fractal dimension significantly decreases away from the lung pleural surface (p < 0.001, n = 25 and p < 0.01, n = 20, respectively). Microfluidic modelling successfully shows that in lung tissue the fluid derived from the blood vessels drains through the interstitium into the lymphatic vessel network and this drainage is different in the subpleural space compared to the intralobular space. When comparing lung tissue from health and disease, human pulmonary lymphatics were significantly different across five morphometric measures used in this study (p ≤ 0.0001). This proof of principle study establishes a new engineering technology and workflow for further studies of pulmonary lymphatics and demonstrates for the first time the combination of correlative μCT and IHC to enable 3D mathematical modelling of human lung microfluidics at micrometre resolution.

Figure 1

3D rendering and histological section alignment. (a) A 3D rendered image of the μCT scanned control human lung tissue block. Scale bars – Blue: 15.84 mm, Green and Red: 21.1 mm. (b) Image a plus an orthoslice section (red) located at the approximate position of the manual microtome sectioning. Blue: 12.2 mm, Green and Red: 21.1 mm. (c) The first of 20 serial stained tissue slices at 20x magnification. Distortion of tissue compared with image (d) can be seen width-wise. (d). An orthoslice view of the newly registered μCT data set to the histological slice.

Figure 2

Feature segmentation. (a) A region of interest (ROI) of the histological slice showing the tissues stained in blue and the lymphatic vessels outlined in brown. A blood vessel is identifiable in the centre of the image by the dense connective tissue surrounding a smooth lumen. (b) The corresponding μCT ROI to image a with the segmented lymphatic vessels outlined in green and the blood vessel marked in red. (c) A rendered 3D image of segmented features within the μCT data set. Scale bars: red = 13.1 mm; green = 10.3 mm; blue = 0.5 mm) White arrow indicates the lymphatic vessel encircling a blood vessel. Lymphatics- Green, Blood vessels- Red, and Interstitial tissue –white.

Figure 3

Schematic Representation of the Model in 2D for Pulmonary Fluid Flow. The blood vessel is shown in red (BV), the lymphatic vessel in green (LV) and the interstitial tissue in white. The vessels are described by Stokes’ flow and the interstitial tissue by Darcy’s Law. The dotted lines represent a flux boundary condition has been applied. Constrains of the model are shown by purple boundaries where an interstitial pressure condition was applied. All other symbols and initial parameter inputs are given in the text.

Figure 4

Spatial Morphometry of Pulmonary lymphatics. (a) The output of the ROIs given from macro-script are shown in red. The background data represents the binary image produced from the Z-stack maximum projection of the lymphatic segmentation of the healthy lung tissue. (b) The volume fraction of 25 lymphatic VOI plotted against their distance from the pleural surface of the lung tissue. (c) The fractal dimension of 20 VOI and their distance from the pleural surface of the lung tissue.

Figure 5

Quantitive morphology of pulmonary lymphatics between control and diseased tissue. A box and whisker plot of the lymphatic volume fraction (a) surface area/tissue volume ratio (b) fractal dimension (c) branch number (d) junction number (e) and tortuosity (f) between the control and diseased tissue. **** = p ≤ 0.0001. ns = not significant. Whisker plots show full range (min-max).

Figure 6

3D model simulation result for flow in an intralobular and subpleural lung geometry. A graphical representation for the control sample of the geometry and solution for the static pressure (Pa) within the lymphatic vessel and blood vessel of the intralobular VOI (a and b respectively), and subpleural VOI (e and f respectively). The interstitial pressure results have been removed from view so the streamlines could be visualised. (c,d,g,h) A close-up view of a part of images (a,b,e,f) respectively. Streamlines of the Darcy’s velocity field into the interstitium are shown in red. The direction of fluid flow is shown by black arrows. The VOIs in (a,b,e,f) have a dimension of 830 × 830 × 830 μm. In the geometry images the blood vessels = blue and the lymphatic vessel = green.