Analysis of the glomerular basement membrane in images of renal biopsies using the split-and-merge method: a pilot study

Digital pathology assessment of kidney glomerular filtration barrier ultrastructure in an animal model of podocytopathy

Background

Transmission electron microscopy (TEM) images can visualize kidney glomerular filtration barrier ultrastructure, including the glomerular basement membrane (GBM) and podocyte foot processes (PFP). Podocytopathy is associated with glomerular filtration barrier morphological changes observed experimentally and clinically by measuring GBM or PFP width. However, these measurements are currently performed manually. This limits research on podocytopathy disease mechanisms and therapeutics due to labor intensiveness and inter-operator variability.

Methods

We developed a deep learning-based digital pathology computational method to measure GBM and PFP width in TEM images from the kidneys of Integrin-Linked Kinase (ILK) podocyte-specific conditional knockout (cKO) mouse, an animal model of podocytopathy, compared to wild-type (WT) control mouse. We obtained TEM images from WT and ILK cKO littermate mice at 4 weeks old. Our automated method was composed of two stages: a U-Net model for GBM segmentation, followed by an image processing algorithm for GBM and PFP width measurement. We evaluated its performance with a 4-fold cross-validation study on WT and ILK cKO mouse kidney pairs.

Results

Mean (95% confidence interval) GBM segmentation accuracy, calculated as Jaccard index, was 0.73 (0.70-0.76) for WT and 0.85 (0.83-0.87) for ILK cKO TEM images. Automated and manual GBM width measurements were similar for both WT (p=0.49) and ILK cKO (p=0.06) specimens. While automated and manual PFP width measurements were similar for WT (p=0.89), they differed for ILK cKO (p<0.05) specimens. WT and ILK cKO specimens were morphologically distinguishable by manual GBM (p<0.05) and PFP (p<0.05) width measurements. This phenotypic difference was reflected in the automated GBM (p<0.05) more than PFP (p=0.06) widths.

Conclusions

These results suggest that certain automated measurements enabled via deep learning-based digital pathology tools could distinguish healthy kidneys from those with podocytopathy. Our proposed method provides high-throughput, objective morphological analysis and could facilitate podocytopathy research and translate into clinical diagnosis.

An ultrathin membrane mediates tissue-specific morphogenesis and barrier function in a human kidney chip

Organ-on-chip (OOC) systems are revolutionizing tissue engineering by providing dynamic models of tissue structure, organ-level function, and disease phenotypes using human cells. However, nonbiological components of OOC devices often limit the recapitulation of in vivo-like tissue-tissue cross-talk and morphogenesis. Here, we engineered a kidney glomerulus-on-a-chip that recapitulates glomerular morphogenesis and barrier function using a biomimetic ultrathin membrane and human-induced pluripotent stem cells. The resulting chip comprised a proximate epithelial-endothelial tissue interface, which reconstituted the selective molecular filtration function of healthy and diseased kidneys. In addition, fenestrated endothelium was successfully induced from human pluripotent stem cells in an OOC device, through in vivo-like paracrine signaling across the ultrathin membrane. Thus, this device provides a dynamic tissue engineering platform for modeling human kidney-specific morphogenesis and function, enabling mechanistic studies of stem cell differentiation, organ physiology, and pathophysiology.

Roles of myosin 1e and the actin cytoskeleton in kidney functions and familial kidney disease

Mammalian kidneys are responsible for removing metabolic waste and maintaining fluid and electrolyte homeostasis via selective filtration. One of the proteins closely linked to selective renal filtration is myosin 1e (Myo1e), an actin-dependent molecular motor found in the specialized kidney epithelial cells involved in the assembly and maintenance of the renal filter. Point mutations in the gene encoding Myo1e, MYO1E, have been linked to familial kidney disease, and Myo1e knockout in mice leads to the disruption of selective filtration. In this review, we discuss the role of the actin cytoskeleton in renal filtration, the known and hypothesized functions of Myo1e, and the possible explanations for the impact of MYO1E mutations on renal function.

Alport Syndrome: Clinical Spectrum and Therapeutic Advances

Alport syndrome is a hereditary disorder characterized by kidney disease, ocular abnormalities, and sensorineural hearing loss. Work in understanding the cause of Alport syndrome and the molecular composition of the glomerular basement membrane ultimately led to the identification of COL4A3, COL4A4 (both on chromosome 2q36), and COL4A5 (chromosome Xq22), encoding the α3, α4, and α5 chains of type IV collagen, as the responsible genes. Subsequent studies suggested that autosomal recessive Alport syndrome and males with X-linked Alport syndrome have more severe disease, whereas autosomal dominant Alport syndrome and females with X-linked Alport syndrome have more variability. Variant type is also influential-protein-truncating variants in autosomal recessive Alport syndrome or males with X-linked Alport syndrome often present with severe symptoms, characterized by kidney failure, extrarenal manifestations, and lack of the α3-α4-α5(IV) network. By contrast, mild-moderate forms from missense variants display α3-α4-α5(IV) in the glomerular basement membrane and are associated with protracted kidney involvement without extrarenal manifestations. Regardless of type, therapeutic intervention for kidney involvement is focused on early initiation of angiotensin-converting enzyme inhibitors. There are several therapies under investigation including sodium/glucose cotransporter 2 inhibitors, aminoglycoside analogs, endothelin type A antagonists, lipid-modifying drugs, and hydroxychloroquine, although targeting the underlying defect through gene therapy remains in preclinical stages.

A comparative study of human glomerular basement membrane thickness using direct measurement and orthogonal intercept methods

INTRODUCTION

Here we report estimates of glomerular basement membrane (GBM) thickness in the Brazilian population performed using direct (DM) and orthogonal interception methods (OIM), and comment on potential sources of variation among estimates made by different laboratories.

METHODOLOGY

A total of 38 patients, ranging from 3 to 78 years of age, 26 (68%) males and 12 (32%) females, were submitted to kidney biopsy procedures for renal disease diagnosis. Glomeruli were diagnosed with minor histological changes by conventional, immunofluorescence and electron microscopy. GBM thickness was estimated using both DM and OIM methods.

RESULTS

Estimates of GBM thickness obtained using DM were higher than those obtained by OIM. However, the application of a correction for non-perpendicular membrane sectioning to DM estimates yielded similar results to those obtained under OIM. The estimated GMB thickness using DM after correction was 289 + 44 nm, versus 287 + 48 nm by OIM. No statistically significant differences were detected in GMB thickness, nor with respect to patient age or sex.

CONCLUSIONS

GBM thickness in the studied Brazilian population measured approximately 290 nm. The application of criteria for estimating the shortest distance between the endothelial and podocyte cell membranes with correction for non-perpendicular membrane sectioning can increase the accuracy of GBM thickness estimates using DM and OIM.

Automatic Segmentation of Pathological Glomerular Basement Membrane in Transmission Electron Microscopy Images with Random Forest Stacks

Pathological classification through transmission electron microscopy (TEM) is essential for the diagnosis of certain nephropathy, and the changes of thickness in glomerular basement membrane (GBM) and presence of immune complex deposits in GBM are often used as diagnostic criteria. The automatic segmentation of the GBM on TEM images by computerized technology can provide clinicians with clear information about glomerular ultrastructural lesions. The GBM region on the TEM image is not only complicated and changeable in shape but also has a low contrast and wide distribution of grayscale. Consequently, extracting image features and obtaining excellent segmentation results are difficult. To address this problem, we introduce a random forest- (RF-) based machine learning method, namely, RF stacks (RFS), to realize automatic segmentation. Specifically, this work proposes a two-level integrated RFS that is more complicated than a one-level integrated RF to improve accuracy and generalization performance. The integrated strategies include training integration and testing integration. Training integration can derive a full-view RFS₁ by simultaneously sampling several images of different grayscale ranges in the train phase. Testing integration can derive a zoom-view RFS₂ by separately sampling the images of different grayscale ranges and integrating the results in the test phase. Experimental results illustrate that the proposed RFS can be used to automatically segment different morphologies and gray-level basement membranes. Future study on GBM thickness measurement and deposit identification will be based on this work.

Measurement of glomerulus diameter and Bowman's space width of renal albino rats

Glomerulus diameter and Bowman's space width in renal microscopic images indicate various diseases. Therefore, the detection of the renal corpuscle and related objects is a key step in histopathological evaluation of renal microscopic images. However, the task of automatic glomeruli detection is challenging due to their wide intensity variation, besides the inconsistency in terms of shape and size of the glomeruli in the renal corpuscle. Here, a novel solution is proposed which includes the Particles Analyzer technique based on median filter for morphological image processing to detect the renal corpuscle objects. Afterwards, the glomerulus diameter and Bowman's space width are measured. The solution was tested with a dataset of 21 rats' renal corpuscle images acquired using light microscope. The experimental results proved that the proposed solution can detect the renal corpuscle and its objects efficiently. As well as, the proposed solution has the ability to manage any input images assuring its robustness to the deformations of the glomeruli even with the glomerular hypertrophy cases. Also, the results reported significant difference between the control and affected (due to ingested additional daily dose (14.6mg) of fructose) groups in terms of glomerulus diameter (97.40±19.02μm and 177.03±54.48μm, respectively).

3D imaging of the early embryonic chicken heart with focused ion beam scanning electron microscopy

Early embryonic heart development is a period of dynamic growth and remodeling, with rapid changes occurring at the tissue, cell, and subcellular levels. A detailed understanding of the events that establish the components of the heart wall has been hampered by a lack of methodologies for three-dimensional (3D), high-resolution imaging. Focused ion beam scanning electron microscopy (FIB-SEM) is a novel technology for imaging 3D tissue volumes at the subcellular level. FIB-SEM alternates between imaging the block face with a scanning electron beam and milling away thin sections of tissue with a FIB, allowing for collection and analysis of 3D data. FIB-SEM was used to image the three layers of the day 4 chicken embryo heart: myocardium, cardiac jelly, and endocardium. Individual images obtained with FIB-SEM were comparable in quality and resolution to those obtained with transmission electron microscopy. Up to 1,100 serial images were obtained in 4 nm increments at 4.88 nm resolution, and image stacks were aligned to create volumes 800-1,500 μm3 in size. Segmentation of organelles revealed their organization and distinct volume fractions between cardiac wall layers. We conclude that FIB-SEM is a powerful modality for 3D subcellular imaging of the embryonic heart wall.

α1β1 integrin/Rac1-dependent mesangial invasion of glomerular capillaries in Alport syndrome

Alport syndrome, hereditary glomerulonephritis with hearing loss, results from mutations in type IV collagen COL4A3, COL4A4, or COL4A5 genes. The mechanism for delayed glomerular disease onset is unknown. Comparative analysis of Alport mice and CD151 knockout mice revealed progressive accumulation of laminin 211 in the glomerular basement membrane. We show mesangial processes invading the capillary loops of both models as well as in human Alport glomeruli, as the likely source of this laminin. L-NAME salt-induced hypertension accelerated mesangial cell process invasion. Cultured mesangial cells showed reduced migratory potential when treated with either integrin-linked kinase inhibitor or Rac1 inhibitor, or by deletion of integrin α1. Treatment of Alport mice with Rac1 inhibitor or deletion of integrin α1 reduced mesangial cell process invasion of the glomerular capillary tuft. Laminin α2-deficient Alport mice show reduced mesangial process invasion, and cultured laminin α2-null cells showed reduced migratory potential, indicating a functional role for mesangial laminins in progression of Alport glomerular pathogenesis. Collectively, these findings predict a role for biomechanical insult in the induction of integrin α1β1-dependent Rac1-mediated mesangial cell process invasion of the glomerular capillary tuft as an initiation mechanism of Alport glomerular pathology.