The application of deep learning for automated segmentation (delineation of boundaries) of histologic primitives (structures) from whole slide images can facilitate the establishment of novel ...protocols for kidney biopsy assessment. Here, we developed and validated deep learning networks for the segmentation of histologic structures on kidney biopsies and nephrectomies. For development, we examined 125 biopsies for Minimal Change Disease collected across 29 NEPTUNE enrolling centers along with 459 whole slide images stained with Hematoxylin & Eosin (125), Periodic Acid Schiff (125), Silver (102), and Trichrome (107) divided into training, validation and testing sets (ratio 6:1:3). Histologic structures were manually segmented (30048 total annotations) by five nephropathologists. Twenty deep learning models were trained with optimal digital magnification across the structures and stains. Periodic Acid Schiff-stained whole slide images yielded the best concordance between pathologists and deep learning segmentation across all structures (F-scores: 0.93 for glomerular tufts, 0.94 for glomerular tuft plus Bowman’s capsule, 0.91 for proximal tubules, 0.93 for distal tubular segments, 0.81 for peritubular capillaries, and 0.85 for arteries and afferent arterioles). Optimal digital magnifications were 5X for glomerular tuft/tuft plus Bowman’s capsule, 10X for proximal/distal tubule, arteries and afferent arterioles, and 40X for peritubular capillaries. Silver stained whole slide images yielded the worst deep learning performance. Thus, this largest study to date adapted deep learning for the segmentation of kidney histologic structures across multiple stains and pathology laboratories. All data used for training and testing and a detailed online tutorial will be publicly available.
Display omitted
In patients of African ancestry, genetic variants in APOL1, which encodes apolipoprotein L1, associate with the nondiabetic kidney diseases, focal segmental glomerulosclerosis (FSGS), HIV-associated ...nephropathy (HIVAN), and hypertensive nephropathy. Understanding the renal localization of APOL1 may provide clues that will ultimately help elucidate the mechanisms by which APOL1 variants promote nephropathy. Here, we used immunohistology to examine APOL1 localization in normal human kidney sections and in biopsies demonstrating either FSGS (n = 8) or HIVAN (n = 2). Within normal glomeruli, APOL1 only localized to podocytes. Compared with normal glomeruli, fewer cells stained for APOL1 in FSGS and HIVAN glomeruli, even when expression of the podocyte markers GLEPP1 and synaptopodin appeared normal. APOL1 localized to proximal tubular epithelia in normal kidneys, FSGS, and HIVAN. We detected APOL1 in the arteriolar endothelium of normal and diseased kidney sections. Unexpectedly, in both FSGS and HIVAN but not normal kidneys, the media of medium artery and arterioles contained a subset of α-smooth muscle actin-positive cells that stained for APOL1. Comparing the renal distribution of APOL1 in nondiabetic kidney disease to normal kidney suggests that a previously unrecognized arteriopathy may contribute to disease pathogenesis in patients of African ancestry.
African polymorphisms in the gene for Apolipoprotein L1 (APOL1) confer a survival advantage against lethal trypanosomiasis but also an increased risk for several chronic kidney diseases (CKD) ...including HIV-associated nephropathy (HIVAN). APOL1 is expressed in renal cells, however, the pathogenic events that lead to renal cell damage and kidney disease are not fully understood. The podocyte function of APOL1-G0 versus APOL1-G2 in the setting of a known disease stressor was assessed using transgenic mouse models. Transgene expression, survival, renal pathology and function, and podocyte density were assessed in an intercross of a mouse model of HIVAN (Tg26) with two mouse models that express either APOL1-G0 or APOL1-G2 in podocytes. Mice that expressed HIV genes developed heavy proteinuria and glomerulosclerosis, and had significant losses in podocyte numbers and reductions in podocyte densities. Mice that co-expressed APOL1-G0 and HIV had preserved podocyte numbers and densities, with fewer morphologic manifestations typical of HIVAN pathology. Podocyte losses and pathology in mice co-expressing APOL1-G2 and HIV were not significantly different from mice expressing only HIV. Podocyte hypertrophy, a known compensatory event to stress, was increased in the mice co-expressing HIV and APOL1-G0, but absent in the mice co-expressing HIV and APOL1-G2. Mortality and renal function tests were not significantly different between groups. APOL1-G0 expressed in podocytes may have a protective function against podocyte loss or injury when exposed to an environmental stressor. This was absent with APOL1-G2 expression, suggesting APOL1-G2 may have lost this protective function.
Chronic kidney disease (CKD) is a complex disease affecting more than 20 million individuals in the United States. Progression of CKD is associated with a number of serious complications, including ...increased incidence of cardiovascular disease, hyperlipidemia, anemia, and metabolic bone disease. CKD patients should be assessed for the presence of these complications and receive optimal treatment to reduce their morbidity and mortality. A multidisciplinary approach is required to accomplish this goal.
Extensive research and policies have been developed to improve access to kidney transplantation among patients with ESKD. Despite this, wide variation in transplant referral rates exists between ...dialysis facilities.
To evaluate the longitudinal pattern of access to kidney transplantation over the past two decades, we conducted a retrospective cohort study of adult patients with ESKD initiating ESKD or placed on a transplant waiting list from 1997 to 2016 in the United States Renal Data System. We used cumulative incidence models accounting for competing risks and multivariable Cox models to evaluate time to waiting list placement or transplantation (WLT) from ESKD onset.
Among the study population of 1,309,998 adult patients, cumulative 4-year WLT was 29.7%, which was unchanged over five eras. Preemptive WLT (prior to dialysis) increased by era (5.2% in 1997-2000 to 9.8% in 2013-2016), as did 4-year WLT incidence among patients aged 60-70 (13.4% in 1997-2000 to 19.8% in 2013-2016). Four-year WLT incidence diminished among patients aged 18-39 (55.8%-48.8%). Incidence of WLT was substantially lower among patients in lower-income communities, with no improvement over time. Likelihood of WLT after dialysis significantly declined over time (adjusted hazard ratio, 0.80; 95% confidence interval, 0.79 to 0.82) in 2013-2016 relative to 1997-2000.
Despite wide recognition, policy reforms, and extensive research, rates of WLT following ESKD onset did not seem to improve in more than two decades and were consistently reduced among vulnerable populations. Improving access to transplantation may require more substantial interventions.
Coding variants in the
gene are associated with kidney diseases in African ancestral populations; yet, the underlying biologic mechanisms remain uncertain. Variant-dependent autophagic and cytotoxic ...cell death have been proposed as pathogenic pathways mediating kidney injury. To examine this possibility, we conditionally expressed APOL1-G0 (reference), -G1, and -G2 (variants) using a tetracycline-regulated system in HEK293 cells. Autophagy was monitored biochemically and cell death was measured using multiple assays. We measured intracellular Na
and K
content with atomic absorption spectroscopy and APOL1-dependent currents with whole-cell patch clamping. Neither reference nor variant APOL1s induced autophagy. At high expression levels, APOL1-G0, -G1, and -G2 inserted into the plasma membrane and formed pH-sensitive cation channels, causing collapse of cellular Na
and K
gradients, phosphorylation of p38 mitogen-activated protein kinase, and cell death, without variant-dependent differences. APOL1-G0 and -G2 exhibited similar channel properties in whole-cell patch clamp experiments. At low expression levels, neither reference nor variant APOL1s localized on the plasma membrane, Na
and K
gradients were maintained, and cells remained viable. Our results indicate that APOL1-mediated pore formation is critical for the trypanolytic activity of APOL1 and drives APOL1-mediated cytotoxicity in overexpression systems. The absence of cytotoxicity at physiologic expression levels suggests variant-dependent intracellular K
loss and cytotoxicity does not drive kidney disease progression.
The mechanism of pathogenesis associated with APOL1 polymorphisms and risk for non-diabetic chronic kidney disease (CKD) is not fully understood. Prior studies have minimized a causal role for the ...circulating APOL1 protein, thus efforts to understand kidney pathogenesis have focused on APOL1 expressed in renal cells. Of the kidney cells reported to express APOL1, the proximal tubule expression patterns are inconsistent in published reports, and whether APOL1 is synthesized by the proximal tubule or possibly APOL1 protein in the blood is filtered and reabsorbed by the proximal tubule remains unclear. Using both protein and mRNA in situ methods, the kidney expression pattern of APOL1 was examined in normal human and APOL1 bacterial artificial chromosome transgenic mice with and without proteinuria. APOL1 protein and mRNA was detected in podocytes and endothelial cells, but not in tubular epithelia. In the setting of proteinuria, plasma APOL1 protein did not appear to be filtered or reabsorbed by the proximal tubule. A side-by-side examination of commercial antibodies used in prior studies suggest the original reports of APOL1 in proximal tubules likely reflects antibody non-specificity. As such, APOL1 expression in podocytes and endothelia should remain the focus for mechanistic studies in the APOL1-mediated kidney diseases.