MOLECULAR DESIGN OF THE KIDNEY FILTRATION BARRIER

Abstract

Kidneys are the central regulators of organismal homeostasis. These organs filter enormous amounts of fluid from plasma; excrete toxic waste products; maintain salt, water, and volume balance; coordinate blood pressure regulation; and maintain the acid–base equilibrium essential for life. Although it has been known for decades that renal glomeruli serve as the site of plasma ultrafiltration and urine production, both the molecular design and function of the kidney filtration barrier have remained elusive. Indeed, the past two decades have witnessed enormous breakthroughs in our fundamental understanding of kidney filtration and the critical role that podocytes, specialized terminally differentiated epithelial cells at the glomerular capillaries, fulfill in the function of the kidney filtration barrier. Here we discuss recent advances in this field that will change the way we think about plasma ultrafiltration in health and proteinuria as a manifestation of glomerular diseases.

INTRODUCTION

The human kidneys filter about 180 liters of fluid every day from plasma. In doing so, the intact kidney filtration barrier tightly restricts passage of proteins based on their biophysical properties, which results in a filtrate that is almost free of albumin or other plasma proteins (1). However, exactly how kidney glomeruli separate the macromolecular components of blood from small solutes has been debated for more than 150 years (2-6).

One million glomerular microvascular units are responsible for ultrafiltration. The primary filtrate then passes along the nephrons where solutes and water are reabsorbed and some additional waste products are secreted. The glomerular filter consists of three layers: a specialized, fenestrated glomerular endothelium, the glomerular basement membrane (GBM), and podocytes that interdigitate to closely enwrap the glomerular capillaries and cover the outer surface of the GBM (7-8). Podocytes are unique terminally differentiated epithelial cells that display primary and interdigitating secondary processes. To this end, they form a 40-nm-wide membrane-like cell junction, called the slit diaphragm, which connects adjacent processes (9). Formation of the slit diaphragm is required for the development of the mature filtration slits. When podocytes are damaged, the intercellular junctions and cytoskeletal structure of the foot processes are altered and the cells take on an “effaced” phenotype (10-11). The majority of chronic kidney diseases start in the glomerulus as a consequence of a limited capacity of glomeruli and podocytes for regeneration and self-renewal, respectively (1). Loss of podocytes, which cannot adequately be compensated, results in scarring of glomeruli, referred to as glomerulosclerosis, and leads to progressive loss of kidney function and, ultimately, end-stage renal disease.

IDENTIFICATION OF GENE DEFECTS OF GLOMERULAR DISEASES REVOLUTIONIZED OUR UNDERSTANDING OF THE FUNCTION OF THE GLOMERULAR FILTRATION BARRIER

About two decades ago, podocytes became the subject of intense scientific scrutiny in kidney research. Landmark genetic studies identified defects in genes expressed by podocytes as the genetic cause of proteinuria and progressive renal disease in humans (12-15). These seminal studies started with the identification of the genetic cause of congenital nephrotic syndrome of the Finnish type. This autosomal recessive disorder is caused by mutations in the NPHS1 gene that encodes for the adhesion protein, nephrin. It should be noted that children with nephrin mutations develop massive albuminuria in utero and a life-threatening renal disease resulting in massive loss of protein in urine right after birth. Interestingly, nephrin is a major constituent of the slit diaphragm, in which nephrin molecules bridge the distance between two adjacent foot processes to form the 40-nm membrane-like cell junction (12,16,17). More recently, cryoelectron microscopy studies have revealed a flexible and multilayered architecture of the slit diaphragm (18). Slit diaphragm formation has been proven to require not only nephrin but also a related Ig superfamily member called neph1—not the other neph family members called neph2 and neph3. Neph1 appears to form homomeric complexes independent of nephrin at the lower part of the junction in close proximity to the GBM, and the nephrin interaction occurs more apically.

Whereas neph1 contains five Ig-like domains (19-20), nephrin is a 1241 amino acid protein with a large extracellular domain containing eight Ig-like domains, one type III fibronectin domain, a single transmembrane region, and a short cytoplasmic tail (9,21). Nephrin is also known to be phosphorylated at several tyrosine residues, which enables it to recruit adaptor proteins subsequently, such as p85 and Nck, and induce signaling to the cytoskeleton or the nucleus (22-26). The identification of signaling functions of nephrin, first demonstrated in cell culture systems (23), triggered a new research impetus into the biology of podocytes (9,27,28), which made rapid advances due to the fact that nephrin and its associated proteins are conserved in evolution (Figure 1). Nephrin and neph1, also called Hibris/Sticks-and-Stones and Roughest in Drosophila melanogaster, or SYG-2 and SYG-1 in C. elegans, are required for a wide variety of developmental processes such as eye development and muscle fusion in the fly (29-31) or synapse targeting in the nematode (32-33). In the Drosophila pupal eye, for example, nephrin/neph1 signaling and adhesion are essential for sorting and shaping of interommatidial precursor cells that surround the lens-secreting cone cells and primary pigment cells. These morphogenetic processes are required for the correct end formation of a fly compound eye (34). Mutations of Hibris (nephrin in mammals) and Roughest (neph1 in mammals) result in the rough eye phenotype, and hence, the name Roughest protein. In C. elegans, synaptic connections can be determined by guidepost signals. For example, SYG-2 (nephrin) is expressed transiently on epithelial guidepost cells during synapse formation. SYG-2 binds SYG-1, the receptor on motoneurons, and directs SYG-1 accumulation and synapse localization. Thus, these model systems were successfully used to better understand nephrin signaling functions in vivo (35-37). Collectively, these studies showed that nephrin-based protein interactions are essential for mediating signal transduction and controlling cytoskeletal rearrangements in podocytes to shape the unique podocyte ultrastructure.

Fig. 1.

The slit diaphragm protein complex controls signaling in podocytes.

In addition to slit diaphragm protein-encoding genes, several cytoskeletal proteins can be mutated in hereditary podocyte disease (Figure 2). These proteins include alpha-actinin 4 (ACTN4) (15), inverted formin 2 (INF2) (38), the nonmuscle class myosin 1e (MYO1E) (39), the RhoGTPase-regulatory protein KANK1 (40), as well as its interactor ARHGDIA (41). Inverted formin 2 (INF2) is a member of the diaphanous subfamily of formin proteins and was shown to regulate actin polymerization-induced RhoA subfamily G proteins (42). Thus, the identification of cytoskeletal-associated genes in genetic forms of proteinuria has highlighted the critical importance of regulating podocyte actin cytoskeletal to establish and maintain an intact glomerular filtration barrier.

Fig. 2.

Identification of gene defects in hereditary nephrotic syndrome reveals the central role of podocytes in albuminuria development.

THE SLIT DIAPHRAGM CONTAINS A MULTIMERIC MECHANOSENSORY PROTEIN-LIPID SUPERCOMPLEX

Mutations in the NPHS2 gene, which encodes the prohibitin (PHB-) homology domain protein podocin, are the most common cause of hereditary nephrotic syndrome and steroid-resistant nephrotic syndrome in children (13,43). Podocin interacts with nephrin at the slit diaphragm (44) and enhances nephrin signaling (23). In C. elegans, the closest ortholog of podocin is the PHB-domain protein MEC-2 (45-46). MEC-2 is part of a multiprotein channel complex expressed in six touch receptor neurons and required to transduce gentle touch in the nematode (47-48). In touch receptor neurons, this channel complex localizes to touch channel puncta along the neuronal process, and MEC-2 has been shown to regulate the MEC-4/MEC-10 ion channels in these puncta. The touch transduction channel is a complex of at least six proteins that multimerize to form megadalton membrane protein complexes that are composed of two DEG/ENaC channel subunits (MEC-4 and MEC-10), two PHB-domain proteins (MEC-2 and UNC-24), and two paraoxonase-like proteins, namely MEC-6 and K11E4.3.

Podocin and MEC-2 share a similar structure with a central membrane-close hydrophobic region, amino and carboxy terminal tails facing the cytoplasm, and PHB-domains that are 50% identical and more than 80% similar (49). These proteins belong to a large family of evolutionarily conserved membrane-associated proteins. The PHB-domain proteins constitute a family of over 1,300 proteins, all of which share an approximately 150 amino acid domain similar to that in the mitochondrial protein prohibitin [PHB-domain (50) or SPFH domain]. Over 360 of these proteins have been identified in animals, many of which have an N-terminal hydrophobic region that results in their location at the inner leaflet of the lipid bilayer. Based on biochemical fractionation in detergent-resistant membrane preparations, a number of the mammalian PHB-domain proteins have been shown to cofractionate with lipid rafts of the plasma and intracellular membranes. Further, podocin and MEC-2 are known to form higher-order multimeric complexes and to bind cholesterol to establish homologous mechanosensory ion channel supercomplexes. Recruitment of cholesterol into the multimolecular complexes is essential for mechanical activation of MEC-4 channels in the worm and TRPC6-based ion channel activity in the podocyte (49). Consistent with a role in the same complex, mutations in the TRPC6 gene have been identified as a cause of focal and segmental glomerulosclerosis in humans (51). However, although molecularly very well characterized, the physiologic role of mechanosensation at the glomerular filtration barrier is still incompletely understood.

THE ADVENT OF OMICS TECHNOLOGIES HAS CHANGED THE WAY SIGNALING CAN BE STUDIED

Given the sophisticated ultrastructure and function of podocytes, it is no surprise that available cell culture models, although valuable tools that truly boosted podocyte research (52), do not express the full spectrum of podocyte cytoskeletal and slit diaphragm genes. Thus, these models are only of limited value in studying complex podocyte signaling networks, slit diaphragm signaling, as well as sophisticated cytoskeletal responses. Therefore, to gain a more thorough understanding of these signaling networks and responses, in vivo assessment is a better research option. Because signaling is highly complex and involves regulatory responses at various levels, this process cannot be adequately addressed by classical one-gene one-protein approaches alone. As a result, with the advent of mass spectrometry (MS)-based proteomics together with deep sequencing of the cellular RNA, analysis of transcriptomes, proteomes, and posttranslational modifications, such as phosphorylation, ubiquitylation, or protease-mediated cleavage, has been enabled on a global scale. In addition, MS offers numerous advantages to studying protein phosphorylation, enabling its quantitative, sensitive, and site-specific measurement on a large scale (53-54). Furthermore, phosphoproteomics approaches in podocytes obtained from mouse models have identified more than 5000 high-confidence phosphorylation sites and provided a deeper understanding of regulatory posttranslational modifications in vivo (55-56). Cross species comparisons also revealed evolutionarily conserved phosphorylation sites (55). Moreover, deep mapping of the podocyte transcriptome by RNA sequencing and proteome by MS-based quantitative proteomics has provided an atlas of protein expression in glomeruli and podocytes with more than 10000 expressed proteins (57-58). Collectively, these omics studies also demonstrate how the integration of multi-omics datasets can identify a framework of cell-type-specific features relevant to organ health and disease. These compiled results can now be combined with datasets on ubiquitylation and degradation as well as detailed information about protein stability and half-lives, thereby ushering in a new era of podocyte research (59-62).

Spectacular advances in imaging technologies allow for visualization of podocyte dynamics in vivo

Despite the fact that diseases affecting the glomerular filtration barrier are a leading cause of end-stage renal failure and despite the enormous progress in understanding podocyte pathobiology, comprehensive models that explain the function of the glomerular filtration barrier are still highly controversial. Several decades ago, electron microscopy studies showed that dextran particles of the size of albumin or bigger did not enter the GBM, leading to the conclusion that the basement membrane may act as a primary filtration barrier (63-64). In contrast, injected human myeloperoxidase was observed to pass through the GBM, but was impeded at the level of the slit diaphragm, thereby leading the authors to conclude that podocyte slit diaphragms are the primary filter (65). This controversy remains a matter of intense scientific debate. For example, current interpretations in terms of a coarse filter at the GBM followed by a fine filter at the slit have difficulty explaining why the glomerulus does not clog (5), and several alternative hypotheses were proposed to explain the function of the size and charge of selective glomerular filtration barriers (6,66). However, convincing scientific evidence was not obtained due to unsurmountable technical limitations and experimental challenges. In attempting to refine detail, all previous studies were hampered by the fact that imaging at ultrastructural resolution was only possible through the use of electron microscopy in fixed (and dead) tissue.

Thus, the introduction of new imaging technologies holds great promise to address this current unsatisfactory situation. For example, mosaic expression of genetically encoded colors allowed the visualization of individual foot processes at ultrastructural resolution. ROSA^mt/mg is a cell membrane-targeted, two-color fluorescent Cre-reporter allele (67). Prior to Cre recombination, cell membrane-localized tdTomato (mT) fluorescence expression is widespread in all cells. Mosaic expression of Cre recombinase in single cells results in cell membrane-localized EGFP (mG) fluorescence expression replacing the red fluorescence. Importantly, mosaic expression of green and red color in neighboring podocytes in living mice allowed—for the first time—visualization of podocyte ultrastructure by conventional fluorescence microscopy without prior fixation (68). Moreover, the application of intravital multiphoton fluorescence microscopy now enables a dynamic visualization of glomerular structure and function over time in the intact, living kidney (69-70). Further, with the development of serial multiphoton microscopy of the same glomeruli over several days, the motility of podocytes and parietal epithelial cells could be visualized even over long periods of time in vivo (71). These technologies will ultimately lead to refinements of ultrafiltration as well as new and profound insights into the molecular design of the kidney filtration barrier since the process of separating macromolecular components from the filtrate can now be visualized in living animals. Finally, morphometric data obtained through superresolution stimulated emission depletion imaging (Figure 3) prior to and after the onset of glomerular disease may allow the study of precise processes that lead to failure of the barrier, thereby potentially providing new possibilities for therapeutic interventions and treatment options.

Fig. 3.

Superresolution microscopy allows for visualization of the kidney filtration barrier.

CONCLUSIONS

Recent spectacular advances have changed the way we think about the molecular design and function of the kidney filtration barrier. These studies have stressed the importance of the podocyte and the slit diaphragm cell junction in maintaining an intact glomerular filtration barrier. Many podocyte-expressed genes have now been identified that cause genetic forms of albuminuria when mutated. However, the exact molecular design and function of the glomerular filtration barrier still remains to be elucidated. Moreover, further studies are urgently needed that will define novel therapeutic interventions and identify new treatment options for inherited as well as acquired proteinuric kidney diseases and that can develop suitable tools to predict patient responses to individualized treatment options. Precision medicine is now entering nephrology.

DISCUSSION

Ziedel, Boston: Wonderful talk and wonderful work. This story has been evolving over some years as we go from the mechanical understanding of the glomerulus—Barry Brenner advances—to the molecular understanding, which is emerging from genetics. I have two questions: (1) Can we do something for these kids? and (2) At this stage, since it sounds like every kid who comes in with proteinuria should be genotyped before we empirically treat them with corticosteroids because someone with this problem is not going to respond to corticosteroids, why should we expose them to it?

Benzing, Cologne: Thank you very much. Regarding your second question, it is very clear that more than 30% of the kids with steroid-resistant nephrotic syndrome have a genetic form of it and should be tested. It's not so obvious in adults, and it will take some time to better understand the development of focal segmented glomerulosclerosis in adults. But, clearly, children should be tested and 30% have a mutation in the podicin gene and there are many, many genes known. Now to answer your first question: Can we do something about this in children with genetic mutations? Transplantation obviously cures disease because you are transplanting a healthy non-mutated kidney. Other than that, we do not have anything. We understand how proteinuria develops, and, of course, we can now think about developing new therapeutic interventions which is currently ongoing. It's ongoing in my department—we have a couple clinical trials running, based on these findings in the laboratory.

Henrich, San Antonio: Beautiful talk. Seventy percent of renal failure in nephrotic syndrome is caused by diabetes in America. Can you make a leap from what you found in this congenital nephrotic syndrome model to something that could potentially be therapeutic in ameliorating or mitigating the proteinuria, the damaged podocytes, and the barrier in diabetes?

Benzing, Cologne: We always thought that proteinuria is the result of holes where protein goes through and we know in diabetes it's identical. But it's really a shortening of the slit diaphragm that causes albuminuria. Later podocytes are lost, and everything goes down. But the initial phase is really shortening of the slit diaphragm; we know, for example, that ACE inhibitors reduce the pressure and by reducing the pressure they can obviously reduce albuminuria. We don't have anything in hand to intervene here other than ACE inhibitors and SGLT2 blockers. I'm pretty convinced that we will be able to address the podocyte and also treat it as the prime determinant of the progression of the disorder, but that will take some time.