The Beginning of the End of the Road for a Renal Glomerulus
The images above are sections of kidney taken from age-matched non-diabetic control animals and type II diabetic animals (18 weeks duration). The central figure in both images is a renal glomerulus. The sections were immunostained with antibodies directed against syndecan-4 (green, monoclonal antibodieectodomain of the syndecan-4 core protein) and alpha-actinin-4 (red). In the control animals, both syndecan-4 and alpha-actinin-4 co-distribute along the walls of the glomerular capillaries, the pattern of staining having the appearance of punctae along the length of the perimeter of the glomerular capillary wall. . In the diabetic animal, the pattern of syndecan-4 staining is disrupted, as indicated by the loss of the green signal. We believe this is due to the disengagement of the interactions between Syndecan-4 with its ligands in the glomerular capillary wall. The narrative below provides further background information.
One of the most sobering journeys I have ever taken a basic scientist was a walk through the LSU Health Shreveport Hemodialysis Clinic and seeing the patients huddled up in their beds undergoing hemodialysis. Although a lifesaving intervention for those who have lost the majority of their renal function due to Chronic Kidney Disease (CKD), a patient is tied to the hemodialysis machine for several hours at a time, for several days a week, sometimes for years on end. As I walked through the clinic, I reflected on several family members, friends, and even colleagues for whom hemodialysis offered them hope for a bit more time on this planet.
The Epidemic of Chronic Kidney Disease
Chronic Kidney Disease (CKD) can be the result of several underlying disease processes, such as poorly controlled hypertension or diabetes mellitus. In some individuals, acceleration of the progression towards the final stages of CKD can occur when two or more co-morbidities exist in an individual (e.g. a hypertensive diabetic). According to the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), 14% of the general population ( approx 46 million people) has some form/stage of CKD. In 2015, NIDDK reported that 661,000 people had entered into renal failure, of these 468,000 were on hemodialysis. The current cost for treating people over the age of 65 for CKD was greater than $50 billion dollars in 2013 according to a current NIDDK report.
A key contributor to the development of CKD is underlying metabolic disease, such as diabetes mellitus. According to data from the American Diabetes Association, in 2015 9.4% of the US population (approx 30 million people) had diabetes mellitus, with 1.5 million new cases being diagnosed every year. This was not always the case and one has to wonder where things started taking a turn for the worse. From a look back in time, according to the Center for Disease Control data, in 1960 there were only 1.59 million people (0.91% of the population) that were diagnosed as having diabetes mellitus. The fact that there has been a proportionate 10-fold increase of diabetes in our population since 1960 is nothing short of disconcerting. If the trend continues, the economic burden in treating the complications arising from long term diabetes mellitus will be staggering.
Early Studies in Diabetic Nephropathy in the McCarthy Lab
The research direction of the laboratory towards studying the effects of CKD on the renal glomerulus was initially set by a serendipitous (i.e. lucky) observation. We had found that an antibody, directed against a basement membrane chondroitin sulfate proteoglycan, specifically stained the extracellular matrix present in glomerular mesangium and not the matrix comprising the glomerular basement membrane (J Cell Biol. 109: 3187-3198). A subsequent study showed that the pattern of staining of this antibody was highly regulated during the development of the kidney (J. Histochem. Cytochem 41: 401-414). Interestingly and germane to kidney disease, we found that the pattern of immunostaining of this antibody in the kidney was a very good marker that demonstrated the progression of CKD in a model of diabetic kidney disease. The pattern of antibody staining readily identified the initiation of the process of mesangial expansion in the glomeruli of diabetic animals (J Histochem Cytochem 42: 473-484) which is considered to be an indicator of the eventual demise of the glomerular structure and ultimately kidney function (Diabetes 38:1077-81). The logical follow-up study was to determine if early intervention in an animal model of type II diabetes via the correction of blood glucose levels with the thiazolidinedione, troglitazone, could prevent mesangial expansion from occurring (Kidney International 58: 2341-2350). The results of that study were extremely positive and showed that earlier intervention with thiazolidinedione treatment prevented the development of mesangial expansion from occurring.
A Slight Detour Leads to a Novel Observation
Within the field of renal biology, it had been understood from studies done in the last millenium that the extracellular matrix which forms the glomerular capilliary wall contributes, in part, to the ultrafiltration barrier function of the glomerular capillary wall (Microscopy and Microanalysis 18:3-21). Key components of the glomerular capillary extracelllar matrix, heparan sulfate proteoglycans (agrin and perlecan), were thought to limit the diffusion of proteins from the bloodstream into the urinary space, due to the net anionic charge density presented by the heparan sulfate glycosaminoglycan chains attached to these proteins. This hypothesis has been explored and tested by several groups over the past forty years; the results of those studies more often than not contributed to further debate about the actual process of ultrafiltration.
Although the focus of our work for years had been on studying how the proteoglycans made by mesangial cells were affected by underlying diabetic nephropathy, an opportunity presented itself where we could develop an animal model that would allow us to explore the role of heparan sulfate glycosaminoglycan in glomerular ultrafiltration. Through the development of this novel mouse model, we thought we could address the nagging questions raised by numerous investigators with regard to the role of heparan sulfate proteoglycans/ glycosaminoglycans in glomerular ultrafiltration. The mouse model was made by breeding a podocyte-specific Cre expressing mouse (Genesis 35: 39-42) with a recently-developed Ext1fl/fl mouse (Science 302, 1044–1046) we developed a mouse model in which glomerular podocytes are unable to assemble heparan sulfate glycosaminoglycans on all heparan sulfate proteoglycan core proteins (Kidney International 74: 289-299).
Serendipity (i.e. luck) again decided to visit the laboratory. Our hypothesis, at the time that we began our studies, was based on the conventional wisdom that existed in the field of glomerular biology. We had hypothesized that the resulting mutant mouse from our animal breeding scheme should develop severe proteinuria within a few weeks after birth. As mentioned in the paragraph above, our hypothesis was based on an existing concept in the renal field that heparan sulfate glycosaminoglycans in the glomerular basement membrane contributed to the charge selectivity of the glomerular ultrafiltration barrier. Thus, by removing the majority of heparan sulfate glycosaminglycans in the glomerular basement membrane via the genetic manipulation in our mutant mouse model, we postulated that the mouse should develop severe proteinuria as the kidneys began handling the ultrafiltration load for the animal after birth. They didn't. In fact the mutant mouse model (the PEXTKO mouse (Kidney International 74: 289-299) ended up living a normal lifespan and does not develop any significant degree of proteinuria. As we began to characterize the phenotype of PEXTKO mouse we discovered that despite the lack of proteinuria, the glomerular podocytes developed what is known as pedicel or "foot process" effacement-i.e. a complete disruption of the normal interactions that occur between the glomerular podocyte and the glomerular basement membrane. Foot process or pedical effacement is seen in many renal diseases, such as diabetic nephropathy. Based on our earlier work, we used morphometry to investigate whether or not the glomeruli in the PEXTKO mouse developed mesangial expansion during the course of their lifespan. Although we found that the glomeruli in the PEXTKO mouse developed hypertrophy, the mesangium itself remained unchanged relative to the overall size of the glomerulus.
The pedicel effacement did suggest the possibility that the podocytes in the mutant animals did suffer from an adhesion defect, mediated by the lack of heparan sulfate glycosaminoglycan chains on the core proteins of cell surface proteoglycans. To test this hypothesis, we developed immortalized podocyte cell lines that were able to synthesize heparan sulfate glycosaminoglycan chains or unable to synthesize heparan sulfate glycosaminoglycan chains ( Kidney International 78:1088-1099). We tested the cells via standard cell adhesion and migration assays and demonstrated that the HS- podocytes were inefficient with regard to attachment to a matrix substrate (fibronectin) and migrated poorly on the substrate compared to wild-type (HS+) podocytes. Immunostaining for the cell surface proteoglycan, Syndecan-4, and cytoskeletal components showed that HS- podocytes had profound differences in their ability to organize focal adhesions and their cytoskeletal organization compared to HS+ podocytes. Immunostaining for Syndecan-4 in renal tissue sections from the PEXTKO mouse showed that the pattern of immunostaining for Syndecan-4 was disrupted in the renal glomerulus compared to control animals, the disruption similar to that seen in the glomerulus shown above, taken from the diabetic animal.
Based on the data derived from the PEXTKO mouse and our in vitro studies using the HS+ and HS- podocytes, we proposed a basic model of how changes to the structure (sulfation, epimerization) of HS (heparan sulfate glycosaminoglycan) that is assembled posttranslationally on Syndecan core proteins present on the basal surfaces of podocyte pedicels affect pedicel organization along the length of the GBM.
In normal conditions, Syndecans (shown in blue/red) are capable of engaging matrix glycoproteins present in the glomerular basement membrane via the pattern of sulfation present on the HS chains (shown in red). The interaction between HS chains and matrix glycoproteins is somewhat promiscuous, since HS is capable of having many different binding partners. Syndecans work alongside integrins (shown in green/yelllow), which are cell surface matrix receptors that have a rather high degree of specificity. The syndecan-integrin pairing promotes the development of a normal adhesive phenotype and normal cytoskeletal organization for the podocyte pedicel. When either the ability to assemble HS chains is lost, as in our PEXTKO model or in our more recent Ndst1-null mouse model (Kidney International 85: 307-3018), these changes to the post-assembly modifications of HS render syndecans incapable of efficient interactions with matrix glycoproteins. In vitro, using our latest model, the Ndst1 null mouse, we were able to show that podocytes are unable to efficiently adhere and migrate on extracullar matrices, the integrin activation status is downregulated (American Journal of Physiology, Renal Physiology 310: F1123-F1135).
A third alternative, that we are currently testing, is that conditions present in diabetes lead to the enhanced shedding of Syndecans from the basal surface of the podoctye pedicel. We believe that the loss of the Syndecan-basement membrane interaction, by whatever means, is associated with foot process effacement in vivo and disruption of the pattern of staining for podocyte syndecan in the glomerulus.
How are these mutant mouse studies relevant to the human condition?
The PEXTKO mouse model (podocytes lack the ability to assemble HS) gave some insight, albeit an extreme model, as to how important the HS on Syndecans is with regard to the maintenance of normal podocyte architecture. The Ndst1-null mouse model (Kidney International 85: 307-3018, American Journal of Physiology, Renal Physiology 310: F1123-F1135) represents a refined model that more closely reflects what is seen in the glomeruli of diabetic animal models. Ndst1 is an enzyme responsible for the N-sulfation of N-acetylglucosamine residues on HS. In turn, when this activity is removed from cells via genetic manipulation, the HS made by the cells is known to be undersulfated. Expression data from genomic screening studies in humans have shown that in humans the overall expression of Ndst1 is significantly decreased in the glomeruli isolated from kidney samples taken from diabetic individuals. We believe that conditions present in diabetes mellitus forces the decrease in expression of Ndst1 which, in turn, would ultimately compromise the ability of podocytes to properly N-sulfate the heparan sulfate present on Syndecans. This would contribute in part, to the foot process effacement that does develop in the glomeruli of diabetic animals.
A key question remains to be answered............
The results of our work with the PEXTKO mutant mice would suggest that the role of HS in ultrafiltration is minimal, at best. However, in our original report (Kidney International 74: 289-299) we noted that at 6 months of age the presence of large lysosomes developing in the proximal tubule epithelial cells, these lysosomes persisting throughout the life of the PEXTKO mice. This would suggest that the loss of HS in the glomerular basement membrane (and podocyte pedicels) had some effect on ultrafiltration, but the net effect could have been masked by an enhanced protein recovery in the proximal tubule epithelial cells (Kidney International 71: 504-513; JASN 27: 482-494).
In order to answer that question, we tried to use an Intravital Two-Photon Microscopy approach to determine to what degree ultrafiltration had been affected in our PEXTKO mutant mouse models (Kidney International 71: 504-513; JASN 27: 482-494). After taking a course on intravital Two-Photon Microscopy at the IUPUI Research Center for Quantitative Renal Imaging we began our studies exploring the possible differences in renal ultrafiltration between wild-type and PEXTKO mice.
The figure above shows two panels of murine glomeruli labeled intravitally with rhodamine-albumin (left) or with a mixture of Alexa488-albumin/150kD rhodamine-dextran. The nuclei in the image were also labeled intravitally using Hoescht 33342 nuclear stain. The arrows in the left hand panel point to two peripheral glomerular capillaries.
The figure to the left is a micrograph taken during the approach of the objective lens towards the outer cortex of the kidney. The kidney was labeled intravitally using a mixture of Alexa488-albumin/150kD rhodamine-dextran and the organ nuclei labeled with Hoescht 333242 nuclear stain. The intravital stain labels all capillaries yellow, including the capilllaries of the vasa recta (the majority of the capillaries in the field). On the right hand side of the field is a larger plexus of capillaries arising from a capsullar artery (a branch of the arcuate artery). The black striations within some of the capillaries are caused by the movement of red blood cells within the capillary proper.
The images show that intravital labeling can be done for the murine kidney. However, our work in this area has been hampered by the fact that the majority of the glomeruli in an adult mouse lie approximately 300µm from the outer capsule of the murine kidney, placing those glomeruli just out of the imaging range of the Zeiss LSM 510 MPM in our Resesarch Core Facility. Because of the relative distance of the glomeruli in the adult murine model combined with the refractive index of renal tissues, image acquistion of actively filtering is difficult and the number of glomeruli capable of being seen are few and far between. In the rat model that is currently in use for most renal-based two photon microscopy studies (Kidney International 71: 504-513; JASN 27: 482-494) many of the glomeruli are within 50µm of the outer capsule of the kidney, making image acquisition easier and the ability to poll large numbers of glomeruli per imaging session feasible.
We are currently trying to solve this problem by refining our surgical approaches in younger mice whose cortex would be relatively thinner than adult mice. We are also working actively to upgrade the Zeiss LSM 510 MPM to a state-of-the-art microscope which has a laser tunable out to 1300nm for deeper penetration into the renal stroma, newer generation detectors to facilitate image capture, and higher scan speeds for video rate imaging.