PATHWAYS TO PROTEINURIA Ramzi Khalil
PATHWAYS TO PROTEINURIA Ramzi Khalil
Pathways to proteinuria ISBN: 978-94-6506-090-3 Cover design: Ramzi Khalil & Ridderprint | www.ridderprint.nl Layout: Ridderprint | www.ridderprint.nl Printing: Ridderprint | www.ridderprint.nl Copyright © 2024, Ramzi Khalil All rights reserved. No part of this thesis may be reproduced or transmitted in any form, by any means without prior written permission of the author.
PATHWAYS TO PROTEINURIA Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Leiden op gezag van rector magnificus prof. dr. ir. H. Bijl, volgens besluit van het college voor promoties te verdedigen op dinsdag 4 juni 2024 klokke 13.45 door Ramzi Khalil geboren te Leiden in 1991
Promotores Prof. dr. P.C.W. Hogendoorn Prof. dr. A.J. Rabelink Co-promotor Dr. J.J. Baelde Promotiecommisie Prof. dr. Y.K.O. Teng Dr. J. van den Born University Medical Center Groningen Prof. Dr. J. van der Vlag Radboud University Medical Center Prof. Dr. S. Florquin Amsterdam Medical Center
Table of Contents Chapter 1. General introduction and outline 7 Chapter 2. Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 17 Chapter 3. Glomerular permeability is not affected by heparan sulfate glycosaminoglycan deficiency in zebrafish embryos 47 Chapter 4. Mutations in the heparan sulfate backbone elongating enzymes EXT1 and EXT2 have no major effect on endothelial glycocalyx and the glomerular filtration barrier 65 Chapter 5. Increased dynamin expression precedes proteinuria in glomerular disease 83 Chapter 6. Transmembrane Protein 14A protects glomerular filtration barrier integrity 103 Chapter 7. General discussion and future perspectives 121 Appendices Nederlandse samenvatting 138 Curriculum vitae 144 List of publications 145 Dankwoord 146
CHAPTER 1 General introduction and outline
Chapter 1 8 Proteinuria and chronic kidney disease Proteinuria is an independent predictor of the progression of kidney injury, cardiovascular morbidity, and overall mortality (1). Under physiological circumstances, no sustained proteinuria is present. Proteinuria only occurs when serum proteins are able to pass the glomerular filtration barrier and when tubular reabsorption mechanisms are saturated or fail. Damage to any of these components can result in proteinuria. Moreover, although the mechanisms of damage and disease vary across the different types of renal diseases, many still lead to proteinuria and can result in chronic kidney disease (CKD). CKD is classified according to the remaining glomerular filtration rate and the amount of proteinuria in the Kidney Disease Improving Global Outcomes (KDIGO) guidelines (2). The prevalence of CKD is expected to rise further due to an aging population. The 2019 ‘Global burden of disease study’ reports that chronic kidney disease rose from rank 29 to 18 between 1990 and 2019 as a cause for disease-adjusted life years (3). In the Netherlands, an estimated 12% of the population suffers from chronic kidney disease (CBS/nierstichting). Treatment of underlying disease and general cardiovascular risk management such as treating hypertension and dyslipidaemia are still the cornerstone of management of chronic kidney disease (CKD) and attenuating proteinuria. A specific treatment for proteinuria might become feasible when the pathways leading to proteinuria are elucidated further. For this, a closer investigation of the structures and mechanisms that normally prevent proteinuria from occurring is required. The glomerular filtration barrier Figure 1 shows an overview of human renal anatomy with each image going into more detail. Most humans have two kidneys as seen in the top left image. They are located in the retroperitoneal space. The parenchyma of the kidney is usually divided in the outer cortex and inner medulla. Each human kidney contains around one million functional units called nephrons. Each nephron consists of a glomerulus and a tubular apparatus, shown in the detail of the sagittal image of the human kidney. The glomerulus, seen in the top right, is a specialized capillary bed that starts with the afferent arteriole which comes from the renal artery, which in turn directly sprouts from the abdominal aorta. 20 to 25% of cardiac output is routed to the kidneys, where it first passes the glomerulus. There, around 20% of passing serum is filtered by passing the glomerular filtration barrier, and the other 80% continues through the efferent arteriole to the peritubular capillaries where both active and passive secretion and reabsorption take place. The entire glomerular capillary bed is lined by the glomerular filtration barrier, shown in the middle image. Below, the detail shows a schematic overview of the distinct structures
General introduction and outline 9 1 of the glomerular filtration barrier. The glomerular filtration barrier is made up of the glomerular endothelial glycocalyx layer (purple), fenestrated endothelial cells (red), glomerular basement membrane (grey), and the visceral epithelial cells (yellow) – or podocytes – with interdigitating foot processes (green). Figure 1. The human kidney, nephron, glomerulus, glomerular capillary, and glomerular filtration barrier
Chapter 1 10 Glycocalyx and fenestrated endothelial cell The first layer of the glomerular filtration barrier is the glycocalyx. As the name implies, it consists mostly of various ‘sugary’ chains, which consist of negatively charged heparan sulphate glycosaminoglycan chains attached to heparan sulphate core proteins and hyaluronan. The glycocalyx has a role in mitigating inflammation, and coagulation.(4) The glycocalyx lines the specialized endothelial cells of the glomerulus. These endothelial cells have the distinguishing feature of being fenestrated with 60 nm pores. The negatively charged glycocalyx covering these pores is considered to be the first barrier between the vascular lumen and the ultrafiltrate that serum proteins such as albumin encounter. In end-stage renal disease, damage to the structural integrity and composition of the glycocalyx is observed.(5) Also, in experimental models where damage to the glycocalyx is induced, proteinuria occurs. As such, it is thought to be a vital part of the glomerular filtration barrier in the protection against proteinuria. Glomerular basement membrane The next layer is the glomerular basement membrane (GBM). It is an extracellular matrix that is proximally deposited by the glomerular endothelial cells and the visceral epithelial cell on the distal end. On electron microscopy, three GBM layers can be distinguished. These are the lamina rara interna on the vascular side, the lamina densa, and the lamina rara externa adjacent to the epithelial side. The GBM mainly contains laminin, collagen type IV, and heparan sulphate proteoglycans. It normally has a thickness between 300 and 400 nm. The GBM has an overall negative charge due to the sulphated glycosaminoglycan chains of the heparan sulphate proteoglycan aggregates. Heparan sulphate glycosaminoglycans Glomerular filtration occurs with both size and charge selectivity (6). Maintaining this selectivity and GFB integrity has long been attributed to heparan sulphate glycosaminoglycans (7, 8). As stated above, they are localized in both the glycocalyx and GBM. Heparan sulphate glycosaminoglycan chains consist of repeating disaccharide motifs that contain a uronic acid and a glucosamine derivative. Theoretically, up to 48 different motifs can be formed. Heparan sulphate synthesis starts with chain initiation, where a tetrasaccharide linkage region is formed and is covalently bound to a core protein. The next phase of HS synthesis consists of chain polymerization, which is dependent on enzymes encoded by the EXT1 and EXT2 genes. These enzymes form a co-polymerase that adds repeating disaccharide units consisting of D-glucuronic acid (GlcAβ4) and N-acetylglucosamine (GlcNAcα4). During chain polymerization, the polymer is also modified
General introduction and outline 11 1 by various sulfotransferases and an epimerase leading to GlcNAc N-deacetylation/Nsulfation, epimerization of GlcA to L-iduronic acid (IdoA), and 2, 3, and 6-O-sulfation. These modifications not only result in an overall negative charge, but also the formation of ligand binding sites, such as FGF-2, antithrombin, chemokines, and cytokines. Thus, when chain polymerization is impaired due to a lack of EXT1 or EXT2, no modification can take place and hence the heparan sulphate is functionally impaired.(9) Podocyte and slit diaphragm Lastly, the distal end of the GFB is covered by visceral epithelial cells or podocytes. Podocytes gained their name from the Greek words podos (ποδος), which means foot, and kutos (κύτος), meaning jar or vessel, and used as a term to describe cells. One of the characteristics of podocytes is the presence of interdigitating foot processes called pedicles. The remaining space between these processes creates so-called filtration pores or slit diaphragms. The diaphragm is not an open connection to Bowman’s space. Adjacent foot processes are connected by nephrin (NPHS1) and NEPH1 (KIRREL1).(10) These proteins connect adjacent foot processes by spanning the slit diaphragm and attaching to the actin cytoskeleton of podocytes. The actin cytoskeleton and dynamin When podocytes are damaged, loss of architectural organization of the cytoskeleton occurs, which leads to retraction and effacement of the podocyte and its foot processes. The pathways and proteins responsible for podocyte cytoskeletal organization are steadily being uncovered. One of these proteins is dynamin, a small GTPase that is primarily known for its role clathrin-coated vesicle budding in neurons. It has now been described to be involved in the turnover of nephrin, regulation of actin, and endocytosis of albumin by podocytes, making it a potential future therapeutical target for preventing proteinuria. (11-14) Tubular reabsorption mechanisms After passing the glomerular filtration barrier, the ultrafiltrate enters Bowman’s space. Afterwards, the ultrafiltrate passes various parts of the tubule, where its content is altered through re-absorption and secretion which occur in both active and passive manners. Normally, when serum proteins such as albumin are indeed able to pass the glomerular filtration barrier, they are re-absorbed by proximal tubular epithelial cells. Tubular reabsorption mechanisms are not only responsible for the re-absorption of filtered protein, but also that of for example urea, bicarbonate, phosphate, glucose, and
Chapter 1 12 sodium. The first part of ultrafiltrate is isotonic to serum, and as such, active processes are required to reabsorb these solutes. Hence, the proximal tubule epithelial cells possess many mitochondria to facilitate these active processes. On the tubular luminal side, these cells are lined with microvilli which form the brush border, which significantly increases the luminal surface area. It is here that solutes and colloids are reabsorbed. Some groups have provided evidence that albumin does in fact pass the glomerular filtration barrier in much greater quantities than previously thought. They claim that tubular reabsorption mechanisms thus play a larger role in preventing proteinuria than has been attributed to them before.(15) Although the relative contribution to preventing proteinuria is controversial, there is a general consensus that adequate tubular reabsorption mechanisms are required to prevent, at least in part, the loss of protein in the urine. A complete loss of tubular reabsorption function results in Fanconi syndrome. This syndrome is characterized by proteinuria and severe acid-base and electrolyte disorders. In children, the most common cause of Fanconi syndrome is nephropathic cystinosis. Nephropathic cystinosis is a lysosomal storage disorder that results in cystine crystal deposition in various tissues. The proximal tubule is usually the first affected site. As stated above, pathological processes in any one of the GFB layers can lead to proteinuria. Although this has long been attributed to the individually affected layer, it is now deemed more likely that the glomerular filtration barrier functions as a whole unit and requires all components to properly function. Moreover, the GFB can be seen as a dynamic barrier with various repair and compensation mechanisms rather than a static barrier that only sieves particles based on size, weight, and hydrostatic pressure. Identifying new targets Identifying which components of the glomerular filtration barrier are needed to maintain proper barrier function is essential to identifying potential therapeutic options. Historically, the function of the various components of the glomerular filtration barrier has been discovered in patients with genetic defects leading to proteinuria.(16) More recently, experimental animal models have been used to identify genes and their encoded products that might be important in GFB function. For example, genetic association studies in proteinuric rats have identified a panel of various genes potentially involved in the development of proteinuria.(17) After identifying these genes in a previously described array, the candidate genes and their encoded proteins were then investigated further to establish whether a direct relation to the development of proteinuria is present. In multiple studies presented here, the method to examine whether a protein has a significant role in the development of proteinuria consisted of assessing whether knocking down mRNA
General introduction and outline 13 1 translation resulted in proteinuria in a zebrafish embryo model. Next, differential mRNA and protein expression are investigated in spontaneously proteinuric rats. In this step, the temporal relationship between the onset of proteinuria and relative loss of expression is assessed. Furthermore, the translation to human proteinuric kidney disease is made by examining the protein expression of the investigated targets in kidney biopsies from patients with proteinuric renal diseases. These methods were used both to investigate known proteins, such as dynamin, but has also uncovered previously unknown proteins to be involved in the development of proteinuria, such as transmembrane protein 14A. Aim and outline of the thesis In this thesis, the pathways leading to proteinuria are explored. To identify potential pathways, elements considered essential are revisited, known pathways are explored further, and new players in the field of proteinuria are identified. First, a zebrafish embryo model to assess both glomerular filtration barrier function and tubular reabsorption mechanisms is presented in Chapter 2. The use of this model for developing new therapeutic options for the rare but devastating disease of nephropathic cystinosis is presented. In Chapters 3 and 4, the loss of heparan sulphate glycosaminoglycans is investigated. Heparan sulphate glycosaminoglycans have long been considered essential for adequate glomerular filtration function. In Chapter 3, a global heparan sulphate glycosaminoglycan deficiency on the development of proteinuria was shown to not affect glomerular filtration barrier function nor tubular reabsorption mechanisms. The used dackel zebrafish embryo mutant has a biallelic germline mutation in the zebrafish homologue of EXT2, resulting in truncated and functionally impaired heparan sulphate glycosaminoglycan chains. Chapter 4 continues with investigating the loss of heparan sulphate glycosaminoglycans in multiple osteochondroma patients, who have a heterozygous mutation in either EXT1 or EXT2. Here, no proteinuria, a specific renal phenotype, or changes to the glomerular endothelial glycocalyx were observed. In Chapter 5, the role of dynamin in proteinuric conditions and human disease is explored. Dynamin has been identified to play an important role in maintaining glomerular filtration barrier structure and function. In this chapter, this role is further specified as a dynamically regulated protective mechanism against the development of proteinuria.
Chapter 1 14 A previously unknown factor in the development of proteinuria is transmembrane protein 14A, which is discussed in Chapter 6. It is described as a protective element in the development of proteinuria through experiments in cell culture, a zebrafish embryo model, proteinuric rats, and human proteinuric kidney diseases.
General introduction and outline 15 1 References 1. Matsushita K, van d, V, Astor BC, Woodward M, Levey AS, de Jong PE, et al. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. Lancet. 2010;375(9731):2073-81. 2. Group CW. KDIGO clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013. 3. Diseases GBD, Injuries C. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2020;396(10258):1204-22. 4. Dane MJ, van den Berg BM, Lee DH, Boels MG, Tiemeier GL, Avramut MC, et al. A microscopic view on the renal endothelial glycocalyx. Am J Physiol Renal Physiol. 2015;308(9):F956-66. 5. Dane MJ, Khairoun M, Lee DH, van den Berg BM, Eskens BJ, Boels MG, et al. Association of kidney function with changes in the endothelial surface layer. Clin J Am Soc Nephrol. 2014;9(4):698-704. 6. Rennke HG, Patel Y, Venkatachalam MA. Glomerular filtration of proteins: clearance of anionic, neutral, and cationic horseradish peroxidase in the rat. Kidney Int. 1978;13(4):27888. 7. Kanwar YS, Farquhar MG. Presence of heparan sulfate in the glomerular basement membrane. Proc Natl Acad Sci U S A. 1979;76(3):1303-7. 8. Kanwar YS, Linker A, Farquhar MG. Increased permeability of the glomerular basement membrane to ferritin after removal of glycosaminoglycans (heparan sulfate) by enzyme digestion. J Cell Biol. 1980;86(2):688-93. 9. Esko JD, Selleck SB. Order out of chaos: assembly of ligand binding sites in heparan sulfate. Annu Rev Biochem. 2002;71:435-71. 10. Miner JH. Glomerular basement membrane composition and the filtration barrier. Pediatr Nephrol. 2011;26(9):1413-7. 11. Gu C, Chang J, Shchedrina VA, Pham VA, Hartwig JH, Suphamungmee W, et al. Regulation of dynamin oligomerization in cells: the role of dynamin-actin interactions and its GTPase activity. Traffic. 2014;15(8):819-38. 12. Gu C, Yaddanapudi S, Weins A, Osborn T, Reiser J, Pollak M, et al. Direct dynamin-actin interactions regulate the actin cytoskeleton. EMBO J. 2010;29(21):3593-606. 13. Sever S, Altintas MM, Nankoe SR, Moller CC, Ko D, Wei C, et al. Proteolytic processing of dynamin by cytoplasmic cathepsin L is a mechanism for proteinuric kidney disease. J Clin Invest. 2007;117(8):2095-104. 14. Soda K, Balkin DM, Ferguson SM, Paradise S, Milosevic I, Giovedi S, et al. Role of dynamin, synaptojanin, and endophilin in podocyte foot processes. J Clin Invest. 2012;122(12):4401-11. 15. Comper WD, Hilliard LM, Nikolic-Paterson DJ, Russo LM. Disease-dependent mechanisms of albuminuria. Am J Physiol Renal Physiol. 2008;295(6):F1589-600. 16. Tryggvason K, Patrakka J, Wartiovaara J. Hereditary proteinuria syndromes and mechanisms of proteinuria. N Engl J Med. 2006;354(13):1387-401. 17. Koop K, Eikmans M, Wehland M, Baelde H, Ijpelaar D, Kreutz R, et al. Selective loss of podoplanin protein expression accompanies proteinuria and precedes alterations in podocyte morphology in a spontaneous proteinuric rat model. Am J Pathol. 2008;173(2):315-26.
Mohamed A. Elmonem1,2, Ramzi Khalil3, Ladan Khodaparast4, Laleh Khodaparast4, Fanny O. Arcolino1, Joseph Morgan5, Anna Pastore6, Przemko Tylzanowski7,8, Annelii Ny9, Martin Lowe5, Peter A. de Witte9, Hans J. Baelde3,Lambertus P. van den Heuvel1,10, Elena Levtchenko1* 1 Department of Paediatric Nephrology & Growth and Regeneration, University Hospitals Leuven, KU Leuven, Leuven, Belgium 2 Department of Clinical and Chemical Pathology, Faculty of Medicine, Cairo University, Cairo, Egypt 3Department of Pathology, Leiden University Medical Centre, the Netherlands 4 Department of Cellular and Molecular Medicine, Switch Laboratory, VIB, University Hospitals Leuven, KU Leuven, Leuven, Belgium 5Faculty of Biology, Medicine and Health, University of Manchester, Manchester, United Kingdom 6Laboratory of Proteomics and Metabolomics, Children’s Hospital and Research Institute “Bambino Gesù” IRCCS, Rome, Italy 7Department of Development and Regeneration, Laboratory for Developmental and Stem Cell Biology, Skeletal Biology and Engineering Research Centre, University of Leuven, Leuven, Belgium 8Department of Biochemistry and Molecular Biology, Medical University, Lublin, Poland 9Laboratory for Molecular Bio-discovery, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium 10Department of Paediatric Nephrology, Radboud University Medical Centre, Nijmegen, the Netherlands CHAPTER 2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction Scientific Reports 2017, 7, 42583
Chapter 2 18 Abstract The human ubiquitous protein cystinosin is responsible for transporting the disulphide amino acid cystine from the lysosomal compartment into the cytosol. In humans, Pathogenic mutations of CTNS lead to defective cystinosin function, intralysosomal cystine accumulation and the development of cystinosis. Kidneys are initially affected with generalized proximal tubular dysfunction (renal Fanconi syndrome), then the disease rapidly affects glomeruli and progresses towards end stage renal failure and multiple organ dysfunction. Animal models of cystinosis are limited, with only a Ctns knockout mouse reported, showing cystine accumulation and late signs of tubular dysfunction but lacking the glomerular phenotype. We established and characterized a mutant zebrafish model with a homozygous nonsense mutation (c.706C>T; p.Q236X) in exon 8 of ctns. Cystinotic mutant larvae showed cystine accumulation, delayed development, and signs of pronephric glomerular and tubular dysfunction mimicking the early phenotype of human cystinotic patients. Furthermore, cystinotic larvae showed a significantly increased rate of apoptosis that could be ameliorated with cysteamine, the human cystine depleting therapy. Our data demonstrate that, ctns gene is essential for zebrafish pronephric podocyte and proximal tubular function and that the ctns-mutant can be used for studying the disease pathogenic mechanisms and for testing novel therapies for cystinosis.
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 19 Introduction Nephropathic cystinosis (MIM 219800) is an autosomal recessive lysosomal storage disorder characterized by the accumulation of the amino-acid cystine in the lysosomes of different body cells. It is caused by pathogenic mutations in the human CTNS gene encoding for cystinosin, the protein transporting cystine out of lysosomes1. In humans, cystinotic infants are born asymptomatic and stay healthy with normal growth parameters until approximately 6 months of life. After 6 months, infants manifest with dehydration, polyuria, polydipsia and rickets. The kidneys are initially affected in the form of defective proximal tubular reabsorption and increased urinary losses of amino-acids, glucose, phosphate, bicarbonate and proteins, or what is known as the renal Fanconi syndrome; however, this is usually rapidly followed by progressive glomerular damage, stunted growth and multiple organ dysfunction2. The aminothiol cysteamine, currently used as a specific treatment for cystinosis, can successfully deplete cystine in the lysosomal compartment and can delay the progression of the disease; however, it does not prevent the renal Fanconi syndrome and does not restore the lost renal function3. Over the past decade much interest has been given to study different pathogenic mechanisms of nephropathic cystinosis in an attempt to find better therapeutic agents targeting mechanisms other than cystine accumulation like autophagy4,5, oxidative stress6,7 and inflammation8,9. A successful mouse model for cystinosis was developed recently10 and was beneficial in revealing many pathogenic aspects of the disease11-15. However, the experimentation on mammalian models is usually time consuming, expensive and limited to a small number of test subjects16. Moreover, the murine model of cystinosis has a milder renal phenotype compared to humans and does not show signs of glomerular dysfunction starting in humans in early childhood17. Zebrafish (Danio rerio) was introduced as an attractive alternative to study pathogenic aspects in many genetic diseases18-23. This isdue to their rapid in vitro development, high fecundity, lower maintenance cost, optical transparency of the fertilized embryo, sequenced genome and the availability of gene down-regulation and gene editing technologies24. Furthermore, they emerged as a promising vertebrate model to study renal biology and associated medical conditions, especially in the fish embryonic and larval stages25,26. The zebrafish embryonic kidney, which is a functional pronephros, consists of a pair of segmented nephrons sharing a single glomerulus and showing astonishing histologic and functional similarities to the human nephron. This structure is formed approximately 24 hours post fertilization (24 hpf) and actual blood filtration starts approximately at 48 hpf27 offering a rapid and simple anatomical model for nephron
Chapter 2 20 patterning28, disease modeling29,30, identification of new genes affecting glomerular function and tubulogenesis16,31,32and drug testing33. In the current study, we investigated the pathological and functional characteristics of the first zebrafish mutant model of nephropathic cystinosis. We elucidated the main pathophysiological defects causing the diseased phenotype, which can be used for targeting novel therapeutic approaches. Results Zebrafish ctns gene The zebrafish ctns(ENSDARG00000008890) is a 10 exon gene in chromosome 11. It corresponds to the coding 10 out of 12 exons of the human CTNS(ENSG00000040531,17p13.2)34. The zebrafish Ctns protein (UniProt F1QM07, 384 aa) has a 60.2% amino-acid identity and 78.5% similarity to the human cystine transporter cystinosin (UniProt O60931, 367 aa), with 75.6% identity and 88.8% similarity in the regions of the seven transmembrane domains (Fig. 1a). The genetic zebrafish mutant line (ctns‒/‒) used in the current study is homozygous for the nonsense mutation c.706C>T (at the 10th base position of exon 8 of the zebrafish ctns gene) leading to a premature stop codon (TAA) and truncated protein at glutamine 236 (p.Q236X). The translation product is thus devoid of the last four transmembrane domains and both lysosomal targeting motifs at the 5th cytosolic loop and the C terminal tail, which is expected to render the protein non-functional (Fig. 1a, b). Up to date, no paralogue gene to ctns has been reported in zebrafish. Morphology of ctns‒/‒ zebrafish larvae We initially evaluated the morphological phenotype of morphant and ctns‒/‒ larvae at 4 days post fertilization (4 dpf) (N=191 and 334, respectively) in comparison to wildtype (wt) larvae (N=152) according to the grading system of Hanke et al., 201316. Here, oedema was graded in four stages. Stage I: no signs of oedema; stage II: mild oedema; stage III: intermediate oedema; and stage IV: severe total body oedema. Seven percent of living morphant larvae showed stage IV oedema and extreme body curvature, while 16% and 24% showed milder oedema (stages III and II, respectively). The morphological changes were less severe in the genetic ctns‒/‒ larvae, as they showed milder pericardial oedema and body curvature (stage II) in 14% of larvae. More severe forms of oedema (Stages III-IV) were very rare (3 and 1%, respectively), while the majority (82%) were not deformed (Fig. 2a and Supplementary Fig. S1 online).
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 21 Figure 1. Alignment of zebrafish Ctns protein and human cystinosin. (a) Amino-acid sequence alignment of the zebrafish Ctns protein and human cystinosin. The site of the genetic zebrafish model truncating mutation (c.706C>T; p.Q236X) is marked in red. Identical amino-acids are denoted by asterisks and similar amino acids by double dots. The seven transmembrane domains are highlighted in grey and the two lysosomal targeting motifs in black. (b) Exon 8 of the zebrafish ctns gene showing the wild-type (wt), the heterozygous (het) and the homozygous (hom)sequences for the c.706C>T mutation. Typical base sequence is marked above each electrophoretogram, while altered sequence is marked below.
Chapter 2 22 Cystine accumulation in ctns ‒/‒ larvae and organs of adult ctns ‒/‒ zebrafish Being a major pathologic feature of cystinosis, we measured cystine levels in both homogenized larvae and adult organs of mutant fish compared to the wt. ctns‒/‒ larvae at 6 dpf accumulated cystine about ten times higher compared to wt larvae (Fig. 2b). We also had a similar increase in homogenates of morphant larvae (data not shown). Furthermore, cystine levels in ctns‒/‒ larvae gradually decreased in response to increasing concentrations of cysteamine in the swimming water (Fig. 2b). Other thiol compounds related to cystine metabolism such as oxidized glutathione (GSSG), total glutathione (GSH) and free cysteine were also evaluated in homogenates of wt and mutant larvae (Fig. 2c-e). ctns‒/‒ larvae showed significantly higher levels of both GSSG and free cysteine compared to the wt; however, GSH was not significantly different. Cysteamine treatment significantly reduced abnormally high GSSG levels. In adult fish, the kidneys of 8 months ctns‒/‒ zebrafish demonstrated cystine concentrationsover 50 times of that detected in wt kidneys. Similarly, the ctns‒/‒ brain accumulated 10 times higher cystine, while the liver and heart accumulated double the amount of cystine in the wt (Fig. 2f-i). Thus, the results show that the inactivation of ctnsgene in zebrafish leads to the failure of cystine metabolism, recapitulating the human phenotype. ctns‒/‒ zebrafish show growth retardation and higher rates of embryonic mortality Next, we investigated if the defects in cystine metabolism had other effects on zebrafish. Therefore, we monitored the developmental stages of both ctns‒/‒ and wt zebrafish embryos in four independent crossings over the first three days of maturation at predetermined time points (3h, 6h, 24h, 48h and 72 hpf). ctns‒/‒ embryos showed significant delay in development at all time points investigated, although the difference was more striking at early time points (≤ 24h)(Fig. 3). Additionally, the percentage of dead embryos during the first 3 days post fertilization was significantly higher in ctns‒/‒ zebrafish (101/363 (27.8%)), when compared to wt (33/322 (10.2%)), P<0.001. Hatching was also relatively delayed in ctns‒/‒ embryos at both 48hpf and 72hpf time points. In a different set of experiments, mortality rates were improved with therapeutic doses of cysteamine (Fig. 3f). Hatching rates were also partially normalized by cysteamine therapy (Supplementary Fig. S2 online). Thus, the deregulation of ctns gene led to overall developmental delay, and increased embryonic mortality that are partially restored by cysteamine.
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 23 Figure 2. Morphology and cystine measurements. (a) Morphology of wild-type and ctns‒/‒ larvae at 4 dpf. Wild-type larva shows normal morphology, while mutant ctns‒/‒ larvae show various degrees of developmental delay and deformity: upper larva show signs of growth retardation in the form of slightly bigger yolk, bulging heart and bent-down head, while the middle and lower larvae show mild and severe deformity, respectively
Chapter 2 24 (bars=1 mm). (b) Cystine content in homogenates of 6 dpf wt or ctns‒/‒ zebrafish larvae. ctns‒/‒ larvae were either free of treatment (N=133) or subjected to 0.1 or 1.0 mM of cysteamine in the swimming water (N=111 and 121 larvae, respectively). Comparison was performed with wt larvae (N=191). (c) Oxidized glutathione (GSSG) content in homogenates of 6 dpf wt or ctns‒/‒ zebrafish larvae (same conditions and larval numbers as cystine). (d) Total glutathione (GSH) content in homogenates of 6 dpf wt or ctns‒/‒ zebrafish larvae.ctns‒/‒ larvae were either free of treatment (N=80) or subjected to 0.1 or 1.0 mM of cysteamine in the swimming water (N=108 and 104 larvae, respectively). Comparison was performed with wt larvae (N=158). (e) Free cysteine content in homogenates of 6 dpf wt or ctns‒/‒zebrafish larvae (same conditions and larval numbers as GSH). (f-i) Cystine content in homogenates of 8-month-old adults (Kidney, brain, heart and liver, respectively) (N=3 of each genotype). Concentrations of cystine and other thiol compounds were expressed as nmol/mg protein. * P<0.05, ** P<0.01, *** P<0.001. ctns‒/‒ zebrafish larvae have increased apoptosis rate that can be ameliorated by cysteamine We further addressed whether, as reported previously in human and mouse tissues,7,11,12 cystinosis triggered apoptosis in zebrafish larvae. Apoptosis was investigated in surviving wt and ctns‒/‒ larvae at 5 dpf using the Acridine Orange (AO) fluorescent dye which binds to DNA of apoptotic cells and spares necrotic cells. Cystinotic larvae were naïve to treatment or treated with 0.1 mM of cysteamine (N=10 for each condition). The apoptotic spots in the untreated ctns‒/‒ larvae were clearly visible and significantly increased compared to wt larvae. Interestingly, the low dose of cysteamine significantly reduced both number and intensity of apoptotic spots, P<0.001 (Fig. 4a-d). We further confirmed the increased rate of apoptosis through the detection of positive staining for caspase-3 by immunohistochemistry in 5 dpf ctns‒/‒ larvae. Apoptotic signals in immunohistochemistry were not restricted to skeletal structures but were also present in internal organs especially in proximal tubules and in the liver (Fig. 4e,f). A higher caspase-3 enzyme activity performed by a luciferase based assay was detected in the homogenates of ctns‒/‒ larvae compared to the wt at 5 dpf, P<0.001 (Fig. 4g). ctns‒/‒ zebrafish larvae have normal locomotor activity In order to evaluate if there are any early behavioural or kinetic abnormalities in the cystinotic zebrafish larvae, we monitored the locomotor activity of 5 dpf ctns‒/‒ larvae (N=56) in comparison to wt larvae (N=52) under light and dark conditions. The quantification of locomotor activity (in actinteg units) did not reveal any significant difference between the two genotypes regardless of the lighting conditions (P=0.416 in light and P=0.279 in dark) (Supplementary Fig. S3 online). Figure 2. Continued
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 25 Figure 3. Early developmental stages of zebrafish ctns‒/‒ embryos and response to cysteamine therapy. Embryonic development was monitored over the first 3 days of life at predetermined time points: (a) 3hpf, (b) 6hpf, (c) 24hpf, (d) 48hpf and (e) 72hpf. The outcomes of four different mating settings, 16 females and eight males from each genotype were used (363 ctns‒/‒ and 322 wt embryos). Percentages of different developmental stages at each time point were calculated per the total number of living embryos for each genotype at each time point. * P<0.05, *** P<0.001 against wildtype percentages using Pearson chi-square test. (f) Effect of different doses of cysteamine therapy on mortality rates of ctns‒/‒ larvae during the first 96 hpf. * P<0.05, ** P<0.01, *** P<0.001 against untreated ctns‒/‒ larvae using student's t test.
Chapter 2 26 Figure 4. Apoptosis in ctns‒/‒ larvae. (a-d) Acridine orange: Five dpf wt larvae and ctns‒/‒larvae, naïve to treatment or treated with 0.1 mM of cysteamine (N=10 for each group), were incubated with Acridine Orange (AO). Fluorescent spots(white arrows) were delineated in high magnification mode and quantified by ImageJ software. (a) A representative tail segment of 5 dpf wt larva (bar= 200µm). (b) A representative tail segment of 5 dpf ctns‒/‒untreated larva (bar= 200µm). (c) A representative tail segment of 5 dpf ctns‒/‒ larva treated with 0.1 mM cysteamine (bar= 200µm). (d) Quantitation of the relative fluorescence intensity of apoptotic spots. Average intensity of untreated ctns‒/‒larvae was set at 100%.*** P<0.001 against untreated ctns‒/‒larvae. (e,f) Caspase-3 immunohistochemistry. (e) Representative images showing increased apoptotic signal over the proximal tubule in 5 dpf ctns‒/‒ larva (left) compared to the negative control (right), bar= 10µm. pt, proximal tubule. (f)Representative images showing increased apoptotic signal over the liver in 5 dpf ctns‒/‒ larva (left) compared to the negative control (right), bar= 30µm. Rabbit serum was used for the negative control sections instead of 1ry Ab.(g) Caspase-3/7 enzyme activity. Quantitation of Caspase-3/7 enzyme activity by a luciferase based assay in the homogenates of 5 dpf wt and ctns‒/‒larvae (On average 60 larvae over 3 separate homogenates for each genotype were used). Results were expressed in luminescence units (RLU)/ µg protein of each homogenate. *** P<0.001.
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 27 ctns‒/‒ zebrafish pronephros shows enlarged lysosomes in proximal tubular cells and partial podocyte foot process effacement Analysis by light microscopy showed no apparent glomerular or tubular abnormalities compared to wt (Fig. 5a, b). Analysis by block face scanning electron microscopy revealed however that proximal tubular epithelial cells (PTECs) in ctns‒/‒ pronephros had numerous and enlarged lysosomes compared to wt (Fig. 5c, d). We calculated the numbers and average surface area of lysosomes in complete cut-sections of wt and ctns‒/‒ larvae at the level of proximal tubules (n= 10 each). Per cut-section lysosomal numbers were higher in the proximal tubules of ctns‒/‒ larvae (68.4±4.7) compared to wt larvae (24.5±3.8), P<0.001. Average lysosomal surface area was also higher in the ctns‒/‒ larvae (1.38±0.1 µm2) compared to the wt (0.51±0.03 µm2), P<0.001 (Fig.5e). On the other hand, cystinotic PTECs did not show cystine crystal accumulation or brush border flattening. The ultrastructural analysis of podocytes of ctns‒/‒ larvae showed partial foot process effacement and narrowed slit diaphragmatic spaces when compared to wt larvae, while glomerular basement membrane appeared to be of normal thickness (Fig. 5f, g). To assess podocyte foot process effacement in a quantitative manner, average podocyte foot process width (FPW) was measured. A previously described formula was used to perform this analysis35. ctns‒/‒larvae showed higher podocyte FPW (0.62±0.09 µm) when compared to wt larvae (0.51±0.05 µm), P=0.033 (Fig.5h). Thus, at the cellular level the ctns‒/‒ larvae also showed some of the human pathological features of the disease.
Chapter 2 28 Figure 5. Morphology of the pronephros of ctns‒/‒ larvae compared to the wt. (a) H&E stained cut-section of a 6 dpf wt larva at the level of the glomerulus and proximal tubules(bar=50µm). (b) H&E stained cut-section of a 6 dpf ctns‒/‒ larva at the level of the glomerulus and proximal tubules showing no apparent abnormality (bar=50 µm). (c) Block face scanning EM image of the proximal tubule of a 4 dpf wt larva (bar=5 µm).Demarcated area was magnified (right) to show size and distribution of lysosomes (asterisks) in the wt (bar=2 µm).(d) Scanning EM image of the proximal tubule of a 4 dpf ctns‒/‒ larva showing intact brush border (bar=5 µm). Demarcated area was magnified (right) to show larger number of lysosomes (asterisks) many of which were significantly enlarged in size compared to the wt (bar=2 µm). (e) Quantitation of the number and surface area of lysosomes in cut sections at the level of proximal tubules in
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 29 both genotypes. (f) Transmission EM image of the glomerulus of a 6 dpf wt larva showing normal foot processes (bar=2 µm). A magnified EM image (right) of podocytes of 6 dpf wt larva showing preserved podocytes slit diaphragms (bar=1 µm). (g) Transmission EM image of the glomerulus of a 6 dpf ctns‒/‒ larva showing partial foot process effacement (black arrows) (bar=2 µm). A magnified EM image (right) of podocytes of 6 dpf ctns‒/‒ larva showing narrowed podocyte slit diaphragmatic spaces (white arrows) (bar=1 µm). (h) Quantitation of podocyte foot process width (FPW) in cut sections at the level of the glomerulus in both genotypes. bb, brush border; bs, Bowman’s space; g, glomerulus; n, nucleus; pt, proximal tubule. * P<0.05, *** P<0.001 between the 2 genotypes using student's t test. ctns‒/‒ zebrafish pronephros shows signs of glomerular disease defective glomerular permselectivity Since many aspects of human and zebrafish cystinosis were similar, we investigated the functional consequences of the disruption of ctns gene in zebrafish larvae. One of the functional tests for the zebrafish kidney is measuring the time required for dextran clearance from the pronephros. In case of a glomerular defect, high molecular weight(HMW)dextran is expected to be lost more rapidly from the vasculature due to impaired glomerular filtration barrier (GFB). Thus, we injected fluorescent labelled 70kDa dextran into the vascular system of 72 hpf larvae (N=20 of each genotype). After 24 hours we monitored the fluorescence intensity of each larva over the retinal vascular bed16. The fluorescence intensity in ctns‒/‒ larvae was significantly lower compared to wt larvae, P<0.001 (Fig. 6a-c). Furthermore, the number of 70-kDa dextran droplets visualized passing through the proximal tubular wall of 72 hpf ctns‒/‒ larvae fixed in 4% paraformaldehyde (PF) 1h after injection was significantly higher when compared to wt larvae denoting also the increased passage of the 70-kDa dextran in the glomerular filtrate (N=10 of each genotype), P=0.031 (Fig. 6). decreased glomerular filtration rate (GFR) Human cystinosis patients develop a slow and gradual decrease in GFR usually starting during childhood. Hence we evaluated the GFR of mutant ctns‒/‒ larvae compared to that of the wt by injecting FITC-inulin into the vascular system of larvae at 96 hpf36. Inulin is freely passing through the glomerular membrane, not reabsorbed and not excreted from the tubular cells, thus is widely used for the assessment of GFR. We monitored the percent of fluorescence intensity decline over 3 fixed anatomical positions in the caudal artery of each larva after 4 hours of injection (Supplementary Fig. S4 online). The percentage of decline of fluorescent intensity in ctns‒/‒ larvae (65.3±5.1%, N=43) was slightly but significantly reduced compared to wt larvae (68.7±4.2%, N=45), P<0.001 denoting the early affection of GFR in ctns‒/‒ zebrafish larvae. The difference was significant in two Figure 5. Continued
Chapter 2 30 independent experiments, P=0.01 and 0.032. These results emphasize the similarity between zebrafish and human cystinosis patients. ctns‒/‒ zebrafish pronephros show impaired proximal tubular function impaired endocytosis of low molecular weight dextran Another functional aspect that we investigated was the tubular reabsorption. Here we carried out a histological evaluation of the number of dextran droplets in the proximal tubular cell wall of fixed larvae after the injection of a fluorescent labelled low molecular weight (LMW) dextran (4-kDa) into the vascular system of 72 hpf larvae in both ctns‒/‒ and wt (N=10 of each genotype). The low molecular weight dextran freely passes the glomerular filtration barrier and is efficiently reabsorbed by the proximal tubular endosomal machinery. Interestingly, the fluorescence over the proximal tubule of ctns‒/‒ larvae was virtually absent compared to the injected wt larvae (P<0.001, Fig. 6), suggesting that proximal tubular reabsorption was defective in the ctns‒/‒ larvae. altered abundance and localization of the endocytic receptor megalin Altered apical abundance of the multi-ligand receptor megalin has been linked with the abnormal endocytosis and defective function of cystinotic PTECs in both mice and human cells11,37. We evaluated the abundance and localization of megalin in the proximal tubules in both 5 dpf ctns‒/‒ and wt larvae, as described previously38. The overall level of megalin present in ctns‒/‒ larvae was about half of that in the wt (Fig. 7a-e); however, the most striking feature was the altered distribution of megalin in the cystinotic PTECs where it accumulated in sub-apical punctate rings or cytoplasmic vacuoles when compared to the wt, where it was more evenly distributed in the apical brush border. This altered localization denotes a defective recycling of megalin from apical endosomes, which may explain, at least partially, the disturbed endocytosis in cystinotic larvae. We also evaluated the overall megalin gene (lrp2a) expression in homogenized 3 dpf and 6 dpf ctns‒/‒ and wt larvae. At the RNA level, there was no statistical significant difference detected between both genotypes(Fig. 7f) similar to zebrafish models of other genetic disorders with impaired proximal tubular endocytosis, such as Lowe syndrome38. The reduced abundance of megalin protein at the proximal tubular brush border in the absence of decreased transcript levels is consistent with the abnormal recycling of the protein rather than a decreased transcription and is similar to the findings in humans37,39.
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 31 Figure 6. Functional evaluation of glomerular permeability and tubular reabsorption of ctns‒/‒ larvae. (a-c) Eye fluorescence assay: peak fluorescence intensity in the retinal vascular bed of ctns‒/‒ zebrafish larvae and wild-type larvae (N=20 each). Fluorescence intensities were evaluated using fixed diameter circles by the ImageJ software. (a) A representative wild-type 4 dpf larva(24h post-injection) (bar=200 µm). (b) A representative ctns‒/‒4 pdf larva(24h post-injection)(bar=200 µm). (c) A diagrammatic
Chapter 2 32 representation of peak fluorescence intensities in the retinal vascular bed of both genotypes. (d-i) Histopathological functional evaluation: (d) A representative proximal tubule of wt larva injected with the 70-kDa labelled dextran(bar=10 µm). (e) A representative proximal tubule of wt larva injected with the 4-kDa labelled dextran (bar=10 µm). (f) A representative proximal tubule of ctns‒/‒ larva injected with the 70-kDa labelled dextran(bar=10 µm).(g) A representative proximal tubule of ctns‒/‒ larva injected with the 4-kDa labelled dextran (bar=10 µm). (h) A higher magnification of the proximal tubules of both genotypes showing internalized 70-kDa dextran within cytosolic puncta that likely correspond to endocytic compartments (marked areas in panels d and f) (bars=5 µm). (i) Quantitation of the number of dextran puncta in both high and low molecular weight dextran injections in both genotypes (N=10 for each genotype and each condition). * P<0.05, *** P<0.001. Figure 7. Megalin expression in proximal tubular cells. (a) Transverse fluorescent image of the proximal pronephric region of wt 5 dpf larva labelled with anti-megalin antibody (bar=10 µm). (b) Higher magnification image of wt proximal tubule (square in panel a) showing mainly the diffuse distribution of megalin at the cellular brush border (bar=3 µm). (c) The proximal pronephric region of ctns‒/‒ 5 dpf larva labelled with anti-megalin antibody (bar=10 µm). (d) Higher magnification image of ctns‒/‒ proximal tubule (square in panel c) showing majority of megalin staining in sub-apical intracytoplasmic vacuoles (white arrows) (bar=3 µm). Outer boundaries of proximal tubules were delineated with green, lumen with red, and nuclear boundaries were delineated with white. (e) Quantitation of megalin protein abundance in proximal tubules of wt and ctns‒/‒larvae (N=5 for each genotype). (f) Quantitation of the megalin encoding lrp2a RNA expression in homogenized larvae of 6 dpf wt vs ctns‒/‒larvae (N= 5 individually separated RNA samples for each genotype). * P<0.05. Figure 6. Continued
2 Cystinosis (ctns) zebrafish mutant shows pronephric glomerular and tubular dysfunction 33 Discussion The last common ancestor of humans and zebrafish was a marine vertebrate that lived approximately 450 million years ago; however, 70% of protein-coding human genes are related to genes found in the zebrafish and 84% of genes known to be associated with human disease have a zebrafish counterpart40,41. In the current study we established and characterized actns‒/‒zebrafish mutant and uncovered many important aspects of the renal pathophysiology resulting in the functional abnormalities of the early larval stage of the ctns‒/‒ zebrafish. The key-findings of the study are: 1) significant cystine accumulation, increased mortality and increased rate of apoptosis in the ctns‒/‒ zebrafish which are partially responsive to cysteamine treatment. 2) early impairment of pronephros affecting both glomerular and proximal tubular function. Significant accumulation of cystine, the main pathologic landmark of nephropathic cystinosis, validated our model and confirmed the pathogenic nature of the mutation. Both morphant and ctns‒/‒ zebrafish larvae showed comparable phenotypes. We preferred to proceed with embryos and larvae of the genetic ctns‒/‒ model, as it is well known that the phenotype severity and toxicity of morphant models are dependent on the morpholino dose injected42, which is not an issue in the genetic model. Although cystinosin in humans is ubiquitously expressed, the expression is especially high in the kidneys43. Interestingly, the adult ctns‒/‒zebrafish kidneys accumulated the highest concentrations of cystine (>50 times the wt) contrasting the murine model of cystinosis in which the highest cystine accumulations were observed in adult liver and spleen10,44. Interestingly, similar to our data in zebrafish homogenates, Wilmer et al., 2011 detected a significant increase in GSSG in cystinotic PTECs compared to wt without alterations in GSH levels45. In their in vitro study PTECs in culture responded to the antioxidant cysteamine treatment by decreasing GSSG and increasing GSH45 while in vivo in fish only a significant fall in GSSG was observed. Cystine accumulation has been shown to cause increased apoptosis rate inhuman tissues and in the mouse model of cystinosis7,11,12. The suggested mechanism is enhanced cysteinylation of pro-apoptotic enzyme protein kinase C delta due to the lysosomal overload and increased lysosomal membrane permeability46. Moreover, oxidative mitochondrial stress and ER stress have been attributed to enhanced apoptosis in cystinotic cells7,12.Similarly, in our zebrafish model the AO rapid screening technique revealed increased number of apoptotic signals in whole larvae. These data were confirmed by caspase-3 immunohistochemistry and enzyme activity in homogenates of the ctns‒/‒ larvae. Importantly, AO apoptotic signals were significantly reduced by therapeutic doses
www.ridderprint.nlRkJQdWJsaXNoZXIy MTk4NDMw