family in the regulation of glomerular permeability. In addition, a role for FAT4 in mammalian planar cell polarity (PCP) signaling has recently been confirmed 95.
Ectopic Notch Activation in Developing Podocytes Impairs Slit Diaphragm Formation and Induces Abnormal Podocyte Differentiation by Aoife Mary Waters A thesis submitted in conformity with the requirements for the degree of Masters of Science Institute of Medical Sciences, University of Toronto
© Copyright by Aoife Waters 2009
Ectopic Notch Activation in Developing Podocytes Impairs Slit Diaphragm Formation and Induces Abnormal Podocyte Differentiation Aoife Mary Waters Masters of Science Institute of Medical Sciences University of Toronto 2009 ABSTRACT
Podocytes are terminally differentiated epithelial cells which regulate glomerular permselectivity by their cell-cell junctions, known as slit diaphragms (SD). Notch signaling regulates podocyte cell fate specification and downregulation of Notch targets occurs with terminal podocyte differentiation. The effects of constitutive Notch activation in developing podocytes on podocyte differentiation and function were determined using a podocyte-specific Cre-lox-p (Neph/Cre) approach. Proteinuria was noted shortly after birth denoting loss of glomerular permselectivity in transgenic mice (CRE;NIC mice). Histologic and molecular analyses of CRE;NIC -expressing mice at onset of proteinuria, show morphologic and cellular changes in podocytes including de-differentiation, proliferation and de novo expression of Pax2. Prior to onset of proteinuria, lower protein levels of key SD proteins are observed while SD mRNA expression is preserved in CRE;NIC mice. Consequently, constitutive Notch signaling in developing podocytes opposes terminal differentiation with deleterious consequences on SD assembly and thereafter, glomerular permselectivity.
ii
ACKNOWLEDGEMENTS I sincerely thank my supervisor, Dr Tino Piscione, for his tireless dedication to my academic development and for allowing me to work on an intriguing model of nephrosis. His mentorship has facilitated my conceptual development for which I am eternally grateful. My gratitude also extends to my co-supervisor, Professor Norm Rosenblum, who has provided support and sound advice throughout my training. His energy and committment to scientific research has been a huge inspiration. I am eternally indebted to Megan Wu and Tuncer Onay, of the Piscione laboratory, for their technical tuition and mentorship. Most of all, I wish to thank Megan for her patience and friendship over the past few years. I sincerely thank my committee members, Dr Jane McGlade and Dr Susan Quaggin for supporting my academic development on this project. Members of the Rosenblum, Egan, Danska, Guidos and Quaggin laboratories are graciously acknowledged for their advice, use of reagents and friendships over the course of this work. Sincere thanks are extended to my medical colleagues at the Hospital for Sick Children, Toronto for facilitating the training of an international colleague. My experience has been enrichened by their dedication to the care of all children, especially those with kidney disease. In particular, I thank my Division chief, Professor Denis Geary for his inspiring energy, committment and friendship. Most of all, I would like to dedicate this work to my family and in particular, my husband, Jonathan, for his understanding and support of my work.
iii
TABLE OF CONTENTS
Page Number
Abstract
(ii)
Acknowledgements
(iii)
Table of Contents
(iv)
List of Abbreviations
(vi)
List of Tables
(ix)
List of Figures
(x)
I. Introduction
1.
1. Podocyte Morphogenesis 1.1.1
The association between normal podocyte morphology
3.
and maintenance of glomerular permselectivity 1.1.2
Overview of kidney development
4.
1.1.3
Podocyte differentiation
5.
1.1.4 Development of the glomerular filtration barrier
6.
1.1.5
7.
Development of the slit diaphragm
2. Molecular Regulation of Podocyte Differentiation 1.2.1 Transcriptional regulation of podocyte differentiation
8.
1.2.2 Posttranscriptional regulation of podocyte differentiation
11.
1.2.3 Integrins and foot process assembly
12.
1.2.4 Molecular regulation of slit diaphragm development
12.
1.2.5 The glomerular slit diaphragm in acquired glomerulopathies
17.
iv
3. Notch Signaling and Glomerulogenesis 1.3.1 The Notch signaling pathway
18.
1.3.2 Expression of Notch during glomerulogenesis
20.
1.3.3 The role of Notch in podocyte cell fate specification
22.
1.3.4 The role of Notch in glomerular vascular tuft formation
23.
II. Ectopic Notch Activation in Developing Podocytes Causes Glomerulosclerosis 2.1
Introduction
26.
2.2
Materials and Methods
27.
2.3
Results
30.
2.4
Discussion
46.
III. Ectopic Notch Activation in Developing Podocytes Impairs Formation of the Glomerular Slit Diaphragm by Decreasing Slit Diaphragm Protein Expression 3.1
Introduction
50.
3.2
Materials and Methods
51.
3.3
Results
53.
3.4
Discussion
76.
IV. Conclusions and Future Directions
82.
List of Original Publications
85.
References
86.
v
LIST OF ABBREVIATIONS ANK
Ankyrin repeats
ATP
Adenosine 5’ triphosphate
cDNA
Complementary DNA
DAPI
4′6-Diamidino-2-phenylindole 2HCl
DDS
Denys-Drash syndrome
Dll
Delta
DMS
Diffuse mesangial sclerosis
EGF
Epidermal growth factor
EPC
Endothelial progenitor cell
ER
Endoplasmic reticulum
FOXC2
Forkhead box C2
FP
Foot processes
FSGS
Focal segmental glomerulosclerosis
Gapdh
Glyceraldehyde-3- phosphate dehydrogenase
GBM
Glomerular basement membrane
GFB
Glomerular filtration barrier
GN
Glomerulonephritis
GSI
Gamma secretase inhibitors
Hbs
Hibris
HBSS
Hank’s balanced salt solution
HES
Hairy Enhancer of Split
IgAN
IgA nephropathy
ILK
Integrin linked kinase
IMPDH
Inosine 5’-monophosphate dehydrogenase
IrreC-rst
Roughest
Jag
Jagged
LFng
Lunatic fringe
LMX1B
Lim domain protein
LNR
Lin-Notch repeats
LTL
Lotus tetragonolobus lectin vi
Mafb
Kreisler
MAM
Mastermind
MCN
Minimal change nephropathy
Mf2
Forkhead domain transcription factor
MM
Metanephric mesenchyme
MN
Membranous nephropathy
NCAM
Neural cell adhesion molecule
Neph-CRE
Nephrin cre recombinase
NICD
Notch intracellular domain
P
Postnatal
PAN
Puromycin aminonucleoside nephrosis
PBS
Phosphate-buffered saline
PCR
Polymerase chain reaction
PDGFRB
Platelet-derived growth factor receptor b
PECAM1
Platelet-endothelial cell adhesion molecule 1
PLCε1
Phospholipase Cε1
POD1
Podocyte-expressed 1
RAM
RBPJ-κ-associated molecule
RBPJ-κ
Recombinant binding protein J-kappa
RT-PCR
Reverse transcriptase polymerase chain reaction
RV
Renal vesicle
SD
Slit diaphragms
SDS
Sodium dodecyl sulfate
SEM
Scanning electron microscopy
SMRT
Silencing mediator of retinoic and thyroid hormone receptors
TAD
Transactivation domain
TEM
Transmission electron microscopy
TGFβ
Transcription growth factor beta
UB
Ureteric bud
vii
VEGF-A Vascular endothelial growth factor-A WT1
Wilm’s tumour-1
viii
LIST OF TABLES
Page Number
Table 1. Expression of Notch pathway components during glomerulogenesis
22.
Table 2. Urine protein quantification by protein:creatinine ratio over time
34.
in wildtype, CRE(-);NOTCH-IC, and CRE(+);NOTCH-IC transgenic mice. Table 3. Analysis of podocyte cell proliferation in glomeruli of CRE(+);
43.
NOTCH-IC and CRE(-);NOTCH-IC transgenic mice (3 mice per age group). Table 4. Median number of slit diaphragms and foot processes per unit (μm)
57.
length of GBM in CRE(-);NOTCH-IC and CRE(+);NOTCH-IC transgenic mice at birth. Table 5. Histomorphometrical analysis of nephrin expression in total number
68.
of WT1+ cells in glomeruli of CRE(+);NOTCH-IC and CRE(-);NOTCH-IC transgenic mice at birth. Table 6. Histomorphometrical analysis of nephrin expression in MYC-
69.
NOTCH-IC podocytes versus MYC-negative NOTCH-IC podocytes in
glomeruli of CRE(+); NOTCH-IC and CRE(-);NOTCH-IC transgenic mice at birth. Table 7. Histomorphometrical analysis of nephrin expression in MYC-positive
70.
NOTCH-IC podocytes versus MYC-negative NOTCH-IC podocytes in glomeruli of CRE(+); NOTCH-IC and CRE(-);NOTCH-IC transgenic mice at one week of age (P7).
ix
LIST OF FIGURES
Page Number
Figure 1. The glomerular filtration barrier.
3.
Figure 2. Morphological features of the glomerular podocyte.
4.
Figure 3. Serial morphogenetic events of the nephrogenic blastema.
5.
Figure 4. Development of the glomerular filtration barrier.
7.
Figure 5. Morphogenetic events of podocyte differentiation.
8.
Figure 6. Electron micrographic features of the podocyte-endothelial interface.
9.
Figure 7. Molecular components of the glomerular slit diaphragm.
13.
Figure 8. The Notch signaling pathway.
19.
Figure 9. Expression of Notch pathway components during glomerulogenesis.
21.
Figure 10. Podocyte-specific expression of MYC-NOTCH-IC in glomeruli of
32.
newborn CRE(+);NOTCH-IC transgenic mice. Figure 11. Histological progression of glomerular lesions in CRE(+);NOTCH-IC
35.
transgenic mice. Figure 12. Transmission electron micrographs (TEM) of CRE(+);NOTCH-IC
37.
mouse glomeruli. Figure 13. Glomerular Wt1, Nphs1, and Nphs2 mRNA expression in CRE(+);
38.
NOTCH-IC transgenic mice. x
Figure 14. Decreased nephrin protein expression in MYC-NOTCH-IC-expressing
41.
podocytes. Figure 15. Analysis of glomerular cell proliferation in CRE (+);NOTCH-IC mice.
42.
Figure 16. Analysis of Pax2 expression in podocytes of CRE(+);NOTCH-IC mice.
45.
Figure 17. Ultrastructural features of glomeruli in CRE(+);NOTCH-IC transgenic
56.
mice at birth. Figure 18. Glomerular Wt1 and Nphs1 mRNA expression in newborn
59.
CRE(+);NOTCH-IC transgenic mice. Figure 19. Glomerular Nphs1, Nphs2, Neph 1-3, Zo1, Fat1 and Wt1 expression
60.
in CRE(+);NOTCH-IC and CRE(-);NOTCH-IC transgenic mice prior to onset of proteinuria (P7). Figure 20. Glomerular Nphs1, Nphs2, Neph 1-3, Zo1, Fat1 and Wt1 expression
61.
in CRE(+);NOTCH-IC and CRE(-);NOTCH-IC transgenic mice at onset of proteinuria (P14). Figure 21. Semiquantitative reduction in nephrin and podocin protein
63.
expression precedes proteinuria in CRE(+);NOTCH-IC transgenic mice. Figure 22. Nephrin loss in transgenic podocytes precedes proteinuria in
66.
CRE(+);NOTCH-IC transgenic mice. Figure 23. Scoring pattern for nephrin staining.
67.
Figure 24. Attenuation of nephrin signal occurs in association with
72.
attenuation of podocin signal in MYC-NOTCH-IC expressing podocytes of
xi
CRE(+);NOTCH-IC transgenic mice at birth. Figure 25. Attenuation of nephrin signal remains associated with
73.
attenuated podocin signal at P7 and P10 prior to onset of proteinuria in CRE(+);NOTCH-IC transgenic mice. Figure 26. Attenuated ZO-1 staining in MYC-NOTCH-IC expressing podocytes
74.
prior to onset of proteinuria in CRE(+);NOTCH-IC transgenic mice. Figure 27. Attenuated synaptopodin in MYC-NOTCH-IC expressing podocytes
76.
in CRE(+);NOTCH-IC transgenic mice prior to onset of proteinuria.
xii
I. INTRODUCTION
1
The mammalian kidney plays a key role in the maintenance of intravascular volume and body fluid homeostasis. Critical to this function, is the ability of the renal glomerulus to deliver a plasma ultrafiltrate to the proximal nephron that is devoid of protein 1. Ultrafiltration is achieved by unique anatomical considerations that include a specialized basal lamina lying interposed between a fenestrated glomerular capillary endothelium and filtration slits formed by the epithelial monolayer comprised of podocytes (Figure 1) 2. Defective glomerular ultrafiltration is characterized by loss of glomerular filtration barrier (GFB) permselectivity and loss of plasma proteins (eg albumin) into the urine (also known as proteinuria). Proteinuria associated with scarring of the renal glomerulus (glomerulosclerosis), leads to end-stage renal disease and affects about 1.2 million people worldwide, at a growing rate of 6-7% annually 3. Defects in any of the three components of the GFB can result in proteinuria 2,4,5. Podocytes, in particular, play an important role in maintenance of GFB permselectivity through regulation of foot process (FP) morphology and slit diaphragm (SD) integrity 5,6. Terminal podocyte differentiation is associated with formation of tertiary FPs and maturation of the SD 7. While mutations in SD components can account for a proportion of proteinuric syndromes, mutations in other podocyte genes such as those regulating podocyte differentiation have also been implicated in proteinuria 8. Maintenance of podocyte differentiation is essential to podocyte function and is challenged in the context of podocyte injury resulting in divergent glomerular morphologies that are dependent on the nature of the inciting injury. In response to injury, podocytes may undergo a number of different fates that include proliferation and de-differentiation or apoptosis. Therefore, understanding the molecular pathways which govern podocyte differentiation may help to elucidate novel molecular mechanisms that regulate podocyte function under normal and pathological conditions. The Notch signaling pathway plays a key role in early podocyte differentiation 8,9. The objective of my research is to determine the effects of sustained Notch activity in developing podocytes on podocyte differentiation and consequently, the effects on SD development and GFB permselectivity.
2
1. Podocyte Morphogenesis 1.1.1
The association between normal podocyte morphology and maintenance of glomerular filtration barrier permselectivity. Podocytes comprise the third layer of the GFB and are highly specialized renal
epithelial cells. Morphologically, they are characterized by cytoplasmic processes (primary processes) which extend out from voluminous cell bodies and further branch into secondary and tertiary foot processes (FPs) which then interdigitate with FPs from neighbouring cells (Figure 2). A highly organized and tightly regulated actin cytoskeleton underlies the unique cytoarchitecture of podocytes as they wrap around the underlying glomerular capillary loops 6.
Figure 1. The glomerular filtration barrier. Blood in the glomerular capillaries is filtered across the fenestrated endothelium and the glomerular basement membrane (GBM) and through the filtration slits between podocyte foot processes to produce the plasma ultrafiltrate (Adapted from Ronco et al 2).
3
Podocytes function to maintain GFB permselectivity, through the regulation of FP morphology. In proteinuric diseases, loss of GFB permselectivity is strongly associated with simplification of FP morphology, (also known as effacement) and is characterized by loss of normal FP cytoarchitecture and interdigitation. Under normal physiological conditions, filtration slits, measuring 25-60nm wide, exist between FP interdigitations, through which plasma is filtered. High resolution electron tomography has shown that filtration slits are occupied by three dimensional multiprotein complexes, known as slit diaphragms (SD) which regulate glomerular permselectivity 2,10, 11, 12. The composition of the multiprotein complexes that comprise the SD is discussed in greater detail in Section 1.2.2.
Figure 2. Morphological features of the glomerular podocyte. (A) Scanning electron micrograph (SEM) of murine glomerulus, illustrating podocytes wrapped around glomerular capillary loops (black arrow). (B) High magnification image of SEM showing podocyte cell bodies (cb) and primary (p), secondary (s) and tertiary foot processes (fp) which interdigitate with foot processes from neighbouring cells (black arrow). 1.1.2 Overview of kidney development Mammalian kidney development begins at 11.5d post coitum in the mouse and at 5 weeks of gestation in humans 13. Induction of the metanephric mesenchyme (MM) follows inductive cues from the outgrowing ureteric bud (UB) (Figure 3). Reciprocal UB-
4
MM interactions ensue with MM-derived nephrogenic epithelial structures forming proximal to the outgrowing UB 13. During this event, the developing nephrogenic structures undergo a series of three primitive transformations from the renal vesicle (RV), to the comma and S-shaped bodies (Figure 3).
Figure 3. Serial morphogenetic events of the nephrogenic blastema. Following condensation of the metanephric mesenchyme (MM), the renal vesicle further morphs into the comma and S-shaped bodies. Podocyte precursors (asterix) line the medial aspect of the proximal cleft of the S-shaped body (Adapted from Saxen et al 13). 1.1.3 Podocyte differentiation Podocyte progenitors form a columnar epithelial cell layer lining the medial aspect of the vascular cleft of the S-shaped body 14, 7 (Figure 3). At this stage of development, podocyte progenitors are attached at their subapical domains by occluding junctions 14,15. Throughout formation of the glomerular capillary loops, podocytes lose their lateral apical attachments and cytoplasmic extensions project from adjacent voluminous cell bodies facilitating encasement of the underlying capillary loops (Figure 4, 5). Primary processes further branch into secondary and tertiary FPs which then interdigitate with FPs from neighbouring cells (Figure 2). FP assembly remains a poorly understood process and whether interdigitations arise entirely as a result of extension of processes from cells that have initially dissociated from each other requires further investigation. Studies investigating cell-cell contact between keratinocytes suggest that cell-cell contacts between adjacent epithelial cells may arise as a result of extension of
5
interdigitated filopodia between two cells 16. Subsequently the filopodia are remodeled such that a simple linear junction morphs into an interdigitating cell-cell junction, while the cell bodies lose their apical attachments resulting in cell-cell contact only at the basolateral aspects of differentiated cells 14. Mechanisms, which govern embryonic podocyte development, are inherently important in understanding the morphological events of podocyte differentiation. In the ensuing subsections, I will review morphological and molecular events that control podocyte development and emphasize the importance of events crucial to formation of the GFB and glomerular SD.
1.1.4 Development of the glomerular filtration barrier Angiogenic factors that include vascular endothelial growth factor (VEGF-A), are secreted by podocyte precursors and lead to the migration and proliferation of endothelial progenitor cells (EPCs) into the vascular cleft of the proximal S-shaped body 7, 22, 23 (Figure 4). Initially, a single capillary loop is formed and during maturation, the capillary loop is further divided into six to eight loops 24. In the developing glomerular capillaries, large undifferentiated EPCs initially occupy the lumina but subsequently some undergo apoptosis while others become flattened and develop intracellular fenestrations which are later required for glomerular permeability and are covered by the proteoglycans-based glycocalyx 7,25. Mesangial cell migration follows formation of the initial capillary loop and subsequently, provides structure to the capillary loops 7,26. During glomerulogenesis, early epithelial precursors express laminin 1 (α1β1γ1) and collagen [α1 (IV) and α2 (IV)] subunits with both podocyte and endothelial precursors having individual basal laminae. During early capillary loop formation, their basal laminae fuse to form a single glomerular basement membrane (GBM) as cells come into close apposition with each other. Subsequently, a shift in both laminin and collagen expression occurs to laminin isoforms that contain the α5 subunit laminin11 (α5β2γ1) and collagens that express the α3 (IV), α4 (IV) and α5 (IV) subunits in the mature glomerulus17-20, 27-29.
6
Figure 4. Development of the glomerular filtration barrier. Following migration and proliferation of endothelial progenitor cells within the proximal end of the S-shaped body (a), a primitive vascular tuft is formed (b) which is surrounded by differentiating podocytes. Fusion of the basal laminae of developing podocytes and differentiating endothelial cells results in the formation of the specialized basal lamina (c) that constitutes the glomerular basement membrane (GBM). (d) In the mature glomerulus, podocytes lie outside the capillary loops, the structure of which is supported by mesangial cells. Electron microscopy demonstrates the podocyte-endothelial interface and fenestrated endothelium adjacent to the podocyte foot processes (Adapted from Dressler et al 30). 1.1.5 Development of the slit diaphragm Originating from subapical junctions interposed between podocyte progenitors, the glomerular SD is located at the basolateral aspect of the podocyte FPs (Figure 6). During podocyte morphogenesis, gradual basal displacement of the interposing cell-cell junction occurs along the lateral margins of apposing cells 15. As podocyte processes interdigitate, the cell-cell junction elongates and undergoes a change in composition and architecture 31. The molecular mechanisms governing this process remain poorly defined and require further investigation.
7
Figure 5. Morphogenetic events of podocyte differentiation. (A) Podocyte progenitors (blue) are simple columnar epithelial cells attached along their lateral margins as they surround the developing capillary (pink). (B) During development, cytoplasmic extensions arise from the basal aspect of podocyte progenitors while cell bodies become dissociated from each other. (C) In the mature glomerulus, foot processes from neighbouring cells interdigitate with each other as they wrap around the underlying capillary loops (Adapted from Quaggin et al 7). The molecular regulation of podocyte differentiation will be discussed in the following section with particular reference to podocyte genes, which when mutated result in defective formation of the GFB with loss of the SD.
1.2. Molecular regulation of podocyte differentiation
1.2.1 Transcriptional regulation of podocyte differentiation Several transcription factors are expressed in early podocyte progenitors and include Wilm’s tumour-1, Wt1 32,33, podocyte-expressed 1 (Pod1) 34, the Lim domain protein, Lmx1b 35, the forkhead domain transcription factor, Mf2 36, kreisler (Mafb) 37 and forkhead box C2 (Foxc2) 38.
8
Figure 6. Electron micrographic features of the podocyte-endothelial interface. Adjacent podocyte foot processes (fp) with intercellular slit diaphragms (arrow) appose the glomerular basement membrane (gbm) and glomerular capillary endothelial cells (endo). Wilm’s Tumour -1 (WT1) Wilm’s tumour-1 (WT1) is a zinc finger transcription factor that encodes a protein that binds to both RNA and DNA 39,40. Within the developing nephron, initial expression is first observed within the renal vesicle (RV) occupying only a subset of cells. As the RV morphs into the comma- and S-shaped bodies, Wt1 expression becomes restricted to podocyte progenitors 32,33. Genetic deletion in mice results in agenesis of both kidneys and gonads, thereby highlighting a crucial role for Wt1 in kidney development 41. Heterozygous Wt1-mice develop severe diffuse glomerulosclerosis in both kidneys by 150 days of age. Electron microscopy demonstrated focal FP fusion and thickening of the GBM at onset of proteinuria in Wt1-heterozygous mice. Other studies also suggest that WT1 may play an important role in human podocyte function. Firstly, WT1 is mutated in 94% of all Denys-Drash syndrome (DDS) patients and the most consistent finding in these patients is the development of a form of glomerulosclerosis called diffuse mesangial sclerosis (DMS) 42,43. Supporting this, is the fact that a common DDS mutation in the WT1 gene can induce the development of glomerulosclerosis when a similar mutation is generated in mice 42-44. Furthermore, WT1 mutations have also been found in 9
patients with nephrotic syndrome and another variant of glomerulosclerosis called focal segmental glomerulosclerosis (FSGS) and in Frasier syndrome, a disorder characterized by steroid-resistant glomerulosclerosis 45. The exact targets and function of WT1 remain undefined mainly due to the diversity of RNA splice forms with the resultant encoded proteins having many different DNA and RNA binding abilities 46,47.
Forkhead box C2, (Foxc2) Forkhead box C2, (Foxc2) belongs to the forkhead-domain family of transcription factors and is expressed during the comma-shaped body stage of nephronogenesis 38. Foxc2 is expressed in podocyte progenitors at the S-shaped body and capillary loop stages of glomerulogenesis. Foxc2 -/- mutant mice exhibit absent podocyte FPs, dilated glomerular capillary loops, absent endothelial fenestrations and clustering of mesangial cells at the glomerular stalk, thereby highlighting a crucial role for Foxc2 in early podocyte differentiation and glomerular development 38. Interestingly, Foxc2 has been shown to physically interact with a Notch transcriptional activation complex to induce transcription of the Notch ligand, Dll4 and its effector Hey2 in murine embryonic endothelial cells 49. Whether Foxc2 regulates glomerular endothelial development in a Notch dependent manner requires further investigation.
Podocyte-expressed 1 (Pod1) and kreisler Podocyte-expressed 1 (Pod1) is a basic helix-loop-helix transcription factor, expressed early in mesenchymal cells surrounding the UB and developing nephric structures and later in podocyte progenitors of the proximal S-shaped body 34,48. Kreisler (Mafb) is expressed in developing podocytes at the capillary loop stage of glomerulogenesis and encodes a transcription factor belonging to the basic domain leucine zipper (bZip) family of transcription factors 37. Genetic mutations in mice for either Pod1 or mafb results in arrest of podocyte differentiation at the single capillary loop stage of glomerulogenesis, thereby highlighting a role for both Pod1 and mafb in podocyte differentiation at the early capillary loop stage of glomerulogenesis 37,48.
10
Within mutant glomeruli, podocytes are columnar, have lost their cell-cell attachments along the lateral margins of apposing cells and lack FPs 37,48, thereby, suggesting that these genes may be important in processes that control the transition from a columnar cell shape to a more complex cellular phenotype.
LMX1B LMX1B is the gene mutated in Nail –Patella syndrome, a disorder characterized by nephropathy, nail hypoplasia and skeletal abnormalities 35,50. LMX1B is a transcription factor encoding a Lim-homeodomain protein. Examination of glomeruli in Lmxb1-/- mice revealed podocytes with defective differentiation as demonstrated by absent FPs and SDs 51. Expression of Lmx1b is restricted to podocytes from the capillary loop stage onwards 52,53. A role for LMX1B in the regulation of FP and GBM proteins has been suggested by the finding of LMX1B binding sites in the promoter regions of genes encoding key SD components, CD2AP and NPHS2 in addition to genes encoding collagens such as COLA3 and COLA4 51,52. These data suggest that LMX1B functions at a later stage of podocyte development than Pod1 and Mafb.
1.2.2 Posttranscriptonal regulation of podocyte differentiation Genetic studies in humans have recently demonstrated that factors other than transcriptional regulation are important for podocyte differentiation and glomerulogenesis. Positional cloning identified phospholipase Cε1 (PLCε1) as a molecule, which regulates podocyte function and development 45. Interestingly, truncating mutations result in a phenotype of DMS while nontruncating mutations result in FSGS 45,54. The resultant phenotypes are reminiscent of the pleiotropic effects of WT1 loss of function in DDS and Frasier syndrome. PLCε1 is localized to the cytoplasm of podocyte cell bodies, intermediate and major FPs 45. During glomerulogenesis, PLCε1 is detected at the S-shaped stage of glomerular development and is highly expressed during the early capillary loop stage as it migrates towards the GBM, where it is primarily located along the basal aspect of
11
developing podocytes 45. Plce1 knockdown in zebrafish demonstrated impaired glomerular permeability with FP effacement and disorganization of SDs
45
. As PLCε1
interacts with IQ motif-containing GTPase-activating protein 1 (IQGAP1), a regulator of cell morphology and adhesion, PLCε1 likely functions to regulate FP formation and morphology.
1.2.3 Integrins and Foot Process Assembly Highly expressed in developing podocytes, integrins are heterodimeric transmembrane receptors (composed of an α and a β subunit) that regulate podocyte function by linking the FP actin cytoskeleton to the extracellular matrix of the GBM 55,56. Integrins α3β1 and α6β1 are the major integrins, which bind laminins while the predominant collagen-binding integrins, are the α1β1 and α2β1 integrins 57. Podocyte specific deletion of the integrin α3 subunit in transgenic mice, develop massive proteinuria within the first week of life and complete FP effacement is evident in newborn mice 58. A supporting role for integrins in FP assembly has been shown in both α3-integrin and α5- laminin knockout mice which fail to assemble FPs 59,60. A similar phenotype is observed in total integrin α3 integrin -/-; α6 integrin -/- double-deleted mice 61
. Most recently, an equally important role for β1 integrin has been demonstrated in
podocyte differentiation 62. Embryos examined at E15 exhibited extensive FP effacement without any changes in the GBM or SD components up until postnatal day 10 62. Podocyte specific deletion of the β1 integrin binding protein, integrin linked kinase (ILK), results in proteinuria, FP effacement, absent SD proteins such as podocin and CD2ap and severe glomerulosclerosis 63,64. Therefore, integrins clearly have an important role in podocyte development and function, of which α3β1 integrin plays a major role 62.
1.2.3 Molecular regulation of slit diaphragm development As podocytes undergo terminal differentiation, assembly of tertiary FPs occurs (Figure 5). Mutations in genes encoding components of the glomerular SD have
12
elucidated the role of SD assembly in regulating FP formation, which ultimately, establishes size-selective glomerular permeability. Figure 7 illustrates the molecular components that constitute the three-dimensional multiprotein complex that forms the glomerular SD.
Figure 7. The molecular components of the glomerular slit diaphragm (Ronco et al 2). The glomerular slit diaphragm is a multiprotein complex comprised of proteins that include nephrin and the family of Neph proteins. P-cadherin and FAT1 and FAT2 are members of the cadherin family of transmembrane receptors that connect the basolateral membranes of the podocyte foot processes. ZO1 is an adaptor protein that links the cytoplasmic domain of Neph1 to the actin cytoskeleton. Nck and CD2AP link the cytoplasmic domain of nephrin to the actin cytoskeleton. Over the past decade, several studies have begun to unravel the molecular composition of the glomerular SD, which includes proteins belonging to the Nephrin/Neph family of immunoglobulin receptors and members of the cadherin family of adhesion molecules such as FAT and P-cadherin 65,66. In addition, proteins are connected to the intracellular actin cytoskeleton by way of key adaptor proteins, which also play a key role in FP assembly and will be outlined in the following section 67,68.
13
Nephrin and Nephs Nephrin, a key component of the mature glomerular SD, is first detected along the lateral margins of podocyte precursors in the late S-shaped body 69. By the early capillary loop stage of glomerulogenesis, nephrin is localized along the basal and lateral margins of differentiating podocytes and later, in maturing glomeruli, nephrin is localized to the basolateral margins of differentiated podocytes. Complete loss of the SD is observed in mice deficient for Nphs1, the gene encoding nephrin 70. While formation of podocyte FPs is observed in Nphs1-/- mice, they exhibit a broader morphology resulting in a narrower filtration slit with an undefined intercellular junction than that observed in their wild-type littermates 70. Nephrin is a type I transmembrane protein of the immunoglobulin superfamily of receptors (185-200kDa) comprised of a short intracellular domain, an extracellular domain with eight distal IgG-like motifs (35nm length) and one fibronectin type-III-like motif
71
. N-linked glycosylation of the extracellular domain of human nephrin has been
shown to be crucial for proper folding of nephrin and thereafter, localization to the plasma membrane 72. At the plasma membrane, the extracellular domain of nephrin has been shown to interact with itself in a homophilic fashion
73
. In addition, nephrin
interacts with Neph1, a structurally related protein (90-110kDa), in a heterophilic fashion that involves a trans-interaction across the plane of SD, connecting adjacent FPs 74. Both proteins also interact at the cytoplasmic side and participate in regulating FP actin cytoskeletal dynamics by way of interactions with their cytoplasmic domains and key adaptor actin-associated proteins that include ZO-1, synaptopodin, α-actinin 4, CD2ap, CASK, IQGAP1, β2-arrestin, Nck and Grb2 67,68,76-81. Nephs are structurally related to nephrin and are comprised of five extracellular domains and are similarly localized to the SD 74,82. Other members of the Neph family include Neph 2 (95-125kDa) and Neph 3 83,84. In vitro data has shown that nephrin can form heterodimers with Neph 1 or Neph2 but that Neph1 and Neph2 themselves do not interact 85. Similar to nephrin, glycosylation of Neph1 is required for correct interactions with nephrin 85. Newborn mice carrying a gene trap mutation in Neph1 develop heavy
14
proteinuria and FP effacement 83. Kidneys from 3 week old mice demonstrate diffuse mesangial hypercellularity, mesangial matrix expansion, in addition to enlargement of Bowman’s space and cystic tubules filled with protein 83. Nphs1-/- mice share a similar phenotype 70,86, thereby highlighting similar roles for nephrin and nephs in the regulation of SD and FP morphology. During glomerulogenesis, tyrosine phosphorylation of the cytoplasmic domain of nephrin occurs transiently during FP assembly 87. Upon engagement of the extracellular domain of nephrin, its cytoplasmic domain is phosphorylated at multiple tyrosine residues by the Src family kinase, Fyn 87-89. Among these residues, interactions between nephrin and Nck are mediated by tyrosine phosphorylation at residues Y1191, Y1208 and Y1232 67,87. Induction of nephrin-mediated actin polymerization requires recruitment of Nck to phosphorylated nephrin at these sites in vitro 87. Podocyte specific deletion of Nck1/2 results in developmental defects in FP formation in vivo confirming a role for nephrin-Nck mediated interactions during FP development 67. Other than a developmental role, reduced tyrosine phosphorylation at these sites has also been shown to precede cytoskeletal rearrangements in the podocyte FP, which predate the onset of proteinuria in puromycin aminonucleoside nephrosis (PAN), a murine model of podocyte injury 68,87. As Neph1-/- mutant mice have similar developmental defects to Nphs1-/- mutant mice, a recent study examined the role of tyrosine phosphorylation of the cytoplasmic domain of Neph1 and regulation of actin cytoskeletal dynamics 68. Like nephrin, engagement of the extracellular domain of Neph1 results in recruitment of an actinassociated adaptor protein, Grb2 and results in the induction of actin polymerization, thereby confirming a regulatory role for the Nephrin-Neph complex in mediating junctional and cytoskeletal dynamics during podocyte FP formation and response to injury 68.
FAT1 and FAT2 The FAT proteins are large transmembrane proteins that belong to the protocadherin superfamily and have been shown to colocalize at the SD with nephrin
15
82,90
. FAT proteins also play a role in the regulation of intracellular actin cytoskeletal
dynamics by binding to Ena/VASP 91-93. A role for FAT1 in FP assembly during podocyte differentiation has been suggested by findings that Fat1-/- mice fail to develop FPs, lack SDs and have proteinuria as well as forebrain and ocular defects 94. Fat 2 -/mice develop proteinuria demonstrating a potential role for this member of the FAT family in the regulation of glomerular permeability. In addition, a role for FAT4 in mammalian planar cell polarity (PCP) signaling has recently been confirmed 95. Fat 4-/mice develop cystic kidneys but whether FAT4 is expressed in the podocyte and regulates PCP signaling requires further exploration. Besides FAT proteins, other cadherin family members have been localized to the SD and include P-cadherin 66. However, no obvious glomerular defect has been observed in P-cadherin deficient mice 96
which suggests that P-cadherin may not be as important as FAT1 and FAT2 for the
permselective function of the SD.
Podocin Podocin is a member of the stomatin family of proteins and is an integral membrane protein, which has been localized to the basolateral surfaces of podocytes 97. Located at the cytoplasmic side of the FP, podocin has been shown to play a key role in SD assembly whereby it interacts with nephrin at the basolateral membrane of the podocyte FP and as a result, targets nephrin to lipid rafts within the plasma membrane where signaling events are subsequently initiated 98. R138Q is a missense mutant variant of podocin (encoded by NPHS2) that has been associated in humans with steroid-resistant nephrotic syndrome. A Nphs2 R140 transgenic mouse, which corresponds to the human R138Q variant, has recently been characterized 99. Newborn homozygous Nphs2 R140; Nphs2 R140 mice develop progressive proteinuria with no initial glomerular abnormalities on light microscopy. While podocytes were initially normal, podocytes subsequently demonstrate focal FP effacement by postnatal day 10. In areas where FPs were preserved, a lack of SDs were noted in contrast to control mice. Therefore, podocin mutations do not appear to interfere with initial FP assembly but may interfere with SD assembly, which
16
can have subsequent deleterious consequences on FP function and thereafter glomerular permselectivity 98.
CD2AP CD2AP is an SH3 domain-containing protein that has been shown to be required for SD assembly and regulation of FP cytoskeletal dynamics 100. CD2AP interacts with nephrin and podocin at the cytoplasmic side of the basolateral FP 101. Newborn Cd2ap -/mice have proteinuria and focal FP effacement within the first week of life 102. Normal FPs were observed in non-effaced areas, thereby suggesting that CD2AP 103-108, does not regulate FP assembly during development. However, given the onset of effacement within the first week of life CD2AP is more likely to play a regulatory role in maintenance of FP cytoarchitecture through its interactions with key SD components such as nephrin and podocin 102,109-112.
1.2.4 The glomerular slit diaphragm in acquired glomerulopathies Studies in both human and rodent models of experimental glomerulopathies have demonstrated altered expression of key SD components 113-120. Amongst the proteinuric disorders themselves, the pattern of distribution reported for key SD proteins such as nephrin has been variable. In MCN, for example, some studies report a reduction or deviation from the normal linear staining pattern for nephrin to a granular distribution pattern while others studies refute a difference 120,121. Rodent models of experimental glomerulopathies have attempted to address the temporal expression of SD molecules over the time course of proteinuria both at the level of mRNA and protein in relation to changes in FP morphology 122. Mice treated with puromycin aminonucleoside (PAN), develop proteinuria 3 days following injection and subsequently recover by day 28 122. In PAN nephrosis, nephrin mRNA expression appears to recover before nephrin protein levels 120. As proteinuria resolves by day 15, protein levels recover as demonstrated by an increase in gold particles and distribution patterns return to patterns observed in control mice 120,122. A compensatory increase in mRNA synthesis, therefore appears to occur
17
following a reduction in protein in the PAN model of nephrosis. Expression of SD proteins may therefore be altered in distribution, quantity and synthesis or as a result of altered cytoskeletal dynamics in acquired glomerulopathies. The reason for altered expression are likely to be multifactorial and dependent on the cellular response to injury in addition to circulating external factors 106.
3. Notch Signaling and Glomerulogenesis
1.3.1 The Notch signaling pathway Genetic and biochemical studies in mice have implicated the Notch signaling pathway in glomerulogenesis as well as establishing the proximo-distal axis of the nephron (including podocytes). The Notch pathway is a highly conserved and ubiquitous signaling system that regulates a diverse array of processes which include cell fate specification, boundary formation, progenitor cell maintenance and various cellular processes that are dependent on both cellular context and developmental stage 123. Notch proteins comprise a family of single transmembrane receptors that mediate short range communication between cells. Ligand binding to receptor triggers a chain of events that leads to the shedding of the extracellular domain (mediated by a metalloprotease) 124 and subsequent cleavage of the transmembrane domain by the enzyme, gamma secretase (Figure 8) 125. Following cleavage, the intracellular domain of Notch (NICD) undergoes nuclear translocation where it associates with the DNA-binding protein, RBPJ-κ and subsequently promotes transcription of target genes (e.g. Hairy Enhancer of Split (Hes)) that control tissue specific differentiation genes 126,127. In the absence of NICD, RBPJ-κ complexes with ubiquitous co-repressor proteins such as SMRT (silencing mediator of retinoic and thyroid hormone receptors) or Mint to repress transcription 128. Limited amounts of NICD are generated upon ligand binding and compete with more abundant repressor proteins to form a NICD/RBPJ-κ complex. NICD/RBPJ-κ is recognized by the Mastermind/Lag3-protein complex and recruits the ARC-L/MED complex, which phosphorylates the NICD/PEST domain triggering NICD ubiquitination by Fbw/sel 10
18
ubiquitin ligase and proteosome-mediated degradation 129. This pathway is known as the canonical pathway.
Figure 8. The Notch signaling pathway. Binding of the Notch ligand (red) on one cell to the extracellular domain (purple) of the Notch receptor on another cell results in gamma-secretase (blue) mediated cleavage of the transmembrane domain of the Notch receptor. This proteolytic processing mediates release of the Notch intracellular domain (ICD) (turquoise) which enters the nucleus and interacts with the DNA-binding protein (RBPJ-κ) (orange). The co-activator Mastermind (MAM) (blue) is then recruited to the RBPJ-κ complex and co-repressors are released. In mammals, at least four mammalian Notch homologues (Notch 1- 4) and at least five ligands [jagged 1(Jag1), Jag2, delta-like1 (Dll1), Dll3 and Dll4] mediate these signaling events. All Notch proteins share a similar domain architecture; the extracellular domain includes multiple epidermal growth factor-like repeats (EGF) (between 29-36 EGF repeats) 130 and three LNR (Lin-Notch repeats) 131. EGF 11 and 12 are essential for ligand binding 123. Specific Notch ligand-receptor interactions have recently been shown to be mediated by a range of proteolytic processing and glycosylation events.
19
The intracellular domain (ICD) contains the RAM (RBPJ-κ-associated molecule) domain, followed by the highly conserved 7 ankyrin repeats (ANK). Lying C-terminal to the ANK domain is the nuclear localization sequences and a PEST domain regulating protein stability. The RAM domain is the main mediator of the NICD/RBPJ-κ interaction, while the main role of the ANK domain lies in MAM recruitment 132,133. The intracellular domains of Notch 1, Notch 2 and recently, Notch 3 have been shown to contain a transactivation domain (TAD) located C-terminal to the ANK repeats 134,135. To date, a transactivation domain has not been reported for Notch 4. Both the ANK and TAD domains are thought to cooperate with Notch proteins to mediate their specific transactivation properties 136.
1.3.2 Expression of Notch pathway components during glomerulogenesis Notch pathway components are dynamically expressed throughout nephronogenesis. Notch 1, Notch 2, Dll1 and Jag 1 mRNA are detected in the RV and its derivative; the expression domain of Notch1 partially overlaps with Notch2 in the Sshaped body 137-139. Notch signalling components particularly, Notch 2, persist in the proximal domain of the late S-shaped body (Table 1). Notch1, Notch2 and downstream transcriptional targets Hes1 and Hey1 are expressed in podocyte progenitors in the Sshaped body but are progressively downregulated during terminal podocyte differentiation such that in mature podocytes, Notch pathway components are no longer expressed 140 (Figure 9). Notch 3 is expressed strongly in the smooth muscle cells of major kidney blood vessels and in the glomerular tuft while Notch 4 transcripts are found in the vascular endothelia. Dll1 expression is restricted to a short stretch of the middle segment of the S-shaped body. Jag 1 overlaps with that of Dll1 but extends to the more proximal part of the S-shaped body, which forms the glomerulus and in the maturing glomerulus, is expressed in the inner region consisting of differentiating mesangial and endothelial cells. Jag1 is expressed in cells adjacent to Notch 2 expressing cells in the developing glomerulus. Jag 2 is expressed throughout the kidney at E16.5 but at much lower levels than Dll1 and Jag1 140. In the mature glomerulus, Dll4 is restricted to the glomerular endothelium 141.
20
Figure 9. Expression of Notch pathway components during podocyte differentiation. Notch 1, Notch 2, Hes1 and Hey1 are expressed in podocyte progenitors at the S-shaped body stage of glomerulogenesis. Downregulation of Notch1 and Hes1 occurs in differentiating podocytes at the capillary loop stage of glomerulogenesis, while Notch2 and Hey1 continue to be expressed. Notch pathway components are no longer expressed in terminally differentiated podocytes.
21
1.3.3 The role of Notch in podocyte cell fate specification The dynamic expression patterns of Notch pathway components during glomerulogenesis suggest that Notch signaling regulates many developmental processes that coordinate the formation and function of the mature glomerulus. Podocyte cell fate specification and formation of the glomerular vascular tuft have been shown to be two Notch regulated processes during glomerulogenesis (Figure 9). Table 1. Expression of Notch pathway components during glomerulogenesis.
Notch
S-shaped
Podocyte
Endothelial
Mesangial
Component
Body
Progenitors
Progenitors
Progenitors
Notch 1
Middle *
+
+
-
Notch 2
Proximal
+
-
-
Notch 3
Proximal
-
+
+
Notch 4
-
-
+
-
Dll1
Middle
-
+
-
Dll3
-
-
-
-
Dll4
-
-
+
-
Jag1
Middle
-
+
+
Jag 2
-
-
+
-
Hes 1
Middle
+
+
-
Hes 5
Middle
-
-
-
Hes 6
-
-
-
-
Hes 7
-
-
-
-
Hey 1
Middle
+
+
-
Hey2
-
-
-
-
Hey L
Middle
-
+
-
Lfng
Middle
-
-
-
* segment of S-shaped body
22
Pharmacologic inhibition of gamma secretase results in nephrons deficient in podocytes, proximal tubules and loops of Henle 142. Supporting genetic studies in mice have confirmed a role for endogenous Notch signaling in podocyte cell fate specification. PSEN, the gene encoding the presenilin enzyme involved in Notch processing, when deleted in murine kidneys, results in loss of proximal tubules (marked by LTL) and podocytes (marked by WT1) 143. Similarly, conditional genetic deletion of Notch 2 in the metanephric mesenchyme (MM) in transgenic mice results in neonatal death on day 2 of life 9. Histological examination of mutant kidneys revealed a paucity of S-shaped bodies, proximal tubules and glomeruli, thereby confirming a role for Notch 2 in determination of proximal nephron fates. In contrast, determination of distal fates was independent of Notch 2 as demonstrated by the presence of E-cadherin and cytokeratin 8 -positive structures representing distal tubules and collecting ducts 9. Conditional homozygous deletion of RBPJ-κ in the MM also results in a reduction of both proximal tubules and Wt1- expressing podocytes, confirming a role for canonical Notch during proximal nephron formation 9.
1.3.4 The role of Notch in glomerular vascular tuft formation Following proximal fate determination, genetic studies in mice have confirmed a role for Notch signaling during formation of the glomerular tuft. Homozygous mice carrying a deletion in EGF 14, the ligand binding domain of the Notch 2 receptor (Notch 2del1), die within 24 hours of birth due to defective kidney development 140. At birth, hypoplastic kidneys with cortical vascular lesions were noted. Kidneys were examined during nephrogenesis, at E13.5, and Notch 2del1/Notch 2del1 kidneys were of normal size. Both control and mutant kidneys exhibited UB growth and branching with condensation of MM and formation of epithelial vesicles 140. Further analysis at E16.5, however, demonstrated smaller kidneys compared to wild-type and heterozygous littermate controls. UB branching and RVs were morphologically normal suggesting that Notch 2 signaling is not required for condensation and epithelialization of early nephrogenic precursors. Morphologically abnormal glomeruli were noted with arrest at the capillary loop stage. An absent capillary tuft consisting of a disorganized clump of cells was
23
observed. Other glomeruli had an-aneursymal like structure that filled the region of Bowman’s capsule with red blood cells. Fewer PECAM1 positive (platelet-endothelial cell adhesion molecule 1) endothelial cells were present in the abnormal mutant glomeruli compared to controls. Migration of some endothelial cells into the cleft of comma- and S-shaped bodies was observed in developing nephrogenic structures of Notch 2del1/Notch 2del1 kidneys 140. In addition to abnormal endothelial cell number, mesangial cell differentiation was also affected in Notch 2del1/ Notch 2del1 mutant mice as was demonstrated by a lack of desmin-positive and Pdgfrb (platelet-derived growth factor receptor b) expression in mutant glomeruli. In the absence of Notch 2, podocytes expressing Wt1 and Vegfa were observed but were clumped together in the centre of mutant glomeruli and did not form the cup shaped epithelial structure observed in controls 140. Overall a quantitative difference in the number of glomerular structures was observed in Notch 2del1; Notch 2del1 mutant mice. Similarly, defective glomerular development has also been noted in combined jagged2/jagged1b zebrafish morphants suggesting that Notch signaling has a conserved role in glomerular vascular tuft formation 144. Notch signaling is likely required at multiple stages of glomerular differentiation such as cellular differentiation of the RVs, endothelial migration, mesangial differentiation and subsequent vascularization of the developing glomerulus.
Summary Podocytes are highly specialized renal epithelial cells, which regulate glomerular ultrafiltration and prevent urinary leakage of plasma proteins by regulation of tertiary foot process morphology and slit diaphragm intergrity. Terminal podocyte differentiation is associated with downregulation of Notch pathway components, which coincides with tertiary foot process assembly and slit diaphragm formation. Maintenance of differentiation is critical to podocyte function. The objective of my research is to determine the role of constitutive Notch activation in developing podocytes on podocyte function, podocyte differentiation and formation of the glomerular slit diaphragm.
24
CHAPTER II:
ECTOPIC NOTCH ACTIVATION IN DEVELOPING PODOCYTES CAUSES GLOMERULOSCLEROSIS
A version of this chapter has been published in the Journal of the American Society of Nephrology 2008 19(6):1139-57
Figure 13. Megan Wu and Tuncer Onay of the Piscione laboratory contributed to work demonstrated in this figure. Figure 16. Megan Wu contributed to work demonstrated in this figure.
25
INTRODUCTION Genetic evidence in mice suggests that podocyte cell fate determination is regulated by Notch signaling during nephrogenesis 9,142,143. Notch signaling controls cell differentiation in multiple developing organ systems 145. A key step in Notch receptor activation is proteolytic cleavage of its intracellular domain (NOTCH-IC) 146, resulting in NOTCH-IC nuclear translocation and complex formation with the transcriptional repressor, Recombination Binding Protein J-κ (Rbpsuh – Mouse Genome Informatics; also known as RBPJ-κ, CSL, CBF1) 127. NOTCH-IC binding to RBPJ-κ recruits transcriptional co-activators, which induce expression of downstream targets, including Hairy/Enhancer of Split (Hes) and Hes-related (Hey) genes 126,127,147 that function as Notch effectors by negatively regulating tissue-specific differentiation 148,149. During proximal nephron specification, it is unclear whether Notch is involved in binary decisions influencing podocyte versus proximal tubular cell fate choices, or subsequently functions in podocyte terminal differentiation during glomerular development. The demonstration that the γ-secretase inhibitor, DAPT, does not prevent podocyte WT1 expression in advanced stage cultured metanephroi argues against a subsequent role for cleavage-dependent Notch activation in podocyte terminal differentiation 142. Consistent with this concept are reports of Notch pathway gene expression patterns in developing podocytes, which reveal a progressive decline in Notch1, Notch2, Hes1, and Hey1 mRNA levels in immature podocytes, and absent expression in podocytes at more advanced stages of glomerulogenesis 137-139. Collectively, these data suggest that high levels of Notch signaling in podocyte progenitors may be essential for initiating differentiation, yet lower or absent levels of Notch signaling in developing podocytes may be necessary for terminal differentiation. Conversely, persistent activation of Notch in developing podocytes may oppose terminal differentiation. To investigate the effect of Notch activation in developing podocytes, I employed a tissue-specific, CRE-loxP-mediated approach to ectopically express NOTCH-IC in embryonic podocytes, and determined the effect of constitutive Notch signaling on podocyte differentiation and post-natal function. My results indicate that ectopic NOTCH-IC expression causes loss of GFB permselectivity, podocyte differentiation 26
defects, and glomerulosclerosis within the first few weeks of life. My data support the concept that down-regulated Notch activity is crucial for podocyte differentiation, and suggest a novel pathogenic role for inappropriate podocyte Notch activation in glomerulosclerosis. MATERIALS AND METHODS
Breeding and genotyping of NOTCH-IC transgenic mouse strains. Neph-CRE/+ mice (kindly provided by Professor SE Quaggin, Samuel Lunenfeld Research Institute, Toronto). These mice contained a 1.25kb fragment of the 5’ flanking and proximal region of the murine Nphs1 promoter (4145 to 8270bp of accession number AF296764) ligated to a Cre recombinase cassette injected into one cell-stage murine embryos 150. IC-Notch1/+ mice, generated by Dr Corinne Lobe (Sunnybrook Hospital, Toronto) harboured a transgene containing the Myc-tagged murine Notch1 intracellular domain lying downstream from a loxP-flanked β-geo and a pCAGGs promoter 151. NephCRE/+ and IC-Notch1/+ mice were crossed resulting in Neph-CRE/+; IC-Notch1/+ transgenic offspring which exhibited constitutive Notch-IC activation specifically in podocytes. All mouse strains were maintained on mixed backgrounds. CRE genotyping was performed by PCR 150. IC-Notch1 genotyping was determined by tail clip lacZ assay 151
. Experiments complied with ethical standards of The Hospital for Sick Children
Research Institute Animal Care Committee.
Glomerular isolation Mouse glomeruli were isolated by perfusion of mouse kidneys with magnetic iron oxide suspended in Hank’s balanced salt solution (HBSS) 152. Kidneys were minced and then digested with collagenase A, DNAse I in HBSS for 30 minutes at 37 oC. The tissue homogenate was then passed through a 100μm nylon sieve and was washed with HBSS. Suspended glomeruli were then isolated with the use of a permanent magnet, which allows tubules and other contaminants to be separated by decantation.
RT-PCR analysis
27
Total RNA was extracted and purified using the RNeasy Micro Kit (Qiagen). cDNA was generated using SuperScript II Reverse Transcriptase (Invitrogen). PCR was performed as described using specific primers. Wt1 (Forward TGCGGCGTGTATCTGGAG, Reverse TTGAAAGGTGAGTGGGAGGAA), Nphs1 (Forward CCCAACACTGGAAGAGGTGT, Reverse CTGGTCGTAGATTCCCCTTG), Nphs2 (Forward TGAGGATGGCGGCTGAGAT , Reverse GGTTTGGAGGAACTTGGGT), 18S (Forward AAGGGCACCACCAGGAG , Reverse GGACATCTAAGGGCATCACAG), Gapdh (Forward CTCATGACCACAGTCCATGC Reverse CACATTGGGGGTAGGAACAC) 153. Semi-quantitative analysis was performed by band densitometry using Image J 154.
Urine protein analysis Mouse urine was obtained by spontaneous expression. For dipstick analysis, 1020 µL of collected urine was spotted onto Labstix Urinalysis Reagent Strips (Bayer) and evaluated by standard colorimetric assay. Urine protein concentration was measured by Bradford microassay (BioRad). Urine creatinine concentration was determined using the Parameter kit (R&D Systems).
Renal histology For routine histology, mouse kidneys were fixed in 10% formalin and processed for paraffin sectioning. Representative sections were stained with periodic acid-Schiff (PAS), imaged by brightfield microscopy, and photographed using an Axioskop microscope (Carl Zeiss).
Transmission electron microscopy (TEM) For TEM, samples were dehydrated in a graded ethanol series followed by propylene oxide, embedded in Spurr-Quetol resin. Sections 100nm thick were cut on an RMC MT6000 ultramicrotome and stained with uranyl acetate and lead citrate, and observed under a transmission electron microscope (CM100, FEI) at 75kV. Transmission electron microscopic images were taken from two sections 100μm apart and three mature glomeruli per level were imaged. Fifty photomicrographs were taken at random locations
28
in the peripheral capillary loops from each glomerulus at a final magnification of 46, 000X. The number of SDs in each imaged field were counted along the apposing glomerular basement membrane (GBM). The GBM length on each micrograph was measured with an image analyzer (Soft Imaging Systems) and analyzed using EXCEL software.
In Situ Hybridization Non-radioactive mRNA in situ hybridization were performed as previously described 138. Plasmids for RNA probe synthesis were kindly provided by J. Kreidberg (mWt1), S. Quaggin (mNphs1), C. Antignac (mNphs2), and G. Dressler (mPax2).
Immunohistochemistry and Immunofluorescence multi-labelling Primary antibodies and lectins were mouse monoclonal anti-MYC (1:1000; Invitrogen), goat polyclonal anti-NOTCH1 (C-20) (1:100; Santa Cruz), rabbit polyclonal anti-WT1 (1:50; Santa Cruz), rat monoclonal anti-CD31 (1:100; Santa Cruz), guinea pig polyclonal anti-nephrin (1:100; Fitzgerald), sheep polyclonal anti-Ki67 (1:100; Chemicon), rabbit polyclonal anti-PAX2 (1:50; Zymed), and FITC-conjugated Lotus Tetragonolobus Purpureas (LTL) lectin (Vector Laboratories, 1:100). Immunoperoxidase staining was performed on formalin-fixed, paraffin-embedded tissue sections. Microwave antigen retrieval was carried out in citrate buffer in four 5-minute cycles at medium-hi setting (Panasonic NN-S758WC, 950W max. output) followed by a 20 minute cooling period at room temperature. Unless otherwise specified, blocking was performed in Universal Blocking Reagent (DAKO). For monoclonal incubations, sections were blocked in 5% rabbit serum for one hour (rat monoclonals), or with the M.O.M. blocking kit (mouse monoclonals; Vector Labs). Primary antibody incubations were carried out at 4°C overnight. Biotin-conjugated secondary antibodies were diluted 1:1000 in blocking reagent, and incubated at room temperature. Immunoperoxidase staining was developed using the Vectastain ABC kit (Vector Laboratories). Dual and triple immunofluorescence antibody staining was performed on PFAfixed frozen sections treated with Proteinase K (Roche) 20 µg/mL (5 minutes, 37°C),
29
washed in 0.1% Triton X-100, and blocked with Universal or M.O.M blocking reagent for polyclonal and mouse monoclonal primary antibodies, respectively. Primary antibody incubations were carried out simultaneously. AlexaFluor 488-, AlexaFluor 594-, or Cy5conjugated secondary antibodies (1:1000; Invitrogen) were used for multiimmunofluorescence labelling. Sections were counterstained with DAPI, and imaged by fluorescence microscopy using a Zeiss Axioskop microscope with an EXFO X-Cite120 120W mercury vapor lamp (Photonics Solutions). Digital photographs were obtained as above, and merged images were obtained using Photoshop v6.0.
Statistical analyses Statistical analyses were performed by Student’s t-test using Statview (v5.0).
RESULTS 1. Podocyte-specific expression of MYCNOTCH-IC in transgenic mice. In vivo effects of constitutive Notch signaling in murine podocytes were determined using a CRE-loxP conditional transgenic approach to induce podocytespecific expression of MYC-tagged Notch intracellular domain (termed MYC-NOTCH-IC) (Figure 10). CRE-dependent MYC-NOTCH-IC protein expression was demonstrated by Western Blot in extracts from Nephrin-CRE/+;NOTCH-IC/+ (hereafter, termed CRE(+); NOTCH-IC) isolated mouse glomeruli but not from CRE-negative, NOTCH-IC/+ transgenic (hereafter, termed CRE(-);NOTCH-IC) or wild-type mouse littermates (data not shown). Immunohistochemistry using anti-MYC or anti-NOTCH1 antibodies to detect MYC-NOTCH-IC protein expression in kidney tissue sections showed multiple podocytes staining positively with anti-MYC (Figure 10E, black arrow) or anti-NOTCH1 antibodies (Figure 10H, black arrow). The patterns of anti-MYC or anti-NOTCH1 immunostaining were nuclear, suggesting nuclear localization of MYC-NOTCH-IC protein. In contrast, glomeruli of CRE(-);NOTCH-IC and wild type littermates lacked anti-MYC (Figure 10C and 10D, respectively) or anti-NOTCH1 (Figure 10F and 10G, respectively) staining in podocyte nuclei. Co-incubation of sections with anti-MYC and anti-WT1 antibodies revealed co-localization of MYC-NOTCH-IC and WT1 in many cells, thereby
30
identifying (MYC, WT1)-double-positive cells as MYC-NOTCH-IC-expressing podocytes (Figure 14F and 14G). We quantified MYC-NOTCH-IC protein expression in podocytes by counting the number of (MYC, WT1)-double-positive and total WT1-positive cells per glomerulus, and observed a progressive increase in the fraction of podocytes expressing MYC-
NOTCH-IC protein over time. At post-natal day 7 (P7), 21±5% of podocytes were
(MYC, WT1)-double-positive (n=22 glomeruli). By P21, this fraction increased to 35± 6% (n=22 glomeruli), and was 51±6% at P28 (n=16 glomeruli). Occasionally, weak anti-NOTCH1 immunostaining was detected within the glomerular capillary tuft or in the cytoplasm of tubular elements (Figure 10G), raising the possibility that CRE-independent (i.e. leaky) MYC-NOTCH-IC expression in non-podocyte lineages might account for glomerular and tubular anti-NOTCH1 immunostaining in transgenic tissues. To examine this possibility, I performed double-labelling experiments on mouse kidney tissue sections using anti-MYC antibody and either anti-CD31 or Lotus Tetragonolobus lectin (LTL) to simultaneously detect cells MYC-NOTCH-IC-expressing cells in glomerular capillary endothelium or proximal tubular epithelium, respectively. MYC-positive cells were detected in glomeruli of CRE(+);NOTCH-IC transgenic mice contiguous but not overlapping with staining for anti-CD31 antibody (Figure 10K). In contrast, MYC-positive cells were not detected in glomeruli of wild type and CRE(-); NOTCH-IC mice also stained with anti-CD31 antibody (Figure 10I and 10J, respectively), suggesting that glomerular capillary endothelium did not exhibit leaky MYCNOTCH-IC expression. Some MYC-positive cells were occasionally detected within capillary lumina outlined by anti-CD31 staining (Figure 10I, white arrowhead); however, since these MYC-positive cells lacked DAPI staining, I suspected that they were anucleate erythrocytes with background staining. I confirmed this by incubating tissues with anti-mouse secondary antibody alone, and demonstrated frequent labelling of DAPInegative cells with anti-mouse secondary within glomerular capillary lumina (not shown). Likewise, I did not detect nuclear anti-MYC staining in tubular structures identified by LTL staining in wild type, CRE(-);NOTCH-IC, and CRE(+);NOTCH-IC mice (Figure 10L, 10M, and 10N, respectively), but observed background staining within tubular cytoplasm (not shown). Thus, I concluded that MYC-NOTCH-IC expression was restricted to the podocyte lineage in our transgenic system, which afforded the
31
opportunity to determine effects of constitutive NOTCH activation exclusively in podocytes.
32
Figure 10. Podocyte-specific expression of MYCNOTCH-IC in glomeruli of newborn CRE(+);NOTCH-IC mice. (A) Breeding scheme to generate transgenic mice that express MYCNOTCH-IC in podocytes. Neph-CRE/+ mice express CRE in podocytes under control of Nphs1 promoter elements. The modular IC-Notch1 transgene is represented pictorially. pCAGGS, CMV enhancer/chicken β-actin hybrid promoter; X, loxP; β-geo, β-galactosidase/neomycin/polyA fusion cassette; MYC, 6X human MYC epitope tag; IRES, internal ribosomal entry site; hPLAP, human placental alkaline phosphatase. (B) Analysis of MYCNOTCH-IC protein expression in lysates of isolated glomeruli as separated by SDS-PAGE. Upper and lower left panels show Western blot results following incubation with anti-MYC and anti-β-actin antibodies, respectively. Black arrowhead in upper panel denotes band position corresponding to MYCNOTCH-IC with approximate molecular weight of 111 kDa. Black arrowhead in lower panel denotes actin band. Upper right panel shows protein ladder (MW Std.) and corresponding protein molecular weights. (C-H) Detection of conditional MYCNOTCH-IC protein in podocytes as revealed by immunohistochemistry using anti-MYC (C-E), and anti-NOTCH1 (F-H) antibodies. Non-serial, representative images are shown of newborn mouse kidney tissue sections from wild type (C, F), CRE(-);NOTCH-IC (D, G) and CRE(+);NOTCH-IC (E, H) mice (original magnification, 400X). Black arrows denote anti-MYC or anti-NOTCH –stained cells. g, glomeruli. Sections were counterstained with hematoxylin. (I-N) Dual immunofluorescence labelling of mouse kidney tissue sections with either anti-MYC (green) and anti-CD31 (red) antibodies (I-K) or anti-MYC (green) and LTL (red) (L-N). Shown are representative images of glomeruli (original magnification, 1000X) from wild type (I, L), CRE(-);NOTCH-IC (J, M), and CRE(+);NOTCH-IC (K, N) mice. For (L-N), anti-MYC staining was detected with AlexaFluor594 secondary antibody, and LTL staining performed with FITC-LTL. Sections were counterstained with DAPI. White arrows, anti-MYC labelled podocytes. White arrowheads, background anti-MYC/antimouse IgG immunoreactivity.
2. Early-onset proteinuria in CRE(+);NOTCH-IC transgenic mice. Effects of MYC-NOTCH-IC expression on post-natal podocyte function as they related to GFB permselectivity were determined by total urine protein:creatinine ratio. Urine protein:creatinine values for CRE(+);NOTCH-IC mice were higher compared with wild type or CRE(-);NOTCH-IC mice at P7, and reached statistical significance at P14 [CRE(+);NOTCH-IC urine protein:creatinine (mg/mmol), P vs. CRE(-);NOTCH-IC, P vs. wild type -- P14: 217±51, P=0.01, P=0.0005; P42: 3359±1567, P=0.05, P=0.07]. For CRE(-);NOTCH-IC mice, random urine protein:creatinine ratios were not statistically significantly different from wild type values, except at P14 [protein:creatinine (mg/mmol) – wild type: 22±3 vs. CRE(-); NOTCH-IC: 67±13, P=0.002]. Since corresponding values at P42 were not statistically different between these two groups
33
(Table 2), I did not consider my finding of proteinuria in CRE(-);NOTCH-IC mice at P14 to indicate a permanent effect on glomerular permselectivity. I also did not observe proteinuria in CRE-expressing mice that did not inherit the IC-Notch1 transgene (data not shown), suggesting that sustained effects on glomerular permselectivity were dependent on combined inheritance of the IC-Notch1 and Nephrin-CRE alleles. Thus, demonstration of proteinuria in CRE(+);NOTCH-IC mice suggested that constitutive expression of MYCNOTCH-IC in podocytes impaired GFB permselectivity. Table 2. Urine protein quantification by protein:creatinine ratio over time in wildtype, CRE(-);NOTCH-IC, and CRE(+);NOTCH-IC mice.
Post-Natal Age:
P0
P7
P14
P42
97 ± 28 (n=3)
34 ± 3 (n=5)
22 ± 3 (n=7)
61 ± 13 (n=4)
54 ± 28 (n=2)
18 ± 3* (n=3)
67 ± 13* (n=5)
108 ± 17 (n=5)
31 ± 17 (n=4)
138 ± 64 (n=4)
217 ± 51*† (n=4)
3359 ± 1567 (n=4)
Wild type Prot:Creat (mg/mmol)
CRE(-);NOTCH-IC Prot:Creat (mg/mmol)
CRE(+);NOTCH-IC Prot:Creat (mg/mmol)
*
P
