CME Slides Forum - A. Woolf

ORCHESTRATING KIDNEY DEVELOPMENT

Adrian Woolf, London, United Kingdom

Chair: Kostas Siamopoulos, Ioannina, Greece

Roser Torra, Barcelona, Spain

woolf

Prof A. Woolf
Nephro-Urology Unit
UCL Institute of Child Health
London, United Kingdom

Thank you very much for asking me to give this talk. Well, the title I was given is, ‘Orchestrating kidney development’ but I’m a clinician, so I’ve subtitled this, ‘Messages from the Clinic’. Really in the next 20 minutes I just wanted to give you a flavour of how gene discovery in kidney and ureter development is beginning to impact into the clinic in terms of genetic diagnoses.

I’m a paediatric nephrologist and in the United Kingdom there are about 800 children with end-stage renal failure and the great majority of them were born with abnormal kidneys, either dysplastic kidneys, hypoplastic kidneys or sometimes even absent kidneys. Of course, there are many more children in the U.K. with these diagnoses who have chronic excretory failure of a lesser severity and the thing that you must remember as well is in the adult population with end-stage renal failure there’s a significant subset who have these diagnoses as well.

I won’t go very much into the anatomy of kidney development but just to remind you that it’s very complex, on the top left here is a human embryonic kidney at about 6 or 7 weeks gestation. You can see a few branches of the ureteric bud that will become the collecting ducts in the ureter and you can see a few little nephrons beginning to form. Then new nephrons continue to form during gestation up to about 34 weeks. At the same time that the kidney is developing in harmony with that, the ureter and the bladder develop. Here is an embryonic bladder at about 5 or 6 weeks gestation. It’s very simple with an epithelium and some mesenchyme and this too has got to mature to generate smooth muscle and a very complex specialised urothelium. Very often we see lower tract abnormalities, as well as kidney congenital abnormalities.

I was thinking about some strategies, how we can bring the new knowledge of genes that orchestrate kidney development based on animal studies really into the clinic. I think the most important thing here and this is just what I’d like you to take away from the talk is I think we need to establish clinics where affected individuals and their families can be assessed in one room by an expert in nephrology, urology if need be and also with a clinical geneticist sitting in the room and you’ll see in my talk why this is very useful. I think as we go to further into human research, we have to think about international collections of cases to look for genes. We’re going to have some very complex whole genome searches, we’re going to be looking for microdeletions in the future. Another challenge will be to understand how phenotypes are determined by the interaction of the genome and environment and we need better animal models as well to mimic human disease.

Since about 1 year ago I’ve been doing such a clinic once or twice a month, in London, U.K. and a Professor of clinical genetics, Professor Raoul Hennekam sits in the room with me to assess these children and their families who are born with renal malformations.

The type of children we’re seeing is a child with a renal tract malformation and perhaps another organ involved, developmental delay, maybe a heart defect or a gut defect. This is what we call a syndromic case or maybe a child with a renal tract malformation and a living sibling or a parent with a known renal tract malformation or thirdly another category is a child with a renal tract malformation and a history of one or more siblings who died before birth or neonatally and they would have had very severe bilateral renal tract malformations. So the idea here is to make genetic diagnoses and to counsel families.

This is a beautiful picture from one of my PhD students Clare Gannon in London. These are two embryonic ureters that have grown up in organ culture and you can see the green is the epithelium that is expressing a specialised gene called uroplakin and the epithelium is surrounded by smooth muscle and these two tissues need to develop and talk to each other for a normal ureter to be made.

The commonest ureter malformation is, of course, vesicoureteric reflux seen here in a cystogram, unilateral reflux with intra-renal reflux here and very commonly we see this together with a dysplastic or a hypoplastic kidney.

A few years ago certain candidates based on mouse experiments became good genes for thinking about human ureteric malformations. The set of genes I’ll tell you about are the uroplakins. These are a set of proteins that coat the surface of the urothelium from the renal pelvis down to the beginning of the urethra. This is what it would look like by a scanning electron microscope, they form these plaques on the urothelial surface. By immunohistochemistry this is a human fetal bladder, you see a thin line coating the surface and there’s a family of these proteins shown in the lower right that interact to form these armour plated plaques.

Now we got interested when Henry Sun in New York showed that mice lacking uroplakin IIIA have severe vesicoureteric reflux and are born with severe hydronephrosis. This is a normal junction of ureter and the bladder and in a mutant mouse they have a sort of golf hole instead of the normal junction. You can see here a neonatal mutant mouse with severe hydronephrosis.

If you’re very clever, you can perform intravenous pyelograms on mice. I know we don’t do it in humans much but in a normal mouse you would see a very faint renogram with most contrast having entered the bladder but in the uroplakin mutant mice you see a delayed clearing of the contrast from the kidneys. This is because in some of the mice either they have vesicoureteric reflux or they have urethers that are actually obliterated, they have a physical obstruction of urine flow before birth.

We began to look for uroplakin III mutations in our clinics in London starting off with children with severe bilateral renal dysplasia and ureteric malformations. Out of 20 of these children we found 4 with de novo uroplakin III missense mutations. Here are the two normal parents in this family and the de novo mutation occurring in conjunction with this severe case of dysplasia.

More recently we found probably a polymorphism in another uroplakin, uroplakin II that causes a frame shift, this can occur in the normal population as well but in this particular patient there was bilateral reflux and severe hydronephrosis.

Uroplakins have been pictured really as forming the armour plating that prevents urine going back into the body from the urinary tract. But we believe that the uroplakin family does more than that we think that they transduce a signal from the urine into the urothelium that maintains the differentiated state of the bladder and the urothelium and the reason we think this is believe it or not that frog eggs are normally coated by a frog version of uroplakin. When a sperm hits a frog egg it fertilises that egg by virtue of interacting and signalling through frog uroplakin. It’s very interesting that in our cohort of patients with uroplakin mutations there are missense mutations just outside the membrane domain in the intracellular domain nearer to phosphorylated tyrosine and the equivalent areas of the frog protein are very important in cell signalling.

This is a human embryonic kidney at 5 weeks gestation, there’s just an epithelium and some mesenchyme. These two tissues will talk to each other over the next few months to generate your nephron tubules, your glomeruli, collecting ducts and we think endothelia as well sometimes can arise within the kidney.

The gene I want to tell you about is called hepatocyte nuclear factor 1ß that many of you may have heard of already. This was first implicated in renal disease about 6 or 7 years ago in the context of something called renal cysts and diabetes syndrome or RCAD. This can occur in an autosomal dominant manner or it can be sporadic and these individuals were first described as having maturity onset of diabetes in the young, uterus malformations and kidney cysts. However, the kidney disease they get is not classic diabetic nephropathy, it’s a developmental defect and these individuals were found to have hepatocyte nuclear factor 1ß or HNF 1ß mutations.

What is HNF1ß? Well, it’s a member of a two gene family that’s ß and α and both of these have been implicated in maturity onset of diabetes in the young. In fact, α much more commonly. They are transcription factors, they have domains that will bind to DNA and they have C terminals that will transactivate DNA to regulate the expression of other genes.

In the early 2000s when the renal cysts and diabetes syndrome was first described the kind of phenotypes that we were seeing and other people described were maybe a kidney of normal size that is polycystic with numerous small cysts or perhaps a kind of small dysplastic kidney with a few cysts. On histology these kidneys either had dysplastic tubules or some of them had glomerular cysts and this is a version of glomerulocystic kidney disease.

This was some work a few years ago from our lab using the human embryo bank at our institute. Here are 3 normal human metanephroi from between about 5 and 10 weeks gestation. The metanephros at that stage expresses HNF but very interestingly in view of the link with diabetes the gene is also expressed in the embryonic pancreas as well.

Now, there have been some very interesting animal studies done more recently. This is work from Gresh et al in the EMBO Journal a few years ago. This is a normal mouse post natal kidney with the cortex and the medulla and the papilla. On the right here is a litter mate kidney which has had the HNF1ß gene knocked out in renal tubules and you see they get a form of polycystic kidney disease in this case mostly in the medulla of the kidney but there are a few cortical cysts as well.

The reason why lack of HNF1ß causes cysts is probably because one of the jobs that this gene has is to upregulate the transcription of ‘anticyst genes’ and two of the genes that it controls the expression of are the gene for autosomal recessive polycystic kidney and uromodulin that many of you know is implicated in another form of human cystic disease, medullary cystic disease.

I just want to give you a few examples of families that we’ve seen in our combined nephrology and genetics clinic to show you, give you a clinical flavour of what this gene means and I think you’ll find it interesting. This first case is a girl of eleven years old at the moment. She had an antenatal diagnosis of a multicystic dysplastic kidney on the right side but postnatally the left kidney was of normal size i.e. that’s extremely abnormal. If you have a non-functioning kidney on the other side, your solitary kidney should be much bigger than normal. So, her solitary functioning kidney was probably hypoplastic. She had persistently raised AST and ALT liver enzymes from the first year of life.

And then she went on rather well. But about two years ago when she was nine she put on weight and her fasting glucoses although they were normal she had raised insulin levels. Then last year she presented with overt really florid diabetes mellitus with polyuria and polydipsia. It was non-ketotic. She has a distant cousin who had diabetes onset at about 13 years of age and we found that this index case has an HNF1ß splice site mutation.

Now, another family that we saw a few months ago is fairly interesting. There are two siblings a girl and a boy, 5 and 3 years old and they both had unilateral multicystic dysplastic kidneys. Both older brother and both parents have normal renal ultrasound scans but this is a key point here, the father was diagnosed with acute gout in his 30s and the grandfather has diabetes mellitus. We found that both index cases with multicystic kidney have an HNF1ß mutation and so does the father who has a history of acute gout but the father has a normal renal ultrasound scan.

If you don’t know what it looks like, this is what a multicystic dysplastic kidney looks like. They’re great big ugly things full of useless non-functional tissue but full of great big cysts.

Edith Potter who was a famous fetal histopathologist about 50 or 60 years ago described in microdissection studies how the normal human ureteric bud would branch serially into collecting ducts and her microdissection studies of cystic dysplastic kidneys showed that these big cysts were attached to malformed ureteric bud and collecting duct derivatives. So, the explanation that links normal collecting duct morphogenesis with multicystic dysplastic kidney is because the HNF1ß gene is normally expressed in the early branches of the collecting duct system.

Here is a human embryo at about 10 weeks of gestation. Everything you see that is kind of purpley brown is expressing HNF1ß and here you can see an example of a beautiful branching collecting duct as it goes from the medulla into the cortex, so one can understand why a mutation here causes this type of multicystic phenotype.

The last family I wanted to tell you about is a little bit different and broadens the spectrum of this syndrome. There was a female now 4 years old born with a left-sided renal tract dilatation which was actually detected antenatally. She would have what we would commonly call a congenital pelviureteric junction obstruction. She has a heterozygous deletion of the HNF1ß gene.

Her brother is 2 years old, he in midgestation was found to have a multicystic dysplastic kidney on one side and a hydronephrosis on the other side subsequently confirmed postnatally as what we would call the pelviureteric junction obstruction. He also has a heterozygous deletion of the HNF1ß gene.

Here’s a picture I’ve taken, this is where I learn all my nephrology now from the eMedicine website. It had a nice picture of a congenital PUJ. So you can see that the phenotypes associated with HNF mutations are fantastically broad.

In the last 7 years up to April 2007 we have requested that Sian Ellard and Coralie Bingham in the Exeter laboratories in the U:K. assess the HNF1ß gene in 80 children with congenital renal disease and so far 14 of the 80 which is about 18% were found to have HNF1ß mutations and about half of those have whole gene deletions. It’s a fantastically common genetic cause of kidney disease in our clinics. Ten years ago it was totally unrecognised.

Who should one suspect has HNF1ß mutations? Classically it would have been in a family with dominant inheritance of cystic renal dysplasia or glomerulocystic disease and diabetes. However, in our cohort in London diabetes is actually rather uncommon and we’re seeing other phenotypes like multicystic kidneys, solitary functioning kidneys and even congenital hydronephrosis. In some of these children one sees raised liver transaminases, or a history of acute gout in reletives.

Genetic testing may cost several hundred Euros, so ok that’s maybe not that different from ordering an ultrasound scan but if you spend that money and come up with a mutation, you provide the family with an answer to their often long-sought question, why has my child been born with a kidney malformation? You know just to have that knowledge is so important. Many of these parents have been wondering if they ate something wrong in pregnancy or whatever. They’ve been wanting to know this answer. The big question is though should we perform genetic and/or renal ultrasound screening of parents, siblings and of course, the next generation? Here nephrologists really need to link up with clinical geneticists in order to weigh up the pros and cons of doing this type of screening tests and you need a geneticist to help with counselling. I don’t believe a nephrologist can do it really on their own very well.

Here I end just hopefully giving you a flavour of what’s going on, translating some basic mouse discoveries into the clinic. I wanted to thank the children and their families, Sian Ellard, Coralie Bingham who do all the HNF sequencing. Raoul Hennekam is my clinical genetics guru, Maria Kolatsi, Dagan Jenkins, Claire Gannon, 3 PhD students who did a lot of the work that I’ve shown you. Thank very much.