Slide 1

Thank you Michael, thank you Gerard. Dear colleagues I’m very happy that so many were able to get out of bed this early, so very good job. I hope we will not bore you then.

Slide 2

You see my topic is molecular mechanisms of calcification and I would like to start briefly with this unofficial logo of KDIGO group which points to the fact that maybe mineral and bone disorder is a systemic disease and part of this systemic disease and potentially related to outcomes is vascular calcification. 

Slide 3

This is what we have been talking about for the last maybe ten years, an example of a CT scan with a heavily calcified coronary artery in a relatively young dialysis patient. This is so prevalent in the dialysis population and so different form the normal population that we wonder how this comes to pass.

Slide 4

Another disastrous disease just as an example is calciphylaxis starting with these painful plaques exulcerating and the bases is probably heavy calcification of small cutaneous arterioles.

Slide 5

We can visualise, especially in animals, these calcifications very nicely. This is microCT scan you can see here the paraffin block of aorta is from uremic rats. You can see it here in a 3-dimensional resolution which is difficult to achieve in the whole human, however we have some clues on how that works. If you look, for example, at the vessel wall with transmission electron microscopy, these are iliac arteries taken out or pieces of iliac arteries taken out at the time of transplantation.

Slide 6

This is how small calcifications start to develop and what we want to know or learn about is how we can make this stage stop and prevent these quite disastrous confluent calcifications that may occur at these vessel sites.

Slide 7

We also have soft tissue calcifications like here in periarticular localisations and I’ll show you this syringe with this milky stuff inside because I will come back to that at the end of my presentation.

Slide 8

I made a question mark because it’s probably, we are wondering where this may come from.

Slide 9

This actually comes from the abdomen and this is a picture that David Goldsmith once showed us here. This is a calcific sclerosing peritonitis and we had a similar patient recently with an ascitis phosphate of 13.3mmol/l in the presence of a serum phosphate of 1.1. We don’t have a real good clue about how this could happen but as I said, I will come back to that later.

Slide 10

I pointed out, I put out three topics for my talk; vitamin K and matrix Gla protein, biomarker and therapeutic target, fetuin A, the injury hypothesis and third part calciprotein particles: a novel concept of crystal clearance.

Slide 11

And I would like to start with the first one.

Slide 12

I’ve shown these slides a couple of times before but it’s important to realise that matrix Gla protein is a vitamin K-dependent protein which is produced in condrocytes and vascular smooth muscle cells and there it acts as a calcification inhibitor.

Slide 13

You’re of course, also aware of the coagulation proteins that are vitamin K-dependent and this works in a very similar way.
Once matrix Gla protein is knocked out arteries are calcified nearly completely everything that is red here around these aortic tree of these mice is calcification and these aortas break and these mice bleed to death at the age of 6-8 weeks. In the humans Keutel-syndrome these are children with a non-functional MGP protein, it’s a rare disease and they are mostly characterised by abnormal cartilage calcifications but we also know that these children already have medial calcifications from autopsy results.

Slide 14

So how does MGP work? It works really within the vessel wall, so if you look at an artery of a wild type animal this is perfectly clean of calcifications. MGP knockouts have heavy medial calcifications along the elastic lamina but if you succeed to over express MGP within the vessel wall, so taking a smooth muscle cell specific promoter, you can prevent calcification completely. But if you over express MGP in the liver by raising serum levels of MGP, it does not prevent calcification so it’s not a target that you can give intravenously, for example. But you can activate it within the vessel wall and this is an important finding for the understating.

Slide 15

There’s a very nice paper by Leon Schurgers out in Blood last year. What this group did was it calcified the vessels of rats by treating them with warfarin. It’s very well known in rodents that warfarin causes directly arterial calcifications. The approach was giving a high dose of warfarin together with vitamin K because by that approach you can leave the coagulation proteins intact without activating peripheral vitamin K-dependent proteins. Then the group was split into 4 more groups. Some groups were treated with a warfarin and vitamin K combination. In the other 3 groups warfarin was withdrawn but in this group only a very low dose of vitamin K was given and in these 2 groups high doses of vitamin K were given.

Slide 16

This is the readout of the experiment. So if you take control vessels, they don’t contain a lot of calcium but they contain a lot of active Gla MGP. Warfarin keeps the vessels calcifying and all the MGP within the vessel wall is inactive. When warfarin is stopped but only low doses of vitamin K are given, it’s not sufficient to reactivate the MGP and the calcification goes on. However, after withdrawal of warfarin if you treat the animals with high vitamin K doses, you can partially regress calcification by in parallel activating the MGP within the vessel wall. So this could really be a treatment approach.

Slide 17

Another approach to influence even MGP-dependent calcifications is to create a model that is hypophosphatemic. In this approach in the Hyp mice Phex deficiency leads to high FGF-23 levels subsequently to hypophosphatemia and in the presence of hypophosphatemia the vessels cannot calcify anymore even if MGP is knocked out.

Slide 18

Now, in the humans this seems to work in a pretty similar way. If you stain, for example, a diabetic vessel in this regard and you stain it for total MGP, you see a lot of MGP around or in the vicinity of the calcification. Here is a calcification stain and here you can see that there is a lot of inactive undercarboxylated MGP in the vessel wall but hardly any active gamma-carboxylated. So in situations of progressive calcification the active MGP or the vitamin K availability may get spent and decreases locally.

Slide 19

Now, there are assays around and the first assay for the matrix Gla protein measures the undercarboxylated so what I showed you in the vessel wall before and actually if you for example, compare controls to dialysis patients, dialysis patients actually have very low undercarboxylated MGP levels and this is probably due to the fact that they are stuck or trapped in the calcified areas and you can see it here compared to augmentation index a measure of vascular stiffness. If you look at several groups of patients; patients from a reference population, patients with coronary artery disease, aortic stenosis, hemodialysis, calciphylaxis, you see that you see MGP levels drop quite tremendously so it’s probably a biomarker of the severity of calcification. There is hardly any overlap between a reference population and patients on dialysis or with calciphylaxis.

Slide 20

Where we have to go with that is correlate these parameters with calcification scores and even better create an assay, they chose a ratio between the active and the inactive MGP.

Slide 21

Second point fetuin-A and vascular calcification.

Slide 22

Most of you are aware of fetuin-A, it’s a liver protein, it’s present in the serum electrophoresis, it has high serum concentrations and it’s down regulated in situations of inflammation. If you look at the formation of hydroxyapatite, so calcium-phosphate crystals, they usually look like this, very solid, a very sharp edge and in the presence of fetuin-A they look softer, smaller, poorer in a way so fetuin seems to interact directly with crystallization processes. If you then look into a vessel wall or on a vessel wall, you see that fetuin-A surrounds calcification – but without being expressed within the vessel wall. So it’s deposited there.

Slide 23

If you then go into the cell culture, you here have vascular smooth muscle cells in the presence of serum or just albumin and the presence of albumin calcification is heavy. If you then start to add increasing concentrations of fetuin-A, you can prevent calcification of these vascular smooth muscle cells quite completely.

Slide 24

But how does that work? I told you that fetuin-A is not expressed within the cells, so this is a vascular smooth muscle cell that is just kept in albumin or in cell culture conditions. If fetuin-A is added here, the fetuin-A actually enters the cells and this is probably taken up by an Exon 2 channel, so calcium channels.

Slide 25

And while it enters the cells or the cells then secrete matrix vesicles which usually look like that, they are membrane coated and contain a lot of calcified calcifications, in the presence of fetuin-A they don’t contain calcifications but they contain fetuin-A which is shown here as an immunogold labelling. This is a tremendous effect. Matrix vesicles are calcifying without the presence of fetuin-A heavily but hardly in the presence of fetuin-A and even the calcification of apoptotic bodies is blunted in the presence of fetuin-A.

Slide 26

So this is a picture many of you have seen before, fetuin-a knockout leads a lot of soft tissue calcifications, organ calcifications like myocardium, skin and lung.

Slide 27

We and others have shown that in dialysis patients low fetuin-A levels predict all-cause mortality, predict cardiovascular mortality but we had one problem in this regard with our argumentation because when we knocked out fetuin in the mice we got calcifications nearly everywhere but not within the vessel wall.

Slide 28

We saw some stones within the vasculature that may obstruct little left vessels but we didn’t see it here and this is why we thought that it is necessary to have a pre-injured vascular system in order to get calcification. This is why we crossbred fetuin-A and ApoE knockout mice in order to create a double knockout mouse.

Slide 29

So does calcification inhibitor deficiency aggravate plaque calcification in this model of atherosclerosis? So we had 3 groups of mice; the control wild type, the atherosclerotic and the double knockouts and we also did 3 interventions we had control animals, animals on a high phosphate diet and uninephrectomy plus high phosphate diet.

Slide 30

So what you can see here in the uninephrectomised mice you have some degree of renal failure. Calcium levels stay about the same, phosphate levels rise with a high phosphate in all the groups and especially when uninephrectomy is performed. PTH rises with high phosphate and after uninephrectomy but cholesterol is very similar in the ApoE knockouts and in the double knockouts the weight development is similar, the mean arterial pressure is similar.

Slide 31

Then we looked at atherosclerosis plaque surface, intimal area and coronary sclerosis and there were hardly any differences between the ApoE knockouts and the double knockouts in all these readouts as you can see here.

Slide 32

But what we could see was that the calcification magnitude looking at aorta, looking at coronaries and also at the ulcerating plaque incidence was significantly higher in the animals on the fetuin-A knockout background so in the double knockouts.

Slide 33

That means once you have atherosclerosis or an injured vessel, you get a lot more calcium deposition. This is visualised here by 2 photon laser microscopy and elastin auto-fluorescence.

Slide 34

Finally calciprotein particles: a novel concept of crystal clearance.

Slide 35

If you put together for example, a calcium phosphate in a test tube without cells, they form in time and temperature dependent mode, they form these nuclei, these calcified nuclei. If you take out for example, small samples out of an artery and you look for calcified –, they look very similar to what you see in the test tube.

Slide 36

You can also visualise it this way with a 3D-tomographic modelling and you can prove that it is calcium phosphate that is contained here. So I come back to this milky solution and actually looking at this milky solution reveals very comparable calcified nuclei in this regard and interestingly, they all contain fetuin but they also all contain albumin in some proportion.

Slide 37

So it seems that the initial formation and inhibition of crystallization are fetuin-dependent because fetuin is strongly apparent in these crystals and albumin and acidic proteins are additionally contained.

Slide 38

They start as an amorphous mass and then they progress to these crystals in the presence of fetuin-A. If only albumin is available, they will not progress into these crystals or calciprotein particles. If you quantify it, you recover only about 40% of available fetuin-A in these crystals but about 70% of the available phosphate. So it’s stuck in these nuclei.

Slide 39

If you look at mineral sedimentation inhibition, there are linear relationships with fetuin-A, it is concentration dependent and there is a synergistic action of albumin and fetuin-A in this regard. So if you for example, look at that bar, you see a maximum inhibition with concentrations of fetuin and albumin.

Slide 40

This is also again fetuin if the ripening of these calciprotein particles is dose-dependent on fetuin-A concentrations and on temperature.

Slid 41

This is the concept of such a calciprotein particle that contains fetuin-A and calcium phosphate crystals and it is estimated currently that fetuin-A and calcium-phosphate ratio is approximately 1:80.

Slide 42

and at last years’ ASN I saw a couple of very interesting posters that looked at the serum of dialysis patients and controls and what they did was they spun down this group from Osaka, Hamano et al they spun down a serum with ultracentrifugation and took care of the pellet of this serum.

Slide 43

Here are 3 examples; 2 healthy subjects and a hemodialysis patient. The pellet was re-suspended here and from the re-suspended pallet fetuin was recovered just in the hemodialysis patient. This fetuin-A was proven to be fetuin-A by western blotting.

Slide 44

This shows what happens with a fetuin concentration in the serum actually. In normals nothing happens but after ultra-centrifugation in hemodialysis patients fetuin serum levels drop and if you then go back to the pallet and you re-solubulise, it the pallet releases calcium and phosphate. This is very, very good proof or idea that these large calciprotein particles circulate in late stage renal disease.

Slide 45

Just as an example, patients from CKD stage 4, CKD stage 5 after ultracentrifugation there’s a reduction rate of fetuin-A levels probably by precipitating these larger complexes by 24%, by 40.1% and this is important for the future I think when we start measuring fetuin levels in these regards.

Slide 46

This is a concept in a way just a conceptual slide, the last slide I have actually. It is currently possible, it’s a hypothesis that has still to be proven but that calciprotein particles are capable of taking up small crystallisation nuclei and transport them off the soft tissue perhaps into the bone perhaps into the – system. I would like o leave you with that. Thank you for your attention.

Slide 47

Chairman: Thank you so much, we have time for a few questions. May I ask you? So calcification maybe primarily localised to the endothelial intimal layer or to the smooth muscle layer. What is your view on fetuin-A and Gla? Do they differ?

Prof. Ketteler: The mechanism I think by which MGP inhibits calcification is not really clear I think. It could be due to keeping the integrity of elastin fibres up, it could be a direct effect, it could be an indirect effect by influencing bone morphogenetic proteins. I’m not aware of any mechanistic view here. Fetuin-A is available in very high serum concentrations and it’s just abundant, so it goes where the calcification starts and it has an interaction on a physical chemical level on the one hand but it is also capable of entering the cell and preventing the matrix vesicle formation. If this then works also at the extracellular level in order to take up crystals and transport them away this has yet to be demonstrated but I think we have now very good preliminary evidence for that.

Question: Markus a lovely talk and it’s fascinating to see how the biology of this is evolving very rapidly into very interesting areas. The warfarin story always interests me because we know we give our patients warfarin for atrial fibrillation or for poor vascular access but only a very small minority seem to have an acute problem with warfarin. It could be, of course that warfarin is associated with faster calcification but if so, I haven’t seen the data to show that definitively. What do you think the real explanation for that is? Have we looked at MCP levels in these patients?

Prof. Ketteler: Not really. I think that my hypothesis on this is that it’s a multiple shot, a multiple injury thing. So if you treat with warfarin, you definitely shut down MGP but if everything else is working, so you have low phosphate, you have good fetuin and so on nothing will happen. I think that this is a combination of injury. But I think it’s still an important one. You know that Janet Haggerty and we are collecting samples from calciphylaxis patients. In the German registry we now have collected 41 patients and actually in 42 of them the calciphylaxis was triggered under warfarin treatment. 42% which is very high I think. So we need to find out who is at risk, who is at danger and we need to find out what is the potential protective role of vitamin K in this regard in those patients who do not need warfarin treatment.

Question: Markus I have another question which is just a hypothesis. You mentioned all the list of vitamin K-dependent proteins like MGP and so on. There is one osteocalcin and you probably know the last publication by Gerard --- group who showed that osteocalcin is a regulator of energy metabolism, regulator of adipocytes, a regulator of adiponectin, blood levels and adiponectin prevents calcification by itself and prevents progression of atherosclerosis and there it seems to be that osteocalcin the – between bone and the vascular system is also something which could be very important.

Prof. Ketteler: Yes absolutely. I think this is an under recognised factor in this regard and if you especially look at populations like the Japanese women are treated for post-menopausal osteoporosis prevention it’s extremely high doses of vitamin K. Now Japanese women are the eldest people in the world, so whether this may have something to do with each other so osteocalcin activation plus MGP activation maybe beneficial in a cardiovascular context is very well possible but is still not investigated enough.

Question: Excellent talk thank you. I just wanted to go back to the vitamin D issue and warfarin. For those patients that you have with calciphylaxis are they on vitamin D? The question is just like in the experimental animals would the combination of warfarin and vitamin D put these patients at higher risk?

Prof. Ketteler: A very important and very difficult question. So I mean vitamin D is a part of the story because high doses may even trigger calcification but lower or some doses may prevent from cardiovascular complications. If we see a calciphylaxis patient who is on high doses or active vitamin D, we reduce it usually or even withdraw it for the time being but whether that’s the right approach we really don’t know. What we did in the past was recommend in patients who were on warfarin to withdraw the warfarin, put them on heparin, wait a couple of days or a week or so and give them higher doses of vitamin K. It will not hurt at all because you cannot over coagulate for example, but this may at least substitute a system that is operative in this danger cascade. But there’s no proof that it really helps.

Chairman: Thank you for this excellent talk.