 Good morning everybody. So today we go to the third day of the summer school. It's getting harder and harder, it seems to me, to start in the morning. But anyway, so as we said before, we started with pathophysiology on Monday. Yesterday we talked a little bit about data acquisition and how to get the data that we need in order to complement or validate our modeling. And today we will start with organ-specific modeling or multi-scale organ modeling also. And we're going to start with an application which, when you talk about flow, might not be the first one you would think of, but it just shows you very, very much how multi-scale, multi-physics comes in and where fluids are really, really important everywhere. So today we're going to start with a talk from René van Donkelei from Technical University in Eindhoven. And this is looking at cartilage and how diffusion, water, hydrostatics and things like that are really, really important. And how it's when you do the modeling, when you look at these things, it's the integration of pathophysiology with different types of modeling, different types of physics that will help you in order to understand things. Thank you very much for coming, René. Please, the floor is yours. Now it should work. Yes. So welcome again to all of you. Good morning. I understand it's early, but let's see how we can dive into cartilage. As was introduced, cartilage is not really the tissue where the fluid flow is the most turbulent, etc. But I would like to show you today that this is one of the, as we call it, solid tissues in which the fluid is really the most important thing. So really fluid is making the function of cartilage. That's what I'll try to explain to you today. And that's why I put the title there, like, Walking on Water. It's not one of the regular titles that I use. But Jerome thought it was the nicest title, so that's why it became there. So really we are walking on water, and I will show you how that works today. So let's take a first look at cartilage. What is cartilage? I don't know how familiar you are with the tissue. But cartilage is the material that is covering the ends of our bones. So everywhere where your two bones meet and where they sort of move, there's cartilage in between. And that's in your fingers, in your hips, in your knees. And especially in your knees, cartilage can become a problem because basically we, all our weight is on our cartilage, right? All our weight is on our knees. And the cartilage is really heavily loaded there. And if I say it's heavily loaded, you have to think about the equivalent of, like, three to four elephants put that on top of one refrigerator, and you will have the load that is at this moment on my knees when I'm just standing here. And that's because the contact area in the joint is really very small. And I'm a little bit overweight, but anyway. So all my weight is on that cartilage, and that is really a heavy load for the cartilage. And not even running or jumping or anything, right? So the cartilage is able to cope with all that load for a really long time. Cartilage ability to regenerate is really small. So if there is cartilage damage, it doesn't really heal automatically. So how does cartilage do that? That is basically the question. Now cartilage, if you look at it, it is a really, if you go into the joint with an arthroscope, and that's the picture that you see on the left. It's really a white, shiny, slippery tissue. And it's slippery because you have to move your joints very easily. And it's white because it doesn't have any blood vessels inside. It doesn't have any nerves also. So that is why you don't feel anything if you just walk. Now the problem comes when the cartilage starts to degenerate. And here you see two pictures of an intermediate and a more severe damaged cartilage in the joint. So what you see there is there is some yellow stuff shining through the cartilage. In the right picture you really see yellow, and the yellow is bone. So all the cartilage there is gum. And basically when you stand, you're standing on your bone, you're touching the bone. And that hurts. That hurts a lot. And the patients that have this kind of disease, which we generally call cartilage wear or joint wear or osteoarthritis, they really have a lot of pain. And that pain leads to disability, lack of joint motion, and really the patients don't want to walk anymore, which is of course not what you want. So does it happen often? Well, as I said, cartilage doesn't really repair. So if it doesn't repair then as soon as you get some kind of damage in your cartilage, the damage will stay there for the rest of your life. And it may only get worse and it may progress very fast or it may progress slowly, but it will not really heal. That's what we think. So suppose that you're 30 years old, you get some cartilage injury, then you will have a cartilage injury until you die basically. Which means that you may be for 60, 70 years, you may be needing drugs or pain medication, and that will only get more severe during your life. So once you get osteoarthritis, you will stay with osteoarthritis until, well, they replace your joint with a metal prosthesis or so. And that means that we get really a lot of people at the age of 65 and older, 80% of all the people have some form of osteoarthritis. And it may start at the young age, like 30 or so, 40 or 50. It's pretty expensive. The drug medication is not that cheap. And as I told you, the treatment is not that easy. Well, say that depending on what kind of treatments we do, we have like 40 to 70, 80% clinical success rate, which means that after five years, pain is less than it was before. Now if you then look at what the result is of the typical treatment, then here you see cartilage. Remember, the cartilage should be interesting. Remember, the cartilage should be really smooth and shiny and slippery. And this is what you get after a surgery. If you go back into the joint a couple of months or years after a surgery, you see that it's really fluffy, not really functional, not really low bearing. It is a fluffy new tissue that has formed there and it's not really functional. So the only thing that we can do is go in, take it away, close the patient and hope for a couple of other years. So this is not what you want. Now, talking about that cartilage in the degeneration, the question is not how is it possible that the cartilage damages, but the question is more how is it possible that some people can really get 80 years old without any cartilage damage. If the cartilage doesn't repair, then how is it possible that this stupid tissue is able to withstand all that load that it has to withstand for so many years without damaging? So for that we have to understand a little bit more about cartilage. So let's go a bit deeper into it. I cut the tissue there, cut the bone through the cartilage and you see that it's just a little tiny white layer of tissue, one to three millimeters thick and as I told you it doesn't have any blood vessels, it doesn't have any nerves, which is why it's a pain-free motion. And then if you make histological slices of it and you stain it for the contents, you would see that it is a, well, it's not so interesting. It's basically an amorphous tissue that has not much structure and the structure that we generally see if you look at this kind of staining is the proteoglycan content. And then you could also do another staining to look at the fibers inside. So the collagen, collagen type 2 fibers are in cartilage and the collagen type 2 fibers, if you look at them a little polarized microscopy, you see a picture like that, which means that fibers in the tissue basically start from the bottom, where the bone is, that the fibers start, they go up, they bend over and they form a layer that is horizontal at the surface. That is what the collagen looks like in cartilage. Now you see that there's about 15% of your total mass of proteoglycans, about 12% is collagen, there's 1, 2% is cells and some other proteins, the most that we have in cartilage is water. So we have 70% water in cartilage about that. Which is very important for cartilage, for the function of cartilage and in the past, like two decades ago, people thought that basically fluid would be the most important thing in cartilage. And if you would take something that had 70% water and 30% solid and you would compress it, then basically what would happen is that the tissue is compressed and water is being expelled from the tissue. And because it happens in that way, your weight will basically go down slower and slower, because the fluid goes down, goes out, and at a certain moment in time, so when the tissue gets smaller and smaller, it will equilibrate and it will stay there at a certain height. So the strain will slowly increase and that's it. But what happens inside is then that initially when you put the weight on, there's a lot of water inside and that water is pressurized. But because the water is pressurized internally and not externally, there's of course a gradient in the pressure and the fluid will move out. That's what happens. And slowly over time when more fluid is flowing out, tissue is compressed and the solid is deformed and the solid is taken over. So now the solid stress has increased in the tissue and this is what we would call a bifasic mechanics. So this is until a couple of decades ago. We thought about cartilage in this particular way. There's water, there's a solid. When you compress it, fluid goes out, solid stress has increased, and the solid will now take up all the load. Now I will show you that this concept is basically wrong, but this is how many people still think about cartilage. And this is what we would call bifasic or poroelastic cartilage. So why is it wrong? I will show you, I will demonstrate you why that is wrong and why water in cartilage is much more important than just that it initially takes up the load and flows out. And to tell you about that, I always bring this polar bear because it's listening to my story just like you very nicely and it's sitting on that mattress. And it is sitting on the mattress in a very nice and elegant way because the mattress basically contains air. So it's not the mattress itself that makes it comfortable for the polar bear, but it's the air inside the mattress that makes it comfortable. Now think about an air mattress. If you bring an air mattress to your holidays and you inflate it, then you feel and you inflate it a little bit more. And at a certain moment in time, that air mattress is okay for a kid, for a child. And the child can sit on it or lie on it and sleep, but if you go on it, then it's too soft. So you have to put in more air, right? And you put in more air until it's okay for you and then you can sleep on it, but when the polar bear comes in and he scares you away, then if he wants to sit on the air mattress he first has to inflate it a bit more because that polar bear is heavier than you are. It's as simple as that. And that means that the heavier the weight on that mattress, the more air you should put in the mattress, otherwise it doesn't function. And it's not just the amount of air that you put in because basically if the air mattress was twice as large and you put in twice as much air, it doesn't make sense, but it's the air pressure that you have inside. So the heavier the weight you want to carry with the air mattress, the more pressure you have to put in before. The air pressure is what is carrying your weight. Now if you open the mattress, then you would always bring this El Chippo air mattress. You can see what is inside and you see that inside that mattress there are these kind of pillars inside. Now what happens is that the air pressure that you have inside that builds inside that air mattress basically is extending the polymer or the rubber. And it's also extending those pillars inside. And the pillars make sure that the mattress stays flat. That it's not going like a balloon because otherwise it would not be easy to sit on. But also it sort of distributes all that load, all that tension in the polymer. So the pillars are important in that mattress. So what does that... I tell you about all this and all this air pressure and stuff but what does it have to do with cartilage? Well, cartilage has a lot of water and it is attracting water. So let's see what the effect is of that as opposed to having a lot of air pressure inside. We just try to suck in a lot of water. And I made this hydrogel in the lab and this hydrogel is really a hydrophilic hydrogel and if you make it it looks like this and if you let it swell, it swells and it really swells like crazy. So it sucks up water, water goes in and it really increases in volume a couple of times. It's not very special. It is a polymer that you could easily create in the lab and think about diapers for instance that really have to take up a lot of water and they get 20 times their thickness if they take up all the water of the babies. That's basically the same stuff. It's a super slurper. So it swells. Now, it swells basically the same as when you increase the air mattress. So the air mattress pressure basically is the same as the fact that that polymer, that hydrogel, is swelling. And that is the same as what happens with the proteoglycans in cartilage because if you look in cartilage you have all these, I told you there are these proteoglycans and what the proteoglycans do in there is they attract water from the outside. Now, the way they do that is proteoglycans are basically very long abundant chains of polymer chains that have side branches and these side branches are called glycosaminoglycans and all these side branches are negatively charged. There are a lot of negative charges inside which means that this whole chain, this whole aggregate is really heavily negatively charged. Now, of course if you touch the cartilage you don't get an electric shock. So what happens in that tissue is there are a lot of ions, positive ions that come in and counterbalance the negative charges. So the total material is neutral. But now you have a lot of these positive ions floating around. So a lot of particles. And if you have a lot of particles somewhere and somewhere else you don't have a lot of particles then the particles would like to move through diffusion and if they can't because they're stuck there for the electrical charges then the only thing that can happen is that the water goes into the tissue with the charges. And that is what happens. That is really what happens in the cartilage. The proteoglycans attract the ions. The ions are a lot of particles and the particles attract all the water. I'm sure you all know about this. But now if you think about it that you have this solid material you have water and you have an ionic phase that attracts the water that is your third phase. And that third phase means that we are now talking about mechanics that we call trifasic mechanics. So we have three phases now. Instead of just water and solid we also have an ionic phase. Now this is what happened in the polymer and this is how I so far compared the air mattress to this hydrogel to cartilage. But the trick is that in the air mattress we have those pillars inside. And if we inflate the air mattress we just inflate it but we also pressurize the mattress. And the pressure that was carrying the load of that polar bear. So could we also sort of pressurize the water if you want. That is the question. Could we do the same thing as in the air mattress? And I told you that in the cartilage we have these fibers, these collagen fibers and they start in the bone and they go up to the surface and then form that superficial layer there. So what would happen to our model system, our polymer if we would also put in fibers there? What would happen then? So we did. So we made the polymer exactly the same stuff and we made a flap of it and then we put in fibers, right? Just only vertical. And we made a version where we put in some fibers we made a version where we put in a lot of fibers and a version where we put in really many fibers. Now all these things were the same size. They were the size of that left figure that you see there. And if you start swelling it they basically swell until the second figure that you see. But if you put in the fibers you will see that it swells a little bit less. And if you put in more fibers it swells less and with the most fibers it swells even less. Now what you should understand now is that it swells less not because it wants to swell less but because it can swell more. Because the polymer that you have there is basically the same polymer. So the stuff that you have there wanted to swell until this size. But if there are fibers inside they still want to swell this size but they can't because the fibers sort of tension and they cannot swell more. So the tissue on the right hand side wants to swell but it can't swell. And that's what we wanted to illustrate at this time with this image and then I did it somewhere in my cupboard and I didn't look at it anymore and that was it for the time being. And then in the building where I'm working we moved floors so I was basically cleaning up my desk and my cupboards and I found this pot that had physiological saline inside and had these four samples. And I took out the first one which is the largest one and I tried to take it out but there was a lot of water inside and I couldn't get it out of the pot so it basically failed. I took the second one out and it was okay but some stuff went off and was kind of broken but still sort of okay. And then the third one and the fourth one. And I ended up with the fourth one I couldn't break it so the first one really crumbled when I looked at it and the fourth one really didn't break so I started putting it under my chair sat on it and tried to break it like this but I couldn't. Why? Because the water inside that fourth one is under pressure. It wants to swell there's a lot of swelling pressure a lot of swelling potential there's a lot of potential for the tissue to attract more water and to hold the water inside that it has inside and it's very difficult to put that water out because of the pressure the force at which it wants to attract even more water. So the water pressure inside that the osmotic pressure inside that stuff with the fibers that was carrying me on my chair and that pressure inside that tissue is really the same pressure that carries your load. So if you stand on your cartilage basically you're standing on a cartilage trying to swell because of all the proteoglycans but it cannot swell because the fibers restrict the swelling and that gives you a really high internal osmotic fluid pressure inside and that pressure is really well capable of carrying your load so at least that is our theory that is a theory that we well we knew that the fibers were sort of okay they were stretched because of the swelling and that would have some effect but when I looked at this stuff I thought that really it may be even more important than we think it is so nice theory good story but can you really sort of prove that so what would happen if you would put that in a model and looked at the effect of these osmotic pressures could you really do that and so we sat down and tried to have that whole theory and put that theory into a computer model of articulated cartilage and see how all these things play together and of course you start with what I called you a biophasic model so we have the solid and we have the water inside but now we know that if you look at the solid it basically has the fibers of course and it has the non-fibre amorphous matrix and that non-fibre amorphous matrix consists of well it's a polymer so it has some stiffness in itself but also it has a really strong swelling capacity and for the fibers well basically we know that fibers if you look at the fiber you can pull it and it's really stiff in tension but if you compress it it doesn't do much until you really have a lot of these collagen fibers piling up and then they have a certain compressive stiffness as well so basically there's build compression stiffness and we have tension in the fibers so they resist tension they resist the swelling therefore and if you put that all together then you come up with a equation for the total stress in your cartilage basically that has all these components inside it has a part that represents the water it has a part that represents the osmotic pressure it has a stiffness that represents the stiffness of the non-fibre amorphous matrix it has a stiffness for the collagen fibers and part of that stiffness is dependent on the fiber orientation which is the tensile part of your fibers so this equation is the basic equation that governs our cartilage material behavior and we call that a swelling because it is swelling fiber reinforced poroviscoelastic model for cartilage why poroviscoelastic and not poroelastic well if you look at these equations and these are the basic equations for the different parts in the model then I would like to draw your attention to the right-hand side where we see that pink kind of color that is for the fibers in tension so the fibers in tension if they are not in tension they don't do anything but if they are in tension then they have a certain stress of course and that stress is partly viscoelastic and we started to recognize that viscoelastic part of the collagen is really important and it adds a second time-dependent property to the tissue next to the water that you have to push out so if you compress something you push out the water it takes time for the water to push out so that is a time-dependent parameter but also the collagen fibers themselves are viscoelastic and that gives you a second time-dependent behavior the other thing is that in the yellow equation for the osmotic pressure there are a lot of constants in there there are two variables that one is the it's called Cx that's the external concentration of solids but in the body it's basically constant it's a body fluid and there's one parameter there the CF which is the fixed charge density number of negative charges that we have in the polymer network in the proteoglycans and this fixed charge density is something that you can measure so the only thing to determine that osmotic pressure is that you have to know in the proteoglycans and you can actually go into the lab and measure that easily that's what people do at the daily basin like tissue engineering so if you can measure that then we have some parameters left in our model and these parameters are all based of determining the composition so that solid fraction the moment the solid fraction is how much basically everything minus the fluid do we have in our tissue and you can see in the top graph that the amount of fluid from the top surface to the bottom surface in cartilage basically changes there's some more fluid on one side also the fixed charge density changes there's more fixed charge density more proteoglycans in the deep zone compared to the surface and also if you look at the collagen the amount of collagen also varies with where you are in your cartilage and the amount of collagen which is that low value there that is also measurable so we can measure how many collagen we have we can measure how many proteoglycans we have how many fixed charge density we have how much water we have in the tissue and then we have everything that we need to put into this equation once we know the behavior of collagen we know the behavior of proteoglycans we have everything so it is all measurable it is composition based and that's the good thing about this there's nothing phenomenological about it we can measure everything that we need and then put it in a model so we have basically no freedom anymore so what would happen could be then if we have no freedom in that sense all the components are determined and we need to of course the parameters in our mechanical equations in our constitutive equations is it possible that we get any kind of data from the literature or that we make ourselves sorry and then fit all these data at the same time together and that's what we tried so we took data from unconfined compression experiments from confined compression experiments from indentation tests from swelling tests where we challenge the tissue with different osmotic baths we did many of these tests and we tried to fit all the data at one time and see what happens is that in all these different tests and there are like 3 or 4 more of these kind of graphs that all fit equally well all these data basically they all fit with just one set of parameters for the cartilage so they gave us some confidence we had a very good description now of how cartilage works and basically it gives us some kind of an idea that cartilage is really a tissue that tries to swell it tries to get in water but it can't swell because of the collagen restricting the swelling and the osmotic pressure that builds up in the tissue the potential of the tissue to suck in more water that is carrying our load that is our idea that's what we started with so now that we have this model could we really check that theory with this model how important really is that osmotic pressure is it like 50% or 60% or even 90% determining the properties of cartilage we didn't really know exactly so we went in with the model and did a test so the cartilage has a superficial layer where the collagen is flat we have its transition zone where the cartilage sort of crosses over and we have a deep area where the cartilage where the collagen basically goes from the bottom to the top and if you compress the tissue then we wanted to see what happens and we thought that we should distinguish what happens in the superficial that middle in a deep zone because it's basically a different kind of tissue so we did it's kind of what you may expect in the beginning at least what I expected osmotic pressure is really, well if you just load the cartilage of course first your fluid flow goes up, your fluid pressure goes up and the fluid pressure it's a dotted line goes down over time as you squeeze out the water so that's what we expected, that's what bifasic mechanics also told us that's not surprising but what we were looking for is what happens over time, what happens in the equilibrium and to be honest I expected that the osmotic pressure would have a significant role and it does, it sort of has 40% of the total load bearing is for the osmotic pressure in that superficial layer and the collagen and the project lightings also have a role to play which is okay but that's only the surface now let's take a look at what happens in the middle and in the deep zone and it turns out that in the middle zone basically all your load has an osmotic pressure like 90, what is it, 97% even and that the collagen and the project lightings they do nothing they have no role to play and if you go to the deep zone you see the same thing the collagen and the project lightings are not strained, not stressed or they don't contribute to the load bearing there but the osmotic pressure is doing everything yeah so really as I told you we're walking on water the water is carrying our load for 92% if I'm just standing here now compare this to what we had in the very beginning the biphasic material so the biphasic material we had of course just like in the cartilage we had the fluid that goes down the fluid pressure builds up and goes down as the fluid leaves but in the biphasic material model it was the solid that was increasing the solid stress and the solid stress was now carrying my weight well that of course is pretty dangerous because if the solid is really strained if the solid is really compressed the solid may also get damaged and if you damage your solid you have a problem because the solid, the project lightings the collagen are hard to heal are hard to regenerate for your body and for the surgeon but in really cartilage it is water and it's pretty difficult to damage water I would say so if the water is carrying our weight we have no problem and that is why cartilage is so really functional that the water is carrying our weight and that is also why the biphasic models to study cartilage biphasic models are really if you want to study the cartilage damage because the damage will depend on what the solid feels and the solid feels really much less than what you would expect based on biphasic mechanics so so that's the difference between these models so that is fun so the osmotic pressure determines all that loading and I already made a link to damage and I tried to explain to you that water pressure really may protect us from damaging our cartilage but of course we do know and that's the first picture I showed you we do know that many people have damaged the cartilage because the load on the cartilage is really really huge so that's why we made a next step we tried to make a next step towards really looking at what is it that damages cartilage and how do we become damaged and what happens then and then we put into our computer model we put a damage model I don't know how familiar you are to modeling damage in tissues but basically what you do is you look at the strains that you get in your tissue like in your collagen or in your solid you look at the strains that you have and when the strains cross a certain level then you may expect that you can strain collagen to 10% then you basically until the strain in the collagen is 10% nothing happens and you basically have your elastic collagen behavior but if you cross that 10% then part of the collagen may start to damage and that part of the damage that is what we sort of store it is an amount of damage that has happened and will not get away so that part of the damage also means that there is some weakening of the collagen and the higher the strain and the strain is on the horizontal axis and there is a K value the K0 is when you start to damage your collagen and then the KC at the far end is when you have damaged all your collagen and between you start damaging your collagen and you damage all the collagen basically the stiffness goes down and we put a very simple linear model there so the more you strain the collagen the more damage it will become and then you want to also look at well not very locally because then you are dependent on really at a certain point maybe some extrapolations of your model so what you want to do with that strain value you don't want to look at it very locally only but you want to look at a sort of average strain value in a certain environment in that tissue and if that average crosses a certain level then you say that there locally you have collagen damage or you have proteoglycan damage for that matter so what we did is for the collagen we said that if the collagen is basically strained to a certain extent in the direction of the collagen wherever we are then if it crosses like 6% and that's based on literature values we start to get damage if it reaches 18% strain it will be completely damaged and for the proteoglycans we said that it's an amorphous matrix we said that not the straining but the deformation itself the shearing that will cause damage and we thought that the shearing would start to cause damage at 30% shear and that it will be completed at 60% shear we had no clue there is no data around but we just did that and now see what happens so we have this model but we also do an experiment of course in the experiment we take cartilage which is attached to bone we load it with an indenter we basically cut slices and then we strain the damage of the collagen and for the loss of proteoglycans from the tissue if you strain for damaging the collagen you get a stain like this and the brown stain means that is damage now you see a brown stain really in the center of the cartilage on the top image not on the surface the surface seems to be like normal then you have this brown blob the cartilage damaged and we have the bottom layer of brown stuff and that brown stuff is always there you shouldn't notice it is because of basically how cartilage develops the tissue below where it's attached to the bone always stains brown so don't pay attention to that but there is a blob there with collagen damage in the center of the cartilage if you load with a higher load you see that that blob of cartilage not only is in the center but also goes to the surface has collagen damage so that is fun there is a really peculiar way that cartilage is getting damaged in the experiment now what does the model say well in the model we basically from the top to the bottom on that vertical axis you see the cartilage surface and the bone and if you go from the surface to the bone and on the horizontal axis you see the total amount of damage that really the total amount of collagen damage is not at the surface but it's below the surface there is an area where you have collagen damage and if you look at the higher loaded tissue you see that really that's the bottom right graph you see that really there is also a lot of damage at the surface really the same as the experiment shows so the model gives us the same Julia pattern for collagen damage as the experiment which is fun so the PSD student that did this he was very happy with it because we had the experimental data earlier than the model and I showed him the experiment and he showed me the model and he said well this is perfect I made a model which is really nice validation and he said well that is fun so I was also very enthusiastic and then he said but there is one thing and it goes up to like maybe 35% of the total collagen is damaged how much collagen damage do you have I said well I don't know I cannot quantify it I can only see where it is but he said well the collagen is damaged for 35% but if I look at my protroglycans they are basically 100% damaged at the surface as soon as I start loading with a certain load I see that the top surface of the protroglycans is damaged and if I increase the load 100% of the protroglycans are damaged in a deeper area and in a deeper area and in a deeper area and he says that I guess that the protroglycans do much more because it is 100% damaged and there is only 30% damage in the collagen and at that time I told him well unfortunately I have to tell you that you have to go back to your model because your model is wrong because we don't see any protroglycan damage at all whatsoever and so I think I don't know what kind of level threshold level he had for protroglycan damage and I think it was like 10% or 20% of the time and he said okay let's put protroglycan damage at 30% or 40% strain before it gets damaged because then we will get less protroglycan damage and he did that and he came back and said well I have the same results it doesn't matter what we do to the protroglycans they are always damaged in the model so there must be protroglycan damage and I told him I showed the picture again and I said there is no damage so we had a sort of a fight between us and then the solution was okay we get this master student you get the master student and you let the master student prove that you are right or wrong that is his task for the next year so her task basically for the next year and what the student shows in the end is yes there is protroglycan damage it's 100% it starts at the surface and if you load it more it gets deeper exactly like the model predicted so why didn't we see that before the reason is that when you do this staining you basically look at where the protroglycans are and if you do that immediately after the loading protroglycans are still everywhere what you have to do is you have to do that loading, that damaging load and you have to put the tissue back in a culturedish for 48 hours and give the protroglycans that are broken just basically time to leave the tissue if you don't give them time to leave they're still there and you don't see it as damage but if you give them time to leave you really see the same pattern as what he predicted with his model so that's very nice and we did a lot more experiments than this and sort of sometimes confirmed, sometimes had some contradicting results because these experiments are not that easy to do but overall we must say that we think that really we have these peculiar patterns these different patterns of protroglycan damage and collagen damage that they are really there we did a lot of different experiments but we never really changed the model the model seems to be pretty consistent and this seems to be what is happening in the tissue okay, so but what does it mean do we really have to take it into account this kind of damage internally so therefore we did the following experiment in the model so we basically loaded our tissue 5% or compressed the tissue 5% with an indenter and on the on the graph you see the reaction force that it takes to compress the tissue 5% over time so you load it 5% you keep it like that and you look at the reaction force and initially and you also compute that and you see that that goes well until 10% compression so there is really no damage in the tissue and you get this graph where you see that when you load it when you compress it 5% more that you really need a lot of force and then it goes to some kind of equilibrium as the water goes out that's fun, it's nice it's just the way you expect but when you start loading 15% you see that there is some damage in your cartilage and now what you can do in your model and you cannot do in the experiment if you can tell the model damage is happening or you can tell the model just continue without any damage of course in the experiment there is always damage but in the model you could say continue without damage and if you would do that if you would think there is no damage then you would get that blue line and if there was damage that red line in that curve and if you continue to 20% strain you see that they really start to deviate as a result of the damage your tissue gets softer now suppose you don't notice because the damage is really internally you don't see it from the cartilage or you think you have nice healthy cartilage and suppose you have that red line because it is damaged but you don't see it and you fit your material parameters to it then you get a completely different fit of your material parameters because you think it is a healthy tissue that you are fitting but really it is determined for a large part by the damage that is happening internally and you didn't take into consideration so yes, it matters what you do there you have to really be very careful with your experiments and really check that you are measuring properties only in the physiological range and if you have damage you have to take into account because otherwise the fit of your model is not really a realistic fit of what is the material properties so then finally I think so then finally we thought that now we have this model and we know that we have all these water and these time dependent properties and stuff can we challenge our model a little bit more with a discussion that I had at some conference with Gerard Eschan basically from the United States Columbia University and he told me that she has this model very nice because no one has a damage model so we have that and it looks great but he also said well you know this is a linear model it is only damage if you cross a certain level basically strain dependent but it cannot really be correct completely at least because we know that damage is also happening dependent on the strain rate so if you damage if you load cartilage faster you get more damage and if you load it slowly you get less damage and we have just a strain dependent cartilage model so we have to do many more experiments and we have to get all these cartilage damage model the collagen much more sophisticated before it gets value because we don't have strain rate dependent damage inside ok I told them but you know that is true we have a strain dependent model but we have strain rate dependent tissue behavior we have fluid inside and viscoelastic collagen but mainly we have fluid inside and because the fluid is inside if we compress it there is a time dependent response because it takes time for the fluid to go out and in that whole transient which can take minutes, minutes, minutes like until equilibrium it may take even half an hour or an hour depending on what you do so the fluid flow is very slow that means that the time dependent response is very slow and very elaborate very important so could it be that even though we have just a strain dependent damage that the strain rate dependent mechanical behavior of the cartilage gives us a strain rate dependent damage could also not be known so we went in and we took a look at what happens to our model if you do strain rate dependent modeling by the way if you look at these tissues I think I forgot to tell you but this here is of course the cartilage with the damage and this here looks like a cartilage damage but that is the meniscus so that's just how unhealthy meniscus looks like and it's not cartilage damage but anyway doesn't matter for the story now let's take a look at cartilage and look at how strain rate would affect the damage development in cartilage so the first thing that you do is just take a very slow loading rate and a faster loading rate if you stop extremely fast 120 millimeters per minute it's 200 percent per second it's nothing like jumping that goes much faster but we have this computer model and it has its limitations so basically this was as fast as we could go at that moment with the loading but you know there is really several times difference between the slow loading rate and the fast loading rate and if you look at what the model predicts then now basically this is what happens if you do a slow loading and the colors represent the damage in the collagen so we're looking at damage development in collagen and we take an indenter and we basically compress the tissue with an indenter now the first thing you will see is that you thought it was actually symmetric model and the damage is not axisymmetric the reason is that these collagen they go up to the bone they cross over and they form the superficial layer but they do that predominantly in one direction but not in that direction so the collagen fiber if you look on top of the cartilage also has a preferred orientation and that preferred orientation in this model is in the plane where you see the red stuff so that is the preferred orientation at the surface and that is why in that direction basically you see collagen damage and not so much in the other direction it has to do with the preferred orientation at the surface of your collagen so the tissue doesn't behave like an axisymmetric model in that sense but it has orientations to it which is important because that means that you have to also look at this in 3D so if you do the slow loading I will show it again what you will see is that basically basically you will see that the damage starts at the edge of where the indenter is and at the edge the damage basically increases until you reach this damage as the equilibrium now let's see what happens on the fast loading on the fast loading if it plays you see that the damage really starts under the indenter and that from under the indenter it goes towards the side of the indenter there is it under the indenter under the indenter goes to the side of the indenter and here it was not starting under the indenter but immediately started at the edge so the way that the damage develops in these tissues is different depending on what your loading rate does and the reason is that the water that is under the indenter basically has to be moved and if you do it very slowly you give that water that is under the indenter time to move and as you give it time to move it basically moves away and the tissue can easily deform but if you do it fast you don't give the water time to move so if you compress it very fast the water is entrapped in the tissue and as it cannot move with respect to the tissue but it is compressed the whole thing, water and tissue together basically go like this so you compress and it goes like whoosh and that means that the whole tissue that is there also deforms and then when the water gets time to move out then under the indenter the tissue basically relaxes and the water goes to the sides but initially strain under the indenter that causes the tissue to become damaged if you do it fast under very slow loading conditions that straining is not that expressive because the water can already start to flow out before it deforms the tissue too much that is what is happening and so you see that the damage development is dependent on the loading rate even though damage itself is not strain rate dependent but it is again the fluid that governs these kind of effects so if you go from like 5 to 120 mm per minute loading and some intermediate speeds you can see these kind of things as intermediate pictures so you see that if you do it fast you have already a lot of damage under the indenter if you do it slower you see that there is not much damage already but you can see that there is not much load at the sides of the indenter earlier now you can also do that for not only different loading rates which is on the horizontal axis but also for different loading magnitudes on the vertical axis and you see that if you do it very slowly and with not that much load that you have no damage and if you do it faster you develop more damage because it gets more red you get more damage but also if you do more excessive loading of course you get more damage but what is sort of dominating these pictures is that if you do it with more load you get more damage but if you do it faster you see that the location of where the damage is happening is sort of changing and here you see that the damage is basically internally more if you have slower loading rates then if you have slower loading rates then faster loading rates and at higher magnitude there is more damage for the faster loading rates than for the smaller loading rates so again, yes there is this loading rate dependent damage developing in the tissue could we can validate that this is a process that we are currently doing well if you do small loading magnitudes, small loading rates you don't see any brownish color if we do higher loading, faster loading we do see that in the intermediate picture so the one that is halfway you see that there is some collagen damage halfway it's not that excessive really like what the picture shows us and if you do the higher loading you really see that there is a lot more collagen damage but also that the collagen damage goes to the surface sort of agreeing with what we found experimentally but of course there are also some parts that we do not really have in agreement with the model and that is that for the faster loading rate we seem to have more damage internally than was predicted by the model and also that red arrow shows that we see some clefts occurring at the surface which we don't know really where they come from there may be artifacts or there may be true but we have to do many more experiments to really validate this bottom line is that both in the experiment and in the model we see strain rate dependencies we see the same kind of patterns in where we find damage but a full really validation I wouldn't still call it like that but we do have a strain rate dependent damage development even though we only have a strain dependent model so if you then put everything together I think I show you how we get to our model and I also show you how important fluid is in this tissue so I guess that when the organizers invited me on a course like this which is mainly on fluid flows they realized that the solid tissue in which fluid is so important is cartilage and I show you that it's probably even more important than we used to think until like five years ago or so that really fluid is governing everything in cartilage but it can only do that because we have the collagen fibers and because the collagen fibers make sure that this tissue wants to swell but it cannot swell if we would let it swell it would be much thicker cartilage but it would not be functional because the fluid would be squeezed out easily and it would be very brittle but for the fact that we allow it to swell a little bit that it wants to swell more but it can't that is why it gets its mechanical properties and we're really walking on water we're walking on that pressure that osmotic pressure inside the water that's how cartilage works and I show you about the damage I show you that with the damage we could even predict what was happening experimentally and that it took really some effort to find out that the model was right experiments needed some reconsideration and that with this really simple damage model of course we have to extend it much more we know but there's not much literature out there to do that so so far we keep it like this until we have more data but we know that with this simple linear model we could even, or viscoelastic model or elastic model we could even have viscoelastic or time dependent damage development in the model so with that I would like to stop I think this was a presentation about water and cartilage I would like to thank you all for your attention I hope I explained to you how cartilage works and that you learned something today because that is the whole purpose of the summer school of course so I hope that you learned this morning also something about a tissue that you may be not the most interested in from your backgrounds it's a very interesting tissue and I would also like to thank the polar bear for helping me to explain all this thank you thank you Rene for this very nice talk is there some questions yeah thank you for your talk just a question about the clinical part of it so it demonstrated that you were able to produce the material in the lab that were reproducing the cartilage function of absorbing the shocks and handling the damages do you think that it will be possible in the future using materials that you create to provide to people that have these kind of problems replacement it's less on the model but more on the optimistic side of you because I'm young, I'm running so I'm maybe damaging my cartilage yeah I understand I think everybody could hear the question so the main part of the question is can we make something to replace the damaged cartilage so what surgeons do nowadays if you have a little bit of pain they send you home because they only do something when you have really a lot of pain which is of course bad already because when you come to the doctor for the first time you probably have still some healthy cartilage around it and when you get really terrible pain all of that healthy cartilage is also gone by that time on the other hand if they do something well you know what you have when you get to the doctor and you don't know what you get when you come back after the surgery so they are really conservative these days but they can try to regenerate your cartilage and they show you that fluffy stuff which is not really the best thing you can have now what the most recent developments are that are being used clinically is not that they so in the end what they do is they take away the whole joint and put in the whole metal prosthesis inside but what they do nowadays is that before they do that they take a kind of say a very small metal thing it looks like a mushroom, it has a cap and a stem a very small thing and they try to put that in your joint to replace just that damaged part of the cartilage but it's a metal thing and if you put that metal inside then if you don't do it really very nicely and in most patients I would say unfortunately what you see is that it's also the other side of the cartilage because that goes against the metal it's really worn away so it doesn't really work and it's only delaying a total knee operation for months to probably sometimes years which is fine now what we are developing ourselves at this moment is a polymeric plug that you don't want to put in the metal but you want to put in the polymer that has properties more like cartilage and has the advantage that we think it's nicer for the opposing cartilage but also that you could do an MRI measurement afterwards so with the metal you cannot put your patient in the MR anymore an MRI is still the main imaging thing for orthopedic surgeons so they want to do an MR after the surgery and they cannot do that with the metal so that's why we put a polymeric thing inside well that is fun but I think it helps a lot of people especially middle aged or older so 50 years old would help those patients but not the very young ones so we are starting now a project where we try to take this all into account so we are starting a project where we want to make an artificial tissue that has these fibers inside and inside the fibers has a swelling polymer and now we recreate exactly that mechanical condition or the mechanical effect that determines the cartilage properties and that is a polymer that we are starting to develop at this moment with a project that was just granted a couple of months ago 2-3 months ago and we can start it after summertime hopefully so that is what we are going to do there are many advantages I can tell you all about that there are also some disadvantages and the main disadvantage is that it's very difficult to attach something like that to the underlying bone it's easier if you have metal you just hammer it in but if you have something that is flexible polymeric has all these fibers it's pretty difficult to attach it to the underlying bone that will be a major challenge for this project one of the main advantages is that the gel that you put in that starts to swell you could also put in cells together with the gel and cells are pretty bad at making new collagen fibers but what they can do pretty okay is make new project like this so as that gel that swelling gel that you can put in as that sort of wears away we hope that the cells that we put in would sort of take over make proteoglycans and fill that network with a natural tissue that is our future sort of goal hope that we get there but it's based on the idea that you need a hydrogel or a swelling material that tries to swell but can't and that gives you your low bearing properties and we're trying to do that also now in our group not now a group in the collaboration with other groups also I do have a question so what you're presented has been very nicely presented and it looks extremely easy of course it's not so would you mind commenting a little bit on the time of research that has been necessary just to arrive to the conclusion that you presented it all started with the PhD project of Wouter Wilson and that was I think in 2002 2001 when it started so let's say 15 years ago and we had at that time still these biphasic models or trifasic we had ions that could swell the tissue but we didn't really understand how it worked and Wouter was in a different project and he was looking at models and he took cartilage as a model very simple and I told him well your project is fine it looks perfect but the cartilage we really need to do something on cartilage because this is not how cartilage works in the way I explained and he was convinced and he said okay I'm going to dedicate my project to make this cartilage model for you let's see how far we get and so he did and he worked for four years on really making this basic cartilage model and it was step by step so we had the fiber reinforcement we had the swelling and then we saw that we needed also physical elasticity in the collagen fibers in the collagen fiber network which was another publication and then we found out and that's also something that I didn't show in this presentation that part of that water not all the water is just attracted by the proteoglycans but part of the water is going into your collagen network the collagen network itself also absorbs some water which is then shielded from osmotic pressure conditions and that seems like a very minor thing but it turned out that we needed to put that into a model to take that into account to really make a nice fit to all the experimental data out there together so the four fits that I showed you that was again a publication so it is really a step by step by step by step thing that we did to come to the model that we had then we had the mechanical model and the next I think six years or so were dedicated to validating it to only finding experimental data to come up with really difficult strange data and to corroborate that to the model to see how good the mechanics were so after five years of model development we had like five years or so of validation studies what we did is you can take cartilage from the bone put it in different saline conditions and see how much they sort of bend as a result what we did is you can cut the cartilage with the scissors or with a knife and see how the cartilage opens over time you can load the cartilage and start to shear it which is a very nice experiment if you ever have time to look at the experimental data of Itai Kohen from Cornell University he has movies on his website where he compresses the cartilage and goes like that and follows the college in orientation and you really see very nice behavior so we all these weird things we did in the lab or we took from the literature we compare our model to all of these things data from also Australia from the group of Asfin Tambaya which is Neil Broome's lab very famous group in cartilage he did really many strange experiments that nobody would think were useful we modeled them all and we were really close in all cases so we had to build really six years of a case for yes we have a good cartilage model in the meantime we started to apply it of course to understand how tissue engineering works and many other things but now in the last five years we started to dig into the damage stuff so we made this cartilage model which is again cartilage damage model which is again a step by step then you do basically a year or so of research of studying to see where you have to put your limitations your thresholds for damage to occur in college and we still don't know for proteic ligands but it doesn't matter it seems that it doesn't matter where you put the thresholds basically get the same results anytime and now we are going into so last year we went into the strain rate dependency and now at this moment we are trying to put another damage mechanism, damage model inside and extend it with a second way in which damage may occur which is not the breaking of the collagen fibers but more like you have this network of collagen and it could also destructure without breaking so if it's entangled then it is basically lower than where it disentangles and it can swell a little bit but it's not really broken the fibers and we are now putting also that into a model so I guess that for the next five years or maybe even longer we will also be developing our damage model and then there will be a very long time in which we have to validate it and in the meantime making a step towards whole joint models which we are already starting and simulations of what happens in humans but it's really really really a tedious thing thank you for the talk I have two questions basically one is regarding the model how important would be to consider a patient specific geometry the patient specific geometry we don't know yet we have had a study which was last year where we started to look at effective different geometries where we have MR data and CT data from patients and we use that to create different models for these patients we thought that especially the roundness of the joint so if you have a very narrow joint or a wider joint that will make really a lot of difference for the way which damage would develop we look at experimental data or clinical data basically and we look at what the model predicts it seems that it doesn't make that much of an effect so that much of a difference so at this moment we say that maybe we don't need to do the patient specific modeling but what makes a lot of difference is where the damage is in the patient so if you have a damage really where the tibia and the femur are in contact really the damage is here or there or at the side and that makes a lot of difference for what is happening next where the initial damage is may not be so much controlled by the shape of the joint but it could be by the accident itself so whether you get that damage because you had a skiing accident or whether someone kicks you at football that determines where you have the initial damage and not so much the shape of the joint we believe at this moment but then where it is that makes a lot of a difference for the progression of the damage so it seems that the shape of the joint is more important for the progression and not so much for where the damage starts at this moment our focus is on the initial part our research on the initial part of the damage and when we have to move on to really progressing to serious damage then we need to account the patient shape but until then I think it's rather unimportant to do that that's the answer and the second one might be regarding the damage so do do you do you consider that maybe cell activity or nutrition might have an effect on damage? yeah for sure we know that some treatments also so if you look at the decision tree for instance for what surgeons will do with your damage then the first thing is they look at age or it's very early on the decision tree because if you're younger than like 35 years they will do sort of regenerative therapies they will drill holes in your bone hopefully there is some bleeding with the bleeding you get some stem cells and these stem cells may make new cartilage if you're over 40 years then these stem cells don't do much their cells are rather becoming rather inactive and they do a different kind of treatment they don't rely on the regeneration of the tissue itself so age is very important because your cells are more active or become less active that also means that if you look at future therapies like the one that I explained to you on the two questions ago that will also depend on the age of the cells the other thing is nutrition so the cartilage cells can do without much oxygen but a lot of the nutrition a lot of the oxygen and the nutrients they diffuse into the cartilage but also because you compress your cartilage and the cartilage relaxes with every step that you take and over a diurnal cycle also there is fluid flow going in and out and that also takes a lot of the transport of nutrients but also of growth factors and larger molecules so again their fluid flow is very important and how the cells respond to that how capable they are of responding still is very important so yes age and the quality of the cartilage of the patient is playing a significant role Hi, thank you the damaged images that they had this central spot of damage make me think about the intervertebral discs that they are also cartilage so can you relate your model to this specific histology with two parts that the vertebral disc have? Yes, yes so we look at intervertebral discs also of course they have a nucleus where the proteoglycan content is much more pronounced so that swells much more but they have an annulus around it that sort of is around the whole nucleus where the collagen fiber network is much more pronounced than the proteoglycans now we use exactly the same model here because it's composition based you basically say on the outside you have more collagen less proteoglycans inside you have more proteoglycans less collagen and now you have created this intervertebral disc you add the collagen fiber orientations to the model you're done so we use exactly the same model there and we also use it to look at damage development in the tissue but more not so much on herniations of proteoglycans that go through the collagen because then you really have to model also the damage and fissures and openings we don't do that at least yet but what we do model is that over time in degenerated discs we know that the nucleus, the proteoglycan with tissue that you get less proteoglycans there and that it sort of fibrillates over time so what you can do is you can compare models where you have a high proteoglycan content with models where internally in the center the proteoglycans are less and there's more collagen coming in in that place which means that they are less flexible that you have less motion with the same of your discs in all directions compression, torsion bending in these discs because they basically behave a little bit stiffer and that then in turn affects also the adjacent discs which will then have to be more loaded and they could then become damaged and etc etc so we're using the same model for interpretable discs thanks a lot for the talk I especially enjoyed your story about your PhD student coming with a model saying all the proteoglycans are damaged and you say bullshit it doesn't fit the data go back and then do the iteration and give the student a chance to prove you wrong I think this is really really important I think especially in modeling but in general I think an attitude of PhD students is like how do you think we should deal with that how should the PhD students deal with that because this student how did he deal or she I don't know what it was like how did he deal with the fact that the model might be wrong but still going it's like what's the dynamics behind it yeah that is one of the most interesting things of supervising PhD students I think but also of science and you should never really you should also be stubborn and trust yourself and trust your own data but you should always be open to the fact that you may be wrong and you should always be open to the fact that other labs may produce some different kind of data than you did now why is that why do they have something else could it be that your model is wrong and of course if it's wrong you have to change it and it is really important not to be really only focused on your own thing but to look out to read the literature to look at what other people do and to accept also that well you make a contribution to science but it's just a very little step and only together we can reach what we want to reach and so you have to be open to understand what other labs show so it could be that the model was wrong for the proteoglycan part we really didn't know what the threshold values were the proteoglycan part is really the most open part in the model on damage so it was the easiest part to discuss also with the student and it could be that now we have this deviatoric strain part that's going in proteoglycan damage it could be that it must that we should have compression or strain or maybe even flow patterns or that there's something completely different governing damage of the proteoglycans we don't know so it was the least certain and therefore also the easiest to discuss part but it was so persistent and I trusted my own experimental data he trusted his model maybe a little bit he over trusted his model he always did but still and so yes we had to find a solution there and this solution was that either he goes into the laboratory and do these experiments to prove that I was wrong but he has like I don't know he couldn't so I gave him a student to do it and then we sort of sat down and we said depending on what the student finds we basically together decide that the student is right so what comes out comes out is the agreement that I had but you have to really always in science stay open to anything that you find out in the literature any data and see how you could explain it either prove that the experiment was done wrong that's also possible that the experiment is completely wrong and that you don't have to trust it but then prove that the experiment is wrong if the experiment is not wrong and it's still producing something that you don't understand but could be important you have to go in and it helps, if you can explain in the end it helps you a lot in understanding what you're doing if you cannot keep it in mind and really do not throw it away just like that it is something that was measured and it is important somehow so it's indeed when something goes against your intuition that's where you really find the differences and also as you say like being critical as a student you have to be critical on your own work but as critical for your supervisor as you are from your own work but it's being this critical which I think is the most important any other questions? very much time for coffee now thanks again thank you