Review the similarities and differences between apoptosis and necroptosis, with an introduction to new tools developed to activate effectors of necroptosis.
Dr Andrew Oberst graduated from Amherst College in 2001. He pursued his graduate studies in Europe, in a collaborative program between the Universities of Rome and Paris, receiving his Doctorate in 2006.
Andrew completed his postdoctoral training at St. Jude Children’s Research Hospital in Memphis, TN, and in 2012 he joined the University of Washington’s Department of Immunology as an Assistant Professor.
Dr Oberst’s research focuses on understanding the pathways of programmed cell death, and how different forms of cell death are perceived by the immune system.
Hello. Welcome to Abcam's webinar on Going Out with a Bang: Understanding the Pathways of Necroptotic Cell Death. Today's principal speaker is Andrew Oberst, Assistant Professor in the Department of Immunology at the University of Washington. Andrew graduated from Amherst College in 2001, and pursued his graduate studies in a collaborative program between the Universities of Rome and Paris. He received his Doctorate in 2006, then completed postdoctoral training at St Jude Children's Research Hospital in Memphis, Tennessee. He joined the University of Washington's Department of Immunology as an Assistant Professor in 2012. Dr Oberst's research focuses on understanding the pathways of programmed cell death, and how different forms of cell death are perceived by the immune system.
Joining Dr Oberst today will be Miriam Ferrer, Product Manager for cellular assays at Abcam. Miriam completed her Biology Degree at the University of Barcelona, and has a PhD from the Vrije University in Amsterdam. After completing her PhD she joined MRC Laboratory of Molecular Biology in Cambridge. I will now handover to Andrew who will start this webinar.
AO: Thanks very much, Vicky, and thanks to Abcam for inviting me to give this webinar today. As Vicky said, I am an Assistant Professor at the University of Washington in Seattle, and this is a photo of me in the mountains near Seattle; it's a wonderful place to live. Unlike most of America, we're not digging out from a bunch of snow right now, so that's also very nice.
I'm going to talk today about cell death and, of course, cells are alive, we know this and they can die. The historical classical way that we've thought about cell death has been the dichotomy between two different forms of death: one, is this unprogrammed process of necrosis and the other is the programmed and very well-described process of apoptosis. When we say 'programmed' we mean a means for a cell to die on purpose, a form of cellular suicide is carried out by specific enzymes within the cell. This is a figure from way back in 1995, so this is a very, it's a classical view of cell death. Necrosis is the unordered killing of cells by a chemical or environmental damage, so essentially you can think of it as something like adding bleach to your tissue culture dish would induce lots of necroptosis. Freeze/thaw is another form of necrosis. Toxins can trigger necrosis.
On the other side of the coin, we have apoptosis and this is a programmed process, as I said, this is a form of cellular suicide that's carried out by and really defined by the activation of the caspase protease. So the caspases are a family of proteases, so protein cutting enzymes, there's 15 proteases that cut their substrate in aspartic acid.
Apoptosis is required for normal development, immunity and tumor suppression is the fundamental part of normal organismal function and of the cellular lifecycle. Several billion cells die in our bodies by apoptosis every day, and apoptosis is defined by these distinct hallmarks and these are, again, a function of the activation of these caspase proteases. When the caspases become active in a cell, that cell will shrink, its membrane will bleb and form these distinctive blistery features. The nucleus will condense and when this happens in vivo, these changes lead to the expression of find-me / eat-me signals on the dying cell, the dying cell is then rapidly recognized by a phagocytic white blood cell. This is depicted here in this artist's rendition of a phagocyte recognizing an apoptotic cell. For this reason, because of this distinctive program associated with apoptosis, we generally think of apoptotic cell death as a tidy and non-inflammatory way to eliminate cells from the organism. So this, again, happens many billions of times in our bodies every day, and that doesn't really cause much of an inflammatory or immune responsive.
I mention this dichotomy between apoptosis and necrosis, remember the classical unprogrammed necrosis, as a historical or classical way of thinking about this, because we now understand that there are other ways for a cell to die in a programmed manner; there are other cell death programmed. Two of those that are receiving a lot of attention now are pyroptosis, and this is mediated by different members of the caspase family and in response to the certain immune stimuli, mainly in macrocrophages. Then this other form of programmed cell death, necroptosis, which is going to be the focus of much of my talk today. Again, from a broad overview perspective, we generally think of apoptosis as a non-inflammatory silent form of death, and we think of these other forms of death as potentially inflammatory, or at the very least a crisis form of death that mainly seem to occur in response to infection or stress when a tissue needs to ring the alarm bells.
I'm going to talk first a little bit about the extrinsic pathway of apoptosis, and then how some of the same signals involved in apoptotic signaling can also trigger this other form of death, necroptosis, that I've just introduced. So extrinsic apoptosis is so-named because it's the activation of the apoptotic pathway by signals that come from outside of the cell. Notable among these is activation of the death receptors, Fas, TRAIL receptor and TNF receptor on the cell surface by the cognate ligands. When these receptors are activated in various ways, they can then activate - recruit and activate - and initiate a caspase-8 and I'll talk a lot more about caspase-8 in the later part of the presentation. Within this apoptotic pathway, caspase-8 activation can then go on to activate members of the Bcl-2 protein family. These proteins have the job of determining whether to, and then eventually going ahead with the permeabilization of the outer mitochondrial membrane. Permeabilization of this membrane allows the release of proteins that are normally sequestered inside that membrane. Among these is the cytochrome C; cytochrome C then binds to an adaptor APAF-1, and activates a full-blown caspase cascade, leading, as I said, to apoptotic cell death.
We can just show you some cells dying by apoptosis, and these are HeLa cells expressing a histone GFP to mark their nuclei, and they've been triggered to undergo apoptosis by a combination of TNF and cycloheximide. This is a very classical apoptotic stem, and you can see here the cells shrink and they bleb; they have the very typical blistering morphology. Then much later, they undergo what we call secondary necrosis where their membranes actually lose integrity. But, as I said, normally in vivo we would expect that these dying cells would be recognized and phagocytosed, taken out by a phagocyte long before they lose membrane integrity, so these cells would be rapidly eliminated and this would be a non-inflammatory process.
Of course, in the process of studying all of these pathways, many knockouts have been made of different members of the pathway. If we think of it as I've introduced it, apoptosis as a way to eliminate cells from the body, you'd expect knocking out full apoptotic enzymes or proteins would lead to phenotypes associated with too many cells, right? You get rid of apoptotic, the apoptotic program and you end up with too many cells.
Usually, this is the case, so if you knockout, for example, Fas, which is one of the receptors responsible for the extrinsic pathway of apoptosis, the phenotype you get is excess immune cells and autoimmunity. If you knockout some of the pro-apoptotic Bcl-2 family you get, again, autoimmunity, in some cases mice that are more prone to tumors, again, associated with an overabundant distal. In certain cases, very interestingly, you get mice with webbed feet and it is because the interdigital space in mice, as in humans, it's sculpted by apoptosis. So if you eliminate some of these Bcl-2 family members, you actually can't sculpt to the digits correctly and you end up with partially webbed feet. If you knockout some of these downstream executioners like caspase-9 and APAF, you get mice that have exencephaly that actually have a brain that extends outside their skull because they have an overabundance of neurons, and the skull can't close around the brain. This, obviously, is embryonically lethal to these mice. So in all these cases, knockout of pro-apoptotic genes and proteins leads to too many cells, as you would expect.
There's an interesting exception to this rule, and that's caspase-8. So remember caspase-8, this initiator of caspase, is activated by the death receptors. If you knockout caspase-8 unexpectedly you get embryonic lethality at around developmental day 11. This seems to be due to a hematopoietic defect and, indeed, the conditional loss of caspase-8 in the lymphocytes, so if you delete caspase-8 in lymphocytes only, using Lox-Cre technology, you end up with T-cells that can't proliferate normally in response to antigens. So, again, this appears to be a growth or a proliferative defect that's manifested by the knockout of caspase-8. So that's interesting and slightly unexpected, and it led us to, it led me during my postdoctoral work to try to address this question: Caspase-8 uniquely among these apoptotic proteins seems to be required for normal development in lymphocyte activation, why is this the case? We now understand why this is the case, and the reason that the caspase-8 is required for these events is that its pathway data is also involved in the suppression of a different form of cell death: necroptosis, which I've already mentioned, also sometimes called programmed necrosis.
The way to program necrosis or necroptosis - and you can use the terms interchangeably - the way that this form of cell death is initiated is again in response to some of these death receptors. But it's activated by a family of kinases, the RIP kinases, receptor interacting protein kinases, and the way that this generally works is that activation of the TNF receptor leads to the recruitment of RIP1. RIP1 participates in NF-kappaB signaling downstream of the TNF receptor, and this is the normal function of the TNF receptor is to activate NF-kappaB and promote an inflammatory cytokine response. But RIP1 can then translocate out of this TNF receptor complex to form a second complex in the cytosol. In that second complex RIP1 can interact with another member of the RIP family, RIP3, and this interaction leads to the recruitment and phosphorylation of a pseudokinase called MLKL, which then translocates to the cell membrane, permeabilizes the cell membrane and leads to cell death. This whole process is suppressed by caspase-8, and so in the absence of caspase-8, when you knockout caspase-8, you get unchecked TNF mediated RIP1- and RIP3- dependent cell death, and this leads to organismal death in the case of the knockout mice.
So I just want to show you a picture of some cells undergoing necroptosis to give you a sense of what they look like, and how this form of death is distinct from apoptosis. Here are, again, some HeLa cells that have now been stimulated with TNF in combination with a caspase inhibitor. Remember, you need to inhibit the caspases to unmask this necroptotic cell death response. You'll see that they're going to die in a very different way to the apoptotic cells, they're going to swell and then actually pop like little balloons. Pop. So you can see this very clear and very distinct morphology associated with RIP1 and RIP3 activation, and how different it is from apoptosis.
Here is a little schematic of how we think of these things, these pathways of working, activation of the receptors can activate caspase-8, which can lead to apoptosis. But the same caspase, caspase-8 can also suppress this other form of death, which can also be activated by the same receptor by blocking RIP1 and RIP3 activation. When caspase-8 is absent, RIP1 and RIP3 lead to this programmed necrotic cell death.
This can be demonstrated very clearly by the fact of the embryonic lethality of the caspase-8 knockout mice, is fully rescued by knocking out RIP3. So the caspase-8 RIP3 double knockout mice are born normally, and they survive and are mostly okay. Furthermore, as I already alluded to, combining TNF receptor stimulation, so activating this receptor in the presence of a caspase inhibitor such as zVAD. So blocking the caspases will lead to the activation of RIP1 and RIP3 mediated necroptosis. This has become a classical way to trigger necroptosis in cells, TNF in combination with zVAD will very nicely block apoptosis and instead sensitize those cells, to necroptotic cell death which depends on RIP1 and RIP3.
Here's a slightly broader view of the necroptotic pathway, and I told you about how the death receptors TNF, TRAIL receptor and Fas can activate necroptosis by RIP1. TLR stimulation is regular TLR3 and TLR4, which can activate the adaptor molecule TRIF, can also trigger necroptosis. Very interestingly, a viral infection can also directly activate this form of death. In the case of DNA viruses this appears to go through the DNA-centered DAI. In the case of other types of viral infection it's not totally clear how this pathway is engaged, but, interestingly, RIP3 knockout mice, so mice that don't have the necroptotic pathway are very sensitive to a number of different types of viral infection. So it's likely that, in fact, engagement of this pathway by viruses is one of the key physiological functions of necroptosis.
I just want to show this to note that what I'm going to be talking a lot about TNF-induced necroptosis, there are many other types of stimuli that can trigger this form of death. As I've already told you, TNF can also - and these other death receptors as well - can also activate NF-kappaB-dependent transcriptional responses and, of course, the TLR stimulation as well of the viral infection can also lead to these types of transcriptional cytokine responses. Furthermore, I've already mentioned apoptosis and how caspase-8 can trigger apoptosis, and, of course, all of these stimuli in certain conditions can also trigger apoptosis, right? These aren't unique necroptotic stimuli, depending on the engagement or in addition of the caspases, these same stimuli can also in some cases trigger apoptotic death. Furthermore, as I already mentioned, the caspases can suppress necroptosis, so this gets very complicated in a hurry. There are all these different stimuli and depending on what's present in the cell, cytokine responses, necroptosis or apoptosis can be engaged.
So we were really interested in studying this complex directly, RIP1 and RIP3, and how RIP1 activates RIP3 to regulate the engagement of necroptosis. We, of course, asked the question: how can we isolate and study these proteins without all of this very complicated other signaling habits? So we turned to a strategy that's been used for the study of the apoptotic caspases, which is inducible dimerization. So the way that this works in the case of the caspases, is that you simply fuse this inducible dimerization domain, shown here in orange, onto your caspase of interest. So you make a chimeric protein that you can then express in your cells, and then you add this dimerizer, a small molecule, that hobbles around rapamycin, to these cells and that will stick two of these dimerization domains together, induce the dimerization of your caspase and lead to its activation and apoptotic death.
We wondered what might happen if we took a similar approach to the activation of RIP3? What if we use a dimerizable RIP3, could we induce cell death and might we be able to uncouple the activation of RIP3 from these very complicated upstream effects? We undertook a very straightforward experimental strategy, we created a fusion protein of RIP3, composed of RIP3, fused to one of these FV domains, these inducible dimerization domains. We expressed this protein in 3T3-NIH cells, which lack endogenous RIP3. Then we added our dimerizer, which we're calling AP1, in this case, to the cells and we simply asked the question, do they die? Very importantly, I want to point out that RIP1 and RIP3, these kinases that trigger necroptosis, normally interact via this domain called the RHIM domain. This is a small domain in the C-terminus of RIP3, and RIP1 also contains a RHIM domain. These RHIM domains have previously been described to form these oligomeric prion-like structures upon activation. So these RHIM domains are sticky little domains that can stick many of them together into these larger prion-like structures that seem to be associated with the activation of RIP1 and RIP3. This will be important, because we're going to be mutating the RHIM domains in some of these constructs, so I just want to point out what the RHIM domain does.
To study this, we made use of a very handy piece of equipment that we have in the lab called the IncuCyte imager, and the IncuCyte is essentially an imaging platform that sits inside a tissue culture incubator, and allows us to image and quantify cells over time and get these very nice cell death curves. So this allows really great kinetic precision of how many cells are dying over time. So what we have here is simply some RIP3 expressing cells, treated with either TNF alone, which normally doesn't trigger a tremendous amount of cell death, it's normally inflammatory cytokines. Or, alternatively, treated with TNF in combination with the caspase inhibitor zVAD, and remember that when you inhibit a caspases in combination with TNF you sensitize cells to necroptotic cell death. So you can see this is just a time course of these cells dying, and when they die they take up a cell impermeable nuclear dye that we put in the media, and they fluoresce very brightly green when they lose membrane integrity. So we'll just play it on a loop. We can count these cells as they die and so at very nice half-hour intervals here we're quantifying how many cells are dying in this population, and so we get these very nice curves. So I'm going to be showing lots more of these kinds of cell death curves throughout the talk. I just want you to understand how they were generated.
We went ahead with our experimental strategy, as I've already mentioned, and we took our RIP3 protein fused with a dimerization domain. We expressed this in cells that don't have any endogenous RIP3. Then we added dimerizer and we were pleased to see that we did, in fact, get to the cell death and this was, we were able to confirm through a number of different ways that this was dependent on the kinase activity of RIP3, the downstream effector MLKL and so on. We got this somewhat limited effect, about 30 to 40% cell death when we added our dimerizer, but this told us that we could seemingly activate RIP3 via dimerization. Now, something that we were a little bit surprised by was that when we mutated the RHIM domain of RIP3, so we'd inserted, put a small tetra alanine mutation into the key activation point of the RHIM domain, we now completely abrogated our ability to activate this construct by dimerization. So, now, remember that this is - we're activating by addition of our small molecule drug, and dimerization of this FV domain, so this was a little bit surprising to us. We didn't expect that that form of activation would require the RHIM domain of RIP3.
As I've mentioned already, the RHIM domain has been described to form oligomeric structures when RIP is activated. We were able to resolve oligomers induced by dimerization of these constructs by DSS crosslinking, so we essentially took cells, activated RIP3 in them and broke them open, added this crosslinking reagent, and then ran them out on a western blot. We were able to track very nicely these RIP3 monomers, the formation of dimers and in the context of the RIP3-1xFV, the version of this that has an intact RHIM domain, we got this very nice formation of oligomers. When we used the version that lacked the RHIM domain, no oligomerization was observed, so we got a nice dimer but no oligomers.
What we think is happening here is simply this, when we dimerize RIP3 we stick two molecules together, but in the absence of that RHIM domain there's no further affects than the simple dimerization of RIP3 doesn't lead to cell death. On the other hand, when we dimerize RIP3 with an intact RHIM domain, we're then able to recruit additional molecules of RIP3 via these RHIM-RHIM interactions, and build these nice RIP3 oligomers which seem to be required, this oligomerization seems to be required for the activation of necroptosis.
So that was very nice, and that made some sense. We thought that we would see what would happen if we added some of the inhibitors that we normally used to modulate TNF-induced necroptosis. So, remember, TNF alone doesn't really trigger cell death unless you combine that with zVAD, and you can just see that here, TNF treatment alone, no death. You add zVAD, a caspase inhibitor, and now you unmask this necroptotic pathway.
To our surprise, however, when we dimerized RIP3, so when we formed these oligomers using this dimerization system that I just outlined. In this case, as I said, we get limited cell death, but if we add zVAD to these cells, we now greatly potentiate the cell death response. So, remember, there's no TNF involved here, this is not a receptor mediated activation of RIP3, it's activation of RIP3 directly in the cytosol using our small molecule. Also, what's somewhat confusing to us, when we did the same thing, dimerize RIP3 in the presence of necrostatin, which is an inhibitor of RIP1, the upstream activator of RIP3, we reduced the amount of cell death we see upon RIP3 dimerization. We were able to demonstrate that these effects are entirely independent of TNF or of receptor engagement, so what that seems to indicate to us is that RIP1 and caspase-8 are actually directly modulating the activation of RIP3 in the cytosol independent of TNF, or receptor engagement.
As I've mentioned already, the RHIM domain is required for the interaction not only between different molecules of RIP3, but also between RIP1 and RIP3. So we were able to show that when you dimerize RIP3 in the cytosol, so when you directly activate RIP3, you recruit RIP1 and caspase-8, so this is an IP experiment when we see very nice recruitment of RIP1 and caspase-8. Indeed, this is potentiated by zVAD, so when we add zVAD to these constructs we get more recruitment of RIP1 and caspase-8. We can see this again in the context of oligomerization here when we look at the formation of these oligomers, dimerization alone gives us some modest oligomerization. When we add necrostatin, the inhibitor of RIP1, that oligomerization is reduced. When we add zVAD, the caspase-8 inhibitor, that oligomerization is potentiated. So this matches very nicely with cell death responses that we saw in the last slide.
That led us, again, to think that perhaps RIP3 is directly modulated by caspase-8 and RIP1 in the cytosol, independent of this receptor signal. So remember that one of the things that I've just showed you is that the RIP1 inhibitor – necrostatin, blocks RIP3-dependant cell death. So we thought that this must mean that RIP1 is promoting RIP3 activation, right? You inhibit RIP1, you get less RIP3 activation in the cytosol, therefore, RIP1 is promoting RIP3 activation. In fact, what we saw was something a little bit different, so we wanted to, of course, make sure that necrostatin wasn't having an off-target effect.
So we did an experiment where we knocked down RIP1 using siRNA, and what we found, in fact, was that the inhibition of RIP1 by necrostatin and the knock down of RIP1 using siRNA, actually had opposite effects. Inhibition of necrostatin, of RIP1 by a necrostatin reduced cell death responses, while a knockdown of RIP1 greatly potentiated them. Again, remember, this is independent of any receptor signaling, this is just directly activating RIP3 in the cytosol. So this told us that the presence of RIP1 is not required for RIP3 oligomerization and activation. We hypothesized that perhaps necrostatin-1 actually acts by turning RIP1 into a dominant negative inhibitor of oligomerization. So, in fact, the inhibition we see here is not because RIP1 is required for RIP3 reactivation, but rather because necrostatin-inhibited RIP1 is blocking that phenomenon.
So this is a little schematic of how we think that might look. RIP3 dimerization we think can recruit RIP1, and once RIP1 is recruited RIP1 can perhaps mediate the suppression of further RIP3 oligomerization by recruiting caspase-8 and flip these suppressive members of the necrosome.
This led us to hypothesize, you know, we wondered, we thought this was very interesting, we kind of think that RIP1 is not only an activated RIP3, but it can also have this inhibitory role. We wondered where this might be important physiologically, and we hypothesized it in the absence of RIP1 and caspase-8 mediated suppression, RIP3 might undergo spontaneous oligomerization in the cytosol. That is to say, while RIP1 can activate RIP3 in some conditions, and our findings implicated RIP1 as a potential inhibitor of RIP3, independent of receptor signaling, just in the cytosol.
So to test this we turned to a different system, a shield-dependent destabilization domain, and so this is a small molecule, sorry, a small protein domain, that you can fuse onto a protein of interest, in this case RIP3, and this will lead to the rapid proteasome-dependent degradation of that protein under normal conditions. But when you add a small molecule ligand of this domain, it will now stabilize the protein fused to the domain. So you can see here these are cell expressing a destabilized form of RIP3, and when we add our shield drug into these cells, RIP3 accumulates over time. We wanted to use this system to test whether in the absence of RIP1, RIP3 accumulation without any additional activation, without any TNF stimulation or anything of that nature, simple accumulation of RIP3 might be enough to kill a cell if it doesn't have RIP1 present.
We first just did some knockdowns looking at TNF-induced cell death and we saw, as we expected there, when we activate with TNF and knockdown caspase-8, we get this great necroptotic response. A knockdown of RIP1 doesn't really have any effect on TNF-induced cell death, because RIP1 is downstream of TNF and we would expect it to be required for a lot of the TNF mediated function.
Interestingly, when we now accumulate RIP3 over time, so we now add our shield drug and just increase the amount of RIP3 in the cells and no TNF stimulation here, just RIP3 accumulating in the cytosol. We now get this very potent cell death response when we knockdown caspase-8, but also when we knocked down RIP1. So not loss of either RIP1 or caspase-8 now leads to this spontaneous receptor independent activation of RIP3 in the cytosol. This was interesting to us and it implies that RIP1 can actually block RIP3 activation independent of receptors, but we wanted to understand better how this might happen, how this inhibitory function of RIP1 might be working? So we hypothesized that RIP1 might cooperate with caspase-8 to prevent the formation and propagation of RIP3 oligomers.
So remember that I already told you that RIP3 oligomerizes when it's activated, and we hypothesized that maybe RIP1 can actually interpolate into those oligomers, recruit caspase-8 and prevent their formation. We reasoned that if this was the case, enforcing RIP3 oligomerization rather than simply dimerizing RIP3 and allowing it to oligomerize by RHIM-RHIM interactions, should remove RIP3 from the control of RIP1 and caspase-8. So if we stick enough RIP3 together, we should now abrogate the ability of RIP1 and caspase-8 to supress RIP3 activators. To do that we created these forms of RIP3 that have two dimerization domains on them, so rather than a single domain, which will simply dimerize RIP3 and allow RIP3 complexes to grow via RHIM interactions, we now put a second FV domain onto RIP3. Which will now have crosslinking of the RIP3 protein in the formation of large RIP3 oligomers, independent of the RHIM domain. We found very nicely that when we used these double dimerization domains, we greatly potentiated the cell death response. So, remember, before with the simple dimerization of RIP3 we were seeing about 30% cell death, now we get nearly 100% cell death, so when we oligomerized RIP3 directly.
Furthermore, very interestingly, we find that this oligomerization is RHIM-independent, so we use a form of RIP3 that completely lacks the RHIM domain, which, as you recall, when dimerized that form of RIP3 doesn't have any effect. If we force RIP3 into oligomers, now we get this very potent cell death response that's completely RHIM domain independent, right? So this chemically enforced oligomerization can substitute for those RHIM-RHIM interactions by adding a second FV domain, now this forces RIP3 into oligomer independent of the RHIM. Furthermore, oligomerization is both necessary and sufficient for RIP3 activation, so this really demonstrates very nicely that, in fact, you need to get oligomerization for cell death; because, remember, the single new version of this construct didn't kill it all.
We can probe this with western blot techniques, again, this is crosslinking oligomers and we see nice oligomer formation both with or without the RHIM domain when we have this 2xFV construct. The version with the RHIM domain interacts with RIP1 and caspase-8; the version without the RHIM domain did not.
But, consistent with our hypothesis, we find that the drug-enforced oligomerization of RIP3 completely removes RIP3 from the control of either caspase-8 or RIP1. So now when we add zVAD or necrostatin, which, remember, modulates the activation of RIP3 oligomers. When we add these drugs to our 2xFV versions, our enforced oligomerization versions of RIP3, we now see absolutely no effect. So, basically, when you force RIP3 of oligomer, RIP1 and caspase-8 are no longer able to observe any kind of control over the RIP3 oligomer. Which is consistent with the idea that the RIP1 and caspase-8 act at the level of RIP3 oligomerization to control RIP reactivation.
How do we think that this actually works physiologically? Let me step back and just give you a little bit of a model. RIP1, as I introduced it, is upstream of RIP3, so during normal receptor mediated activation of RIP3, RIP1 is ubiquitinylated and activated and there's a complex series of molecular modifications to RIP1. RIP1 can then recruit RIP3 and phosphorylate it, and this appears to potentiate the oligomerization of RIP3. Recruitment of MLKL and the induction of programmed necrosis, so during receptor mediated necroptosis RIP1 is activated, RIP3 is activated, MLKL is activated and you get cell death. It's the way we think of this as a relatively linear pathway.
But, remember those RHIM domains are sticky, they're oligomerizing a prion-like little domain, and so we hypothesized it in the cytosol of healthy cells, RIP3 can probably run into other modules of RIP3 with some regularity. When this happens, these RHIM domains can oligomerize and lead to RIP3 phosphorylation. When this happens in the absence of RIP1 or caspase-8, you would obviously expect to get cell death. So to prevent that, we hypothesized that RIP1 is recruited to these spontaneously formed RIP3 oligomers, recruits caspase-8 and FLIP and actually suppresses the formation of necroptosis, inducing RIP3 oligomers via probably the degradation of these complexes mediated by the cIAPs. Although exactly how that works is not totally clear. So in this model, what's interesting to keep in mind, is that RIP1 can act, and in some cases as an activator of RIP3. But, importantly, in other cases, and at a steady stage, RIP1 is actually suppressing RIP3 activation.
If you're interested in this idea, and if you want to see some really very elegant genetic models that are consistent with this idea that have also been published recently, you can, and I'll just steer you towards this very nice recent review by Ricardo Weinlich and Doug Green, published in Molecular Cell recently. These are former colleagues of mine, and they know what they're talking about, so take a look at that if you're interested in learning more about this whole idea.
So, with that, I would just like to close by thanking the people who did the work, and in particular Susana Orozco, a fantastic graduate student in my lab here at the University of Washington. I just want to acknowledge my collaborators and in particular Nader Yatim and Matthew Albert. Nader is a graduate student in Matt Albert's lab at the Institut Pasteur in Paris; they've been very close collaborators throughout this process. Igor Brodsky at the University of Pennsylvania, Sid Balachandran at Fox Chase Cancer Centre and Steve Tait at the Beatson Institute in the UK. With that, I will turn the microphone over to Miriam from Abcam who's going to talk a little bit about some of the Abcam products that they have available to study some of these pathways.
MF: Thank you, Andrew, for such an interesting and comprehensive presentation. We have received some questions, but I just want to remind you that you can keep submitting the questions to Andrew, to the Q&A panel located at the right hand side of your screen. So hello again, and I would like to take this opportunity to talk to you about some Abcam products that might be of interest to you. If you're interested in immunotherapies, we would like to invite you to our upcoming webinar on February 3rd, hosted by Dr George Prendergast from the Lankenau Institute. This one-hour webinar will provide a novel summary of the future of immunochemotherapy, focusing on the convergence of cross-disciplinary themes. As with all Abcam webinars attendance is free, but registration is required.
Let's now move on to necroptosis. I would like to make you aware of the range of RabMAb antibody products for necroptosis markers. A few of the markers available at the moment include RIP, Caspase-8 and MLKL. RabMAb antibodies are rabbit monoclonal antibodies with high affinity and high specificity, and have been extensively validated in multiple relevant applications such as western blot and IHC, as well as in several species. Our scientists are currently working on bringing to you soon all the RabMAb products against RIP3 and its phosphor-counterpartner, and also a specific mouse MLKL antibody.
I would like to highlight our RabMAb antibody against phosphorylated MLKL serine 358. This is a highly specific antibody against the human protein, and it's ideal for western blot and IHC experiments. On the publication highlighted on the right in Molecular Cell last year, Wang et al. described how the antibody was developed and how they studied the specificity of the antibody.
If you are looking for general cell death markers that can be used to detect both apoptosis and necroptosis, we have a range of kits that might be useful for you. If you want to detect caspase-8 activation, we have a range of products available. You can detect activation by flow of fluorescent microscopy using a FITC or rhodamine staining, or through a simple microplate reading experiment using colorimetric or fluorometric detection. If you want to easily differentiate between live and dead cells, our simple assay for live and dead cells can give you the answer in just 10 minutes. You can visualize the cells by flow cytometry or in a fluorescent microscope.
Other products for assessing cell death that you might find useful are assays to detect changes in the mitochondrial membrane potential. There are several dyes you can use such as TMRE, JC-1 or JC-10. All these dyes are very easy to use and quick to visualize. The image on the slide shows the staining of mitochondria with PMRI, which can be used in both adherent and suspension cells.
When studying cellular pathways it is essential to use the right tools, so we have a range of inhibitors and activators of necroptosis pathways that you can use to ensure that you are looking at the right cellular activity.
As you have seen from Andrew's presentation, there are several necroptosis markers which are apoptosis markers as well. If you are interested in a quick and simultaneous detection of several of these markers, why not use an antibody array to quickly check whether caspase-8, Fas or TNF-alpha have been activated in your samples. You can detect up to 43 markers at the same time in four to eight cell or tissue lysates.
So thank you for your attention, and without further delay I will pass the microphone back to Andrew who is ready to answer your questions.
AO: Thanks very much. We've had several questions come in and I think I can get to the answers. One of the first questions, and I think this is a great one, is does necroptosis cause inflammation in vivo? It's a great question and we - as I presented, necroptosis and compared to apoptosis, I'd say that necroptosis is an inflammatory form of cell death. The problem is that it hasn't really been easy to demonstrate that, so there aren't yet, and we're starting to develop them, but there aren't great tools to really detect necroptosis in vivo. So one of the only tools that we've had up to now has been with knockout mice, and you can take a RIP3 knockout mouse and you can give it the same stimulus as you'd give a wild-type mouse and see whether it has a different response to see whether the inflammatory response is reduced when you eliminate the necroptotic pathway. That seems to be the case with some stimuli so, for example, a TNF-induced sterile septic shock if you just inject a bunch of TNF into a mouse it will get very sick, and its body temperature will drop and it'll eventually die.
That response seems to be partially abrogated in RIP3 knockout, so that's a condition where it seems that part of that inflammatory response does depend on RIP3. Now, the challenge there is to discern whether all of that effect is due to necroptotic cell death, as opposed to potentially some other functions of RIP3 that maybe we don't understand yet. Could RIP3 have a cell path of a transcriptional role, for example? So just now with some of the reagents that Miriam has described, we're now starting to be able to directly look at more necroptotic cells in things like tissue sections. So using, for example, that phospho-MLKL antibody, so it's a good question, but not one that I think that we really firmly know the answer to yet, despite many reviews that present necroptosis as inflammatory.
Let's see, another, just a sort of technical question: What kind of dye was used to detect necroptotic cells? That's a dye called SYTOX Green, it's a cell-impermeable DNA dye, and very important you'll want SYTOX with an X at the end, S-Y-T-O-X. There's another version of that same dye that's cell-permeable that's called SYTO Green, S-Y-T-O, and that's the best dye that we've found for these types of assays, it's incredibly bright and it really lights up cells once they die.
Another question here: If my microarray results show upregulation of RIP3 and MLKL, can I assume that the pathway of cell death is by a programmed necrosis, or do I need other genes for verification? That's a good question, and I don't think that you can assume that. Both RIP3 and MLKL are interferon responders, so interferon stimulation will upregulate RIP3 and MLKL, but it won't necessarily trigger necroptosis. In certain conditions interferon signaling can trigger, or at least sensitize cells in apoptosis. How exactly that works, again, it's not totally understood, but I think that the upregulation in the genes is not enough to determine that a cell is undergoing necroptosis. You need to obviously determine whether your cell is dying, and if your cell is dying then something like a phosphor-MLKL reagent, or using a knockout for RIP3 or MLKL would be the best way to really determine that necroptosis is at play in those cells.
A question here: I only recently discovered the existence of necroptosis, so it's quite new to me and I wonder what are the possibilities of detecting it in paraffin tissue sections? Right now, really, the only option is the phospho-MLKL antibody that Miriam's talked about, that's as good as we have right now. Hopefully, in the near future, Abcam will be able to give it something like a phospho-RIP3 antibody that will work really well in paraffin tissues. As far as I know, the phospho-MLKL is really the best verified reagent at this point. Notably, it only really works in human tissue sections, so that's a challenge. But, again, with the diligent work from the folks at Abcam, hopefully more reagents to study these pathways will become available.
Another question here: Does necroptosis happen in the brains and neurons? I don't really know the answer to that question. The brains and neurons do not seem to express very much RIP3, at least at steady state, but whether in some disease conditions that the pathway might be engaged, I don't really know.
Another question here: Can necroptosis be induced independent of RIP kinase activation? I think that I would say that the answer is no, but it depends how you define necroptosis I suppose. I mean, necroptosis was initially described in studies as the context of the RIP kinase, as a sort of a defining characteristic of necroptotic cell death. So RIP kinase independent necroptosis almost seems like a contradiction in terms. With that said, it certainly is possible that there are other ways to activate MLKL, the downstream effector of necroptosis, independent of the RIP kinases. How that might work is not clear, but you certainly could get a programmed form of death and it depends on MLKL, but it's independent of RIP3. To my knowledge, that hasn't been well-described at this point.
Do all cells express RIP3? No, they do not. RIP3 is highly expressed during embryonic development in many cell types, but many cells down-regulate RIP3 as adults. Hematopoietic cells consistently express a good amount of RIP3 and, as I've already mentioned, RIP3 is interferon inducible and so it can potentially be upregulated in some cells. Importantly, very importantly, for the study of this pathway, many of the tissue culture cells that we routinely use lack RIP3. So, for example, HeLa cells, 293T cells, HCC1-15 cells, all these sort of standard, heavily-used laboratory cell lines have lost RIP3 expression. Why that's the case is a separate question, but if you just take your normal HeLa cells and try to induce necroptosis in them, nothing will happen. So there are a couple of commonly used cell lines in addition to hematopoietic cell lines like Jurkat cells. You can also try HT-29 cells, that's the human cell line and it does express RIP3 and it can be induced under necroptosis. But just check in the literature before you start trying to do necroptosis studies on any old cell line, because many of them have lost RIP3 expression.
Can you differentiate between pyroptosis and apoptosis in vivo or in vitro; how can you measure both? I would argue that you can't easily differentiate between apoptosis and pyroptosis. They're both caspase-dependent, these quote unquote specific inhibitors of the different caspases tend to not actually be very specific at all. I think that using knockout animals or now with the advent of CRISPR technology, knocking out specific effectors of the pathway is really your best bet for differentiating between those two forms of cell death.
Is it known how PLR3 or 4 binding leads to activation of RIP3? Yeah, the downstream adaptor protein TRIF has a RHIM domain, so it has the same domain that activates RIP1 and RIP3, and so TRIF activation can directly activate either RIP1 or RIP3, interestingly. So that can be RIP1-dependent or RIP1-independent, depending on what's going on in the cell.
Is there any less cell death when using IETD-FMK rather than Z-IETD-FMK? So IETD-FMK is a specific inhibitor of caspase-8, and we observe comparable amounts of cell death using either of those inhibitors. We also have done that by knocking down caspase-8 that’s not so greatly sensitized in that case. So that all points to it being a specific effect on caspase-8. With that said, zVAD can be somewhat aspecific, so something to keep in mind.
Finally, I think this is the last question: Is there evidence with the recruitment of the RIP3 oligomers to the bad complex by RIP1, rather than monomer or dimer RIP3? What is known about the MLKL downstream targets and is possible to get a good predictor? Oligomerization does need to be required for the potent recruitment of RIP1 and caspase-8, so I think you do need to oligomerize RIP3 to get that complex to really form. Downstream MLKL targets are currently described as being the membrane, so MLKL can directly translocate to the membrane and poke holes in it essentially. So whether that - my guess is that there are some other factors involved, it's not just MLKL in the membrane, that there are probably some modulators of MLKL activation at the membrane, but that has not yet been well-described. Phospho-MLKL, as I said, is about as good as we have now for MLKL activation, so that's the best reagent that has been developed to date. Hopefully, some better or additional versions of those types of reagents will be developed shortly.
A couple of final late-breaking questions here. Is there a relation of necroptosis signaling with calcium? If there is one I don't think it's well-defined. Necroptosis - yeah, I guess I would have to take the fifth on that one.
Finally, did you experience an additive response when you stimulate membranes with membrane receptors as well? So I guess this would be in the context of the dimerization studies. When you stimulate a membrane with membrane receptors, you get pretty limited cell death anyway. When you stimulate with membrane receptors in combination with zVAD, you get nearly 100% cell death with either the membrane receptors, but you also get 100% cell death with any of our activatable RIP3s. So the answer is essentially no, because when you induce necroptosis by either pathway, it's nearly 100% and so you can't really add to that. I think that if you titrated things carefully you could probably get some kind of additive effect there, but we didn't really explore that carefully.
Final question: Any relationship between mitochrondrial dysfunction and the RIP-kinase pathway? Probably, but it has not really been described terribly well. There's some indication that mitophagy, so the illumination of damaged mitochondria might be able to activate RIP3 in some context, but how that might work is not clear. So that's it, I think I've covered most of the questions here, and I thank you all again for your attention, and I thank Abcam for giving me the opportunity. Please feel free to contact me if you have any further questions or comments, or concerns.
Thank you, Andrew, and thank you Miriam for your presentation. We had quite a lot of questions today and there are still questions coming in as we speak, so for those questions that are not answered, our scientific support team will contact you shortly. If you have any questions about what has been discussed in this webinar or have any technical enquiries, our scientific support team will be very happy to help you. They can be contacted at firstname.lastname@example.org. We hope you have found this webinar informative and useful to your work. We look forward to welcoming you to another webinar in the future. Thank you again for attending and good luck with your research!