Session 4: Molecular View of Adaptive Immunity
Transcript of Part 2: The Immunological Synapse: Signaling and Function
00:07.3 Hello. I'm Michael Dustin, 00:09.3 from the University of Oxford 00:11.1 and New York University School of Medicine. 00:13.2 Welcome to the talk on immunological synapses, 00:16.1 Part 2 - Signaling and function. 00:20.1 So, this is a movie 00:22.1 of pre-immunological synapses forming 00:25.1 and you can see this 00:27.2 bullseye-like pattern 00:29.2 emerging from an initial less organized or even inverted pattern. 00:35.0 So, this is the process of immunological synapse formation. 00:38.1 We want to... 00:40.1 we'll define this a little bit more generally, 00:42.3 and then I will go on to talk about signaling, 00:45.1 effector function mediated by the immunological synapse, 00:48.0 signal amplification at the immunological synapse, 00:51.2 and applications to autoimmunity and cancer therapy. 00:56.2 So, in terms of definition, 00:58.2 the concept of a synaptic basis of T cell activation 01:02.1 basically went back to the 70s and early 1980s, 01:04.3 and there's a review by Mike Norcross in 1984 01:08.0 that describes a synaptic basis of T cell activation, 01:11.1 Bill Paul and Bob Seder 01:13.1 used the term in 1992 in a cell review, 01:15.2 but the first appearance in peer-reviewed publications 01:17.2 is in the late 1990s 01:19.1 with work from Avi Kupfer 01:21.3 describing supramolecular activation complexes 01:24.2 and our work describing... 01:26.2 you know, in reference to Kupfer 01:28.2 and our own work 01:30.3 referring to immunological synapses 01:32.2 and that final bullseye pattern 01:34.2 of a mature immunological synapse. 01:36.1 So, what wrote was: 01:37.2 The immunological synapse is characterized by a specific pattern of molecules 01:40.3 in the contact; 01:42.0 LFA-1 is localized to the periphery of the contact"... 01:45.2 what Kupfer described as the pSMAC, 01:47.1 or peripheral supramolecular activation complex... 01:49.1 while the T cell receptor is localized to the center of the contact... 01:53.1 or what Kupfer defined as 01:55.3 a central supramolecular activation complex. 01:59.2 This is described in these references from '98, 02:01.2 and later on, 02:04.2 in collaboration with David Colman, 02:06.0 who is a neuroscientist, 02:07.2 we basically thought about a few criteria 02:10.0 that were common to the neural and immunological synapses: 02:12.2 first, that cells remain individuals, 02:15.3 which is part of the neuronal doctrine; 02:18.0 that there bona fide adhesion molecules mediating these junctions, 02:21.2 basically cadherins in the neural synapse 02:24.1 and integrins in the immunological synapse; 02:26.1 that these are at least provisionally stable junctions 02:29.0 -- obviously, in the nervous system, 02:30.2 synapses can be stable for years, 02:32.2 in the immune system 02:34.2 the cells are extremely motile in the steady state, 02:36.3 so when they stop for a few hours 02:39.1 that's quite a stable... 02:41.2 quite a contrast in their behavior 02:43.2 and a relatively stable junction, 02:45.1 although it's much less stable than what you'd see in the nervous system 02:48.0 in many cases; 02:49.2 and you have directed secretion taking place 02:52.0 across this interface, 02:53.3 and I'll talk about that a little bit when we describe effector functions. 02:58.0 So, in terms of signaling mechanisms 03:00.2 or the actual triggering process, 03:02.1 the T cell receptor, 03:06.0 which is, again, at the heart of the immunological synapse, 03:08.1 provides the immunological specificity, 03:10.1 utilizes what's referred to as non-receptor tyrosine kinases, 03:12.0 so it utilizes a couple of... 03:14.2 several different kinases, 03:16.2 but the first two are the Src family kinase Lck 03:21.2 and the Syk family kinase ZAP70, 03:23.1 or zeta-associated protein 70, 03:24.2 as the major signal initiators 03:27.2 which are recruited to phosphorylated cytoplasmic domains 03:32.1 of this multisubunit complex. 03:36.0 This signaling process is modulated 03:38.1 by a number of other proteins, 03:39.3 and this paper published a couple of years ago 03:42.2 from James and Vale 03:45.2 reconstituted this process 03:48.1 in non-lymphocytes 03:50.1 using a combination of the T cell receptor, 03:52.1 the kinases I just described, 03:54.0 and then the modulators 03:56.1 CSK, a kinase that actually 03:59.1 inhibits Lck by phosphorylating it, 04:01.1 CD45, a phosphatase 04:04.1 that antagonizes the CSK-mediated inhibition, 04:07.3 and together this system 04:09.2 would basically set up a regulated basal state 04:11.3 from which T cell receptor engagement 04:13.2 would trigger signaling, 04:15.1 so this seemed to be a minimal 04:17.2 set of molecules required 04:20.0 for reconstituting T cell receptor signaling 04:21.3 in a non-T cell. 04:23.3 But when we go back to T cells 04:25.1 and basically look at these molecules, 04:26.3 we can see that they're very highly organized 04:29.1 in the immunological synapse, 04:30.3 so a number of the experiments that I'm going to describe 04:34.0 utilize a reconstituted system 04:36.2 in which the antigen presenting cell 04:37.2 is replaced by a supported planar bilayer, 04:40.1 kind of schematized here, 04:43.0 which... again, a technology developed in Harden McConnell's lab 04:45.1 at Stanford, 04:46.2 where an artificial bilayer 04:49.1 is deposited on a glass substrate. 04:51.1 This is a very optically ideal system. 04:53.1 We can put purified molecules 04:55.1 in a laterally mobile form in the substrate 04:57.3 and T cells can assemble an immunological synapse. 05:00.2 We can then use optical methods 05:03.1 that are only applicable to this type of interface, 05:05.2 like total internal reflection fluorescence microscopy, 05:08.1 in which you're bouncing a laser beam 05:10.2 off of the interface 05:12.0 and creating this very shallow evanescent wave 05:13.3 that excites fluorescence 05:15.2 within about 200 nanometers of that substrate, 05:17.1 so it's very synapse-specific in this case 05:19.1 and, for example, 05:20.2 we can then look at signaling processes 05:22.1 like the recruitment of ZAP-70 05:24.0 to these T cell receptor clusters, 05:25.3 which are actually the active component 05:27.2 in the signaling immunological synapse. 05:29.2 So, you have this large central cluster, 05:31.2 which is actually not active in signaling, 05:34.0 but these small peripheral clusters 05:35.2 are the places where most the 05:38.2 ZAP-70 is being recruited, 05:40.0 and most of the other evidence of this tyrosine kinase cascade 05:43.1 is localized. 05:44.3 We can also see that these microclusters, 05:46.3 based on this work from Rajat Varma, 05:48.3 which was done in my lab, 05:50.2 where the... 05:52.3 if you look at CD45, 05:55.1 the signal here, 05:56.2 in relation to the T cell receptor 05:58.1 on these planar substrates, 06:00.0 with an MHC-peptide complex 06:01.2 that triggers this T cell, 06:03.2 that you, very early on in the contact, 06:05.1 or later, once you have this mature 06:08.1 bullseye-like immunological synapse, 06:09.3 that you have CD45 exclusion from these microclusters, 06:14.2 and this is where the ZAP-70 is being recruited. 06:16.2 So, while one hypothesis 06:20.0 for triggering through the T cell receptor 06:21.2 is that the local exclusion of these phosphatases 06:24.2 like CD45 06:26.1 could be critical to this, 06:28.1 and this is an idea that was first floated by Tim Springer 06:30.1 and has been expanded on extensively 06:32.2 by Simon Davis and Anton van der Merwe, but... 06:35.1 and certainly supported experimentally, 06:36.2 at least as a kind of a... 06:40.2 in a descriptive fashion, here. 06:45.1 So, another remarkable thing about 06:48.2 both the immunological synapse 06:50.1 and this total internal fluorescence microscopy 06:52.2 as a method to study it 06:54.2 is that we can detect single MHC-peptide complexes 06:57.0 and we can actually basically start to the study 06:59.2 the sensitivity of the T cell 07:01.2 and directly demonstrate this single molecule recognition process. 07:04.2 So, these are single MHC-peptide complexes 07:06.2 that are being captured, 07:08.0 they're diffusing around in this planar bilayer substrate, 07:10.1 and the ones that are immobile here 07:12.1 in this red area, 07:13.2 which is actually the T cell receptor, 07:15.2 are effectively being captured by the T cell receptor. 07:17.3 The T cell has about 50,000 receptors on its surface, 07:21.2 maybe over 10,000 of those are in this interface, 07:24.2 and those are there waiting 07:26.3 in a proper environment defined by the adhesion molecules 07:30.2 to very efficiently capture these molecules. 07:32.1 So, to point out that the LFA-1/ICAM-1 adhesion system 07:34.3 is in place in this system, 07:36.2 it's just not labeled in this context. 07:37.1 So, you'd only achieve this very high sensitivity 07:39.3 with the adhesion system in place 07:42.0 and actually, if we looked at it, 07:44.1 it would be forming this ring-like pSMAC. 07:47.1 There's not a large cSMAC in this case 07:49.2 because the cSMAC size is linearly related 07:51.3 to the MHC-peptide density 07:53.1 and at very low MHC-peptide densities you don't see much. 07:55.2 We'll talk more about that in talk 3. 07:58.1 So, with this single molecule, 08:00.2 you know, recognition process, 08:02.0 we looked for CD45 exclusion 08:04.2 to see if this CD45 movement from those sites 08:07.2 was seen even when the T cell receptor 08:09.1 was just seeing a single MHC-peptide complex, 08:11.0 and what you can see from this plot is that we did the... 08:14.2 well, we did these measurements 08:16.2 of where the single MHC-peptide complex was engaged, 08:19.0 and then at the point of engagement, 08:21.2 and then around that point, 08:23.0 and then basically did a local comparison, 08:25.1 set that to 1, 08:26.3 and then basically looked at basically 08:28.2 what was there... 08:30.1 err, the surrounding area was set to 1, 08:31.2 and then we'd look at what was happening in that central region. 08:33.1 And what you can see is that there's a lot of noise in these measurements, 08:35.1 but we can still see exclusion of CD45, 08:38.0 on the order of 10% 08:40.3 from these sites of single MHC-peptide engagement. 08:43.1 So, again, from a... 08:45.1 just kind of a phenomenological standpoint, 08:47.0 this CD45 exclusion 08:48.3 is happening at points where the T cell receptor is engaged, 08:51.2 even by single molecules. 08:53.0 We know, again, that signaling is being initiated, based on ZAP-70 recruitment. 08:57.2 So, one model for T cell receptor triggering 09:00.1 that accounts for the single-molecule sensitivity 09:03.0 is this CD45/phosphatase exclusion model, 09:05.3 again van der Merwe and Davis would be sources 09:09.0 to look for more information on that, if you're interested. 09:11.1 There's also... there are other ways to do this, though. 09:15.1 So, Art Weiss has demonstrated recently that 09:17.1 Csk inhibition can also trigger T cell receptor signaling 09:20.0 using a pharmacological agent. 09:22.1 So, either CD45, 09:26.3 as the sort of negative regulator of local tyrosine phosphorylation, 09:28.3 or CSK to regulate Lck 09:31.2 will achieve this rapid activation. 09:34.3 Balbino Alarcon and colleagues 09:37.1 basically have been studying 09:39.1 conformational changes in the T cell receptor 09:40.3 and evidence of that. 09:42.0 This is, again, 09:44.2 an ongoing area and no one has a structure 09:46.2 for a complete T cell receptor 09:48.1 with the signal transduction components, 09:49.2 so we don't know at this point if this is right, 09:52.3 but he has signatures of these conformational changes. 09:55.2 Kai Wucherpfennig and earlier, even, Larry Stern, 09:58.2 have described the... 10:00.3 kind of a conformation of the cytoplasmic domains 10:03.0 attached to membrane lipids, 10:04.2 and this modification of the sequestration by lipids 10:07.2 in the inner leaflet of the membrane 10:09.1 can also be used as a triggering mechanism. 10:10.3 And finally, Cheng Zhu 10:12.2 has come up with some very interesting data 10:15.0 suggesting that the T cell receptor, again, 10:16.2 in a kind of a conformation change mode, 10:18.3 acts as a catch bond 10:20.1 that in response to force 10:22.1 undergoes conformational changes 10:23.3 that both increase the strength of its interaction 10:25.3 with the MHC-peptide complex 10:27.1 and contribute to the signaling process. 10:30.2 So, again, there are a number of things going on in this area 10:33.2 and we don't... 10:35.0 really, the jury is still out on 10:37.0 what combination of those mechanisms, 10:39.0 or even unknown mechanisms 10:40.2 that people haven't thought of, 10:42.0 are actually acting to account for this high sensitivity. 10:44.2 So, this centripetal movement, 10:46.2 again, of the MHC-peptide complex, 10:49.3 the T cell receptor, 10:51.1 and adhesion molecules is very dramatic, 10:53.2 and this movie from Ron Vale's group 10:55.2 generated by Adam Douglas 10:58.2 using the Jurkat cell line 11:00.1 expressing fluorescent GFP-actin 11:04.0 beautifully illustrates this centripetal movement 11:07.1 using a method of speckle microscopy, 11:09.0 which is based on spinning disk confocal microscopy. 11:13.0 So, this is clearly an engine 11:16.0 for the formation of these centripetal patterns, 11:18.2 and the differential association of the T cell receptor 11:22.0 and the integrins with actin 11:23.3 probably then account for this bullseye-like patterning. 11:26.0 In talk 3, we'll also discuss how the ESCRT machinery 11:29.2 may be important for actually the final deposition 11:31.3 of the T cell receptor in the center. 11:34.2 So, effector function of immuno... 11:37.0 so, there are two kinds of... 11:38.1 I've talked about signal integration 11:39.3 and the signal integration 11:41.2 can accomplish transcriptional changes 11:43.1 and cell cycle regulation 11:45.0 that leads to the proliferative burst, 11:46.3 but once that proliferative burst is over 11:49.1 the T cells become effector cells 11:51.1 and they have to do things, 11:52.1 they actually can't just proliferate and things, 11:54.1 they actually have to kill targets or help B cells, 11:56.0 and this synapse also helps in this stage. 11:59.1 And this is a fantastic movie 12:01.1 done by a method called 12:03.1 laser lattice light sheet microscopy, 12:06.1 which was developed by Eric Betzig at Janelia Farms, 12:10.2 and this Ritter et al paper with Gillian Griffiths 12:12.1 and Jennifer Lippincott-Schwartz 12:14.0 beautifully shows T cells 12:16.1 with a green fluorescent actin probe 12:18.1 and orange markers of their cytolytic granules, 12:21.1 which contain agents that will kill this target cell, 12:25.0 and also the centriole 12:26.2 -- the microtubule organizing center -- 12:28.1 marked with this little dot here. 12:29.2 So now, if you watch this movie, 12:31.3 this T cell really goes after this target cell, 12:35.2 forms an immunological synapse... 12:37.3 you can see, with the clearing of actin... 12:40.1 kind of a bright actin structure at the periphery 12:42.2 and then a clearing of actin towards the center, 12:44.1 where those granules and the microtubule organizing center 12:46.1 dock right up to that interface 12:48.3 and deliver cytolytic agents that will result in, again, 12:52.1 death of the target cell. 12:54.1 So, this is a beautiful example of effector function. 12:55.2 Again, this was published this past year. 12:58.1 So, in terms of thinking about... 13:00.2 like, how important is it that you do that to kill a target? 13:04.0 So, we were able to do some experience 13:05.3 in collaboration with Yuri Sykulev 13:07.2 where we basically could kind of measure 13:09.3 the efficiency of the cytolytic T cell 13:12.0 by comparing how many granules it released, 13:15.1 which can be measured through a serine esterase release assay, 13:18.0 and how efficiently it killed the target. 13:20.2 So... and you can kind of use target killing 13:22.3 over granule release 13:24.0 as kind of the measure of efficiency, 13:25.2 and then set up situations 13:27.2 where the T cell either formed an immunological synapse 13:30.0 or had a broken immunological synapse, 13:31.2 so it could basically... 13:32.3 you know, pharmacologically repair or break the synapse 13:35.3 depending upon, you know... 13:37.3 experimentally, and what we found is that 13:40.3 basically forming an intact immunological synapse 13:42.2 improved killing about 3-fold. 13:44.0 And you can say, "Well, that's not that much...". 13:45.3 Some biologists have a cut-off at 5-fold, 13:47.3 but we actually think that 13:49.2 this could be very important in the context of, 13:52.2 say, killing in vivo, particularly tumor cells, 13:54.1 where it can take up to 6 hours for a T cell 13:56.3 to kill a tumor cell based on work from [unknown], 13:59.0 so a 3-fold change in efficiency 14:01.0 would take things from 6 hours to 18 hours, 14:03.1 and then, you know, 14:05.0 you're actually starting to deal with timeframes 14:07.1 that are relevant to the replication rate of the tumor cell. 14:10.1 So, if you want to keep up, basically, 14:12.1 it's probably good to have an intact synapse, 14:14.0 and that seems to be the case, 14:15.1 as I'll show you in a moment. 14:17.0 So, you know... 14:18.2 the other thing I wanted to talk about 14:20.3 is the actin, you know, 14:22.1 which again highlights 14:24.1 both this bullseye-like formation, 14:25.3 the formation of this junction 14:29.1 that the cytotoxic T cell uses to grab onto the target cell, 14:33.1 also amplifies the signal, it turns out, 14:36.0 from the T cell receptor 14:37.2 downstream of that early tyrosine kinase cascade, 14:40.1 and this is work from Sudha Kumari in my group 14:44.1 that kind of uses a similar kind of actin probe 14:47.1 to that used by Ritter et al 14:49.0 to follow T cell receptor microclusters 14:51.1 and then look at the, you know, 14:53.1 very specifically at the amount of filamentous actin at these sites, 14:57.1 and one way to basically 15:00.0 take these kinds of movies, 15:01.1 which can be a little bit difficult to deconvolve, 15:03.1 and illustrate events that are happening over time... 15:06.1 this view is called a kymogram, 15:08.0 which gives you essentially this region here, 15:11.1 this line effectively is the top, 15:14.0 for the T cell receptor and for the actin, 15:15.3 of this graph, 15:18.0 and then you basically just run that same region 15:20.1 at different times for both signals, 15:23.1 and what you can see is that within this very complicated pattern in the actin 15:27.1 you have very steep diagonals, 15:30.0 which means things are moving very fast, 15:31.3 they're doing a lot of movement in a short time, 15:34.1 but then these kind of more shallow... 15:36.1 these kind of somewhat larger angles 15:38.2 going down in this direction... 15:40.1 they basically reflect slower movement, 15:43.3 and those are tracking perfectly with the T cell receptors. 15:45.3 So you have these 15:48.0 foci of actin that are following the T cell receptor very closely 15:50.2 and what Sudha was able to show 15:53.1 is that these actin foci 15:55.2 were dependent upon a protein called 15:57.1 Wiskott–Aldrich Syndrome protein, 15:58.2 which is deficient in a primary immunodeficiency, 16:00.1 where you have B cell and T cell dysfunction, 16:02.2 and also platelet dysfunction. 16:04.2 It's basically an actin regulator 16:08.1 that regulates the so-called Arp2/3 complex 16:10.0 that makes branched actin networks, 16:12.1 and if you don't have WASP, 16:14.1 you don't basically form these actin foci 16:17.1 at the T cell receptors. 16:18.1 But a lot of your other actin polymerization is fine 16:20.2 and you can actually form a synapse just fine, 16:22.1 so that was a bit confusing initially, 16:24.1 but looking at that again, 16:26.0 at just that actin at those T cell receptor foci 16:28.1 was critical for understanding this. 16:29.2 And Sudha went on to show 16:31.1 basically that phospholipase C-gamma recruitment 16:33.1 is critically dependent upon these actin foci, 16:35.1 and phospholipase C-gamma 16:37.1 is a critical sort of, you know, 16:40.3 nexus or focal point for T cell receptor signaling. 16:43.0 So, activating PLC-gamma 16:44.3 gives you both calcium and Ras activation, 16:46.2 so it's both important for acute signaling 16:49.2 and also for some of the sustained signaling 16:51.1 that leads to transcription, 16:52.3 and this is critically dependent upon these actin foci. 16:55.1 So, these actin foci 16:57.0 are not just forming the whole synapse... 16:58.2 or F-actin is not just forming the whole synapse, 17:01.0 but there are subtypes of actin in the synapse 17:03.1 that are driving this signal amplification process. 17:06.1 And this is just a kind of a movie that kind of illustrates, again, 17:11.1 the remarkable things that are happening in the synapse with actin. 17:13.2 These little dots are basically 17:16.0 microcontact-printed regions of a T cell receptor ligand 17:19.3 and the system is back-filled 17:22.1 with the adhesion molecule ICAM-1, 17:23.2 so the cell is spreading on the ICAM-1, 17:25.2 actin is polymerized at the T cell receptor, 17:27.0 and then it's forming these intricate kind of spirals, 17:30.0 which basically, with the actin, 17:31.3 kind of distributing to the integrins, 17:33.3 and the integrin is kind of taking this actin 17:36.0 that's being generated at the T cell receptor 17:37.2 and then kind of modifying it. 17:39.0 So, there's a, you know, 17:40.3 quite a bit to learn about the specifics of this, 17:43.0 but at this point one thing we feel that we know 17:46.2 is that this is very important for signal amplification. 17:50.2 So, I want to talk a little 17:52.1 about the applications of the immunological synapse 17:53.2 at this point. 17:55.0 So, having told you some things about the way it's working... 17:57.1 so, what is the significance of this? 18:00.0 So, we think that there are a couple of particular things... 18:02.1 I mean, one recent work 18:04.2 with Kai Wucherpfennig's group on autoimmune disease I want to point out, 18:07.2 and also a recent collaboration we did 18:10.1 with a group of cancer immunologists who use radiation therapy 18:12.2 in combination with a new type of biologic 18:15.0 called a checkpoint blockade drug 18:18.3 that essentially... 18:21.1 again, and these checkpoint blockade antibodies 18:23.3 have really revolutionized melanoma therapy at this point, 18:27.1 so kind of really changed the game in a lot of ways. 18:30.3 So... but then first i want to basically 18:34.0 define one other term 18:36.1 or at least mode of interaction 18:38.1 between T cells and substrates, 18:39.2 and this is... 18:41.2 essentially to compare a synapse 18:43.1 and what we refer to as a kinapse. 18:44.3 So, just to illustrate what this is... 18:46.1 so, we have a synapse on this side, 18:48.2 which is... kind of, maybe a movie you've seen before. 18:51.1 It's radially symmetric, the T cell receptor goes towards the center. 18:54.1 This is showing the T cell receptor signal in this case. 18:56.1 But you also see cells 18:59.2 that essentially start to form a symmetric synapse, 19:01.1 but then they break symmetry 19:03.0 and they start to move, 19:04.1 and this one actually is moving kind of down and towards me, 19:07.2 so what you'll basically start to see 19:10.0 is that it starts to have a trailing edge and a leading edge. 19:12.2 And we can describe the cytoskeletal dynamics 19:17.2 of Kupfer's supramolecular activation complex 19:20.1 as having characteristics like a lamellipodium, 19:22.2 a lamella, 19:24.3 which is basically a site of adhesion molecule 19:27.0 concentration for the pSMAC, 19:28.3 and then the cSMAC being sort of like 19:31.1 the uropod in this migrating cell. 19:32.2 So, we think that the machinery 19:34.1 is actually quite similar, 19:35.2 so when a cell is basically moving 19:37.1 or engaged in this, 19:38.2 what we refer to as a kinapse, a moving junction, 19:41.2 it's actually integrating signal in much the same way 19:43.3 as the cell that's forming the radially symmetric synapse, 19:46.1 but this cell will basically 19:48.2 move away from the place where it started out, 19:50.2 whereas this cell will pretty much stay in place 19:52.2 because the radially symmetry balances all the forces 19:54.2 and the cell will stay. 19:56.1 So essentially, this is an issue of, you know, 19:58.1 Should I stay or should I go? 19:59.3 The cells do these different behaviors in different contexts in vivo, 20:03.0 and we've found that this is 20:06.2 potentially useful in both the autoimmune and cancer immunotherapy context, 20:09.2 so we think about it. 20:10.2 And we know genetically, 20:12.1 based on a paper by Tasha Sims in 2007, 20:16.2 again, a postdoc in my lab, 20:19.1 that essentially the symmetry breaking 20:22.0 is driven by a protein kinase called protein kinase theta, 20:25.1 and the reformation of a stable synapse, 20:29.0 from this state going to a stable synapse, 20:31.2 requires WASP, 20:34.0 the kinase I described before. 20:36.0 And again, it's probably in synergy with the integrin system 20:38.2 to basically, essentially, 20:40.3 repair the broken pSMAC 20:42.1 and basically get you back to a symmetric state. 20:44.2 So, this heatmap... 20:47.1 it's complicated, but the punchline is that basically 20:50.1 autoreactive T cells, and these are T cells 20:52.1 from patients that recognize bona fide autoantigens 20:54.2 in multiple sclerosis and type I diabetes, 20:58.0 and comparing these to T cells 21:00.1 that are responsible for killing 21:02.3 flu-infected T cells in influenza... 21:04.2 or sorry, influenza-infected target cells, 21:07.2 you essentially... 21:10.2 these form fantastic synapses, 21:11.2 so the red and orange all say that these are, 21:14.0 you know, attributes of stable synapse formation, 21:17.1 and then as you basically go into these different clones, 21:20.1 which are indicated up here, 21:21.3 in the multiple sclerosis 21:23.2 or type I diabetes setting, 21:25.0 you see, you know, 21:27.2 progressive essentially 21:30.1 defects in this immunological synapse formation process. 21:33.2 So, I guess... and this is just kind of illustrated here... 21:35.2 you have essentially kinapses 21:37.2 forming with these 21:40.2 MS-specific T cells 21:41.3 and type I diabetes-associated 21:44.0 autoreactive helper T cells. 21:45.3 So, this is... and again, 21:47.3 this is a characteristic that appears 21:50.0 to come from the antigen recognition process, 21:52.1 so self-antigen recognition in this context, 21:54.0 driving a weaker synapse, 21:56.1 and we think one possibility is that 21:58.1 this failure of stable synapse formation 22:00.2 may make these more difficult to regulate. 22:02.0 That's our hypothesis at this point 22:04.1 and I think that this is something that we're looking to test in the future. 22:08.0 So... cancer... 22:10.2 so, uhh, immune evasion 22:13.0 has recently been made by cancer biologists, 22:16.0 this review by Hannahan and Weinberg, 22:18.2 who kind of revisited the hallmarks of cancer... 22:21.1 so, cancer immune evasion 22:24.1 is now considered a hallmark of cancer. 22:26.0 So, this is... 22:27.1 and this is a big change over the past decade 22:29.1 in how people think about this, 22:30.3 and a lot of this change 22:32.2 has been driven by a couple of advances 22:34.2 based on these therapeutic agents 22:36.3 that basically remove inhibitors 22:40.2 of the immune system 22:42.1 or neutralize inhibitory pathways 22:43.3 that are probably normally protecting us from autoimmunity 22:46.2 or immunopathology, 22:47.2 but in the context of cancer 22:49.2 can be allowing the cancer to evade the immune system. 22:51.2 You block these and the immune system 22:54.0 suddenly will see the tumors 22:56.1 much more clearly and destroy them. 22:58.0 So, again, a change in opinion. 22:59.3 And this is one of those systems, 23:01.1 so, an antibody called Ipilimumab 23:03.2 that's targeting a protein called CTLA-4, 23:06.1 which is cytotoxic T lymphocyte antigen-4, 23:10.1 and this is an image from work by 23:13.2 Jim Allison and Jackson Egan, 23:16.2 basically showing a fluorescent CTLA-4 23:19.2 moving to the immunological synapse. 23:21.2 So, this is a synaptic marker, 23:25.1 it's a regulator of effector T cells, 23:27.1 it's generally thought of as an inhibitor 23:29.2 -- when it's blocked by antibodies 23:31.3 it then increases activation -- 23:33.2 and essentially also it's used by the regulatory T cells 23:38.1 I mentioned in part 1, 23:40.2 which buffer responses and therefore, 23:43.0 because CTLA-4 is an effector molecule of the regulators, 23:45.3 inhibiting it actually... 23:47.2 it neutralizes some of the function of the regulatory cells too. 23:50.2 And finally, CTLA-4 cross-linking, 23:53.1 though, has kind of a side effect. 23:55.1 I mean, all drugs have side effects, 23:56.2 and in this case it's synapse destabilization. 23:59.1 So, again, CTLA-4, 24:01.2 perhaps because it's involved in an axis 24:04.1 that regulates protein kinase C-theta, 24:05.3 which I mentioned before is a synapse breaker, 24:07.2 may enhance the activation of PKC-theta 24:09.2 and in doing so essentially 24:12.2 destabilize these synapses. 24:14.1 And we've actually seen this is in a model of 24:18.1 cancer therapy in the mouse, 24:19.3 where we're setting up a model 24:22.0 of a breast carcinoma 24:23.2 implanted in the flank of a rodent 24:25.1 that then grows over a period of 24:27.2 about 12 days and metastasizes to the lung, 24:30.2 so it's a model that basically 24:33.1 recapitulates key events 24:36.2 in the development of breast carcinoma, 24:37.2 although it doesn't start from the orthotopic site 24:40.2 like the ducts of the mammary glands. 24:43.1 It basically is... 24:44.3 we're taking a cell line 24:46.2 and putting it into the animal. 24:48.1 So, this tumor is susceptible... 24:50.1 and we can then view the T cells within these tumors 24:54.2 using a method called two-photon intravital microscopy, 24:57.1 in which an anesthetized animal 24:59.1 is basically placed on the microscope stage, 25:01.1 kept warm, 25:03.2 kept alive, 25:05.0 but under anesthesia, 25:06.2 and the tumor can be surgically exposed and imaged 25:10.2 using pulsed infrared lasers 25:12.1 that basically allow you to image deep into the tissue 25:14.1 and do timelapse imaging in real time. 25:17.1 And these are some images 25:19.0 and this is basically just illustrating 25:21.0 our therapeutic model. 25:22.2 So, if we put in anti-CTLA-4, again, 25:25.1 an anti-mouse version of Ipilimumab, 25:28.1 what we find is that, as in some patients, the therapy fails, 25:32.3 so the tumor continues to grow, you have high metastases. 25:35.3 The blue cells here are the tumor cells, 25:37.2 the red are some tumor vasculature, 25:39.3 which, again, is famously abnormal, 25:42.1 but nonetheless the T cells are getting in, 25:44.1 but what you can see the T cells are doing 25:46.2 is they're highly motile. 25:47.2 Even in the tumor, the T cells are moving. 25:50.1 This means they're forming kinapses; 25:51.2 they're not forming synapses. 25:54.0 However, if we combine the anti-CTLA-4 25:56.3 with radiation therapy, 25:57.3 which was the therapeutic modality 26:00.0 that Sandra Demaria and Silvia Formenti 26:03.0 were focusing on in this case, 26:05.3 now we see that the T cells 26:07.2 -- the green cell, the T cells -- 26:09.1 within the tumor are arrested on the tumor cells, 26:12.0 so they're now forming synapses, 26:13.1 and this is correlated with tumor shrinkage 26:15.1 and low metastasis. 26:17.1 So, what we went on and then showed... 26:19.2 and these are basically... these just illustrate the two modes, 26:22.2 so the low motility in the tumor is associated with the, 26:25.1 again, effective combination therapy, 26:28.0 whereas the high motility is associated with the anti-CTLA-4, 26:31.2 again, kind of side effect. 26:32.3 We found that we could basically recapitulate, 26:35.1 even in the combination with radiation therapy, 26:38.1 we could recapitulate this increase in motility 26:41.2 by inhibiting another receptor called NKG2D, 26:45.0 and this is... 26:46.2 you know, I don't want to... 26:48.1 obviously the details of this are important, 26:49.2 but this is a, kind of, 26:51.0 what's referred to as a co-stimulatory receptor. 26:53.2 So, CTLA-4 antagonizes the activity of CD28, 26:56.3 which is one type of co-stimulator. 26:58.2 NKG2D is in another class of co-stimulators 27:02.1 on the CD8 T cells, 27:03.2 and what appears to happen is that 27:05.3 when you irradiate the tumors they upregulate ligands for NKG2D 27:09.1 and that additional co-stimulatory pathway 27:12.1 stabilizes the synapse, 27:14.1 so unlike CD28 that activates PKC-theta 27:16.2 and inhibits synapse formation, 27:17.3 NKG2D is delivering activating signals 27:20.3 that are stabilizing the synapse without activating protein kinase C-theta, 27:24.1 we believe, 27:26.1 and this is leading to essentially a reversal 27:28.3 of the therapeutic effect. 27:30.2 So we think that what we're essentially... 27:32.2 and this is just illustrated here. 27:34.1 So, if you block the NGK2D, 27:36.1 you kind of increase the lung metastasis 27:38.2 versus the effect of ionizing radiation 9H10, 27:42.1 which is the anti-CTLA-4 antibody... 27:44.1 so, no lung mets, 27:46.1 and basically the antibody to NKG2D, you get the lung mets back. 27:50.2 So, essentially the behavior of the T cell in the tumor 27:54.2 correlates very strongly with our clinical outcome 27:58.0 in this rodent model, 27:59.3 and this approach is in human trials 28:03.1 and, you know, 28:06.0 is showing promise in conjunction with Ipilimumab. 28:08.2 So, this is sort of an illustration 28:11.3 of how we can use information 28:13.2 about the immunological synapse 28:15.1 to think about what's happening 28:17.2 in the context of immunotherapy, 28:20.1 and this is a movie of the situation 28:22.2 where the combination therapy 28:24.2 has been very successful in eradicating the tumor, 28:26.2 no tumor cells, lots of T cells. 28:28.2 So, at least in the immediate post-tumor site period, 28:32.0 this is probably what we want to see. 28:33.2 So, there are a number of factors 28:35.2 that will basically be controlling this, 28:37.1 but I think one of the key things that we're, 28:40.1 you know, kind of coming back to 28:42.1 is that the radiation therapy 28:44.1 is basically triggering an innate injury response, 28:47.1 it's upregulating NGK2D ligands on the tumor cells, 28:51.0 and this is allowing the immune system 28:54.1 to get back on target in the presence of the therapy 28:57.1 with anti-CTLA-4 28:59.1 to eradicate or at least significantly slow down the growth of the tumor. 29:03.2 So, this anti-CTLA-4 side effect, 29:05.1 its synapse destabilization, 29:06.3 can be counteracted in this case by irradiation therapy, 29:09.2 but we think the more specific thing that's doing 29:12.1 is engaging this other signaling pathway 29:14.1 that stabilizes the synapse and, you know, 29:17.3 essentially this what's referred to as a stress induced molecule, 29:19.2 the NGK2D ligand, called RAE-1. 29:24.0 If you're interested in the detail, it's in this paper. 29:27.0 So again, that's an example of using 29:29.2 information about the immunological synapse 29:31.1 to think about mechanisms in disease and disease treatment. 29:36.2 I'd like to acknowledge many investigators 29:39.1 who contributed to this at both... 29:42.3 particularly NYU and more recently Oxford, 29:45.0 where we've moved in the last two years. 29:47.0 Again, there are many other things 29:50.2 that you can follow up on 29:52.2 in terms of citations in the talk 29:54.1 and, you know, I guess, Happy Explorations. 29:57.2 So, thank you. Bye-bye.