Intracellular trafficking

Video Transcript

• 00:00 - 00:19:  Hi, thank you for joining the fourth session of Spotlight on Neuroscience, our month-long virtual conference.
• 00:19 - 00:26:  We have three wonderful talks for you today, followed by a captivating panel discussion being led by Dr. David Rubenstein.
• 00:26 - 00:31:  My name is Sian Constantine and I'm the Strategic Marketing Manager for Neuroscience at Abcam.
• 00:31 - 00:36:  Before we begin, I'd like to run through a couple of housekeeping notes.
• 00:36 - 00:38:  All attendees are automatically muted.
• 00:38 - 00:43:  However, please feel free to submit your questions in the Q&A box at the bottom of your screen.
• 00:43 - 00:48:  And these questions will be addressed during the dedicated Q&A session after each talk.
• 00:48 - 00:53:  If we don't get to a question, we will get the response over to you shortly after the event.
• 00:53 - 00:55:  So please don't panic.
• 00:55 - 01:06:  I'm also excited to inform you that Abcam is approved as a provider of continuing education programs in the clinical laboratory sciences space by the ASCLS PACE program.
• 01:06 - 01:16:  PACE credits are available for this event and a link to request credit will be sent in the chat box at the end of the event.
• 01:16 - 01:25:  So now I'd like to give you a brief overview of Abcam and how we operate in the neuroscience space.
• 01:25 - 01:34:  So Abcam really wants to advance the needle in science and particularly in neuroscience in these three pillars.
• 01:34 - 01:44:  So we're looking to continue to create research tools that can help you achieve your mission faster in the lab.
• 01:44 - 01:50:  We have three areas that we're particularly keen to do this in. One is the neurodevelopment space.
• 01:50 - 01:57:  Another is the neurobiological processes space. And finally, the neurological diseases space.
• 01:57 - 02:11:  So we've got many examples over the years of how we've collaborated with industry and academia to make this happen and to make sure that the right research tools are available to you to buy from our catalog.
• 02:11 - 02:26:  We want to make sure we continue to do this. So if you have any ideas, if there are any particular research tools that you think would be good to have or you want to chat to us about what you think about our products, please do contact us at neuroscience@abcam.com.
• 02:26 - 02:38:  We'd love to collaborate further with you. So the research tools that we generate and put in the catalog for our customers span multiple product types.
• 02:38 - 02:47:  And traditionally, we've been most known for antibodies. And whilst that is still a big part of our catalog, it's by no means the only part.
• 02:47 - 03:02:  And certainly in the neuroscience space, we'd like to build out more. So what we are doing quite well with the antibody technology we've always had is to build out the ELISA kits and the proteins and peptides.
• 03:03 - 03:16:  What we would now like to do more of for neuroscience is think about the cellular biochemical assays, cell lines, lysates, and multiplex microRNA immunoassays.
• 03:16 - 03:27:  So really keen to work with you. There are multiple products already in the catalog, and I encourage you to browse through these different product types, especially if you didn't know that they were available.
• 03:27 - 03:39:  It'd be great for you to be aware of them, but also send us any feedback that you might have. Two products in particular I can draw your attention to in the cell line space are the iPSC neurons.
• 03:39 - 03:47:  We have glutamatergic neurons and skeletal myocytes that are particularly useful for neuroscience research.
• 03:47 - 04:07:  So please do check those out. If there's anything you're particularly interested in that we don't have right now, and maybe it'll take a little bit too long for us to develop in a collaboration in a catalog, then please do also check out our customization services where we can build customized products for your research.
• 04:07 - 04:16:  This slide just gives you a flavor of some of the different product types we have specifically to support your research in intracellular trafficking.
• 04:16 - 04:24:  So on the bottom right-hand side, you can see the iPSC glutamatergic neurons I mentioned a moment ago.
• 04:24 - 04:33:  But also you can see the breadth of the catalog across the other areas, such as the ELISA kits and the Altitude cell lines.
• 04:34 - 04:41:  So now I would like to take a moment to introduce the moderator for today's event, Professor David Rubenstein.
• 04:41 - 04:49:  Professor David Rubenstein is the Professor of Molecular Genetics and a UK Dementia Research Institute Professor at the University of Cambridge.
• 04:49 - 05:03:  He is Deputy Director of the Cambridge Institute for Medical Research. Professor Rubenstein earned his MB, CHB, BSc, Med, Hons, and PhD degrees from the University of Cape Town.
• 05:03 - 05:13:  He came to Cambridge in 1993 as a senior registrar in genetic pathology and was the first person to complete formal training in this field in the UK.
• 05:13 - 05:20:  His research is focused on the field of autophagy, particularly in the context of neurodegenerative diseases.
• 05:20 - 05:33:  His laboratory pioneered the strategy of autophagy upregulation as a possible therapeutic approach in various neurodegenerative diseases and has identified drugs and novel pathways that may be exploited for this objective.
• 05:33 - 05:44:  Professor Rubenstein's contributions reveal the relevance of autophagy defects as a disease mechanism and to the basic cell biology of this important catabolic process.
• 05:44 - 05:51:  Thank you. I will now hand the mic over to David.
• 05:51 - 05:59:  Thank you so much for introducing the conference and thanks to Abcam for organizing everything.
• 05:59 - 06:14:  I'm delighted that we have three world-leading speakers whose work I really admire who are going to cover different domains relevant to membrane trafficking and neurodegeneration.
• 06:14 - 06:22:  And so thank you all three of you for agreeing to participate in this afternoon's event.
• 06:22 - 06:33:  The first speaker is Thierry Galli. He's a group leader at the Institute of Psychiatry Neurosciences in Paris.
• 06:33 - 06:44:  And he's very well known for his research focusing on the roles of SNARE proteins in the context of neuroscience, particularly neuronal cell differentiation.
• 06:44 - 06:59:  Most recently, he's proposed a new mechanism for membrane expansion and neurite growth based on non-vesicular transport of lipids at ER plasma membrane contact sites and secretory reticulophagy.
• 07:00 - 07:07:  And welcome, Thierry, and we all look forward to hearing your talk this afternoon.
• 07:07 - 07:15:  Thank you very much, David, and thank you to Kayleigh and the people at Abcam for organizing this.
• 07:15 - 07:18:  So I'm going to share my screen.
• 07:18 - 07:27:  I hope it works without any problems, and we'll see in a second if that's coming up.
• 07:27 - 07:34:  It's taking a little bit of time, apparently.
• 07:34 - 07:36:  Yeah, okay, here it is.
• 07:36 - 07:39:  So if you see the screen well, I'm going to start.
• 07:39 - 07:41:  Indeed, David.
• 07:41 - 07:44:  Thierry, we don't see your shared screen yet.
• 07:44 - 07:45:  Okay.
• 07:45 - 07:50:  There we go again.
• 07:50 - 07:51:  There.
• 07:51 - 07:54:  Looks great.
• 07:54 - 07:56:  What about now?
• 07:56 - 07:57:  This is great.
• 07:57 - 07:58:  Thank you.
• 07:58 - 08:00:  Okay.
• 08:00 - 08:01:  Thank you.
• 08:01 - 08:02:  Sorry.
• 08:02 - 08:03:  All right.
• 08:03 - 08:06:  So without waiting more, I'm going to tell you about.
• 08:06 - 08:21:  So what we've been interested in my team for quite some time now is to try to understand how certain cells, and particularly here today neurons, are able to grow tremendous amounts of their membrane.
• 08:21 - 08:58:  And so this little drawing on the left here depicts the case of motor neurons, which extend axons to very long distances in humans, but think even larger mammals, giraffes, and other animals.  Single sell neurons are able to send projections very long distances. Then within the brain if the distance is not, if the axons don’t go that far, via ramifications as depicted here in the case of dopaminergic neurons,
• 08:59 - 09:08:  The length of the axon can be tremendous, and this requires the addition of membrane.
• 09:08 - 09:13:  There are several mechanisms for adding membrane to the plasma membrane.
• 09:13 - 09:22:  David referred to previous work that we did about lipid transfer from the endoplasmic reticulum to the plasma membrane.
• 09:22 - 09:30:  And today I will tell you some of the latest on mechanisms that involve the secretion of material.
• 09:30 - 09:40:  So if you imagine a secretory vesicle here in red and the plasma membrane in gray, then the plasma membrane grows.
• 09:40 - 10:03:  If you add a secretory vesicle as depicted here, this will increase, obviously, the surface of the plasma membrane, provided that endocytosis, which we'll also hear about today, is limited and there is a scheme where exocytosis exceeds endocytosis.
• 10:03 - 10:08:  In that case, then the plasma membrane grows.
• 10:08 - 10:16:  So there are several pathways of exocytosis of membrane addition, several pathways of so-called secretion.
• 10:16 - 10:25:  And this little figure depicts two main mechanisms.
• 10:25 - 10:47:  The one that we all learned in textbooks is on the left-hand side of this little slide and depicts the so-called conventional secretory pathway, whereby lipids and membrane, transmembrane proteins, and secreted proteins are produced in the endoplasmic reticulum down here.
• 10:47 - 11:07:  And then they are translocated through the Golgi apparatus and then packaged into secretory granules, or they can also go into endosomal compartments and eventually end up at the plasma membrane here.
• 11:07 - 11:23:  And this is really the classical secretory mechanism that we all learned about and is responsible for the release of neurotransmitters, neuropeptides, peptidic hormones, insulin, etc.
• 11:23 - 11:44:  But what emerged in orange in the last 10 to 15 years is the fact that other compartments of endosomal nature could also fuse with the plasma membrane and release their content into the extracellular medium.
• 11:45 - 11:55:  And this is particularly the case of late endosomes, also called multivesicular bodies, because they include what are called intraluminal vesicles.
• 11:55 - 12:15:  And these compartments are also able to fuse with the plasma membrane and release their content into the extracellular medium. And the mechanisms are relatively different, particularly how proteins and lipids can end up here rather than there is likely very different.
• 12:15 - 12:40:  So, I will tell you today mainly about a pro-SNARE protein, which I will introduce in the next slide in greater detail. But this is just to show you an old piece of data where we indeed could show that this protein, which at the time we called TI-VAMP and which is now referred to by everybody as VAMP7,
• 12:40 - 12:55:  was involved in the secretion of these little vesicles that you see here that can be observed in the extracellular medium, not so much when you interfere with this VAMP7-dependent secretory pathway.
• 12:55 - 13:23:  And, in fact, here to impair the VAMP7 secretory pathway, we expressed a small piece of the so-called longing domain of the molecule. And when you do that, you also see that the release of a protein, which is bovine serum albumin, which has been pre-internalized into late endosomes and lysosomes, is impaired by the expression of the longing domain.
• 13:24 - 13:41:  So, these are old data that already indicated that VAMP7 plays an important role in this pathway. Now, I would like to connect this protein first with its basic function and secondly with autophagy.
• 13:41 - 14:04:  First of all, VAMP7, which I just alluded to, is a vesicular SNARE. And you can see here another vesicular SNARE in blue in this prototypic model. And what these molecules do is interact with target SNAREs, which are their favorite partners in a different membrane.
• 14:04 - 14:33:  So, what you see here is a vesicular SNARE in blue interacting with target SNAREs that, for instance, would be at the plasma membrane. The formation of this complex drives the intimate docking and then fusion of the secretory vesicle with the plasma membrane. And you can see the opening of a pore and the release of the content into the extracellular medium.
• 14:34 - 14:55:  So, the basic function of VAMP7 is similar to this VAMP2 that you see here. It's to mediate fusion. Now, several groups that are listed here, including seminal work that David Rubenstein carried out in his lab 10 years ago, actually showed that VAMP7 is involved in autophagy.
• 14:55 - 15:24:  And this was really the first evidence that this SNARE, which we were finding involved in secretory mechanisms, also played a role in autophagy. At that time, what David found is that VAMP7 is involved in a very early event of autophagy where so-called precursor vesicles can fuse together to form this membrane called the phagophore that then encapsulates elements of the cytoplasm.
• 15:25 - 15:39:  That, at the time, everybody thought were only destined for degradation. Then, we and others indeed pointed to the fact that from here, you could also go to secretion.
• 15:39 - 15:57:  I just want to point to another little piece of data from Beaujolais Diuretics Lab here in a recent paper showing that VAMP7 is the only vesicular SNARE that includes a so-called LIR, LC3 interacting region.
• 15:57 - 16:14:  And this is a sequence that allows it to interact with a very important protein called LC3, a key protein that is a marker of autophagosomes.
• 16:14 - 16:34:  So, previously, we found that VAMP7 was involved in axonal growth, again, by expressing this aminoterminal longing domain that I showed you was inhibiting secretion, also by silencing the expression with siRNA.
• 16:34 - 16:50:  And both affecting VAMP7 with these approaches had a profound effect on axonal growth. I don't know why there's a phone call at the same time. Sorry about that.
• 16:50 - 17:04:  And so, putting all this together, we were curious to now ask the question, is there a connection between the role of VAMP7 in autophagy and the role of VAMP7 in neurite growth?
• 17:04 - 17:33:  So, we started experiments where we were interested in looking at the effect of interfering with autophagy and looking at the development of neurons in culture. So, these are hippocampal neurons grown in culture here, and they are stained in red for tau, which is a marker of axons, and beta-tubulin, which is a marker of dendrites in green.
• 17:33 - 17:56:  The first experiment that we carried out consisted of diluting some elements of the neuronal culture medium to induce a partial starvation of the neurons and in conditions where we could demonstrate that there was induction of autophagy, and I don't have time to show you all the details there.
• 17:57 - 18:16:  And we wanted to see what was the impact on axonal growth, dendritic growth, and on neuronal differentiation. The surprise was that neuronal polarity in this condition was preserved, and that the axons, which here are stained with tau in red, were longer than in the control condition.
• 18:17 - 18:35:  Then we treated this very same culture in the same conditions, control and with diluted medium. We treated them with an inhibitor of autophagy, and we could see that, indeed, that blocked the induction of autophagy by diluted medium.
• 18:35 - 18:58:  But what we found is that spouting also had a very profound effect on growth and polarity. And what you can see here is that now, instead of having only one axon in this case (the top cases), instead of having only one axon, we had neurons with so-called supernumerary axons.
• 18:59 - 19:14:  So you can see here that you have two red branches. And the other thing is that these branches, these axonal branches, tended to be shorter than in the control situation or with the diluted medium.
• 19:14 - 19:28:  And what you can see in the bottom right panel is that we could reproduce this by inducing autophagy with drugs like rapamycin, which induced longer axons here in red. And, of course, spouting plus rapamycin,
• 19:28 - 19:45:  we had, like I told you previously, now multiple axonal branches. So this is summarized here. By inducing autophagy by medium dilution or rapamycin, we get longer axons. By inhibiting autophagy, we get multiple axons.
• 19:45 - 20:02:  So we wanted to understand what was going on here because, obviously, autophagy was impacting greatly neuronal differentiation. And we thought that there might be a connection with VAMP7 in this pathway.
• 20:02 - 20:16:  So by video microscopy, we looked at the trafficking in the axon of VAMP7 together with the only transmembrane protein involved in autophagy, which is ATG9.
• 20:16 - 20:32 :  And very interestingly, we could observe many vesicles that would travel in the growing axons that were positive for both VAMP7 and ATG9. Some were trafficking retrogradely, as you can see some of these yellow spots are going retrogradely.
• 20:32 - 20:29:  But very interestingly, there were also many particularly small puncta that were also going anterogradely.
• 20:39 - 20:54:  And this was also the case when we looked at VAMP7 and LC3. LC3 is this molecule that gets targeted and goes into autophagosomes.
• 20:54 - 21:10:  And so you can see several VAMP7, LC3, anterogradely moving structures. So to us, this was a strong indication to basically try to go forward and try to understand what was going on.
• 21:11 - 21:16:  So we then moved to a simplified model, which is a cell line that can grow neurites called PC12 cells.
• 21:19 - 21:20:  And PC12 cells can be induced to grow neurites by treatment with NGF.
• 21:25 - 21:30:  And we made, by CRISPR, VAMP7 knockout and ATG5 knockout.
• 21:30 - 21:33:  And what you can see here is that the VAMP7 knockout, as the case for neurons I showed you previously,
• 21:33 - 21:37:  very much impaired in the growth of neurites.
• 21:37 - 21:41:  Whereas the ATG5 knockout is a little bit like the spouting, growing longer and more ramified neurites.
• 21:41 - 21:55:  So ATG5 knockout leads to a complete blockade of degradative autophagy.
• 21:57 - 22:05:  We then treated these cells with rapamycin and we observed something very similar to what I showed you in the neurons.
• 22:05 - 22:10:  You can see that the longest branch, which is here in purple, tended to be longer.
• 22:11 - 22:20:  There was no more of this effect in the VAMP7 knockout. And this effect of rapamycin was also observed in the ATG5 knockout.
• 22:20 - 22:31:  And this is quantified and summarized here. So there is a paradoxical effect of rapamycin, which in fact induces the longest neurite to be longer.
• 22:31 - 22:37:  And this is also the case in the ATG5 knockout, but not in the VAMP7 knockout.
• 22:37 - 22:44:  And now if we compute all of the long branches and the longest neurite plus all the branches,
• 22:44 - 22:50:  now we don't see the effect of rapamycin anymore. We can see only the effect of the genotypes.
• 22:50 - 22:59:  So strong impairment of growth in the VAMP7 knockout and strong increase of growth when you abrogate degradative autophagy.
• 23:00 - 23:05:  So this led us to try to understand the effect on secretion more specifically.
• 23:05 - 23:17:  And to do that, we recovered the medium conditioned by VAMP7 knockout and ATG5 knockout cells upon NGF differentiation.
• 23:17 - 23:21:  And we compared that with the medium conditioned by wild type cells.
• 23:21 - 23:24:  And we precipitated all the proteins.
• 23:24 - 23:39:  And then we did mass spectrometry with the appropriate number of replicates, etc. And I will start by showing you what happens when you compare the ATG5 knockouts with the wild type.
• 23:39 - 23:47:  What you can see in this volcano plot on the right side are molecules that are more secreted by the ATG5 knockout.
• 23:47 - 24:04:  So we see several markers of the autophagy pathway that normally should go into degradation, particularly SQSTM1, which is in fact P62, LC3A, LC3B, GABARAP, and CALCOCO.
• 24:04 - 24:08:  But another family of proteins emerged in this.
• 24:08 - 24:17:  And these are the reticulons, RTN3, RTN1, RTN4, and the atlastins, atlastin1 and atlastin3.
• 24:17 - 24:25:  And these are molecules of the endoplasmic reticulum. And I will show you more detail about RTN3 in a second.
• 24:25 - 24:32:  Now, what's important is that RTN3 and atlastin3 were much less secreted by the VAMP7 knockout cells.
• 24:32 - 24:42:  And of course, now, if you compare the VAMP7 knockout and the ATG5 knockout, so those that don't grow neurites with those that grow extended neurites,
• 24:42 - 24:53:  now you can see that the ATG5 knockout secretes much more of all these molecules compared with the VAMP7 here.
• 24:53 - 24:56:  So the VAMP7 cells are very impaired in the secretion of all of these molecules.
• 24:59 - 25:01:  So let's focus now on reticulon 3.
• 25:01 - 25:15:  So reticulon 3 is this molecule here. There are several splicing variants, and it's inserted in the endoplasmic reticulum membrane via four membrane-spanning regions.
• 25:15 - 25:25:  These are not transmembrane. And then it has this very long aminoterminal tail, and another form is shorter and doesn't have this aminoterminal tail.
• 25:25 - 25:38:  So the long form of RTN3 was previously shown by Ivan Dikic in autophagy, so-called reticulophagy, so a pathway that degrades ER tubules by autophagy.
• 25:38 - 25:43:  And that's not the case for the short form.
• 25:43 - 25:51:  So we went on to try to understand what kind of RTN3 we were looking at.
• 25:51 - 26:05:  And here, to cut a long story short, this is the short form of RTN3 that is secreted in a VAMP7 and ATG5 independent manner.
• 26:05 - 26:08:  You can see the band at 25 kilodaltons.
• 26:08 - 26:13:  So I'm going to take you through this western blot, which are important here.
• 26:13 - 26:20:  But before I do that, I want to point out the fact that here we looked at the whole secretome,
• 26:20 - 26:25:  but in this particular experiment, we looked at extracellular vesicles.
• 26:25 - 26:37:  So because RTNs and atlastins are associated with membranes, we thought that they ought to be secreted in a form that would be associated with membranes.
• 26:37 - 26:42:  So we precipitated the extracellular medium to recover extracellular vesicles.
• 26:42 - 26:53:  So we looked at the total expression of proteins in cell lysates, and we looked at what is secreted by wild-type VAMP7 knockout and ATG5 knockout.
• 26:53 - 27:07:  And these cells, we also looked at what happens when they are treated with rapamycin, which inhibits degradative autophagy, as I told you, and bafilomycin, which prevents lysosomal degradation.
• 27:08 - 27:19:  So first of all, what is important here in wild-type cells, you can see that there is secretion of RTN3, a little more with rapamycin, and much more when you prevent degradation,
• 27:19 - 27:23:  and much more if you look in ATG5 knockout cells.
• 27:23 - 27:36:  So there's a clear confirmation of the proteomics, also clear confirmation of the proteomics that this is VAMP7-dependent because you see much less in all of these treatment conditions in the VAMP7 knockout.
• 27:36 - 27:53:  What is interesting is that LC3-2, this molecule that is targeted to autophagosomes, can also be secreted, particularly if the cells are treated with bafilomycin and now the degradation is prevented.
• 27:53 - 28:04:  This is also VAMP7-dependent, and you don't see it secreted in ATG5 knockout for the reason that, in fact, LC3-2 is not produced.
• 28:04 - 28:14:  But what you see secreted in ATG5 knockout is, in fact, LC3-1, which also is sent out by these cells.
• 28:14 - 28:27:  Now, what you can see is that P62 also is released, and you can see here it's also sensitive to bafilomycin, and this is also VAMP7-dependent,
• 28:27 - 28:35:  and you can see the tremendous accumulation of P62 in the cells where degradative autophagy is prevented.
• 28:35 - 28:46:  So there is, indeed, VAMP7-dependent secretion of P62, reticulons, and LC3-2.  The secretion of RTN3 is definitely increasing in ATG5 knockout cells.
• 28:46 - 28:55:  This is occurring in extracellular vesicles, and I'm not going to go into further detail about this experiment.
• 28:55 - 29:10:  Very interestingly, at the time that we were wrapping up revisions of this paper, a paper about the group of Volker-Ockey came out showing that ATG5 knockout neurons overexpress VAMP7,
• 29:10 - 29:19:  reticulon 3, and reticulon 4, reticulon 1, and so you see that, very interestingly, by a completely different approach,
• 29:20 - 29:24:  the group of Volker was also able to point out to the very same molecules.
• 29:24 - 29:41:  So here, what is likely to happen is that these molecules are less secreted, and then they tend to accumulate in the neurons because in the ATG5 knockout,
• 29:41 - 29:45:  they are less degraded due to the absence of ATG5.
• 29:47 - 30:01:  So to connect with the work of other people in the field, particularly that of Jay Debnath, who also described a pathway where exosomes, extracellular vesicles,
• 30:01 - 30:15:  are able to release elements of the autophagy pathway, indeed, we think that we are looking at something rather similar to what Jay Debnath has found in non-neuronal cells.
• 30:15 - 30:27:  So we wanted then to go a little further with the complementary approaches, and what we've developed in my lab in recent years are recombinant antibodies,
• 30:27 - 30:40:  and you can see here the clone F1-1, which we developed against VAMP7, and you can see by yeast-to-hybrid that it interacts only with VAMP7 and not other VAMPs,
• 30:40 - 30:54:  and by co-expression with GFP-VAMP7, you can see that it binds GFP-VAMP7, and, of course, there's also a cytosolic pool of this synthetic antibody.
• 30:54 - 30:59:  Interestingly, when we express this synthetic antibody into cultured neurons,
• 30:59 - 31:03:  we completely prevent the overgrowth of the axon,
• 31:03 - 31:05:  as you can see in the quantification,
• 31:05 - 31:08:  and here we prevent the overgrowth
• 31:08 - 31:11:  that we observe otherwise when we dilute the medium,
• 31:11 - 31:14:  as I showed you at the beginning.
• 31:14 - 31:18:  We also went back to one of the original approaches,
• 31:18 - 31:21:  which I showed you impairs the growth of the axon,
• 31:21 - 31:25:  which is the expression of the longing domain of VAMP7,
• 31:25 - 31:30:  and we did this together with treatment with rapamycin or not,
• 31:30 - 31:34:  and this is showing you the localization of RTN3,
• 31:34 - 31:39:  which is present in a substructure of the ER,
• 31:39 - 31:41:  these ER tubules,
• 31:41 - 31:45:  and what you can see is that if you treat with rapamycin
• 31:45 - 31:49:  and the neuron expresses the longing domain,
• 31:49 - 31:52:  now there is a dramatic change in the appearance
• 31:52 - 31:53:  of reticulon three,
• 31:53 - 31:57:  which now decorates all of the endoplasmic reticulum,
• 31:57 - 32:01:  so not just the tubules, but you can see the ER sheets
• 32:01 - 32:04:  and, in fact, all of the endoplasmic reticulum,
• 32:04 - 32:07:  so we think that this is yet further proof
• 32:07 - 32:10:  that by impairing VAMP7,
• 32:10 - 32:15:  there are profound effects on molecules of the ER
• 32:15 - 32:16:  like RTN3,
• 32:16 - 32:20:  so to summarize what I have told you here,
• 32:20 - 32:24:  we think that downstream of the kinase mTOR,
• 32:24 - 32:30:  which is extremely important to control autophagy
• 32:30 - 32:33:  and cell growth, there are two mechanisms,
• 32:33 - 32:38:  the degradative autophagy and particularly ER phagy,
• 32:38 - 32:41:  which was known previously whereby pieces of the ER
• 32:41 - 32:46:  are degraded by a fusion of autophagosome and lysosome.
• 32:46 - 32:49:  On top of that, we think that these pieces of ER
• 32:49 - 32:51:  that include reticulons and atlastins
• 32:51 - 32:54:  can end up in a mixed compartment
• 32:54 - 32:59:  of late endosomal and autophagosomal nature,
• 32:59 - 33:01:  late endosomal and phagosomal nature,
• 33:01 - 33:03:  and that those are released
• 33:03 - 33:06:  and that they participate in neurite growth
• 33:06 - 33:09:  and they release this extracellular vesicle
• 33:09 - 33:13:  which contains elements of the ER.
• 33:13 - 33:14:  Now, what can be the effect
• 33:14 - 33:17:  also of these extracellular vesicles?
• 33:17 - 33:20:  So we collected the medium conditioned
• 33:20 - 33:24:  by these different cells, VAMP7 knockout,
• 33:24 - 33:26:  wild-type ATG5 knockout,
• 33:26 - 33:30:  and we gave this conditioned medium to naive cells,
• 33:30 - 33:32:  and what we could see is that the ATG5
• 33:32 - 33:34:  knockout conditioned medium
• 33:34 - 33:38:  tended to inhibit the growth of naive cells.
• 33:38 - 33:42:  So just to remind you, ATG5 knockout cells overgrow,
• 33:42 - 33:45:  but they release elements that inhibit,
• 33:45 - 33:50:  just as if there was a negative feedback loop
• 33:50 - 33:53:  provided by this secretion.
• 33:53 - 33:56:  So we wanted, and these are now unpublished data
• 33:56 - 33:57:  which I want to share,
• 33:57 - 33:59:  we wanted to know a little more
• 33:59 - 34:02:  about these extracellular vesicles.
• 34:02 - 34:05:  So we separated large extracellular vesicles
• 34:05 - 34:08:  and typical exosomes by centrifugation,
• 34:08 - 34:12:  and we showed by dynamic light scattering
• 34:12 - 34:15:  that indeed at 15,000 G,
• 34:15 - 34:20:  we have rather large 186 nanometer,
• 34:20 - 34:24:  400 nanometer extracellular vesicles,
• 34:24 - 34:26:  and at 200,000 G,
• 34:26 - 34:30:  we collect rather small extracellular vesicles
• 34:30 - 34:32:  in the size of typical exosomes.
• 34:32 - 34:35:  We analyzed them by Western blotting.
• 34:35 - 34:37:  We showed that typical exosomal markers
• 34:37 - 34:40:  like HSC70 and TSG101
• 34:40 - 34:46:  are present mainly in the exosomes of the 200,000 pellet
• 34:46 - 34:48:  and very little in the 15,000.
• 34:48 - 34:51:  We could also show that in this prep
• 34:51 - 34:54:  that RTN3, which I told you previously,
• 34:54 - 34:57:  is secreted more by ATG5 knockout
• 34:57 - 35:00:  and less VAMP7 by VAMP7 knockout.
• 35:00 - 35:03:  Indeed, we could show that this is also the case
• 35:03 - 35:06:  with the large and small extracellular vesicles,
• 35:06 - 35:10:  and you can see here that, in fact,
• 35:10 - 35:17:  both the ATG5 large EVs and small EVs secrete RTN3,
• 35:17 - 35:22:  and you can see here that exactly like what I showed you
• 35:22 - 35:24:  previously for the total secretome,
• 35:24 - 35:28:  RTN3 is more secreted by the ATG5 knockout,
• 35:28 - 35:30:  less by the VAMP7,
• 35:30 - 35:33:  and so this is a clear indication
• 35:33 - 35:38:  that the RTNs and atlastins are very likely to be,
• 35:38 - 35:43:  indeed, in large and small extracellular vesicles.
• 35:43 - 35:48:  We tested both of these pellets,
• 35:48 - 35:51:  of these extracellular vesicles on naive cells,
• 35:51 - 35:53:  and both of them are inhibitory,
• 35:53 - 35:56:  whether they correspond to typical exosomes
• 35:56 - 35:59:  or to larger extracellular vesicles,
• 35:59 - 36:02:  so we think that now it's going to be really important
• 36:02 - 36:05:  to try to understand the molecules
• 36:05 - 36:10:  that are signaling in these extracellular vesicles.
• 36:10 - 36:12:  So, indeed, we propose that,
• 36:12 - 36:16:  on top of the typical degradative ER phagy,
• 36:16 - 36:21:  there is also a mechanism to release elements
• 36:21 - 36:25:  of the endoplasmic reticulum, such as RTN3,
• 36:25 - 36:27:  via late endosomes.
• 36:27 - 36:28:  One of the big questions is,
• 36:28 - 36:34:  how do these ER elements go into late endosomes?
• 36:34 - 36:36:  Do they go by engulfment?
• 36:36 - 36:38:  And this is a possible mechanism
• 36:38 - 36:42:  because it has been seen by other authors previously,
• 36:42 - 36:47:  or could they go also by a SNARE-dependent fusion?
• 36:47 - 36:51:  And this is also a possibility that some ER-derived vesicles
• 36:51 - 36:53:  could fuse with late endosomes,
• 36:53 - 36:58:  and then this would bring molecules like RTN3
• 36:58 - 37:00:  and atlastins into late endosomes,
• 37:00 - 37:04:  and from there, they could be included
• 37:04 - 37:06:  into intraluminal vesicles
• 37:06 - 37:09:  and then be secreted as extracellular vesicles.
• 37:09 - 37:13:  So, and here, this pathway is important for neurite growth.
• 37:13 - 37:16:  When you kill degradative autophagy
• 37:16 - 37:19:  and you leave only this pathway,
• 37:19 - 37:22:  then the neurites grow more, but at the same time,
• 37:22 - 37:25:  the release of this extracellular vesicle
• 37:25 - 37:27:  provides a negative signal.
• 37:27 - 37:28:  So, we think that in development,
• 37:28 - 37:30:  this can be really important,
• 37:30 - 37:35:  but also in certain diseases
• 37:35 - 37:42:  where there could be an over-secretion of these elements.
• 37:42 - 37:45:  So, with this, I would like to wrap up
• 37:45 - 37:50:  and thank all the people who have participated in this study.
• 37:50 - 37:54:  Particularly, this was the work of Jose Wozniacki,
• 37:54 - 37:58:  and that's included in a paper that came out in 2020
• 37:58 - 38:02:  with great help and contribution by Sebastien
• 38:02 - 38:04:  and also Francesca,
• 38:04 - 38:09:  and now this project is continued also by Soumya in the lab.
• 38:09 - 38:14:  And these are the collaborators that have worked with us on this,
• 38:14 - 38:18:  particularly excellent work of our imaging platform
• 38:18 - 38:23:  and also the proteomics platform here at NICAR Hospital.
• 38:23 - 38:26:  With collaboration with the lab of Guanggu Shui,
• 38:26 - 38:29:  I didn't have time to show you the lipidomics that we have done,
• 38:29 - 38:32:  which also is extremely interesting in the ATG5
• 38:32 - 38:34:  and the VAMP7 knockout,
• 38:34 - 38:39:  and also Marissa Colombo and Claudio Fader
• 38:39 - 38:43:  with whom we've worked for years on autophagy.
• 38:43 - 38:45:  And these are the funding bodies,
• 38:45 - 38:48:  and I also would like to take this opportunity
• 38:48 - 38:53:  to advertise the opening of a post-doc position in my lab.
• 38:53 - 38:57:  And with this, I thank you very much for your attention.
• 39:01 - 39:02:  Terry, thank you very much.
• 39:02 - 39:04:  That was really nice and interesting.
• 39:05 - 39:09:  There are already some questions that I've been sent.
• 39:10 - 39:14:  So, is that okay for me to start?
• 39:15 - 39:19:  The first question is, how does LC32 reach the late endosomes?
• 39:19 - 39:22:  What about P62?
• 39:22 - 39:25:  Yeah, I think that these are some of the questions
• 39:25 - 39:28:  that we are trying to ask at the moment.
• 39:28 - 39:35:  And as you have seen, we can see the secretion of LC32
• 39:35 - 39:37:  in the wild-type cells.
• 39:37 - 39:42:  So, this has to be coming from autophagosome,
• 39:42 - 39:46:  and so obviously there has to be a communication
• 39:46 - 39:49:  between autophagosome and late endosome
• 39:49 - 39:52:  so as to provide LC32 to the late endosome.
• 39:52 - 39:54:  That's option number one.
• 39:54 - 39:58:  Option number two is what Jay Devnat has proposed,
• 39:58 - 39:59:  that there would be...
• 40:00 - 40:07:  a pool of LC32 that could be lipidated at the level of the late endosome. But in that case,
• 40:09 - 40:16:  indeed, that would not explain how P62 would end up there. And since we see secretion of P62,
• 40:16 - 40:22:  I would be tempted to think that rather model hypothesis one, there is a communication
• 40:22 - 40:26:  between autophagosome and late endosome. And this is something to explore now.
• 40:27 - 40:36:  That's interesting. Another question is, can you speculate about the role of this process
• 40:37 - 40:45:  in various diseases, whether, you know, secretory reticulophagy or
• 40:46 - 40:51:  secretion of... unconventional secretion, these types of routes is, for instance,
• 40:52 - 40:56:  contributing to the types of spreading events that people...
• 40:56 - 41:06:  Yeah. So basically, the way I think this at the moment is that we need to think a little bit out
• 41:06 - 41:16:  of the box in terms of secretory mechanisms and not to consider only the classical secretory pathway.
• 41:16 - 41:23:  But if you consider that molecules from the cytoplasm, but also from the ER and possibly
• 41:23 - 41:32:  from other compartments, because there was also publication about mitovesicles being released,
• 41:32 - 41:36:  so elements of the mitochondria, there might be elements from the Golgi apparatus.
• 41:37 - 41:47:  So, there were recent papers showing secretion of metalloproteases being secreted by late endosomes.
• 41:47 - 41:55:  And so, and, you know, with an mTOR pathway, this is the lab of Dimitriades who had a paper in
• 41:55 - 42:04:  molecular cell showing that by, like in our case, inhibiting mTOR leads to secretion of elements
• 42:04 - 42:11:  from the Golgi apparatus via autophagosome and late endosomes. Now, those could be impacting
• 42:11 - 42:19:  extracellular matrix, could have autocrine effects, could have paracrine effects. Randy
• 42:19 - 42:28:  Scheckman showed that secretion in extracellular vesicles of cyclin D1 has an impact on neural
• 42:28 - 42:36:  stemness in target cells. So, I think that this is further expanding the possibilities of
• 42:37 - 42:43:  exosomes and extracellular vesicles to carry signals, if now they are also transporting
• 42:43 - 42:50:  elements of the ER, mitochondria, Golgi, etc. And so, we need to look into
• 42:51 - 42:58:  this kind of signals, because they can really have a strong impact in neurodegenerative disease and
• 42:58 - 43:03:  cancer. That's, I think, these are two areas where they could have a strong impact.
• 43:04 - 43:11:  Alfonso Martin Pena has a comment. He says, great talk. In regards to the VAMP7 rate of
• 43:11 - 43:17:  movement, what is the range of speeds, either anterograde or retrograde, and whether the
• 43:17 - 43:22:  speed changes with the different modifications that are genetic or pharmacological?
• 43:22 - 43:32:  Yes. So, we went through that, and we have all the quantification in the Wojnacki et al. Cell
• 43:32 - 43:37:  Reports 2020 paper. It's one of the supplemental figures. And there is an interesting effect of
• 43:38 - 43:48:  treating the neurons with rapamycin, and so impairing, you know, regulating autophagy
• 43:48 - 43:54:  has an effect on axonal transport. So, Erika Alsbaugh looked at this in great detail,
• 43:54 - 43:59:  particularly retrograde transport, but we also looked at anterograde transport.
• 43:59 - 44:05:  In the case of VAMP7, both of them are affected by rapamycin, for instance.
• 44:06 - 44:14:  Okay. An anonymous question that's just come in is, could you expand a bit more on how these
• 44:14 - 44:27:  secreted proteins like RTN3 inhibit membrane growth? So, we don't know yet what are the
• 44:27 - 44:36:  elements in exosomal vesicles that might be having this autocrine-paracrine effect. We don't
• 44:36 - 44:44:  know if these are the RTNs, the atlastins. The recent report showed that impairing
• 44:45 - 44:52:  VAMP7-dependent secretion of exosomes impaired the release of microRNA. So, you know,
• 44:52 - 44:57:  these extracellular vesicles, they include a lot of goodies. They include lipids, they include
• 44:57 - 45:03:  proteins, and they include micro and nucleic acids. So, each of these elements could be
• 45:03 - 45:09:  important. But what I think is really critical is that in a condition where there is overgrowth,
• 45:09 - 45:15:  so by preventing, so this is really what's important, by preventing degradative autophagy,
• 45:15 - 45:22:  you get overgrowth. In this condition, the neuronal cell is basically secreting,
• 45:23 - 45:30:  releasing outside elements of the endoplasmic reticulum and of other compartments,
• 45:30 - 45:36:  particularly endoplasmic reticulum. And those elements have this kind of effect.
• 45:36 - 45:42:  And now we need to investigate which one. I cannot say at this point that it's RTN3.
• 45:42 - 45:47:  The only thing which is interesting is that we get the effect with both the large and the small EVs,
• 45:47 - 45:53:  and both of them have RTNs, whereas only the small ones have the typical exosome markers.
• 45:53 - 45:58:  So, there has to be something, you know, if anything, we need to look at the large EVs because
• 45:58 - 46:04:  they have something different from the small EVs that make them really interesting.
• 46:04 - 46:09:  Dr. David M. Fields Great, Thierry. So, I think that wraps up
• 46:09 - 46:13:  the questions for this talk. That was fascinating. Thank you very much.
• 46:13 - 46:15:  Dr. Thierry M. Moulton Thank you. Thank you, David. Thank you.
• 46:15 - 46:19:  Dr. David M. Fields The next speaker is Ira Milosevic,
• 46:20 - 46:26:  who's an associate professor at the Wellcome Center for Human Genetics at the University of Oxford.
• 46:27 - 46:34:  She's an expert cell biologist, particularly working on neurons, and the present focus of
• 46:34 - 46:39:  her group is to understand the mechanisms of synaptic vesicle acidification and recycling
• 46:39 - 46:45:  using physiological and cell biology-based approaches. Her group is also working on
• 46:45 - 46:49:  how selected synaptic functions, for instance, plasticity, can be improved,
• 46:50 - 46:55:  and an example of this would be by lifting negative regulation of neurotransmission.
• 46:56 - 47:01:  So, welcome, Ira. Thanks very much for participating, and we look forward to your talk.
• 47:01 - 47:05:  Dr. Ira Milosevic Thank you, David, very much for the invitation,
• 47:05 - 47:11:  and thank you for Abcam and colleagues for organizing this meeting.
• 47:11 - 47:20:  So, I'm starting to share my presentation, and I'm just going through options. Yes.
• 47:21 - 47:22:  I hope you can see my slides now.
• 47:23 - 47:24:  Yes, we can.
• 47:26 - 47:32:  So, what I would like to talk today is about some thoughts that we have about,
• 47:33 - 47:39:  or ideas that we have about possibilities to improve synaptic transmission in the failing
• 47:39 - 47:45:  brain or in the aging brain when aging and failing coincide with each other.
• 47:46 - 47:54:  And I would like to start with a somewhat thought-provoking idea that I personally think
• 47:54 - 48:03:  that the brain is a product of evolution, and despite that, we don't really know the principles
• 48:03 - 48:11:  that are governing its evolution. And nonetheless, we have started playing with augmenting some of
• 48:11 - 48:15:  the selective brain functions, and not now. I mean, we have been doing this for centuries
• 48:15 - 48:21:  with improved diet and exercise and so on, but more recently, we are also including
• 48:21 - 48:27:  technological aspects like this verbal electronics, neurostimulation, and neurotechnology
• 48:28 - 48:35:  that is more and more discussed. And for a couple of decades already, we are playing with
• 48:35 - 48:40:  neurotrophics, different types of pharmacological agents that are also affecting brain function.
• 48:40 - 48:49:  And one would, like when one thinks about it, there is obviously a very important relevance
• 48:50 - 48:55:  for this type of research, either in aging or the type of neurodegenerative diseases like
• 48:55 - 49:01:  Parkinson's disease. So, there is certain plasticity in the brain, and we are hoping
• 49:01 - 49:09:  to understand how we can play in a way that we can increase selective brain function,
• 49:09 - 49:17:  particularly in the case when the brain starts to fail. So, the level on which we are kind of
• 49:17 - 49:25:  trying to address this is on the contact points between the neuronal cells where the communication
• 49:26 - 49:34:  between neurons happens. So, our neuron system relies on the quantum neurotransmitter release
• 49:34 - 49:41:  and synaptic vesicle recycling. And it has to be so since the billions of small synapses are often
• 49:41 - 49:47:  very far away from the cell body. So, the small synaptic vesicles that are filled with neurotransmitters
• 49:47 - 49:54:  that fuse with the postsynaptic side need to be locally remade and relatively fast because
• 49:54 - 50:00:  neurotransmitters are a fast process. And this is a well-studied process for almost five decades
• 50:00 - 50:07:  right now. And I would just like to emphasize the key players there, small synaptic vesicles,
• 50:07 - 50:13:  and bring your attention that besides the lipids, there are a couple of proteins that are
• 50:13 - 50:19:  important for the physiology of synaptic vesicles. Obviously, synaptobrevin 1-2, those are
• 50:19 - 50:25:  SNARE proteins which are important for fusion, and I'll just talk more in a second. But then the way
• 50:25 - 50:31:  how synaptic vesicles get filled with neurotransmitters is by the action of V-ATPase that
• 50:31 - 50:37:  pumps the protons and then those protons get exchanged for neurotransmitters using neurotransmitter
• 50:37 - 50:44:  transporters. And then there are also peripheral proteins like RAP3 or synapsin that are basically
• 50:44 - 50:50:  deciding the destiny, the fate of synaptic vesicles once after they are remade through the process of
• 50:50 - 50:57:  endocytosis. To just remind you a little bit of how exocytosis of small synaptic vesicles and
• 50:57 - 51:05:  general secretory vesicles work. So on the vesicle side, we have 1,2-synaptobrevin that contributes
• 51:05 - 51:10:  one SNARE domain towards the formation of the SNARE complex. And on the plasma membrane,
• 51:10 - 51:15:  there are two proteins, syntaxin1 and SNAP25, with SNAP25 contributing two SNARE domains.
• 51:16 - 51:25:  And often on the plasma membrane, it's believed that synapsin and SNAP25 make clusters,
• 51:25 - 51:31:  which are also considered like a hot zone where secretory vesicles fuse. It's just like
• 51:31 - 51:38:  shown in the illustration here on the right. So my lab is working on different aspects of the synaptic vesicle
• 51:38 - 51:45:  cycle. So we are playing a bit with exocytosis. Our bread and butter studies still come
• 51:45 - 51:50:  from the aspects of studies of endocytosis, particularly clathrin-mediated endocytosis that
• 51:50 - 51:56:  is depicted here. A couple of years ago, we started to be interested in what happens
• 51:56 - 52:02:  after the new synaptic vesicles are formed. How do they acidify? How do they get refilled?
• 52:02 - 52:08:  And basically, how do they mature by maturation? I mean, how do they acquire peripheral proteins
• 52:08 - 52:16:  and basically how the fate of them is decided, whether they will go in the cluster of reserved
• 52:16 - 52:23:  pool of vesicles or they will immediately fuse with the plasma membrane to secrete neurotransmitters
• 52:23 - 52:33:  again. And so given that these processes are quite well studied, the way how we look at it
• 52:33 - 52:41:  is from the perspective of how can we improve certain aspects of synaptic transmission
• 52:41 - 52:47:  when it starts failing, so that the neurons and synapses stop performing optimally.
• 52:48 - 52:55:  And today, I was planning to tell you two strategies out of three. I'll just also
• 52:55 - 52:59:  mention the third strategy that we have in the lab is kind of looking at how endocytosis
• 53:00 - 53:08:  helps or it's linked to protein homeostasis and also neurodegenerative diseases.
• 53:08 - 53:14:  But what I will tell you today is pretty much like four stories, two of which are published.
• 53:14 - 53:20:  I will go very fast through them and some unpublished data that we have along the lines. So the strategy,
• 53:21 - 53:26:  the first strategy that I was thinking to discuss today is how to play with negative regulation
• 53:26 - 53:33:  of neurotransmission. And I will explain in a second what do I mean by it. So as I pointed
• 53:33 - 53:40:  before, the synaptic vesicles are basically coming to the particular sites where the
• 53:40 - 53:47:  SNARE protein syntaxin and SNAP25 are enriched. However, there is a group of proteins that is
• 53:47 - 53:54:  basically cytosolic and has a SNARE domain nonetheless, and it's able to bind syntaxin and
• 53:54 - 54:03:  SNAP25. And for one of these proteins called tomosyn, it's been suggested that it basically
• 54:03 - 54:11:  competes for the binding site of vesicles by titrating syntaxin and SNAP25, which is in the
• 54:11 - 54:20:  plasma membrane. So we work on, we kind of were interested if there are more proteins like tomosyn
• 54:20 - 54:26:  and we came across a protein called amisyn, which is also a syntaxin binding protein, also cytosolic,
• 54:26 - 54:34:  and very few studies have been basically made with this protein. So, so few that I can kind
• 54:34 - 54:42:  of put them on one slide and give you a complete overview of what's been known. So the Scheller's
• 54:42 - 54:49:  lab in the early 2000s has identified this protein as, and suggested that it may regulate SNARE
• 54:49 - 54:55:  complex. And a few years later, Bob Burgoyne's lab has linked it to exocytosis both in
• 54:55 - 55:01:  syntaxin-independent and syntaxin-dependent manners. Further studies have been suggesting it's
• 55:02 - 55:11:  like mutations have been found to be linked with autism and diabetes. And more recently,
• 55:11 - 55:18:  a nice study has been shown also that amisyn is helpful to control the fusion core of large
• 55:18 - 55:25:  dense core vesicles in insulin-secreting cells. So it took us a little bit of time to build,
• 55:25 - 55:31:  to start working with amisyn first, because we had to build tools, we cloned it, made the
• 55:32 - 55:39:  anti-amisyn antibody, a specific one, and also generate transgenic mice. So when we started
• 55:39 - 55:43:  working on this a couple of years ago, the information that we have from the literature
• 55:43 - 55:48:  is that this is a small cytosolic protein of 24 kilodaltons that has a SNARE domain that's
• 55:48 - 55:56:  known to bind syntaxin. And we went to repeat some of the experiments and expand on them.
• 55:56 - 56:01:  And indeed, we found that this protein makes a SNARE complex with syntaxin-assembly 5
• 56:01 - 56:05:  readily, not only the SNARE domain, which is shown here, but also the full-length protein.
• 56:06 - 56:09:  Part of this research on this protein has been limited by
• 56:10 - 56:16:  capabilities to express the protein before because it has some somewhat unusual
• 56:17 - 56:22:  amino acids and codons. So we played a lot to find the right expression system. But once we
• 56:22 - 56:29:  were able to express the whole length amisyn, the different types of experimental possibilities
• 56:30 - 56:34:  opened up. So we were able to study how this protein actually forms the SNARE complex
• 56:35 - 56:45:  and is it able to basically compete with synaptobrevin 1 and 2 in this type of aspect.
• 56:45 - 56:51:  So the black line here shows the formation of the SNARE complex by three recombinantly expressed
• 56:51 - 56:56:  proteins: syntaxin 1, synaptobrevin 5, and a short version of synaptobrevin. And then in a
• 56:56 - 57:03:  whole system, we would add increasing amounts of a full-length amisyn. And as more amisyn we add,
• 57:04 - 57:10:  the more we see the inhibition of the classical formation between synaptobrevin and syntaxin,
• 57:10 - 57:17:  suggesting that this protein is basically interfering with the classical
• 57:17 - 57:22:  SNARE complex formation. And not only that, but it also interferes with the fusion,
• 57:22 - 57:28:  because when these recombinant proteins are put in the liposomes, some of them are fluorescent,
• 57:28 - 57:35:  so we can measure the fusion between them. We see that increasing amounts of amisyn added in such
• 57:35 - 57:45:  a system inhibit the fusion of those liposomes. So next, we were interested to see where
• 57:46 - 57:51:  amisyn is present in the cell, and previous studies suggested that this is a soluble protein.
• 57:51 - 57:55:  When we cloned it and expressed it in PC-12 cells, the neurosecretory cell line,
• 57:55 - 58:02:  we found it to be enriched in the plasma membrane. And then subsequent experiments
• 58:02 - 58:07:  with neurotoxin showed us that this enrichment is actually not due to the SNARE domain. So it must
• 58:07 - 58:14:  be another part of amisyn which recruits a protein to the plasma membrane. So we looked
• 58:14 - 58:22:  at analogies and found high analogies with exocyst subunit Sec3. We tried to crystallize the
• 58:22 - 58:28:  protein; we were not successful in that aspect, but based on homology and protein prediction
• 58:28 - 58:35:  studies, we postulated that the N-terminal of amisyn is basically a pH domain. And
• 58:38 - 58:45:  when we made a specific mutation in this pH domain, we found that the protein does not bind
• 58:46 - 58:50:  to the plasma membrane anymore. I'll tell you a little bit more in detail about how we
• 58:50 - 58:55:  discovered this binding. So just to conclude this slide, amisyn is a conserved,
• 58:55 - 59:02:  vertebrate-specific protein; also, very little work is done in lower organisms like
• 59:02 - 59:09:  C. elegans, or no work is done. So amisyn, we found, interferes with the SNARE complex
• 59:09 - 59:14:  assembly and liposomal fusion. Interestingly, when you stimulate those neurosecretory cells,
• 59:14 - 59:20:  amisyn tends to translocate to the plasma membrane very rapidly. And we found that this
• 59:20 - 59:26:  is interestingly not done in a calcium-dependent way. As you know, exocytosis is a calcium-sensitive
• 59:26 - 59:31:  process, and yet this recruitment of amisyn to the plasma membrane is not. So we were curious
• 59:31 - 59:40:  about what actually recruits amisyn to the plasma membrane in the fusion sites. We turned to lipids
• 59:40 - 59:47:  next and performed some dot-blot experiments in which we figured out that the molecule that
• 59:47 - 59:52:  is actually recruiting amisyn to the plasma membrane is PIP2. And this experiment I'm quite
• 59:52 - 59:59:  fond of. It's a co-sedimentation experiment in which we combined artificially made liposomes
• 60:00 - 60:06:  with different compositions of PIP2 together with recombinantly expressed amisyn, and then subjected it to
• 60:06 - 60:10:  centrifugation. So if amisyn binds to the liposome, we would find it in the pellet.
• 60:10 - 60:17:  If not, it stays inside the supernatant. And when liposomes contain no PIP2, then all protein
• 60:17 - 60:22:  basically stays in the supernatant. Nothing is found in the pellet. But as we increase the amount of PIP2
• 60:22 - 60:32:  in those artificial liposomes, we also increase the binding of amisyn to the liposome,
• 60:32 - 60:36:  suggesting that PIP2 is important for the recruitment of amisyn to the plasma membrane.
• 60:38 - 60:45:  In the next line of work that was started by Julian Lin in the lab, previous work was led by
• 60:45 - 60:52:  Ilona Kontradiuk. We also generated amisyn mutant mice, and those mice are viable.
• 60:53 - 60:57:  And I'm not going to go into the details. We are preparing the story about it at the moment.
• 60:57 - 61:03:  But I just want to bring your attention to what basically lack of amisyn does to
• 61:04 - 61:12:  synaptic release and in general release probability. So without amisyn, we see
• 61:12 - 61:17:  more miniature EPSC, meaning that more synaptic vesicles are fusing. I'm not showing you this
• 61:17 - 61:23:  data here, but what I'm showing you is one example of EPSC in a wild type in comparison to the
• 61:23 - 61:32:  littermate control. So in general, we see much higher EPSC, so more secretion from the neurons
• 61:32 - 61:38:  that are missing amisyn. And also, since the animals are viable, and a lot of protein is
• 61:38 - 61:45:  actually expressed in the hippocampus, we analyzed some of the hippocampal circuits and also saw that
• 61:45 - 61:50:  the release probability in these animals is increased. To me, it's kind of fascinating
• 61:50 - 61:57:  how much plasticity is there, how much more synaptic transmission can be increased just in
• 61:57 - 62:02:  the absence of one protein. I'm now sorry that I didn't put heterozygotes in there because they
• 62:02 - 62:08:  would be somewhere in between, which suggests that it doesn't really have to be the absence of the
• 62:08 - 62:14:  whole protein; even the removal of part of the protein has an effect. So the follow-up that
• 62:14 - 62:20:  we do along this line, and also with our recent joining of the Wellcome Centre for Human
• 62:20 - 62:29:  Genetics, is to basically look for specific examples within the human population where people have
• 62:30 - 62:43:  particularly incredible skills in some fields like mathematics, music, or super memory, and yet often
• 62:43 - 62:50:  they do suffer from defective social interactions,
• 62:50 - 62:56:  defective behavior, and restricted interests. The reason why this attracted me here is that the
• 62:56 - 63:02:  behavior studies that we perform on amisyn mutant mice, where we found that certain
• 63:02 - 63:12:  aspects of brain function, like learning and memory, may be enhanced while those mice
• 63:14 - 63:19:  express a lot of repetitive behavior and also have defective interactions. So much more studies
• 63:19 - 63:26:  have to be done; those are just preliminary findings, but one line of work would be going to explore
• 63:26 - 63:36:  this negative inhibition of exocytosis and try to see how, just to summarize, how we can elevate
• 63:36 - 63:42:  transmission. So I made a little conclusion slide and also to indicate how we believe amisyn works
• 63:42 - 63:48:  and competes with synaptic vesicles for their fusion, which I also think is important to limit
• 63:48 - 63:53:  the amount of fusion, so to fuse the vesicles that are already primed, but possibly the ones that
• 63:53 - 63:58:  need to come, because calcium is still around. There is another layer of regulation by just having those
• 63:58 - 64:04:  vesicles finding the right site of fusion. So as I mentioned to you before, amisyn is a vertebrate
• 64:04 - 64:11:  specific protein that is recruited to the plasma membrane upon stimulation, but interestingly in a
• 64:11 - 64:17:  calcium-independent way, but PIP2-dependent way. It interferes with the SNARE complex assembly and
• 64:17 - 64:22:  liposomal fusion. I didn't show you the data in the cells, but we also showed it in cells; it kind of
• 64:22 - 64:28:  interferes with the, I did show you that when we decrease the amount of amisyn, when we put more
• 64:28 - 64:36:  amisyn inside the cell, we actually can block endocytosis. The elevated amisyn levels reduce
• 64:36 - 64:42:  the number of released secretory vesicles and the lack of amisyn alters synaptic vesicle release
• 64:42 - 64:47:  probability and leads to social and repetitive behavior abnormalities. And with that, I would
• 64:47 - 64:54:  like to move and turn to the second strategy that we consider in the lab on how we can improve
• 64:54 - 65:00:  synaptic transmission by playing with organelle acidification and synaptic vesicle refilling.
• 65:00 - 65:07:  And there, I would like to bring your attention to the VATPase, a proton pump that is present on
• 65:07 - 65:11:  all organelles that are acidified, but also including secretory vesicles and small synaptic
• 65:11 - 65:19:  vesicles. And why VATPase is important for synaptic vesicle function. Somehow nature defined,
• 65:20 - 65:27:  designed the whole system in a way that the VATPase pumps protons, and then those protons get
• 65:27 - 65:34:  exchanged for neurotransmitter in a way that which neurotransmitter is to be enriched
• 65:34 - 65:39:  in a synaptic vesicle depends on neurotransmitter transporters that that vesicle has.
• 65:39 - 65:46:  And just to advertise a study that we have on bioRxiv at the moment, we are also
• 65:46 - 65:56:  looking in collaboration with Young Labs and SEVA on how many neurotransmitter transporters each
• 65:56 - 66:01:  individual synaptic vesicle has; that work is under a new neuron at the moment. So to go back
• 66:01 - 66:07:  to our regulation of VATPase, we were interested in the regulation of VATPase on synaptic vesicles.
• 66:07 - 66:14:  And as any protein, transmembrane protein is regulated by transcription and the protein
• 66:14 - 66:20:  stability, as well as lipids in this case, phospholipids and phosphoinositol 3,4-bisphosphate.
• 66:20 - 66:25:  But what most of us know from the tech, like textbook knowledge, is that there is
• 66:25 - 66:34:  a dissociation between this V1 domain or V1 sector from the V0 sector and some regulatory subunits
• 66:34 - 66:44:  V1C as well that have their own kinetics. So, to study the activity of the VATPase,
• 66:44 - 66:52:  together with Rang Haryan, we're developing assays to look either at the acidification
• 66:52 - 66:59:  through the fluorescent molecules, which are fused to the transmembrane domain of some
• 66:59 - 67:05:  synaptic vesicle proteins, and they are pH-sensitive moieties, or by using voltage dyes
• 67:05 - 67:13:  that are the dyes that are sensitive to voltage differences in membrane charges.
• 67:13 - 67:19:  And the work that I would like to emphasize along this line is how we believe
• 67:20 - 67:26:  RAB-connectin links the acidification of vesicles and recruitment of some peripheral
• 67:26 - 67:30:  proteins. But before I go there, I just want to summarize in one slide
• 67:31 - 67:38:  the paper and the model that we propose on how we think the VATPase is blocked
• 67:38 - 67:46:  while the vesicle is coated with clathrin. So, this is just an animation that summarizes
• 67:46 - 67:53:  the model that we have. We adopted the model that Tom Kirchhausen and collaborators developed
• 67:53 - 67:59:  at Harvard on how the clathrin-coated vesicle works and put VATPase in the center of one of these hexagons.
• 68:00 - 68:07:  It's too large to fit into the pentagons on the VATPase of the clathrin coat, and the adapter
• 68:07 - 68:15:  proteins also help to recruit the clathrin tripods. And in order for six
• 68:15 - 68:22:  tripods to fit in, the regulatory subunit V1H, displayed in green, needs to be pushed in.
• 68:22 - 68:28:  So we actually use a lot; we combine our data with a lot of what's present in the literature, and
• 68:31 - 68:36:  we are working a little bit more along the lines to prove this model. So,
• 68:37 - 68:42:  an interesting set of data that we gained while studying acidification and neurons
• 68:43 - 68:49:  came from the fact that when we block the acidification of synaptic vesicles or
• 68:49 - 68:54:  general acidification of vesicles by different inhibitors like bafilomycin or salinomycin,
• 68:55 - 69:01:  we also seem to block the recruitment of peripheral protein RAB3, which is a small
• 69:01 - 69:06:  GTPase involved in synaptic vesicle trafficking, endocytosis, and particularly synaptic plasticity,
• 69:06 - 69:11:  that we are also interested in. So, this small molecule cycles between the surface of synaptic
• 69:11 - 69:19:  vesicles and the cytoplasm, and as such, again, I won't go into the biology of RAB3,
• 69:19 - 69:24:  but it's really essential for the trafficking of small synaptic vesicles as well as for acetylcholine.
• 69:25 - 69:33:  So the experiments that showed that when you stimulate the neuronal synapse,
• 69:34 - 69:38:  neurons, particularly look at the synapse and the amount of RAB3 in the synapse,
• 69:39 - 69:46:  and you can see that RAB3 tends to diffuse from the synapse upon stimulation, and then in a couple
• 69:46 - 69:51:  seconds afterwards, it comes back, presumably when new synaptic vesicles are made, as this is
• 69:51 - 69:56:  rather strong stimulation of 300 detection potential that is applied there. So what we did,
• 69:56 - 70:03:  we performed the very same experiment in the presence of bafilomycin, an inhibitor of VATPase,
• 70:03 - 70:08:  and it's the same data; I mean, it's a very similar data set, it's just normalized a
• 70:08 - 70:14:  little bit differently, but in normal wild-type neurons without bafilomycin treatment,
• 70:14 - 70:19:  we see that RAB3 diffuses from the synapse and then basically recovers and comes back,
• 70:19 - 70:25:  while in the presence of bafilomycin, somehow, this recruitment of RAB3 back to the synaptic
• 70:25 - 70:31:  vesicle is not very efficient. So we were puzzled by that, and we're looking for the
• 70:31 - 70:36:  potential links of what can be linking this acidification and the recruitment of peripheral
• 70:36 - 70:42:  proteins, or how does the surrounding of the vesicle know that the vesicle is now
• 70:42 - 70:49:  acidified and full of neurotransmitters and can be sorted to be either in a recycling pool,
• 70:50 - 70:56:  reserve pool, or fused with the plasma membrane again. And then in the literature, we came across
• 70:56 - 71:08:  the protein RAB-connectin 3-alpha, and also we looked more into the control of VATPase and went to
• 71:09 - 71:15:  study the literature in yeast, where this association and dissociation have been originally discovered.
• 71:15 - 71:23:  So yeast also have a RAVE complex, which is needed for the VATPase assembly
• 71:23 - 71:28:  and the stability of the VATPase. And one of these key components of the RAVE complex is the
• 71:28 - 71:36:  protein called RAV1 and RAV2. So RAV1 corresponds to mammalian RAB-connectin 3-alpha, while RAV2
• 71:36 - 71:43:  is mammalian RAB-connectin 3-beta. And this protein has been shown to bind to both
• 71:44 - 71:51:  RAB3-GAP and GAF. So RAB3, GTP-GTP exchange factor, and also the small GTPase that activates
• 71:51 - 72:00:  proteins for RAB3. So that brought our attention in combination with another paper
• 72:01 - 72:09:  from Gus Smith's lab that in 2012 suggested that the RAB-connectin, the gene name is DMXO2,
• 72:09 - 72:16:  is basically interacting with the VATPase. Further studies have actually, again, this is a protein that
• 72:16 - 72:23:  not much is known. It's a rather large protein of 340 kilodaltons. It contains a large number of
• 72:23 - 72:32:  these WD40 domains and presumably may be involved in different multiprotein complex assemblies.
• 72:32 - 72:38:  It's been also suggested to have a link in endocytosis and Wnt signaling as well as Notch
• 72:38 - 72:43:  signaling, which may explain why it's essential for life. So without RAB-connectin, animals are
• 72:43 - 72:50:  dying at embryonic day seven, and mutations in these proteins have more recently been found to
• 72:50 - 72:55:  be linked with mental retardation, hearing loss, and endocrine polyneuropathy syndrome.
• 72:56 - 73:04:  So as with amisyn, and given that not much work has been done in vertebrate and
• 73:04 - 73:11:  mammalian cells in general with this protein, we first went to clone this large DNA, which was
• 73:11 - 73:18:  three times larger than the EGFP vector, and upon cloning it and making the antibodies,
• 73:18 - 73:24:  we found that indeed, this is a synaptic vesicle protein, so it accumulates largely on the small
• 73:24 - 73:38:  synaptic vesicles. But in addition to that, it localizes very nicely with RAB3 which is a classical marker for small synaptic vesicles, when you look at the cell body, we also see
• 73:38 - 73:43:  a punctate staining in the cell body of neurons, suggesting that RAB-connectin is also present
• 73:43 - 73:47:  on other organelles of the endosomal pathway. We are able to purify
• 73:47 - 73:52:  small synaptic vesicles as well as clathrin-coated vesicles, and we find RAB-connectin to be present
• 73:52 - 73:58:  in both fractions. We also found it to be present already on the clathrin-coated pits.
• 73:58 - 74:04:  In this particular case, to study, because clathrin-coated pits are very transient inside
• 74:04 - 74:12:  the cell, we took advantage of dynamin knockout cells that accumulate the clathrin-coated pits,
• 74:12 - 74:17:  and then looked to see whether RAB-connectin co-localizes with those structures, and we found it to be
• 74:17 - 74:25:  permanently expressed there. So I don't have time to go into all the details that kind of help us
• 74:25 - 74:32:  define a hypothesis that we work on in the lab. So we think that VATPase is a quality
• 74:32 - 74:37:  controller during endocytosis. The reason for that is that in an independent study, we found
• 74:37 - 74:42:  it to interact with some endocytic protein, including endophilin, and that basically also
• 74:42 - 74:50:  promotes stable VATPase activity. And the reason why we believe that this is the case is that
• 74:52 - 74:59:  from a theoretical standpoint, each VATPase has one to two VATPases, and if they
• 74:59 - 75:05:  don't have a properly integrated VATPase, the small synaptic vesicle that is being reformed
• 75:05 - 75:09:  through endocytosis is pretty much useless, because then it cannot be filled with neurotransmitters.
• 75:09 - 75:16:  So it would make sense at some point during the process to ensure that there is a functional
• 75:16 - 75:22:  VATPase over there. And as Guy Smith suggested, that RAB-connectin binds to one of these regulatory
• 75:22 - 75:26:  subunits, we won't see that also; it's the latest one to come inside the complex.
• 75:26 - 75:30:  And we found that RAB-connectin was also interacting with the machinery for endocytosis;
• 75:30 - 75:35:  we believe that this may play a part in this quality control during endocytosis.
• 75:35 - 75:40:  But it could be that there is more to that protein. So as I mentioned before, it shows
• 75:40 - 75:49:  embryonic lethality, yet we can develop in vitro, we can make neurons that develop normal synapses
• 75:49 - 75:55:  in the absence of RAB-connectin 3-alpha. And since we have synapses, we went to analyze them and see
• 75:55 - 76:01:  how they look. So we noticed that, obviously, when we did the first electron microscopy,
• 76:01 - 76:08:  they differ quite a bit from the synapses of the normal wild type animals, in a way that they have
• 76:08 - 76:14:  an average of a little bit less small synaptic vesicles, accumulation of those endosomal-like
• 76:14 - 76:20:  structures. But also the vesicles are not normally distributed inside the synapse; they often look
• 76:20 - 76:26:  like scattered around. So this type of clustering or packing density seems to be affected. And also,
• 76:26 - 76:32:  many of them don't seem nice and round, but often seem oval. So they don't have this nice
• 76:32 - 76:39:  and round structure as typically synaptic vesicles have. When we examined some synapses by
• 76:40 - 76:44:  physiological or electrophysiological methods, I won't go into the details with everything here
• 76:44 - 76:48:  and EPSCs and so on. I'll just show you the result of miniature EPSCs, which I find interesting,
• 76:48 - 76:52:  because it's not only that the frequency of miniature EPSCs is reduced, but also the amplitude.
• 76:53 - 76:59:  Amplitude basically tells us how much neurotransmitter, indirectly, how much of neurotransmitter
• 77:02 - 77:06:  those synaptic vesicles retain, and being reduced suggests that there are difficulties or issues
• 77:06 - 77:13:  with filling. So given the link of RAB-connectin with the VATPase, we next explored whether
• 77:15 - 77:21:  the acidification is altered in the absence of RAB-connectin 3-alpha. And that indeed seemed to be the case.
• 77:22 - 77:32:  So we employed classical fluorine assays. And here, so the, I think I explained already earlier how the
• 77:32 - 77:38:  fluorine works. I will just move on to the experiment that we did along this line. So we expressed
• 77:38 - 77:43:  RAB-connectin 3-alpha, I'm sorry, we expressed the RAB-fluorine in this case in wild-type and
• 77:43 - 77:48:  RAB-connectin 3-alpha knockout neurons, and then subjected those neurons to ammonium chloride.
• 77:48 - 78:00:  And that basically decreases the amount of fluorescence in the model. And we found that most of the
• 78:00 - 78:06:  fluorescence, or like crunches it on the surface. And we found that most of the fluorescence in this case is not
• 78:06 - 78:14:  quenched, presumably because it's enclosed inside recycling organelles. So indeed, endocytosis or
• 78:14 - 78:19:  vesicle acidification, both of them can be altered. But in order to prove whether it's really a
• 78:19 - 78:25:  vesicle or endocytosis, or maybe both, we perform an additional set of experiments for which we needed
• 78:25 - 78:31:  to develop one method first. So previously we purified synaptic vesicles from either
• 78:31 - 78:39:  mouse or rat brains. Traditionally, purification is done from one cow brain or two pig
• 78:39 - 78:46:  brains. However, in this case, because animals were lethal, we didn't have so much material.
• 78:46 - 78:52:  So we could culture neurons as I showed you before. So we collected those neurons and made
• 78:52 - 78:59:  synaptic zones out of them, then broke those synaptic zones and made them interact with the
• 78:59 - 79:08:  beads that were decorated with anti-SNAP-2, the synaptobrevin 2 antibody. And then those
• 79:08 - 79:18:  beads were subsequently washed. And this dataset actually made my day when Zora Farsi and
• 79:18 - 79:23:  Sudhoya Govindeshankaran, who worked on this project, showed it to me for the first time. It also shows
• 79:23 - 79:31:  that those isolated synaptic vesicles are active as they can acidify as measured by the voltage
• 79:31 - 79:39:  dye, fluorovolts in this particular case. So we then proceeded to isolate small synaptic vesicles
• 79:40 - 79:49:  both from wild type as well as the RAB-connectin-3 knockout. And notice that the vesicles that are
• 79:49 - 79:57:  isolated from knockout do not acidify properly. Or I would say, rather, that there are two
• 79:57 - 80:00:  populations; one seems to be acidic.
• 80:00 - 80:08:  One seems to acidify okay, and the other one doesn't seem to acidify at all. So on average, the amount of acidification is decreased.
• 80:08 - 80:20:  We did not, however, see that endocytosis is affected when we performed these classical experiments using dyes that are not pH sensitive, like FM1-43.
• 80:20 - 80:35:  Now, to go back to this previous idea that we had on when we block acidification with bafilomycin, that bafilomycin does not properly recycle at the synapse.
• 80:35 - 80:48:  Interestingly, one of the proteins whose distribution is not properly maintained in the RAB-connectin-3 knockout is exactly RAB3.
• 80:48 - 81:00:  This also came as a surprise. I showed you previously how those synapses develop normally. We had an example where you could not really see a difference between wild type and RAB-connectin-3 knockout neurons.
• 81:00 - 81:15:  In contrast to RAB-connectin-3, here in green, you see a lot of small synaptic vesicles, a lot of synapses in the wild type, but somehow it looks like there are many fewer synapses.
• 81:15 - 81:25:  Actually, the number of synapses does not change; what changes is that they are often depleted of RAB3.
• 81:25 - 81:40:  The last dataset that I would like to share with you today is that classical experiment in which we would stimulate the wild type neurons and see how RAB3 dissociates and then recovers from the synapse.
• 81:40 - 81:56:  We did the same experiments in wild type and RAB-connectin-3 knockout mice, and noticed that in this particular case, RAB3 does not recover, at least in the timeframe in which we're looking.
• 81:56 - 82:12:  Basically, after stimulation, it does not come back. So also keep in mind that the baseline is already lower, as you can see in the normalized first part, so there's less RAB3 to start with, and then even that small amount can dissociate,
• 82:14 - 82:32:  but it does not dissociate with synaptic residuals efficiently after the stimulation is done. This is a specific case for RAB3. We tested RAB27, RAB26, and RAB5 and did not find any difficulties in recruiting these proteins after stimulation.
• 82:32 - 82:48:  So, just to conclude my second part of the talk, we believe that RAB-connectin-3 is a VATPase quality controller during endocytosis that promotes stable VATPase activity, as well as recruitment of RAB3 to small synaptic vesicles.
• 82:48 - 83:08:  Lack of it impairs recycling and neurotransmission, but interestingly enough, endocytosis. Some of the previous studies, sometimes it's tricky to study endocytosis if you're using pH-sensitive probes, as this can result from defective acidification or defective endocytosis.
• 83:08 - 83:31:  Also, the block of synaptic vesicle acidification impairs the recruitment of RAB3, as I mentioned, and interestingly, the synapses show variable resilience to failure in use with quality VATPase activity. I can discuss this a little bit more if you are curious what I mean by that.
• 83:31 - 83:51:  And the goal here, as in the previous, is to see if we can, during this decay of synaptic function in aging disease, improve it by promoting synaptic vesicle acidification, by refilling, or simply by reducing the level of synaptic vesicle competitors, as I showed before.
• 83:51 - 84:11:  We are particularly interested in hearing loss. So we have a line of work in the lab that involves the hair cells in our ears, as well as learning and memory. And with that, I would like to thank you.
• 84:12 - 84:35:  First, I would like to thank people who performed this nice work in the lab, Sindhuya Gaurishankaran, who almost single-handedly, along with me and another colleague, Zora Farsi, who was a PhD student in Ranghad Yadav's lab, analyzed the RAB-connectin function, as well as Jolin Lin and Ilona Kontradiuk for the first part of work on Amisyn.
• 84:36 - 84:48:  Next, I would like to thank our funding, particularly in Germany, DFG and Alexander von Humboldt, different types of SFBs, Schramm and Engelhorn, and more recently, John Black and John Feldschers' presentations.
• 84:49 - 84:57:  I would also like to mention that the PhD and postdoc positions are available now that we have moved and are ready to restart again. Thank you very much.
• 84:57 - 85:04:  Thank you, Ira. That's very interesting. And there are quite a few questions, so I hope it's okay to ask a few.
• 85:05 - 85:19:  The first question was, how do VATPase inhibitors affect neurotransmission? I think I know what the answer is, but maybe you can start with that.
• 85:20 - 85:35:  So obviously, acidification is affected; the refilling is affected; neurotransmission is affected. There are a couple of studies along this line. We expanded them a little bit further by doing repetitive stimulation.
• 85:35 - 85:44:  So what surprised me is that when we put bafilomycin on those neurons, I thought that neurons would be completely silent.
• 85:45 - 85:54:  But interestingly, they were not immediately completely silent after this classical, let's say, three-time detection potentials; some synapses were still active.
• 85:54 - 86:09:  So we went on to repeat the stimulations. And there it seems to be an interesting phenotype that some synapses are just more resistant to the block of acidification than the others are.
• 86:10 - 86:15:  And I think that's an interesting part that we would like to follow up a little bit more.
• 86:15 - 86:34:  So can I follow up? Do you know how the decreased acidification mediated by the bafilomycin actually affects the RAB3 recruitment to the synaptic vesicles? Do you understand what the mechanism is?
• 86:35 - 86:58:  So we believe, and this part, we are kind of building on studies that were published in the early 2000s as well, which suggest that in order to recruit RAB3, there is a group of proteins that are helping that, like RAB3-GAPs, that are helping the recruitment, right?
• 86:59 - 87:19:  And the RAB3-GAP is directly bound to RAB-connectin. That's how RAB-connectin originally was identified in a pulldown with RAB3-GAP. So we believe that RAB-connectin interacts with VATPase on one side, and then on another side recruits RAB3-GAP that directly helps the recruitment of RAB3.
• 87:20 - 87:26:  So this may not be the whole story, but at least it's a hypothesis under which we work.
• 87:26 - 87:42:  Okay. The second question was, and I think you mentioned in your slide, but it might be worth just reminding the audience, are there pathologies associated with the effects of synaptic vesicle acidification or lack of RAB-connectin 3A?
• 87:43 - 87:57:  Yes. So I mentioned RAB-connectin, right? I mentioned that mutations in RAB-connectin are linked to hearing loss, mental retardation, and this interesting PEMS syndrome, delayed onset of puberty as well, and mental retardation in that group of patients.
• 87:58 - 88:20:  I think more interestingly, there is, obviously, if you block acidification refilling, there will be quite a lot of pathology that results both in brain development if it happens early on during development, but I think the increasing amount of evidence also accumulates
• 88:20 - 88:37:  that defective refilling with neurotransmitters, which results in part from defective acidification, may cause issues in neurodegenerative diseases as well, like Parkinson's disease.
• 88:41 - 88:59:  The last question I've got here is, in addition to synaptic vesicles, does RAB-connectin associate with other organelles that acidify? Maybe one can sort of expand that question: is RAB-connectin always on RAB3-associated vesicles?
• 89:00 - 89:16:  No, and that's an interesting question. I think a really cool part as well because RAB-connectin interacts with VATPase, so we were just curious as we saw on that neuronal cell body, we saw a bunch of puncta there.
• 89:16 - 89:33:  So we went to make a co-localization study with the markers of early endosomes like RAB1, RAB5, RAB7, for RAB11, for recycling endosomes, and lysosomal markers as well, and we found RAB-connectin to be present on all of these organelles that acidify.
• 89:33 - 89:48:  So the thing is: what does RAB-connectin do on these other organelles, right? Obviously, it does not recruit RAB3, which is specific for small synaptic vesicles, but we think that it may have a role in recruitment of some other proteins.
• 89:48 - 90:04:  So I think it would be interesting to know which ones for lysosomes we would be actively; I don't know, I'm looking for collaborators who would help us to nail this down on lysosomes because I think it's quite interesting to know how does the surrounding
• 90:04 - 90:17:  lysosome sense the extent to which lysosomes are acidified, or during diseases, when the lysosomes are not properly acidified, what kind of changes on its surface and how is that directly affected by acidification.
• 90:17 - 90:27:  Interesting. So thank you very much for your talk and your new insights and all the unpublished data. That's really cool. Thank you, Ira.
• 90:27 - 90:28:  Thank you.
• 90:29 - 90:43:  Our next speaker is Anna-Marie Kureva, and I've known Anna-Marie for a long time. Actually, I knew her work even before I met her, and I think I met her first maybe 20 years ago almost.
• 90:43 - 91:02:  You know, she's been, and she is one of the leaders in protein degradation and autophagy, starting with her work when she was in Fred Dice's lab where she characterized chaperone-mediated autophagy, and subsequently she's made many seminal contributions in autophagy,
• 91:02 - 91:08:  but maybe particularly in forms of selective autophagy.
• 91:08 - 91:11:  She's a recognized leader, therefore.
• 91:11 - 91:24:  She works on protein degradation and its intersection with the biology of aging and age-related disorders, particularly neurodegenerative diseases.
• 91:24 - 91:36:  She's based at the Albert Einstein College of Medicine where she started a laboratory, and now she's co-director of the Institute for Aging Research, and I'm very grateful that she could join us this afternoon.
• 91:36 - 91:39:  Thank you, Anna-Marie.
• 91:39 - 91:41:  Thank you very much, David.
• 91:41 - 91:44:  So, it's such a great pleasure to be here.
• 91:44 - 92:05:  Thank you, Abcam, for organizing, and in particular, thanks to David for making trafficking a relaxed concept because contrary to the previous speakers, I'm going to be talking less about vesicular trafficking and more about trafficking of proteins to lysosomes for degradation.
• 92:05 - 92:08:  So, I think you can already see my screen.
• 92:08 - 92:11:  Is that correct?
• 92:11 - 92:13:  Can you see the slides?
• 92:13 - 92:14:  Yes, we can.
• 92:14 - 92:15:  Terrific.
• 92:15 - 92:21:  So, as David mentioned, we are interested in selective forms of autophagy.
• 92:22 - 92:38:  Autophagy doesn't need a definition for this group, as we have been hearing already from the previous speakers, but what I want to illustrate is that there are very different ways in which cells can do autophagy, depending on how they deliver the materials to be degraded to lysosomes.
• 92:38 - 92:54:  We have been hearing about the macroautophagy process, these LC3-mediated studies that Terry was referring to, in which the materials are sequestered inside double-membrane vesicles, they fuse with the lysosomes, and they can be in bulk or they can be very selective.
• 92:54 - 93:07:  Invaginations in the surface of the lysosomes or the endosomes can also internalize materials for degradation in these compartments in what is known as microautophagy and endosomal microautophagy.
• 93:07 - 93:20:  But what I will be mostly talking about is trafficking directly of single proteins across the lysosomal membrane for degradation through this process that is known as chaperone-mediated autophagy.
• 93:20 - 93:27:  And then I will highlight a little the relation with this form of selective microautophagy.
• 93:27 - 93:40:  So chaperone-mediated autophagy that I will be referring to as CMA for short, is a very simple form of autophagy, where you just need the secondary lysosomes, not vesicles forming anywhere.
• 93:40 - 93:57:  And what it degrades are single proteins that all contain a targeting motif. And actually, we have developed tools and you can access them through here, free to help identify the presence of these motifs in your whole proteins.
• 93:57 - 94:07:  The motif is important because it's recognized by a chaperone, HSC70, and when it's recognized it will target this protein through degradation through this pathway.
• 94:07 - 94:26:  For the lysosome, the most important component for this pathway is the lysosomal-associated membrane protein type 2A. As many of you might know, the LAMP2 gene undergoes splicing to give three variants, A, B, and C, that only differ in the cytosolic and transmembrane
• 94:26 - 94:32:  tails of these LAMPs, and only LAMP2A is essential for CMA.
• 94:32 - 94:50:  The process, as I say, is very simple. The chaperone recognizes the protein, brings it to the tail of LAMP2A, inducing multimerization of the single transmembrane protein into a translocation complex; the protein gets unfolded, and assisted by a form of HSC70 that resides
• 94:50 - 94:56:  in the lysosomes, the protein is then translocated and rapidly degraded.
• 94:56 - 95:08:  CMA is constitutively active in almost all the cells in your body, but it can be maximally upregulated in response to different stressors, as the ones that you see here.
• 95:08 - 95:25:  We have found that CMA, like any other component of the proteostasis machinery, has common functions to these components, so it contributes free amino acids by breaking down proteins inside the cell, and it also contributes to quality control. Damaged proteins or altered
• 95:25 - 95:31:  proteins can be selectively targeted for degradation and elimination from the cell.
• 95:31 - 95:45:  But something that we have been very interested in recently is that CMA degrades proteins, not only because they are damaged, but it can degrade fully functional proteins in order to terminate their function.
• 95:45 - 96:03:  Many labs now have identified that there is some kind of cell type specificity in which proteins are degraded through CMA for this regulatory function. So, for example, the well-studied validated substrates, but our labs are not there, so a bunch of enzymes are degraded
• 96:03 - 96:19:  in this pathway in order to terminate the flux through these metabolic pathways, glycolysis, lipogenesis, lipolysis. Transcriptional programs can also be regulated by the degradation of either the transcription factors or the inhibitors of the transcription factors.
• 96:19 - 96:29:  Activation, reprogramming. So just by modulating the protein in this selective manner, it contributes to regulate other functions inside the cells.
• 96:30 - 96:46:  So our incursion into the neurodegeneration field came actually from the realization that many of these prone-to-aggregate proteins that all of you are very familiar with in neurodegeneration contain a CMA targeting motif, and that doesn't mean anything because, as you know,
• 96:46 - 97:01:  CMAs can be degraded through many different pathways, but the wild type of this protein, similar to degradation through macroautophagy, as David's lab has shown beautifully, can also, a fraction can also undergo degradation through CMA.
• 97:01 - 97:16:  So one thing that we start to see developing in collaboration with terrific neuroscientists that help us along the way, is that the pathogenic variants, either mutations or post-translational modifications that you find in the patients of these proteins
• 97:16 - 97:30:  are still identified by the chaperone that brings them to the surface of the lysosomes when they undergo this CMA degradation, but unfortunately these proteins get stuck at the lysosomal membrane, and do not translocate.
• 97:30 - 97:42:  This is bad because that will decrease the amount of the protein degraded through this pathway, but also will compromise the degradation of all these array of proteins that I showed you in the previous slide.
• 97:43 - 97:57: So we find that the pathology is a combination of the proteotoxicity of not being able to eliminate the pathogenic protein, but also the loss of function of the inability to regulate these other proteins that occur in the cell.
• 97:57 - 97:59: And are degraded through CMA.
• 97:59 - 98:14: And of course, all this work that we have done through the years we have always done it in vitro, in neuronal cultures or even in isolated lysosomes. So, more recently we've been very interested to see if this blockage of CMA by pathogenic variants of these
• 98:14 - 98:29: proteins, also of course in vivo. And in order to do that, we have to develop new tools. So we have developed fluorescent reporters, that when these proteins arrive to the lysosomes via CMA, they will highlight and they will fluoresce, and then
• 98:29 - 98:35: Suzanne and Aurora developed a transgenic mouse model that contains this reporter.
• 98:35 - 98:50: And in collaboration with Jose Javier Bravo Cordero, we've been using intravital imaging and all kinds of sophisticated imaging to follow this process. So for example, this is yes, an image of the brain, and you can see here this panther, and this panther really corresponds
• 98:51 - 99:07: to the arrival of the substrate to lysosomes, because as you can see here, it co-localized with LAMP2, with cathepsin, so this was just a validation. But having these animals allows us, and this is for example a hippocampus region, so it allows us to really understand
• 99:07 - 99:23: the normal basal CMA activity in different regions of the brain, and also how does it change in response to pathology or to stress. So for example here in the hippocampus you can see cells with very high levels of CMA, a lot of panther, others with lower
• 99:23 - 99:38: levels, and we can even go to the single-cell resolution so we highlight the membrane, but then as you go inside you can see the presence of panther, not only CMA activity but where it is happening inside the cells.
• 99:38 - 100:02: So we are taking advantage of those models to really answer this question whether pathogenic proteins, like tau, can compromise CMA activity in vivo. So Aurora and Mathieu crossed this CMA reporter mouse model with a model of proteotoxicity related to Alzheimer's, so this is a triple transgenic that expresses a pathogenic ATP, pathogenic
• 100:00 - 100:06: tau, and presenilin 2. And if you look here, so this is the control animal that's
• 100:06 - 100:11: not the stress tau, the pathogenic one has the human form of tau, and when you look at
• 100:11 - 100:15: the panther in the neurons, and probably the quantification is easier to follow, you have
• 100:15 - 100:21: a dramatic decrease in CMA activity in the neurons, while less panther. But for example
• 100:21 - 100:25: if we look at astrocytes in this case, we don't see at this particular stage, we don't
• 100:26 - 100:31: see a decrease in CMA activity, suggesting that the inhibitory effect of these pathogenic
• 100:31 - 100:36: proteins was primarily happening at the level of the neurons. Of course our ultimate goal
• 100:36 - 100:42: is to show that that's also happening in patients, this decrease in CMA. We cannot use
• 100:42 - 100:49: the reporters yet in people, but what Mathieu has been doing is using mathematical modeling
• 100:49 - 100:55: to develop an algorithm that allows us to infer CMA activity based on the expression
• 100:55 - 100:59: of the network of components that we know that participate in this pathway, by giving
• 100:59 - 101:06: them a weight and a directionality. So, for example, using single-cell RNA-seq from brains
• 101:06 - 101:11: of control individuals, healthy individuals, and individuals in two different stages of
• 101:11 - 101:16: Alzheimer's disease, these are Braak stages, we've been analyzing, for example, how is
• 101:16 - 101:22: the CMA transcriptional program in excitatory and inhibitory neurons. And if we calculate
• 101:22 - 101:28: the CMA score, you can see how in the neurons there is a gradual decrease in this score
• 101:28 - 101:33: with the progression of the disease, but if we look at the astrocytes or other forms of
• 101:33 - 101:39: CLIA, this decrease is not there, so this resembles very much what I just showed you
• 101:39 - 101:44: in the mouse models of disease. Of course, this is still an index, so it's just inferred
• 101:44 - 101:49: and we have to keep validating, but something that got me excited is when Mathieu plotted
• 101:49 - 101:55: the CMA score against different types of pathology, I'm just showing here a bit like Bardem,
• 101:55 - 102:01: but we have done for other pathologies, there is an inverse correlation. So, the lower is
• 102:01 - 102:07: your CMA activity, the higher is the pathology and the faster it progresses. So, at least
• 102:07 - 102:13: it's just some examples that we think that these pathogenic proteins also compromise
• 102:13 - 102:20: CMA activity in vivo and in patients. So, what is this compromise? Why these proteins
• 102:20 - 102:26: are blocking CMA? So, just to follow with the examples of Tau, Benha and Mathieu in
• 102:26 - 102:32: the lab started in collaboration with Lugano just over this weekend and started trying
• 102:32 - 102:37: to understand or try to figure out what is different when you have a pathogenic form
• 102:37 - 102:43: of Tau. So, the normal form of Tau will bind to the LAMP2A, this receptor multimers and
• 102:43 - 102:48: the protein will translocate and we can follow those events. For example, this is lunatic
• 102:48 - 102:54: electrophoresis and you see here the monomers of LAMP2A and you see here the multimers.
• 102:54 - 102:59: You add regular Tau, you can see major changes, but when you add pathogenic variants, for
• 102:59 - 103:04: example, either mutations in Tau or acetylated forms that have been shown to be toxic in
• 103:04 - 103:13: the disease, you see an increase in the clustering of this LAMP2A. We also had the luxury to
• 103:13 - 103:19: get brains from patients and I'm just showing here two controls and two Alzheimer's, but
• 103:19 - 103:25: we did 9 and 10. We isolated those lysosomes and sure enough, when we look at how is the
• 103:25 - 103:30: status of this LAMP2A, we still see, I mean, we see a very similar clustering to the one
• 103:30 - 103:35: that we can recapitulate in vitro. So, this led us to try to understand why this clustering
• 103:35 - 103:42: happened. We found that these translocation complexes that are accumulating are not transporting
• 103:42 - 103:48: anymore and that we can detect that Tau is kind of stuck in the middle of these complexes.
• 103:48 - 103:54: We wanted to know why is Tau not crossing when normal Tau is one of the fastest substrates
• 103:54 - 103:58: that we have ever found. So, we decided to look at the chaperones because if you remember,
• 103:58 - 104:03: I mentioned you have the cytosolic HSC70, that is the one that brings the substrate,
• 104:03 - 104:09: but you also have HSC70 in the lumen of the lysosomes. And we thought that because this
• 104:09 - 104:15: one has been shown by the Dice lab that contributes to complete the translocation, maybe we should
• 104:15 - 104:19: have studied the interaction with these chaperones. And the main difference is that this one is
• 104:19 - 104:25: a neutral pH, right, in the cytosol. Meanwhile, this one is an acidic pH. So, we decided to
• 104:25 - 104:32: see how is the binding of HSC70 to Tau, wild-type normal Tau, and what was very interesting is
• 104:32 - 104:39: that as you can see here, the lower is the pH, the higher is the interaction of HSC70 with Tau.
• 104:39 - 104:46: So, in a way, this luminal HSC70 that will be at this lower pH will have an advantage versus
• 104:46 - 104:51: the cytosolic one, and we think that this higher affinity might contribute to complete the
• 104:51 - 104:57: translocation. So, then could that be related to the problems that we saw with the translocation
• 104:57 - 105:02: of pathogenic Tau? So, if we do the same experiments of binding, but now using the
• 105:02 - 105:08: acetylated, and we also did it with the mutant forms. So, when you check the binding to HSC70
• 105:08 - 105:13: and neutral pH, there is some decrease in binding, but it's really nothing to die for.
• 105:13 - 105:19: But look what happened. When you start lowering the pH, you completely lose this advantage. You
• 105:19 - 105:24:  lose this gradient, this increased affinity. So, we think that that's probably what is contributing
• 105:24 - 105:30: to this Tau just getting stuck in the middle of the translocation complex and not being able
• 105:30 - 105:36: to cross the membrane. So, basically, just to summarize what I showed you, we think that
• 105:36 - 105:42: normal forms of these proteins are degraded by CMA in the same way that they are degraded by
• 105:42 - 105:48: other pathways. When you block this pathway, you start building up of these proteins,
• 105:48 - 105:54: but also when you have the pathogenic variants, they have an inhibitory effect on CMA. Something
• 105:54 - 106:00: that David mentioned is that we found a while ago that CMA activity decreased with age.
• 106:00 - 106:06: So, we start questioning whether this decrease with age in CMA or the decrease imposed by the
• 106:06 - 106:12: pathogenic proteins on CMA will have some consequences for the normal functioning
• 106:12 - 106:18: in the brain of the neurons. So, to test that, what Mattia has done is to develop
• 106:18 - 106:25: CMA-incompetent mouse models in different neuronal types. And because LAMP2A is the limiting
• 106:25 - 106:31: component for this pathway, what we have done is to eliminate LAMP2A either under a cankinase
• 106:31 - 106:37: prepromoter or a TH. An interesting fact is that when you block this pathway, contrary, for example,
• 106:37 - 106:43: to the blockage of macroautophagy that you have very rapid neurodegeneration in a matter of weeks
• 106:43 - 106:49: and after the animals are born, in this case, the degeneration is much more gradual.
• 106:49 - 106:55: So, these animals are six months of age, wild type, and they knock out, and you can see that
• 106:55 - 107:01: they develop symptoms in this case of excessive clustering and other behavioral changes related
• 107:01 - 107:07: to neurodegeneration. Since the neurons are not dying, we can at least, at least at the beginning,
• 107:07 - 107:11: we can study their functions in collaboration with the cell search we have done, for example,
• 107:11 - 107:16: in the TH models recordings for dopamine uptake. And if you compare wild type with the ones that
• 107:16 - 107:22: do not have CMA, you clearly have a functional deficit in those neurons. And when we look at
• 107:22 - 107:28: proteostasis in those different models, we can see if you look at prone to aggregate proteins,
• 107:28 - 107:33: such as tau, wild type, and alpha-synuclein, we have very high levels of these proteins. Even other
• 107:33 - 107:38: degradation pathways seem to be perfectly fine, and we have characterized that extensively,
• 107:38 - 107:43: and you start seeing aggregation of proteins, for example, here of alpha-synuclein.
• 107:44 - 107:50: So, I don't want to give the idea that CMA is responsible for taking care of the aggregates,
• 107:50 - 107:55: because we know that the mechanism of translocation requires unfolding.
• 107:55 - 108:00: What we think is happening and why we see aggregates when we block CMA is because CMA
• 108:00 - 108:06: contributes to maintaining a low level, so the right levels, of these prone to aggregate proteins.
• 108:06 - 108:12: So, if those levels cannot be maintained normal, if you block CMA, there is going to be an increase
• 108:12 - 108:18: in the levels of the soluble protein, and everything is going to start precipitating.
• 108:18 - 108:24: So, in a way, what we think is that what the blockage of CMA or the inhibition of CMA that
• 108:24 - 108:30: happens with age is displacing this gradient of soluble proteins toward aggregation.
• 108:30 - 108:36: And to put it in a more scientific term, that the salt and the water, if we look, for example,
• 108:36 - 108:42: at the super saturation score of the proteins in the brain of control mice, and mice with
• 108:42 - 108:50: effective CMA, you see the shift to a lower solubility. So, we also wanted to know what
• 108:50 - 108:56: is aggregating. I mean, is there a particular part of the proteome that is more dependent on
• 108:56 - 109:02: having functional CMA? So, we did quantitative proteomics of aggregates comparing those
• 109:02 - 109:05: ones that, of course, when you don't have macroautophagy versus the ones that, of course,
• 109:05 - 109:11: when you don't have CMA. And in green are the proteins that aggregate preferentially when you
• 109:11 - 109:17: don't have CMA. And it's a bunch of enzymes related to metabolites and proteins related
• 109:17 - 109:22: to calcium regulation, trafficking, some of the synaptic proteins that we've been
• 109:22 - 109:28: hearing in previous talks. So, we think that what we see in this phenotype is a combination
• 109:28 - 109:34: of the proteotoxicity that results from the presence of these misfolded proteins or aggregated
• 109:34 - 109:41: proteins that are not timely degraded, but also a loss of function because many of these proteins
• 109:41 - 109:46: that have to be normally turned over, all of a sudden they end in these aggregates and trapping
• 109:46 - 109:52: other proteins, and that contributes to the loss of, for example, neuronal function that we see
• 109:52 - 110:00: in these models. So, then, basically, we have convinced ourselves that this decrease of CMA
• 110:00 - 110:07: with age or in pathology is important and has important consequences. So, the other thing is,
• 110:07 - 110:14: does this aggregate in the course of disease? I mean, so far, I have shown you we eliminated CMA
• 110:14 - 110:18: in a wild-type, otherwise wild-type model, and then we see already accumulation of the
• 110:18 - 110:24: wild-type forms of tau and the wild-type forms of synuclein. So, what happens when what you have
• 110:24 - 110:33: there is a mutant or a pathogenic variant? So, to see that, Nadia utilized the same triple transgenic,
• 110:33 - 110:38: so this is presenilin 2, so it's a very gradual and slow course of degeneration. So, you can
• 110:38 - 110:44: see here, for example, tau aggregates and how they accumulate in this mouse model with age.
• 110:44 - 110:50: And now, if on this pathogenic background, we impose the decrease of CMA, you can see how it
• 110:50 - 110:56: accelerates the disease, the pathology. So, in this case, I'm showing tau aggregates, but in blue
• 110:56 - 111:02: is always the triple transgenic. In orange is the triple transgenic, where we have, in addition,
• 111:02 - 111:09: block CMA, and you can see that different forms of tau pathogenic variants start accumulating,
• 111:09 - 111:14: and even Aβ. When we look at the brains of the triple transgenic at a stage that we don't see
• 111:14 - 111:22: much Aβ, we can see that our triple transgenic with deficient CMA already has this accumulation
• 111:22 - 111:28: of Aβ. So, we think this is interesting because something that was very disappointing for us
• 111:28 - 111:35: when we start comparing the proteome of the brains of Alzheimer's disease patients with the
• 111:35 - 111:41: brains of the triple transgenics that we all use, or even our CMA-incompetent models, the ones that
• 111:41 - 111:46: only have a blockage of CMA but don't have the pathology, when you look at this coincidence
• 111:46 - 111:51: mapping, so yellow is supposed to be, they look very similar. And as you can see, there is very
• 111:51 - 111:56: little yellow, so these models are not recapitulating what we see in the patients.
• 111:56 - 112:01: But now, when we do this combination, when we impose this decrease in proteostasis, in this
• 112:01 - 112:07: case, CMA, that is going to happen with age, you can see that the brains of these animals, at least
• 112:07 - 112:14: at the proteomic level, resemble much more those of Alzheimer's. So, we think that the combination
• 112:14 - 112:19: of the gradual decrease of CMA with age plus the inhibitory effect that many of these pathogenic
• 112:19 - 112:26: proteins have on CMA is what probably is also happening in the patients. And then, and I think
• 112:26 - 112:32: it's a great fit because we've been carrying beautiful work on unconventional secretion,
• 112:32 - 112:38: so something that we also wanted to know is if the blockage of CMA, in addition to this aggregation,
• 112:38 - 112:44: could also contribute to the release of some of these pathogenic proteins that, as you know,
• 112:44 - 112:49: is a typical feature of the disease. So, to test that, we were fortunate to collaborate
• 112:49 - 112:55: with Brad Hyman, who has this beautiful reporter that is expressing a pathogenic form of tau
• 112:55 - 113:00: and a fluorescent protein, but when they are expressed in the neurons, we inject them in
• 113:00 - 113:05: different parts of the brain, they get clipped, so you end with the two proteins. So, in the area
• 113:05 - 113:11: that you inject, for example, here, you are injecting here, and you see your GFP and your
• 113:11 - 113:17: tau protein. However, if there is propagation of the pathogenic tau, you will see the appearance of
• 113:17 - 113:22: only the red protein, in this case tau, in the distance, and this is what he has shown in this
• 113:22 - 113:29: beautiful model. So, Enrique and Mathieu wanted to know what happens if you do the same assay of
• 113:29 - 113:35: pathogenic tau and we look at the propagation when you don't have CMA that is going to happen
• 113:35 - 113:40: gradually with age. So, they did those studies. This is in the wild type, this is the knockout,
• 113:40 - 113:46: and as you can see here, when you look at the recipient cells, you can see that the ones when
• 113:46 - 113:51: you don't have CMA, there is much more release of tau, and we measure it by different methods, but
• 113:51 - 113:57: you also see that the recipient cells accumulate much more tau, and when you look at the temporal
• 113:57 - 114:02: cores from a distance, normally the distant cells accumulate tau in a wild type, and then
• 114:02 - 114:08: eventually it declines. In this case, if you don't have CMA, they keep accumulating this tau.
• 114:08 - 114:14: So, we think that basically this kind of support that the blockage of CMA is resulting in the
• 114:14 - 114:20: release of this tau among other things, but then how is this going out? We tried very hard to see
• 114:20 - 114:26: docking of these lysosomes directly into the plasma membrane, but we completely failed,
• 114:26 - 114:33: and then we started thinking about the relation of CMA with other autophagy pathways, and I mentioned
• 114:33 - 114:38: very quickly in the introduction that, as we have here, you can load, obviously, proteins in
• 114:38 - 114:44: these multivesicular bodies, proteins from the cytosol, and there is this LC3 beautiful dependent
• 114:44 - 114:50: model that we just here, but we also found that in a completely LC3-independent manner,
• 114:50 - 114:57: this HSC17 protein that recognized the targeting motif for CMA can also target the same proteins
• 114:57 - 115:03: using the same motif to these multivesicular endocytic bodies. So, we kind of thought that
• 115:03 - 115:09: because there is this crosstalk among pathways, when you block CMA, maybe part of these proteins
• 115:09 - 115:15: are now loaded in higher amounts into these multivesicular bodies, and actually, if we isolate
• 115:15 - 115:21: these endosomes, multivesicular bodies in normal mice, and we, same for acetylated tau, you see
• 115:21 - 115:28: very little amount of acetylated tau here. When you block CMA, you start seeing much more of this
• 115:28 - 115:34: oligomeric acetylated tau, and what was really exciting is that collaboration with Alice O'Goode,
• 115:34 - 115:40: and we also got brains from the patients, and if you compare the late endocytic compartment
• 115:40 - 115:48: of the healthy individuals versus Alzheimer's patients, you can see much more accumulation of this tau
• 115:48 - 115:54: in the multivesicular bodies of the patients. So, we thought that maybe this rerouting was
• 115:54 - 115:59: happening. We wanted to test it, and what we did was to measure the release of tau relative to
• 115:59 - 116:06: control conditions. So, this is the amount of tau in the media that was released from the neurons that
• 116:06 - 116:13: we consider control. If we block macroautophagy, we don't see higher release. If we only block
• 116:13 - 116:20: formation of these multivesicular bodies, we don't see major changes here. If we block CMA,
• 116:20 - 116:25: we see the release the same way that we saw in vivo, so now you are releasing tau, but if we
• 116:25 - 116:32: block CMA, and at the same time, we block this trapping of these proteins into multivesicular
• 116:32 - 116:38: bodies, you attenuate the release, suggesting, or at least this is our working hypothesis,
• 116:38 - 116:44: that this rerouting is probably the one that then results in the release in the extracellular media.
• 116:45 - 116:51: So, just to summarize a little, I showed you that CMA activity decreased with age,
• 116:51 - 116:56: and that's enough when it happens in neurons to result in progressive neurodegeneration
• 116:56 - 117:02: and accumulation of toxic proteins. When it happens in the context of disease,
• 117:02 - 117:08: it accelerates disease progression and accelerates propagation of the disease.
• 117:08 - 117:13: So, what can we do about that? So, in collaboration with Evry Scala Theotes,
• 117:13 - 117:19: we have developed some drugs that are able to activate CMA selectively. So, for example,
• 117:19 - 117:25: they cross the blood-brain barrier, and using the dendral model, I'm showing you here in a healthy animal,
• 117:25 - 117:29: remember PANTA with CMA, so when you get these compounds, you activate CMA.
• 117:30 - 117:35: So, we wanted to know if we can use those compounds, at least to slow down the course of
• 117:35 - 117:41: the disease. So, Mattia used several models of tau pathology, and in this case, I'm showing you this
• 117:41 - 117:49: PS19, that as many of you might know, it develops synaptic loss, gliosis, and tau pathology,
• 117:49 - 117:55: and cognitive impairment by six months. We wait till then, and then at seven months,
• 117:55 - 118:01: we added the CMA activators, and basically, if you compare, these are wild types,
• 118:01 - 118:07: these are the disease model, and the disease model with the drug show lower hyper locomotion,
• 118:07 - 118:13: that is characteristic of the disease. This decreasing spatial memory seems to be
• 118:13 - 118:18: prevented, and then also engagement, clustering, we did a bunch of tests.
• 118:18 - 118:23: But I wanted to look at the brain, of course. We wanted to know if we were changing
• 118:23 - 118:29: the proteotoxicity, so again, three different areas of the brain. If we are staining for normal
• 118:29 - 118:35: tau, you can see how the PS19 accumulate a considerable amount of this tau protein,
• 118:35 - 118:41: but when we put the animals with the drug, we can see a very dramatic reduction, and we think that
• 118:41 - 118:46: as a consequence of that, for example, if you look at the gliosis characteristic of this animal,
• 118:46 - 118:52: it is also a reduction, suggesting that the drugs might contribute to the improved phenotype.
• 118:53 - 118:58: Just to summarize a little of what I wanted to convey, I think this trafficking of proteins
• 118:58 - 119:04: directly across the lysosomal membrane for degradation, it's important to the maintenance
• 119:04 - 119:09: of this metastable proteome. Unfortunately, this process decreased with age, and I also
• 119:09 - 119:14: showed you how in some pathologies, there is a further decrease, and we think that that can
• 119:14 - 119:21: contribute to loss of neuronal function and proteotoxicity, and in the context of disease,
• 119:21 - 119:27: to accelerate the progression of the disease and the propagation of the pathology, and in the last
• 119:27 - 119:32: slide, I just showed you our early attempts. Of course, this is just animal models so far,
• 119:32 - 119:38: so there is still a long way for the clinic, but at least it's encouraging that we can eliminate
• 119:38 - 119:44: part of this proteotoxicity by upregulating this pathway, and with that, this is the most important
• 119:44 - 119:50: slide. I tried to mention the people that were involved in the work, and we can only do that
• 119:50 - 119:55: with terrific neuroscientists all around, because I'm just a cell biologist, so anything
• 119:55 - 119:59: related to neurons is too much for me, and we are very fortunate to have
• 120:00 - 120:04: this funding, I would say, supporting our work, so thank you so much for your attention.
• 120:05 - 120:09: Thank you very much, Ana Maria. That's a real tour de force, and I think it makes a
• 120:10 - 120:16: lovely summary of the importance of CMA, particularly in the context of neurodegeneration.
• 120:18 - 120:24: I have a few questions that have been sent, so I think one important question that was raised
• 120:25 - 120:32: by an anonymous attendee was, what is the fraction of proteins degraded by CMA versus
• 120:32 - 120:40: other degradation pathways? So that's the one million question, and I think we even probably
• 120:40 - 120:46: gave her entrenched to know that, because I think this, we have to think about protein degradation,
• 120:46 - 120:52: and I know David is more than in agreement with me, as a very dynamic process, so it doesn't mean
• 120:52 - 120:56: that there is, this protein is always going to be degraded by this pathway. If you think about
• 120:56 - 121:02: proteins, they have many different ways. You can ubiquitinate these proteins, you can trap them in
• 121:02 - 121:07: macroautophagy, you can eliminate them through CMA, and what we think is important to understand
• 121:07 - 121:14: is, in which conditions, or in which cell types, one of these mechanisms is predominant, and they
• 121:14 - 121:19: all co-exist. I mean, we shouldn't think about these pathways as one goes up, the other goes
• 121:19 - 121:25: down. There is co-existence of these pathways, and still, you may have pools of proteins that
• 121:25 - 121:29: are, depending on their location, depending on their conformation, depending on the translational
• 121:29 - 121:35: modifications, might go through one pathway or the other. Based on the motif, it looks very good,
• 121:35 - 121:41: because when you look at the presence of that motif, canonically, it's in about 40 percent of
• 121:41 - 121:47: your proteins. It's a good number, but also, you can make the motif, I didn't have time to say that,
• 121:47 - 121:52: through post-translational modifications, because what HSC70 recognized is the charge
• 121:52 - 121:58: on that pentapeptide. So, you can destroy it, or you can make it, so that kind of expands the pool
• 121:58 - 122:02: of proteins that, at a given time, can go through this pathway. But I would suggest that, more than
• 122:02 - 122:08: quantitatively, I think it's the timing and the conditions in which each of these proteins is
• 122:08 - 122:14: degraded by each of these pathways is probably more important, at least for the physiology of
• 122:14 - 122:21: these processes. You sort of touch on the idea that there's probably compensation and
• 122:21 - 122:25: sort of some, how can I say, crosstalk. And I was wondering,
• 122:28 - 122:37: is anything known about the signals that would result if you block CMA, for instance,
• 122:37 - 122:42: that might enhance flux through some other protein degradation pathways?
• 122:43 - 122:49: Yes, that's a great question, and I think we are very interested in that, because I think
• 122:49 - 122:55: we could leverage in order to do interventions, right? If we understand when you block a pathway
• 122:55 - 123:00: and the other gets upregulated, what is the mechanism? We might be able to directly activate
• 123:00 - 123:05: that mechanism without having to block the other pathway. What we have found so far in the
• 123:05 - 123:11: crosstalk between CMA and lysosomal microautophagy and macroautophagy is that the
• 123:11 - 123:17: compensation is not universal, and it's really depending on the cell type. So, for example,
• 123:17 - 123:22: when we eliminate CMA in the liver, we don't see aggregates. We don't see accumulation of
• 123:22 - 123:28: like proteotoxicity per se. We see loss of function because we lose the regulatory function,
• 123:28 - 123:33: but we don't see accumulation of proteins. And what we found is that the proteasome
• 123:33 - 123:38: and macroautophagy get nicely upregulated, so they can take care of quality control.
• 123:38 - 123:41: Unfortunately, you still lose the regulatory effect, because as you know,
• 123:42 - 123:47: the time of activation, the circadian time of activation of macroautophagy and the proteasome
• 123:47 - 123:52: is very different from the CMA. So you basically don't have the ability to eliminate the function
• 123:52 - 123:58: of a protein by CMA. To do that by macroautophagy, you will do it, but at a different time.
• 123:58 - 124:03: But we don't, at least for quality control purposes, there is this compensation.
• 124:03 - 124:07: However, in neurons, and now we have many other cell types, like in retina,
• 124:07 - 124:13: in metabolic stem cells, T cells, so we don't see this compensation. So if you block CMA,
• 124:14 - 124:19: macroautophagy, it doesn't, it's not inhibited, but it doesn't do this extra mile. It doesn't
• 124:19 - 124:25: do what the liver is doing, and we really don't understand why, because in the liver, we link
• 124:25 - 124:30: this macroautophagy crosstalk with CMA. One of the mechanisms is probably through TFEB,
• 124:30 - 124:35: because TFEB can be degraded through CMA. So if you don't degrade, you block CMA,
• 124:35 - 124:39: you have higher TFEB, and hopefully some of it goes to the nucleus. So that was kind of
• 124:39 - 124:45: one of the mechanisms. That should also happen in neurons, but we don't see that. Does it mean that
• 124:45 - 124:50: neurons are less dependent on TFEB, and that's why they cannot do this compensation, or that
• 124:50 - 124:56: compensation does not only depend on the TFEB? So those are very important questions that we are
• 124:56 - 125:00
• really trying to resolve, because it will be very nice to see what is the signaling that allows for
• 125:00 - 125:06: this compensation. Excellent. I think the last question that I'd like to ask, I've just got in
• 125:06 - 125:14: here, so the attendee says, first of all, I want to say the research and presentation sound
• 125:14 - 125:25: incredibly interesting, and the person wanted to understand why CMA decreases with age, and wondered
• 125:25 - 125:32: whether this is predominantly due to decreased expression of LAMP2A as one gets older.
• 125:32 - 125:37: Another great question, and sometimes for these questions, I really shouldn't generalize. I'm going
• 125:37 - 125:44: to tell you what we know so far, but we learn as we go. I've been a very liver-centric person
• 125:44 - 125:48: all my life, and all of a sudden, we have all these tissues, and I'm trying to learn.
• 125:48 - 125:54: So in many tissues, like for example in the liver, in the skeletal muscle, in kidney, we have seen
• 125:54 - 126:01: that the decrease of CMA is not transcriptional, and for a long time, I mean, we saw that there
• 126:01 - 126:07: was this reduction in levels of LAMP2A, but it was not coming out in any microarrays or RNA-seq,
• 126:07 - 126:12: and so then we started thinking that was a trafficking problem. Maybe, I mean, LAMP2A is a
• 126:12 - 126:17: very heavily glycosylated protein, so we thought, well, if with age there are some alterations in
• 126:17 - 126:22: Golgi trafficking, maybe you have less LAMP2A arriving, or it's less glycosylated and becomes
• 126:22 - 126:28: unstable, but it turns out that this is still trafficking fine, but what we found is that once
• 126:28 - 126:33: it reached the lysosomal membrane, the changes in the lipid composition of the membrane that
• 126:33 - 126:39: happened with age made this protein very unstable, because we found that in order for LAMP2A to be
• 126:39 - 126:45: turnover, like any other protein in lysosomes, it goes to specific microdomains, and they're
• 126:45 - 126:51: kind of unweighed, gets cleaved, and that's how you renew your LAMP2A, but in the case of aging,
• 126:51 - 126:56: we have seen that there are these big ceramide patches, and with the lipidomics of the isolated
• 126:56 - 127:01: lysosomes, and you have your LAMP2A trapped and degraded there, so we think in that case,
• 127:01 - 127:06: definitely post-translational, but then, for example, there are studies in T-cells
• 127:06 - 127:10: that start to show that there is a problem with the transcriptional regulation, for example,
• 127:10 - 127:15: in response to reactive oxygen species, so we start to think that maybe in the basal
• 127:15 - 127:22: decrease in CMA might be post-translational, but the ability to induce CMA decrease with age,
• 127:22 - 127:27: maybe because of this transcriptional or splicing or other mechanisms, and we are currently really
• 127:27 - 127:33: investigating those ones, yeah. Brilliant. Ana Maria, that was really lovely, and thanks for
• 127:33 - 127:40: taking the questions so nicely. We now have a panel discussion, which I'm a little bit scared about,
• 127:42 - 127:48: so please could Thierry and Ira also switch on their cameras and unmute.
• 127:52 - 127:59: So, we've been given five questions, and when I look at them, they're all very difficult questions,
• 128:00 - 128:08: and I wondered whether each of you want to choose a question, maybe starting with Thierry,
• 128:08 - 128:13: he's at the greatest time to recover, and so, and, you know, just give your thoughts on one or two of
• 128:13 - 128:18: the questions that you think are appealing. Unfortunately, it puts Ana Maria in the hardest spot,
• 128:18 - 128:23:  well, I'll probably have to go last, so it's me in the hardest spot, but, you know, why don't
• 128:23 - 128:27:  you just kick off with something that you think is of interest in the questions that you want to
• 128:27 - 128:35:  offer some thoughts about. You can read the question out that you think you'd like to answer.
• 128:36 - 128:42:  I'm not sure I see the questions. Okay, so the five questions are these. What are the most
• 128:42 - 128:47:  significant unanswered questions in the field? The second is how do you see research with
• 128:47 - 128:53:  intracellular trafficking future developments, and that's sort of related to how do you fix
• 128:54 - 128:58:  intracellular trafficking research to progress in the next five years, and how confident are you
• 128:58 - 129:02:  that this field of research will drive new discoveries? And I think the second part of
• 129:02 - 129:10:  the question we won't bother with, we all think so, that's why we're doing it. The third question
• 129:10 - 129:15:  was what are the most significant challenges we face in this field and how do you approach
• 129:15 - 129:22:  addressing them? And the last one is what advice do you give to existing researchers in the space
• 129:22 - 129:29:  or newcomers choosing to focus on intracellular trafficking, specifically in the context of
• 129:29 - 129:38:  neuroscience? Okay, so I would like to answer the first question you mentioned, David,
• 129:39 - 129:46:  by connecting some of the things that Ana Maria showed, which of course I was extremely interested
• 129:47 - 129:56:  in, which relates to secretion of tau, of course, and some of the results that we also
• 129:57 - 130:05:  obtained and which I showed. So I think one of the challenges now is to understand what the signals
• 130:05 - 130:18:  are for this decision-making kind of process where proteins that were not described as
• 130:18 - 130:24:  secreted proteins in the first place get secreted in the end. And there is a common
• 130:26 - 130:33:  thing that came out in what I showed and what Ana Maria showed is that by affecting certain
• 130:33 - 130:40:  processes of autophagy, now molecules that should be degraded end up outside of the cell.
• 130:41 - 130:49:  So we need to, I think, both in the case of CMA blockade and degradative autophagy blockade,
• 130:49 - 130:55:  in both cases, you know, I think it's amazing that both of us are coming out with a similar
• 130:55 - 131:02:  concept that, okay, it's not degraded, then it's thrown outside. And when I say it's thrown outside,
• 131:02 - 131:08:  it sounds like there's no specificity when I think there is. And the question is, what is
• 131:08 - 131:16:  this specificity? Because what we are seeing being secreted unconventionally is not a repertoire
• 131:18 - 131:27:  of what the cells express. It is a subset. And, you know, I've been spending days and weeks and
• 131:27 - 131:33:  months looking and trying to understand, is there a signal there? And I don't see it in our case.
• 131:33 - 131:40:  So in the case of Ana Maria, of course, with CMA, there is the KFERC signal. So that's really
• 131:40 - 131:48:  interesting because there is there a signal to go in the CMA pathway. But then with the other
• 131:48 - 131:58:  mechanism, what is or what are the signals? I think, you know, that's a major, major goal.
• 131:58 - 132:06:  And then, you know, connecting with IRA, I think that really understanding the trafficking,
• 132:07 - 132:13:  the targeting of the V-ATPases and in the different organelle where it is, and then all of its
• 132:13 - 132:21:  regulation is extremely important because this, the acidification and the proton pumping
• 132:21 - 132:27:  is really so central to everything that we heard this afternoon. I mean, the V-ATPases, you know,
• 132:27 - 132:33:  I think since long time is probably one of the most important protein complexes, not just one
• 132:33 - 132:40:  protein, but it's a protein complex. And so how it's regulated is extremely important, both for
• 132:40 - 132:47:  autophagy and for synaptic vesicle recycling. So I see those as major challenges.
• 132:48 - 132:53:  If I may add, how do they change with aging, right? And how do they change in certain diseases?
• 132:53 - 133:00:  And how can they even drive certain diseases? Yeah, absolutely. Absolutely.
• 133:01 - 133:05:  IRA, do you want to add thoughts on one of the other questions?
• 133:06 - 133:10:  Sure. Maybe I can expand a little bit on this one because the next question is
• 133:10 - 133:16:  also about future developments, right? So the way how I also see it is this,
• 133:17 - 133:21:  I mean, I'm just going to possibly a little bit rephrase what Thierry was saying,
• 133:21 - 133:25:  this coordination of trafficking, intracellular trafficking with the other processes inside the
• 133:25 - 133:35:  cell. I think this is possibly something that the future work would be, at least this is something
• 133:35 - 133:42:  that I see as an interesting aspect, as well as the regulation of these individual transport
• 133:42 - 133:48:  steps. So we know a lot about transport steps, but we don't really know molecular, I mean,
• 133:48 - 133:56:  I often like to think in molecular terms, how are they actually regulated? And so along this line,
• 133:57 - 134:02:  I think lipids actually play quite a bit of role. And we always talk proteins, proteins,
• 134:02 - 134:10:  proteins, simply because methodology for other factors in the cells is not as developed as the
• 134:10 - 134:16:  one for proteins. So I'm very fond of sphingolipids and phosphatidylinositols, as you see in
• 134:16 - 134:21:  my talk. So I'm always trying to keep an eye on that because I think we often miss an elephant
• 134:21 - 134:29:  in the room just by focusing on proteins. Yeah, that's a little bit. And also, I see that this
• 134:29 - 134:34:  would be the questions that the field will be addressing in the coming years, probably through
• 134:34 - 134:43:  the light of diseases, as this is where the interest is most, I mean, it's the biggest or
• 134:43 - 134:51:  the need is more imminent. And yeah, I hope this helps. Anna Maria, do you have, do you want to
• 134:51 - 134:55:  have, do you want me to go first? So you've got time to think, you've just been talking and working,
• 134:56 - 135:01:  you're welcome to go first. Okay, please go ahead, David. So, I mean, I think that, you know,
• 135:01 - 135:09:  ultimately, one of the aspects we challenged with is trying to show physiological and disease
• 135:09 - 135:16:  relevance. And I think the one group of disease or major group of diseases that really sort of still
• 135:17 - 135:25:  is underexplored in general, actually, in terms of the sort of the hardcore biology,
• 135:25 - 135:30:  particularly in the membrane trafficking, are psychiatric diseases. And I think that
• 135:30 - 135:37:  we really need to take on board that they're very common, and they're truly devastating. I mean,
• 135:37 - 135:42:  the serious psychiatric diseases hit people, you know, in their late teens, often for the first
• 135:42 - 135:48:  time, and the disability can be profound and the impact, we don't have to rehearse. So it's,
• 135:48 - 135:53:  you know, and so I think that it's very likely that understanding membrane trafficking
• 135:55 - 136:01:  will be really illuminating for these diseases. And I think one of the challenges that we all face
• 136:01 - 136:09:  when we do our sort of rather reductionistic cell biology experiments, and I guess animal
• 136:09 - 136:17:  experiments as well, is that we often not working in what I'd call the physiological, even the
• 136:17 - 136:24:  disease range of effects. And that's just because often one can't see things in that range. But I
• 136:24 - 136:31:  think the challenge is sort of marrying our basic biology with a very big dynamic range,
• 136:31 - 136:38:  so sort of experimental systems, trying to get that into a context where it becomes relevant for
• 136:38 - 136:43:  diseases and physiological processes that we're interested in. And so I think that
• 136:43 - 136:49:  that sort of remains a big challenge. And I think that, you know, clearly neurodegenerative diseases
• 136:49 - 136:56:  are important, and there's tons that is left to understand, but psychiatric diseases are very,
• 136:56 - 137:03:  are maybe even more important numerically and have a bigger devastation over many more years.
• 137:03 - 137:09:  And I think we need to sort of really push ourselves to see, as a field, to see if we can
• 137:09 - 137:13:  sort of relate the more basic science to an understanding of those conditions and hopefully
• 137:13 - 137:22:  some type of therapeutic effects. May I? Go for it, Thierry. Okay. Yeah, no,
• 137:23 - 137:29:  referring to what you just said, and also the comment of IRA on lipids, several of the
• 137:30 - 137:39:  psychotropes and other drugs acting in psychiatric diseases are amphipathic molecules, and they like
• 137:40 - 137:45:  to insert into membranes, and they have a slow mode of action, you know, things like
• 137:45 - 137:50:  clozapine and chlorpromazine and these classes of molecules. So, of course, they act on receptors,
• 137:51 - 137:58:  but I think there are data already out there that they probably affect to a great extent
• 137:59 - 138:07:  lipids. And, you know, just to name one, which is a major drug, kind of, which is alcohol,
• 138:07 - 138:13:  and it acts on lipids, and the brain is particularly sensitive because, you know, brain is a lot of,
• 138:13 - 138:19:  is a lot of fat and a lot of lipids. So, I think there's an important challenge there.
• 138:19 - 138:26:  And, of course, if lipids are affected in a way or another, then membrane trafficking
• 138:26 - 138:31:  is affected. And so that's another, that's, I completely agree with IRA so that this is a
• 138:31 - 138:37:  challenge, and a difficult one, because we cannot manipulate lipids like we manipulate genes
• 138:37 - 138:44:  with CRISPR and secondarily proteins in this manner. IRA, what's your thoughts?
• 138:49 - 138:55:  No, I would like to second David's opinion about psychiatric diseases. I think this is really,
• 138:56 - 139:04:  it's really important to try to link our research to them. I believe the reason why
• 139:04 - 139:09:  neurodegenerative diseases are coming up first is simply because they're a little bit simpler,
• 139:09 - 139:17:  you know, and the links are more obvious, at least at the moment. And I do second your opinion
• 139:17 - 139:24:  that there will be some important links between trafficking and neuropsychiatric diseases.
• 139:26 - 139:32:  Ana Maria? I was just going to comment that it's interesting that when I read the question,
• 139:32 - 139:36:  the first thing that I thought is like limitations, lipids, because we don't have the tools,
• 139:36 - 139:41:  and that's something that we've been hurting. But just to put it also in the whole picture,
• 139:41 - 139:47:  I mean, completely agree about the interactions. I was also delighted to see as Thierry was presenting,
• 139:47 - 139:52:  I was like, oh, we are going to the same direction here. So, that was very exciting.
• 139:52 - 139:56:  But one of the things that I think, because we are all cell biologists and
• 139:56 - 139:59:  necessarily don't think about that, is cellular energetics.
• 140:00 - 140:06:  Anything that we are all looking at is so ATP dependent. Any post-translational modification,
• 140:06 - 140:11:  it's going to depend on your acetyl-CoA, right? So, those are things that, as David mentioned,
• 140:11 - 140:16:  we do it very reductionist, we make it very simple. But once you put it in an organism,
• 140:16 - 140:24:  we are all exposed to all these changes, to these metabolic constraints. And then we simplify our
• 140:24 - 140:28:  models because we don't want to incorporate, but I think that's going to have a very big impact.
• 140:28 - 140:34:  And I hope that as metabolism, proteostasis, cell biologists start getting more and more
• 140:34 - 140:39:  interactive, we will have new biology, or at least new ideas of the regulation of some of
• 140:39 - 140:45:  these processes and what are the limitations. So, I thought that is something. And regarding the
• 140:45 - 140:50:  diseases, I agree with Ira and David that, you know, we go for neurodegeneration because it's
• 140:50 - 140:56:  like, okay, I have this misfolded protein, so at least I can follow to see what it's doing. But even for
• 140:56 - 141:02:  those ones, we always complain we don't have good models. And again, we don't have good models
• 141:02 - 141:08:  because our animals are very young and they degenerate. These are diseases of aging. So,
• 141:08 - 141:12:  we are not incorporating, for example, in this case, the aging on those models. And I think,
• 141:12 - 141:18:  or the comorbidities. I mean, you have Alzheimer's and diabetes are continuously related.
• 141:18 - 141:23:  So, I understand why we have to be simplistic in our models. But I think for resolving the
• 141:23 - 141:29:  questions of translation of our findings, I think we have to move to a more complex model.
• 141:29 - 141:36:  And at the same time, we cannot use models that would take 10 or 15 years to show the
• 141:36 - 141:43:  symptoms, right? Because even if that existed, that would be very difficult.
• 141:43 - 141:48:  I think a way that, I mean, we've been talking a lot in the proteostasis meetings that we have
• 141:48 - 141:53:  for aging, at least. We were talking about that because, I mean, we are not getting any younger.
• 141:53 - 142:00:  It's becoming a major problem. But we still don't know even what is the baseline level in population
• 142:00 - 142:05:  of these trafficking events, of these, you know, autophagy and all those things.
• 142:05 - 142:10:  So, there is now some initiatives trying to look at these cohorts of patients or people,
• 142:10 - 142:16:  in general, healthy people that have been followed for years and try to start not only
• 142:16 - 142:21:  taking the typical thing, oh, yeah, we do the GWAS and we will get everything, but to start
• 142:21 - 142:26:  doing cell biology on those cohorts. And I think that's going to probably change a little the way
• 142:26 - 142:32:  that we see things from the, even we may discover new biology by having this population diversity.
• 142:34 - 142:39:  I may bring one more point, is basically the robustness of the system, you know. Sometimes
• 142:39 - 142:44:  the system can compensate a lot, but it comes to some critical point where it can't compensate
• 142:44 - 142:52:  anymore. And so, what are those components on which the system starts to fail? And why does
• 142:52 - 142:58:  it fail on those components? I think would also be something that would probably become more clear in
• 142:58 - 143:03:  the years to come. And then on a cellular level, why certain neurons are more sensitive than the
• 143:03 - 143:07:  others, like dopaminergic neurons and so on, when it comes to neurodegeneration or Parkinson's
• 143:07 - 143:16:  disease and so on. And something that Maria just mentioned. So, we tend to simplify, you know,
• 143:16 - 143:21:  we tend to simplify because our systems are complex, our diseases are complex, the models
• 143:21 - 143:27:  are complex, the technology is complex. So, in order for people to understand, we tend to sometimes
• 143:27 - 143:35:  oversimplify. And like every metal that has two sides, right? The oversimplification is sometimes
• 143:35 - 143:41:  not necessarily good because, I mean, it is indeed helping to understand the one aspect, but
• 143:43 - 143:51:  the other aspects tend to get ignored in a way because something that, yeah, I do think that
• 143:52 - 143:59:  certain discoveries and results could be integrated. So, the question of integration
• 143:59 - 144:04:  as well comes along. So, I was wondering on that, just to know the thoughts of the panel,
• 144:05 - 144:10:  of how much computational modeling can help us with that? Because, I mean, we can still do the
• 144:10 - 144:15:  simplistic models, but experimentally. But then when you start putting these inputs, I mean, there
• 144:15 - 144:21:  are some nice examples in the autophagy field, for example, of modeling. And then you can, you know,
• 144:21 - 144:25:  you add this thing and then you see how it responds. I mean, it still requires a lot of
• 144:25 - 144:30:  cell biology done, but I was wondering if that can help us first with integration of
• 144:30 - 144:37:  different inputs and also to start thinking a bit more globally. There are models out there on,
• 144:37 - 144:46:  you know, modeling the Golgi apparatus, modeling membrane trafficking. Yeah, there are things
• 144:46 - 144:53:  coming up and we need to challenge, we need to use them and to challenge them with experiments.
• 144:53 - 145:03:  So, that's, you know, that's to be done and probably not done enough by, you know, any of us,
• 145:03 - 145:09:  I think, I'm afraid, but we should definitely, yeah. I know of some models that are already
• 145:09 - 145:20:  interesting in the field. We also keep an eye on it, but also to see how we can integrate this data
• 145:20 - 145:30:  from genomics that I also, like, touched upon in my talk. It's handling large data sets, you know,
• 145:30 - 145:36:  and there is a whole new world in there that we are discovering at the moment.
• 145:40 - 145:43:  Are there any other points you want to cover? I think we've actually
• 145:44 - 145:49:  touched on all the questions in a way. Well, there was the advice to the younger researchers.
• 145:49 - 145:56:  We didn't touch upon that. Well, there's advice and there's advice. No, I mean, there's generic
• 145:56 - 146:01:  advice about how to build a career and that might take us another few hours. I think that's…
• 146:03 - 146:10:  David, can I just say one thing? Train in biochemistry, if there is one advice.
• 146:10 - 146:16:  I'm glad you say that. So, there's nothing wrong with learning how to do good biochemistry,
• 146:16 - 146:20:  because that's the call for healing mechanisms. Absolutely.
• 146:20 - 146:28:  I think the other thing that, you know, I think Anna Maria's point about the modeling and the
• 146:28 - 146:34:  sort of types of comments that Iris made about looking at other types of data, I think is
• 146:34 - 146:41:  important. And Anna Maria's point about trying to sort of link this to doing cell biology on humans,
• 146:41 - 146:47:  if we can find ways of doing that, is going to be a big opportunity in the future.
• 146:47 - 146:54:  I also think that, you know, the progress that has been made, I'm going to say in AI,
• 146:54 - 146:59:  but in the sort of general field, if you just look at what the AlphaFold people have done,
• 146:59 - 147:05:  it's stuff that, you know, 10 years ago, at least I thought was unachievable. I think that,
• 147:06 - 147:14:  you know, the potential for sort of marrying different fields with, I was going to say,
• 147:14 - 147:24:  with biochemistry is immense. And I think that, you know, there's a huge amount to be done and
• 147:24 - 147:30:  very important work still to be done. And, you know, getting back to Iris' point, trying to
• 147:31 - 147:40:  marry simplicity and complexity is really going to be a big challenge moving forward and very
• 147:40 - 147:48:  exciting. So I think that, you know, this field will last for much longer than five years. It
• 147:48 - 147:56:  probably, you know, ultimately it can last forever. But I think that moving forward towards
• 147:56 - 148:04:  these challenges and breaking these new frontiers, and also sort of, you know, the new methods and
• 148:04 - 148:09:  techniques that really create opportunities, again, that most of us couldn't have dreamed of
• 148:09 - 148:16:  when we were doing our PhDs. So I think it's a very exciting time. And actually, you know,
• 148:17 - 148:22:  the types of things that companies are doing by making various reagents and tools available
• 148:23 - 148:28:  really has also transformed things because I can remember when I was a student, I made everything
• 148:28 - 148:35:  myself. You know, there weren't kits for the things I needed, unfortunately, most of the
• 148:35 - 148:44:  things I needed. So I think that that also helps science a lot. So are there any final remarks
• 148:44 - 148:51:  before we hand back to the Abcam hosts? Basically, I think that what we said, learn
• 148:51 - 148:58:  biochemistry, learn computational biology, and don't expect that things that look weird is because
• 148:58 - 149:02:  they are not real, because I think the word unconventional for all the cell biology that we
• 149:02 - 149:07:  are discovering is coming more and more, like unconventional secretion, unconventional autophagy.
• 149:07 - 149:12:  So I think just be ready and enjoy what is different because we still have a lot of biology
• 149:12 - 149:19:  to discover in this pathway. That's absolutely right. I would maybe like to add that, in my
• 149:19 - 149:26:  opinion, the fun only starts now, right? Because now we have tools, I mean, not all of them,
• 149:26 - 149:30:  right? We discussed that there is some limitation, particularly in the field of lipids and the large
• 149:30 - 149:39:  data and so on, and obviously, but we have more tools than ever before. And we have some basic
• 149:39 - 149:45:  understanding on how the systems work, right? So now I think the fun starts where we can use
• 149:45 - 149:51:  this knowledge to basically address the problems like certain diseases. And in that aspect, I would
• 149:51 - 149:57:  also like to acknowledge Abcam for organizing this meeting. I think the link between industry and
• 149:57 - 150:06:  academia should be strengthened, and in a way that like, in more collaborative ways, we are always
• 150:06 - 150:13:  open to that. Because there is no way that one group or one person will solve this all, right?
• 150:13 - 150:22:  So this is going to be, like intracellular trafficking is complex, and it's going to require
• 150:22 - 150:34:  group work to sort it out. Absolutely agree. Good. So it just remains for me to thank,
• 150:35 - 150:41:  again, Abcam for organizing this, and really thank Thierry, Ira, and Anna-Maria for their excellent
• 150:41 - 150:50:  talks and nice answers to interesting questions. So thank you very much, and have a good day.
• 150:50 - 150:52:  Thank you, David, and thank you all.
• 150:53 - 150:54:  Thank you from my side as well.
• 150:55 - 151:02:  Great. Well, thank you, David, for closing that down and passing it on to me. So I think all
• 151:02 - 151:07:  that's left now to do is to really, on behalf of Abcam, to thank all our attendees for joining
• 151:07 - 151:13:  today's event. We really appreciate your feedback. We'll be conducting a short poll, which should
• 151:13 - 151:19:  appear on your screen. We're also going to put information in the chat box on how you can request
• 151:19 - 151:25:  PACE credits for those of you that are interested in that for today's event. And finally, I'd like
• 151:25 - 151:32:  to thank our speakers, Thierry Ghali, Ira Milosevic, and Anna-Maria Kuvera for giving
• 151:32 - 151:37:  their wonderful talks. I mean, I found them all really fascinating. I really enjoyed learning
• 151:37 - 151:42:  about these unconventional biological crises that you talked about, and I really look forward to
• 151:42 - 151:48:  seeing how your respective stories develop. And, you know, there was a lot of, you know,
• 151:48 - 151:52:  collaboration with industry, and we at Abcam are always happy to collaborate and support
• 151:52 - 151:58:  our scientists out there in the field. And so, and lastly, of course, a special thanks to
• 151:58 - 152:05:  David Rubenstein for moderating today's session. So last thing to say is that we look forward to
• 152:05 - 152:09:  welcome you all at our next session on neurodegenerative diseases, which takes
• 152:09 - 152:15:  place tomorrow, October the 27th. Please know that a separate registration for this event
• 152:15 - 152:20:  is required, and the link has been placed in the chat for you to register. Thank you all,
• 152:20 - 152:28:  and enjoy the rest of your day.