Scientist Stories: Seema Lakdawala, Assembly of Viruses

Welcome to the influenza webinar series hosted by the Human Vaccines Project. The Human Vaccines Project is a nonprofit public-private partnership with the goal of decoding the human immune system to accelerate the development of vaccines and immunotherapies. In order to make real progress towards the development of a universal influenza vaccine, collaboration among researchers around the world is essential. We hope this monthly webinar will foster the scientific foundation needed to advance influenza research. This series of webinars will be broadcast on the third Tuesday of each month and posted online. For a copy of today's webinar, please visit the Human Vaccines Project website at www.humanvaccinesproject.org/talks. Following the presentation, we will have time for questions from the audience. Please type your questions into the questions box in the GoTo interface. I will read these questions following the presentation. Before we begin, I'd like to call your attention to an upcoming symposium on innovative vaccines for resistant infectious diseases and emergent threats being hosted by the New York Academy of Sciences on May 22nd in New York. The symposium will include talks from some of the world's leaders in vaccine development, including Wayne Coffey, CEO of the Human Vaccines Project. Please note the early bird deadline for registration is April 19th. And now it is my pleasure to introduce Dr. Seema Lakdawala. Dr. Lakdawala is a microbiologist who synthesizes biochemistry, microscopy, and animal models in her study of influenza virus assembly and pathogenesis. She is currently an assistant professor at the University of Pittsburgh where she runs her own lab. Dr. Lakdawala and her team studies the assembly of influenza genomic RNA and the viral and environmental properties which promote airborne influenza transmission with the goal of developing comprehensive surveillance system to assess the pandemic potential of influenza strains. Today, Dr. Lakhtiwala will tell us about seeing is believing: from intracellular influenza viral RNA assembly to persistence of influenza viruses in aerosols. Go ahead, Seema.

Thank you so much. Thank you for the introduction and for the invitation to tell you all a little bit about my work at the University of Pittsburgh. Hopefully all of the movies play. I'm sure I don't need to tell this audience about influenza viruses, but I just want to give everybody sort of a basic introduction. This is a cross-section of an influenza virus, right? It's an envelope virus. There are two surface glycoproteins. I'm sure many of the webinars in the past have covered this. The first glycoprotein is the hemagglutinin, or HA, and there are currently 18 subtypes of the HA protein, and it binds to cell surface receptors.. The other glycoprotein is the NA. It destroys sialic acids to minimize aggregation, and there are 11 known subtypes. I'm sure everyone knows, but when we talk about influenza viruses like H1N1, we're talking about a hemagglutinin of subtype 1 with a neuraminidase of subtype 1. So you can appreciate the large variation that exists in nature between 18 subtypes of HA and 11 subtypes of NA. The genome, which is what my lab is really concerned with and has studied for quite some time now, is composed of negative sense RNA segments. So there are 8 negative sense RNA segments that make up the influenza A virus genomic material, and each segment is sort of shown schematically here where the RNA coats these gray spheres, which are the nuclear protein, and the 3' and 5' ends are bound by the polymerase, a virally encoded polymerase complex. What's necessary for a fully infectious virus is that you need one copy of all 8 segments because they encode the 11+ proteins that make up the viral machinery to overtake a cell and produce more influenza viruses once it's inside of the cell. And so this incorporation of all 8 segments into one single virion is sort of an interesting concept I'm going to talk about today of how do you get one copy of all 8 and and how is that sort of mechanism ensured, and what kind of things are we studying in my lab to tease this apart? So there have been 4 pandemics in our last century. My lab is really interested in sort of the emergence of pandemics and the part of flu assembly that is important for these properties. So going back 100 years, 1918, there was obviously the H1N1 Spanish flu that circulated in the human population until 1957, where there was an H2N2 outbreak until 1968, and there was an emergence of an H3N2 strain in 1968, so referred to as Hong Kong flu. And then in 1977, there was reemergence of the 1957 H1N1 strain that circulated in the human population until 2009, where there was the swine flu 2009 pandemic. And so, you know, my lab is interested in how pandemic viruses emerge. And so many of these pandemics emerged through a process known as reassortment. And so what is reassortment? Well, there's two viruses, and because the genomes of influenza viruses are segmented and I said that you need all 8 segments to make a fully infectious virus. What happens is when 2 different strains of flu infect the same cell, the genomic segments can swap and you can get a variety of different types of viruses that emerge from a single co-infected cell that is in where the genomic materials have sort of intermingled. This process of reassortment has produced most of the known influenza pandemics, the 1957, 1968, and 2009 pandemic. It also is responsible for a lot of the genetic diversity that we see in nature for influenza strains. And it's— there is a bottleneck. So experimentally, if you do a co-infection experiment in vivo or in vitro, so either in animals or in cells, you will not get out all 256 possible combinations if you use two influenza strains. And what drives this bottleneck is really an interesting phenomenon that we're interested that my lab is looking at in terms of looking at constraints at the genetic segment level, looking at how different segments assemble into a virus and whether that is leading to this reassortment bottleneck. So as I mentioned, my lab is interested in the sort of larger question of how pandemic influenza viruses emerge. And the way I think about this and the way we've been thinking about it for the past few years is that it's like a racetrack. At the starting line, there are all these influenza viruses that are found in animal species. So flu is not only a human pathogen, it's also an animal pathogen, which is why it always poses such a large pandemic risk and a huge public health burden. In these animal species such as chicken or waterfowl or pig species, the virus is constantly reassorting, right? This process of co-infection and intermingling of the segments.. And these newly emerging viruses have to overcome certain hurdles in order to become epidemiologically successful in the human population because we're ultimately interested in human influenza pandemics. So these viruses have to overcome the first hurdle, which is they have to have access. They have to be able to infect a human host. And so it's easier for some animal viruses than others to infect a human host. This leads to a smaller pool of viruses, and these viruses have to overcome the second hurdle, which is to replicate efficiently within the human host. Do they use the right receptors? Can they internalize? Do they use the right host machinery for replication? All of those can sort of limit this bottleneck, and this— only a few viruses can overcome this hurdle. So again, we're narrowing down. And the last hurdle, which is probably one of the biggest hurdles, is that they have these viruses have to be able to transmit between the air, between people, and only viruses that can do all three of these tasks could become a pandemic. And so my lab is focused primarily on these last two hurdles here: replication inside of the cell as well as airborne transmission. And so I'll tell you two stories today, one on replication, which we'll start with, and then one on transmission. And so this is the sort of a rough schematic of the influenza viral life cycle. So influenza viruses bind to cell receptors, the sialic acids on the cell surface of a plasma membrane of a cell. They get internalized. Upon low pH, the HA will facilitate fusion of the viral membrane with the endosomal membrane, and it'll release the genomic material into the cytoplasm. Influenza is an RNA virus unlike other RNA viruses in that it replicates in the nucleus. And so the viral RNA are transported into the nucleus, which is a site of viral RNA synthesis and replication. And these newly synthesized viral RNA segments have to then transport out of the nucleus and assemble into a complex of all 8 towards the plasma membrane for budding, right? As influenza buds from the plasma membrane. And so there's a lot of questions we didn't know about this. And one of the things that I've been interested in answering for a long time is how do influenza viral RNA segments sort of assemble inside of a cell? And so I might refer to viral RNA as vRNA, and this is the genomic RNA, so the negative sense, not any of the mRNA or the crRNA, which is the complement replication intermediate. So the model, after I worked on this for quite a number of years, that we proposed after using a variety of imaging techniques was that influenza viral RNA segments seem to replicate in the nucleus and form subcomplexes of more than 2 but less than 8. And these subcomplexes export together, and then they hitch a ride on endosomes. And these endosomes undergo dynamic fusion or fission events, and that helps facilitate a complex of all 8. And we've been, you know, we use this to sort of study viral RNA segments, but this opens a lot of questions as to the host proteins that are important for the transport machinery and the transport of viral RNA from the nucleus all the way to the plasma membrane for budding.. And so for the past few years, we've been interested in examining some of the host proteins involved in this process. And so we started looking at the endosomes. So these are actually Rab11a-containing recycling endosomes and also the cytoskeletal proteins that are important for transport. So this just shows that viral RNA— so this is actually immunofish, so which is immuno immunofluorescence combined with fluorescent in situ hybridization. So here we're looking at the viral RNA segment for PA that encodes for the PA protein, and here is RAB11A, which is a small GTPase that marks recycling endosomes that typically go apically in differentiated airway epithelial cells. And it's previously been shown by other groups that RAB11A can specifically bind to PB2. And again, this is the the viral RNA. This is sort of the genomic architecture of influenza, and there are three polymerase components, and PV2 is one of the viral polymerase components that is thought to bind to Rab11a. I do want to highlight that, you know, I sort of draw different— that there are these regions that are exposed. And so there was work that we published a couple years ago now showing that the viral RNA are actually, have a different architecture, that there are regions of the RNA that are exposed and that are not bound by nuclear protein. I'm not going to talk about that today, but I'm happy to take questions if anybody has any questions about that. So we wanted to start looking at Rab11a, and one of the first questions we wanted to ask was whether viral infection could alter the transport dynamics of Rab11a. Does it change the movement of Rab11a if it's hijacking it for its own purposes, then presumably it is. And so in order to actually image endosomal movement inside of a cell, because it happens so quickly, we actually utilized a sort of emerging, emerging, but sort of novel microscopy techniques. And we constructed a light sheet microscope that was developed by our collaborators at the National Institute of Biotechnology and Bioengineering at the NIH, Hari Shroff, and it is a light sheet microscope, and it's actually a dual view light sheet. To conventional spinning disk confocal microscopes is that in the spinning disk confocal, you are illuminating an entire sample, right? So imagine you're trying to image a golf ball.. And in a confocal, what you'll do is you have a flashlight and that's your laser source. And you shine this flashlight on this golf ball and you illuminate the entire ball, but you only take a single plane. You only capture light from a single plane and it'll look like a circle. And then if you move through the golf ball in 3 dimensions in a confocal and you reconstruct it, you'll get what looks like an American football, right? It's an oblong because there is a distortion in the Z-plane. So the XY axis, the XY resolution is similar, but it's almost 2 to 3 times in the Z direction. And then light sheet microscope, what happens is instead of using a flashlight, we're using a very thin sheet of light. And so we are illuminating a very single plane, and that is the exact same plane that we are imaging. And so there isn't a lot of photo damage to the rest of the sample as you capture all of the different planes. Because of this decrease in photodamage, we have an imaging rate that is 30 times faster than a regular spinning disk confocal for a similar signal-to-noise ratio. So the movies I'll be showing you today, if I forget to mention it, then we're doing an entire cell volume, which is 50 slices, in about 500 milliseconds. Right, so that's less than a second to get an entire cell volume. And so we're doing really high temporal imaging rates.. Also, a power of the light sheet microscope is that because of this dual view, we have one view on one side and another view on another side. We can capture— we can go back to isotropic resolution because we can now use— we can fuse and render the images such that we can remove the distortion in the z-axis. And so now we have similar resolution in x, y, and z. And so this allows for us to track things with high accuracy inside of a cell. Okay, so this is the light sheet microscope that we built at the University of Pittsburgh, and this work was done by Amar Bhagwat in my lab, and Eric Dutroux helped him, and Amar built the whole system. So here is our pipeline. So these are cells that we have that stably express this GTPase RAB11A that is GFP-tagged. These are A549 cells that stably express them, and what we What we can do is we can have a field of cells and we can isolate a single cell, demarcate the coverslip, identify the nucleus. And RAB11A, this is typically where there is no nuclear, no signal because RAB11A doesn't go into the nucleus. We can then identify spots and we can track these spots using an imaging software known as Amaris. And this will give us tracking in X, Y, and Z dimensions. And then we could export all of this position and time data and do even more analysis using custom algorithms that we've developed in MATLAB. And so we can now sort of decipher the movement of the intracellular movement of host proteins, and I will talk about viral RNA as well, using this pipeline and giving some quantitation to the movement. And so what do we do to move? Look, how do we analyze this movement? So the first that we look at is track length, which is how long is the track. And so you can see this is sort of a spot and it moves a certain distance. The displacement is actually how far the spot has traveled, and the duration is the amount of time. We can then use these values to give us information on the speed as well as the velocity of a track.. And so we can use those parameters to tell us, you know, how fast foci are moving and the, and as well as the velocity of which they are moving. In addition, you can see that not all tracks are the same, right? So here are 3 tracks that we typically see where there is some local movement, some long-range movement, and somewhere in between. We also look at a parameter known as straightness, which is how direct is this path.. And this is typically from 0 to 1. And so track 1 here would have a straightness of about 0. Track 2 would have a straightness of about 1. And track 3 would be somewhere in the middle. We also look at a rest coefficient, which is how often do spots sort of have local motion along a track. So they'll do some spots will go local, have long range, and then more local motion. So how often does a spot halt along its track?. And we think this is important for understanding how viral RNA segments may all come together en route to the plasma membrane. So within a given cell— so I just told you our pipeline for imaging a single cell, we image multiple cells, but within each single cell there are thousands and thousands of tracks that we can capture and analyze within a 5-minute movie. Again, we're capturing our entire volumetric imaging rate is about half a second, 700 milliseconds, using both arms. And so we have lots of temporal information even in a 5-minute movie. And so we have thousands and thousands of tracks, and we want to analyze all of these tracks but also compare them between cells, right? So we understand the cell-to-cell variability. And so to do this, instead of comparing histograms of different parameters between cells, we also want to compare these to different conditions. We use something known as the cumulative frequency. So we've taken all of the histogram of frequencies of a set of tracks inside of a cell. We plot them as a cumulative frequency and then we take the 50% value of the cumulative frequency so that we can distill all of this information of all of these tracks inside of a cell down to one value that we can then compare between cells of the same treatment group or between different treatment groups. So here is just another movie because it's fun to watch movies. So this again is the GFP-tagged RAB11A and these are uninfected cells and you can just sort have an appreciation for the amount of movement that happens sort of apically as well as around the entire cell, and that the majority of the RAB11A tends to be bound to the apical surface of the plasma membrane. And so going through some of these parameters that I talked about, right, looking at duration and track speed and straightness and arrest coefficient, what you can see is that influenza infection— so now here we infected with an H1N1 virus, the WSN strain, which is a lab adapted adapted strain. You can see that influenza infection produces RAB11A vesicles that tend to be longer. So the movement is now longer, slower, and more arrested in the presence of influenza viral infection. And so how do we think viruses are doing that? I'm not going to go through too much more data, but we have some really interesting data where we think that virus, viral RNA is actually outcompeting certain motor proteins. And so here, if you imagine, RAB11A is a small GTPase, and it facilitates binding to different motor proteins on an endosome by binding to these inner— these, um, what are called FIPs, which are these sort of interacting domain proteins. And so RAB11A will bind to multiple FIPs that either bind to dynein or kinesins that move on microtubules. It can also bind to myosin, which moves on actin. So it has the ability to move along both microtubules as well as actin filaments. And perhaps this— for this audience, I should, I should mention that, you know, endosomes don't just move by binding to a single motor, right? They don't move as we walk in a very directional manner. They actually have a similar amount— they actually have differing amounts of plus-end and negative-end motors. So dynein and kinesin are found on an endosome, and they tend to do this rock-and-back motion. And by changing the proportion of motor proteins, you can alter the dynamics and the movement of these endosomes. And so what we think is happening, and we have some evidence for this, is that the viral RNA segments are outcompeting certain FIT proteins or motor proteins, and that is producing, changing the amount of different motors that are on these endosomes. In particular, we think it's it's altering the amount of dynein that are on the centrosomes. But we really, going back to some of our original questions, are, you know, what other host proteins are involved? And we said we wanted to assess the impact of cytoskeletal proteins on influenza viral RNA transport. And this has been done a little bit in the literature, and people have used drugs to look at the role of microtubules and actin during viral infection. And You know, there's some discrepancy about how large of an effect these different cytoskeletal proteins have on viral replication, and so we wanted to take a more mechanistic viewpoint on this and look in live cells as to what happens. So we wanted to define the cytoskeletal proteins that facilitate viral RNA transport, and we started with microtubules. And the reason for that is because RAB11A is thought to primarily move along microtubules in many cell types. Even though it can move along actin, it tends to prefer to move along microtubules, especially for its apical transport from the microtubule organizing center out. And so if you treat cells— that these are our stable cell lines— either with DMSO or with nocotazole, which is a drug that will depolymerize microtubules, what you hopefully will notice is that the amount of microtubule localization is sort of disrupted. It's pushed to the periphery and it's no longer concentrated inside of the cell.. It also changes the apical distribution. So this is a sort of confocal slice. And so we started with nocotazole, and again here I'll just show you some movies. So this is untreated cells. These are stable expressed cell lines. These are uninfected, and you see it's nice dynamic movement. And now if I— if we treat these cells with nocotazole and do imaging, what you will hopefully notice, a very stark difference in the movement. So this movie is actually playing, but the RAB11A is not moving as dynamically as in untreated cells. And also, when we have looked at the titers or replication capacity of the 2009 pandemic H1N1 virus in cells treated with nicotazole, and we haven't seen the difference at multiple MOIs in MDCK cells or in A549 cells, is showing you here that our drug treatment is working. We've also done this in differentiated airway epithelial cells. So these are human bronchial epithelial cells that are grown at an air-liquid interface on transwells, and you can treat with nocotazole, which breaks down all the microtubules that are in the beautiful cilia of these cells. And we see no difference in growth of the 2009 pandemic H1N1 virus in this system either. So this is really striking to us because clearly there is a large impact on the movement of RAP11A when we treat with nocotazole or drugs that depolymerize microtubules, but there doesn't seem to be any impact on viral replication. And so we wanted to visualize the dynamics of viral replication or viral transport in cells treated with nicotinamide monophosphate. And for that, we turned to a virus that I developed many years ago, which the PA protein, which again is one of the components of the polymerase complex that is bound by the 3' and 5' ends of the viral RNA, is bound to GFP. And these cells, so here are cells that have this PA-GFP virus, so it can infect cells, it turns them green, and we can stain for the nuclear protein. And we find that roughly about 85% of all of the foci, almost 90% in A549 cells, that are GFP positive, this PAGFP positive, also are NP positive, suggesting that it is a good surrogate for viral RNA to study viral RNA dynamics in live cells.. And so here are again two cells, so this is DMSO treated, so these are now A549 cells that are infected with the 2009 pandemic virus that it has a PA that is fused to GFP, so full-length GFP. Oops, I want to play this movie. Yep, okay, so if you treat with DMSO, so these are, you know, normal cells, this is you're looking at viral RNA movement now in the GFP, as you see it's very dynamic. If you have these cells and you treat with nocotazole, you actually still see very dynamic movement of viral RNA segments. And so there clearly is, as you would expect from the viral titer data, that viral RNA are still being transported around the cell and packaged efficiently. If we look at the dynamics of the transport, we see that the movement of the PAGFP vesicle foci within a cell, we see that the dynamics are very similar in the absence of microtubules. And so. There's no real change in track displacement, a little bit slower track speeds, and a little bit less straight. But there's no difference in the amount of times it arrests. And this is in stark contrast when you look at the movement of RAB11A vesicles when you treat with nocautisone. And so we believe that there is a microtubule and RAB11A independent mechanism of transport. This is also confirmed a little bit by looking at the colocalization of RAB11A. This is again stable cell lines expressing Rap11a-GFP that we then infected with this WSN H1N1 strain. And we did, this is immunofish, and we stained for the viral RNA segment here, PA. And we can then quantify and do fine colocalization on fine Z-stacks. So we do confocal imaging where we take fine Z-stacks, and then we can colocalize spots inside of a cytoplasm. To look at things that are related. And so if you have untreated or DMSO-treated versus nicotinamide treated, you see that the amount in the nicotinamide treated group that are co-located RAB11A-GFP and viral RNA go down, and this is resulted by an increase in the amount of viral RNA foci alone. So to summarize this part of my talk, Hopefully you see the convincing data that influenza viral RNA segments alter the movement of Rab11A to become slower and less directed. Rab11A movement is microtubule dependent. However, intact microtubule filaments are not necessary for influenza viral replication or viral RNA transport, and that there's only a slight difference when you depolymerize the move— there's only a slight difference in the movement of viral RNA when you depolymerize microtubules. And so this sort of helps identify that viral infection alters the transport machinery associated with RAB11A recycling endosomes, and that there are redundant mechanisms of transport, right? There clearly is a RAB11A and microtubule-independent mechanism of transport that we are actively going after. And so coming back now to our track, I'm going to switch directions and talk about the third hurdle and some work that my lab has been doing to sort of to understand the viral properties that promote airborne transmission. And so how does flu transmit? For those in the audience who don't think about transmission as much as I do, you can break this down into 3 ways. So there is contact transmission. This can be direct or indirect. So either direct is in the presence of a contaminated— in the absence of a contaminated surface, so either kissing or shaking hands. And indirect involves a contaminated surface. The other is respiratory droplet transmission, and so you can see in the sneeze diagram that there, when you breathe out or when you sneeze, you just expel a wide range of aerosols, and there are some fine aerosols and there are these larger droplets, and these larger droplets will fall out of the air, and so this is known as droplet spray transmission, where these larger droplets can fall onto either people, the mucous membranes, and cause disease, or aerosol transmission, which are these fine mist aerosols that you can breathe into your lower lungs. And typically these aerosols are less than 5 microns in size. And so we've been thinking about sort of how influenza transmission happens and some of the parameters that modulate influenza transmission quite a bit since I was a postdoc with Kanta Subbarao at the NIH. And so you can imagine that there are, you know, you can break transmission up into 3 different zones, right? There is the donor, And so all the things that are happening within that donor, the virus that's replicating within that donor, where it's replicating, are all important for thinking about airborne transmission. Then there is the environment, because in order for transmission to actually function, these expelled viruses have to remain infectious in the environment for a considerable amount of time. And then there is the recipient and, you know, the immune status of that recipient. Are they susceptible? Are they resistant to infection? And these are all modulation of any of these three parameters could impact the transmissibility of influenza viruses. And so the story I'm going to tell you about today is sort of looking at the role of environment because this is a critical aspect. Because you can imagine that we, by understanding how persistent and how the environment manipulates expelled influenza viruses and droplets, we can start to think about non-pharmaceutical intervention strategies to limit the spread of influenza in during seasonal infections or during future pandemics. And so we started with the idea of relative humidity because that's been proposed in the literature quite a bit to impact transmissibility of influenza viruses. And so the prevailing dogma, not just with viruses or influenza viruses but even with bacteria and fungus, is that microbes are are less pathogenic at mid-range relative humidity conditions. And so this is sort of an old chart, and I just want to break this down. So in the wintertime, if you are in a temperate region and you— it's cold like in Pittsburgh and you have the heat on, you're somewhere within this 20 to 30% relative humidity range because the heater is drying out the air. In the summertime, you tend to have the AC on and so you have, you know, the humidity is a little bit wetter and it's about 40 to 50%. And so this mid-range is thought to be this optimal environment in which we can live in. So we wanted to ask a very simple question: whether relative humidity impacts the viability of seasonal influenza viruses, human seasonal influenza viruses. And we did this in collaboration with Lindsay Marr at Virginia Tech who is an aerobiologist and has published a couple of papers on this topic over the over a number of years. And this work was done by Karen Kormuth in my lab, a postdoctoral fellow. And so working with engineers is really fun because they can build all sorts of neat things to test hypotheses. And so we went back to some old information on rotating drum vessels. And so Lindsay and her team constructed a rotating drum with tunable relative humidity. So we have this drum, it's it's in a biosafety cabinet in my lab on occasion. And we can first condition the drum to different humidity conditions. And we can, with dry air and wet air, and then we can let, we can nebulize virus into the drum after it's been conditioned. And at the end of the drum on the other side, there is an environmental port so we can monitor the humidity and temperature. And then we can close off the drum and we can rotate it. And I'll show you this little movie here of it rotating. We can open up the port on the back and then collect aged aerosols after 1 hour in these different humidity conditions. And this nebulizer will make very small submicron aerosols. So in order to study this, right, this is, you know, this question has been asked in the past, and most people have used either lab-adapted strains of influenza viruses, or they've grown viruses in MDCK cells or in eggs, and those don't really mimic the physiological droplets that we breathe out or sneeze out as humans because we have a lot of mucus in our airways. And so in order to recapitulate this, we actually use differentiated airway epithelial cells. So these are human bronchial epithelial cells. These are grown in a 3-dimensional culture, so they're on a transwell. They are differentiated at an air-liquid interface, and they form mucus because they're at this air-liquid interface. They form a lot of really interesting mucus that we can collect, and we call this airway surface liquid. And we can combine this airway surface liquid with viruses that we have grown in, um, MDCK cells, a very traditional tissue culture method, and then we can nebulize this mixture into the rotating drum and we can ask questions about how relative humidity will impact the stability of the 2009 pandemic H1N1 virus in aerosols at different humidity conditions. And we did a range of humidity conditions. And I think you will notice from this graph that there was no difference. So I'm just showing you here, these first few bars are indoor heating, and these are sort of the, what we would consider the optimal indoor AC conditions. And so this was very striking to us in that we didn't see any decay of influenza viruses after an hour at these different relative humidity conditions in suspended aerosols. We also used Phi6, which is a surrogate that can be done down at Virginia Tech, to test whether this airway surface liquid is really the one that was protecting it. So Phi6 is also an enveloped bacteriophage, and you'll see that if you supplement Phi6 media with airway surface liquid from human bronchoepithelial cells, you see no decay. And when you have just media without any of the airway surface liquid, you do have decay at this mid-range relative humidity condition, which would suggest that the virus is less stable. And we also needed to show that the nebulization didn't impact the virus suspension, and so we did studies looking at the titer pre- and post-aerosolization of the nebulized material, and we didn't see any differences in titer. So nebulization didn't have an impact. So this is really interesting because in aerosols it suggests that viruses that are released into these small aerosols are highly persistent in our environment, and we wanted to correlate this with droplets because again fomites and large droplets that fall out of the air can also facilitate the spread of influenza viruses in a community. And so we used a chamber system where we can manipulate the relative humidity within a chamber by different by having different saturated salts within that chamber. So it's just a desiccated chamber. We have a relative humidity and temperature logger that is constantly monitoring the temperature and humidity within this chamber. And we can make small 1 microliter droplets onto just a tissue culture plastic dish and do it in 3 replicates to see whether in droplets humidity has an impact on the stability. If we just take MDCK grown viruses, again this is the 2009 pandemic H1N1 virus, you see that there is a loss in titer. Again, this is raw titers here at mid-range humidity conditions. I'm probably going to show most often now log decay, which is virus viability goes down or titer goes down, log decay goes up. If you see this same data as plotted as log decay, you see this sort of V-shaped curve. This is what has been published in the literature and found by previous groups. We then wanted to ask whether the airway surface liquid could protect viruses from decay similar to what we saw in aerosols. We removed airway surface liquid from uninfected human bronchial epithelial cells at this 3-dimensional airway cultures and combined it with MDC-K grown virus. And we did the same sort of droplets at a range of different humidity conditions. And what you can see is in black are the droplets where you are incubating the viruses with the airway surface liquid from epithelial cells, and in red are the MBC Coagulovirus that I just showed you before. And clearly there is less decay, so again this is log decay, in droplets that contain in human airway mucus, suggesting that airway mucus can provide some sort of microenvironment that protects the virus from decay in these environmental conditions. But these are all supplemented with mucus, and so we wanted to ask whether the lung environment can change the relationship between viruses and humidity. And so now we grew viruses in HBE cultures, and so viruses, especially human seasonal viruses, grow quite well in these cell culture models. And they will, when the virus that we collect from them will already be coated in this mucus environment in these airway, the proteins that are sort of expelled from airway epithelial cells. And we did similar experiments where we looked at the stability of viruses in droplets at different humidity conditions when they were grown in these HB cells. And so here I'm just showing you the raw titers. So there'll be two graphs of raw titers and then I'll show you the log decay decay. And as we've done this over and over again, what we've noticed is that when we believe that viruses are sensitive to different humidity conditions, that sensitivity, we're sort of marking a benchmark of 2 logs decay. So here in these MDCK grown at 29, the 2009 pandemic grown in MDCK cells at 75% and 85%, there's 2 logs less in these compared to a no-chamber control and compared to low relative humidity conditions. If we take virus of the 2009 pandemic that's been grown in these HBE cells, we no longer see this decay at these humidity conditions, and that you can see here in the log decay format. So in seasonal, so we did a seasonal H3 antivirus, is actually quite interesting. We did, we see some decay but not this too long decay in in the 2000, in H3N2 virus from 2009, this Perth H3N2. But if you grow them in HB cells, you no longer see any decay. Again, they're highly stable. They seem to be in droplets at a range of relative humidity conditions. And then when we looked at flu B, it was quite interesting in that flu B sort of was different in that we did again see this greater than 2-fold log decay in MDCK-grown B viruses. So for those who, who may not know, during an influenza season, not only is there influenza A circulating, but also there's influenza B circulating. And then we also see this decay greater than 2 logs in viruses, influenza B virus that are grown in HB cells, suggesting that influenza B viruses perhaps are not protected by an interaction with the airway surface liquid, and that there may be components or virus-specific factors that are promoting this type of protection. And so I'm going to end with a model, and what we think is happening here is sort of like trying to, you know, bring all this data together, and we have some other ones, some other data as well, how this sort of, you know, impacts transmissibility, because we do know from epidemiology data as well as animal model data that at mid-range relative humidity conditions there is a decrease in transmission. And so what we think is happening is that perhaps it's more about the physics of the droplet. So at low relative humidity conditions, the virus is sort of remains in these smaller droplets for longer and can facilitate more long-range transmission. Whereas higher droplets at higher RH, the droplets are able to take on more water and fall out of the air at a faster rate. And we know that sometimes contact, indirect contact, as well as fomite transmission are less efficient at long-range transmission than aerosol transmission. So we think primarily during a season, aerosol transmission will drive transmissibility. In some context. And so just to sort of sum up, my lab at the University of Pittsburgh has sort of two areas that we are focused in. We're looking at the role of influenza virus assembly and reassortment, as well as properties that promote airborne transmission of influenza viruses. And our ultimate goal is that we can combine these two areas to get a better assessment of the pandemic potential of circulating zoonotic viruses and really to help help, you know, have a way of screening emerging pandemic threats in a more sophisticated manner. And so this— acknowledgments— this is my lab. I've tried to highlight everyone's work. This is our funding sources. Again, Amar developed the light sheet microscope and constructed it here at our lab in collaboration with Hari Mike Meyerberg and the folks at the University of Pittsburgh Airway Tissue Core have been fantastic in getting us airway, these human airway epithelial cells. Also, we have a great collaboration with Lindsay Martin at Virginia Tech to study the persistence and longevity of influenza viruses and aerosols and droplets, and that work was done by Karen. If you want to see our movies and some of the things that we're working on, we always try to post new movies from our DICEMIM at our website. Wattawallab.com. You can follow us on Twitter. All of the code that we've generated to analyze the tracks of viral RNA as well as rabble RNA in 3 dimensions is available on our GitHub site. These are again our funding sources. And I want to just highlight this past year myself and Ed Hutchinson sort of published a series of educational material influenza viruses. So Ed made This very beautiful coloring page, and I made a connect the dots to sort of highlight the new architecture of the influenza genome. And we have a word find. So if you know any teachers or you teach undergrads or high school students and you want to explain influenza to them, we have a couple of materials that can be used. This is open source, and so it should be freely available to everybody. There's also a fact page that we made. So I would love for people to use it.

Um, and with that, I will take your questions. Yes, thank you, Sima. That was so interesting. I loved your pictures and videos. Um, at this point, we can't take any questions. Just type them into your question box, and I will read them. And while we're waiting for others to kind of get their stuff typed in, I'm curious if you could give us a little bit of information on the RAB11 independent pathway that you talked about in your first section.

You said you guys are chasing that down, and I'm interested what you found so far. Yeah, actually, you know, it's interesting. There was a really nice paper from Malvina Fox's group where they have shown that the endoplasmic reticulum, and I think it's using Rab11A, but that viral RNA can actually also transport along the endoplasmic reticulum membrane, right? So if you imagine, there's a lot of membrane inside of a cell, right? So there are these endosomes, there's the ER, there's the Golgi, and so it sort of is sparking this this thing that we're thinking about, in that perhaps the viral RNA just want a membranous scaffold. And the ER is quite interesting because using lattice light sheet microscopy, folks in Eric Betzig's group has shown that the ER can actually expand large spaces within the cytoplasm very rapidly, and they don't need microtubules or cytoskeletal proteins to do this. How that sort of like forming its tubular sheets and like sort of branching off and anchoring points is sort of an interesting biophysical property of the endoplasmic reticulum. And so right now what we're doing is developing pipelines to visualize endoplasmic reticulum on our DICEM, similar to the work that's been done by Eric Betzig and his group, and then combining that with our fluorescent viruses to see is the viral RNA on these endoplasmic reticulums, is RAB11A there, in the absence of RAB11A is that what it's being transported along And so that's what we're chasing right now.

We do think that they're membranous vesicles, but yeah. Got it, awesome, thanks. And we do have another question. I think it says, what do you think makes H1N1 different to single it out for the protective effect by the lung mucus microenvironment?

Yeah, that's a really interesting question. So this has been a really fun experiment, and one, some slides I took out is that we've also started to look at the longevity of the response, and we see that there is subtype-dependent longevity. So H1N1 viruses tend to behave a specific, different way and how long they're persistent in droplets than H3N2 viruses as well as influenza B viruses. And what we think is happening is we think that there are specific viral properties and we're very interested in the HA protein. So we've done some initial mass spec analysis of our airway surface liquid and we find that there is a lot of, of course, as you would expect, sialic acid and proteins that are heavily glycosylated there, as well as mucin, which is heavily glycosylated, and influenza HA binds to sialic acid receptors. And, you know, H1 and H3 would have similar sialic acid preferences, but the avidity of which they bind may be slightly different. And so we're looking at how the HA/NA balance of H1 versus H3 is help facilitating their stability within these droplets and how that could also be responsible for this transmission properties that we see. It is quite interesting in animal models, we have some data suggesting that the H1N1 pandemic virus is heavily transmissible, right? So that virus will transmit very efficiently even within a very short period of time. That's been shown by other groups as well as us, that within 2 days you'll get 100% transmission of the 2009 pandemic H1N1 virus. And that's not true with seasonal H3N2 viruses. And so So understanding how sensitive these viruses are and the differentiation between these viruses and their sensitivity in the environmental conditions, it might be really important for teasing apart these differences that we see in the transmission.

Got it. It looks like there are no other questions. So Seema, I want to thank you again for a wonderful talk, and thank you to everybody who joined.

And this will be available on the Human Vaccines Project website later today if you want to watch.

So thank you. Thank you all so much.