Building multicellular structures during development: New roles for Toll receptors (Part1): Jennifer Zallen

00:00:14.00 Hi.
00:00:15.00 My name is Jennifer Zallen and I am a professor at Sloan Kettering Institute and an investigator
00:00:19.00 with the Howard Hughes Medical Institute.
00:00:22.00 And today I'd like to talk to you about building multicellular structures during development.
00:00:26.00 So, each one of us is a multicellular animal that contains billions of cells, each with
00:00:31.00 their own intrinsic properties, and these cells have to work together to build the characteristic
00:00:36.00 shapes and features of different animals.
00:00:38.00 So, our bodies contain many different kinds of cells and these cells often have to move
00:00:42.00 some distance from where they're born to where they need to end up in order to carry out
00:00:46.00 their functions in the animal.
00:00:48.00 So, in my lab we're interested in understanding how cells get where they need to go during
00:00:52.00 development and how many cells work together to build functional 3-dimensional structures.
00:00:57.00 So, one way to think about this problem is to imagine a crowd of people at a busy train
00:01:02.00 station.
00:01:03.00 Each person has somewhere where they're going, but how they get there can be quite complicated.
00:01:07.00 So, to get where you're going and navigate through this complex environment, you need
00:01:11.00 to know where you are in space at multiple levels.
00:01:15.00 For example, you need to know where you are relative to the 4 or 5 closest people around
00:01:19.00 you, but you also need to take into account the overall flow of the crowds, the push and
00:01:23.00 pull of crowds of people as they move through the station, as well as know where you are
00:01:28.00 on a map relative to your final destination.
00:01:30.00 And all three levels of information are necessary to take the most efficient path to where you're
00:01:36.00 going.
00:01:37.00 So, in my lab we're interested in understanding how cells navigate as they move through the
00:01:42.00 body during development, and to get to the places where they need to end up, cells receive
00:01:50.00 spatial information from many different sources.
00:01:52.00 So, for example, cells can interact with other cells in their immediate environment, cells
00:01:58.00 generate forces in order to move and they in turn are exposed to the forces generated
00:02:04.00 by cells around them, and finally each cell needs the equivalent of a compass that tells
00:02:08.00 it where it is relative to a global map of the entire animal.
00:02:12.00 Shown here is a map of the Drosophila embryo, the fruit fly embryo.
00:02:17.00 So, all three types of information are necessary to build the elaborate structures that are
00:02:23.00 generated during development, as well as to enable the regeneration and repair of these
00:02:28.00 structures in adult animals.
00:02:31.00 so, today I'm going to talk about an organization problem faced by many animals, which is how
00:02:36.00 to build a body axis that's elongated from head to tail.
00:02:41.00 And this is a process that happens in fish, it happens in frogs, and it happened to you
00:02:46.00 or you wouldn't be watching this talk.
00:02:48.00 And during this process, all of these animals have to reorganize an amorphous mass of cells
00:02:53.00 that can be hundreds of cells or more in size into one that's highly elongated along one
00:02:58.00 axis to produce the characteristic elongated shape of the body plan.
00:03:03.00 Now, in my lab, we study this problem in the fruit fly Drosophila, shown here on a banana.
00:03:10.00 Drosophila is an excellent system for studying developmental biology; flies are very easy
00:03:16.00 to grow and work with in the lab, and importantly the structural changes that build the head-to-tail
00:03:21.00 body axis happen at the surface of the embryo, where we can directly visualize what the cells
00:03:26.00 are doing by live imaging.
00:03:28.00 Now, it's important to understand this process not only because many different types of cells...
00:03:35.00 different types of tissues have to elongate during development such as the lung, the gut,
00:03:39.00 and the spinal cord, but also if we can understand how cells self-organize to form complex structures
00:03:45.00 during development, we would be that much closer to helping to fix structural defects
00:03:50.00 that arise during human disease.
00:03:53.00 So, this is a picture of the Drosophila embryo undergoing elongation, and the embryo before
00:03:58.00 elongation is shown here at the top.
00:04:01.00 It's about half a millimeter long and at this stage the embryo is basically just a single
00:04:05.00 monolayer of cells surrounding a ball of yolk, so all the cells are located at the surface
00:04:10.00 of the embryo.
00:04:12.00 And the head is to the left, the tail is to the right, and as... during elongation, the
00:04:19.00 germ band indicated by the yellow line undergoes a major reorganization in less than one hour
00:04:24.00 to more than double in length along the head-to-tail axis and at the same time it shrinks to half
00:04:30.00 its initial height along the perpendicular axis.
00:04:33.00 And because the embryo is encased inside a protective membrane that surrounds the embryo,
00:04:39.00 as it gets longer it's forced to wrap around on itself so that the head ends up next to
00:04:44.00 the tail.
00:04:46.00 Now, Ken Irvine and Eric Wieschaus showed that these structural changes occur primarily
00:04:53.00 through organized cell movements, so while the embryo as a whole is elongating from left
00:04:58.00 to right in these images, individual cells within the embryo are moving up and down in
00:05:03.00 a coordinated fashion, and this is just shown in these pictures, here.
00:05:07.00 And what you can see is that cells that... basically, rows of cells intercalate to form
00:05:14.00 fewer, longer rows.
00:05:15.00 So, for example, if you look at the cells in this yellow row, here, by the end of the
00:05:19.00 process the yellow cells that were all originally next to each other in the row become separated
00:05:24.00 by neighboring cells from the green and purple rows.
00:05:28.00 So, to understand how cells reorganize to get from here to here, Dene Farrell and Ori
00:05:34.00 Weitz in the lab came up with a computer algorithm that can automatically identify and track
00:05:40.00 cells in time-lapse movies.
00:05:42.00 And what they found is that individual cells not only navigate as solitary cells through
00:05:50.00 this tissue, but they actually engage in collective cell behaviors that enhance the speed and
00:05:56.00 efficiency of elongation.
00:05:58.00 So, this slide shows the two types of cell rearrangements that contribute to elongation.
00:06:03.00 Each row of [images] shows the same group of cells from three time points from a time-lapse
00:06:08.00 movie.
00:06:09.00 And the top panels show this process of local neighbor exchange, discovered by Thomas Lecuit
00:06:14.00 and colleagues, where a single vertical edge between two cells, shown in red, and is replaced
00:06:21.00 by a horizontal contact, here in yellow, between the top and bottom cells that were previously
00:06:27.00 not in contact with each other.
00:06:30.00 In addition, through live imaging of many cells, we were able to find that these cells
00:06:35.00 also engage in this unexpected collective rosette process, where many linked edges,
00:06:43.00 shown here in red, shrink simultaneously, leading to the formation of a multicellular
00:06:48.00 rosette structure where as many as 10 cells can come into contact at a single point.
00:06:53.00 And these structures always form and resolve in a strictly directional fashion, so they
00:06:58.00 form through the contraction of vertical edges and they resolve through the formation of
00:07:01.00 these new horizontal edges, shown here in yellow.
00:07:04.00 And what this does is convert a group of cells that was initially tall and thin into one
00:07:09.00 that's now short and wide.
00:07:10.00 And you can imagine, if many cells are doing this throughout the tissue, that this would
00:07:13.00 cause an overall elongation of the embryo.
00:07:18.00 So, to show you what this looks like, this is a movie of axis elongation in a Drosophila
00:07:22.00 embryo where each polygon is a cell and the cells are outlined by the green fluorescent
00:07:28.00 protein tagged to a membrane marker.
00:07:32.00 And the head is to the left, the tail is to the right, and the cells that we're interested
00:07:36.00 in are in the center.
00:07:38.00 And as the embryo elongates, the cells at the bottom will internalize, but we focus
00:07:42.00 on the cells that remain at the surface, which drive this elongation.
00:07:45.00 And you can tell that the embryo is elongating because cells are flowing to the right, out
00:07:49.00 of the field of view, as the tail is moving further away from the head.
00:07:53.00 And what's driving this elongation is groups of cells within this tissue are converging
00:07:56.00 along the vertical axis and resolving in the horizontal direction through the formation
00:08:01.00 of these transient, multicellular rosette structures, where many cells come into contact
00:08:06.00 at a single point.
00:08:08.00 Now, these behaviors are reiterated throughout the tissue and are associated with a doubling
00:08:13.00 in the length of the body axis.
00:08:16.00 Now, since our discovery of this rosette mechanism for tissue elongation, rosettes have now been
00:08:22.00 shown to be a general mechanism for elongating many different types of tissue sin vertebrates,
00:08:29.00 including the chick and mouse neural plate, which are structures that have to elongate
00:08:33.00 substantially to form the spinal cord.
00:08:36.00 And these behaviors suggest that there's communication between cells; these cells are not just navigating
00:08:41.00 on their own, but they're communicating with each other to form these groups that are able
00:08:45.00 to rearrange more efficiently than the cells could achieve on their own.
00:08:49.00 And we would now like to understand how these cells generate the forces that allow them
00:08:54.00 to move and rearrange within the sheet, and how these movements are spatially organized
00:08:59.00 so that each rearrangement inexorably moves the head further away from the tail.
00:09:04.00 So, to identify the molecular basis of these behaviors, we started by looking for molecules
00:09:09.00 that are asymmetrically localized in the plane of the tissue and correlate with the direction
00:09:15.00 of cell movement, and could potentially be generating the forces that move these cells.
00:09:19.00 And this slide shows the first two molecules that we've found.
00:09:22.00 So, here you're looking at a field of around 30 cells, and in particular the myosin II
00:09:27.00 motor protein, shown here in green, localizes to the surface... the cortex of cells, but
00:09:33.00 it's not uniform throughout the cortex.
00:09:35.00 It's higher at these vertical edges between cells, creating the impression of vertical
00:09:39.00 lines in this image.
00:09:41.00 Now, the movement of myosin II motors along actin filaments generates contractile forces
00:09:47.00 within the cell, and so this suggests that contractile forces that are planar polarized,
00:09:52.00 or oriented consistently within the plane of the tissue in this vertical direction,
00:09:56.00 could help drive cell rearrangements during elongation.
00:10:00.00 Par-3 has the opposite pattern -- it localized to the horizontal edges of cells -- where
00:10:06.00 it stabilized adhesion between cells or the ability of cells to stick to each other, and
00:10:11.00 it also excludes myosin, keeping myosin at those vertical edges.
00:10:15.00 So, through work from my lab and others, we've now identified more than 10 proteins that
00:10:22.00 include both signaling and structural proteins within cells that are required to organize
00:10:26.00 these contractile and adhesive proteins within cells and are necessary for the embryo to
00:10:32.00 elongate.
00:10:33.00 So, to look at the dynamics of how these proteins organize during elongation, Sergio Simoes
00:10:41.00 in the lab made movies of embryos where myosin II is shown here in green and Par-3 is labeled
00:10:46.00 in red, and time is indicated in minutes and seconds.
00:10:50.00 And at the beginning of this movie, myosin II is mainly in the cytoplasm, so these circles
00:10:54.00 you see are exclusion from the nuclei of cells, and Par-3 is present in dots at the newly
00:10:59.00 forming junctions between cells.
00:11:02.00 And what you can see is that, as the embryo gets ready to elongate, myosin II, in green,
00:11:08.00 moves from the cytoplasm to the vertical edges of cells and Par-3, in red, becomes restricted
00:11:12.00 to the horizontal edges.
00:11:14.00 And what was striking about these movies is that myosin II is not just asymmetrically
00:11:20.00 localized within cells, but it often forms these long, extended cables that can connect
00:11:25.00 multiple pairs of cells within the tissue.
00:11:28.00 And this suggests that mechanical contractile forces aligned across many pairs of cells
00:11:34.00 could generate the forces that drive rosette formation and cell rearrangements during cell
00:11:40.00 elongation.
00:11:41.00 So, to test this idea, Rodrigo Fernandez-Gonzalez in the lab used laser ablation to cut individual
00:11:49.00 edges between cells in order to measure their mechanical properties.
00:11:52.00 So, in this first movie, Rodrigo used a UV laser to cut a single interface between cells
00:11:59.00 that's contracting on its own to drive local neighbor exchange.
00:12:03.00 And you can tell that that edge was under tension, because the two vertices attached
00:12:07.00 to the cut edge move apart when that edge is ablated, and the speed and the distance
00:12:12.00 of that retraction is proportional to the amount of tension that was acting on that
00:12:16.00 edge right before he cut it, assuming that viscoelastic properties are uniform throughout
00:12:21.00 the tissue.
00:12:22.00 So, this retraction distance over time is plotted here in this chart.
00:12:27.00 Now, the movie on the right shows what happens when Rodrigo uses the laser to ablate a single
00:12:32.00 edge, but one that's part of a multicellular cable that's going to contract to form a rosette.
00:12:39.00 And what you can see is there's a much greater displacement of the two vertices attached
00:12:42.00 to the cut edge, as well as the surrounding cells, so these edges in cables are under
00:12:47.00 more tension, and, in fact, shown here in this plot, you can see that edges in cables
00:12:51.00 are under twice as much tension as single vertical edges that contact on their own,
00:12:57.00 and both types of vertical edges are under more tension than these horizontal edges,
00:13:01.00 in green, which have much less myosin.
00:13:05.00 So, these results tell us that mechanical forces are anisotropic within the tissue.
00:13:09.00 They're highest in the vertical direction and where they're highest in the sheet is
00:13:15.00 in multicellular cables that are forming rosettes.
00:13:19.00 And we've shown that these forces can travel up to 10 microns in the tissue and promote
00:13:25.00 cell rearrangement and axis elongation.
00:13:28.00 So, we now have a lot of information about the cell interactions that occur during axis
00:13:34.00 elongation as well as the polarized force-generating machinery that actually drives these cell
00:13:40.00 movements.
00:13:41.00 But if we take a step from thinking about what's going on inside the cells to looking
00:13:45.00 at what's happening throughout the entire tissue, a major unanswered question is, what
00:13:50.00 orients these polarities across hundreds of cells so that they all move in the same way
00:13:55.00 to promote tissue elongation?
00:13:58.00 And this suggests that an additional position remains to be discovered that provides a global
00:14:02.00 map that orients these polarities so that the cells all move in the same way to move
00:14:07.00 the head further away from the tail.
00:14:10.00 Now, the only clue that we had to the nature of this map came from Ken Irvine and Eric
00:14:16.00 Wieschaus over 20 years ago, showing that striped patterns of gene expression along
00:14:21.00 the head-to-tail axis are necessary for elongation.
00:14:24.00 So, like us, flies have differences in cells along the head-to-tail axis, and in flies
00:14:31.00 these differences take the form of stripes because the body of the fly is segmented.
00:14:35.00 In particular, Eve and Runt are two transcription factors that are expressed in 7 stripes along
00:14:42.00 the head-to-tail axis, and this striped expression pattern is critical for their function in
00:14:48.00 elongation.
00:14:49.00 So, embryos fail to elongate normally when either Eve or Runt stripes are missing, or
00:14:55.00 when either of these proteins is uniformly expressed throughout the embryo.
00:14:59.00 So, these proteins don't just have to be there, but they have to be present in stripes in
00:15:03.00 order for the embryo to elongate.
00:15:06.00 And we found that these stripes are not only necessary for elongation, but they're also
00:15:09.00 sufficient to polarize cells.
00:15:12.00 And the evidence for this came from experiments that I did with Eric Wieschaus where we generated
00:15:16.00 mosaic embryos that have an ectopic source of Eve or Runt protein, shown here in blue.
00:15:21.00 So, for example, in this mosaic embryo a myosin marker, shown here in green, and it localizes
00:15:28.00 mainly to vertical edges of cells in the normal region of this embryo, but when you get to
00:15:32.00 this clone of cells that has ectopic Runt protein, you can see that myosin polarity
00:15:37.00 reorients to follow the clone boundary.
00:15:39.00 And, similarly, in this embryo that's mosaic and has an ectopic source of Eve, you can
00:15:44.00 see that myosin polarity, in green, reorients to follow the boundary of the Eve clone.
00:15:50.00 So, in each case, the orientation of cell polarity is dictated by the juxtaposition
00:15:56.00 of cells that have different levels of Eve or Runt activity.
00:16:00.00 So, these results suggest a model where differences in gene expression along the head-to-tail
00:16:05.00 axis of the embryo provides the spatial information that organizes cell movements during axis
00:16:12.00 elongation.
00:16:13.00 But I mentioned that Eve and Runt are transcription factors, which means that their job is to
00:16:17.00 localize to the nucleus of cells, bind to DNA, and regulate the expression of other
00:16:23.00 genes.
00:16:24.00 And so because of their role in the nucleus, this model is ultimately unsatisfying, because
00:16:29.00 what we really want to know are what are the targets of Eve and Runt that localize to the
00:16:34.00 surface of cells and allow cells to read the spatial cues in their environment that tell
00:16:38.00 cells which way to move.
00:16:40.00 So, to identify these targets, Adam Pare, a postdoc in the lab, used RNA sequencing
00:16:45.00 to compare gene expression profiles between embryos where both Eve and Runt were knocked
00:16:50.00 down by RNA interference with water-injected controls.
00:16:55.00 And he was looking for proteins that are targets of Eve and Runt, that are expressed in stripes,
00:17:00.00 and that localize to the surface of cells, where they could directly regulate cell polarity
00:17:06.00 and behavior.
00:17:10.00 Using this approach, Adam identified three transmembrane proteins in the Toll receptor
00:17:14.00 family as targets of Eve and Runt during axis elongation -- Toll-2, Toll-6, and toll-8.
00:17:21.00 So, Toll was originally identified in Catherine Anderson's lab in Drosophila and it's now
00:17:26.00 known to represent a large family of proteins in both vertebrates and invertebrates.
00:17:32.00 And these proteins have a conserved molecular structure including a single transmembrane
00:17:36.00 domain and a long series of extracellular leucine-rich repeats, as well as conserved
00:17:42.00 sequences in the cytoplasmic domain.
00:17:44.00 Now, these Toll-family receptors are perhaps best known for their roles in regulating the
00:17:51.00 innate immune response to pathogens such as bacteria, viruses, and parasites in the body.
00:17:58.00 And, in particular, this slide just illustrates some of the phenotypes of mouse mutants that
00:18:05.00 are lacking individual Toll-family receptors, and these mutants are defective for a subset
00:18:10.00 of immune responses in the animal.
00:18:14.00 And the signaling pathway downstream of these Toll-like receptors is very well characterized,
00:18:19.00 with the major output of this pathway being the movement of an NF-κB transcription factor
00:18:25.00 from the cytoplasm to the nucleus of cells, where it activates the expression of genes
00:18:30.00 involved in the inflammatory response.
00:18:33.00 However, while some of the Toll-family receptors in Drosophila are required for immunity, the
00:18:40.00 functions of most of the members of this family in Drosophila were not known.
00:18:45.00 So, since these genes are expressed in stripes, Adam first wanted to confirm that they really
00:18:51.00 are regulated in some way by Eve and Runt, and that's shown by the images on this slide.
00:18:56.00 So, the top rows of images show pictures of a wild type of normal embryo, and you can
00:19:02.00 see that Toll-2 is expressed in many thin stripes along the head-to-tail body axis,
00:19:07.00 and this pattern changes to become fewer, broader stripes in embryos that are missing
00:19:12.00 Eve or Runt.
00:19:13.00 Now, Toll-6 and Toll-8 are expressed in a slightly different pattern -- they're expressed
00:19:18.00 in several broader stripes -- and this striped expression is reduced in eve mutants and is
00:19:25.00 increased and is in a more uniform pattern in runt mutants.
00:19:30.00 Now, if you look at these... now, these Toll receptors are expressed in a way... in a striped
00:19:36.00 pattern that requires Eve and Runt activity for their normal pattern, and you can tell
00:19:39.00 that these three Toll receptors are expressed in slightly different striped patterns.
00:19:44.00 And putting these patterns together, you get something that looks like a positional code,
00:19:48.00 where nearly every cell expresses a different combination of one, two, or three receptors
00:19:54.00 along the head-to-tail axis of the embryo.
00:19:56.00 So, the expression patterns of these three Toll receptors relative to the segments of
00:20:01.00 the animal are indicated here, and you can see that, when you combine these three expression
00:20:06.00 patterns, nearly every cell expresses a different type or concentration of these three receptors
00:20:14.00 than their immediate neighbors to the head or to the tail.
00:20:18.00 And the evidence for this type of positional code comes from types of experiments shown
00:20:23.00 here in the bottom.
00:20:24.00 So, here, Adam looked at the mRNA expression patterns of Toll-2 and Toll-8, and you can
00:20:31.00 see that the bright Toll-2 stripes in red are largely complementary to the Toll-8 stripes
00:20:36.00 in green.
00:20:38.00 And this image on the right is an image of Toll-8 protein fused to a derivative of the
00:20:44.00 green fluorescent protein, and you can see that... it's also expressed from its endogenous
00:20:49.00 regulatory sequences and you can see that this Toll-8 protein fusion is expressed in
00:20:54.00 stripes and localized to the membranes of cells.
00:20:59.00 So, these three Toll-family receptors meet our first two criteria.
00:21:03.00 They're expressed in stripes downstream of Eve and Runt and they localize to the surface
00:21:07.00 of cells where they could be directly involved in sensing spatial information in the environment
00:21:12.00 that determines the direction of cell movement.
00:21:15.00 So, to ask if these proteins are required for elongation, Adam Pare used genome engineering
00:21:22.00 to generate embryos that are defective for one, two, or three Toll-family receptors.
00:21:28.00 And what he found is that, although you get a reduction in elongation when you remove
00:21:32.00 two receptors, you have to remove all three in order to see a strong defect.
00:21:37.00 So, working together with Zack Mirman and Dene Farrell in the lab, they made movies
00:21:42.00 of embryos that are missing one, two, or three receptors, and this slide shows a movie of
00:21:47.00 a wild type embryo on the top and an embryo that's completing lacking Toll-2, -6, and
00:21:53.00 -8, shown here on the bottom.
00:21:55.00 And time is indicated in minutes and seconds relative to the start of elongation, so this
00:21:59.00 movie starts 10 minutes before the embryo is going to elongate.
00:22:02.00 And the cells are color-coded based on the rearrangements that they've gone through;
00:22:07.00 so, they'll start off as purple and then they turn blue as they go through one rearrangement,
00:22:12.00 and then green, yellow, and red as they go through more and more rearrangements within
00:22:17.00 the sheet.
00:22:19.00 And if you focus on the top movie, you can see that the wild type cells are very active,
00:22:24.00 so the cells are turning colors.
00:22:25.00 You see many green cells, followed by yellow and then red cells appearing in the tissue,
00:22:30.00 and this is accompanied by elongation, so cells are flowing to the right, out of the
00:22:34.00 field of view, and the embryo is doubling in length.
00:22:36.00 In contrast, if you look at the bottom movie, the Toll-2,6,8 triple mutant is much more
00:22:40.00 sluggish.
00:22:41.00 There's many more purple and blue cells remaining late in the process and, as a result, the
00:22:46.00 elongation of the embryo is reduced by 30% in these Toll-2,6,8 triple mutants.
00:22:53.00 So, these Toll receptors are required for cell movement and fewer movements occur when
00:22:58.00 they're missing, and when we followed the movements that still occur in these mutants,
00:23:02.00 we found that the cells that do rearrange in this background are moving in a way that's
00:23:08.00 less directional.
00:23:10.00 So, for cell rearrangements to produce elongation, the rearrangements have to occur directionally
00:23:16.00 with respect to the head-to-tail axis of the embryo, otherwise the cells would just be
00:23:19.00 changing places and you wouldn't get a net elongation.
00:23:22.00 So, all the rearrangements have to occur through the contraction of vertical edges, separating
00:23:27.00 these cells that were originally in contact, and the formation of cells with new horizontal
00:23:31.00 edges between cells that were not previously in contact.
00:23:36.00 And when we plotted the rate of errors in this second step of edge assembly, you can
00:23:42.00 see that in wild type embryos these rearrangements occur directionally 90% of the time.
00:23:49.00 In contrast, when we looked at the error rate of edge assembly in embryos missing one or
00:23:55.00 two Toll-family receptors, what you can see is that the frequency of defects goes up to
00:23:59.00 about 20%.
00:24:00.00 And in embryos missing all three Toll-receptors, cells rearrange in a way that's not directional
00:24:08.00 a third of the time.
00:24:09.00 So, for example, in this schematic, a vertical edge will shrink to a point and the new edge
00:24:14.00 that forms also is the exact same edge as the edge that was there originally.
00:24:20.00 So, even though these cells went through the trouble of going through a cell rearrangement,
00:24:25.00 no net elongation occurs as a result, and the defects in the Toll-triple mutant are
00:24:29.00 similar to the defects in Eve and Runt triple mutants.
00:24:33.00 So, these defects in the orientation of cell movement are suggestive of a defect in cell
00:24:39.00 polarity, and in fact Adam found that these mutants have defects in the localization of
00:24:45.00 the force-generating machinery within cells.
00:24:48.00 So, I mentioned earlier that Par-3 protein, shown here in magenta, is planar-polarized
00:24:54.00 -- there's a difference between horizontal and vertical edges in the plane of this tissue
00:25:00.00 -- and you can see that there's more Par-3 at edges that are close to horizontal and
00:25:04.00 less Par-3 at vertical edges.
00:25:05.00 And Par-3 acts in these horizontal domains to stabilize adhesion between cells and also
00:25:10.00 to restrict myosin so that myosin is at vertical edges.
00:25:14.00 And what you can see from this image is that Toll-triple mutants, Par-3 planar polarity
00:25:20.00 is severely disrupted.
00:25:22.00 Par-3 now localizes to edges of a variety of orientations, so you now see many more
00:25:28.00 cells with Par-3 not just as horizontal edges but at most or all the edges of that cell.
00:25:37.00 And Adam sees a similar defect in the localization of myosin II.
00:25:41.00 So, this tells us that these cells are not just moving in the wrong direction, but they
00:25:47.00 actually seem to have completely lost their direction.
00:25:50.00 They don't have the capacity to polarize their force-generating machinery in a way that would
00:25:55.00 allow them to move in an organized way.
00:25:58.00 So, what I've told you so far is that we have these three Toll receptors that are required
00:26:04.00 for cell polarity and for cell rearrangement during axis elongation.
00:26:09.00 Now, if these are the key receptors that link the striped information provided by the Eve
00:26:15.00 and Runt striped transcription factors to the organized cell behaviors that establish
00:26:21.00 tissue structure, then the key missing experiment is that these proteins should be sufficient
00:26:26.00 to polarize cells, not just required for this process.
00:26:29.00 So, for example, I mentioned that expressing an ectopic source of Eve is sufficient to
00:26:33.00 repolarize cells in the vicinity.
00:26:36.00 And so we wanted to know, can an ectopic source of a Toll receptor have this same effect,
00:26:42.00 if these receptors are the critical targets of Eve and Runt during axis elongation.
00:26:46.00 Specifically, we wanted to know if, in a field of cells that's not normally polarized, if
00:26:51.00 you put in a stripe of a Toll receptor protein, is that enough to polarize these cells.
00:26:57.00 So, to address this question, Athea Vichas, a graduate student in the lab, expressed the
00:27:02.00 Toll proteins in stripes in the late embryo, at a stage when these cells are not strongly
00:27:08.00 polarized.
00:27:09.00 So, you can see this control stripe, shown here in blue, does not induce a strong polarity
00:27:14.00 of myosin II, in green, in this image, but in contrast, in embryos that express Toll-2
00:27:21.00 in this stripe, in magenta, you get a strong enrichment of myosin II, specifically at the
00:27:26.00 boundary where Athea induced a difference in Toll-2 expression levels.
00:27:33.00 So, this result tells us that the Toll receptors are not only required for polarity, but they
00:27:38.00 are sufficient to induce polarity when you introduce a local discontinuity or a stripe
00:27:44.00 of Toll activity.
00:27:46.00 So, if these are the key receptors that establish cell polarity in the embryo, this raises the
00:27:53.00 question of, what are the ligands that they bind to to allow cells to sense their orientation
00:27:58.00 with respect to the overall axes of the animal?
00:28:02.00 And in the immune system, these cells, the mammalian Toll-like receptors, detect foreign
00:28:07.00 molecules in the animal that are presented on invasive pathogens.
00:28:11.00 But this is in the embryo in the absence of infection and so there aren't predicted to
00:28:16.00 be any pathogens present, and so we proposed that maybe, in this case, these receptors
00:28:21.00 might be detecting receptors expressed on other cells.
00:28:25.00 For example, if the Toll receptors interact homophilically with other receptors in the
00:28:31.00 same stripe, then interaction between these receptors might recruit Par-3 to horizontal
00:28:37.00 edges of cells.
00:28:39.00 Alternatively, if these receptors interact heterophilically, between stripes of cells
00:28:44.00 that express different receptors, then interactions between these receptors might locally activate
00:28:49.00 myosin and induce these cells to rearrange.
00:28:52.00 Now, the striped expression of these Toll receptors suggest that they act as a positional
00:28:57.00 code and we're starting to decipher this code using studies in cultured cells.
00:29:03.00 And these studies, Adam working together with Chris Fincher in the lab, expressed the Toll
00:29:07.00 receptors in Drosophila S2R+ cells to ask, can these receptors mediated adhesion?
00:29:14.00 And what they found is that cells expressing one receptor have a preferential affinity
00:29:18.00 for cells expressing a different receptor, resulting in the formation of these long chains
00:29:22.00 of cells that express alternating Toll-family receptors.
00:29:27.00 So, this type of interaction, along with some other evidence in cultured cells, is consistent
00:29:33.00 with the idea that these receptors participate in heterophilic interactions -- they interact
00:29:37.00 with different members of the same family -- suggesting that they could mediated interactions
00:29:42.00 between neighboring stripes of cells along the head-to-tail axis of the embryo.
00:29:47.00 So, what I've told you about today is that the global map that provides spatial information
00:29:53.00 that orients the movements of hundreds of cells with respect to the head-to-tail body
00:29:58.00 axis of the embryo is provided by three members of a conserved Toll receptor family that interact
00:30:04.00 heterophilically to mediate interactions between neighboring stripes of cells, and these interactions
00:30:11.00 give cells information about their orientation with respect to the body axes and help to
00:30:17.00 polarize the force-generating machinery that propels cell movements within the sheet and
00:30:22.00 leads to these local and collective cell rearrangements that drive tissue elongation.
00:30:29.00 So, these results identify a new function for Toll-family receptors in development,
00:30:36.00 in addition to their known role in detecting... in telling the difference between self and
00:30:40.00 non-self in the immune system.
00:30:42.00 These receptors can also tell different types of cells apart in the embryo in a mechanism
00:30:48.00 that gives cells spatial cues as they move to their final destinations.
00:30:54.00 And we would now like to understand how this code works -- how do cells tell the difference
00:30:59.00 between neighboring cells that might express different types of Toll receptors or even
00:31:03.00 different levels of the same receptor?
00:31:06.00 And how do these heterophilic interactions between receptors at the cell surface signal
00:31:11.00 to the force-generating machinery that drives cell movement?
00:31:14.00 In particular, I mentioned that the major output of Toll receptor signaling in the immune
00:31:19.00 system is the activation of NF-κB-dependent transcription of genes involved in the immune
00:31:25.00 response in the nucleus of cells.
00:31:27.00 But the rapid timescale of the axis elongation, combined with these striped patterns and predicted
00:31:32.00 localized interactions between Toll receptors raises the possibility that these receptors
00:31:36.00 might have a signaling output that doesn't go through the nucleus, but could potentially
00:31:42.00 act in a fast-acting way at the cortex of cells to result in a rapid execution of spatially
00:31:50.00 guided cell movements.
00:31:51.00 And a better understanding of these novel functions of Toll-family receptors in Drosophila
00:31:56.00 can provide insight into the mechanisms by which Toll-like receptors regulate processes
00:32:02.00 such as tissue regeneration and wound-healing in mammals.
00:32:05.00 Now, finally, I just wanted to acknowledge the amazing group of people in my lab who
00:32:11.00 did the work that I told you about today.
00:32:13.00 And they've made many important discoveries about how cells assemble to form tissues.
00:32:20.00 And I'd like to thank the people whose work I had time to talk about today, as well as
00:32:24.00 the people whose work I didn't have time to talk about today.
00:32:27.00 And thank you very much for listening.

This Talk

Speaker: Jennifer Zallen

Audience:

Educators of Adv. Undergrad / Grad
Researcher
Educators

Recorded: August 2016

Talk Overview

A major challenge in developmental biology is to understand how cells move and interact to form complex multicellular structures. Dr. Zallen’s laboratory uses multidisciplinary approaches to study how tissue architecture is dynamically established and remodeled during development. This talk describes the discovery of a system of spatial information that organizes cell movements in the developing fruit fly embryo to elongate the body axis from head to tail. Interestingly, the global positional code is provided by an ancient family of Toll-related receptors that are widely used for pathogen recognition in the innate immune system. This research elucidates general principles that link molecular signals to the physical forces and collective cell behaviors that establish tissue structure, and may help to reveal how errors in these processes contribute to birth defects and cancer in humans.

Speaker Bio

Jennifer Zallen

Jennifer Zallen

Dr. Zallen is a Member and Professor at Sloan Kettering Institute and an Investigator of the Howard Hughes Medical Institute. She received her A.B. in Biology from Harvard University and her Ph.D. with Dr. Cori Bargmann at the University of California, San Francisco. She did her postdoctoral research with Dr. Eric Wieschaus at Princeton University… Continue Reading

Leave a Reply Cancel reply