Cell Adhesion, Signaling and Cancer

Transcript of Part 2: Discovery and Characterization of a Focal Adhesion Protein Implicated in Tumor Progression

00:00:03.26	Hello, my name is Mary Beckerle,
00:00:05.24	and I'm a professor of Biology and Oncological Sciences at
00:00:08.26	Huntsman Cancer Institute at the University of Utah.
00:00:11.04	In this part of my presentation, I want
00:00:15.13	to describe to you some work in my laboratory
00:00:19.07	on the discovery and characterization of a focal adhesion
00:00:22.22	protein that is implicated in tumor progression.
00:00:28.01	As we've discussed in part one, cells are surrounded by
00:00:34.19	other cells and extracellular matrix
00:00:37.05	material and lots of soluble cues and factors that influence
00:00:42.01	their behavior in really dramatic and important ways.
00:00:45.13	And one really important type of informational cue comes from the information
00:00:55.10	that cells get from interacting with the extracellular matrix.
00:01:00.01	Extracellular matrix interaction occurs at specialized zones of the cell surface,
00:01:09.12	which in cultured cells are called focal adhesions.
00:01:12.11	These focal adhesions, as we discussed last time,
00:01:16.28	are areas that are rich in integrin adhesion receptors,
00:01:22.17	and they are playing a really important role
00:01:26.23	in bidirectional transmembrane communication.
00:01:32.21	The focal adhesions sense extracellular matrix,
00:01:36.12	but they also are involved in sensing cues that are present on cell surfaces.
00:01:41.03	And they also, as we will discuss in part three,
00:01:44.12	can sense physical stress such as mechanical stimulation.
00:01:51.22	The responses that cells launch as a result of the signaling through
00:01:58.20	these specialized regions of the membrane are really
00:02:01.13	very diverse and include control of cell growth or cell death,
00:02:05.26	cell motility, cytoskeletal organization,
00:02:09.27	and these signals can even result in changes in
00:02:14.05	gene expression, illustrating that there must be some mechanism by
00:02:18.01	which the molecules present at these adhesion sites
00:02:22.09	are able to communicate with the nucleus.
00:02:25.06	Integrins are the primary receptor for extracellular matrix
00:02:30.22	that are located at these specialized adhesion zones.
00:02:36.11	They are concentrated at these focal adhesions.
00:02:39.11	They mediate transmembrane bidirectional
00:02:41.29	signaling, and a really interesting challenge has been to try and understand
00:02:47.01	how they signal to effect so many really important cell behaviors.
00:02:52.08	Because unlike growth factor receptors,
00:02:54.22	they themselves don't have any intrinsic catalytic activity.
00:02:58.18	Rather they seem to operate by recruitment of a large collection of
00:03:04.11	cytoplasmic proteins to the cytoplasmic face
00:03:11.04	of the plasma membrane, and it is these proteins
00:03:13.27	which really facilitate integrin signaling function.
00:03:18.02	There are over 50 components that are present
00:03:22.04	with integrins at these focal adhesions,
00:03:25.28	and these are highly dynamic structures where proteins
00:03:30.01	are coming and going. There are both structural components
00:03:33.23	and catalytic constituents. Here you can see an example
00:03:37.09	of the evidence of one catalytic constituent, and that
00:03:42.03	is tyrosine kinases and their substrates
00:03:45.28	because here we are labeling the cell with an antibody
00:03:48.26	directed against phosphotyrosine.
00:03:50.29	And you can see that these bright patches here
00:03:55.25	where phosphotyrosine is accumulated represent
00:03:58.26	the focal adhesions where integrins would also be concentrated
00:04:03.10	and where you can see there is a very close connection with the ends
00:04:07.22	of the actin filaments, the stress fibers that terminate at these sites.
00:04:13.14	One protein that is one of the many proteins present at these focal adhesions
00:04:20.11	is a protein that was discovered in my laboratory,
00:04:23.02	several years ago, called zyxin.
00:04:25.15	This was a really interesting discovery process, I think,
00:04:31.03	and one that was very serendipitous.
00:04:33.01	In fact, I was making a lot of different antibodies against focal adhesion
00:04:38.05	constituents. At that time we only knew of three focal adhesion constituents,
00:04:42.15	and I was trying to look for new ones.
00:04:44.02	And I was making antibodies against some candidate proteins
00:04:47.24	and one of the rabbits, prior to immunization
00:04:50.12	actually was producing an antibody that already stained
00:04:55.04	focal adhesions of cultured cells.
00:04:56.22	And so I followed what this antibody was recognizing,
00:05:02.07	and it turned out it was recognizing a protein that was
00:05:05.04	previously not appreciated as a focal adhesion constituent.
00:05:07.29	And this protein now has the name zyxin,
00:05:12.04	which comes from the Greek root zeugos, which means to join together,
00:05:17.14	because it is at places where actin filaments are joined to the plasma membrane.
00:05:20.21	This is a very interesting protein that had a number of
00:05:25.20	unusual structural features: highly enriched in proline residues
00:05:30.12	and it also had 3 C-terminal double zinc finger motifs called
00:05:36.16	the LIM domain. The LIM domain was first identified in three transcription factors
00:05:42.25	Lin-11, Isl-1 and Mec-3. And these are all homeodomain proteins
00:05:49.11	that contained two N-terminal LIM domains.
00:05:54.14	In collaboration with Mike Summers and Dennis Winge
00:05:57.17	we solved the structure of the LIM domain,
00:05:59.19	and we determined that it is really a double zinc finger structure
00:06:04.10	with two sort of independent modules, each of which binds a zinc atom.
00:06:10.24	And surprisingly, although many zinc fingers are involved in direct
00:06:17.24	nucleic acid binding, it's quite clear from our analysis that
00:06:22.13	LIM domains serve as protein binding interfaces.
00:06:26.03	However, I should point out that this structure of
00:06:30.09	the LIM domain is very, very similar to the GATA-1 DNA binding domain,
00:06:35.09	so there remains a possibility that in addition to protein-protein interaction,
00:06:39.17	LIM domains may be able to associate with nucleic acids,
00:06:43.05	although this has not been demonstrated directly.
00:06:46.28	There are now a number of LIM domain proteins that have been identified-
00:06:52.12	more than 50 human proteins, and they are
00:06:57.04	functionally quite diverse. Some of them have
00:07:00.00	catalytic features, such as serving as kinases,
00:07:04.00	others are present exclusively in the nucleus and are known
00:07:08.01	to play a role in regulation of gene expression
00:07:11.28	and many others are found in association with the cytoplasm.
00:07:16.20	And what I'll tell you today is that this protein zyxin,
00:07:20.11	seems to be a multifunctional protein that resides
00:07:24.00	prominently at the focal adhesions, but also
00:07:26.27	has... where it seems to regulate cell motility
00:07:30.10	and perhaps even cell death, but it also has
00:07:36.05	the capacity to move into the nucleus, so it is an amazingly interesting protein
00:07:41.02	that is a candidate for participating in direct communication
00:07:44.21	between the cell surface and the nuclear compartment.
00:07:48.28	Zyxin also, we now have come to appreciate,
00:07:55.04	may be implicated in a particular type of cancer: Ewing's sarcoma.
00:08:00.11	I want to tell you a little bit about Ewing's sarcoma, and then
00:08:04.22	about the connections between this focal adhesion protein
00:08:08.28	and its properties and the progression
00:08:11.09	of Ewing's sarcoma that we now understand from pre-clinical model studies.
00:08:17.16	Ewing's sarcoma is a very devastating childhood and young adolescent
00:08:24.26	tumor.  It's a small, round, blue cell tumor
00:08:28.22	that you can see histologically is very homogeneous
00:08:33.18	in terms of the way it appears at the cellular level.
00:08:37.23	It's a tumor of the bone, extremely rare, but very, very aggressive.
00:08:43.24	And although it is rare, because it impacts children and young people
00:08:48.13	there's been a fair amount of emphasis on trying to understand its cause
00:08:54.08	and develop better treatment strategies.
00:08:56.21	Right now, we are faced with the challenge that metastatic
00:09:01.01	Ewing's sarcoma is really a lethal disease.
00:09:05.12	Here you can see some data looking at survival
00:09:08.22	from Ewing's sarcoma if the disease was localized,
00:09:15.23	where there's a really good, greater than 80% survival rate,
00:09:20.13	versus if the tumor has spread.
00:09:25.23	And many, many cases of Ewing's sarcoma have micro-metastases
00:09:28.25	at the time of initial detection, and
00:09:31.02	you can see that once one has metastatic disease
00:09:35.04	there is a very, very poor prognosis.
00:09:38.00	Even more strikingly, I think is the fact that we really don't have effective therapies
00:09:45.06	to treat this disease.  And here you can see
00:09:47.11	the outcomes for patients after relapse from their initial
00:09:54.00	disease, and those treated with the most aggressive therapy really
00:10:00.18	ultimately do not do much better than those that receive palliative care.
00:10:05.05	So unfortunately this is although a rare tumor, an extremely aggressive and
00:10:11.08	difficult to treat tumor.  We understand now that
00:10:16.10	Ewing's sarcoma is caused by a reciprocal chromosome translocation,
00:10:21.00	and 11;22 translocation that results in the productive fusion of
00:10:26.13	a region of the EWS gene with a region of the FLI1 gene.
00:10:32.17	And here you can see one gets out of that fusion a chimeric protein,
00:10:39.00	the EWS-FLI1 chimera, which brings together a transactivation domain from EWS with
00:10:46.05	a DNA binding domain from FLI1.
00:10:50.17	This protein is believed to function as an aberrant
00:10:54.20	transcription factor that we know is responsible for
00:11:00.19	causation of Ewing's sarcoma.
00:11:02.21	In the lab we can demonstrate that EWS-FLI1 expression is sufficient to behave
00:11:11.17	as an oncoprotein, and it's sufficient to transform cells.
00:11:14.26	Here are on the left you can see cells grown in soft agar
00:11:19.10	when they are expressing FLI1, and you can't really see any colonies developing here.
00:11:23.26	But when those cells are expressing the EWS-FLI1 fusion protein,
00:11:30.27	you can see many colonies growing
00:11:34.16	in this soft agar cloning assay.
00:11:37.19	So these cells display anchorage independent cell growth
00:11:41.27	and behave as transformed cells by this classical assay.
00:11:47.17	So a number of groups have been really interested in trying to understand
00:11:52.05	what happens within cells when you express EWS-FLI1,
00:11:56.13	and there's been a recent report by Amsellem and colleagues
00:12:01.26	in which they explore what is the difference between gene expression
00:12:07.18	patterns in 3T3 cells and in 3T3 cells that are programmed to express the EWS-FLI1
00:12:15.14	oncoprotein. And their analysis identified several genes which were
00:12:22.14	misregulated or displayed altered expression patterns
00:12:26.21	when EWS-FLI1 was present. And one of these was
00:12:32.10	the focal adhesion protein zyxin that I mentioned
00:12:35.17	was discovered by my laboratory.
00:12:38.10	You can see here if we look at the level of protein in a 3T3 cell control
00:12:44.04	using an anti zyxin antibody, you see
00:12:46.17	this much protein, and a typical 3T3 cell expressing
00:12:51.04	EWS-FLI1 shows this dramatic reduction in the level of the zyxin protein.
00:12:56.27	And here you can see this again by immunocytochemistry,
00:13:00.10	where in the control cells you can see zyxin nicely concentrated
00:13:04.18	at these focal adhesions, whereas in the
00:13:08.26	3T3 cells that have been transformed with EWS-FLI1,
00:13:13.19	there's both an altered morphology and a loss of zyxin at the focal adhesions.
00:13:20.10	Interestingly of course, when you put EWS-FLI1, a transcriptional regulator into cells,
00:13:29.03	you are expecting to have many, many dramatic changes occur
00:13:33.04	within those cells. So seeing that zyxin is down and it is
00:13:35.20	correlated with this change in morphology
00:13:37.10	is one thing, but knowing that it is responsible, and to what extent
00:13:41.24	it might be responsible for that change in morphology
00:13:43.28	is an open question that needs to be addressed directly.
00:13:48.01	And it was addressed by these investigators by taking these
00:13:52.05	3T3 cells that are transformed with EWS-FLI1,
00:13:57.14	which again show this really dramatic altered morphology
00:14:01.08	from this really well spread, fibroblast-like morphology
00:14:07.00	to a much more apparently motile and non-spread morphology.
00:14:17.06	And what they were able to do was to ask, okay,
00:14:19.13	in these cells, if we put back a wildtype zyxin,
00:14:23.11	what happens? And remarkably you can see this really dramatic phenotypic
00:14:29.13	reversion where the simple re-expression of this one gene, the zyxin
00:14:34.21	gene, causes a phenotypic reversion to something that looks
00:14:38.03	much more like the untransformed cells.
00:14:41.11	So this provided an indication that the zyxin protein
00:14:45.22	was not only reduced in expression, and in response to EWS-FLI1
00:14:51.20	transformation, but also might be responsible for some of the
00:14:56.13	transformed phenotypes that is associated with EWS-FLI1 expression.
00:15:02.23	So my laboratory has been interested in trying to understand what is the function of zyxin.
00:15:08.07	And we took an approach of generating a targeted disruption
00:15:12.20	of the zyxin gene in order to completely eliminate all expression of zyxin
00:15:17.27	and then be able to study the consequences of loss of zyxin
00:15:22.01	expression for cell function.
00:15:24.19	And here you can see that we have successfully generated cells
00:15:29.21	that lack zyxin.  Here's wildtype cells, mouse embryo fibroblasts,
00:15:34.29	expressing zyxin.  You can see nice focal adhesions
00:15:38.01	staining. And over here, you can see... you can't see. There is a cell here.
00:15:42.24	And it has been labeled with anti-zyxin antibody and there is no zyxin
00:15:46.28	protein. And this is confirmed by the Western blot here showing that
00:15:52.03	the targeted gene disruption was effective
00:15:55.04	and completely eliminates the zyxin protein.
00:15:58.06	We’ve now begun to characterize those cells
00:16:03.01	and trying to think about how the changes in those cells
00:16:06.20	might be relevant for this Ewing's sarcoma
00:16:11.16	phenotype that was observed by Amsellem and colleagues.
00:16:14.27	First thing that we looked at since we knew that zyxin was present at these
00:16:20.13	focal adhesions where actin filaments are tethered to the plasma membrane
00:16:24.05	is whether the actin cytoskeleton was disturbed in the zyxin null cells.
00:16:29.23	And here you can see that although the zyxin null cells
00:16:35.09	display actin stress fibers-so here are control cells
00:16:39.22	and then null cells-both display actin stress fibers, so we can't
00:16:44.15	conclude that zyxin is absolutely essential
00:16:47.24	for the establishment of these actin arrays in cells.
00:16:51.14	But if we perturb this system and challenge it with a Rho kinase inhibitor,
00:16:57.03	and remember now Rho signaling is important for building
00:17:02.07	of robust stress fiber arrays, and Rho acts in part through
00:17:06.21	the activity of the Rho kinase. So if we perturb Rho kinase signaling
00:17:13.12	with a Rho kinase inhibitor- if we treat wildtype cells with a Rho kinase inhibitor for short
00:17:20.06	periods of time under conditions where we don't see dramatic reduction in the stress fibers.
00:17:25.21	If we look at null cells, we see that the stress fibers are completely eliminated
00:17:31.22	when zyxin isn't there. And this would be
00:17:34.08	the ultimate outcome for the wildtype cells too
00:17:38.06	if we continued this treatment for a longer period of time.
00:17:41.11	So these results illustrated that although the actin stress fiber arrays are
00:17:49.10	present in both wildtype and zyxin null cells
00:17:52.14	that the arrays that are present in the zyxin null cells are less robust
00:17:57.02	and are more sensitive when stress fiber pathways are compromised.
00:18:03.10	And this is actually very interesting because there is a long history illustrating that
00:18:11.18	stress fiber content is inversely related to motility.  That is that
00:18:15.13	the bigger and more robust the stress fibers are
00:18:19.05	the less motile the cells are.  So you can imagine then that
00:18:23.19	if the zyxin null cells have reduced stress fiber content
00:18:28.29	or less robust stress fiber content
00:18:31.12	that one might anticipate that the cells would show enhanced motility.
00:18:36.05	And indeed that is exactly what we see.  We've looked at
00:18:39.24	the zyxin null cells and evaluated their motility properties,
00:18:43.22	in a number of different ways. And here is just one experiment
00:18:47.12	where we looked in a monolayer wound assay.
00:18:51.17	We took wildtype and zyxin null cells, grew them
00:18:56.17	in a monolayer, and then scratched a wound in that monolayer,
00:19:01.16	measured that wound gap, and then monitored the recovery of that wound as the cells
00:19:09.29	migrated into the cleared area.
00:19:12.13	And this is just an endpoint analysis shown here,
00:19:16.11	but after eight hours you can see that the zyxin null...the wildtype
00:19:19.28	cells still have a pretty substantial gap in this wound,
00:19:23.05	whereas the null cells show complete recovery of the wound.
00:19:29.09	And a blind analysis of a large number of wounding assays
00:19:34.15	and quantitation of the velocity of cell migration revealed that
00:19:38.04	the zyxin null cells actually migrate about twice as fast as the wildtype cells,
00:19:46.26	consistent with the idea that the loss of zyxin could be involved in tumor progression
00:19:55.08	and motility. Zyxin null cells also seem to display migratory properties that are
00:20:02.11	uncoupled from extracellular matrix cues.
00:20:05.05	We talked a lot in the first part about how cells normally get
00:20:09.17	substantial signals from their environment and how that has an impact on
00:20:14.16	the signaling output and the behavior of the cells.
00:20:17.14	And in this experiment in which we've examined and compared
00:20:22.21	wildtype cells and zyxin null cells
00:20:24.08	for their migration properties in a Boyden chamber, a transwell migration assay,
00:20:30.27	we have some indication that the zyxin null cells are migrating
00:20:35.26	at sort of maximum speed regardless of any extracellular matrix cues.
00:20:41.23	And this is a really striking phenotype.
00:20:44.27	Here you can see in the open bars the wildtype
00:20:47.28	cells and what we've done here is to put cells on the top of this Boyden chamber
00:20:53.26	and then put different amounts of extracellular matrix protein, in this case fibronectin,
00:21:00.24	on the underside of that filter and allowed the cells to migrate through.
00:21:04.20	And what you can see is that wildtype cells don't migrate very well at all
00:21:10.00	when there is no fibronectin present, no sort of signal,
00:21:13.09	nothing to grab onto on the other side of that filter.
00:21:17.19	And as the fibronectin level increases, you can see that the
00:21:22.18	wildtype cells begin to migrate effectively
00:21:26.11	through the pores and onto the other side of the filter.
00:21:29.29	Interestingly the null cells in comparison, start out, even
00:21:34.25	with no extracellular matrix cues, migrating at maximal velocity
00:21:39.28	and seem unaffected by increasing fibronectin concentration.
00:21:45.25	So what this suggests is again consistently showing us
00:21:50.03	that the zyxin null cells are migrating more effectively than wildtype cells.
00:21:54.02	But also suggests that in the absence of zyxin
00:21:57.25	the cells are sort of almost primed for migration.
00:22:01.04	They appear to be behaving as if they have a signal
00:22:04.19	that there was matrix present when there really isn't any matrix present.
00:22:08.13	Now going back to the Ewing's sarcoma model system then,
00:22:13.13	what is the impact of the expression of EWS-FLI1 oncoprotein
00:22:21.05	in 3T3 cells vis-Ã -vis migration?
00:22:24.14	And does restoration of zyxin expression also suppress a migratory phenotype?
00:22:33.01	So here you can see work again from Amsellem and colleagues
00:22:37.05	where we are looking at migration of normal 3T3 cells
00:22:42.09	following sort of tracks of cells and measuring
00:22:46.08	net displacement. And here you can see when we put in
00:22:51.04	the EWS-FLI1 transgene into those cells in the blue bars,
00:22:56.23	we see a dramatic increase in motility,
00:23:00.27	consistent with what you'd expect for a transformed cell.
00:23:05.06	And interestingly... and we see down here
00:23:10.07	that zyxin levels have fallen there as I showed you before.
00:23:15.18	Now interestingly when you put zyxin back into these cells
00:23:19.10	what are the consequences for cell motility?
00:23:21.06	Well, you see a restoration of the more slow, wildtype motility.
00:23:27.06	So once again it appears that the changes in zyxin levels
00:23:32.27	are contributing very substantially to the motile phenotype of these cells.
00:23:38.06	And I think that this...some insight into how this
00:23:42.08	occurs may come from work of Yu-li Wang's lab, where he followed individual cells
00:23:50.07	that were plated on a polyacrylamide substratum
00:23:55.05	into which he embedded fluorescent latex beads
00:23:57.29	so he could really look at how much force
00:24:00.19	is generated at various adhesion points,
00:24:04.08	tethering points between the cell and the substratum.
00:24:10.23	And he noticed a very interesting thing, and that is that when you
00:24:17.03	compare the traction stress maps, where they are really measuring
00:24:20.21	the deflection of these beads with the location and intensity
00:24:24.28	of expression and localization of a GFP-zyxin,
00:24:29.24	what they observed is that where you have really, really high traction forces
00:24:35.11	at this leading edge here shown in this deeper red color,
00:24:38.00	you see that the area is really devoid of zyxin.
00:24:42.01	So, and then you can follow over time a single, the development of a single
00:24:48.18	focal adhesion and see that this focal adhesion starts out
00:24:54.05	with no zyxin and then the zyxin level goes up, up, up
00:24:58.14	and then plateaus. And strikingly, as the zyxin levels are going up here
00:25:03.13	the traction stress is going down.
00:25:06.20	So clearly there is an inverse relationship between
00:25:09.15	the presence of zyxin at these focal adhesions
00:25:12.21	and traction stress, and this could be a very interesting functional explanation
00:25:19.11	for why the cells that lack zyxin are more migratory,
00:25:24.25	that is they have the capacity to generate a greater or more sustained
00:25:29.05	traction stress.  And that is something that we're all very interested in looking at
00:25:33.01	in more detail in the future.
00:25:35.00	So in summary of this part of the presentation then,
00:25:38.03	I described the Ewing's sarcoma as a lethal, devastating disease
00:25:46.02	of childhood, fortunately rare which is caused by
00:25:51.14	a reciprocal translocation that gives rise to the EWS-FLI1 oncoprotein.
00:25:57.05	EWS-FLI1 expression in a model Ewing's sarcoma system
00:26:03.18	results in decreased zyxin, and our knockout studies
00:26:08.22	showed that actin stress fibers are less robust in zyxin null cells,
00:26:13.11	and that decreased zyxin is associated with enhanced cell motility.
00:26:17.23	So how is this enhanced motility and impact on the actin cytoskeleton achieved?
00:26:26.02	Well, we learned something about this actually
00:26:29.19	by a sort of surprising link by studying Listeria,
00:26:35.02	an intracellular pathogen that is really the subject of Julie Theriot's
00:26:39.12	iBioSeminar presentation, so you can learn much more about
00:26:43.26	Listeria by looking at her presentation.
00:26:46.00	But, studies of Listeria provided some important insight into zyxin function
00:26:53.20	and its mechanism of action.
00:26:54.23	Just briefly, Listeria is an intracellular bacterium
00:27:00.08	that invades its host and is taken up by phagocytosis.
00:27:04.12	It's a foodborne bacterium that can cause really
00:27:09.29	devastating disease, particularly in immunocompromised individuals.
00:27:13.25	But the cell biology of this organism is that it
00:27:16.25	gets into these phagosomes and manages to
00:27:20.22	escape from this phagosomal compartment
00:27:23.18	and gets released into the cytoplasm where it can replicate. And for the
00:27:27.08	purposes of our discussion, it harnesses
00:27:29.17	the host cell machinery for actin assembly,
00:27:32.02	growing this nice long actin comet tail,
00:27:36.23	which allows it to move within the host cell and actually generate these
00:27:43.07	microspike projections which enable it to invade neighboring cells.
00:27:47.21	And in this way, it completely is able to transmit itself from one
00:27:51.14	cell to the next without ever leaving the protection of the intracellular
00:27:56.25	environment. And from a variety of investigations
00:28:02.02	we now understand that there is a single protein present on the Listeria surface,
00:28:07.09	the ActA protein, which is both necessary and sufficient for pathogenicity of this bacterium.
00:28:13.22	And it's necessary and sufficient for the ability of these bacteria to generate
00:28:20.11	these actin comet tails and thus move within the host cell and be transmitted
00:28:26.03	from one cell to the next.  The ActA protein
00:28:31.00	is a protein that is anchored in the bacterial surface via a membrane anchor
00:28:36.18	and then has a long protein sequence area which
00:28:43.07	is present in the host cytoplasm.
00:28:45.29	And genetic analysis via mutational studies and deletion studies
00:28:51.03	have shown that there are two really critical functional domains in ActA,
00:28:55.18	which are important for its ability to cooperate with host cell machinery to build
00:29:02.09	this actin comet tail on the surface of the bacterium.
00:29:06.02	One is the domain here which is critical for the nucleation of actin assembly.
00:29:09.21	And the other is a series of proline repeats here which are critical for acceleration
00:29:15.04	of actin assembly on the bacterial surface.
00:29:18.06	And so, in collaboration with Daniel Louvard's lab, we made a bunch of
00:29:23.09	antibodies against Listeria ActA with the idea of trying to look for what endogenous cellular
00:29:29.02	proteins ActA might be mimicking on the surface of the bacterium, in order to
00:29:34.24	recruit host machinery and stimulate the elaboration of this actin comet tail.
00:29:39.25	And when we did this we had the surprising result
00:29:45.18	that antibodies against Listeria ActA recognized
00:29:49.03	just a single eukaryotic protein, a single human protein,
00:29:53.27	at least in our experiments. And that protein turned out to be zyxin.
00:29:57.27	And interestingly, it turns out that
00:30:01.01	this really important proline rich domain in ActA,
00:30:04.01	which is critical for acceleration of actin assembly,
00:30:06.15	is almost completely conserved in zyxin,
00:30:10.10	giving rise to the idea that zyxin might be able to stimulate actin assembly.
00:30:16.07	And indeed zyxin shares with ActA the ability
00:30:20.07	to interact with proteins in the Ena/VASP family,
00:30:23.08	which are critical regulators of actin assembly
00:30:27.19	that recruit profilin and actin and by multiple mechanisms
00:30:32.26	enhance actin assembly. So, returning back to this Ewing's sarcoma model then,
00:30:41.13	we see that in model Ewing's sarcoma cells that are expressing the EWS-FLI1
00:30:47.03	transgene, that's associated with loss of zyxin,
00:30:54.14	decreased actin stress fibers, which are associated with loss of zyxin,
00:30:58.15	as is increased motility. And so, I think it's
00:31:03.16	going to be extremely interesting to explore further
00:31:06.21	what is the role of zyxin in human patient
00:31:10.10	samples, and evaluate whether or not loss of zyxin
00:31:14.23	is correlated with more aggressive disease.
00:31:18.05	Interestingly, in the original Amsellem experiments
00:31:23.10	they also, in addition to looking at motility,
00:31:26.02	looked at whether zyxin has an impact on anchorage independent
00:31:31.25	cell growth of this EWS-FLI1 transformed fibroblasts.
00:31:35.19	And they found that it had really quite a dramatic impact.
00:31:40.11	So here you can see again normal 3T3 cells
00:31:43.26	that are not transformed don't grow in soft agar,
00:31:49.06	and so you don't have any colonies.
00:31:51.21	If you express, if you transform those cells with a control
00:31:57.19	vector, you don't see any change.
00:32:00.15	If you express more zyxin in those cells you don't see any change.
00:32:04.08	But look, as we saw in the original soft agar cloning experiment that I showed you
00:32:09.11	earlier, when you express EWS-FLI1 in those 3T3 cells, you see
00:32:15.12	a dramatic increase in the ability of those cells to grow in
00:32:19.14	soft agar.  And you can suppress that ability by re-introducing zyxin into those
00:32:25.19	cells. So this was, in addition to the motility
00:32:29.18	phenotype, I think a very striking piece of data that
00:32:33.28	suggested that zyxin might also be influencing
00:32:37.21	either cell proliferation or apoptosis and contributing to this suppression
00:32:45.07	of anchorage independent cell growth.
00:32:48.15	Interestingly, I told you that it is very, very difficult to
00:32:54.11	treat Ewing's sarcoma, so there are many
00:32:56.11	many investigators really working hard to identify new strategies
00:33:01.17	for therapeutic intervention with this disease.
00:33:05.26	And one interesting new concept that several groups are working on
00:33:10.29	is the use of an antibody directed against a cell
00:33:15.21	surface component, CD99, which is highly expressed
00:33:19.29	on Ewing's sarcoma cells, human tumor cells.
00:33:24.21	And it turns out that if you take this anti CD99 antibody and
00:33:29.14	expose Ewing's sarcoma cells to this antibody, the clustering of that
00:33:36.10	cell surface marker somehow induces apoptosis of the Ewing's sarcoma cells
00:33:43.03	by a mechanism that at this point in time really isn't understood.
00:33:46.20	I mean we really don't understand yet what CD99's normal function is.
00:33:50.08	But because it is so highly enriched on Ewing's sarcoma cells
00:33:53.19	and because engagement of CD99 causes apoptosis, it's
00:33:58.25	really being thought of as a potential therapeutic intervention for Ewing's sarcoma.
00:34:05.10	And interestingly, in the course of trying to understand
00:34:09.18	how clustering and engagement of anti-CD99
00:34:13.23	with anti-CD99 antibodies causes apoptosis.
00:34:18.14	investigators found that zyxin is a critical componentin that apoptotic pathway.
00:34:27.21	And here you can see some flow cytometry analysis looking at
00:34:33.03	the ratio of viable cells to apoptotic or necrotic cells
00:34:36.15	before or after anti-CD99 treatment, and
00:34:44.14	with normal zyxin levels or suppressed zyxin levels.
00:34:48.18	And here you can see that in the normal cell population
00:34:53.06	76% of the cells are in the viable channel here.
00:35:00.01	And if you engage those cells with the anti-CD99 monoclonal antibody,
00:35:08.24	which is called 0662, you can see a shift from the viable channel to
00:35:14.26	the apoptotic channel consistent with the idea that
00:35:19.15	the monoclonal antibody can induce apoptotic signaling
00:35:22.23	in these cells. Now that doesn't shift at all when
00:35:29.04	you use a control antisense construct,
00:35:33.12	but if you use a construct that is specifically designed to knock down zyxin levels,
00:35:38.04	what you see is that you have a reduction in the apoptotic response
00:35:46.16	suggesting that zyxin is at least partially involved or partially required
00:35:52.00	for the ability of anti-CD99 to induce apoptosis.
00:35:56.09	And so, this leads to an interesting addition to the
00:36:04.03	prior thinking about zyxin function that in addition to
00:36:08.17	contributing to regulation of motility,
00:36:12.28	that it may actually be a pro-apoptotic factor under some circumstances
00:36:17.29	and loss of zyxin might contribute to increased survival.
00:36:21.16	So this is quite interesting in the sense that
00:36:24.06	we know that expression of EWS-FLI1 causes reduction in zyxin expression in this
00:36:30.26	3T3 model, and that would be then associated with
00:36:34.06	increased motility and increased survival based on the cell based studies.
00:36:39.26	And so I think a very important future goal is to really now
00:36:44.07	go and test whether zyxin expression level is an indicator of tumor stage
00:36:50.23	and in addition, whether or not it will predict therapeutic response.
00:36:56.18	For example if zyxin levels are extremely low,
00:36:59.22	that might give us the insight that those tumors
00:37:04.08	have greater metastatic potential because
00:37:08.07	of the increased motility associated with loss of zyxin.
00:37:10.29	Interestingly, many investigators are now starting to find zyxin down regulated
00:37:18.08	in a variety of tumors, and here's an example showing that it is down regulated
00:37:24.12	in invasive bladder cancer.  Here's the superficial tumor where the brown staining
00:37:29.18	indicates there is still robust zyxin expression.
00:37:32.11	And in this invasive carcinoma you see
00:37:35.09	very little expression of zyxin, if any at all.
00:37:38.15	And these investigators did a very extensive analysis to look
00:37:42.26	at association of a number of potential biomarkers
00:37:48.14	with tumor stage and grade and found that expression levels for zyxin,
00:37:54.02	as well as a couple of other markers, were directly associated with tumor stage and grade.
00:37:59.09	So, this could be another important and valuable
00:38:03.18	insight that allows clinicians to be able to identify tumors
00:38:08.28	and diagnose them with a higher degree of resolution using
00:38:12.28	these types of biomarkers to further refine
00:38:16.06	the grade of the tumor. Now I mentioned in the beginning that focal adhesions also are
00:38:22.11	really important for communicating with the cell nucleus,
00:38:25.24	and that these signals that emanate from focal adhesions
00:38:30.26	also influence gene expression.  And I think one of the biggest challenges
00:38:34.22	that we face as cell biologists is
00:38:38.04	knowing that that occurs, not really understanding what is the mechanism
00:38:42.13	of action by which these events that are happening at the cell surface
00:38:47.05	are communicated to the nucleus to directly regulate
00:38:51.01	gene expression.  And I am sure there are going to be a number of mechanisms
00:38:57.01	involved here, but one thing that I like to point out is that the zyxin protein
00:39:02.25	may be a candidate for communicating with the cell nucleus.
00:39:07.00	And indeed a student in my lab recognized a number of years ago
00:39:11.12	that the zyxin protein has sequence that looks very similar to
00:39:16.16	a nuclear export signal, showing these very highly ordered leucine residues
00:39:22.11	that are in similar positions to nuclear export signals
00:39:26.13	in the Rev protein of HIV1 as well as IkB.
00:39:31.17	So we looked to see whether if we deleted that leucine rich region of zyxin if it
00:39:37.16	had an impact on the subcellular distribution of zyxin,
00:39:41.10	and here you can see expression of wildtype zyxin in cells. And at steady state
00:39:46.18	you really see the zyxin concentrated in these focal adhesions,
00:39:50.03	and very little zyxin if any detectable by
00:39:52.13	immunocytochemistry within the nucleus.
00:39:54.11	But, if we delete just these 17 amino acids from zyxin, this leucine rich region,
00:40:00.06	and now ask where is that protein concentrated,
00:40:03.23	what we see is this really dramatic shift in the subcellular distribution of zyxin with
00:40:09.07	the protein migrating really substantially from these focal adhesions and accumulating
00:40:14.28	in the nuclear compartment.
00:40:17.26	So to test further whether or not this
00:40:22.06	sequence of zyxin really functions as a nuclear export signal,
00:40:27.05	we did a classic experiment where we took that zyxin sequence
00:40:31.26	and hooked it onto a carrier protein, in this case I am showing you
00:40:36.28	here GST, glutathione S-transferase, and asked
00:40:42.03	whether or not if we inject normal GST or GST tagged with this zyxin sequence
00:40:49.07	directly into cell nuclei, whether the zyxin sequence will support
00:40:54.06	the nuclear export of that GST protein.
00:40:57.27	GST is too large to diffuse through nuclear pores, so if you inject
00:41:02.06	it into the nucleus in the absence of a functional nuclear export signal,
00:41:06.00	it will be retained in the nucleus.
00:41:07.17	And that's exactly what you see with the unmodified GST.
00:41:12.10	You microinject it into the nucleus, and this is a fluorescently tagged version of GST,
00:41:17.05	put it into the nucleus, and it stays there.
00:41:20.09	But if you take just, in this case, amino acids 319 to 335 of zyxin
00:41:26.07	and attach that to the GST through genetic engineering,
00:41:31.05	express that protein, label it, and inject that protein into the nucleus,
00:41:34.29	you see that very, very rapidly that protein winds up in the cytoplasm,
00:41:41.09	illustrating that that short sequence of zyxin has the capacity
00:41:45.03	to serve as a nuclear export signal
00:41:46.26	and transport GST from the nucleus to the cytoplasm.
00:41:50.24	So, this was very exciting because it really suggested that the zyxin protein had a nuclear export signal
00:42:00.21	and perhaps was present in both the focal adhesions
00:42:04.29	and the nuclear compartment and could be part of the
00:42:07.18	machinery that really functionally linked the cell surfaceto the nucleus.
00:42:12.15	However, of course, when you are cutting pieces of proteins up
00:42:15.14	and putting them onto other proteins, things can happen that are unanticipated,
00:42:20.21	and so we really wanted to have a direct mechanism to look at whether zyxin ever
00:42:26.14	goes into the nucleus. And this was also important
00:42:28.20	because, as I pointed out, at steady state we really don't see zyxin
00:42:31.25	in the nucleus, so we developed an assay to report
00:42:36.17	whether zyxin goes into the nucleus.
00:42:39.05	And, I think, this is an experiment that David Nix in my lab
00:42:43.08	did when he was a graduate student,
00:42:44.12	and I think it is really amazing that this experiment worked.
00:42:47.25	He decided to test whether or not zyxin goes into the nucleus
00:42:53.21	by injecting a cocktail of two different antibodies
00:42:57.17	into, directly into the nuclei of individual cells.
00:43:02.07	And the two antibodies were an anti-zyxin monoclonal antibody
00:43:06.23	and a matched control antibody that didn't recognize any cellular proteins.
00:43:11.25	And once again these antibodies, which were labeled with different color fluorochromes,
00:43:18.00	are too large to diffuse out through nuclear pores.
00:43:22.00	So you would expect them to be retained in the
00:43:23.28	nucleus unless they are specifically exported
00:43:27.21	by binding to another protein partner.
00:43:31.08	So here we start out, we inject these two antibodies into cell nuclei.
00:43:36.24	Zyxin is, as I mentioned, concentrated in the cytoplasm,
00:43:39.20	and if zyxin didn't shuttle into the nucleus, you would expect
00:43:44.22	over time for the antibodies to remain trapped in the nucleus
00:43:48.02	and for the zyxin to remain in the cytoplasm.
00:43:51.23	In contrast, if there is shuttling going on and zyxin
00:43:55.22	is moving from the focal adhesions into the nucleus and back
00:43:59.26	out again, then we might expect one of two things.
00:44:03.08	We might expect zyxin to itself become trapped in the nucleus
00:44:09.09	over time along with its antibody partner.
00:44:13.05	Or, alternatively, we might imagine that zyxin could migrate into the nucleus,
00:44:18.17	bind the specific anti-zyxin antibody,
00:44:21.22	and because the zyxin harbors a nuclear export signal,
00:44:25.17	it could actually extract that zyxin partner from the nucleus
00:44:31.25	and facilitate its nuclear export.  And surprisingly,
00:44:35.16	this is exactly what we saw.
00:44:38.17	So here you can see a cell into which we have injected
00:44:43.07	control antibody in A and anti-zyxin antibody together into the nucleus of the cell
00:44:49.24	in B. And this is a very early time point after injection, and you can see
00:44:54.06	that both antibodies are prominently concentrated within
00:44:58.07	the nucleus of the cell.
00:45:00.08	And if we wait over time, and then look, what we see (and this is a different cell, of course)
00:45:05.26	the control antibody marks the site of injection. It is retained in the nucleus.
00:45:11.28	But look over here, the anti-zyxin antibody which was co-injected into the nucleus of the cell
00:45:18.18	is completely missing from this nucleus
00:45:21.05	and has now begun to populate these focal adhesions.
00:45:25.17	So this was a very, very striking result
00:45:28.13	which illustrated to us that protein that was in the cytoplasm,
00:45:32.06	zyxin protein in the cytoplasm could move into the nucleus.
00:45:35.08	And that protein once dwelling in the nucleus
00:45:38.22	could come out of the nucleus and go back to the focal adhesions,
00:45:42.15	really, clearly suggesting a mechanism for communication
00:45:46.15	between these two cellular compartments.
00:45:49.17	And we can now show using leptomycin, an agent that inhibits
00:45:56.15	nuclear export, that indeedif we treat cells with leptomycin to inhibit
00:46:02.01	nuclear export that zyxin begins to accumulate in the nuclear compartment.
00:46:08.25	And if we look at large populations of cells,
00:46:10.16	we see this occurs in an asynchronous fashion
00:46:13.16	illustrating that there must be some physiological signals
00:46:16.17	that stimulate the release of zyxin from the focal adhesions
00:46:21.24	and the import of the protein into the nucleus.
00:46:25.21	And of course, a really important unanswered question
00:46:29.26	is what those signals might be.
00:46:32.26	So what could be the functional significance of nuclear shuttling of zyxin?
00:46:39.00	Well, I think there are two general possibilities. One is that the zyxin protein
00:46:43.20	could itself serve as a nuclear factor. You could imagine that the zyxin
00:46:48.15	protein is sort of retained at focal adhesions
00:46:51.14	where it's awaiting some sort of signal
00:46:55.11	or information from the extracellular environment
00:46:59.02	and under certain conditions would be released,
00:47:02.20	allowed to move into the nuclear compartment, and perhaps even
00:47:06.05	participate directly in regulating gene expression or some other nuclear function.
00:47:11.00	And this is a really intriguing idea, although to date
00:47:18.16	there's not really any direct evidence for a role for zyxin
00:47:24.14	within the nucleus, and particularly not in regulation of gene expression.
00:47:29.19	Though, I'll point out that the zyxin protein, remind you that the zyxin protein
00:47:34.16	has these LIM domains, which are found also in many transcriptional regulators,
00:47:40.22	so there remains, I think, an intriguing possibility that zyxin
00:47:44.09	could be directly influencing some nuclear activity.
00:47:47.21	The other alternative is a more general one that suggests that zyxin is, again,
00:47:55.10	sitting at these focal adhesions here, and then in response
00:48:00.18	again, to some signal, perhaps it could move into the nucleus
00:48:04.10	and recover a protein out of the nucleus,
00:48:09.13	returning it to the cytoplasm, and thus downregulating
00:48:11.28	the activity of that factor. Or alternatively
00:48:15.01	it could be a chaperone that really is designed
00:48:20.24	to retain a nuclear factor in the cytoplasm.
00:48:24.24	So for example, zyxin could be sitting at the focal adhesion
00:48:28.20	holding onto the protein that has a nuclear localization signal.
00:48:31.17	When that protein is bound to zyxin, if it ever went into the nucleus it would be carried
00:48:36.07	rapidly out because of the NES on zyxin.
00:48:38.11	But under certain conditions you could imagine that that protein would be
00:48:42.09	specifically released from zyxin, allowed to move into the nucleus
00:48:46.07	to activate some nuclear function.
00:48:48.29	So again, we now understand that this zyxin protein
00:48:52.15	clearly shuttles between the focal adhesions,
00:48:54.19	these really important signaling zones that are sensitive to extracellular cues.
00:48:59.22	It shuttles between those zones and the nuclear compartment.
00:49:04.27	And I think the next challenge is to try and identify what the
00:49:11.00	specific signals are that stimulate the release of zyxin
00:49:16.09	into the nucleus, and what its specific function there might be.
00:49:20.00	Interestingly, it is now appreciated, just in the last
00:49:24.10	several years, that many LIM domain proteins
00:49:27.19	have this very interesting sort of duality
00:49:31.14	in terms of their subcellular distribution.
00:49:33.24	All of these proteins listed here are proteins that reside both on
00:49:41.08	the actin cytoskeleton or in the focal adhesions, and also have
00:49:45.24	the capacity to move into the nucleus.
00:49:48.03	So I think there's a class of these molecules
00:49:50.25	that are going to be playing a really interesting
00:49:53.17	role in regulation of nuclear function
00:49:56.13	and in communication between the cytoskeleton and
00:50:00.14	the nucleus, real candidates for intracellular communication.
00:50:03.26	So I'll close there and just acknowledge some of the people in my lab
00:50:07.17	that participated in the work I described.
00:50:11.08	I think it's a very exciting time in which we are appreciating
00:50:16.00	much more the important roles of the proteins
00:50:19.24	that are present at these specialized adhesive zones,
00:50:24.09	and beginning to see not only how these proteins
00:50:27.14	influence cell behavior such as motility, perhaps
00:50:31.05	apoptotic signaling, perhaps communication with the nucleus,
00:50:35.22	but also are beginning to be able to put these pathways
00:50:40.24	into a broader context and appreciate how disturbance
00:50:44.17	of the functions of these molecules in these pathways
00:50:47.05	may impact on human disease.