Tip:
Highlight text to annotate it
X
So, let's move on now to consider how KSHV mechanistically is linked to KS pathogenesis.
And the beginning of wisdom here is to understand a fundamental aspect of herpesvirus biology,
which is that like all herpesviruses, KSHV has two transcriptional programs:
a latent program and a lytic program.
So, what does that mean? Latency is a state in which the viral genome persists in the nucleus
in an episomal form. There's no integration in this form of the life cycle.
So, there's a circular plasmid maintained in the nucleoplasm.
And the important thing is that gene expression from this plasmid is highly restricted.
So, the KSHV genome is 165 kilobases, it has over a hundred open reading frames,
but during latent infection, only a handful of these open reading frames are expressed.
In fact, the main open reading frames that are expressed are shown in the boxes here,
there's one other gene down here that's expressed in latency,
but fundamentally, it's a highly restricted program,
less than 5 percent of the genome is actually transcriptionally active during latency.
As a result, there's no virus production and the cell survives.
Now, in that sense, it's a cryptic infection, but it's not irreversibly cryptic.
Because the whole genome is still retained in the nucleus,
there is the possibility to re-awaken gene expression from the rest of the genome.
And when that happens, the virus enters its so-called lytic cycle.
So, the lytic cycle is the cycle that is perhaps more familiar to students
who've had microbiology, this is the phase in which all the open reading frames of the virus are expressed.
They're expressed in a developmentally and temporally regulated fashion,
immediate early genes to late early genes, DNA replication, late genes.
So, there's this sequential, or cascade of gene expression, but the net result of that is that all the viral open reading frames are expressed,
viral DNA is replicated, progeny, virions, are assembled, the cell is killed and viruses released.
Ok, so, these are the two cycles
and we're going to consider in this segment what latent infection contributes to KS development
and in the final section, what the lytic infection might contribute.
It's important to know that in KS tumors, advanced KS tumors,
most of the spindle cells, all of the infection in a KS tumor is trained on the spindle cell compartment.
The infiltrating inflammatory cells are by and large not infected.
So, it's really the endothelial cells of the spindle cell compartment that harbor the virus
and in fact, 95 percent of those cells are latently infected, or even more, sometimes.
There is lytic infection going on in the tumor, it's limited to a very small subset of spindle cells,
and we'll come back to that at the end of the lecture because I think it is important,
it's not a negligible aspect of this, and we'll come back to that.
So, because the latency program is so restricted, involving only a handful of genes,
there's been a lot, and because it's on in every tumor cell in the tumor,
people have been interested in understanding what the latency program is all about,
what are the genes that are expressed and what do they do?
So, here I'm showing a schematic of the major latency locus and you can see that there are four open reading frames here
and these include a gene down here called kaposin, that we'll have a great deal more to say about it.
And another gene over here called FLIP.
These two genes are pro-inflammatory genes that I'll be telling you a great deal more about.
There's also a cyclin D homologue expressed in the latency program.
A fully active cyclin D homologue. And then of course the protein LANA, that I mentioned earlier.
LANA is the antigen in our original immunofluorescence test for infection.
Now we know that LANA plays a very important role in the maintenance of the latent episome
and I'll be discussing that in brief later on.
A more recent discovery is that the latency program also includes a dozen pre-microRNAs
that are processed to engender 17 mature microRNAs.
These are all products of a latent transcription unit.
Ten of the 12 pre-microRNAs are found in the intron of that unit.
I won't have more to say about these microRNAs in today's talk,
they're a subject, as you might imagine, of very active investigation in our lab and others.
And suffice it to say we really don't know yet what all of these do,
but it's virtually certain they're playing an important role in either the maintenance or the elaboration of latency.
Alright, now one of the dirty little secrets of KSHV virology is that latent infection of many cells in culture
doesn't produce a very striking phenotype.
So, it turns out that once you learn the tricks of how to grow KSHV,
it's possible to infect practically any adherent cell in culture with the virus.
The default pathway is latency. And generally, if you're working with established or already immortalized cell lines,
there's no particular phenotype that's linked to that latent infection.
And when latent infection is carried out on a lot of primary cells, again, we don't typically see immortalization.
So, this is strikingly different from the situation in Epstein-Barr virus.
Some of you may know that EBV is a very powerfully immortalizing virus,
the latency program of EBV is very strongly immortalizing.
After, to this day, EBV is still the only reproducible method for immortalizing human B-cells from peripheral blood.
B-cells have a very short lifespan in culture unless you infect them with EBV,
but every time you do, an immortalized line will grow out.
So, the EBV latency program is powerfully immortalizing.
KSHV's latency program doesn't seem to be that way.
And for a while, we were quite despondent that the latency program had virtually no phenotype that we could dissect in the lab.
But that turns out not to be true.
It was initially shown by Gary Hayward and his colleagues and his colleagues at Johns Hopkins
that primary endothelial cells that are infected with KSHV do have a phenotype,
and I'm going to show you an example of that on the next slide.
This is the work of Claudia Grossmann in our lab,
but I want to emphasize the primary discovery here was by Gary Hayward at his group at Hopkins.
If you take primary endothelial cells, in this case, dermal microvascular endothelium,
and do either a mock infection, above, or a KSHV infection, below, and wait a week,
what you discover is that after a week, so, in the first few days of infection, there's a little bit of lytic replication
and spread throughout the culture.
That's ultimately extinguished and all the remaining cells are latently infected.
And what you can see is that the normal endothelial cells have a sort of cobblestone morphology,
but when they're, each of the cells is sort of more or less cuboidal,
but when the cells are infected with KSHV, within a week, of the latency program being established,
the cells undergo this very exuberant, or dramatic, spindling phenotype.
They become extremely elongated, just like the spindle cells I showed you earlier in the KS biopsy.
It's important to realize, though, that these cells are not immortalized.
In our hands, and I think in the hands of most people in the field,
these cells will hang around for a week or two in culture, and then they'll die.
We have never observed outgrowth, long-term outgrowth, of these cells in culture.
So, the phenotype is one of cell shape, a cell shape change.
That shape change is due to a massive reorganization of the actin cytoskeleton.
So, here is a phalloidin staining of uninfected and infected cells,
and you can see that the actin cytoskeleton has undergone a rearrangement into these long tubules,
or cables, of actin. But again, I want to emphasize, the cells are not immortalized.
Sometimes, they survive a little longer than their uninfected counterparts, sometimes not.
I don't know for sure whether there's a survival phenotype, but there's certainly not immortalization.
Now, one of the interesting things about having a cell culture phenotype
that is so strikingly reminiscent of one aspect of the in vivo phenotype
is that because there's very limited gene expression in latency,
we can hope to identify which viral genes are linked to this process by cloning the individual latent genes,
one by one, into retroviral vectors and using them to infect primary endothelial cells
and then waiting to see if the cells infected by these individual retroviruses containing individual KSHV genes
will undergo a shape change.
And this is an experiment that was conducted by Claudia Grossmann in my lab
in the, about 2004, 2005, and the result is very gratifying.
It shows that a single KSHV gene, that encoding vFLIP, is responsible for the spindling phenotype.
So, you can see here that the cells infected with the empty vector
or vectors containing any of the other KSHV genes, failed to change the shape of the endothelium,
but cells infected with the vFLIP retrovirus strongly undergo the spindling phenotype.
That spindling phenotype, by the way, is cell autonomous,
conditioned medium from these cells, if put on fresh endothelial cells do not cause them to spindle,
and it's due to expression of the FLIP gene.
So, let's review the NF-kappaB activation pathway.
NF-kappaB, as you may know, is a transcription, a heterodimeric transcription factor,
which exists in the ground state in the cytoplasm
and it is retained in the cytoplasm by its binding to an inhibitor called IkappaB.
So, when IkB binds to the two subunits of NFkappaB, that subunit is retained in the cytoplasm
because of masking of its nuclear import signals.
But that process is regulated. It's regulated by a kinase called IKK,
which is a heterotrimer of three subunits, alpha, beta and gamma.
IKK phosphorylates IkB and that phosphorylated form of IkB undergoes degradation in the proteasome
and the loss of IkB allows the release of the active transcription factor,
whose nuclear import signals are now available to the cellular machinery.
It can be imported into the nucleus and turn on the promoters of a large number of cellular genes.
Interestingly, the genes that NFkappaB controls are largely genes that involve pro-inflammatory signaling,
so a huge number of cytokines and chemokines.
Also, certain growth factors, VEGF being among them.
And many factors involved in cell survival and the prevention of apoptosis.
So, NFkappaB is a pro-inflammatory and anti-apoptotic, or pro-survival,
transcriptional program that involves hundreds of genes.
It turns out that FLIP, and this is not our work again, is an inhibitor of IKK, an activator of IKK.
It binds to the NEMO subunit and triggers constitutive phosphorylation of IkB
and constitutive activation of NFkappaB.
That results in a pro-inflammatory and anti-apoptotic signal cascade.
Now, how do we know that that activity of FLIP is responsible for the shape change that I just showed you?
We know that because it's possible to develop antagonists of the IKK/IkB signaling pathway.
And one such antagonist is something called the IkB super-repressor, or IkB-SR.
This is a variant, a mutant, of IkB, in which the two serines that are phosphorylated by IKK
are mutated such that the molecule cannot be phosphorylated
and therefore, if overexpressed, all the NFkappaB in the cell is retained in the cytoplasm
in a form that can never be delivered to the nucleus.
So, in order to ask whether the NFkappaB inducing activity of FLIP is responsible for the shape change,
Claudia did a simple experiment.
She transfected HUVEC cells, infected HUVEC cells with a retrovirus that expresses the IkB super-repressor
and then superinfected them with either a control virus or the FLIP virus
and asked about the morphology.
And the results were very gratifying.
In cells that did not get the IkB super-repressor, FLIP expression led to the characteristic spindling that I showed you earlier.
But in cells that did get the super-repressor, that spindling morphology change was totally ablated.
So, clearly, it's the NFkappaB induction by FLIP that is responsible for the spindling phenotype.
So, let's summarize the FLIP effects. Latent FLIP expression triggers a morphologic change
in endothelial cells that makes them look just like a KS spindle cell.
The shape change is cell autonomous and is due to the induction of NFkappaB.
Now, of course, the induction of NFkappaB in these endothelial cells not only induces a cell autonomous shape change,
whose biochemistry we're still trying to understand,
but also triggers the release of a large number of paracrine signaling molecules,
cytokines, chemokines, VEGF, et cetera.
Pro-inflammatory and pro-angiogenic molecules.
And I think the relationship of this to the biology of KS that I've mentioned earlier
is now obvious, this is one of the things that links viral infection
to the production of a pro-angiogenic and pro-inflammatory microenvironment.
But I want to emphasize that neither authentic KSHV latency nor FLIP expression itself
will immortalize cells in vitro.
They may confer survival benefit on endothelium
and it's known from the work of Ethel Cesarman and others that in primary effusion lymphoma,
the survival of the lymphoma cells is absolutely dependent on FLIP,
so there is certainly, in the B-cell compartment, FLIP is very important in extending cell survival.
We haven't seen such a dramatic effect in the endothelial cell.
So, that's just worth keeping in mind, that we don't see immortalization here,
and I'm going to come back to that in a minute.
Alright, so, FLIP is a pro-inflammatory gene embedded in the latency program
and it turns out it's not the only one.
A few years ago, we discovered a second gene in the KSHV latency program
that is also a pro-inflammatory gene.
So, this locus was discovered by Rob Sadler, who was then a post-doc in my lab in the early to mid-90's,
and Rob discovered that latency, that one of the regions that was transcriptionally active in latency
encoded a 1.4 kilobase transcript that included the K12 open reading frame, shown here,
but in fact, upstream of the K12 open reading frame, a short 60 amino-acid open reading frame,
were many, were two sets of GC-rich 23 nucleotide repeats.
So, that, these clusters of repeats were called DR1, direct repeat 1 and direct repeat 2.
And initially they were annotated by the computer as being non-coding.
Because there were no AUG codons anywhere in this region.
There were neither AUG codons nor stop codons anywhere,
so the computer said that the only thing that was being expressed as a protein was open reading frame K12,
a sixty amino acid, highly hydrophobic membrane protein.
But it turned out that that wasn't true, that Rob showed that these direct repeats
themselves could also be translated in infected cells
and the translation was initiated not at AUG codons, but at variant CUG codons,
which of course, the mammalian ribosome is capable of using as initiation codons at low efficiency.
In fact, there are several products from this locus now. The product of open reading frame K12 itself
is called, we now call kaposin A, but two other proteins can be made.
What we call kaposin B, which is the product of just translating DR2 and DR1,
followed by a stop codon, and then, if you initiate translation in a different frame,
you can make something called kaposin C, which translates the repeats of DR2 and DR1
fused in frame to the membrane segment of K12.
We call that protein kaposin C.
We don't understand much about kaposin C yet, but kaposin B has been taken up by Craig McCormick in the lab
and Craig showed that kaposin B is an adapter molecule in signal transduction.
And he showed that by exploring its interaction partners in a yeast two hybrid screen,
from which he recovered a very interesting kinase called MAP kinase associated protein kinase 2, or MK2.
So, it turns out that the interaction between kaposin B and MK2 is direct.
If we make purified kaposin B in E. coli and purified MK2 and mix them together,
in this case the kaposin B is fused to GST, and in this GST pull-down format,
we see very efficient pull-down of MK2 by the GST kaposin B protein
and this binding is referable to the DR2 segment, not the DR1 segment.
So, what is MK2? MK2, it turns out, is an effector of a signaling pathway called the p38 MAP-kinase pathway.
So, let's talk a little bit about that pathway.
The p38 MAP kinase pathway is a signaling pathway
that is devoted to sensing inflammation and other stresses in the microenvironment.
So, some of those stresses include oxidative stress or osmotic stress,
but also a bunch of cytokines in the microenvironment, like IL-1 or TNF, IL-6.
Signaling through that pathway can activate a cascade of kinases called the MAP kinases
that converge on a p38 MAP kinase in the cytoplasm.
Phosphorylation of that kinase leads to its import into the nucleus,
where it can associate with the protein of interest for us, which is MK2, the target of kaposin B.
So, MK2 is itself a kinase, it resides in the nucleus in an inactive state.
When p38 is activated it binds to it in the nucleus, phosphorylates it.
That complex has some nuclear substrates, which I won't talk about today,
but it is also very efficiency exported back into the cytoplasm,
where it can phosphorylate other cytoplasmic targets that I will talk about.
So, one thing you can see is that signaling through this pathway, including the things that activate inflammatory signals,
like IL-1 and TNF, lead to the cytoplasmic export of MK2.
And phosphorylation of cytoplasmic targets.
And an early clue that the phosphorylation of MK, that the interaction of kaposin and MK2 was real
came from this localization experiment.
If you look at cells that are transfected with kaposin B, and just ask where it resides in the ground state,
it resides in the nucleus, as does MK2, here.
And if you apply TNF or LPS, inflammatory signaling molecules that lead to MK2 exiting the nucleus,
you can see that, of course that process isn't a hundred percent efficient,
there's still nuclear MK2, but you can see this obvious cytoplasmic blush of staining
that is not present in the ground state. Well, when you co-express kaposin B,
the same thing happens, that same cytoplasmic blush occurs with LPS or TNF,
just as it does for MK2.
So, the two proteins colocalize in the nucleus and they are co-transported under these pro-inflammatory signaling conditions.
And in fact, even more interesting, because when you examine the sequence of kaposin B,
it doesn't appear to have any classical nuclear localization signals
for import or export from the nucleus.
And we believe that it's shuttling is being governed by its interaction with MK2.
Now, how does kaposin B affect MK2's kinase activity?
In principle, it could either be an inhibitor or an activator,
but when we do an immune complex kinase assay, that is to say we transfect cells with kaposin B,
and now immunoprecipitate MAP-kinase associated protein kinase 2, and give it a substrate,
in this case, HSP27, we can see that cells that are expressing kaposin B have enhanced MK2 kinase activity
compared to the controls that don't express kaposin B.
So, kaposin B's interaction is an activating event, kaposin B activates MK2 kinase activity.
So, what does that mean? Well, I haven't told you yet what MK2's cytoplasmic kinase activity is for.
It's probably for many things, but one thing that it is clearly for is to govern the half-lives of cytokine messenger RNAs.
So, some of you may know that cytokine messenger RNAs were discovered back in the 80's
by Bob Kamen to have AU-rich elements in their 3 prime UTRs
that make their messages extremely unstable.
And for many years it's been known that certain growth factors, like Myc, and oncogenes,
but also many cytokines, have these AU-rich elements, or AREs, in their 3' UTR
and these confer great instability on the messages so that at the ground state, they're present at very low levels.
The new information, in the last 5 to 7 years, is that that instability is regulated.
And it's regulated by a cytoplasmic machinery whose members are still being enumerated,
but it turns out that one of the components of that regulation is MK2.
MK2's cytoplasmic kinase activity blocks the degradative machinery
that normally chews on ARE containing messages
and leads to their post-transcriptional stabilization in the cytoplasm,
and that in turn, since many of these ARE-containing messages are cytokine encoding mRNAs,
should lead to an enhanced release of secreted cytokines.
So, before I show you the experiment, let's look at the logic here.
What this pathway really is is an amplification scheme
in which inflammatory signals from the microenvironment, proceeding through this pathway,
lead to more inflammatory cytokine production.
This has made the whole p38/MK2 pathway an interesting target for the pharmaceutical industry
which has been trying to develop inhibitors of this pathway as potential anti-inflammatory agents.
Many such compounds have been developed. Interestingly, in animal testing,
they seem to predispose to staphylococcal and other infections,
which is why they haven't yet become successful drugs in the clinic.
But it speaks to the role of this pathway in inflammation and host defense.
Ok, so does kaposin B really have the activities that would be predicted from this?
Well, let's first think about how this might relate to KS pathogenesis.
It turns out that AREs, many ARE containing molecules have roles in KS pathogenesis.
IL-6 is known to affect spindle cell survival,
TNF-alpha, very abundant in KS lesions. A particularly interesting one is gamma-interferon.
Gamma-interferon, we showed a few years ago, is a weak inducer of the lytic cycle of KS,
so can promote KSHV growth.
But, the most interesting thing about gamma-interferon is that about a decade ago,
the biotechnology industry cloned this gene and expressed it
and was interested in exploring its use as a therapeutic.
And one of the things they were interested in was would gamma-interferon be a therapeutically useful drug for KS?
The drug, I think was given to 3 patients, two of whom had a florid exacerbation of KS
and that aspect of drug development was dropped.
So, up regulation of gamma-interferon is linked to the progression of KS
and of course, a final gene that's very important in KS pathogenesis is VEGF.
VEGF is controlled by an ARE and VEGF is a very potent pro-angiogenic factor.
So, I think there's lots of reasons to believe that things that manipulate this pathway
could have very dramatic effects on KS.
But all of this is only relevant if kaposin B expression really does stabilize ARE containing messages.
So, how can go about looking at that?
And Craig McCormick in the lab developed, adopted this assay that had been developed elsewhere,
in which an AU-rich element is cloned downstream of a globin gene that's under the control of a tet-operator.
That gene is then co-expressed with kaposin B and transfected into HeLa cells
that have a tet-off configuration of the tet activator.
Tetracycline or doxycycline is then added, the gene is turned on, I'm sorry, the gene for the globin gene is turned off
by the addition of doxycycline, and then RNA is harvested at various points
and probed on a Northern for beta globin.
So, this is a nice system because you can assess the stability of a message
without toxic drugs like actinomycin D.
The globin gene is turned off by the addition of doxycycline
and the half-life of the RNA is measured.
So, here's the experiment in the top panel. You can see that beta-globin itself,
with no ARE is a stable RNA and that stability is not regulated by kaposin B.
When you put an ARE from, in this case GM-CSF, but we can do this for other AREs,
you can see that as Kamen showed years ago, RNA stability is greatly impeded.
RNA is turned over very rapidly. But, that turnover can be partially inhibited
by co-expression of kaposin B. And as you would surmise from that,
cells expressing kaposin B have exaggerated production of GM-CSF and IL-6 and several other cytokines.
Now, we don't yet know exactly how kaposin B's interaction with MK2 triggers MK2 up regulation,
but here's a simple model.
Craig mapped the binding site for kaposin B to the central domain of the molecule.
It turns out that MK2, which exists in the ground state in an inactive form,
is inactive because the molecule folds back over on itself
and its C-terminal inhibitory domain inhibits the catalytic domain.
We know that kaposin B binds to the catalytic domain,
so we proposed that its action here is by displacing that inhibitory loop and leading to up regulation of kaposin B.
The story is more complicated than that
and to date, we don't have any direct biochemical evidence in vitro that this is the sequence of events
but it's our working model as we move forward.
Alright, so I think we can understand that the expression of FLIP and kaposin
can help us understand the pro-inflammatory phenotype of a KS lesion
and as students of the disease, that's very satisfying.
But as virologists, there's a deeper paradox here.
And the question involves why does KSHV want to do this?
We teach our students in graduate school that viruses usually want to evade
or block inflammatory and immune responses
and if you look elsewhere in the world of virology, particularly among the pox viruses and other large, DNA viruses,
there are many factors that are devoted specifically to blocking these things.
Soluble TNF receptors, soluble secreted cytokine binding proteins,
things that impede cytokine and pro-inflammatory signaling.
And the reason for that is clear: that viral evolution is principally driven by one thing,
which is spread; it's about enhancing viral replication, release and spread,
from one cell to another, one tissue to another, one individual to another, one population to another.
It hurts us to admit this as physicians and as human beings,
but disease, virally associated diseases, which is why we study these viruses,
is a sideshow in viral evolution. Disease usually only afflicts a minority of infected subjects,
there are examples of diseases that promote spread, respiratory infection
is very efficiently spread by causing disease,
but for most viral diseases, disease is a side show.
And that is certainly true of KS. KSHV is not, is spread in the population,
is not materially enhanced by the 1 in 10,000 people who's going to develop KS.
So, here we have a paradox and we have to presume that somehow,
promoting a pro-inflammatory environment is somehow beneficial to KSHV spread.
How might that operate? Well, one idea is that having lots of B-cells in the microenvironment
of an infected cell might provide a favorable microenvironment, or milieu, for viral dissemination.
That's a provocative and sort of attractive idea,
but at the moment it's totally conjectural and we don't have an experimental system that yet validates the idea.
So, I think it's important to remember that these pro-inflammatory molecules are playing some as-yet-undermined role
in the natural history of KSHV infection,
despite the fact that they are playing a role we can understand in the pathogenesis of the disease.