Skin Stem Cells: Their Biology and Promise for Medicine

Transcript of Part 3: Cancer: Hijacking the Wound Repair Mechanisms Used by Stem Cells

00:00:14.02	My name is Elaine Fuchs.
00:00:15.24	I'm a professor at Rockefeller University in New York City, and I'm also an investigator
00:00:21.04	of the Howard Hughes Medical Institute.
00:00:23.23	All of the tissues of our body must be able to replenish dying cells and repair wounds.
00:00:30.25	We cannot get rid of those fundamental mechanisms; they're essential to life.
00:00:36.19	And yet it's those mechanisms that cancers hijack in order to take advantage of...
00:00:44.01	of, effectively, the ability of our stem cells to be mobilized and repair wounds,
00:00:49.22	this time doing it uncontrollably.
00:00:54.23	So, we study specifically squamous cell carcinomas.
00:00:59.08	These are the second most common cancer worldwide.
00:01:05.09	When we consider that squamous cell carcinomas are not only found in the skin,
00:01:10.04	but also head and neck, cervix, esophagus, cancers of the anal-genital tract, and some types of lung,
00:01:18.19	breast... and breast cancers, as well as skin cancers, now we're talking about
00:01:22.26	the sixth most deadly worldwide cancer.
00:01:27.09	And yet there's very few effective therapies for squamous cell carcinoma treatment.
00:01:34.12	So, what we've learned in the course of the field's studies on skin stem cells is that
00:01:43.18	squamous cell carcinomas of the skin can originate from skin stem cells.
00:01:49.22	They're there for the long haul, so they can acquire the mutations that ultimately
00:01:55.09	will lead to cancers and tumors.
00:02:00.06	So, some years ago, my laboratory isolated and characterized the tumor initiating cells,
00:02:08.19	the so-called cancer stem cells, if you will, that are giving rise to squamous cell carcinomas,
00:02:15.00	in this case using our favorite model system, the laboratory mouse.
00:02:20.23	The way in which we examined this was to generate a mouse that we could effectively
00:02:28.24	create by chemical carcinogenesis or by genetics skin cancers or squamous cell carcinomas.
00:02:36.24	And we isolated those squamous cell carcinomas, fractionated their cells using
00:02:42.18	fluorescence activated cell sorting, and then performed serial transplantation of each of the pools
00:02:49.02	that we isolated from the FACS machine, tested them one by one, individually,
00:02:55.02	to find out which one or ones had activity that would lead to the generation of
00:03:02.00	a new squamous cell carcinoma in a host recipient mouse.
00:03:06.26	We got this down to almost the near single-cell level, where we could take a single cell isolated
00:03:13.12	from a squamous cell carcinoma and introduce it into a host recipient mouse,
00:03:19.01	and get a squamous cell carcinoma.
00:03:21.19	That told us that we had a so-called cancer stem cell.
00:03:25.08	What does it look like?
00:03:26.20	So, we learned that the cancer stem cells produce a new set of transcription factors
00:03:33.05	that are different from the normal stem cells.
00:03:36.05	So, the cancer stem cells produce ETS2, ELK3, AP1, KLF5, and SOX2 and 9,
00:03:44.00	whereas the normal hair follicle stem cells produce Sox9, Tcf4, Tcf3, Nfatc1, Nfib, and Lhx2.
00:03:55.08	There's very little overlap in the transcriptional regulators between the hair follicle stem cells
00:03:59.21	and the cancer stem cells.
00:04:02.28	One of the interesting new sets of genes that are important in the squamous cell carcinoma
00:04:09.25	cancer stem cells are ETS and the ELK binding sites.
00:04:14.27	These are present in nearly all squamous cell carcinomas -- super-enhancers, those are the
00:04:19.21	regulatory elements that are controlling the activity of the stem cells -- and those factors
00:04:26.26	turn out to be phosphorylated and superactivated by a particular kinase, the MAP kinase
00:04:33.10	that is downstream from, in this case, oncogenic RAS or hyperactivated RAS,
00:04:39.08	which is frequent in squamous cell carcinomas.
00:04:42.25	So, the Ras/MAP kinase pathway, then, phosphorylates and activates ELK3 and ETS2.
00:04:50.14	And we don't yet exactly know how it works, whether it recruits histone acetylases,
00:04:56.19	or whether it increases the DNA binding affinity or the nucleosome binding affinity for the transcription factors,
00:05:03.27	but what we do know is that the phosphorylation and superactivation
00:05:09.13	of these transcription factors is critical with regards to altered,
00:05:14.27	malignant program of gene expression produced by these stem cells.
00:05:20.08	And we can also take advantage of cloning one of these regulatory elements that is
00:05:26.00	turned on in the squamous cell carcinoma stem cell.
00:05:29.15	And we can use that to express a GFP -- fluorescent green protein -- in mice.
00:05:35.28	And we just use a fluorescent red protein as a control.
00:05:41.10	But what we find is that the regulatory element, now, is specifically expressed in
00:05:50.22	the squamous cell carcinoma cells.
00:05:54.01	So, we know that the skin stem cells from squamous cell carcinoma originate from
00:06:00.05	the wild type skin stem cells.
00:06:02.07	But they don't look anything like the wild type skin stem cells.
00:06:07.25	They have a completely different program of gene expression.
00:06:11.09	They're expressing elevated levels of cyclins, which force these cells into a superproliferative state.
00:06:17.21	They express transforming growth factor alpha, a growth stimulator
00:06:22.22	for the squamous cell carcinoma cells.
00:06:25.22	They express certain cell survival genes that are long-term survival genes not normally
00:06:31.13	expressed by the hair follicle stem cells or by the epidermal stem cells.
00:06:35.22	They express genes which give... which are important for epithelial-to-mesenchymal transitions,
00:06:43.14	for invasion.
00:06:45.16	They express VEGF-A that recruits blood vessels to these stem cells, and so on and so forth.
00:06:54.25	These are, effectively, a list of genes that are implicated in many types of cancers.
00:07:01.26	And yet they're not properties of normal stem cells.
00:07:05.25	So, the cancer stem cells, as we're learning, reside at the tumor-stroma interface.
00:07:11.25	And that's very similar to what I described to you for the normal stem cells.
00:07:17.03	Epidermal stem cells and hair follicle stem cells reside at the epidermal-dermal interface.
00:07:23.15	And so too, now we're finding that squamous cell carcinoma stem cells reside at the interface
00:07:29.20	between the epithelium and the stroma.
00:07:31.24	But, now, the stroma is entirely different.
00:07:35.06	There are immune cells.
00:07:36.25	There are altered fibroblasts.
00:07:39.01	There are blood vessels.
00:07:41.10	There are a whole variety of different changes in the tumor microenvironment that
00:07:47.21	distinguish the stem cell microenvironment from its normal counterparts.
00:07:53.07	So, there's heterogeneity, also, that arises in the tumor microenvironment, and that turns out
00:08:00.03	to influence their behavior.
00:08:02.20	Wherever there is a blood vessel that comes up next to the tumor-stroma interface,
00:08:09.02	those stem cells, now, respond very differently, because the perivasculature contains a growth factor
00:08:15.22	called transforming growth factor beta, which is a member of the BMP superfamily
00:08:21.14	of signaling factors.
00:08:23.14	And now that signaling crosstalk between the perivasculature and the cancer stem cell
00:08:30.05	changes the properties of the cancer stem cells.
00:08:33.04	And it does so in a very important way.
00:08:35.26	First, the stem cells become slower cycling as a consequence of receiving this growth factor.
00:08:43.05	But more importantly, these stem cells become invasive.
00:08:48.26	And that is a very nasty thing.
00:08:51.02	So, in order to be able to study this, we engineered a mouse to mark the TGF-beta-sensing
00:08:57.12	stem cells of the tumor in red.
00:09:00.20	And we also engineered the mouse so that we could track the behavior of these TGF-beta-sensing stem cells
00:09:09.24	by activating them and their progeny -- following their progeny -- in green.
00:09:15.18	And so as you can see from the tumor, here, the green cells, the progeny, of the TGF-beta...
00:09:22.12	of the TGF-beta-sensing cells became invasive, and broke down the basement membrane of the tumor
00:09:30.26	and started to invade the stroma.
00:09:34.19	What's interesting is that this stroma is where the blood vessels are that
00:09:40.04	basically prompted this initiative effect.
00:09:42.13	And so we think that it's these cells that are then the ones that get into the bloodstream
00:09:48.07	and could be the roots of metastasis.
00:09:53.08	Another interesting and important aspect from these studies turned out that the TGF-beta-responding
00:09:59.16	stem cells have an increased resistance to chemotherapy.
00:10:03.13	So, the drug of choice for treating squamous cell carcinomas of the head neck is cisplatin.
00:10:10.17	And here we treated the tumor with cisplatin, and by day 3 you can see the blue cells,
00:10:16.27	which are the dead cells, are the majority of the tumor.
00:10:21.06	But the red cells that are here, the TGF-beta-responding cells, did not undergo apoptosis.
00:10:28.23	Are they doing something?
00:10:30.03	Perhaps they're just sitting there in the tumor.
00:10:32.09	So, in order to be able to determine this, we washed away the cisplatin, and we activated
00:10:39.07	the lineage-tracing marker so that these cells are now green, and all their progeny are green.
00:10:45.15	If these cells are responsible for regrowing the tumor, the tumor should be green.
00:10:52.00	And so we did that experiment.
00:10:55.06	And what we found is a green tumor.
00:10:57.16	And what this tells us is that by stem cells receiving this change in their microenvironment,
00:11:04.17	that prompts them to receive a TGF-beta signal, causes them to become slower cycling, invasive,
00:11:11.19	and resistant to chemotherapy.
00:11:14.21	So, what's the mechanism involved?
00:11:18.21	And here, there's probably multiple mechanisms.
00:11:21.06	We've looked at one of those.
00:11:23.16	And it involves a factor called NRF2.
00:11:26.17	This is a master regulator of a pathway called the glutathione pathway, and I'll get to that
00:11:32.19	in a moment.
00:11:34.05	NRF2 is normally kept at bay by virtue of its association with an inhibitory protein,
00:11:41.24	KEAP1.
00:11:43.03	And in response to TGF-beta, p21 is upregulated.
00:11:48.04	It competes for the binding of KEAP1 to NRF2.
00:11:53.15	And that then turns NRF2 into a stabilized activator of the glutathione regulatory pathway.
00:12:01.28	The glutathione metabolism genes then go up in the cancer stem cells.
00:12:07.01	And the ones that we identified that were up in the TGF-beta-responding stem cells
00:12:12.00	are those shown in red.
00:12:14.01	So, what's the glutathione pathway doing?
00:12:16.26	The glutathione pathway is what normal, healthy cells use to clear out nasty stuff like cisplatin,
00:12:25.13	in the course of chemotherapy, or reactive oxygen species, in the course of radiotherapy.
00:12:32.26	And so by upregulating the glutathione pathway... normally, it clears out all sorts of
00:12:39.13	nasty stuff from healthy stem cells.
00:12:41.18	But here it's clearing out all sorts of nasty stuff from the cancer stem cells, the cells
00:12:49.01	that we would like to get rid of.
00:12:51.01	So, is this relevant?
00:12:53.04	I've told you about studies in mice -- are they relevant to human?
00:12:56.20	So, here we looked at the glutathione expression program in squamous cell carcinomas of humans.
00:13:04.21	And what we learned is that the glutathione pathway is upregulated in a small cohort of
00:13:10.26	human patients with head and neck squamous cell carcinoma.
00:13:14.18	And when we looked at the prognosis of that small cohort, that cohort with
00:13:19.28	upregulated glutathione pathway genes turns out to be the cohort with the poorest prognosis,
00:13:27.15	suggesting that this is an important finding that could help us in terms of future treatments of cancer.
00:13:34.18	So, NRF2, the good and the bad.
00:13:39.05	NRF2 is what is normally contained in broccoli, in blueberries, and all the things that
00:13:46.15	we're told are really good for our health.
00:13:49.00	They are if we're healthy.
00:13:51.21	But for tumor patients or cancer patients, antioxidants may not necessarily be a good thing.
00:13:58.25	And it's something that is really going to bear the need for future studies, to be able
00:14:03.15	to explore deeper into this.
00:14:05.28	So, how do we go from taking the findings that we've made to, perhaps, the clinics
00:14:13.11	or to new therapies for the treatment of squamous cell carcinoma.
00:14:19.01	Well, to this end, we've engineered isogenic strains, or isogenic cells, that contain
00:14:27.08	the ability... a sensor that allows us to look at whether or not the cells are experiencing
00:14:32.12	TGF-beta or not.
00:14:34.22	We've also engineered the cells such that the... one type of squamous cell carcinoma cell
00:14:41.07	has the TGF-beta receptor on its surface, and we've excised the TGF-beta receptor
00:14:47.21	from the other cell, making it or rendering it unable to respond to TGF-beta.
00:14:53.02	So, here's an example.
00:14:55.11	The red reporter shows us nuclear red if TGF-beta signaling is on.
00:15:00.10	And you can see that if we co-culture the TGF-beta receptor positive and negative cells
00:15:06.00	in a dish, what we find is that when we add TGF-beta the cells that have the receptor
00:15:12.22	show nuclear red -- TGF-beta is on and signaling in those cells --
00:15:17.15	the green ones, TGF-beta is unable to signal.
00:15:20.25	So, now we can treat the cells with cisplatin.
00:15:24.04	And what we see on the left-hand panel is that if you don't add any TGF-beta and
00:15:31.01	you just add cisplatin, you kill all of the cells, as evidenced by the fact that they round up
00:15:37.01	and die.
00:15:38.06	On the right-hand side, if you treat the cells with TGF-beta, those cells able
00:15:43.10	to respond to TGF-beta are surviving the cisplatin, much the same as what I described to you for the tumor.
00:15:51.25	So, we can also see gamma-H2AX, a sign of DNA damage.
00:15:56.17	And again, DNA damage occurs in the cells that cannot respond to TGF-beta,
00:16:04.09	and the cells that can respond to TGF-beta are refractory.
00:16:07.04	So, what this tells us is that, basically, now we can start to think about therapeutics
00:16:13.19	that would be able to compromise the ability of these slow cycling cancer stem cells
00:16:19.26	that have high levels of glutathione to be able to kill those cells in the presence of, perhaps,
00:16:26.09	combinatorial drugs for... for these cells.
00:16:31.04	We also think that it might be possible, through combinatorial drugs and the use of TGF-beta inhibitors
00:16:39.08	in a very co... in a very specific way, to first kill off the bulk of the tumor cells
00:16:43.28	with cisplatin, for instance, and then mobilize those cancer stem cells
00:16:50.24	that are resistant with, perhaps, TGF-beta inhibitor drugs or antibodies, and then hit the tumor
00:16:57.28	again with cisplatin to wipe out the mobilized, activated cancer stem cells.
00:17:03.11	So, these are ideas that at the moment are just ideas, but we'd like to be able
00:17:08.25	to take these ideas to the practice, and ultimately to translational medicine.
00:17:14.11	But there are many mutations, genetic mutations, that exist within these human squamous cell carcinomas,
00:17:22.24	these solid tumors.
00:17:24.17	And ultimately, we'd like to know which ones are cancer-driving and which ones aren't.
00:17:30.15	So again, my laboratory has said whether we could harness the power of fly and worms,
00:17:37.20	and take advantage of fly and worm genetics, only apply it to the laboratory mouse.
00:17:43.03	Could... to be able to screen through and find out which of the genes that have been
00:17:48.13	mutated in cancer are actually causing the cancer.
00:17:52.17	And so here, two of my former lab members
00:17:57.22	-- Slobodan Beronja, who's running his own laboratory at the Fred Hutchinson Cancer Center,
00:18:02.16	and Geulah Livshits, who was a graduate student and is now postdoccing in a cancer laboratory --
00:18:10.00	came up with a method of carrying out rapid genetics in mice.
00:18:15.08	What they did was to take advantage of the fact that lentivirus only gets into the
00:18:21.09	very first cell layer that it sees, which, right after gastrulation, is the single layer
00:18:27.09	of surface ectodermal cells that are going to give rise to the epidermis, the hair follicles,
00:18:33.01	the mammary gland, the corneal covering of the eye, and the sweat glands.
00:18:38.15	And what they did was to, then, put the mother mouse under anesthesia at nine and a half days
00:18:45.15	of development, right after gastrulation, with her pups, and basically use ultrasound
00:18:51.19	to find the pups, and then use a microinjection needle to be able to inject high-titer lentivirus
00:18:59.07	just in the amniotic sac, the fluid that the embryo is bathing in.
00:19:03.14	So, in a non-invasive way, we expose the embryo to lentivirus.
00:19:10.09	We now seal up mom, let the pups develop.
00:19:13.09	In this case, we analyze the pups six days later.
00:19:16.11	And what we find is that the lentivirus is now integrated stably into the mouse genome.
00:19:22.27	And it only infected the cells of the epithelium, and did not get into the inner layers beneath it,
00:19:30.26	the dermis.
00:19:32.02	And what we find is that the red embryo, in this case carrying... the lentivirus carried
00:19:38.10	a red fluorescent protein, is now nicely infected over the entire mouse embryo epithelium.
00:19:46.20	And when you look at a cross-section of the tissue, you see that the epidermis and the hair follicles,
00:19:52.01	but no other cells of the skin, basically, were stably transduced with lentivirus.
00:19:59.13	So, what this allows us to do is put any gene we want, or down regulate any gene that we want,
00:20:05.14	or CRISPR/Cas edit any gene that we want, specifically in the skin epithelium
00:20:12.00	in a matter of days.
00:20:13.20	That used to take my laboratory years to accomplish.
00:20:18.03	Ten years later, we've basically been able to accelerate the pace at which we do research
00:20:25.00	in the laboratory.
00:20:26.22	We've now carried out rapid functional genetic screens... genes for... screening for,
00:20:34.17	what are the oncogenes in that group of different human mutations that are found in squamous cell carcinomas?
00:20:41.13	Who are the tumor suppressors?
00:20:44.11	What are the microRNAs that are drivers of squamous cell carcinoma?
00:20:51.18	And what genes are important in regulating the balance of growth and differentiation
00:20:56.18	in the skin epithelium?
00:20:58.23	These technologies and the advances that we've been able to make... carrying out even
00:21:03.27	whole genome-wide screens in mice, which were previously, just a few years ago, thought to be possible
00:21:10.14	only in worms or in fruit flies.
00:21:14.17	So, I hope I've been able to give you the insights, today, of why my laboratory studies the skin,
00:21:21.04	and why we continue to use it as an important model to be able to uncover
00:21:27.23	the basis of genetic disorders, and also the basis of cancers.
00:21:34.04	And for the goal, the broader objective, of basically making advances in the treatments
00:21:41.20	of human genetic diseases.
00:21:44.12	The work that we do is focused on mouse.
00:21:47.25	And there's always that need to be able to draw parallels between mouse and human.
00:21:52.18	But those kinds of approaches are now possible.
00:21:55.27	This is my laboratory.
00:21:57.17	It's an international laboratory of researchers from around the world who get together
00:22:03.20	with a common goal.
00:22:05.07	And that is to work together to be able to advance our knowledge of biology and, ultimately,
00:22:12.20	to make advances that can be useful in the clinics, and in the biotechnology and pharmaceutical industries.
00:22:20.19	Thank you.