stem cells and cancer

>> we're going to get going. the first speaker is david solomon, ph.d. from state university of new york in albany, postdoctoral fellow at roche institute of molecular biology, six years as a staff fellow at the national institute of dental research, then saw the

light and came to nci and now he's the head of tumor growth factor section, he will talk about stem cells. >> thank you, terry. i'd like to thank terry, it's been several years since i've been going this, my swan song, retiring after 40 years of

service. i'm hoping to extend my say in science and doing other things like teaching. my interest started in graduate school as a developmental biologist, i moved into mammalian and to cancer institute, the same time it was

coming to light with barry pearson and others making observations that really cancer is just a caricature of development, an abnormal profile of normal development, so what can understand the genetics and epigenetic factors that regulate normal embryonic and tissue

development, we might be able to understand cancer. and like developmental biology, cancer rests at its inception in the understanding of those cells that derive cancer like normal development, that's stem cells. it's quite clear from a number of different disciplines not

only has the implication with respect to understanding cancer and treatment but other types of diseases some of which i've highlighted in this article from "time" magazine, and obviously cancer itself. so let's look at this observation, about 1800s there

was a developmental biologist, who said that ontogeny, normal development, recapitulates biologeny, basically evolution. and back in the '70s and '80s of last century, the observation was made tumors are caricatures of normal tissue development, normaldevelopment.

maybe oncogeny partially recapitulates ontogeny, temporally or spatially. a perfect example are testicular carcinomas called embryonal, highly aggressive, what lance armstrong had, but they have the capacity to differential into all the cells types derived from

germ layers, calledder calledseratomas, benign. it says cancer is a flexible and plastic disease. two other observations came to approach with respect to looking at cancer as a problem in developmental biology, the embryonic microenvironment where

beatrice mintz and barry pierce started, reprogram tumor cells to become normal cells in a lineage-restricted fashion and differentiate so they can reverse the tumor genic phenotype. there's something about those environments that can reprogram tumor cells.

the reciprocal is the case, the tumor microenvironment or the niche can actually reprogram adult tissue stem cells, induce pleury potential cells, which i'll get into later, the actual tumor initiating cells in any given cancer, this paratime here seems to be very true for a

number of different types of carcinoma and sarcoma, to a lesser degree lymphoma and leukemia. development especially early embryonic development is important, where the most pleury potent and adaptive and most primitive stem cells exist, and

that's in the stage called the blastocyst, day 4 in mammalian development, a body of cells within the blastocyst called the inner cell mass which gives rise to a population of cells called the epiblast cells, those cells have the properties to give rise to three germ layers.

people have derived embryonic stem cell lines from these inner cell mass epiblast-like cells, both from mice and pigs, of course you heard about the contentious research about using human embryonic stem cells. very recently, this resulted in the award of a nobel prize,

made a seminal observation, you could take adults differentiated cells like fibroblasts or epidermal skin cell or hepatocytes and reprogram them to become more like embryonic stem cells, and these cells are called induced pluripotential cells, they use the same genes

important in embryonic development. we'll get back to that pointed later. these are different sets of pluripotent cells that renew. in the adult there are tissues and organs, brain, mammary glands, bone marrow an skin

where some populations of stem progeny have been identified, there's a good deal of understanding as to how these cells give rise to a lineage commitment with respect to differentiation in these various organs. cancer is shown that cancer and

tumors per se have a subpopulation of cancer stem cells that can arise either from fully differentiated tumor cells de novo or from preexisting normal stem cells in that particular organ where the cancer has developed, or progenitor cells that have been

transformed. and the survivl of those cancer cells in the tumor as well as normal somatic stem cells depends upon the niche, the surrounding other types of cells, fibroblasts, endothelial cells, neuronal cells, infiltrating lymphoid cells,

macrophage or neutrophil, that supply cytokines and growth factors to maintain and keep those cells in a functionally active state. so it's important to realize we're talking about a community effect here. and cancer is dysfunctional

communication. and that means the niche and the tumor cells are not communicating correctly as they would normally be in normal tissue development. now, there are several models that have been used to explain the heterogeneity that exists in

cancer with respect to cancer stem cells. the first says it's a stochastic event, any tumor cell has the potential to be randomly transformed to give rise to a cancer stem cell following immortalization and subsequent transformation.

this is the stochastic, clonal, and then the progenitor and fully differentiated cells. this event says that the cancer is clonal, it keeps giving rise to stem cells and they themselves can potentially give rise to other differentiated cells.

another theory says maybe it's not that simple. maybe there's really a method to the madness. cancer cells, one property is their ability to self renewal, that's not true in the population of tumor cells, they divide but don't have renewal

capacity that's indefinite. stem cells do. that's the problem with cancer treatment today. it doesn't have a way of it's treating the bulk differentiated tumor cells. so by asymmetric division, and this is now the classic cancer

stem cell model, or hierarchical model he have a preexisting subset of cells that can asymmetric really divide and give rise to cancer stem cells or noncancer stem cells, the more differentiated progeny, okay? and this can follow from the

transformation that either normal population of tissue stem cells or their progeny, which are transit amplifying cells, from a primitive or more quasi-differentiated progenitor cell, that's an important piece that can explain the heterogeneity that exists in

various types of cancer. and then there's a third model that actually says, well, actually this is a modification of the classic cancer stem cell model, that is there there's a bidirectional conversion that maybe a cancer stem cell or progenitor cell can go backwards

and give rise to cancer stem cells themselves so there's an inner conversion or meta plastic state that says that there's another possibility of regenerating the cancer stem cell pool from preexisting progenitor cells that are present within the tumor, much

like you have in normal organs. and bob weinberg has basically validated that model, and one of the factors that drive that is epithelial mesenchymal what are some of the factors that can regulate the frequency and the actual population size of a cancer stem cell population

or progenitor population in the tumor? okay, so as i said, tumors can arise from normal populations, either stem cells with progenitor cells in an organ, and they are obviously giving rise to downstream progeny cells, or progenitor cells, that

can be malignantly transformed. now, these cancer stem cells and progenit can have genetics as well as -- and/or epigenetic modifications that occur temporally and spatially, and that have accumulated over time and they can actually contribute to the process of tumor

progression which means tumor aggressiveness as well as subsequent metastatic metastasis that kills the patient and not the primary tumor. that occurs through a process called intravasation, exit of the tumor cells into lymphatic

and extravasation, into organs where the cancer cells colonize. most of the evidence suggests that it's the cancer stem cells that are the ones that actually are involved in seeding the metastatic sites. now, there are contextual signals that the cancer stem

cells are exposed to from the tumor itself and/or christian the niche, endothelial cells that can contribute to and maintain the process of these genetic and epigenetic alterations in the cancer stem cell population, so it's a very

dynamic system. you have to think of it, not only treating the cancer possibly, understanding stem cells and target the therapy, the surrounding niche cells which actually support those finally, the immunological status of the host in which the

cancer is developing is absolutely important and very exquisite because tumors can arise under indifferent immunological states, these states neither positively or negatively modify the presence or absence of cancer stem cells. and the phenotype that these

cells exhibit in a temporal and spatial way, general observations. what are some of the properties that are shared by normal tissue stem cells and cancer stem cells? well, first, the first important property is their ability to

self renewable. and that is these are tissue-specific stem cells and organs are self-renewing by a processes of asymmetric cell difference, not symmetric, through the lifetime of an animal, in rapidly turning over a stem cell population like in

the gut, or bone marrow, or in a slow process in certain organs like the breast or pancreas, and they actually maintain these tissue-specific stem cells for their progeny. this specific differentiated cells in those particular organs or tissues.

cancer stem cells, the tumor initiating cells, exist and at they are capable of recapitulating the tumor, and have the same property of self renewing. and the self renewal maintains president stem cell population in an asymmetric way, whereas

the more differentiated tumor cells divide in symmetric division. finally, differentiation itself of both normal tissues and cancer tissues gives rise to heterogeneous populations, and these recapitulate the phenotype of the tumor as well as the

organ, stem cells and cancer stem cells, unlike normal stem cells which undergo periods of regulated quiescence wherer there not dividing, and they do get reactivated under certain conditions, like wound healing or inflammation, this is a very controlled process.

cancer stem cells tend to be more quiescent except for when they asymmetrically divide and give rise to themselves in the more differentiated population, but their quiescent problems are more extended than normal tissue stem cells, and that property actually is detrimental because

it contributes to the intrinsic chemo and radioresistance of these cells, versus the normal bulk tumor cells which are more differentiated. and finally, both cancer cells and normal tissue stem cells use very similar regulatory pathways to control self renewal and

differentiation of those cells into the heterogeneity of the original tissue or the tumor itself, these are temporally and spatially controlled in both types of conditions. so let's look at normal embryonic development. i think it's an important point

to understand since these are the genes that are the bad news genes. normally embryonic development is really important for cancer progression is really occurring early, day 6.5, and there are a variety of genes that are important like nanog, oct3/4,

which have the potential to give rise to all of the germ layer of embryo, and these germ layers give rise obviously to specific subsets of different tissues. and they are signaling molecules, that are involved in regulating the derivation of these different germ cell

populations. and interestingly enough, a number of these signaling pathways here and down here are active in cancers especially the more aggressive cancers that are very difficult to treat, and my interest is in breast cancer so triple-negative breast cancer

has a higher propensity to exhibit these embryonic genes, luminal types are less aggressive. the primitive streak is the process that gives rise to these three germ layers. the physical process whereby it's the overlying epiblast,

ecto derm is invaginated, epithelial cells convert to mesenchymal, giving rise to the endodetermine, mesoderm and endoderm. they can go backwards and give rise to epithelial cells, mesenchymal epithelial transition.

there's emt and met, very die dynamic that occur early in the embryo at an early stage. and they are temporally and spatially regulated, there's a concentration dependent effect of genes with respect to how they can function. and that's basically

morphogenic. different genes are turned on post fertilization in the embryo at different times, different levels of expression. it shows you that sort of example that i'm talking about of concentration and temporal regulation, such as nodal and

you can see the genes expressed at high levels, some are expressed at lower levels. there's a sequential process of gene expression, they have to be turned on at the right time and at the right place. signaling pathways like the beta-catenin or cripto-nodal,

turned on at a specific time, there's a concentration dependency effect when to be turned on and how much they are turned on and they shut down, and there's gradation in color, with respect to linear demarcation, that's an important phenomena is contributes to

cancer. a number of these pathways actually cross-talk with each othero it's really complicated but shows you the genes have a very dynamic and and how might they contribute to cancer, so let's look at this. ths is from a meta analysis of

40 studies of 40 human embryonic stem cell lines looking at genes most frequently expressed, consistently expressed. i tried to highlight the ones that are of interest because they are called the trinity or pluripotentiality that maintain the hierarchy of primitive stem

cells, nanogs, oct4, and sox2. there's another gene that's particular called tgrf, or cripto, genes our lab discovered. these genes are consistently found to be important for maintaining pluripotentialalitiened y andrenewal.

these can co-occupy promoters of a number of different genes, this just shows you an example, let's say of nanog or sox2 or myc, and promoters occupied, and if you look at these promoters individually, there's a tremendous amount of overlap of co-occupancy by not just one of

these guys but at least two or three, like nanog, sox2 and oct4, interacting to regulate a downstream panoply of genes, like a pyramid effect, shown here in a simple cartoon. you have these three genes, nos gene, activating genes that maintain embryonic stem cell

pluripotency, that's a positive effector. these genes are also silencing genes which are basically involved in differentiation. so they are keeping these guys on but some of these guys off. genes that are involved in the terminal differentiation of

let's say ecto derm or meso derm, that occurs through the cooperation with the poly gene or polycomb gene complex. as i said, these three genes, nanog, sox and oct 4, can occupy promoters and specifically interact with each other and form heterodimerric complexes.

you can see level of complexity and things can be finely tuned and regulated with respect to transcriptional control. ips cells were derived by yamanaka and other groups using the same sets of genes important in early embryonic development, that is oct4, sox2, klf4 and

c-myc. the function of myc which is not good, a known gene in cancer that's bad news, can be replaced by lin28. using these subsets of genes they were able to show that you could actually transfect these quartets of genes into fully

differentiated cells, in this case mouse embryonic fibroblasts, push them backwards and they became more embryonic in appearance and they are called induce the pleuri potential stem cells, and can be redirected using conditions to differentiated cells.

this is reverse differential. that process is occurring by transition, when you want to get a cancer stem cell, cancer is originating from carcinoma by definition, derived from epithelial cells, going back to a mesenchymal or primitive state which is indicative of a cancer

stem cell, and that process is controlled by bmt, the reverse process. i hope i'm showing you symmetry and how they are interconnected. what is the cancer stem cell phenotype? first of all, self renewal, tumor initiation is a low level

of tumor cells. the more aggressive this stem cell is, the lower the number of cells you need to use to innoculate to get a tumor. the more differentiated cells lack that property. signaling pathways that are used in stem cells, cancer stem

cells, are the same or similar to stem cells, pathways used in embryonic stem cells. they recapitulate the original histotype of tumor, that's one of the properties of cancer stem cells, they have increased dna repair, they are quiescent, they are resistant to apoptosis and

those properties in conjunction with upregular regulation of drug transporter make them resistant to radio and chemotherapy, they don't respond to normal conventional chemotherapies and actually emt itself is the process that gives rise to these cells and induces

and facilitates ma metastasis. this shows information recapitulating the idea cancer stem cells can either originate from normal tissue stem cells through transformation, they can arise from the progenitor progeny transamplifying, or there's a population of mature

tumor cells that have the potential to be mutated and become cancer stem cells. but these two populations here that in general are the more dynamic and more aggressive with respect to their ability to give rise to the cancer stem cell population and to more

aggressive type tumors. these usually occur through mutations in the long term, but generally before mutations occur you have epigenetic alterations that set them up for those mutations. so here are some examples, in the skin or the breast.

you can have transformation of different cells in the lineage, let's say of the skin, that give rise to different types of tumors, papillomas, basal cell carcinomas, or triepithelial derived from different lineage in the normal pathway. the same is true of breast

cells derived from stem cell cancers, triple-negative, or give rise to differentiated luminal cells and these are basically lobular carcinomas, more treatable. the same phenomena is true for lymphoma, leukemia and colon so several studies have actually

shown in various types of cancers, and my interest is in breast cancer in particular but also true in let's say glioma and bladder cancer, these genes, the nos genes as well as their target genes that i was talking about, that they control, are reexpressed in very malignant

types of cancers associated with these type tissues. so that's quite clear, the likelihood of the cancer being more aggressive and expressing embryonic phenotype is bad news. the more differentiated the phenotype and tumor, the less probability is that you're going

to have that kind of profile, genetic profile, transcriptome profile. so let's look at breast cancer, per se, okay? so this is just looking at nanog, oct4 and klog. these are genes associated in tumors that are high grade, and

associated with poor prognosis. if you look at the genes expressed in this stem cell-like model, expressed in early embryonic development, these are in tumors that are very aggressive, usually, and the genes that are expressed in more differentiated tissues are

silent. so that makes sense. what about breast cancer, which is my field of interest. so this is a disease that occurs in a quarter of a million women, 40,000 deaths occur yearly. it's second only to lung cancer in frequency, it is a lifetime

risk, one out of eight women developing breast cancer obviously higher in people with brca1 and brca2 mutations, they give rise to nasty tumors with basal and triple-negative. the stage of menarche, if it's early that's a risk factor, because it's the length of

estrogen exposurery determines your propensity, also parity, whether you've had children, nulle parity women, looking at nuns, let's say, have a higher risk of developing breast cancer because they never had kids. if you have children later in life, single pregnancy, that's a

risk factor. you've had thesecond pregnancy interestingly enough it diminishes the risk factor. obesity, there was a study done women that these women have a 70% increased risk of developing breast cancer. adiposeites in a person that's obese is not a good thing.

it's going to lead to your risk of breast cancer being increased, that's also true for colon cancer or other types of as i've said, there are genetic mutations or genetic familial breast cancer that account for 5 to 10%, usually due to mutations within the brca1 and brca2 gene,

these are regulating basically dna repair. and they are very aggressive tumors. so we can histologically classify breast cancer. now there's a new way, charles perot did this, brilliant. a transcriptome analysis and

defined the phenotype of the cancer by molecular profile and what genes were expressed. he broke tumors into six major classes, luminal a or b, and lack her-2b, an antibody used to treat it. there's a subtype call normal breast cancer which looks more

like adipose tissue, interestingly enough. and then you start getting into the bad news tumors. the basal like tumors, her-2 enriched, more undifferentiated, meta plastic, these tumors here specifically basal like and claudin low are extremely

resistant to those sorts of therapies, triple-negative breast cancers because they lack estrogen receptor, progestin receptor and her-2, that's why they are called triple-negative, more of a mesenchymal phenotype, more meta plastic. so what about progression in

breast cancer and association with some of the nos target genes, there's a positive correlation with grade and er-negative activity, and more basal-like or triple-negative-like tumors, and the expression of these nos-like targets.

so they have a greater tendency, er negative tumors, grade 3 tumors and basal cell-like tumors, to express embryonic genes, when they are reactuated, es genes or es signatures, those patients not only have poorer survival but also they are metastatic pre-survival is

significantly shorter than populations of cancers where the genes that are not expressed. here's another one for survival, cumulative survival, nanog and oct4, if you express that alone your survival is worse, and if you express them together it's even worse.

so in breast cancer. this is another study. here is another study showing staging and expression of some of these genes, you can see going from a stage 1 breast cancer to stage 4, the propensity or proclivity for exhibiting higher level of

expression of nanog, oct3/4 and sox2 increases when you use normal staging of breast cancer. and interestingly enough, if you look at the gene temporal transcriptome analysis in the normal embryo around day 6.5, the day i was talking about, with vagination and germ layer

formation, if they mirror that gene set expressed in the early embryos, their recurrence rate is far greater than tumors that don't express that sort of gene signature set. here is three separate studies that show that. if they are expressing genes

associated with differentiated cells they have a better survival rate, so it's just the opposite of that which you would expect. that's not true for emt-related genes, interestingly enough. it's embryonic genes collectively that's important in

this context. you have to understand normal breast development to study breast cancer, hard in a human, so we used a mouse. most of the development of the mouse breast occurs post natally. puberty in the mouse occurs

between four to six weeks, that's when you see the outgrowth of a duct-like system of glandular epithelial cells. animal ages, the glands becomes filled with arborized system of ducts, okay? that these ducts are actually -- they terminate with terminal end

buds, i'll get to that in a minute. if the animal undergoes pregnancy, okay, estrogen, progesterone and prolactin are involved. it's involved in ductal elongation and growth. when you see extensive side

branching in early pregnancy, progesterone primarily in the early stages of pregnancy and then prolactin are involved in that process. side branching and alveolar development, these are the structures that make the milk protein.

then you get the full blown lactation where the gland is chock full of al alveoli, wherethe mammary gland is, producing milk and secreting milk. when they drop off, the gland involutes. if you look in detail, in the virgin there's structures at the

end of the ducts that are terminal end buds that contain populations of basal cells called cap cells in the terminal end buds, the cancer stem cells. the underlying epithelial cells are the lu minal cells, they give rise to the ducts, and the myoepithelial cells, and this

structure during pregnancy, early pregnancy, under the influence of progesterone and prolactin develop alveoli, the mature milk-producing structures in the glands. that phenomena during normal ductal elongation and alveolar development not only is

controlled by systemic hormones but those hormones in fact control the production of local growth factors and cytokines in the different subpopulations in the mammary glands. so for example, there are luminal progenitor cells, erpr positive, when you expose them

progesterone will induce specific growth factors such as ligands or win 4, involved in driving proliferation of mammary basal stem cells, adjoining the luminal layer. there's the er/pr negative, these give rise to alveolar progenitor cells.

this is a dynamic system where cells are communicated in an interesting way, controlled by a social community of cells being regulated by systemic hormones of in cancer, unlike normal breast development with interplay of cellment, cancer cells tend to produce their own

growth factors, they are in a self stimulating mode. they don't care about these other supporting cells for survival in the sense of meeting growth factors that would normally be used to maintain normal mammary gland structure. so they drive these progenitor

cells to become pr negative luminal progenitor which become aggressive because they have propensity to become more basal-like in phenotype. the lineage in the mouse mammary gland has been studied, there are various markers, i won't go through them, that describe the

primitive basal stem cell population, more mesenchymal, and there's a common bipotent population in the postnatal gland, higher in prenatal gland, this population gives rise to myoepithelial cells, and basal cells, this population also can give rise to luminal progenitor

cells that have the potential to differentiate either the ductal luminal cells or alveolar luminal cells, and there are markers that define each one of these subsets. you can get transformation of any one of these subsets, and the earlier you get

transformation if you transform early mammary stem cell, usually you end up with a basal tumor, triple-negative, more differentiated cells if they become transformed they are less malignant, more luminal and more treatable.

here are common subsets of markers across different types of cancers, there are markers, there's prominent, there's cd44, which is involved in cell adhesion interestingly enough and other regulatory processes. drug transporters here and here that have been used to mark stem

cell populations in different types of cancer, more recently aldahide hydrogenase, to mark different populations of stem cells in different types of tumors. and in breast cancer population aldh positive tends tore it more progenitor like and

less basal like, it's the downstream progeny. still not a fully differentiated lineage cell, still bad news. these are markers that can define stem cells in those distant types of scenarios. here is a comparison of mouse and human markers, some of which

are shared, that has been elucidated in the breast. human versus mouse. and you can see here there are a number of integrants that have actually been shown by cd29, cd49f and cd61, beta one, alpha six and beta three integrant overexpressed in stem cells,

probably involved in maintaining the niche in the cooperative environment, and cd44 down here in humans which is common, as in cd49 in the mouse. you can see in the mouse you can have transformation of any of these populations, whether it's the primitive stem cell or the

progenitor population or the more fully differentiated population, and actually experimental models have been produced in transgenic settings, called gems, if you overexpress certain genes known in the etiology of mouse and human breast cancer, also normal mouse

mammary development such as wnt1 or p53 null they give rise to basal tumors, meta plastic, nasty tumors, they look like a triple-negative breast cancer. if you start mutate brca1 or brca2 or overexpress notch 1 or 2, they are also triple-negative.

they are a little less aggressive than these, but they are still a very aggressive tumor that resembles to a large degree some of the triple-negative breast cancers down here. then if you use ras or myc alone and transform mammary

epithelial, you're transforming the more differentiated luminal cells whetherrer in alveolar or ductal and you get more of a tumor that looks like a luminal tumor in the human. and so here is the comparable human scenario. lineage-specific cells in human

if you transform the more primitive basal cell you get claudin tumors, more aggressive and mesenchymal. here by brca1 or other types, they are triple-negative, styl mesenchymal. her-2, luminal a and b derived from cells more differentiated

and less aggressive. there's treatments for her-2 with herceptin but they are more emt is a process which contributes to genesis of tumor stem cells, generates tumor stem cells because you're taking an epithelial cell making it a mesenchymal cell, those cells

that are derived exhibit stem cell-like properties and acquire the properties of metastasis, and aggressiveness, mesenchymal, resistant to drug and radiotherapy, and they are very immuno-- resistant to immunotherapy. the phenotypes associated with

these states, epithelial cell, that converts to a mesenchymal cell, all these cell adhesion proteins that are necessary, these are lateral proteins that maintain basal polarity, and cadherins junction. these proteins are indicative of fibro plastic mesenchymal cell.

these proteins actually suppress these guys in the epithelial compartment and turn on these s in the -- gives rise to mesenchymal phenotype, and this process is not unique to cancer because you have emt occurring in carcinogenesis, fibrosis, so it's not unique to cancer.

factors that stimulate emt, multiple cytokine and growth factors, some of which i listed here, wnt signaling, tgf beta, notch, factors to contribute, such as the bmt, local factors involved in this process. yes? >> (inaudible)

>> correct. >> (inaudible). >> yes. so you can use -- they have an immunocompromised mouse now, i forget the name, where you can take a small population like that represents 1% to 2%, and raise it up to 5% in this

immunosuppressed model. yes, it's been done for melanomas. i forget. a guy who did this is in the university of wisconsin, i forget his name. ryan -- i'm sorry, i'm losing it.

>> yeah, if you make the animal more severely immunosuppressed, you can blow that population up, so that in itself says that there's a system you're testing for stem cells, it's important, and maybe limiting your ability to detect those stem cells, depending upon the immunological

data that posts. here is a reverse situation. if you overexpress octor nanog, okay, you get the expression of more mesenchymal like genes, slog and snail and mesenchymal genes and cadherin. if you knock down these genes, you actually do the reverse, you

get the appearance of a more epithelial-like gene, so this has been done in experimental context, in breast cancer stem and reciprocally, there are human cell lines here that when treated with factors that induce emt, they exhibit more stem-like properties.

so you can find ways of actually spanning population in vitro, genetically as well as epigenetically. and embryonic development in the breast there's a up in of different tissues, a series of progressions from going to normal breast to hyperplasia,

carcinoma in situ and invasive carcinoma, a number of genes that have been identified associated with the various steps of progress. i'm not going to go through them but you can see a lot of them are very -- are genes involved in early embryogenesis and emt.

here is carcinoma progression relationship to the breast. you have niche cell types, cancer associated fibroblast, macrophages, endothelial cells, these guys are dumping out cytokines and growth factors, a whole series of them. bad news.

they contribute to this progression. if you transform the breast, let's say with radiation, you usually get hyperplasia. but there's a form called atypical which is bad news, higher risk of converting to ductal carcinoma in situ, when

it becomes invasive it's an invasion that starts from the periphery of the tumor, called the invasive front, and this is where the cell was more mesenchymal, and this is where emt is occurring, at the invasive front of the carcinoma, seen in other types of tumors.

that process is controlled by niche factors secreted by some support cells that are brought into the environment of the tumor or the mammary gland. the cells undergoing emt become mesenchymal, become invasive, get into the stroma, intravisate and they can disseminate.

in breast cancer organs that are tend to be colonized are lung, brain, liver and bone marrow, the process of colonization circulating tumor cells depends upon these are more mesenchymal, to colonize these end organs they have to revert back to a more epithelial state, so what

happens is there's an met process, those cells now redivide back in those sites of colonization. so you can understand that process, you may have a respective target. i've looked at a protein called cripto, gpi linked, cell

associated or it can be secreted, so it's functioning in an autocrine, soluble or tethered in exposome, multi-functional chaperone protein that binds to nodal and can bind to grp 78 and alk 4, the cfc domain, this protein is expressed in the embryo, i won't

go through this, it's important in meso derm and endoderm formation, at this point they die, so it has a very important role there. how does this signal? t's a chaperone protein for presenting nodal and gdf 1 and gdf 3 factors in the tgf beta

family, the alk 4 receptor normally binds to r2b which transphosphorylates, you get activation of the pathway, cripto can also act in the be a accepts of nodal and gdf and snad, become grp78, co-receptor for this pathway, and this pathway, and then you get the

activation of ras, they are downstream effectors, and finally cripto itself can act as inhibitor of tgf beta it's a ying and yang situation with respect to level of expression of this protein. there are negative effects of this pathway, with the capacity

to bind nodal or cripto, or cripto alone and sequester it. and shut it down. and if we look in normal human tumors, this is a meta analysis from probably 20 different labs showing expression, from 30 to 100% in frequency, in breast, colon, cervix, lung, nonsmall

lung, pancreatic cancer, showing you staining in different tumors for this growth factor. so it might be an interesting target for therapy, and we've developed an antibody conjugated, phase 1 clinical study showing partial response, it may be interesting diagnostic

therapy in plasma or serum. you look at the -- if you overexpress this gene it takes normal mammary epithelial cells and makes them mesenchymal, they form tube or duct-like shat's in vitro. in vivo in you make a transgenic mouse, what you see here is that

there's hyper branching occurring in these transgenic mice compared to the normal litter mates that are wild-type that aren't transgenic, you can see that here and here, hyper branching, secondary and tertiary branching, a lot have structures ductal and nodules

and these animals will develop papillary adenocarcinomas. so what about breast cancer, human breast cancer? interestingly enough, if you look at the molecular subtypes, l a and b, higher levels in the her-2 positive, highest level of expression in the basal

or triple-negative breast here is an indication, association with that. i'm getting to the end right here. terry, i know yre getting nervous. how do we treat? we can treat with conventional

cytotoxic and radiotherapy, but that only wipes out bulk cells, still leaves cancer cells. you could still get metastasis. target cancer stem cell and use conventional therapy, chemo and radiotherapy, you might be able to get a more greater remission response to the tumor itself,

the primary tumor, as well as to metastasis itself if you wipe out that cancer stem cell population, or the niche. here are the pathways targeted and inhibitors put into phase 1 clinical trials, that target car stem cells, and i'm not going to go through them but

just to show you there are things in the clinic and out there there are using these embryonic signaling pathways as targets reactivated in cancer the niche itself, macrophages, t cells with subtypes, adipocytes. companies have developed compounds against some of those

factors that are secreted by the niche cells, the infiltrating lymphoid cells or activated fibroblasts. so you can imagine taking these kinds of drugs and putting them together with the conventional drugs that are out there that have their day in the sun and

not very good, unfortunately, in response rate because you get reactivation of a tumor that's nasty eventually, after you treat the patient. that's shown here. so here is an arborist taking the tree down, gets to the stomach and says the hell with

you're killing the tumor but what comes back is a tumor that's more aggressive, look at the tree now. it's bad news. that's what basically happens in resistance to chemotherapy and radiotherapy. if you now attack the root,

which is the stem cells, that generate the bloody tree and after you chop the tree down maybe you can get rid of it, and also maybe if you poison the soil in which the tree is growing, that is the niche, you could accelerate that process. you have two additional target

zones, other than the direct bulk tumor, you have the cancer stem cells and their progeny, as well as the niche. and i thank you. [applause] and i didn't dwell a lot on my particular research which is cripto.

i used it as an example but tried to give you an overview of the relationship to normal development, its importance in cancer stem cell genesis and cancer progression. >> moving on, the microphone. >> that would help.

>> okay. so after this next lecture we'll be going over to the pathology last week so we're scheduled to do it this afternoon. our next speaker is dr. verma, he got his ph.d. in india, i first met him when i was a professor at george washington

university, subsequently i came to nci and so did he. so currently he's in the division of cancer control and population sciences, chief of the epidemiology and genomics research program, he's going to talk to us today about epigenetics.

>> thanks a lot, terry. i'm happy 26 years ago, (indiscernible) we used to practice our seminars. i'm glad he came. as terry mentioned, we are epidemiologists, we work with risk assessment of identifying those factors, things which can

contribute to disease when we make prediction models in 2022 we expect almost 18 million will be survivors, we have some success by screening problem. if we can detect cancer early, then we have option that we can do

some therapeutic or approaches. except ovarian and pancreatic, other cancer takes long time, you have more option to do something. colon cancer has been studied more, and now we know that based on that age, certain kind of

markers will appear. when biology of the cancer came up, those kinds of information are there, which we want to utilize time to time. and here also it is shown different stages during the cancer development are also shown in major kind of cancer,

and when risk assessment came, previously people used to work with one gene, then genome wide association studies, tag the snps, it's time you saw in one chromosome, one marker, as time passes by more and more markers, so gradually we make progress, started getting information

because genetics was dna based, easy to do and human genome map was made and after that complete genome was sequenced, a number of markers associated with cancer were identified, those are the basis of our pharmacogenomics structure and other biology which is coming

out of that. now in cancer genome we know that somatic mutations occur in the body, approximate number knowing different kind of organ sites, in childhood, adulthood, how many are those, at the same time we know that in cancer major are 12, now if we identify

gene or some feature we can tell which pathway is affected, what we have to do regarding that. now comes epigenetics, genomic sequence does not change, it tells of our behavior. what is epigenetics? these are stabilized gene expression by several mechanisms

so you know about genetic code, and now methylation code and histone code are two codes, very important in this indication. so genetic information tells you what you can do, what epigenetics will tell, when to start something, when to stop. and the name suggests epi means

above, genetics, time to time you'll see in popular magazines some articles that your dna is not your destiny. the reason is your environment and your lifestyle affect you more in terms of developing disease or being healthy. these studies were done on mono

zygotic twins, several twin cohorts are also there. when the twins were followed, initially found cardiovascular, later on cancer, established the fact there is some mechanism than genetics, and that gave rise to pharmacogenetics, pharmaco epigenetics,

genome-wide studies, snps, epigenome, components which are in epigenetics and i want to discuss those, those are followed. nucleosome, 146 base pairs, around is four kinds of histones are wrapped around, and tease are in terms of octomers, and

these are types of histones, stems from each of the thing that regulates gene expression. these are the components of epigenome, these are components like methylation code, the promoter region, the sequences are there, sometimes c gets methylated, it becomes five

methylated, it starts behaving differently. another thing histones, they are neutralize child of nucleic acid, they get modified by methylation, phosphorylation, ubiquitination. these enzymes of proteins enrolled in methylation or

histone modification or nonhistone protein, those also have incorrect size, like polycomb. this looks complex. some consequences, these are the repeats, usually people don't tell any importance but in fact by epigenetics we knew they

continue to genomic instability, cancer development. methylation of dna, they kinds of events, initiate methylation or maintain methylation, or modifications will come, about histone i told you, along with that noncoding rna especially microrna also affect cancer

in genomic imprinting it's been seen especially in igf. in promoter region, cpg sequence are unmethylated. (indiscernible) but if genes are hyper methylated, then it gets inactivated, at the same time some promoters have already methylated, they get

hypomethylated like oncogene, both play a significant role in one zygotic twins, the study was done, and in schizophrenia, it plays a role. age, lifestyle, environment, other things, mutate dna from cells in the body and sequence, sequence will be the same,

epigenomics is dynamic. now, throughout the life starting from prenatal condition and then babies or children, lifelong, epigenetics changes have been studied now. which factors are affecting, at which state, initially maternal and other effects, but later on

it is lifestyle or diet or other things. we prove in many several or hundreds of thousands of people we collect samples of other different times when you can select and analyze, that's how things are done. you can imagine if one pregnant

lady is smoking, three generations are affected, her own, the baby, those stem cells as you heard in the previous lecture. mostly epigenetic changesry are quite early in time have more effect later on, what kind of how many diseases they will have

developed. and development plays an important role, many things in the environment like arsenic, benzene, cadmium, chromium, it's been observed they are related, epigenetically. in fact, if you see that phenomena, exposure to

radiation, they happen simultaneously, short time transcription factors of cell development they are affected. if they are for very long time, genetic come. if you reduce something, that is better. here i have shown how many --

these things have information available for different cancer, how many genetic changes are there. epigenetics, the pioneer work from peter jones, university of southern california, andrew fine feinberg at johns hopkins.

if you see the structure, enthuse nucleosomes are apart, if you see the bottom part they are compact, that color shows methylation which comes in the region at the same time they lose circulation part, and then the nucleosomes are so tight that transcn factors for

many other factors which are needed for gene expression, they can not get in, gene gets inactivated, that's the basic mechanism. it works. when methylation occurs, it happens gradually, if you do some kind of experiment you have

to expense both, then only you get correct information for that. also for whole genome global hypomethylation, they are already methylated, hypomethylated, but in promoter region for specific genes they get hyper methylated.

time of exposure during the lifetime histone modifications they increase. methylation, three situations can has been. abnormal increase, abnormal decrease or no change. in tumor suppressor genes, one or both allele is methylated,

all imprinting gene, one is hypomethylated, another is replaced by recombinase, decrease also it was seen, retrovirus used to be delivered recombinant genes and many times those vectors will not work. in the early '90s epigenetics was not developed.

that kind of situation was in abnormal protooncogenes. no change, chemical induced or deactivation. those genes enrolled in pathways, that you can tell. similarly, ten years ago microrna, they did not think in

humans we can find, but now in different cancer, cancer specificmicrornas, you can get, 20 to 22 nucleotide long, and they are very stable. not only they are used, they are profiled, used to screen populations, associated with certain kind of cancer to

identify which population is at high risk but those genes which code for microrna, the polymorphism you can use for identifying high risk population. here it is shown that micrornas are sequence , they are used,

activation or inactivation depending on which system you're studying. histones as i mentioned, these are the four histones, all modifications are shown here. in different cancer those modifications have been identified.

so we initiated one program called nis road map program seven years ago, the idea was the whole genome sequence was done, same way we wanted to do methylation profiling, microrna profiling, what is dna's mapping, chromatin accessibility, that was

completed last year. also these histone modifications to characterize monoclonal antibodies, we gave funds to companies to raise those, distributed to investigators so they can characterize those. and now we know that at different locations also you can

identify which histone modifications are there and you can follow. that profiling helps in differentiating from control from disease. and since this field is a new in 1997, 1998, after that one book i edited, i updated about cancer

epigenetics, since most you cannot implement unless you do epidemiology, different components or factors like modifiable factors, alcohol, smoking, radiation exposure, all of that, how they work, that i compiled. and then for all these studies

we want to have such biomarkers which you can collect noninvasively, we're last year this book came on noninvasive early diagnosis and prognosis of this year in january this cancer epigenetics came, risk assessment, diagnosis, treatment and prognosis, everywhere this

implication is there, one time i was interviewed by "nature," and they asked me why do you think epigenetics is so good, what do you think, and i explained epigenetic changes are reversible and there are have an edge over genetics. the edge, if you identify

mutation, the most you can use that information to follow up a patient or new people who do not have to to identify that you are going to have cancer. these epigenetics marks which i call methylation histone, are reversible. it has a lot of implication in

therapy. next year again i was interviewed by "nature," and then i mentioned that successful approval of first generation of drugs, so this reversal of epigenetic changes, then people started making drugs for therapy, so four of them have

been approved for cancer treatment. that was also made a breakthrough. now the emphasize epigenetics searched very much. when you plan such experiments in methylation, then to reduce the false negative and false

positive results, you have to think about methylation content in that area, methylation level, different points and profile and certain profile, so then you'll get better results, better planning will give you good results. also at johns hopkins

steve baylin has sequenced, all sequences identified by him, also he has identified -- isolated some compounds in lung cancer which are doing very well. example is suppose methylation comes in h3, phosphorylation, change is permanent, at

different points identified. since we have to in this research collect samples, during cancer development tumor cells shed some cells, or now this concept coming for exosomes also, they can carry dna, rna, microrna, all those, hyperplasia, dysplasia,

carcinoma, collect those and follow up in the same patients longitudinally if you can follow them individually. s have been identifieduntil different cancers a set of which are used as biomarkers for screening. i have compiled those. this table shows that.

at the same time, four different genes, they have been categorized into different categories based on how we can use this for prevention, or modifying those changes. so those things are also going on, so these are examples of those just which have been

developed, as cited, and jean-pierre is very much interested, they are also working very well. as it happens with any drug there are some side effects but they are working in clinic. here is some of those compounds, natural compounds shown which

are histone circulation activated, they work as inhibitors, whether for methylation changes, different compounds have been identify and belongs to different categories. short chain fatty acids, later on in histone, for different enzyme, different forms have

been identified. so same thing all genes epigenetics and genetics going on simultaneously, especially in pharmacy, they are making it a specific enzymes, so those kinds of research is also going on. in combination therapy they are working very well, and sometimes

in those patients where no therapy is working, if you treat with epigenetic inhibitors, that works like radiosensitizers and patients start getting response. and they can attack or work on those genes enrolled in cell cycle differentiation, migration, now i will give you

two examples, trials are going on at nci and in this case 35% are there who did not give response to any kind of therapy. it was mixed, and day 10, that was followed and they started giving response. similarly another batch valproic acid, more than 50 patients were

treated, they were also not giving response. if you go to clinical trial site at nci and put histone inhibitors, you'll get information. histone inhibitors, solid solid tumors or lymphoma or breast cancer or other kinds of

tumors, these are different examples, how many people are enrolled, what is their history, how many are responding or not, that kind of information you can get. at the same time for histone inhibitors, that one for methylation inhibitor, 51

studies going on in this case, it's working in clinic as well. companies like bristol meyer, smithkline, american society of clinical oncology, you will see the potential drugs which may be coming soon. in therapy when you think about genetics, think about

environment and epigenetics also, and these are those four drugs approved by f.d.a., so they are approved and are being used, and these are -- this is one important nonsmall cell lung cancer, efficacy was reported, similarly in this case epithelial tumor cells,

demethylating agents were used. a few examples of different cancers, which genes are affected most, in colorectal cancer at early stage, only those genes are shown which are regulated epigenetically. at the same time to correlate genetics with epigenetics, k-ras

mutation was identified at the same time methylation profiling was done, combined, and then based on that variation of different people who were in cohort about which people did not know which disease they had, whether they have colon cancer other not, those were

identified, so a group of genes have been identified and one thing is called, this phenotype was identified in colon cancer, in colon cancer what happens is left colon behaves differently, left side is different than right side. in this cancer certain regions

are called simple phenotypes, (indiscernible) more reports are coming so people can target those. just based on methylation profiling, healthy people could be distinguished from each other. in lung cancer proteomic markers

were combined, steph bellenski has characterized those techniques, samples, you can take those samples and analyze profiling. this one is for asbestos exposure and mesothelioma, you can study simultaneously and identify high risk population.

in esophageal cancer a number of genes in the same patients colored at different times, as you see some genes get methylated early, some late, some also, based on that percent methylated regions are calculated. methylation can be correlated

with survival, and this example is for pancreatic cancer. the doctor wants to know family history, lifestyle, other kinds of information, how that is used, before something has long standing alcoholic, basic mutation, then that person is at high risk of pancreatic cancer,

that's how it is used. with the treatment followup also this methylation profiling is useful, like shown here for the breast cancer. in 15 or 20% of cancer infectious agents are enrolled, now other cancer coming up, hepatitis b,c, for liver cancer,

h. pylori, they play important role, and in that also complete methyloma of hpv, their methylation epigenetic revelation is important. why you want to understand that, because even if histopathology or other tests are not given, positive or negative, this

asymptomatic healthy areas, you will know high risk of developing cancer. in this different stages, those have been characterized very well, within that early genes or late genes, when they are methylated, how much they are methylated, all that has been

sorted out and that has been very useful. since in 15 to 20%, these infectious agents are enrolled. until specific population methylation will be there. that may be some isolated case but otherwise general screening markers are there like this.

these are for hepatocellular carcinoma, now many kinds of cells during development (indiscernible). mostly psa is used in prostate, below 4 nanograms is the level, it is difficult for a doctor to tell whether a person should be treated for prostate cancer or

not. epigenomics was one company this started this, quest diagnostics acquired that and now do the test. it is much more with much more sensitivity and specificity. in bladder cancer, if you take cells to follow methylation

profiling, here it is shown, exfoliated cells you can use. hypermethylation for different cancer, you can do simultaneously, peter laird identified such techniques that many cancers especially follow-up of treatment for prognosis these technologies are

now, in cancer it is considered that about 30 to 40% cancer you can prevent, by changing diet or other things, some natural food or bioactive food components are such that they have properties of methylation inhibitors, so that area is also coming up. just lie changing methylation,

without adding carcinogens such models have been made for liver at the same time like tomato, apple, coffee, broccoli, tea, all these natural food of grapes, they have been characterized, center of automated medicine, they are food such being characterized,

histone modification properties, methylation properties, these are both. sometimes by changing diet you can do several things, physical exercise or physical activity, lifestyle is important, no smoking is important, came way these are natural compounds, you

hear more about that. i mentioned the road map program, such that in 2006 mandate came 1.5% of budget in such problems, in which more than two institutes are enrolled, that mechanism or that program is affecting many diseases, so in that epigenome

and microbiome two are identified at a time. as i mentioned, like genome, sequence that epigeno is completed, epigenetics in several other diseases also, it is there. in these programs we gave initial money and then later on

instituted developed it further. this is one example, $219 million, we put, and some mapping centers were identified which did methylation, and then specific cancer, kidney, lung, data correlation center was established, some data came to different data centers, and nis

they were kept, data was so much we had to develop our own coordination center. technology development since this changes with time, different points we have more amount. and similarly, no known marker identifies, as last year this

whole road map was colored. if you get a chance to search, you can see those articles that how it was completed and how much it is important. what it has, information about protocol, assay standards, analysis tools, 92 comprehensive epigenome datasets were done.

for genome, you take any cell from any organ site and sequence, it will be the same. for epigenome we knew many kinds of cells we had to do, so 120 kinds of cells for that mapping was done. if you read about that, different kind of findinglities,immunological or

other cells, why we did that at least we have something to compare with. supposing some disease you are seeing some changes, this database will work for that. it has lot of implications, and during that course very highly impactful papers were published

by this program, in "cell," "nature," "science" as shown here, now after nih started that, international level, international human epigenome consortia, that was also initiated, and i represent nih in that, ihec, that map for different cell types, as well as

model organisms they are doing and they are correlating aging as well, epigenetic changes for normal development they are needed, but if some abnormality is there, it is disturbed, so that is going on and they have plans to do at least thousand epigenome, that is going on.

and in our meeting in 2010, now it has developed more, at the time itself so many countries participated, and funding were many, from european union was there, and along with that affymetrix, and "nature" and "science" published about ihep, to know how best to utilize that

this is that program, and exchange of views and data. now why we are interested more and more, we want to use epigenetic markers that whether we can use this as new risk factor to identify why certain risk have more cancers like african-american prostate cancer

is more than white, native hawaiian has less, similarly japanese populations migrate to u.s., once they come here they have colon cancer, there we want to utilize that. the cohort case control studies whether we can improve existing markers or not.

how we can integrate genomic and epigenomics, how can we use this information for better define cancer subcategories. now you must heard about this program, precision medicine initiative, president obama around march or april announced that since everybody responds

differently, science about continue, science should be done in which a person is followed using all these omics or new approaches, proteomics, metabolomics, all those, so we're doing now this precision medicine initiative, and then follow sentinel technologies

whole genome sequences, epigenome, metabolome, all those you do and follow and how that works, that also you can do, read about that in march, the new england journal of medicine, or google if you put precision medicine initiative, that is the way we're utilizing whatever new

information or omics information is coming, and integration or implication in health. these are the collaborators with whom i have worked, and now i'll take questions. >> yeah, equilibrium happens, like histone modifications, when

they are modified, undergo histone has (indiscernible) will be modified. and that also can be methylated, single, dimethylation or trimethylation, methylation of dna goes also, but in disease it depends whether whatever is there, that is affecting

histone, or it is affecting dna promoter. and then they try to diverse, like for genetic we have the repair mechanism, similarly for epigenetic data, normal modification is there but if affected for a longer time then it is there.

so it is not necessary, they happen simultaneously, but it depends on microenvironment and that organ site of those tissues, different cell type, that is the reason that we have done epigenome mapping for different types of cells as well.

sometimes i mentioned about exosomes, they will come in the circulation, circulating cells are there. t we want to use for the most you can tell whether someone has cancer or not cancer, what to tell that you have lung cancer you have to go

to the organ site and then you compare and then you can tell. so systematically. when you go to a doctor he has the history and follow the changes going on. so that's how we want to use. >> thank you.