[Music] can you talk about risk and gene expression profiling there's um other ways in which we can talk about risk and one of them is this gene expression profiling and you probably will hear a bit more from from dr morgan which can you know it's another way to look at at prognosis and try to separate those patients that actually may have a more aggressive form of myeloma so that's one one of the ways there's a there's actually a commercial assay that's available too for doing this through this you know sky company that can test for for the high risk markers now one one of the things to think about with these two things is that uh again number one prognosis is not determinate it's not a black and white you have this you're going to do well you have this you're going to do bad we're still going to work together through those things and and uh you know the the the other thing is that these are only approximations because there's many other things that come into play you know if you come to the clinic and a person has you know severe diabetes and heart failure it's gonna be a lot harder to get going through the through the completion of the treatment so all of that plays out on how we're able to propose different treatments uh for for patients what is gene expression profiling or gep gene expression profiling turns out to be a really useful tool it's quite hard to do but i think it's now commercially available gene expression profiling doesn't look at the protein it looks at that rna that takes the information from the dna to the protein and you can use it to do a variety of things you can use it to classify myeloma into six broad groups makes it very easy to do that but what it was classically used for was to make prognostic schools and there's one called the gp 70 which is basically a 70 gene school that predicts 15 of people with myeloma who really don't do well with current treatments that are or should be the subject of clinical trials we should be doing those trials which are called risk stratified trials to really try and improve the outcome for that group because the reality is in the last 20 years their outcome really hasn't improved hardly at all in contrast the low risk patients are now doing really well so there are other tests there's one called the emc 92 which is a very similar test the problem has been accessing those tests and so only some centers have taken the trouble to establish them and so most people don't have that information i think it will change with these dna panels because they're easier they're more robust people are investing in the technology and so i'm hopeful that that's going to be the way that we risk stratify patients in the future can you explain what a gene expression profile is one way to look at that is also instead of just looking at the dna code is to look at these rna molecules these are the molecules that are expressed from the genes so the genes express the rna the rna then is translated into the protein the more rna you make the more protein you get if we can try to understand the characteristics of tumor cells and how much of a gene is being expressed that could be useful information too and what i'm showing here what's called a heat map let me explain that we now have the capability of looking at how every gene in the cell is being expressed and we have about 17 000 genes in our genome so there's about seventeen thousand genes that give us all the functions we need to be a human being different genes are expressed in different cells so a muscle cell will express genes a liver cell a plasma cell they'll express their sets of genes but what we really wanted to understand was how does a normal plasma cell express genes compared to a myeloma plasma cell they're both plasma cells you would expect them have certain genetic characteristics in common but the genes that they express may be different in the tumor cell than in the normal plasma cell and the heat map that i'm showing here demonstrates every gene that i'm looking at is a row so a different gene each row is a different gene every column is a different sample and you'll notice on the left side there's a label there that says npc those are normal plasma cells and you'll notice there's a set of genes that are red that means they're expressed at high level there's a set of genes that are green they're expressed at a very low level but what we really want to do is compare that then to the myeloma gene expression pattern for those same genes and look what happens in myeloma all those genes that are red in normal plasma cells aren't expressed very well in myeloma all those genes that are not expressed much at all in normal plasma cells are now highly expressed in myeloma so we're understanding a pattern here there's a pattern of genes that are changing in their expression that distinguish a myeloma plasma cell from a normal plasma cell and that tells us something about the biology of this disease now look at those myelomas and the columns and you'll notice that even though there's a block of red in the block of green those two columns are not identical in every myeloma brings home my point every myeloma is different so even though there are some common genes that are often upregulated or genes that are down regulated in their expression every myeloma is a little different and really we need to understand that in fact you could take that pattern there and you could say i've got a pattern where else do we see patterns that distinguish outcomes one of those is in a grocery store barcode so i'm giving you an example here of a pattern these are two grocery store bar codes they have some things in common some things that distinguish them it's a barcode look at the bottom of the barcode you'll notice that the bars all look very similar that means those two products in the grocery store probably share something in common but look as you move up the pattern the pattern starts shifting which says wait a minute those two barcodes probably are defining two different products in the grocery store but something that they have in common because the pattern is similar turns out one of those code the scans says well this is wheat bread and one of them is white bread they share a common feature they're both bred but they are distinct one is white one is wheat let's think about that application of a barcode that might distinguish a patient who will respond to a therapy versus a patient that doesn't respond to a therapy and can we in fact use a scanner to scan a pattern of gene expression that will distinguish a response and a non-response for a particular therapy and that's work that we've done in other labs have done as well so here's an example of a tumor one myeloma this myeloma was being treated with borteson and this is just a laboratory treatment of the myeloma and you can see in this graph what you're seeing is a hundred percent of the cells are alive when we started but we started giving the this tumor in the lab or velcade which a lot of you are familiar with and you'll notice that at 100 percent as we increase the concentration of the drug we had fewer and fewer cells surviving so that by the time we got to a relatively low concentration of drug we killed all the myeloma cells that would tell me that that myeloma right there had a pretty good response to bortesmid because we killed all the cells now the question is if i test other myelomas what happens well uh oh what you can see here is i've got a lot of myelomas that look like that first one where they die very quickly in the drug but look at all the myelomas that aren't responding very well that's a demonstration of the heterogeneity of myeloma from different patients some of those patients myelomas responded killing all those tumor cells to bortismid very quickly but there are a lot of myelomas that didn't respond well to that at all how can i tell the difference what if i could generate a barcode of a heat map of a gene expression that might distinguish a responder from a non-responder so what we did is we collected on the left side all the guys that responded and on the right side of this graph all the myelomas that didn't respond and then we put them in these groups and said let's look at the gene expression pattern of responders and non-responders and ask the computer if it can identify a set of genes that will distinguish those two and here is a result i now have another heat map but my heat map is not distinguishing normal myeloma from normal plasma cells from myeloma plasma cells the pattern on the left are all the myelomas that responded to the drug the pattern on the right are all the myelomas that did not respond to the drug that could be useful information so for example if a new patient walks in the door and i say you know what before we even treat you let's do an expression analysis of your tumor and if we generate a heat map of one tumor and it looks like that you say gee that pattern matches the ones on the left that are responders this will be a good drug for you to try if your pattern look look like the one on the right which were all non-responders you might say you know based on your gene expression pattern you might not respond to the drug maybe we should find other drugs that you will respond to so gene expression profiling is a way of looking at all the messenger rna that a myeloma tumor makes it can be done in many different ways sort of the traditional way has been something called microarray which is one genomic test for looking at it a more modern way would be through next generation sequencing but either way you do it you you still get the gene expression profile which is a list of all the genes in the cell and how much of each one there is and that has been done and it's been shown that it really serves to divide up myeloma patients into different groups which correspond pretty closely to the fish in fact but it gives you more information than the fish so it doesn't only tell you about the you know the chromosome translocation or you know you know maybe gains or losses of chromosomes but it really tells you about the behavior of the cell you know is it making a lot of genes associated with cell division which would be something we'd associate with a poor prognosis and in fact you can you can get much more sophisticated than that and develop prognostic indices that can be used to say okay this patient's got high-risk disease because the tumor has the expression of all these different genes there are two tests the sky 92 and the my prs but there have been issues with insurance reimbursement and i believe that my prs is no longer available i do not believe that the sky 92 is widely used but it is a useful test and can tell you a lot of information about the tumor how is gene expression profiling or gep done gene expression profiling can be done um with what's called microarray and that's the test like the my prs or the sky 92 and it's sort of somewhat older like 10 10 years old whereas most people now who are doing gene expression at least for for research would use sequencing which is a more robust technique i don't think there are any commercially available sequencing tests for gene expression in myeloma the way it works with gene expression profiling and myeloma is that we actually take a genomic approach meaning we look at all the genes and we don't just look at a select few and mind you of all those genes there's typically only a few of them that are really interesting but we still look at all of the genes and but then for instance for the sky 92 test they focused in on just 92 out of the 30 000 to to make the score that they used to determine prognosis and it was a similar number for the my prs test often you'll see gene expression profiling summarized with a heat map with green and red sometimes blue and yellow you know the green often is low the red is high same with blue and yellow and what you're doing in that case typically is you're taking just myeloma patients so you don't have normal cells in there you don't have other tumors you just have myeloma patients and you're looking within it within a group of myeloma patients how do the genes differ in their expression and you'll find that you know certain patients have a very characteristic profile they all express the same kind of genes and we can that's how we group patients into different biologic types and those biologic types often correspond with the genetics so you know a classic example would be the 1114 chromosome translocation well there's a gene on chromosome 11 called cyclin d1 that gets turned on by that translocation and so when you look at the gene expression profiling you see that in those group of patients there's sky-high expression of cyclin d1 can gep help guide which treatments to use that's sort of the holy grail of sort of genomic testing is to identify models that'll tell us which patients will respond to which drug um and there's very provocative data from dr chapman in the united kingdom who identified a profile and if depending on how patients scored they were either likely to respond to proteasome inhibitors like velcade or kiprolis or they were likely to respond to imminence like revlimid uh or thalidomide and that signature that he used i think was like six genes a very small number of genes and so he's i think looking to validate that in another set of patients to see if it's real and then if it is it could be a useful tool but currently there isn't anything that we can we don't have a practical test right now