Sonia S. Kupfer, MD, discusses the impact of anatomic location on colon cancer disparities, and how looking at location might inform on preventitive strategies. She also provides a definition of the proximal and distal colon, and looks at prevelance data rates from the American Cancer Society in different demographics. Also covered are the ongoing and future studies of the role of vitamin D in minimizing cancer risk.
[MUSIC PLAYING] SONIA KUPFER: So over the next 40 or so minutes, I want to take you through a whirlwind tour of how anatomic location might impact colon disparities. And what we learned from looking at anatomical location in particular might help to inform preventive strategies. So just so we're all on the same page, we all know what we're talking about, we define the proximal and distal colon kind of arbitrarily. But basically, proximal is from the splenic flexure, transverse colon, ascending colon, and cecum. And then the distal, it refers to anything distal to the splenic flexure. So that includes the descending colon and sigmoid and rectum. These are, as I said, somewhat arbitrary, and when I talk about these different studies, there's probably a gradient. But for ease of studying and for defining, this is what we're going to be talking about. So why is it even important to think about this kind of thing? So from my perspective, I'm very interested in colon cancer and, in particular, colon cancer disparities. So this is data from the American Cancer Society showing that African Americans, both men and women, have the highest rates of proximal colon cancer. They have the highest rates of overall colon cancer, but specifically also of proximal cancer. And so you can see here from the blue and the green lines representing African American males and African American females. Now, when you talk about cancer disparities, you can run the gamut in what could possibly be causing this. So of course, access to care, uptake of screening-- all of these sorts of things play a role. What hasn't received as much attention has been potential environmental or biological explanations that could potentially also be modifiable. And I just want to make the point before moving on that I'm going to be talking about cancer here. But of course, for those interested inflammatory bowel diseases, you have the obvious differences in the proximal and distal colon, in terms of, for example, risk of ulcerative colitis, where some individuals will just have disease in the rectum and some will have disease throughout. So some of these studies may also impact other gastrointestinal diseases. So what is the risk for proximal cancers? And there have been a number of large-scale studies to try and address at least the epidemiology. So this was a large database study done by Dr. Lieberman, who uses a really nice endoscopic database. So he has a large number of patients and basically can look at distributions among men and women and across races. So he, in this case, looked at 21,000 individuals who had polyps that were considered large. He doesn't have pathology, so we don't know for sure that these are adenomas. But we can presume that as they get bigger, most likely they have some at adenomatous component. So the things that were independent risk factors for proximal large polyps was African American race, being female, and then as you got older. And so this is something that is a theme that keeps coming up, is that African American race is associated with proximal location. So another study that was done at BU looked at their endoscopic database, but here they could actually look at pathology. So they included a large number of African Americans and non-Hispanic whites. And they make the point that this is a safety net hospital. So these are individuals where hopefully you've gotten rid of some of the other barriers, such as access. And so you can actually look at what might still be residually different between different races. And so you can see that rates of advanced neoplasia by race and location, that African Americans had substantially higher rates of proximal advanced neoplasia, which is obviously what we're interested in, in terms of cancer prevention. It didn't quite reach statistical significance. But certainly the trend was in a very suggestive direction. And this is probably in part because of sample size. But again, the theme keeps developing, that proximal neoplasia seems to be higher in African Americans. And in this case, maybe you've gotten rid of some of the other barriers. So we're really potentially looking more at biological differences or environmental differences. So was a recent study in Gastrointestinal Endoscopy that looked at the Kaiser Permanente health care system in Southern California. And basically they looked at the rates of post-screening colon cancer. So these are individuals who get cancer after they've been screened. And they looked at it for colonoscopy versus sigmoidoscopy. And they found that African Americans had the highest rate of post-screening colon cancer if sigmoidoscopy was the initial test. And you can see that the hazard ratio here was 1.71. So again, this is yet another other large-scale study that underscores the impact of proximal cancers in this population. And there's other studies that I don't have time to go through that basically also show the same trend. So if we take a step back and we ask, well, so what could be the differences in this proximal cancer risk in African Americans? We have to start with-- how about different modes of carcinogenesis or different pathways? So just to remind you, as we go from normal mucosa on the left to an invasive cancer on the right, there are two main pathways that we can arrive at a cancer. So the first is chromosomal instability, which is the most common carcinogenic pathway. It probably takes-- we don't exactly know this, but somewhere between 8 and 20 years to go from normal to cancer. And you can see that there are different steps along the way with loss of APC, then KRAS activation, and then loss of P53. And so we call these cancers microsatellite-stable, and that will become obvious when I talk about the next pathway. And these tend to be left-sided. Not always, obviously. But if you think about the sort of prototypical disease that we think about, a hereditary disease called familial adenomeatous polyposis, where individuals get thousands of polyps and develop cancers very early, the distribution of that tends to start on the left side and then move proximally. So that's kind of what we think is how chromosomal instability works. And then the other pathway, which is about 15% of cancers, is the mutator phenotype. And here on your path from normal mucosa to cancer, what's happening is that you're losing mismatch repair. And that can either happen because you have a germline mutation-- in Lynch syndrome, for example-- or in individuals as they age, they can get methylation of the MLH1 promoter. Along the way, you have a mutator phenotype, so you're mutating a lot of different genes, including BRAF, P53, and a whole host of other mutations. So this mutator phenotype may take a shorter time to develop cancer, two to eight years. And this is then defined as microsatellite-instable, or MSI-High. These tend to be right-sided cancers and also tend to have a better prognosis. And so one of the thoughts is, well, could it just be that there are different pathways of carcinogenesis among African Americans that might predispose them to developing proximal cancers? So we looked at this in a consortium that was started a few years ago called The Chicago Colorectal Cancer Consortium. It's a number of different medical centers in Chicago. We had incident cases as well as prospective cases. And you can see the breakdown of the colorectal cancers. And we had a nice number of African American cases. Those that had biological samples was fewer, but still substantial when you think about what series are out there where you also have biological data. And in addition to tumors, there was clinical data questionnaires and, as I mentioned, biospecimens. So when you look at the differences between African American and whites in terms of cancer, the things that pop out are, again, we're seeing this proximal location. So this is in line with other studies that I showed previously. And that was significant. And then interestingly, there was a higher rate of lymphocytic infiltrate in African Americans, and then a higher stage at presentation. And then when did a basically multivariate regression analysis to try and look at what are independent risk factors for proximal microsatellite-stable tumors, we found that lymphocytic infiltrate was one of those predictors, so with a high odds ratio. And this is just to show you an example that a lymphocytic infiltrate basically are T-cells that are seen within the tumor. And these could either be a response to the tumor itself or, potentially, inflammation that is then leading to cancer. And so we're left with this question. Why do African Americans have a higher risk of proximal colon cancer? And microsatellite instability, which tends to be on the right side, is probably not what's going on here. In fact, there was a recent study using samples from North Carolina, from a population-based study, that found that perhaps African Americans even have lower rates of microsatellite instability, and that might lead them to having more aggressive cancers. So I don't think microsatellite instability is really the answer. So let's move on to what other possibilities there could be. So could it be different genetic predisposition? Could gene expression be different? Could the environment be different? Or could the response to the environment be different by race? How about the microbiome? And then really, the million-dollar question is, how do any of these differences predispose to developing cancer? So I'm just going to take you through a few of these and show you some of the work that I've done, but also some of the work that others have done in this area. And just to note that not a lot has been done looking at differences by race, and so the studies are few and far between. So some of the studies I'm going to show you are just looking, in general, how do the right and the left colon differ? And we can see what we can learn from that. So just to start off, the right and the left colon are actually two different organs. They come from two different embryological origins-- so the proximal colon from the midgut, the distal colon from the hindgut. So they have different origins. They have different blood supplies. And in fact, I'm not going to go through all this, but the last step, actually, for forming the colon is really the fusion of the two parts, so where the proximal and the distal colon come together. So what about functional differences? So they clearly have different roles. So the proximal colon tends to be more saccular. We all know that the wall of the cecum is quite thin compared to the sigmoid, for example. It's an acidic environment. It is a reservoir. Its main function is water and electrolyte resorption. And then it also has a role in fermentation. And this is where the bacterial composition probably plays a big role. The distal colon is more tubular, has more of a neutral pH. It's more of a conduit. Less resorptive function, although there still is some resorption. And then has 30% less fermentation than the proximal colon. So there are functional differences. So could these differences somehow be different among races and perhaps also predispose to cancer? So what about genetics? So this is kind of like the 40,000-foot view of predisposition. And so it's a place to start, because it's something that we can test. So when we did this looking at our large series of DNA from African Americans, we looked at differences between rectal cancer and colon cancer. We didn't separate it by proximal and distal. And we found one association on chromosome 11q23 that was associated primarily, or only, with rectal cancer in African Americans. And that actually was something that had been seen in other populations. So probably genetics might be playing a role. However, no other consistent results have been found with anatomic location. And probably, here, the issue is that as you know, these odds ratios are very low. These effect sizes are low. So you need hundreds of thousands of people to really show significant differences. So it may just be an issue of power. What about gene expression? So no study has looked at differences in gene expression between African Americans and European Americans. But there was a study published in 2003 that looked at differences between the proximal and the distal colon. And they basically took colonoscopic biopsies from individuals with the inherited Lynch syndrome, as well as individuals who were sporadic-- meaning they were normal, sorry. And all of these were presumed European Americans. So this is normal tissue, I should point out. Not from tumors, but normal tissue, just looking at differences between the right and the left. They also looked at embryonic tissue as well, just to see when these differences are sort of established in development. And basically, they ran a microarray and saw that there were obvious differences between the right and the left. In fact, they could, with 100% accuracy, define a tissue as being either from the right or the left colon. And you can see that there was, in terms of genome Y, there was 70% increased gene expression in the distal colon and 30% increased in the proximal colon. So they had some examples of perhaps this might be contributing to cancer on one side or the other but really couldn't make any definitive conclusions. They also found, interestingly-- I don't have it up here-- but that among the embryological tissue, embryonic tissue, that a lot of these changes happened during development. So there must have been other forces, be they epigenetic or environmental, that might define right versus left. And so the one thing that they did say is that the genes tended to cluster for the proximal in what you would expect-- so ion, anion, and electrolytes-- as well as some other interesting categories that you see listed here. And that the distal, the genes were much more involved in cell-cycle control, DNA replication, DNA damage and repair. Again, these are sort of more correlations and clustering analyses. So what this exactly means for cancer risk is still yet to be determined. So what about the microbiome? As we all know, the colon has the largest burden, or carries the largest number of bacteria. And if you look at the proximal versus distal colon in terms of the microbiome, you can see here that the proximal colon has a high concentration of substrates. As I mentioned, it has an acidic pH and rapid bacterial growth. Versus the distal, which has a lower substrate availability, has a neutral pH, and a relatively slower bacterial growth pattern. And so could the microbiome somehow be different among different populations? And could that influence risk of cancer? So this was a study that was done by Yunmei Wang in [INAUDIBLE] lab. Some of you might have been involved with that and can chime in if I'm explaining this incorrectly. But basically, they took wild-type mice and they looked at the microbial composition in the proximal versus distal colon, looking under germ-free conditions and then in normal conditions. And they had mice from two different labs. And then they also compared the microbial composition to stool samples that were obtained from these mice. And so basically what this is showing is the different composition and relative abundance of the bacteria. And of course, as I think other studies have borne out is that really individuals clustered together primarily, but also that the location of where the mice were caged also made a difference. So these three over here were from University of Chicago, and these were from another lab. The interesting thing is when they then plotted sameness-- so using this principal components analysis-- they interestingly found, again, the blue here represents the University of Chicago. This is Jackson Lab here. So found that where the mice were housed was the first big separation. But then, you might not be able to see it, but here, interestingly, the proximal colon often was closer-- so more similar to the stool sample versus the distal colon. And that was a pattern that was seen at both labs for all mice. And over here is just a graphical representation of the diversity, which was higher in the proximal colon and was similar in the stool, which is this bar here, compared to the distal colon. So this is some of the things that I had already described previously. And then what they did is they took RNA samples from- it was for RNA as well as for protein-- from the proximal and distal colon of these mice under germ-free conditions and normal conditions, and then looked at gene expression. And in this case, they were interested in looking at toll-like receptors. And they looked at TLR2, which is in Figure A, and at TLR4, which is in Figure B. And they found that under germ-free conditions, the proximal and distal colon, there was no difference. And actually, the expression was quite low. But when microbes were introduced, you can see that for TLR2, the proximal colon expression was significantly higher compared to the distal colon expression, and that for TLR4, it was the opposite effect. And so the conclusion here was that-- and I should say that this also was seen with protein levels here. And so this suggests that the microbiome does shape the host response and that there are differences in the proximal and distal colon-- underscoring again that the proximal and distal colon microbial composition is different. So as I said, we don't really know how this is in humans. The few studies that have been done have looked at individuals who have undergone bowel prep, so they may have disrupted their microbiome. And so something that had been started in [INAUDIBLE] lab and that we recently acquired additional samples was for this healthy colon study. And so the idea here is to take individuals who have not undergone a bowel prep, perform a colonoscopy, and obtain samples from various parts of the colon, including a biopsy so we can look at gene expression. And you can see the different places that we've gotten samples from. We have recruited seven individuals. The majority of them, four, are African Americans, and three, European Americans. We've obtained aspirates, brushes, and then biopsies from these different locations. And the data acquisition and analysis is currently in progress. So it'll include microbiome analysis, host gene expression, and then perhaps also some metabolic parameters, including bile acids and butyrate quantification. So hopefully this will give us some idea, in a native state without a bowel prep, how the microbiome might shape the host response. And this is a first step in terms of understanding, perhaps, how microbial composition could differ between populations, although it is understandably a small sample size. So then what about diet and lifestyle factors? So clearly there have been a number of associations with dietary factors and colon cancer. And so these include diets high in red meats or processed meats, cooking meat at high temperatures, low intake of fresh fruits and vegetables, physical inactivity, obesity, tobacco use, and heavy alcohol use. But if you look by anatomic location, there's really no consistent associations. And probably part of that is that a lot of these studies just lump colon cancer in one big pot and don't really look at specific associations by anatomic location. And so that's probably in part why we're not able to find consistent associations. And the other thing is that we don't always include a lot of minority populations in these kinds of large epidemiological studies. So I think the jury's still out, and probably we need better and bigger studies to try and address if diet and lifestyle could be different. An interesting study, though, that was done by Dr. O'Keefe at the University of Pittsburgh was to ask the question, why do African Americans have much higher rates of colon cancer compared to native Africans? So what are the forces at work there that could predispose? And he hasn't-- so I should say what he did here is that he recruited individuals, both in Pittsburgh, that were of African descent and European descent, and did diet recall and colonoscopy and blood samples and a lot of different measures. And then similarly, he had age- and gender-match controls in South Africa. So that was the native African population. And he looked at differences, shown here as differences between the African and the American diet. And so again, this might be a little bit small. But the things that I wanted to point out is that native Africans had much lower daily kilocalorie intake. There was also significantly lower fat and saturated fat intake. And then some of these other things-- there were some vitamin differences. He also thing included in this breath hydrogen and breath methane tests. So this is a-- lactulose is given, and then you measure the exhaled hydrogen and methane. And this is a measure, sort of an indirect measure, of the bacterial load. And he found that breath methane was really significantly higher in native Africans. What maybe didn't get as much attention in this study is the difference between African Americans and European Americans. They weren't really powered to look at differences. And in fact, if you look at the diets, they were relatively similar. The main difference is, again, if you look down here, in terms of the fasting breath hydrogen and breath methane, that while native Africans had lower hydrogen and higher methane, there were also differences between here-- these are the African Americans and these are the European Americans. And I'll show you another graph where this is shown graphically. But basically, the African Americans had the highest of all hydrogen breath expulsion parts per million and had the lowest breath methane, suggesting that the bacterial compositions in African Americans are different from native Africans, but also different from Caucasian Americans. So this is the graphical representation of the hydrogen lactulose breath test results and the methane breath test results. And you can see here that the native Africans, for the hydrogen, were pretty similar to the Caucasian Americans, whereas there was a significant difference here in exhaled hydrogen among African Americans. And similarly, while not as big of a difference, you can see that across time, that the African Americans had the lowest exhaled methane. And then they also were able to look at colon proliferation in these individuals. And here they did look somewhat by location. So you can see that across the board, regardless of where you're looking in the colon, that native Africans had the lowest amount of proliferation in their colon. And interestingly, they describe it as, the native Africans' colons looked so wonderful. They were like in awe of these colons, which was a funny description. But then when you look at African and European Americans, shown here, you can see that the confidence intervals are crossing, or the error bars are crossing, but that there is at least a suggestion that perhaps there are some differences by different sites in the colon. Again, this was a small study. It wasn't powered to look at this as an endpoint. But it does raise some interesting possibilities, in terms of differences between individuals who are eating pretty similar diets. So in the last part of my talk, I'm going to be focusing on vitamin D. And really, the reason that I do that is vitamin D, of all cancers, has shown the greatest protection in colon cancer. And so there's a lot of interest in using vitamin D as a chemopreventive agent. And so this is just one study. There's a lot of studies out there that summarize epidemiological studies looking at serum vitamin D level and reduction in colon cancer odds ratios. So you can see here that as you increase your serum vitamin D, your odds of developing colon cancer go down. And this has been shown for adenomas as well as for cancers. And as you probably know, African Americans of all US populations have the lowest serum vitamin D levels. So this is data from the NHANES that looks over here, across the various age groups, that of all US populations, African Americans have the lowest. And this is likely due to skin pigmentation, and less so probably due to what they take in orally. However, the only way that we can supplement it in an effective way and quickly is by giving oral supplementation. So there's a lot of interest in using vitamin D to potentially address some of the colon cancer disparities. However, it's not really known how these two might interact. And so we've looked at this in a number of different ways. So just to remind you of vitamin D metabolism, very briefly-- so vitamin D3 is converted in the liver to 25-hydroxy vitamin D, which is what we measure in the serum. And then it's converted in the kidney to 1,25, which is the active metabolite and has what we think are most of the biological actions. 25 can also be converted locally, by the colonic epithelium by an enzyme called CYP27, into 1,25. And so far less is known about what the role of this sort of local conversion has in terms of cancer prevention. The 1,25 is then intercellularly bound to be the vitamin D receptor, which then couples with RXR and binds DNA and is responsible for gene expression of a number of different targets. And we think in cancer, that the main ones are in terms of its effect on differentiation, apoptosis, growth inhibition, and decrease in proliferation. And then the only other thing I want to mention is that the first metabolism step is mediated by this by this enzyme, CYP24A1. So we first looked at this by asking, again, a more of a genetic hive view question-- could variants in the vitamin D receptor be associated with colorectal cancer? And could these differ between populations? So that would be sort of one explanation as to why perhaps individuals could be at higher risk and, if they have less vitamin D, that that could be exacerbated. So basically we looked at a large series of African Americans from the University of Chicago, from North Carolina. We also had Europeans from a large series in Spain. And basically, this is called a Manhattan plot, where you're looking at location on chromosome 12 down here, which is where the vitamin D receptor is located. And then the negative log 10 p-value. So as you go up on the scale it becomes more significant. And when you're looking at the number of tests that we did, the threshold for significance is way up here. So as you can see, clearly none of these variants reached statistical significance. So we concluded from this that it's probably not variation in the vitamin D receptor that is responsible for, at least in part, the increased risk of colon cancer. So in a follow-up study to this, we looked at other genes in the vitamin D pathway. So this includes the GC protein and CYP24A1, as well as a number of other ones. And there was some suggestion-- and here we looked by location in the colon. And there was some suggestion that these two variants, which are quite rare in frequency. Well, not that rare, but to 8% to 10% frequency in the African American population-- that they might predispose to left-sided colon cancer. However, when we adjusted for the number of tests that were done, these didn't reach statistical significance. So again, this may be an issue of power. And so it may be that the more variance you have within the entire vitamin D pathway, the more at risk you are of not getting the protective effects of vitamin D. But there's really, as yet, not a clear association with location in the colon. So I'm not sure that this is going to be the full answer. So then coming back to the vitamin D metabolism, we then ask the question, well, so if it's not genetic variation within the vitamin D receptor or any of the other pathway genes, could it be that people, individuals, have different responses to 1,25 that we can measure in terms of gene expression? And so this basically bypasses a lot of looking at all of these from a very high point of view and really gets down to, you know, what is the gene expression? What is the vitamin D actually doing? And so here, the gene expression could be influenced by genetics, but it could also be influenced by different environment. And so it's kind of a nice phenotype to look at. So in order to do this, we wanted to have a model in which to study this. And so what we did is came up with a short-term in-vitro culture system, or ex-vivo, however you want to term it. And basically, we published this this year. But basically, we found that we could maintain normal colonic architecture in these ex-vivo cultures. We basically take biopsies and keep them alive for six to eight hours. And then we also did some Ki-67 staining and found that the normal proliferation within the crypt base was maintained. So we thought that at least it's not a perfect system, so obviously you can't look at much longer response. But it was a starting point, to start asking these kinds of questions. And so the study design was basically getting biopsies. And in this case, we started just with the rectosigmoid colon. And we treated them in this culture with vitamin D or ethanol and then extracted the RNA and ran them on expression arrays. And So we first looked at overall transcriptional response to 1,25 vitamin D. So this is called a volcano plot. And basically, it's plotting meaningful change of gene expression on the x-axis, and then again, the negative log p-value. So as you go up, you're getting more significant gene expression differences. And then as you go across, you're looking at downregulation and upregulation. So we were very happy, actually, to find that the top upregulated genes in our ex-vivo culture system were CYP24A1, which we would expect, since that's the first enzyme to break down 1,25. And then another one, CD14, which is a known vitamin D target. And then when we looked across the genome, we found that there were 883 differentially expressed genes at a low FDR, which is a False Discovery Rate, of less than 1%, and that they were sort of equally split between those that were upregulated by 1,25 and downregulated. And then we asked the question, because we had almost equal numbers of African Americans and European Americans, we asked, well, so even to the same amount of 1,25, are there interethnic differences? Are there genes that respond differently to 1,25 by race? And we found that there were a number of genes here that reached statistical significance after we adjusted for the number of tests that we were doing, and they're listed here. And this linear model included various types of covariates that we thought could influence gene expression. So that included serum vitamin D, age, gender, and then principal components, which is meant to deal with other variation that we can't measure. And so just as an example, this was an interesting gene that popped out as showing differences in response in the two populations. This is EphA2. This is a receptor tyrosine kinase. And in the normal colon, this protein is expressed at the top of the colonic crypts. And so there's been a suggestion of its involvement in crypt maturation. And it's also been found to be overexpressed in colon cancer. So this was a nice candidate of perhaps something that could explain at least some differences in response to vitamin D. And the box plot here just shows you that in European American biopsies that were treated with ethanol, the response to vitamin D, which is shown here, was greater than what's shown in African Americans. So more work is needed now to understand how these genes might be involved in predisposition to colon cancer, but certainly there were some nice candidates. So then that has now raised more questions and some hypotheses about how can we link vitamin D, anatomic location, and African Americans? So I have a number of different hypotheses that I've come up with recently and kind of given them some names, but this is a work in progress. So the regional response hypothesis-- so as I mentioned, what has been found is that colon cancer risk is associated with serum vitamin D. So you could say, well, maybe it's just differences in serum vitamin D that are affecting risk of colon cancer, and if you gave everyone vitamin D, that all these differences would disappear. And so the question here is, if that's true, are there differences in response to vitamin D in the proximal versus distal colon? And I'm going to show you some data that at least has suggested that yes, this is true. And if this is true, then is there a dose response? So can we find that there are different genes that respond differently by different doses of vitamin D in the proximal and distal colon? And this is not, as yet, known. Then a separate hypothesis, but not mutually exclusive from the first, is that are there in interethnic differences? So even if you gave the same amount of vitamin D and everyone's vitamin D level was the same, could there still be differences in response to vitamin D for a number of different reasons, be they genetic or other environmental factors? And so I call this the interethnic response hypothesis. And as I've shown in the previous few slides, I think that this may actually be going on, as well. So there could be differences between African Americans and European Americans in response to the same amount of vitamin D. And this is at least true-- and I showed that data-- for the distal colon. We don't know yet for the proximal colon. And then the final hypothesis, which I think we have to consider, is a microbial response hypothesis. So we know that vitamin D and the microbiome, there's a cross-talk. And so the question here is much broader. How do microbiome-vitamin D interactions differ by anatomic location and race? And how does this then potentially impact risk for colon cancer in the two different locations? So let me show you some preliminary data that addresses some of these hypotheses. So this is a study, a very early study, where we took four female African Americans-- and again, this is using our ex-vivo culture system. And we looked now at differences in response between the right and the left colon. And interestingly enough for our top-responding genes-- in this case, CD14-- we found that yes, in fact, the right and the left colon differ in their response to vitamin D, where the right colon in this gene is downregulated, whereas in the left colon, it's upregulated, and that's significant with only four individuals. And CD14 is a monocyte marker, but it's also expressed on the colonic epithelium. So could this be reflecting, again, a certain microbial environment where you want to downregulate CD14, which basically acts along with TLR4 to induce a response to bacteria? And perhaps in the last colon, where it's not as important or has a different role, that you see this upregulation. So these are open questions that I don't have answers for, as yet. When we looked at CYP24A1, the results were at least suggestive but not statistically significant. So again here, you can see that in both segments of the colon, there's upregulation of CYP24, which is what we would expect because of the metabolism. But you can see that the degree of upregulation in the right differs from the left. And again, this is a trend, and we obviously need more individuals. But again, suggests that vitamin D acts differently in the right versus the left colon. And then I just want to make the point-- one question could be, well, is it just because vitamin D receptor expression is wildly different in the two segments of the colon? And I'm sorry I don't have the error bars, but the p-value is not significant here, so that it's pretty neutral in terms of expression in the right and in the left colon. And what I'm showing here is just in ethanol. So this is not a response to vitamin D. This is just kind of the surrogate for baseline vitamin D receptor status. And I've looked to see if other studies have investigated this, if there is a known gradient in the human colon of vitamin D receptor expression. And the studies are all over the place. So I got the geodatabase to work, but it looks like there's really no gradient that I can completely tell-- although some people have suggested that vitamin D receptor expression might be higher in the right colon. So then one of the problems, as I said, for the ex-vivo culture system is that it's really very short-term. So we can only look at six-hour expression. When you're trying about gene expression, it's a direct transcriptional target. So six hours might be OK to look at early responses, but we really can't look at later responses or other measures-- so cellular phenotypes, like proliferation. So this is where colonoids would be extremely helpful, because you could get colon tissue from various populations, from different places in the colon. And you could grow these up and you could you treat them for longer periods of time, and look also at things like proliferation and apoptosis. So this is just to show you one of our early successes. We've been having trouble recapitulating this. But basically, this is what these colonoids look like on day five. They have still this spherical shape. And then by day 15, they're starting to have these little out-pouchings. And by 22, it's just a whole glob of cells. And basically, this is a mini colon epithelium that's growing. And then you can separate these out and keep them growing. They can be frozen and can be used in the future. So this really holds a lot of promise for asking and investigating the kinds of things that we're interested in. However, this technology is just still somewhat in development. But we're hopeful that soon we're actually going to get it to work for a long period of time. However, the core here has gotten mouse colonoids to work very well. So I'm going to show you some data from mouse colonoids, because those are much more consistent. So basically, these are mouse colonoids it's from APC Min mice-- two of them, so it's still small numbers. But again, asking the question about vitamin D response by location in the colon. And what's very nice here is that the pattern seemed to reflect what we're seeing in our human tissue, so that CYP24, while upregulated in both the proximal and the distal, there's more upregulation in the distal colon in response to vitamin D. And similarly, CD14, we see much greater expression in response to 1,25 in the distal colon versus the proximal colon. And again, I just show this to show you that it doesn't seem to be a difference in vitamin D receptor expression at baseline that's influencing these differences. So again, raising the very interesting hypothesis that it may be the microbiome in the local environment that impacts the response to vitamin D. The other thing I just wanted to point out about the mouse colonoids, in contrast to our human samples, is that you get much more robust fold change. So you may not have seen it, but our upregulation in our ex-vivo cultures, like 1.5, twofold increase. And here you're getting huge relative fold change. So it seems like a much more stable kind of a culturing system to ask these types of questions, so we're very eager to extend this now to human samples. And so I'll just leave with a few open questions. So how does differential vitamin D response impact colon carcinogenesis by anatomic location and race? So I think there's some leads, but we don't really have this whole thing worked out. And then importantly, what are the interactions between vitamin D and the microbiome by anatomical location and race? And I think it'll be important, again, to ask these questions in a diverse population and really looking at different locations in the colon. And then finally, the million-dollar question would be, well, can we use vitamin D supplementation or other drugs that mimic vitamin D activity to alter risk by anatomic location? So could we envision a time where we just treat African Americans with vitamin D and see if that reduces their proximal colon cancer risk, and thereby hopefully reduce disparities? In the process, hopefully we'll learn a lot about vitamin D in the colon, which will be shorter-term goals than this loftier long-term goal. And so our ongoing and future studies involve genome-wide differences in response to vitamin D in the proximal versus distal colon. So we're getting a larger sample size than we currently have. We're optimizing the human colonoids to study response by location and race. As I mentioned, the healthy colon study, we have those samples and are waiting now for the analysis so we can start to make some of the correlations. And it'll be interesting to see, for example, how at baseline the microbiome differs in the right versus the left, and how gene expression in those locations is shaped, potentially, by the microbiome. And then finally, a randomized trial of vitamin D supplementation in African Americans is currently ongoing. Marc Bissonnette is the PI. And so this will hopefully give us, at least for the distal colon, some markers of how vitamin D supplementation might impact and might be influenced also by the microbiome. And so this work is done with a lot of people, so I want to acknowledge, in particular, Brandon, who unfortunately couldn't be here today. He's a technician who's been doing all of the ex-vivo and colonoid studies. The Di Rienzo Lab in Human Genetics, with whom I work very closely, David Witonsky, who is a bioinformatician, and the Di Rienzo Lab members who listen to my talks every few weeks and give good ideas. Candace and Jean, who have the colonoid core and have really helped to get that going, and have provided us with the mouse colonoids. Kristen, and Katie before her, who have recruited a lot of these patients. Mark and Jann for really helpful discussions. My collaborators in the Colorectal Cancer Consortium, my former mentor Nathan and Xavier. And then funding from the NCI, the DDRCC, and then the American Cancer Society. So with that, I'll finish. And thank you for your attention. [APPLAUSE]
