Session 9: The Immunology of Organ Transplantation
Transcript of Part 2: Taming and Tracking the Human Alloresponse
00:00:14.12 Hi, I'm Megan Sykes. 00:00:15.23 I'm a professor at Columbia University, where I'm the Director of the Columbia Center 00:00:20.28 for Translational Immunology. 00:00:22.11 Today, I'm going to tell you about some of our research on taming and tracking the human alloresponse. 00:00:30.11 So, while we've made great advances in the field of transplantation, the success is 00:00:38.06 still limited by: one, the complications of the drugs that we use to avoid rejections; 00:00:44.18 and secondly, by chronic rejection, which remains a problem, even with the use of 00:00:50.04 optimal immunosuppression. 00:00:52.07 So, in our field of transplantation, the induction of immune tolerance has become a major goal, 00:01:00.00 because this state of tolerance would overcome both of these limitations. 00:01:06.03 What do I mean by tolerance? 00:01:08.12 Tolerance implies long-term graft acceptance without immunosuppressive therapy. 00:01:14.21 But importantly, with an otherwise intact immune system, that can recognize 00:01:20.13 foreign antigens and protect one from infections. 00:01:25.21 Successful tolerance induction, really, in my view, requires intentional, planned 00:01:33.02 immunosuppression withdrawal, achieving tolerance in a high proportion of individuals. 00:01:37.28 It's been known for some time that the rare person removing themselves from immunosuppression 00:01:44.03 may not reject. 00:01:45.14 The vast majority do. 00:01:47.07 But there are examples of tolerance induction through that... that rather dangerous and 00:01:55.06 inefficient process of immunosuppression withdrawal. 00:01:59.18 I'm referring, when I say tolerance induction is successful, to a process where it works 00:02:05.23 in most people, and it's intentional. 00:02:09.06 Okay, so how do we get to clinical trials of tolerance induction. 00:02:14.03 Of course, it begins with experimental models. 00:02:16.20 But if you look in the literature, you'll find actually hundreds or even thousands 00:02:22.05 of models of tolerance in rodents. 00:02:25.11 Unfortunately, almost none of have been successfully applied in humans. 00:02:30.18 And there are a number of steps that we need to take before we go to humans. 00:02:34.20 Because if... if we're going to... if we think we have tolerance, we're removing the standard of care. 00:02:41.26 We're removing the chronic immunosuppressive therapy that is otherwise used to prevent 00:02:47.13 rejection. 00:02:50.00 This is necessary if we want to have a normal... normally function immune... functioning immune system. 00:02:56.24 So, that's why one of the main reasons why we want tolerance, so that we don't have to have 00:03:00.25 that chronic immunosuppression. 00:03:03.13 But if we're going to remove that standard of care, we need to know that it works. 00:03:07.18 And this requires, first of all, in stringent rodent models, showing that it works. 00:03:13.09 And that usually involves extensively MHC-mismatched skin grafts, which are very difficult to 00:03:19.25 get acceptance of. 00:03:21.10 Secondly, we need to have other animal models. 00:03:24.27 It's very difficult to go from a mouse to a human in withdrawing immunosuppression. 00:03:31.00 We really need large animal models that are closer to humans. 00:03:34.14 And thirdly, it's desirable to have some experience with whatever drugs or agents you're 00:03:41.06 going to use in your tolerance protocol in other clinical settings. 00:03:45.23 As it happens, the approach that we've... 00:03:48.07 I and my colleagues have worked on for many years, involving hematopoietic cell transplantation 00:03:53.19 and induction of mixed hematopoietic chimerism, has come closest to meeting these criteria, 00:03:59.26 and has gotten into clinical trials. 00:04:03.19 So, if we're going to use hematopoietic cell transplantation to induce tolerance, 00:04:11.00 we can't use it in the usual way that it's used, for example, in a patient with leukemia or lymphoma. 00:04:17.12 In those settings, the transplant, of bone marrow or other hematopoietic cells, is done 00:04:23.15 after very heavy treatment of the recipient aimed at eradicating as many of the cancerous cells 00:04:29.28 as possible. 00:04:31.25 In a patient who doesn't have cancer, who needs an organ transplant, we can't justify 00:04:37.22 doing such toxic treatments. 00:04:39.16 Instead, we have to develop a way of preparing that recipient for their bone marrow transplant 00:04:45.23 -- that has the capacity to re-educate their immune system and induce tolerance -- 00:04:51.16 that is far less toxic. 00:04:53.05 It has to not eliminate the recipient's bone marrow cells, so that if you failed to get 00:05:00.14 engraftment of your donor bone marrow you still have a functioning bone marrow and normal hematopoiesis. 00:05:08.02 And nevertheless, this treatment has to be strong enough to overcome the very strong 00:05:13.14 immune resistance in the recipient to the donor. 00:05:17.14 And we know we've succeeded in doing that if we achieve mixed chimerism. 00:05:21.00 And what I mean by mixed chimerism is coexistence of donor and recipient hematopoietic elements. 00:05:29.18 The donor ones aren't rejected, and so you see them in the circulation, 00:05:34.10 and the recipient ones have not been killed off by the conditioning, 00:05:38.24 and so you see them together with the donor cells. 00:05:42.02 Now, I mentioned in my introductory lecture that graft-versus-host disease is 00:05:47.11 a major complication of HLA-mismatched and even -matched hematopoietic cell transplantation. 00:05:53.26 It has some benefit in treating leukemia... in malignant diseases, but is absolutely 00:05:59.19 an unacceptable complication to introduce in somebody who doesn't have a malignant disease 00:06:05.23 but needs an organ transplant. 00:06:07.26 So, this is a big challenge: to cross HLA barriers, avoid GVHD, overcome the host resistance 00:06:16.07 of the donor, and do all of this with minimal toxicity. 00:06:21.17 It took many years for our groups at Mass General Hospital to reach a point 00:06:27.07 where we could actually try that. 00:06:28.22 And our pathway to clinical trials of tolerance induction using hematopoietic cell transplantation 00:06:37.11 actually involved rodent models, shown on the left side of this slide, 00:06:44.28 some involved in studies of treating leukemias and lymphomas 00:06:48.24 that ended up bringing us to a mixed chimerism approach 00:06:51.09 with non-myeloablative conditioning. 00:06:53.23 Meanwhile, our studies in rodents had shown that we could induce organ tolerance with 00:06:58.18 a non-myeloablative regimen for mixed chimerism induction. 00:07:03.04 And our studies in leukemias and lymphomas actually led us to an approach that 00:07:07.26 we could try in patients. 00:07:09.12 Because, in that setting, you can sometimes go directly from rodents to humans because 00:07:17.11 the patients who will try an experimental protocol have failed all other possibilities, 00:07:22.08 if they have a very advanced malignant disease. 00:07:26.11 And so we had the opportunity to try this mixed chimerism approach in some of 00:07:31.18 these patients who... in whom it was used as a platform for immunotherapy. 00:07:36.19 However, on the organ transplantation side, we didn't go straight from mice to humans. 00:07:41.21 We actually had a non-human primate model in the middle, which is very similar to humans 00:07:47.12 in how it responds to transplants, and was a very critical step in allowing us to do 00:07:53.17 bone marrow transplants for tolerance induction in patients with no malignant disease. 00:07:59.24 The protocols that we tried at Mass General for inducing tolerance in these patients underwent 00:08:06.22 several iterations. 00:08:07.25 A total of ten patients were transplanted under these three protocols. 00:08:14.15 The first one... and these were all supported by the immune tolerance network of the NIAID... 00:08:21.04 and the first one is shown here, and all of them have in common that they utilized non-myeloablative doses 00:08:28.00 of cyclophosphamide; local irradiation to the thymus; 00:08:32.19 a monoclonal antibody against CD2 that is given to deplete the recipient and donor T cells in vivo; 00:08:39.06 and then a combined kidney and bone marrow transplant on day 0. 00:08:44.12 And the post-transplant immunosuppression has involved a calcineurin inhibitor 00:08:48.25 for a period of 9-12 months. 00:08:52.02 The second iteration of the protocol brought in some steroids for a very short period 00:08:58.20 after the transplant to avoid an engraftment syndrome, as well as treatment with a B cell depleting agent, 00:09:06.16 rituximab, to avoid antibody-mediated rejection. 00:09:11.20 And the third iteration involved even more rituximab, several treatments with it, 00:09:17.11 and a slightly longer period of steroid treatment, but was otherwise similar. 00:09:23.05 This work has been published already, in these two papers shown here. 00:09:27.16 And I'll just very briefly summarize the clinical results. 00:09:31.14 Out of ten patients, seven were removed from immunosuppression successfully for periods 00:09:37.28 of years. 00:09:39.07 And their... their outcomes are shown here. 00:09:41.28 These first four patients had the first two regimens in the first ITN trial. 00:09:49.07 And that first patient is now more than 14 years off immunosuppression, doing very well. 00:09:55.27 This second patient is more than eight years off. 00:09:59.04 These other two patients had several years of no immunosuppression, but eventually returned 00:10:04.08 to immunosuppression, unfortunately, due to a low-grade chronic rejection. 00:10:12.13 In the second trial, where we added more B cell depletion to avoid this low... 00:10:19.03 very low-level antibody-mediated rejection that led to these patients returning to immunosuppression... 00:10:25.20 in the second trial, we added more rituximab, and these patients actually are now more than 00:10:30.12 seven years post-transplant and doing very well without any antibody-mediated rejection. 00:10:36.17 Now, I mentioned this term, mixed chimerism, where the donor and recipient cells coexist 00:10:42.14 in the patient. 00:10:43.14 And this is illustrated here for... for these four patients in the second trial. 00:10:50.21 And what you see is, in multiple bone marrow lineages -- lymphocytes, granulocytes, monocytes, etc -- 00:10:57.09 we see a contribution of the donor in the circulating population, but only 00:11:05.02 for a very short period of time, for a period of 1-2 weeks. 00:11:10.06 So, this is very transient chimerism, and it's very different than what we can achieve, 00:11:16.06 for example, in rodents, where the chimerism persists forever. 00:11:19.16 But we knew from our non-human primate studies, and from some other studies in patients 00:11:25.10 with malignancies, that transient chimerism, when achieved with a kidney transplant at the same time, 00:11:30.18 could lead to tolerance. 00:11:32.24 So, in the lab, we've been studying what the mechanisms of this tolerance are. 00:11:38.13 And one of the things that we've observed in the patients who got this treatment is 00:11:42.24 that there's a marked enrichment for what we call regulatory T cells among the T cells 00:11:49.15 that initially come back after the transplant. 00:11:52.22 So, this slide here shows you the percentage of T cells in the CD4 lineage that have 00:12:00.26 this regulatory cell phenotype over time post-transplant. 00:12:04.04 And each type of symbol represents an individual patient. 00:12:09.08 And what you see is that these percentages of regulatory cells in that first year post-transplant 00:12:16.17 are very, very high compared to the pre-transplant level, which you see here is very, very low. 00:12:24.06 It's normally a very small percentage of CD4 T cells, but it goes way up after the transplant. 00:12:28.28 It eventually comes down to normal over a period of about a year. 00:12:33.18 Now, in... if you look at the actual absolute numbers of those regulatory T cells, 00:12:39.17 you can see, over here, that they are in fact depleted at one week post-transplant, but that 00:12:45.27 within a couple of weeks they come pretty close to the pre-transplant baseline level. 00:12:51.06 In contrast, the non-regulatory T cells, the conventional CD4 T cells 00:12:55.27 -- here, you see their pre-transplant values, and here you see post-transplant -- 00:13:00.12 they remain low for a very, very long time. 00:13:03.06 And that explains why we see such a marked enrichment of the regulatory cell population 00:13:08.28 among those CD4 T cells. 00:13:11.17 And these are bona fide regulatory T cells, because FOXP3 is the transcription factor 00:13:18.16 that is a master regulator of the Treg program. 00:13:22.20 And demethylation of the... of the FOXP3 region in the genome is a hallmark of a... 00:13:29.12 of a bona fide Treg. 00:13:30.24 And we can see here, in this bottom plot, that the level of TSDR methylation 00:13:36.27 actually correlates very well with the percentage of regulatory T cells we detect. 00:13:40.28 So, these are really Tregs. 00:13:42.10 Why are they so enriched? 00:13:45.02 Well, it looks like there's a few things going on. 00:13:48.24 But the main one is probably that they are expanding in the periphery after we 00:13:56.14 deplete the T cells, the vast majority of the T cells. 00:13:59.24 So, it seems that they... some of them are spared from the depleting T cell antibody, 00:14:06.09 and the ones that remain undergo a lot of proliferation. 00:14:09.12 And we can see this here in this upper right plot, where we're looking at the percentage 00:14:14.27 of these regulatory T cells that express Ki67, which is a marker of proliferating cells. 00:14:21.00 And you can see that it's way up at 1 and 2 weeks post-transplant compared to baseline levels. 00:14:28.13 So that... and this is something that happens when you deplete lymphocytes, 00:14:32.12 that the ones that remain undergo proliferation. 00:14:35.01 So, that's one of the major mechanisms of this enrichment. 00:14:40.02 Well, what role do these Tregs play in the tolerance that we see in these patients? 00:14:44.10 Well, we can look at this by looking at alloresponses in in vitro mixed lymphocyte reactions and 00:14:51.23 cytotoxic T lymphocyte assays. 00:14:54.18 And what we have found in all of our tolerant patients is that they become very hyporesponsive 00:15:01.11 toward their donor. 00:15:02.15 Here, we're looking at their proliferative response, in red, to the donor, 00:15:07.04 and in blue, to a third party individual to... that they're not tolerant to. 00:15:12.27 You see that the response to the donor is markedly reduced. 00:15:16.26 And this this particular sample was taken one year after a transplant. 00:15:21.10 But what you see over here is that when you remove the regulatory T cells, the Tregs, 00:15:26.15 from that patient's cell population, and now measure the response to the donor, 00:15:31.08 a response is revealed. 00:15:32.27 So, this indicates that those regulatory T cells were suppressing the anti-donor response. 00:15:38.22 However, in this same patient, when we went back and did similar studies at 8 and 18 months post-transplant, 00:15:45.08 the result was a bit different. 00:15:47.09 Now, you still see that patient is hyporesponsive to the donor -- you can't even see a red symbol here, 00:15:54.01 because there's no response to the donor -- but now depleting the regulatory T cells 00:15:59.23 doesn't really reveal any anti-donor reactivity. 00:16:02.17 There's still very little response there, even though we've enhanced the response 00:16:06.27 to third party by depleting Tregs. 00:16:09.07 So, this suggested to us that we no longer depended on regulatory T cells to see 00:16:15.15 this donor-specific hyporesponsiveness at 18 months post-transplant, and that something else 00:16:21.13 might be going on. 00:16:22.23 And we hypothesized that maybe that large number of alloreactive T cells present 00:16:27.18 prior to the transplant had been deleted of those that recognized the donor. 00:16:33.14 So, to summarize these functional assays, tolerance in seven of seven cases in this study 00:16:40.22 was associated with the development of donor-specific unresponsiveness in 00:16:45.20 these in vitro assays. 00:16:47.16 We saw a regulatory T cell enrichment in all of these patients after the transplant. 00:16:52.26 And in some assays, we could see that those regulatory T cells were playing a role in 00:16:57.05 suppressing the anti-donor response. 00:16:59.20 But when we looked late, more than a year post-transplant, we did not see a role 00:17:04.18 for regulatory cells in suppressing that response in the longer term. 00:17:10.13 So, this made us think that perhaps the long-term tolerance might be deletional. 00:17:16.27 Okay. 00:17:18.04 Well, there's... deletion is one possibility, that the donor-specific T cells actually got 00:17:23.13 eliminated. 00:17:24.13 So, here, if we think of the red cells as the ones that recognize the donor, 00:17:29.21 if they're deleted they're actually gone from the immune repertoire after the transplant. 00:17:33.16 But another possibility is that they're anergic, meaning that they persist but they 00:17:39.23 simply don't respond when stimulated through their T cell receptor. 00:17:43.01 They're in the state of anergy. 00:17:45.03 And functionally, with the kinds of assays that I've just told you about, 00:17:48.09 there was no way to distinguish those two possibilities. 00:17:51.25 So, we really wanted to find a way of distinguishing them, and actually seeing what happens to 00:17:58.07 alloreactive T cells. 00:18:00.08 And this is a big challenge, because T cells recognizing a given set of alloantigens 00:18:06.15 on an MHC-mismatched donor represent a very large number of cells and specificities. 00:18:14.10 It's thought that 1-10% of T cells respond to a given donor. 00:18:19.09 And this is thought to be due to recognition of thousands of different peptide/MHC specificities 00:18:25.13 on an allogeneic MHC. 00:18:28.23 And all the studies that had been done show that there's no particular predictable 00:18:36.01 dominant immune response. 00:18:37.10 And so there's really no way to pick out clones that you can track over time with tetramers, 00:18:43.11 for example. 00:18:44.13 So, the approach that we used was really facilitated by the development of a technology for 00:18:50.10 high-throughput sequencing of the hypervariable region of the T cell receptor beta chain, 00:18:56.22 known as the CDR3. 00:18:58.23 And this is the... the part of the T cell receptor that is most specific for the peptide 00:19:05.27 that is seen by that T cell receptor. 00:19:09.06 And this hypervariable region is formed by the rearrangement of the V, the D, and the J segments 00:19:16.08 of the T cell receptor, along with N insertions that give it additional diversity. 00:19:22.17 And a commercially available platform was developed for actually sequencing up to 00:19:30.18 millions of these unique sequences simultaneously. 00:19:35.15 And this led us to hypothesize that high-throughput CDR3 sequencing of transplant recipient's 00:19:43.13 donor-reactive T cells prior to a transplant would allow us to identify 00:19:49.13 the repertoire of TCRs... of clones that recognize the donor's alloantigens. 00:19:55.11 And that we could then carry out such sequencing after the transplant to track the fate 00:20:00.17 of those T cells. 00:20:02.23 Using this approach, we actually succeeded in developing a method for tracking 00:20:09.20 a patient anti-donor T cell response and obtaining evidence for clonal deletion 00:20:16.02 as a mechanism of tolerance in the patients that I've been speaking about. 00:20:20.10 And what this assay involves is taking patient lymphocytes, whole PBMCs; 00:20:27.14 labeling them with a CFSE dye, which is a fluorescent dye that dilutes each time the cell divides, 00:20:34.21 so the level of CFSE staining is a marker of how much a given cell has divided; 00:20:40.22 and stimulating those in a co-culture for six days with donor PBMCs that have been irradiated, 00:20:47.02 so they can't divide, and also labeled with a different dye, a violet dye; 00:20:53.16 co-culturing for six days, collecting the cells, and specifically sorting the recipient, the responder cells, 00:21:02.00 that have divided -- those that have diluted their CFSE dye -- and separately sorting, on a FACS sorter, 00:21:11.24 CD4 and CD8 cells of that recipient that have divided in response to donor antigens. 00:21:20.09 And what we did is then subjected each of these populations, these divided cell populations, 00:21:25.18 to high-throughput sequencing of the T cell receptor CDR beta... 00:21:30.15 CDR three region. 00:21:32.01 And also did the same thing on CD4 and CD8 cells from the unstimulated T cell population 00:21:37.24 of that patient. 00:21:38.24 And this is all prior to the transplant. 00:21:41.22 And then we can actually define a sequence as alloreactive... donor-reactive if there... 00:21:49.03 if it's expanded more than fivefold in this mixed lymphocyte reaction compared to 00:21:55.05 its frequency in the unstimulated population, shown over here. 00:22:01.09 So, this... this is a way of identifying a repertoire, a set, of T cells that we call 00:22:08.00 a fingerprint of the anti-donor alloresponse. 00:22:13.24 And this is what... when we developed this assay, we tested it on our tolerant patients. 00:22:19.14 And what we found was that in all three of the tolerant patients who we studied 00:22:25.16 there was a significant... a statistically significant decline in the frequency of donor-reactive 00:22:32.18 CD4 and CD8 cells in the circulation, over time, after the transplant. 00:22:37.22 And we saw this in all three patients compared to the pre-transplant level. 00:22:44.03 We also had one patient in this trial who failed to achieve tolerance, who got the same treatment 00:22:49.22 but rejected the kidney after the immunosuppression was withdrawn. 00:22:56.06 And what you see here is that this patient did not show any significant reduction 00:23:01.24 in the frequency of anti-donor clones in the circulation. 00:23:06.06 So, it suggests that this method actually distinguishes the tolerant from the non-tolerant patients. 00:23:14.13 We've also tested this method on patients who don't get a tolerance-inducing regimen 00:23:19.15 but who just get a kidney transplant with conventional immunosuppression. 00:23:24.25 And some of our typical results are shown here. 00:23:28.22 Interestingly, we don't see any reduction... these are two different individuals, two different recipients, 00:23:35.04 looking at the frequency of donor-reactive clones over time. 00:23:41.02 Here's pre-transplant, and here's post-transplant. 00:23:43.17 You can see that in both of these patients there's a statistically significant increase 00:23:48.11 in the number of circulating donor-reactive CD4 clones after the transplant, showing you 00:23:54.05 the stimulation of the immune response by the transplant. 00:23:58.09 So, that helps to validate this assay as showing us something very biologically meaningful. 00:24:06.01 And what we were able to conclude from this study is that high-throughput deep CDR sequencing 00:24:11.24 of recipient's donor-reactive T cells pre-transplant enables identification of a specific set of 00:24:19.23 donor-reactive T cells. 00:24:22.01 And these donor-reactive clones can then be tracked in the post-transplant period to 00:24:27.19 tell us something about what's going on immunologically. 00:24:32.14 And our studies indicate that we are identifying biologically relevant T cells with this pre-transplant MLR, 00:24:40.12 because their frequency goes up in a conventional transplant recipient 00:24:44.06 after the transplant. 00:24:46.11 And our data suggests that in the tolerant patients who get 00:24:49.05 combined kidney and bone marrow transplantation, 00:24:52.11 deletion of donor-reactive T cells is a long-term mechanism of tolerance. 00:24:56.28 And in studies I didn't have time to go through, this deletion seems to be the result 00:25:02.25 both of global T cell depletion with the conditioning and specific exposure to the donor antigens. 00:25:09.20 In contrast, expansion of circulating donor-reactive clones is detected in conventional transplant 00:25:15.22 recipients. 00:25:17.07 And so far, this deletion analysis has outperformed the functional assays that I referred to earlier 00:25:25.14 because functional studies actually showed donor-unresponsiveness in the patient 00:25:29.28 who failed tolerance in addition to those who succeeded, suggesting that that patient 00:25:35.16 was demonstrating anergy, at least under the conditions of our in vitro assay, whereas this 00:25:41.19 deletional assay actually distinguished the tolerant from the non-tolerant patient. 00:25:47.06 Okay. 00:25:48.24 I'm just going to spend a few minutes talking about how this TCR tracking method can also 00:25:54.26 be used to better understand what's going on within an allograft. 00:26:00.20 And this study actually involves patients who receive intestinal transplants. 00:26:06.26 And at our center, at Columbia, our patients are actually followed by surveillance biopsies 00:26:13.28 of the intestinal allograft through a stoma that is created at the time of transplant, 00:26:19.28 because the symptoms of rejection can be quite nonspecific. 00:26:23.15 And doing these surveillance biopsies is a way of making sure that we're on top of 00:26:31.08 a rejection if it does occur. 00:26:32.25 So, it's looked at histologically. 00:26:35.03 So, this approach has actually given us an opportunity to look not only in the circulating 00:26:42.04 cell populations of these patients but also at what's going on in the graft biopsy specimens 00:26:49.00 in real time. 00:26:50.00 And we've taken advantage of this to look at... within the graft at the 00:26:57.01 alloreactive T cells that we've identified with the method I just spoke about. 00:27:01.13 So, just to give you a bit of background, intestinal transplant outcomes are... 00:27:06.28 are not as good as we would like them at this point. 00:27:11.07 And there's a lot of rejection that occurs. 00:27:13.25 And particularly in patients who get intestinal transplants alone. 00:27:18.13 Some patients don't just get an intestinal transplant; they get a liver transplant with it, 00:27:22.19 because their liver has failed for a variety of reasons, often due to chronic TPN 00:27:29.06 used to treat the intestinal failure. 00:27:32.10 But what you see in this slide is that the patients who get multivisceral transplants 00:27:36.11 -- liver, pancreas, stomach, and everything along with the intestine -- 00:27:40.07 actually have lower rejection rates than patients who get intestines alone. 00:27:44.16 So, we hypothesized that this might have to do with the interplay of graft-versus-host 00:27:52.04 and host-versus-graft reactivity in these patients. 00:27:56.20 Now, I should say that the intestine comes with a very big load of lymphocytes. 00:28:01.25 And it's known that intestinal transplantation can cause graft-versus-host disease. 00:28:07.13 So, we've looked at this in our patients, and hypothesized that, in fact, 00:28:14.26 lymphocytes from that graft may go into the circulation, and that may be a marker of patients who 00:28:22.11 won't have rejection. 00:28:24.15 And this can occur without graft-versus-host disease. 00:28:27.16 And in fact, when we investigated this hypothesis, it turned out to be the case. 00:28:32.01 What we found in a lot of these patients, and particularly those who got the multivisceral transplants, 00:28:38.03 shown with the circles... we saw very high levels of donor chimerism 00:28:44.00 in the circulation, spontaneously, without any bone marrow transplant. 00:28:48.24 And most of these patients did not have graft-versus-host disease. 00:28:52.13 In 14 patients shown here, 8 showed this mixed chimerism in the circulation, but only one 00:28:59.05 had graft-versus-host disease, and it was very mild and self-limited. 00:29:03.05 Interestingly, many patients did not have this macrochimerism. 00:29:08.20 And we define macrochimerism as more than 4% donor T cells in the circulation at its peak. 00:29:15.20 And it's most commonly the patients getting intestinal transplants, the ones here 00:29:20.00 with triangles, did not get macrochimerism. 00:29:23.03 But what was striking is this association down here. 00:29:26.16 We observed that the patients who have this macrochimerism, as we've defined it, 00:29:31.14 have much lower graft rejection rates than those who don't have macrochimerism, 00:29:37.00 consistent with the hypothesis that we started out with, that this macrochimerism may protect the patients 00:29:44.19 from rejection, and can occur without graft-versus-host disease, as we've seen. 00:29:50.06 Now, as I mentioned, we also study the grafts, and we can do flow cytometry with 00:29:55.08 multiple parameters to actually look at the replacement of donor cells by the recipient within the graft. 00:30:02.20 And we can look at all sorts of different subsets of cells within those mucosal biopsies 00:30:08.07 over time. 00:30:09.22 And this is just an example of one such biopsy, where we're looking at different lymphocyte subsets, 00:30:14.13 and we have specific markers that... this antibody that goes up in the y-axis distinguishes 00:30:22.10 recipient cells, whereas those that are negative for that antibody are donor-derived. 00:30:26.23 So already, you're seeing mixed chimerism in this intestinal graft. 00:30:32.20 And what we noticed is that there was a highly variable rate of replacement of 00:30:39.09 the donor T lymphocytes that come in that graft by recipient T cells from patient to patient. 00:30:47.19 And that there was an association with rejection, and development of donor-specific antibodies, 00:30:54.28 with more rapid replacement by the recipient. 00:30:58.21 So, this part of the slide is showing you the rate at which... the percentage of recipient cells 00:31:06.15 in these different T cell subsets. 00:31:08.16 And you can see that it's quite high quite early on in these patients who undergo rejection. 00:31:15.02 Each line is a different patient. 00:31:17.09 In contrast, patients who don't have rejection, or have a DSA-negative rejection, 00:31:22.24 the rate of replacement of donor T cells by the recipient within the graft is very slow. 00:31:28.20 Very interesting. 00:31:29.20 And this part of the slide on the right just represents this in a different way, 00:31:35.04 making the same point. 00:31:36.16 So, what... what's going on here? 00:31:39.21 It looks like donor cells appearing in the blood and recipient cell... recipient T cells 00:31:46.00 not replacing the donor cells in the graft is associated with less rejection. 00:31:51.20 Well, our original hypothesis was that, in fact, all of this is reflecting a balance 00:31:57.17 between the graft-versus-host response, caused by T cells in the graft, 00:32:02.18 and the host-versus-graft response, which is a systemic immune response. 00:32:08.02 And using this T cell receptor tracking method that I just spoke about, we could actually 00:32:13.08 look at this in both directions: in the graft-versus-host and host-versus-graft directions. 00:32:18.08 So, this is the same assay that I mentioned earlier, and now we're doing it in both directions 00:32:23.28 on pre-transplant donor and recipient cells to identify the GvH and the host-versus-graft 00:32:31.23 T cell repertoires. 00:32:33.24 And now we can interrogate these biopsy specimens for these clones. 00:32:39.13 And what we found was quite striking. 00:32:42.11 In the early period post-transplant, particularly in those patients who have slow replacement 00:32:48.17 of their graft T cells by the recipient, there's a marked expansion of graft-versus-host reactive 00:32:55.13 T cells within the graft. 00:32:57.20 It's... they're much more frequent than what we see in the lymphoid tissue prior to transplant, 00:33:04.06 for example, shown in the black bars. 00:33:06.09 So, there's a huge expansion of GvH-reactive CD4 and, over here, CD8 cells 00:33:12.08 in the graft compared to what was in the donor lymphoid system. 00:33:18.14 And this was interesting. 00:33:19.23 And we wondered why that was. 00:33:21.27 Our analyses, our flow cytometric analyses, included analyses of the antigen-presenting cell 00:33:29.20 populations in the graft. 00:33:32.23 And what we found was, in contrast to T cell replacement rates, which were extremely variable, 00:33:38.24 as I showed you... 00:33:40.11 patients with rejection tended to have rapid replacement of T cells by the recipient, 00:33:46.01 whereas those without rejection had slower replacement... in contrast to all that, all of the patients 00:33:51.21 showed quite rapid replacement of antigen-presenting cells, of myeloid cells, 00:33:57.03 with a dendritic cell phenotype by the recipient. 00:34:01.05 This one is at day 16. 00:34:03.11 Almost all the APCs are recipient-derived, whereas the T cells are still mostly donor. 00:34:08.25 There's very few recipient ones. 00:34:11.00 So, that was quite a uniform finding: early replacement of donor APCs by the recipient. 00:34:17.26 And that could explain this expansion of graft-versus-host-reactive cells in the graft. 00:34:23.07 They're having recipient antigens presented to them on these recipient APCs that enter the graft, 00:34:29.11 expanding this GvH response. 00:34:32.04 And we think this is very protective. 00:34:34.20 We have additional studies that I don't have time to take you through, but that GvH response 00:34:40.26 also goes into the circulation, contributing to the macrochimerism. 00:34:45.18 The other thing that we can see with this technique... we can interrogate the biopsies 00:34:50.06 for host-versus-graft clones, and what we see is something that hasn't been shown before. 00:34:56.07 During a rejection of a human allograft, there's a huge enrichment of host-versus-graft alloreactive 00:35:02.19 T cells within those grafts. 00:35:05.26 And that may be obvious, but there's some dogma from the literature that most T cells 00:35:11.02 infiltrating a graft during a rejection are bystanders; they're nonspecific. 00:35:16.02 That is clearly not the case here. 00:35:18.04 A very high percentage of these T cells during a rejection are host-versus-graft reactive, 00:35:24.16 as defined by our TCR tracking method. 00:35:27.13 This goes down as the rejection resolves, but it's still an enrichment for... 00:35:33.02 those host-versus-graft cells persist long-term in these patients, and we think they may pose 00:35:38.27 a constant risk for rejection. 00:35:41.10 So, I'll end here with a summary. 00:35:44.25 We have found that there's a direct correlation between early region and accelerated replacement 00:35:50.09 of donor T cell populations in the graft by recipient T cells that look like those 00:35:56.19 in the circulation. 00:35:57.25 They have a blood-like phenotype. 00:36:00.27 Host-versus-graft clones predominate among those host T cells within rejecting grafts. 00:36:05.12 They persist at lower levels long-term. 00:36:08.07 And what I didn't show you is that they changed their phenotype long-term; 00:36:11.28 they look more like tissue resident lymphocytes, and they seem to seed the entire gut. 00:36:16.20 And we think these pose a constant threat of rejection. 00:36:20.27 Thirdly, in contrast to the highly variable replacement rate of donor T cells by the recipient 00:36:26.06 in the gut, antigen-presenting cell replacement is uniformly rapid. 00:36:30.25 And finally, we think these rapidly immigrating recipient APCs are driving the local expansion 00:36:37.12 and activation of GvH-reactive T cells coming with the graft, and that these 00:36:42.15 may actually control the host-versus-graft-reactive clones, curbing rejection and replacement of 00:36:49.00 donor T cells by the recipient within the graft. 00:36:52.12 So, I'm going to end there. 00:36:54.13 Obviously, I've talked about a lot of different studies, and that's involved a huge number 00:36:59.23 of people, both at Columbia and originally at Mass General, 00:37:05.09 where we did the clinical trials of tolerance induction. 00:37:09.15 And the intestinal transplant studies have involved many people in the lab, 00:37:15.23 but also in... on the clinical side as well. 00:37:18.15 So, thank you very much for your attention, and I'll stop there.