ALBERT

Description: Key Points CMV reactivation fundamentally resets posttransplant CD8 reconstitution, resulting in massive expansion of CMV-specific CD8 Tem. CMV reactivation is associated with defects in the underlying TCRβ immune repertoire.

Description: Abstract 2998 Introduction: GvHD remains the most deadly complication of HSCT despite current prevention strategies. To address the unmet need for better GvHD control, we have created a non-human primate (NHP) model with which to rigorously test mechanism and efficacy of novel therapeutics. In this study, we determined whether a novel combination of mTOR inhibition (with sirolimus) and CD28:CD80/86 costimulation blockade (with belatacept) could control GvHD. Here we show for the first time that these two agents combine synergistically to prevent both the clinical and immunologic manifestations of primate aGvHD. Methods: Rhesus macaque recipients were irradiated (9.6 Gy in 2 fractions at 7cGy/min), and then transplanted with G-CSF-mobilized PBSC from a haplo-identical donor (1–5×108 TNC/kg). Recipients were treated with either sirolimus alone (n = 4, troughs targeted at 5–10 ng/mL), belatacept alone (receiving weekly doses of 20 mg/kg), or combination therapy. Clinical GvHD was monitored using our previously described NHP grading scale (Miller et al., Blood 2010), and multiparameter flow cytometric analysis was performed. Results: Untreated controls (n = 5) developed rapid, severe histopathologically-proven aGvHD and succumbed rapidly (MST = 7 days). Recipients treated with either sirolimus or belatacept alone were partially protected from the clinical manifestations of GvHD. Sirolimus-treated recipients (n = 6) developed predominantly GI disease (with diarrhea but no elevation of bilirubin) and had an MST of 14 days (Figure 1). Recipients treated with belatacept alone (n = 3) developed primarily liver aGvHD (bilirubin rapidly rising to 6–30 × normal with histologically-confirmed lymphocytic infiltration) and an MST of 11 days. In striking contrast, recipients treated with combined sirolimus + belatacept (n = 5) demonstrated neither uncontrolled diarrhea nor hyperbilirubinemia at the timed terminal analysis (1 month post-transplant). We employed multiparameter flow cytometry to determine the immunologic consequences of sirolimus and belatacept on T cell proliferation (using Ki-67 expression) and cytotoxity (using granzyme B expression). We found that the clinical synergy observed with combined therapy was recapitulated immunologically. Thus, while untreated aGvHD was associated with rampant CD8+ proliferation (with 83 +/− 14% Ki-67+ CD8+ vs 4.7 +/− 0.6% pre-transplant), sirolimus or belatacept as monotherapy both partially controlled proliferation (35 +/− 3% and 65 +/− 23% Ki-67+ CD8+ with sirolimus or belatacept, respectively). Combined sirolimus + belatacept dramatically reduced proliferation (to 8 +/− 3%, favorably comparing with 13% Ki-67+ CD8+ T cells using standard Calcineurin Inhibitor/Methotrexate (CNI/MTX) prophylaxis). Sirolimus and belatacept both also partially controlled GvHD-related T cell cytotoxicity. Thus, while untreated aGvHD was associated with excessive granzyme B expression in CD8+ T cells (82 +/− 2% granzyme Bvery high CD8+ cells vs 0.3 +/− 0.2% pre-transplant) sirolimus or belatacept monotherapy also partially controlled cytotoxicity (8 +/− 1% and 35 +/− 1% granzyme Bvery high with sirolimus or belatacept, respectively). Combination therapy dramatically reduced the proportion of these cells, to 1.5 +/− 0.8 % granzyme Bvery high, favorably comparing with 4% granzyme Bvery high using CNI/MTX. The ability of sirolimus, belatacept, or the combination to control Ki-67 and Granzyme B expression closely correlated with survival (Figure 2A, B) supporting a pathogenic role for these highly proliferative and cytotoxic cells in aGvHD pathology. Moreover, significant co-expression of granzyme B in the Ki-67+ cells was observed (Figure 2C) suggesting that dual-positive Ki-67/Granzyme B cells may mark a pathogenic population, amenable to tracking in the peripheral blood. Implications: These results reveal a previously undiscovered synergy between sirolimus and belatacept in the control of primate aGvHD, and provide support for future clinical investigation of this novel prevention strategy. They also identify CD8+/Ki-67+/Granzyme Bvery high dual-positive T cells as a potentially sensitive biomarker of GvHD pathogenesis, amenable to monitoring in either the blood or in GvHD target organs. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 2550 We have developed a novel, MHC-defined rhesus macaque model of total body irradiation-based haploidentical hematopoietic stem cell transplantation. This model has permitted us, for the first time, to perform a rigorous study of the cellular and molecular basis of uncontrolled primate GvHD, and to evaluate the efficacy of a novel, clinically-relevant T cell costimulation blockade-based immunosuppressive regimen to control this disease. We have found that after unprophylaxed haploidentical transplant, severe GvHD developed, which was characterized by rapid clinical decline, and widespread T-cell infiltration and organ damage, with histopathologic evidence of disease in the lungs, the liver, and the GI tract. Mechanistic analysis revealed activation as well as possible counter-regulation, with rapid, CD8-predominant T-cell expansion and accumulation of both CD8+ and CD4+ granzyme B+ effector cells as well as FoxP3pos/CD27high/CD25pos/CD127low CD4+ T-cells. In addition, CD8+ cells downregulated CD127 and BCl-2 and upregulated Ki-67, consistent with a highly activated, proliferative profile. A cytokine storm also occurred, with GvHD-specific secretion of IL-1Ra, IL-18, and CCL4. The combination of CD40/CD28 costimulation blockade (using a monoclonal antibody against CD40 and the CTLA4Ig fusion protein) and mTOR inhibition with sirolimus (Costimulation Blockade and Sirolimus, “CoBS”) resulted in striking protection against GvHD. Thus, at the 30-day primary end-point, CoBS-treated recipients demonstrated 100% survival compared to no survival in untreated recipients. Long-term analysis revealed that CoBS treatment resulted in mean survival increasing from 11.6 to 62 days (p

Description: Abstract 2549 Introduction: Given the emerging importance of sirolimus as a therapuetic for graft-versus host disease (GvHD), it is critical to rigorously define the mechanisms by which this agent impacts T cell immunity after hematopoietic stem cell transplantation (HSCT). Therefore, we have used our novel rhesus macaque model of haploidentical HSCT and GVHD to probe the mechanisms of sirolimus-mediated GvHD prevention when given as a monotherapy. The insights gained from this study will facilitate the rational design of sirolimus-containing combinatorial therapies to maximize immunosuppressive efficacy. Methods: Transplant recipients were prepared with 8Gy total body irradiation and were then infused with MHC-mismatched donor leukopheresis products(n=3, avg. 6.5×108 TNC/kg, 3.4×107 total T cells/kg). Recipients received sirolimus monotherapy (serum troughs 5–15 ng/mL) alone as post-transplant immunosuppresson. Clinical GvHD was monitored according to our standard primate GvHD scoring system and flow cytometric analysis was performed to determine the immune phenotype of sirolimus-treated recipients compared to a cohort of recipients (n= 3) that were given no GvHD immunoprophylaxis. Results: Sirolimus modestly prolonged survival after MHC-mismatched HSCT compared to no immunosuppression (〉19 days versus 6.5 days in the untreated cohort, with GvHD confirmed histopathologically at the time of necropsy). We found that sirolimus significantly inhibited lymphocyte proliferation in transplant recipients: The ALC remained suppressed post-transplant (eg ALC of 0.46 × 106/mL on day 15 post-transplant versus 4.3 × 106/mL pre-transplant, with recovery of other leukocytes: WBC=5.1 × 106/mL, ANC=2.6 × 106/mL). These results suggest that sirolimus can have a profound impact on lymphocyte proliferation, inhibiting GvHD-associated lymphocyte expansion by as much as 200–300-fold compared to untreated controls. Sirolimus had a similar impact on CD4+ and CD8+ subpopulation expansion. Thus, while CD4+ T cells and CD8+ T cells expanded by as much as 300-fold and 2000-fold, respectively, without sirolimus, the expansion of these cells was significantly blunted with sirolimus, with maximal expansion of CD4+ and CD8+ T cells being 4- and 3.6-fold, respectively compared to the post-transplant nadir. Sirolimus-treated recipients also better controlled the upregulation of the proliferation marker Ki-67 on CD4+ or CD8+ T cells. Thus, while untreated recipients upregulated Ki-67 expression by as much as 10-fold after engraftment, (with 〉80-98% T cells expressing high levels of Ki-67 post-transplant versus 5–10% pre-transplant) sirolimus-treated recipients better controlled Ki-67 expression (17-40% Ki-67-high CD4+ and CD8+ T cells post-transplant). While the impact of sirolimus on T cell proliferation was profound, it failed to completely inhibit activation of T cells, as measured by both Granzyme B and CD127 expression. Thus, when effector CD4+ and CD8+ T cell cytotoxic potential was measured by determining expression levels of granzyme B, we found that sirolimus could not downregulate this key component of immune function and GvHD-mediated target organ damage: Granzyme B expression in both CD4+ and CD8+ CD28-/CD95+ effector T cells was unchanged despite sirolimus monotherapy. Down-regulation of CD127 expression, which identifies activated CD8+ T cells in both humans and rhesus macaques, also demonstrated resistance to sirolimus treatment. Thus, while a cohort of recipients that were treated with combined costimulation blockade and sirolimus maintained stable CD127 levels post-transplant, and untreated animals demonstrated total loss of CD127, up to 60% of CD8+ T cells in sirolimus-treated recipients down-regulated CD127, consistent with breakthrough activation of these cells despite mTOR inhibition. Discussion: These results indicate that while the predominant effect of sirolimus during GvHD prophylaxis is its striking ability to inhibit T cell proliferation, sirolimus-based immunosuppression spares some cellular signaling pathways which control T cell activation. These results imply that therapies that are combined with sirolimus during multimodal GvHD prophylaxis should be directed at inhibiting T cell activation rather than proliferation, in order to target non-redundant pathways of alloimmune activation during GvHD control. Disclosures: No relevant conflicts of interest to declare.

Description: We have developed a major histocompatibility complex–defined primate model of graft-versus-host disease (GVHD) and have determined the effect that CD28/CD40-directed costimulation blockade and sirolimus have on this disease. Severe GVHD developed after haploidentical transplantation without prophylaxis, characterized by rapid clinical decline and widespread T-cell infiltration and organ damage. Mechanistic analysis showed activation and possible counter-regulation, with rapid T-cell expansion and accumulation of CD8+ and CD4+ granzyme B+ effector cells and FoxP3pos/CD27high/CD25pos/CD127low CD4+ T cells. CD8+ cells down-regulated CD127 and BCl-2 and up-regulated Ki-67, consistent with a highly activated, proliferative profile. A cytokine storm also occurred, with GVHD-specific secretion of interleukin-1 receptor antagonist (IL-1Ra), IL-18, and CCL4. Costimulation Blockade and Sirolimus (CoBS) resulted in striking protection against GVHD. At the 30-day primary endpoint, CoBS-treated recipients showed 100% survival compared with no survival in untreated recipients. CoBS treatment resulted in survival, increasing from 11.6 to 62 days (P 〈 .01) with blunting of T-cell expansion and activation. Some CoBS-treated animals did eventually develop GVHD, with both clinical and histopathologic evidence of smoldering disease. The reservoir of CoBS-resistant breakthrough immune activation included secretion of interferon-γ, IL-2, monocyte chemotactic protein-1, and IL-12/IL-23 and proliferation of cytotoxic T-lymphocyte–associated antigen 4 immunoglobulin-resistant CD28− CD8+ T cells, suggesting adjuvant treatments targeting this subpopulation will be needed for full disease control.

Description: In this study, we used the rhesus macaque model to determine the impact that AMD3100 has on lymphocyte mobilization, both alone and in combination with G-CSF. Our results indicate that, unlike G-CSF, AMD3100 substantially mobilizes both B and T lymphocytes into the peripheral blood. This led to significant increases in the peripheral blood content of both effector and regulatory T-cell populations, which translated into greater accumulation of these cells in the resulting leukapheresis products. Notably, CD4+/CD25high/CD127low/FoxP3+ Tregs were efficiently mobilized with AMD3100-containing regimens, with as much as a 4.0-fold enrichment in the leukapheresis product compared with G-CSF alone. CD8+ T cells were mobilized to a greater extent than CD4+ T cells, with accumulation of 3.7 ± 0.4-fold more total CD8+ T cells and 6.2 ± 0.4-fold more CD8+ effector memory T cells in the leukapheresis product compared with G-CSF alone. Given that effector memory T-cell subpopulations may mediate less GVHD compared with other effector T-cell populations and that Tregs are protective against GVHD, our results indicate that AMD3100 may mobilize a GVHD-protective T-cell repertoire, which would be of benefit in allogeneic hematopoietic stem cell transplantation.

Description: Graft versus host disease (GvHD) is the most common complication of hematopoietic stem cell transplant (HCT). However, our understanding of the molecular pathways that cause this disease remains incomplete, leading to inadequate targeted strategies for prevention and treatment. To address this, we determined the gene expression profile of non-human primate (NHP) T cells during active and partially controlled acute GvHD (aGvHD), in order to accomplish two goals: 1) uncover important genetic drivers of aGvHD and 2) identify novel, targetable pathways for optimal aGvHD prevention. Utilizing microarray technology, we measured the gene expression profiles of flow cytometrically purified CD3+ T cells from NHP recipients of MHC partially-matched HCT in three treatment cohorts resulting in increasing degrees of survival: 1) no immunoprophylaxis (No Rx, MST = 7.5); 2) sirolimus monotherapy (MST = 17) tacrolimus-methotrexate (Tac-Mtx) dual prophylaxis (MST = 49). Arrays were performed on T cells purified on Day +14 post-transplant (unless terminal analysis occurred earlier due to severe disease). This comparison allowed us to determine the impact of both mTOR and calcineurin inhibition on the molecular pathways dysregulated during GvHD, and to determine which genes and pathways remained dysregulated despite prophylaxis. Pathways identified by this strategy may contain new therapeutic targets unaffected by current immunoprophylactic approaches. We found that the transcriptional profile of donor T cells from HCT recipients with unprophylaxed GvHD was characterized by significant perturbation of pathways regulating T cell proliferation, effector function and cytokine synthesis (Figure 1a). By identifying pathways unaffected by sirolimus or tac-mtx therapy (Figure 1b), we discovered multiple potentially druggable targets not previously implicated in the pathophysiology of aGvHD. These targets prominently included the hedgehog and the aurora kinase A pathways. Utilizing a murine aGvHD model, we demonstrated that pharmacologic inhibition of these pathways could mitigate disease and improve survival (Figure 2a,b). These data provide the first identification of the T cell transcriptome of primate acute GvHD and the hedgehog and aurora kinase A pathways as novel potential targets for prevention of this disease. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

Description: Ex-vivo expanded natural regulatory T cells (nTregs) are emerging as promising cellular therapeutics for the prevention of allograft rejection and graft-versus host disease. However, their widespread translation to the clinic has not yet occurred, in part because of the lack of a clear understanding of their in-vivo quantitative and functional dynamics. Here we have used our translational non-human primate model to answer some of the critical questions related to the survival, stability and immunologic compatibility of Tregs with other immunosuppressive therapies. By infusing escalating doses of CFSE-labeled, autologous CD4+CD25++CD127-/lowFoxP3+ Tregs, we determined, for the first time, that in primates, the accessible, peripheral Treg pool is quite large: ∼74± 13 x106 cell/kg: 〉 20-times larger than the number of Tregs that have previously been infused into patients (Figure 1). Tregs infused in the absence of concomitant immunosuppression had a surprisingly short in-vivo half-life, even when transferred into autologous recipients (T1/2=2.1± 0.3 days): Tregs infused at dose of 20x106 cell /kg were undetectable in either the blood, bone marrow or lymph nodes within 2 weeks after their infusion, despite their initial trafficking to all three sites early after infusion. In addition to their short half-life, infused Tregs quickly underwent rapid phenotypic alteration, with significant loss of both CD25 and FoxP3 expression within 6 days of transfer, suggestive of a potential concomitant loss of suppressive function (Figure 2). By Day +10, only 28.0± 2.7% of infused CD4+CFSE+ cells were CD25+FoxP3+ compared to 74.2± 6.6% on the day of infusion (p〈 0.01). While treatment with tacrolimus was able to rescue the loss of CD25 expression, it did not improve the rapid loss of FoxP3 expression that occurred in the infused cells, resulting in a half-life that was not improved compared to Tregs infused without immunosuppression (2.6± 0.2 days, p = ns compared to no immunosuppression) as well as a similarly rapid loss of FoxP3 expression (25± 9 % FoxP3 expression on day +10 post-infusion). In striking contrast, rapamycin extended the persistence of the infused Tregs beyond day 35 (T1/2 = 3.9± 0.8 days, p〈 0.01 compared to no immunosuppression) and significantly improved the CD25+FoxP3+ phenotype of the adoptively transferred cells, with 70± 6% retaining expression of both CD25 and FoxP3 at day +10, p= 0.01 compared to no immunosuppression) (Figure 2). In addition, compared to animals receiving Tregs either without immunosuppression or in the presence of tacrolimus, rapamycin treatment resulted in significantly more infused Tregs in the bone marrow and lymph nodes when measured one week after transfer (in the bone marrow, 6± 0.7 % vs 1± 0.6 % without immunosuppression and 2.4± 0.4 % on tacrolimus, p〈 0.01 compared to either treatment). These results provide the first quantitative evaluation of the availability of the accessible Treg pool in primates, and suggest that Tregs undergo a significant and potentially deleterious phenotypic degradation after transfer. While tacrolimus was unable to fully rescue this degradation, rapamycin significantly improved both the survival and phenotypic stability of these cells. These results suggest that even natural Tregs may be phenotypically and functionally unstable after transfer, and that adjunctive strategies to stabilize function and half-life, including mTOR inhibition, should be strongly considered when using these cells as an adoptive immunotherapy. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 1008 Regulatory T cells (Tregs) have been shown to be potent inhibitors of autoimmunity, and to be capable of suppressing alloimmune responses that occur during both allograft rejection and graft-versus host disease. However, they have yet to gain widespread use clinically, due in part to the fact that it remains extremely costly and difficult to produce them in sufficient numbers and with sufficient suppressive capacity to significantly impact the alloimmune response. Here we have used our established non-human primate model to demonstrate that significant Treg expansion (up to 600-fold in 21 days) can be maintained, and suppressive capacity enhanced by exposing Treg cultures to a short burst of sirolimus at the end of the culture period. Using a highly sensitive and specific in vitro CFSE-MLR assay we show that Tregs significantly inhibit allo-proliferation of multiple T cell subpopulations including both CD4+ and CD8+ T cells (3.2 and 2.7-fold inhibition of proliferation, respectively), as well as their CD28+CD95+ and CD28-CD95+ subpopulations (2.2 and 2.1 and 1.9 and 2.7-fold inhibition of CD4+ and CD8+ subpopulation proliferation, respectively). Tregs were able to combine in vitro with the newly FDA-approved CTLA4-Ig analog belatacept to enhance the inhibition of alloproliferation that occurred with either agent alone (4.8-fold inhibition of CD8 T cell proliferation with Tregs + belatacept, compared to 3.0-fold or 1.9-fold inhibition of CD8 T cell proliferation with Tregs or belatacept alone, respectively). Importantly, we have found that the suppressive activity of ex-vivo expanded Tregs could be further enhanced by pulsing with sirolimus. Thus, while long-term culture of Tregs in the presence of sirolimus (1–1000 nM) profoundly inhibited Treg expansion (50–800 fold inhibition of expansion when cultured in the presence of 1–1000 nM sirolimus), a 48 hour pulse of sirolimus (100 nM) on days 20–21 of culture completely preserved Treg yields while doubling their suppressive function against CD8 proliferation when compared to unpulsed Tregs, p