ALBERT

Description: The nuclear receptors (NR) retinoid X receptors (RXRs) exert immunomodulatory functions to control inflammation and metabolism via homodimers and heterodimers with several other NRs including retinoic acid receptors. IRX4204 is a novel, highly specific RXR agonist in clinical trials that potently and selectively activates RXR homodimers but not heterodimers. Here, we show that in vivo IRX4204 was compared favorably with FK506 in abrogating acute graft-versus-host disease (GVHD), which was associated with inhibiting allogeneic donor T cell proliferation, reducing T helper 1 differentiation and promoting regulatory T cell (Treg) generation. Recipient IRX4204 treatment reduced intestinal injury and decreased IFN-γ and TNF-α serum levels. Transcriptional analysis of donor T cells isolated from intestines of GVHD mice treated with IRX4204 revealed significant decreases in transcripts regulating pro-inflammatory pathways. In vitro, inducible Treg differentiation from naïve CD4+ T cells was enhanced by IRX4204; in vivo, IRX4204 increased the conversion of donor Foxp3- T cells into peripheral Foxp3+ Tregs in GVHD mice. Using Foxp3 lineage tracer mice in which both the origin and current FoxP3 expression of Tregs can be tracked, we demonstrate that IRX4204 supported Treg stability. Despite favoring Tregs and reducing Th1 differentiation, IRX4204-treated recipients maintained graft-versus-leukemia responses against both leukemia and lymphoma cells. Notably, IRX4204 reduced in vitro human T cell proliferation and enhanced Treg generation in mixed lymphocyte reaction cultures. Collectively, these beneficial effects indicate that targeting RXRs with IRX4204 could be used as a novel approach to prevent acute GVHD in the clinic.

Description: Chimeric antigen receptor (CAR) T cells targeting CD19+ hematologic malignancies have rapidly emerged as a promising, novel therapy. In contrast, results from the few CAR T-cell studies for infectious diseases such as HIV-1 have been less convincing. These challenges are likely due to the low level of antigen present in antiretroviral therapy (ART)-suppressed patients in contrast to those with hematologic malignancies. Using our well-established nonhuman primate model of ART-suppressed HIV-1 infection, we tested strategies to overcome these limitations and challenges. We first optimized CAR T-cell production to maintain central memory subsets, consistent with current clinical paradigms. We hypothesized that additional exogenous antigen might be required in an ART-suppressed setting to aid expansion and persistence of CAR T cells. Thus, we studied 4 simian/HIV-infected, ART-suppressed rhesus macaques infused with virus-specific CD4CAR T cells, followed by supplemental infusion of cell-associated HIV-1 envelope (Env). Env boosting led to significant and unprecedented expansion of virus-specific CAR+ T cells in vivo; after ART treatment interruption, viral rebound was significantly delayed compared with controls (P = .014). In 2 animals with declining CAR T cells, rhesusized anti–programmed cell death protein 1 (PD-1) antibody was administered to reverse PD-1–dependent immune exhaustion. Immune checkpoint blockade triggered expansion of exhausted CAR T cells and concordantly lowered viral loads to undetectable levels. These results show that supplemental cell-associated antigen enables robust expansion of CAR T cells in an antigen-sparse environment. To our knowledge, this is the first study to show expansion of virus-specific CAR T cells in infected, suppressed hosts, and delay/control of viral recrudescence.

Description: In order to understand the mechanisms underlying the increased rates of stem cell transplant rejection in patients with sickle cell disease (SCD) we have developed a mouse model of transplant rejection utilizing the Berkeley SCD mouse. We find that Berkeley SCD mice exhibit a significantly higher rate of transplant rejection when transplanted with fully allogeneic bone marrow from an SJL donor after non-myeloablative conditioning and peri-transplant T cell costimulation blockade. Thus, while control C57BL/6 (BL/6) mice or hemizygous litter-mates exhibit 80–85% engraftment rates, only 20% of Berkeley SCD mice engraft with donor marrow. Transplant rejection in the SCD mice is mediated by CD8+ and/or NK1.1+ cells, as depletion with antibodies directed at either of these cell-surface molecules increases engraftment rates to 75–90%. To dissect the mechanistic underpinnings of these observations, we performed a multiparameter phenotypic and functional flow-cytometric analysis of leukocytes from the peripheral blood, lymph nodes and spleen from SCD mice, and compared them to the two non-SCD controls. We observed that SCD mice displayed striking phenotypic differences compared to normal mice, but that their hemizygous litter-mates were indistinguishable from the BL/6 controls, underscoring the sickle-specific nature of our observations. SCD mice showed minimal phenotypic variation in the peripheral blood, but the spleens and lymph nodes from SCD mice were highly abnormal. Sickle mice exhibited ∼10-fold more splenic NK and NK-T cells than either the BL/6 or hemizygous controls. These cells produced large quantities of both TNF-α and IFN-γ, potentially contributing to the inflammatory milleu that is associated with SCD. Furthermore, while absolute numbers of splenic T cells were not different in sickle mice compared to controls, their phenotype was skewed toward a highly activated state. Thus, while many (40%) splenic memory T cells from normal controls exhibited a Central Memory (Tcm) phenotype, there were fewer Tcm cells in SCD mice (6%), and a skewing of the memory repertoire towards a more activated Intermediate Memory phenotype (71% in SCD mice compared to 25% in controls). Finally, in ELISpot analysis, lymph node cells from naive sickle mice, having never been previously exposed to allo-antigen, were able to secrete IFN-γ when briefly co-cultured with donor SJL stimulators, a phenomenon not observed with either control BL/6 or hemizygous controls, again, pointing to the increased inflammatory milieu and potential for rejection in the sickle mice. In summary, we have observed evidence for a multi-faceted increase in immune activation in SCD mice, which correlates with their increased transplant rejection. These results point both toward possible therapeutic targets aimed at decreasing SCD-mediated transplant rejection and toward future studies in SCD patients, in order to determine whether similar mechanisms of immune activation also occur in the patient population.

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 350 Introduction: Leukapheresis is a widely utilized modality for collecting hematopoietic stem cells (HSCs). While collection of CD34+ cells with stem-cell activity is the primary goal of most mobilization and leukapheresis procedures, these cells only represent ∼1% of most leukapheresis products. The profile of the non-CD34+ cells is likely influenced by the choice of mobilization strategy, and has the potential to profoundly impact the post-transplant immune milieu of the transplant recipient. Two of the most critical of the CD34-negative cell populations that are collected during leukapheresis include effector and regulatory T cells. Thus, in evaluating mobilization regimens, the impact on these regimens on the mobilization of each of these T cell populations into the peripheral blood should be rigorously evaluated. Methods: We used a rhesus macaque model to determine the impact that mobilization with AMD3100 (a.k.a., Plerixafor or Mozobil®)+ G-CSF (“A+G”) had on peripheral blood CD4+ and CD8+ effector T cell populations as well as on FoxP3+/CD4+ T cells. Three rhesus macaques were mobilized with 10ug/kg SQ of G-CSF for five consecutive days prior to leukapheresis. AMD3100 was administered at 1mg/kg SQ in combination with the last dose of G-CSF two hours prior to leukapheresis. Leukapheresis procedures were performed for two hours using a modified CS3000 Plus cell separator. A peripheral blood sample was taken before cytokine therapy, just prior to leukapheresis following mobilization, one hour during leukapheresis, and at the end of the procedure. These samples were analyzed by multicolor flow cytometry using a BD LSRII flow cytometer. Results: Bulk, effector, and regulatory T cell subpopulations were analyzed flow cytometrically. The proportion of total CD3+ T cells remained stable during mobilization and apheresis: Thus, CD3+ T cells represented 77% of peripheral blood lymphocytes prior to mobilization, and 69% post-apheresis). The balance of CD4+ to CD8+ T cells was also relatively stable. Thus, for one of the three animals tested, the CD4+ and CD8+ proportions remained unchanged after apheresis. For two animals, the average CD4+ % decreased from 67% prior to mobilization to 52% post-apheresis. In these two animals, there was a reciprocal increase in the % of CD3+ T cells that were CD8+ (28% pre-G+A to 40% post-apheresis). The CD28+/CD95- naïve (Tn), CD28+/CD95+ central memory (Tcm) and CD28-/CD95+ effector memory (Tem) subpopulation balance of CD4+ and CD8+ T cells was also determined, by comparing the relative percentages of each subpopulation post-apheresis with their relative percentages prior to mobilization. Compared to their pre-G+A percentages, the post-apheresis CD4+ percentages of Tn, Tcm and Tem were 92%, 93% and 160%, respectively. Thus, the relative proportions of Tn and Tcm CD4+ cells decreased post-apheresis, while the relative proportion of CD4+ Tem increased compared to cytokine administration. For CD8+ T cell subpopulations, the post-apheresis proportions of Tn, Tcm, and Tem compared to their pre-G-CSF proportions were 99%, 70% and 130%, respectively–thus demonstrating the same direction of change as observed for CD4+ T cells. The most striking change in T cell subpopulations occurred in the CD4+/FoxP3+ compartment. The proportion of CD4+ T cells expressing FoxP3 increased by an average of 600% when post-apheresis samples were compared to pre-mobilization samples (FoxP3+ cells were 9.6% of CD4+ T cells post-apheresis versus 1.5% pre-GCSF). An average of 32% of these FoxP3+ CD4+ T cells expressed high levels of CXCR4. CXCR4 expression has been previously documented on human FoxP3+ T cells (Zou et al., Cancer Res, 2004), but this is the first observation of high level expression of CXCR4 on macaque FoxP3+ CD4 T cells, or of their ability to be efficiently mobilized with AMD3100. Discussion: These results suggest that treatment with AMD3100 and G-CSF may mobilize T cell subsets into the peripheral blood that could have beneficial effects during allo-transplantation. The combination of an increase in Tem cells, which have been observed to have decreased ability to cause GvHD (Zheng et al., Blood 2008), along with FoxP3+/CD4+ T cells, which may have regulatory functions, suggests that A+G mobilization could produce an apheresis product with a beneficial CD34-negative cell profile for allogeneic transplantation. Disclosures: No relevant conflicts of interest to declare.

Description: T stem cell memory (Tscm) represents a subset of T cells whose phenotype sits between those conventionally thought of as naive and those designated central memory. Functionally, they exhibit the capacity for self-renewal and can generate other memory cell subsets. As such, Tscm cells replenish and support the memory cell pool during infections, and contribute to immune reconstitution in patients following hematopoietic cell transplantation (HCT). Their fate and biological role in Graft-Versus-Host Disease (GVHD) is not yet understood. To address this question, we determined the dynamics and tissue-specific homeostasis of Tscm following allogeneic HCT in the translational non-human primate model of GVHD. In this study, we tracked Tscm cells by flow cytometry at baseline, and longitudinally after MHC haploidentical HCT. Recipients were divided into 2 large groups based on clinical outcome: "Primary GVHD": Animals receiving subtherapeutic GVHD prophylaxis (either no prophylaxis, or monotherapy with either Rapamycin or CTLA4-Ig; neither of which significantly prevented GVHD); "Breakthrough GVHD": Animals receiving combined Rapamycin/CTLA4-Ig or Tacrolimus/Metotrexate (both of which prevented primary GVHD but were eventually associated with breakthrough disease). We have previously shown that reconstituting T cells in animals with primary versus breakthrough GVHD display highly significant differences in their immunophenotype and transcriptome profiles. We have now discovered that primary GVHD is associated with a robust expansion of Tscm cells that peaks at day 5 post HCT (Fig 1A and B) and is associated with a decrease in the number of naive T lymphocytes. The Tscm compartment subsequently exhibited a contraction that accompanied an increase in the number of differentiated memory/effector T cell subsets (not shown). In contrast, animals that initially controlled GVHD maintained a delayed and prolonged expansion of Tscm cells (Fig 1A and B) that was associated with the significant preservation of naive T lymphocyte counts. Importantly, upon development of breakthrough GVHD, recipients demonstrated a transition of Tscm cells toward effector cells (Fig 1C). At the time of necropsy, animals with both primary and breakthrough GVHD displayed similar Tscm tissue distribution profiles, with these cells residing in the peripheral blood 〉 lymph nodes 〉 spleen 〉 bone marrow 〉 colon ~ liver ~ lungs. These data suggest that during GVHD, naive T cells transit to memory/effector lymphocytes via Tscm, and that one of the key functions of GVHD immunoprophylaxis may be to limit the rate at which naive T cells enter into the Tscm and subsequent effector T cell pool. Figure 1. T stem cell memory dynamics during acute GVHD. A-B) The frequencies (A) and absolute numbers (B) of CD8 T stem cell memory cells identified as CD3+CD14/20-CD8+CD45RA+CCR7+CD95+ lymphocytes and tracked longitudinally by flow cytometry following allogeneic HCT in primary (red circles) and breakthrough (black open circles) GVHD. Data display mean ± SEM. *p

Description: The PD-1/PD-L1 pathway plays an important role in regulation of alloimmune responses and in induction and maintenance of peripheral tolerance. Because GVHD is driven by donor T cells and PD-L1 expression can be markedly elevated on T cells during activation, we investigated the functional significance of PD-L1 expressed by donor T cells in regulating murine models of acute GVHD. PD-L1 expression was up-regulated on donor CD4 and CD8 T cells during GVHD. We considered the possibility that PD-L1 expression on activated donor T cells might inhibit GVHD by down regulating donor anti-host T cell responses, consistent with PD-L1 co-inhibitory activity when expressed on host parenchymal cells during GVHD. Surprisingly, T cell mediated GVHD lethality was markedly reduced in recipients of PD-L1-/- compared to WT donor T cells in both B6 to BALB/c model of GVHD(P

Description: Acute graft-versus-host disease (aGVHD) develops in more than half of patients after allogeneic hematopoietic cell transplantation (allo-HCT) despite poly-pharmacy immunoprophylaxis. Importantly, the dysregulated pathways responsible for this breakthrough disease remain largely unidentified. Thus, the discovery of these pathways represents one of the critical challenges for the field of allo-HCT. To address these needs, we have developed a model of aGVHD in rhesus macaques, which allows us to study the mechanisms of aGVHD both in its untreated state and in a variety of immunoprophylactic settings. Using a systems-based approach, we have created both a multiparameter flow cytometric and transcriptomic map of the immune landscape of aGVHD in allo-HCT recipients, in comparison to two critical control groups: (1) healthy untransplanted controls, and (2) those receiving autologous transplantation. We find that recipients of allo-HCT receiving 1) no immunoprophylaxis 2) monotherapy with CTLA4Ig or 3) monotherapy with sirolimus develop early fulminant aGVHD with multi-organ disease (Figure 1a-grouped as 'Primary GVHD'). The immunophenotype of T cells from the Primary GVHD cohort exhibits an effector/memory phenotype with robust proliferation and acquisition of cytotoxic function. Transcriptomic analysis reveals enrichment of Th1-associated transcripts (IL12RB2, CCR5, CXCR3) as well as programs of proliferation early in the post transplant period (Figure 1b). Flow cytometric data confirms an increase in the number of CD4 and CD8 T cells producing the Th1 cytokine, IFN-g at this time-point (Figure 1c). In contrast, standard-of-care Tacrolimus/Methotrexate (Tac/Mtx) as well as novel CTLA4Ig/sirolimus combination immunoprophylaxis (CoBS) both significantly improved survival of animals after allo-HCT. However, similar to human patients undergoing allo-HCT, these recipients often developed clinical signs of breakthrough aGVHD (starting around day 30 post-transplant) characterized by both gastrointestinal and skin pathology. This cohort was thus termed the "Breakthrough GVHD" cohort (Figure 1a). Unexpectedly, despite the presence of breakthrough clinical aGVHD, the Tac/Mtx and CoBS cohorts were still able to control programs of T cell proliferation, effector phenotype acquisition and Th1 cytokine skewing. However, both transcriptional and flow cytometric profiles demonstrated enrichment for molecules that reflect Th17/Th22 skewing (RORC, IL17A, AHR, and IL22) (Figure 2a) and production of IL17a (Figure 2b). These results suggest that while current methods of immunoprophylaxis are able to limit both T cell proliferation and Th1 polarization, breakthrough Th17/Th22 pathway activation occurs despite these therapies. These data suggest that emphasis should be placed on exploration of pharmacologic inhibitors of IL17/IL22 for the prevention/treatment of breakthrough aGVHD. Disclosures No relevant conflicts of interest to declare.