Search

Hematopoietic stem cells (HSCs) are capable of replenishing the entire blood system when needed and transplantation of HSCs remains as one of the most effective, curative treatments for patients with genetic diseases and hematopoietic malignancies. In vitro culture is an essential process for ex vivo expansion and modification of HSCs, however engraftment levels of cultured HSCs cannot be accurately estimated. This is mainly due to lack of reliable cell surface markers representing functionality of HSCs after culture, which also limits the resolution of molecular analyses. We have previously shown that HSCs are vulnerable to endoplasmic reticulum (ER) stress responses fueled by accumulation of unfolded / misfolded proteins (Miharada et al., Cell Rep. 2014). Importance of ER stress suppression is also evident in vivo, as proliferative FL-HSCs fail to expand upon ER stress induction when natural molecular chaperone, bile acid, is reduced (Sigurdsson et al., Cell Stem Cell. 2016). Thus, ER stress elevation severely impairs the potential of HSCs, however usual marker profile is no longer representative of their functionality. In this study we aimed to discover the key signature and novel markers that represent functional retardation of HSC under activation and stress induction. Initially we compared gene expression profiles of fresh and 14-days cultured Lineage-Sca-1+c-kit+(LSK) CD48- (CD48-LSK) cells from mouse bone marrow using microarray analysis, since CD48 has been reported to enrich functional HSCs after in vitro culture (Noda et al., Stem Cells, 2008). We discovered abnormal up-regulations of genes frequently associated with mast cells (MC) in cultured CD48-LSK cells, and identified Cd244 as one of the top upregulated genes. CD244 is a member of the slam family of genes but is considered to be redundant with other slam markers in isolating HSCs from untreated mice. Indeed, freshly isolated CD150+CD48-LSK cells are negative for CD244. However, after 14-days in vitro culture with stem cell factor (SCF) and thrombopoietin (TPO), majority of CD150+CD48-LSK cells were positive for CD244. After shorter (7-days) culture, we found that CD48-LSK cells could be subdivided to CD244+ and CD244- populations (CD244-HSC and CD244+HSC). CD244-HSCs expressed high levels of HSC-related genes such as Fgd5, Hlf, Fhl1 and thrombopoietin receptor Mpl, In contrast, CD244+HSCs expressed MC-related genes, e.g. Cpa3, Gzmb and Mcpt8. In transplantation settings, CD244+HSCs showed no engraftment while CD244-HSCs showed long-term engraftment revealing them as functional stem cells. Since our and other groups have demonstrated that induction of ER stress impairs potential of mouse and human HSCs, we asked if ER stress induction would lead to the elevation of MC signature. Using an ER stress inducing chemical, thapsigargin, we could see increased ratio of CD244+HSCs within CD48-LSK cells. Conversely, the addition of TUDCA, a bile acid known to suppress ER stress, resulted in decreased frequency of CD244+HSCs. These findings strongly indicate that ER stress could be influencing the number of non-functional HSCs. To further substantiate the connection to ER stress and MC signature we analyzed a knock out mouse model of the ER stress modulator Trib3 (Trib3-/-) that is known to show an abnormal differentiation towards mast cells. Trib3-/- HSCs expressed MC genes including Cpa3 already at the steady-state condition. The number of CD244-HSCs after 7-days culture was significantly lower than control mice, and showed poor long-term engraftment potential in transplantation settings. To further elucidate the key molecular changes that impair HSCs, we compared gene expression profiles between fresh HSCs and CD244+/CD244-HSCs after 7-days culture. Gene expression comparison between CD244+ and CD244-HSCs independently confirmed the enrichment of MC cell related genes including Granzyme B (Gzmb), known to have negative impact on HSC potential (Carnevali et al., J Exp Med. 2014). Moreover, the Rel-A pathway was significantly lower in CD244-HSCs compared to fresh HSCs, suggesting a potential implication of NF-kB signal in the first alterations in HSCs during in vitro culture. We conclude that the induction of a MC cell signature fueled by ER stress is critical for normal HSC potential, and CD244 is a novel marker predicting the functionality of activated HSCs and allowing more detailed molecular analysis of activated HSCs. Disclosures No relevant conflicts of interest to declare.

Invariant natural killer T (iNKT) cells are an interesting subpopulation of T cells that can potently inhibit graft-versus-host-disease (GVHD) through the production of Interleukin 4 (IL-4), while also carrying anti-tumor potential (Leveson-Gower et al. Blood. 2011;117:3220-9; Schneidawind et al. Blood. 2014;124:3320-8). Murine iNKT cells differentiate during thymic development into three distinct sublineages, named according to the classification of conventional T cells: Th1-like iNKT (iNKT1) cells, Th2-like iNKT (iNKT2) cells, and Th-17 like iNKT (iNKT17) cells (Brennan et al. Nat Rev Immunol. 2013; 13:101-1). In this study we investigated the immune regulatory and anti-tumor potential of iNKT1, iNKT2 and iNKT17 cell subsets. Thymic iNKT1 cells, iNKT2 cells and iNKT17 cells were isolated from 8-10 week-old FVB/NJ mice by flow cytometry based on PBS-57-CD1d-Tetramer and a combination of cell surface molecules (iNKT1: ICOS- PD1- CD27+ CD24-; iNKT2: ICOS+ PD1+ CD27+ CD4+ CD24-; iNKT17: ICOS+ PD1+ CD4- CD27- CD24-) and purity was confirmed by intra-nuclear staining for the transcription factors PLZF and RORγT. RNA sequencing analysis determined that iNKT1 cells were the main subset expressing proinflammatory and cytotoxic genes, such as Interferon gamma (IFN-γ), Fas Ligand (FasL), Perforin, and Granzyme B (Gzmb), whereas IL-4 was expressed by iNKT2 cells and, at a lesser extent, by iNKT17 cells. To assess the immuno-regulatory potential of the three iNKT sublineages, we employed a murine major histocompatibility complex (MHC)-mismatched bone marrow transplantation model. BALB/c (H-2Kd) recipients were lethally irradiated with 8.8 Gy; on the same day, 4 x 106 TCD-BM cells and 1 x 106 conventional CD4 and CD8 T cells (Tcon) from FVB/NJ (H-2kq) mice were injected intravenously. Additionally, 5 x 104 purified iNKT1, iNKT2 or iNKT17 cell subsets from FVB donors were injected. A significant survival benefit was observed when iNKT2 (p=0.017) and iNKT17 (p=0.033) cells were adoptively transferred compared to mice that only received TCD-BM and Tcon, whereas there was no survival benefit in the group that received iNKT1 cells. In addition, body weight was improved in mice that received iNKT2 (day +41: p=0.009, day +49: p=0.005 and day +59: p= 0.005) or iNKT17 (day +59: p= 0.006) compared to mice that received iNKT1 cells. Clinical GVHD scores were also improved in mice that received iNKT2 (day +41: p= 0.012, day +51: p= 0.005) or iNKT17 (day +28: p=0.05, day +51: p=0.007) compared to mice that received iNKT1 cells. Interestingly, we found that even 1 x104 iNKT2 (p= 0.008) and iNKT17 (p= 0.04) significantly suppressed GVHD. As iNKT1, iNKT2 and iNKT17 have a very different gene expression profile, we tested the ability of the sublineages to kill a B-cell lymphoma cell line transduced to express high levels of CD1d (A20-CD1d) in vitro and found that iNKT1 cells killed A20-CD1d cells significantly better than iNKT2 (p=0.006) or iNKT17 (p=0.0001) cells. These findings are in line with the sequencing data mentioned above, showing that iNKT1 cells express a more inflammatory phenotype. In summary, we demonstrate here that only iNKT2 and iNKT17 cells protect from GVHD, whereas iNKT1 cells have cytotoxic function. To our knowledge, this is the first study to show functional differences between the iNKT sublineages, suggesting that iNKT1, iNKT2 and iNKT17 cells have diverse functions. Therefore, these data provide new biological insights, which will be useful for developing iNKT cell-based cell therapy. Disclosures Chang: Spring Discovery: Membership on an entity's Board of Directors or advisory committees; Epinomics and Accent Therapeutics: Equity Ownership, Membership on an entity's Board of Directors or advisory committees.

Human invariant natural killer T cells (iNKT) are a rare population of lymphocytes that bridges innate and adaptive immune functions. iNKT cells have different potent effector functions and subsets, but this heterogeneity is mostly understood by a superimposed framework of classic T cell markers. This may help explain contradictory studies of iNKT cells in diseases for which they are thought to play a pathogenic role. In this study, we present data suggesting new markers to understand human iNKT cell heterogeneity and function, as well as monitoring this in the clinical context of allogeneic stem cell transplantation (ASCT) and graft-versus-host disease (GVHD). In both preclinical murine models and in correlative clinical studies, iNKT cells are associated with less GVHD and better immune reconstitution following ASCT. We performed bulk whole transcriptome sequencing on iNKT cells from patients at day +30 after ASCT, some who developed and some who did not develop GVHD. Gene expression signatures of the iNKT cells from each individual are grouped based on a transcriptome library of reference cell types (CIBERSORT). Using this approach, we distinguished patients who will develop GVHD as predominately having an activated cytotoxic cell gene signature, whereas patients without GVHD had a CD4+ T cell gene signature within purified peripheral blood iNKT cells. Using two different high throughput single-cell RNA sequencing (sc-seq) platforms to determine differential gene expression on the single cell level, we first examined the gene expression signatures of primary human iNKT cells from healthy donors. Activation with CD3/28 microbeads of primary iNKT cells promoted two main differential transcriptional profiles in iNKT cells. One profile resembles conventional CD4+ T cells and associated Th2 cytokine profile (IL2, IL4), whereas the other profile has genes associated with cytotoxicity and inflammation (TNF, IFNG, GZMB). Confirmation by flow cytometry showed that while CD8 surface expression did not well differentiate the pro-inflammatory and cytotoxic subsets, but both CD94 and KLRG1 better associate with Th1 and cytotoxic responses. CD94+ iNKT are restricted to CD4- cells and KLRG1 expressed in both CD4+ and CD4- populations. By comparison, the analysis of sc-seq of normal healthy subjects and post-transplant patients with and without GVHD resulted in the identification of three major iNKT populations. The population most associated with GVHD showed expression of a pro-inflammatory profile enriched for CXCR4,CD94, HLA-DRB1/A1 and decreased KLRB1. Whereas, post-transplant patients with healthy GVHD-free and relapse free immune reconstitution showed a more T cell like profile but with increased expression of CXCR4, CD94, KLRB1 and decreased HLA-DRB1/A1. Using flow cytometry of ex vivo expanded iNKT cells, we found the expansion of CXCR4+HLA-DR+ that can be CD4+ or CD4- and that these cells produce high levels of the cytokine TNF-a, IFN-g and IL-4. Importantly, iNKT cells that are CXC4+HLA-DR+KLRB1low make significantly less IL-4 and these are enriched in the GVHD setting. Using flow cytometry, we measured iNKT cell phenotypes in a pilot cohort of 48 individuals; 10 healthy controls, 10 patients with no complications or relapse after ASCT and 18 patients with GVHD of which 9 were steroid refractory. In addition, we evaluated 11 patients with new onset T1D. The frequency of the cytotoxic CD94+ iNKT cells were significantly increased in all patients following and in T1D compared to healthy controls. Whereas, patients with GVHD showed an increase in HLA-DR+CXCR4+KLRB1low iNKT cells. Increases in the late activation marker HLA-DR distinguished corticosteroid refractory (HLA-DR++) from steroid responsive patients at the time of initial GVHD diagnosis. For patients who responded to steroids or further therapy, we found that a reduction in HLA-DR associated with treatment response. In a second 24 patient cohort, we found that the presence HLA-DR+CXCR4+CD161low cells at day +30 was statistically greater in patients who would go on to develop GVHD as compared as those who did not (student T test, p To conclude, we have used different methods to discover and validate groups of iNKT cells with distinct functions, and the associated phenotype correlate with clinical inflammation both following ASCT and in T1D. Disclosures No relevant conflicts of interest to declare.

T cells contribute to host-tumor interactions in patients with monoclonal gammopathies. Expansions of CD8+CD57+ T-cell receptor Vβ–positive (TCRVβ+)–restricted cytotoxic T-cell (CTL) clones are found in 48% of patients with multiple myeloma and confer a favorable prognosis. We now report that CTL clones with varying TCRVβ repertoire are present in 70% of patients with Waldenström macroglobulinemia (WM; n = 20). Previous nucleoside analog (NA) therapy, associated with increased incidence of transformation to aggressive lymphoma, significantly influenced the presence of TCRVβ expansions (χ2 = 11.6; P < .001), as 83% of patients without (n = 6) and only 7% with (n = 14) TCRVβ expansions had received NA. Clonality of CD3+CD8+CD57+TCRVβ+-restricted CTLs was confirmed by TCRVβ CDR3 size analysis and direct sequencing. The differential expression of CD3+CD8+CD57+TCRVβ+ cells was profiled using DNA microarrays and validated at mRNA and protein level. By gene set enrichment analysis, CTL clones expressed not only genes from cytotoxic pathways (GZMB, PRF1, FGFBP2) but also genes that suppress apoptosis, inhibit proliferation, arrest cell-cycle G1/S transition, and activate T cells (RAS, CSK, and TOB pathways). Proliferation tracking after stimulation confirmed their anergic state. Our studies demonstrate the incidence, NA sensitivity, and nature of clonal CTLs in WM and highlight mechanisms that cause anergy in these cells.

INTRODUCTION The recent success of checkpoint blockade immunotherapies in diverse solid tumors has prompted the evaluation of these treatments in hematologic malignancies such as acute myeloid leukemia (AML). It is critical to identify the patient and disease subsets that could respond to such therapies. Infiltration of tumors by cytotoxic T lymphocytes (CTLs) has been associated with better prognosis and responses to checkpoint inhibition. We hypothesized that the presence of a substantial fraction of activated CTLs and natural killer (NK) cells in the blood and bone marrow samples of hematologic tumors could indicate a preexisting active immune response potentially targeting the tumor cells. Moreover, the density of the immune infiltrate could shape and be shaped by the expression of cancer-germline and leukemia-associated antigens (LAAs), antigen-presenting machinery (APM) and immunosuppressive genes by the tumor cells. Here, we examined these immunological properties of hematological tumors in large-scale gene expression datasets to identify immunologically active patient subsets. METHODS Curated set of 9,544 transcriptomes collected across 36 hematological malignancies (HEMAP), including 1,858 AML cases was utilized to identify subsets of patients with existing, potentially tumor-directed immune responses. Additional multi-omics datasets of 173 AML patients from The Cancer Genome Atlas (TCGA) were integrated to gain insight into the genetic landscape of immunologically active patients. Cytolytic activity (geometric mean of GZMA (granzyme A) and PRF1 (perforin) transcript levels, Rooney et al., Cell 2015) was used as a marker of immunologic activity. Cytolytic activity was correlated to the expression levels of all transcripts, gene sets from collections such as MSigDB and manually curated gene sets representing the APM (HLA-A, -B, -C, B2M), 145 known cancer-germline antigens as well as established LAAs such as WT1 and PRTN3. Furthermore, we used an in silico flow cytometry approach, CIBERSORT (Newman et al., Nat Methods 2015), to infer the relative fractions of 22 immune cell subpopulations from the gene expression data to dissect the immune cell composition of the samples. RESULTS Cytolytic activity showed high correlation with other transcripts expressed in activated CTLs and NK cells (e.g. GZMB, GNLY, KLRB1, CD8A, CD2; Spearman's R ≥ 0.7) as well as lymphocyte activation-related gene sets across both the HEMAP and the TCGA AML datasets, validating it as a robust and specific metric of active cellular immunity. When correlated to the CIBERSORT immune cell populations, cytolytic activity was positively associated with CD8 T cells and showed a negative correlation to the proportion of M2 macrophages. High levels of cancer-germline antigens were associated with decreased expression of components of the APM and low cytolytic activity, suggesting HLA downregulation as a mechanism of immune evasion by cancer-germline antigen-expressing tumor cells. We observed extensive heterogeneity in the cytolytic activity between different diseases and subtypes within the same disease, most prominently in AML. In AML patients, complex karyotype and unfavorable prognosis were correlated with high cytolytic activity, indicating biological similarity of the immune-infiltrated tumors. Furthermore, TP53 mutations, genome fragmentation and immune checkpoint transcripts such as CD274 (PD-L1), PDCD1LG2 (PD-L2), CTLA4 and LAG3 were enriched within the complex karyotype cluster in the TCGA AML dataset. In contrast, mutations in NPM1 and FLT3 showed a modest but significant negative correlation to cytolytic activity. CIBERSORT analysis revealed that AML cases with low cytolytic activity preferentially had enrichment of an eosinophilic phenotype in addition to increased M2 polarization of macrophages. CONCLUSIONS Using large-scale transcriptomics approaches, we were able to identify patient subsets with variable levels of immune cytolytic activity within hematologic malignancies. Furthermore, we identified connections between the cytotoxic immune response and genetic properties of AML tumors. These observations have potential clinical implications, as the choice of patients to clinical trials receiving immune checkpoint blockade immunotherapies would require careful consideration in light of the observed immunological heterogeneity. Disclosures Mustjoki: Bristol-Myers Squibb: Honoraria, Research Funding; Pfizer: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Ariad: Research Funding.

BACKGROUND: Alloimmunization to minor blood group antigens is more common in sickle cell disease (SCD) than in the general population. Causes include the high frequency of transfusion as well as phenotypic differences between the largely African-American SCD population and mostly Caucasian blood donors. Alloimmunization, as well as the risk of alloimmunization, results in higher costs as well as reduced availability of and delays in obtaining antigen-matched blood to meet clinical needs. We used a candidate gene approach to determine if specific genetic signatures were associated with alloimmunization and thus could identify patients at high risk for alloimmunization and therefore requiring prospective antigen matching for transfusion. METHODS: Patient data were previously collected through a multicenter study to identify genetic factors associated with SCD complications (Duke and UNC; DU) and through the Cooperative Study of SCD (CSSCD). All subjects were adults with HbSS genotype (DU mean yrs 34.9 ± 12.1, CSSCD mean yrs 29.1 ± 9.5, p < .0001). The DU data set was 43.7% male while the CSSCD data set was 37.1% male (p = .18). A genome-wide association study (GWAS) was conducted with the Illumina 610 BeadChips for ­­­390 individuals (215 DU, 175 CSSCD). Quality control processing of these data was published previously (Solovieff et al, 2010). Additionally, imputation of the DU data set was conducted using IMPUTE2 (Howie et al, 2009) and a global reference panel from 1000 Genomes in order to have consistent genotypes across both studies. Based on previous literature, we studied 255 SNPs in the HLA locus as well as 684 SNPs in 53 genes previously associated with immune responses, totaling 939 SNPs. Additive logistic regression was performed separately for the DU and CSSCD data sets with PLINK (Purcell et al, 2008), controlling for age at enrollment. Results from the two data sets were combined by meta-analysis, using the invariance weight method in METAL (Willer et al, 2010). False discovery rate (FDR) q-values were generated using PROC MULTTEST in SAS v9.4 (SAS Systems). Contingency tables were produced using SAS and association between candidate SNPs and specific antibodies was tested using PLINK. RESULTS: Alloimmunization rates were different in the DU and CSSCD cohorts (30.7% vs. 18.9%, respectively; p = .0075) but were not associated with age or number of transfusions > 0 in either cohort. The most significant genetic associations with alloimmunization (FDR q = .28) occurred for 46 SNPs in the HLA locus and 10 other genes (Table 1). Two of these SNPs were missense mutations in GZMB (granzyme B (granzyme 2, cytotoxic T-lymphocyte-associated serine esterase 1)), one of which (rs8192917) has previously been associated with acute kidney transplant rejection and acute GVHD after stem cell transplant. A missense mutation in TRIM31 (E3 ubiquitin-protein ligase) was also associated with alloimmunization (p = .01). TRIM31 regulates Src-induced anchorage-independent cell growth and has been implicated in antiretroviral defense, cancer and ADHD. Within the DU cohort, we conducted a sub-analysis among only antigen-negative individuals, in whom immunization rates were 15.8%, 14%, 4.5%, 10.2%, 11%, and 8.3% to E, C, K, S, Jkb and Fya, respectively. Regression analysis of all 939 SNPs identified nominally significant associations predicting specific antibody alloimmunization. A missense mutation in PRKCQ (rs2236379) was associated with both anti-E and –C alloimmunization. Additional missense mutations were identified as follows: in TGFBR3 and GZMB with anti-C; in CFH, andTLR6 with anti-K, and in TLR10 with anti-Jkb. CONCLUSIONS: We have identified several immune response genes that are also associated with increased risk of alloimmunization in SCD (nominal p ≤ .006, FDR q = .28). Some associations may arise from the prevalence of alloimmunization to specific antigens. For example, PRKCQ was associated with development of alloantibodies to both E and C antigens, which reside on the same nonglycosylated RH protein. Expanded studies may identify those patients who would benefit from intensive or specific antigen-matching transfusion protocols. TableGenes Associated with AlloimmunizationGene or Locus (ordered by descending p value)SNPs with FDR q = 0.28 SNPs with p ≤ .006GZMB3 (2 missense)3TGFBR242TLR311HLA locus262TGFBR332PRKCQ3CD801VAV21TRIM312 (1 missense)RNF391ZNRD11 Disclosures No relevant conflicts of interest to declare.

The therapeutic benefits of allogeneic hematopoietic stem cell transplantation (allo-HSCT) for hematologic malignancies are primarily derived from anti-leukemia effect that is mediated by T cells in donor grafts. Unfortunately, these T cells also mediate graft-versus-host disease (GvHD), the major complication of allo-HSCT. We have recently reported that blockade of alloreactive donor T cell trafficking to the GvHD target organs significantly reduces GvHD in both a MHC fully-mismatched (B6 (H-2b) → Balb/c (H-2d)) (75% vs 0% overall survival) and a minor-mismatched (B6 (H-2b) → B6x129 (H-2b)) allo-HSCT models (100% vs 10% overall survival) compared to WT T cells (Choi et al., Blood 2012). In this study, we examined if inhibition of VLA-4, which is required for transendothelial migration and access to GvHD target organs, could reduce GVHD. VLA-4 consists of two subunits α4 (CD49d) and β1 (CD29). To genetically eliminate VLA-4 from allogeneic donor T cells, we generated Tie-2 cre+ α4-/- mice (B6, H-2b, CD45.2+) in which α4 is knocked out by Tie-2 cre, which is expressed in hematopoietic cells. Splenic pan T cells were isolated from these mice and T cell-depleted bone marrow cells (TCD BM) from congenic B6 mice (CD45.1+). 5x106 TCD BM and 5x105 splenic pan T cells were transplanted into lethally irradiated allogeneic Balb/c recipient mice (H-2d, CD45.2+). We found that VLA-4 deficient T cells significantly reduced GvHD compared to WT T cells (Fig. A). VLA-4 deficient T cells had no significant effect on donor engraftment and achieved a complete donor chimerism. In addition, recipients transplanted with VLA-4 deficient T cells had significantly better histopathology score (Fig. B). Similar results were observed when donor T cells were infused at day 11 after allo-HSCT. While we speculate defective T cell trafficking to GvHD target organs would be the primary reason for the reduced GvHD in the recipients of VLA-4 deficient T cells, we examined other possible mechanisms as well, such as T cell alloreactivity and proliferation. We found that VLA-4 deficient T cells proliferated at the same rate as WT T cells in the presence of anti-CD3/CD28 antibody-coated beads. However, VLA-4 deficient T cells proliferate less in the presence of allogeneic antigen presenting cells (APCs) compared to WT T cells. Interestingly, VLA-4 deficient T cells upregulate CTLA-4 and GZMB significantly more than WT T cells after in vitro activation. These data suggest that not only T cell trafficking to GvHD target organs but also altered T cell phenotype and function might help explain the observed reduced GvHD. Genetic deletion of α4 in T cells will also alter expression of α4b7. Since antibodies to b7 have also been shown to reduce GvHD (Waldman et al., Blood, 2006) our current studies will focus on determining which of these pathways (α4b1 vs. α4b7) is most important in mitigating GvHD by using genetic knockout models, antibodies to b1 and b7 and small molecule inhibitors of α4b1 vs. α4b7. In conclusion, our data suggests that VLA-4 represents a promising therapeutic target for future efforts to mitigate GvHD after allo-HSCT. In addition, this strategy can be exploited in other diseases and settings in addition to allo-HSCT, such as solid organ transplantation, chronic inflammatory diseases and autoimmune diseases. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

When released from cytotoxic T lymphocytes and natural killer cells, Granzyme (Gzm) serine proteases induce programmed cell death of pathogen-infected cells and tumor cells. The Gzms rapidly accumulate in the target cell nucleus by an unknown mechanism. Many of the known substrates of GzmA and GzmB, the most abundant killer cell proteases, bind to DNA or RNA. Gzm substrates predicted by unbiased proteomics studies are also highly enriched for nucleic acid binding proteins. Here we show by fluorescence polarization assays that Gzms bind DNA and RNA with nanomolar affinity. We hypothesized that Gzm binding to nucleic acids enhances nuclear accumulation in target cells and facilitates their cleavage of nucleic acid-binding substrates. In fact, RNase treatment of cell lysates reduced cleavage of RNA binding protein (RBP) targets by GzmA and GzmB. Moreover, adding RNA to recombinant RBP substrates greatly enhanced in vitro cleavage by GzmB, but adding RNA to non-nucleic acid binding proteins did not. For example, exogenous RNA enhanced GzmB cleavage of recombinant hnRNP C1 (an RBP) but not LMNB1 (a non-RBP). In addition, GzmB cleaved the RNA-binding HuR protein efficiently only when it was bound to an HuR-binding RNA oligonucleotide, but not in the presence of an equal amount of non-binding RNA. Thus, nucleic acids facilitate Gzm cleavage of nucleic acid binding substrates. To evaluate whether nucleic acid binding influences Gzm trafficking in target cells, we incubated fixed target cells with RNase and then added Gzms. RNA degradation in target cells reduced Gzm cytosolic localization and increased nuclear accumulation. Similarly, pre-incubating Gzms with exogenous competitor DNA reduced Gzm nuclear localization. The Gzms form a monophyletic clade with other immune serine proteases including neutrophil elastase (NE) and cathepsin G (CATG). Upon neutrophil activation, NE translocates to the nucleus to drive the formation of neutrophil extracellular traps (NETs). NE and CATG, but not non-immune serine proteases such as trypsin and pancreatic elastase, also bind DNA with high affinity and localize to the nucleus of permeabilized cells. Consistent with this finding, competitor DNA also blocks the nuclear localization of NE. Moreover NE and CATG localization to NETs depends on DNA binding. Thus the antimicrobial activity of NETs may depend in part upon the affinity of these proteases for DNA. Our findings indicate that high affinity nucleic acid binding is a conserved and functionally important property of serine proteases involved in cell-mediated immunity. Disclosures: Lieberman: Alnylam Pharmaceuticals: Membership on an entity’s Board of Directors or advisory committees.

Abstract 104 Natural killer (NK) cells are innate lymphocytes important for early host defense against infectious pathogens and surveillance against malignant transformation. Resting murine NK cells regulate the translation of effector molecule mRNAs (e.g. granzyme B, GzmB) through unclear molecular mechanisms. MicroRNAs (miRNAs) are small non-coding RNAs that post-transcriptionally regulate the translation of their mRNA targets, and are therefore candidates mediating this control process. While the expression and importance of miRNAs in T and B lymphocytes has been established, little is known about miRNAs in NK cells. Here, we utilized two next-generation sequencing platforms, Illumina GA and SOLiD, to define the miRNA transcriptomes of resting and IL-15-activated primary murine NK cells. Using Illumina GA sequencing, we identified expression of 297 mature miRNA and 64 miRNA* sequences in NK cells at read counts of 1 to 507,369. The 5 miRNAs with the highest abundance (percent normalized read count) in resting NK cells were miR-21 (31.6%), miR-16 (8.3%), miR-142-5p (6.5%), miR-142-3p (3.8%), miR-24 (3.7%). Detection of known miRNA was similar using SOLiD sequencing in both resting (r2=0.79) and IL-15-activated (r2=0.75) NK cells, with most of the discordant detection arising in miRNAs with low expression (1-10 read counts). These results were further validated using real-time RT-qPCR and microarray profiling. A bioinformatics pipeline was developed that identified 37 novel miRNA genes that encoded 26 novel mature miRNA sequences, including miRNAs processed from snoRNAs. Of 84 miRNAs with significantly altered expression levels (Illumina, normalized read counts) following IL-15-activation, miR-223 demonstrated the greatest fold decrease (2.3, P=1.03E-10). The decrease in mature miR-223 following 24 hours of IL-15 activation was confirmed in SOLiD sequencing (7.7 fold-decrease, P=1.0E-10), qPCR (2.5 fold-decrease, P=0.01), and microarrays (4.6 fold-decrease, P=0.003). IL-15 induced reduction in mature miR-223 was time-dependent with fold decreases of 1.5 (NS) at 8 hours, 2.2 (P=0.003) at 24 hours, and 8.1 (P=0.012) at 48 hours, assessed by real-time qPCR. Decreases in mature miR-223 expression early following activation likely occur via regulation of the miRNA processing machinery, since primary and precursor miR-223 transcripts exhibited increased expression after 8 hours (2.1±0.3 fold increase, P=0.03) and 24 hours (2.5±0.3 fold increase, P=0.03) of IL-15-activation. MiR-223 is computationally predicted to bind to the murine GzmB 3'UTR sequence, and we therefore evaluated the ability of miR-223 to target the murine GzmB 3'UTR in vitro. Overexpression of miR-223 in 293T cells decreased luciferase protein signal controlled by the GzmB 3'UTR by 44±2% (P=0.002) without affecting negative controls (perforin 3'UTR or no 3'UTR). Moreover, mutating the miR-223 predicted binding site within the murine GzmB 3'UTR abrogated miR-223 mediated regulation. These data suggest that miR-223 directly targets the murine GzmB 3'UTR, and thereby contributes to tonic suppression of GzmB mRNA translation in resting NK cells. Thus, the sequenced and validated miRNA transcriptome provides a valuable framework for further elucidation of miRNA expression and function in NK cell biology. Disclosures: No relevant conflicts of interest to declare.

Abstract 681 Granzyme B (GZMB) is a serine protease with important roles mediating target-cell apoptosis induced by natural killer cells and cytotoxic CD8+ T cells. The polymorphism rs8192917 (A55G) in the GZMB gene has been reportedly associated with the ability of effector cells to secrete GZMB. In this study we analyzed the impact of GZMB polymorphism on transplant outcomes in patients undergoing unrelated HLA-fully-matched myeloablative bone marrow transplantation (BMT) through the Japan Marrow Donation Program. The GZMB genotypes were retrospectively analyzed in a cohort of 360 pairs of patients with hematologic malignancies and their unrelated donors. The presence of the G/G genotype in the donor side was associated with a significantly higher incidence of relapse (62% vs. 32%, P=0.04; Fig 1). Multivariate analysis revealed that the donor G/G genotype, which is expected to produce lower levels of GZMB after stimulation, (Girnita et al, Transplantation 2009) was an independent risk factor for relapse (hazard ratio, 4.17; 95% confidence interval, 1.63 to 10.70; P=0.003). The host GZMB genotypes did not significantly influence the transplant outcomes. These results suggest an association of low GZMB secretors with disease relapse. These could therefore be useful in selecting the donor and creating therapeutic strategies for improving the final outcome of allogeneic BMT. Disclosures: No relevant conflicts of interest to declare.

Regulatory T cells (Treg) contribute to the maintenance of self-tolerance and have been demonstrated to both suppress autoimmune diseases and mitigate GvHD in mouse models. Major obstacles for their routine use in human clinical trials to reduce autoimmunity or GvHD include the low number of Treg found in the peripheral blood in the resting state (5–10% of CD4+ T cells), low yields and efficiencies of purification and severe limitations in ex vivo expansion without downregulating Foxp3 expression. The master gene responsible for the normal development and suppressor function in vitro and in vivo of Treg is Foxp3 which is exclusively expressed in Treg. It remains unclear what regulates the expression of Foxp3 in Treg. Recently, selective CpG motifs within the Foxp3 locus have been reported to be demethylated in Treg and heavily methylated in CD4+CD25-T cells. Thus, we sought to determine if an FDA-approved demethylating agent, Decitabine, could be used to enhance expression of Foxp3 via an epigenetic effect and convert CD4+CD25- T cells into Treg. We incubated human CD4+CD25-FOXP3- T cells with CD3/CD28 beads (1:1 bead to cell) and hIL-2 (50u/ml) for 2–3 days and treated them with or without 0.1–10 uM Decitabine for an additional 3–4 days. Optimal expression of FOXP3 was seen in those cells incubated with 1–5 uM Decitabine. Real time RT-PCR demonstrated levels of mRNA for FOXP3 that were comparable to that seen in bead-activated natural Treg (10–12 fold increase above baseline). This resulted in increased expression at the protein level in 60% of treated cells (5% of PBS-treated cells). Decitabine also induced Foxp3 expression in 80% of murine CD4+CD25- T cells stimulated two days with CD3/CD28 beads and hIL-2 (10 u/ml), followed by two days of Decitabine treatment. The CD4+CD25-T cells expanded up to two fold during the stimulation and Decitabine treatment. One mouse spleen generated more than 5 x 106 Treg-like cells which is ten fold more than the number of naïve natural Treg. In addition, these Decitabine-treated CD4+CD25- T cells (dcT) showed suppressor function in a mixed lymphocyte reaction. We next tested whether these cells were able to suppress GvHD in an allogeneic BMT model. Mice that were transplanted with dcT showed significantly higher survival rate and maintained their weight better. The Treg surface markers, such as GITR, CTLA4, CD27, CD69 and CD101, were upregulated in the dcT. Activation of Treg has been known to upregulate GzmB but not GzmA or Prf 1. We found that four fold more of dcT expressed GzmB than PBS-treated cells did (95.4% vs. 24.5%). Interestingly, Decitabine treatment induced GzmA expression in over 50% of the cells while about 5% of PBS-treated cells and natural Treg expressed the gene. Using Prf 1 KO and GzmB KO mice, we found that the suppressor function of the dcT is partially dependent on Prf 1, but not on GzmB. To identify cytokine(s) responsible for the suppressor function of these dcT, we analyzed supernatant from these cells and found that the level of IL-10, which has been known to be upregulated in Treg and involved in suppressor function, was upregulated in dcT. The gene expression profiles were also compared between dcT, naïve natural Treg, and bead-activated natural Treg. We found that dcT expressed high level of CTLA-4, GITR, CD25 and Prf 1 which have been known to be upregulated in natural Treg. In summary, Decitabine-treatment of CD4+CD25- T cells enhanced Foxp3 expression. These dcT showed suppressor function in both in vitro and in vivo and shared part of gene signature of natural Treg.

RUNX1 is implicated in over 30 different translocations in human acute leukaemia. RUNX1 can either activate or repress transcription of key regulators of cell growth and differentiation through binding to promoters or enhancer elements. The ETV6/RUNX1 (TEL/AML1) fusion resulting from t(12;21) translocation is the most common chromosomal aberration in paediatric cancers (25% of ALL). The ETV6 part of the fusion protein contains domains interacting with the mSin3, N-CoR and HDAC-3 corepressors. A part of the RUNX1 gene involved in the fusion carries DNA-binding domain. RUNX1 regulates haematopoietic myeloid cell differentiation and transcriptional activation but the role in lymphoid development is not yet fully understood. We hypothesize that ETV6/RUNX1 causes pathological differentiation block in lymphoid cells via repression of target genes of RUNX1 by histone deacetylation. Therefore, we studied the role of histone deacetylase inhibitors (HDACi). We have previously confirmed specific effect of HDACi (valproate-VPA, Trichostatin A-TSA) on ETV6/RUNX1 leukaemic cells in comparison with lymphoblastic leukaemias with different mechanism of leukaemogenesis (BCR/ABL and PDGFRβ/ETV6). To prove the direct effect of HDACi on ETV6/RUNX1 in vitro, we utilized a target gene of RUNX1, granzyme B (GZMB). To determine whether ETV6/RUNX1 represses GZMB via direct interaction with RUNX1-binding site at GZMB promoter, luciferase activity was measured in HeLa cells transfected with pcDNA3.1-ETV6/RUNX1Myc and compared with HeLa with pcDNA3.1 empty vector. Cells were transfected with pGZMB-luc or pGL3-basic to normalize the luciferase activity (pGZMB-luc/pGL3-basic). Fold change of ~3 RLU indicated that GZMB was downregulated by ETV6/RUNX1. To test the direct effect of HDACi on ETV6/RUNX1, after incubation of HeLa cells with VPA and TSA, luciferase activity was monitored again. Repression activity was reduced in treated transfected HeLa cells to 53% after VPA administration and 49% after TSA administration when compared to untreated cells. We subsequently used the effect of HDACi on ETV6/RUNX1-positive leukaemic cells for identification of (ETV6/)RUNX1 target genes in the lymphoid progenitor cells. Expression profiling analysis confirmed changes in the expression of group of genes differentially expressed in ETV6/RUNX1-positive and -negative leukaemic cells. In selected group of genes with known role in the cell cycle regulation (JunD, ACK1, PDGFRB and TCF4) expression changes were confirmed by quantitative expression analysis (qRT-PCR). In established HeLa model direct interaction of ETV6/RUNX1 with promoter sites of these genes will be confirmed. This work shows for the first time direct transcription repression by ETV6/RUNX1 on GZMB gene model and further it shows reversibility of this effect by HDACi. These data support our hypothesis that histone deacetylase inhibitors affect ETV6/RUNX1-positive cells by direct disruption of its repression activity, and that treatment with HDACi may release pathological differentiation block caused by ETV6/RUNX1 chimeric transcription factor.

Peripheral T-cell lymphomas (PTCL) constitute a heterogeneous and aggressive group of tumors whose pathogenic alterations remain largely unknown. They show great morphologic, immunophenotypic and clinical differences. Expression profiling has been demonstrated to be a useful tool for the molecular classification of tumours. We used DNA microarrays (Affymetrics, Lymph_DX) to look for molecular differences in peripheral T-cell lymphomas. We analysed the expression of 19 pathologically confirmed PTCL (11 PTCL, unspecified (u) and 8 angioimmunoblastic, AILT). A time-course of stimulation with anti-CD3 and anti-CD28 was performed in CD4+ and CD8+ isolated peripheral blood lymphocytes, and the gene expression of tumors was compared with resting and stimulated normal T-cells. Cases of classical Hodgkin’s lymphoma, with which PTCL may be confused, were included. Gene expression of PTCL revealed several clusters of genes with biological significance showing differential expression within the tumors. We defined a “Cytotoxic Response” cluster characterized by the expression of cytotoxic molecules PRF1 and GZMB and GZMH, a “B-cell-related genes” cluster which included B-cell specific genes PAX5 or CD19, and an interesting “Resting-cells genes” cluster. These three clusters were negatively correlated with a “Proliferation” signature. On the other hand, the “Proliferation” signature significantly correlated with the expression of TNFRSF8 (CD30) (r=0.63, p=0.003) and the expression of CD3G (r=0.51, p=0.022). Other interesting genes were also found associated specifically to the proliferation of PTCL. Comparison between PTCLu and AILT showed differences in the expression of CD21 and other genes reflecting the higher presence of dendritic cells in AILT, and also revealed differences in expression of relevant chemokines, such us CXCL5, CXCL13 and CCL19. Intriguingly, compared with normal CD4 and CD8 T-cells, PTCL showed lower expression of most of the typical genes associated with T-cell biology, suggesting that PTCL often lose the common genetic program of T-cells. PTCL expression profiles are difficult to interpret due to the significant proportion of other infiltrating cells accompanying the tumor. Our results showed that microarrays are a helpful tool to dissect the PTCL expression profile, identifying those genes expressed by infiltrating cells and those expressed by tumor cells.

We read with great interest the recently published article by Wagner et al[1][1] illustrating the constitutive expression of granzyme B (GzmB) and perforin in human polymorphonuclear neutrophils (PMNs). The authors reported for the first time that GzmB and perforin are expressed not only in natural