ALBERT

Description: Conventional multilevel inverter topologies like neutral point clamped (NPC), flying capacitor (FC), and cascade H bridge (CHB) are employed in the industry but require a large number of switches and passive and active components for the generation of a higher number of voltage levels. Consequently, the cost and complexity of the inverter increases. In this work, the basic unit of a switched capacitor topology was generalized utilizing a cascaded H-bridge structure for realizing a switched-capacitor multilevel inverter (SCMLI). The proposed generalized MLI can generate a significant number of output voltage levels with a lower number of components. The operation of symmetric and asymmetric configurations was shown with 13 and 31 level output voltage generation, respectively. Self-capacitor voltage balancing and boosting capability are the key features of the proposed SCMLI structure. The nearest level control modulation scheme was employed for controlling and regulating the output voltage. Based on the longest discharging time, the optimum value of capacitance was also calculated. A generalized formula for the generation of higher voltage levels was also derived. The proposed model was simulated in the MATLAB®/Simulink 2016a environment. Simulation results were validated with the hardware implementation.

Description: Key Points High ST2 and TIM3 at day 28 after allogeneic HCT were associated with nonrelapse mortality and overall survival at 2 years. Low day 28 L-Ficolin was associated with VOD/SOS and high CXCL9 correlated with chronic GVHD.

Description: Introduction We report the phase I data from ongoing phase I/II study of combination of targeted agents sorafenib, vorinostat and bortezomib in poor-risk AML. Our findings from a previous phase I study, performed at Indiana University, of the combination of sorafenib and vorinostat in patients with AML suggested that two major groups of patients may benefit most from this targeted regimen, patients with FLT3-ITD mutation, and those with complex or poor-risk cytogenetics (monosomy 5 or 7). In addition, our findings were suggestive of a synergistic action obtained by inhibition of p52NFKB, a down-stream target of proteosome inhibition. With the hypothesis that addition of proteasome inhibitor bortezomib would be of benefit towards such synergism, a phase I/II clinical trial combining bortezomib with sorafenib and vorinostat was initiated. The phase I data is reported here. Methods The phase I portion of the trial utilized a traditional 3+3 design on five cohorts to determine the MTD of the combination. Eligibility required age ≥18, a confirmed baseline diagnosis of AML by the revised guidelines of the International Working Group for AML, and included untreated disease in elderly or relapsed/refractory disease in all ages, monosomy 5,7 or complex cytogenetics or positive FLT3-ITD mutation, ECOG PS 0-2, and adequate kidney and liver function. Dose limiting non-hematologic toxicity was defined per the CTCAE v4.0 criteria. Hematologic toxicity was prolonged cytopenia with 42 days after discontinuation of therapy. The treatment was given in cycles, with each cycle consisting of 2 weeks treatment followed by 1 week off. Dose and/or administration schedule of drugs were escalated between the cohorts. Results Seventeen patients were enrolled on the phase I portion. Fifteen patients completed at least one cycle of treatment and 2 were taken off earlier due to disease progression. The median age was 51 years (24-73), and 10 (59%) patients were male. Sixteen patients had prior therapy at time of enrollment and 59% were heavily pretreated (≥3 lines of therapy) including stem cell transplantation in 29%. Fifty nine percent had FLT3-ITD mutation and 53% had poor-risk cytogenetics. No DLTs were seen in all 5 cohorts and MTD was not reached. The safe dose for phase II was determined at sorafenib 400 mg bid, vorinostat 200 mg bid (both for 14 days), and bortezomib 1.3 mg/m2IV on days 1,4,8,11, every 21 days. Most common grade 1-2 toxicities were diarrhea (59%), nausea (41%), vomiting (24%) and rash (18%). Majority of toxicities were grade 1. Response was observed in 6 patients (40%) with 4 achieving a complete remission (27%). All responders had relapse/refractory disease. Conclusion The combination of sorafenib, vorinostat, and bortezomib when given with a 2-week on, 1-week off schedule is safe with minimal side effects, and tolerable as an outpatient regimen for the treatment of poor-risk AML. Encouraging responses with this regimen are seen in these patients. The phase II portion of the study is currently ongoing. Correlative studies to further elucidate the molecular attributes of efficacy of this regimen in poor-risk AML are underway. Updated results will be presented. Disclosures No relevant conflicts of interest to declare.

Description: Background: Clonal heterogeneity is a known issue in multiple myeloma (MM) and the emergence of drug resistant clones is responsible for the incurability of the disease. Multiple studies of bulk CD138+ bone marrow samples have attempted to stratify MM patients into smaller, more distinct, patient risk groups based on molecular phenotypes. Recently, single cell RNA sequencing (scRNA-seq) technology has been applied in MM to identify cell clones. This leads to a new question: can we classify patients with scRNA-seq data guided by previously defined subtypes, and how do the single cell results correspond with the classification? Methods: We developed a novel, deep transfer learning framework to predict MM patient subtypes in patients with scRNA-seq based on patient classifications from microarray data. While the problem of scRNA-seq batch corrections has been intensively studied using transfer learning, there has been less work on similar comparisons between scRNA-seq and patient-level data. To address this issue, we utilized domain adaptation, a specific transfer learning approach, to combine scRNA-seq profiles and patient-level microarray data using a multitask learning framework. Figure 1 illustrates our computational framework. Its aim is to classify both cells and patients (with scRNA-seq data) according to patient level classifications derived from previous gene expression profiling studies for MM. Specifically, we adopted the 10-subtype classifications derived from microarray data1. Patients with scRNA-seq were summarized into a single vector by averaging gene counts across all the cells. Gene expression profiling data (including scRNA-seq and microarray) for MM patients from multiple studies were input into the transfer learning network consisting of 5 hidden layers. The last hidden layer was used to calculate the maximum mean discrepancy (MMD) between the patients from scRNA-seq and microarray to integrate the datasets. The datasets in this study are summarized in Table 1. Two microarray datasets (GSE19784, GSE2658) and one scRNA-seq dataset (GSE117156) were obtained from NCBI Gene Expression Omnibus. IUSM data were locally generated. One microarray and one scRNA-seq dataset were used in training and testing. GSE19784 was split into 80% training and 20% testing. GSE117156, due to the smaller sample size (11 patients), was split into 90% training and 10% testing. We ran 20 rounds of random cross validation using TensorFlow on a GTX1080 GPU. The expression profiles of patients and single cells from all datasets (GSE19784, GSE117156, GSE2658, IUSM) were input into the trained model after each round of cross validation to produce low-dimensional representations and predictions for each training, testing, and validation sample. Results: We found that our model was able to identify signals in the data based on expression profiles from patient-level and single cell data. The patient classification labels can be consistently reproduced in a held-out test set of patients as well as in a validation cohort of microarray data from 559 MM patients (GSE2658) and scRNA-seq from 4 MM patients from IUSM (Figure 2). These results show that the model can learn the subtypes across multiple datasets and platforms. The 4 IUSM patients tended to cluster similarly to their individual CD138+ cells after training, while GSE2658 patients still maintained some separation between MM subtype clusters (Figure 3). The single cells from our cohort of 4 patients did not necessarily classify to the same subtype as their patient. Conclusions: We found that a domain adaptive classifier can be trained across scRNA-seq and bulk gene expression profiling data from MM patients to integrate data and transfer knowledge. These models showed that single cells within a patient do not necessarily match the patient level molecular characteristics. Not surprisingly, similar results have been found in other cancer types2. As our novel framework is further refined and more patients are sequenced, we expect more unique insights into both inter- and intra-tumor MM heterogeneity. References: 1. Broyl A, Hose D, Lokhorst H, et al. Gene expression profiling for molecular classification of multiple myeloma in newly diagnosed patients. Blood. 2010;116(14):2543-2553. 2. Patel AP, Tirosh I, Trombetta JJ, et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science. 2014;344(6190):1396-1401. Disclosures Abonour: Celgene: Consultancy, Research Funding; BMS: Consultancy; Takeda: Consultancy, Research Funding; Janssen: Consultancy, Research Funding. Roodman:Amgen: Membership on an entity's Board of Directors or advisory committees.

Description: Objective: Multiple myeloma (MM) is the second most common blood cancer in the United States. Clinical and pathological indicators are inadequate for accurate assessing prognosis of MM patients1. The objective of this study was to develop a robust computational model for predicting prognosis of MM patients by inferring transcription factor (TF) activities based on global gene expression measurements. Methods: The workflow is illustrated in Figure 1. Gene expression data from MM patients were retrieved from the Multiple Myeloma Research Foundation CoMMpass study2 (MMRF) and Gene Expression Omnibus (GEO). MM-specific TF-target gene relationships were derived from the ENCODE database3 and co-expression analysis was based on MMRF data. The activity for each TF was inferred using a rank-based enrichment approach based on the target gene expression measurements in individual patients4. A prognostic model was developed using Cox regression analysis based on the inferred TF activities that demonstrated the highest association with patient survival. The TF model performance was further evaluated using data from four independent MM cohorts, in which gene expression levels were measured using different platforms. Results: Based on MMRF data (767 patients), the inferred activity of four TFs, BATF, E2F1, MYBL2, and RAD21, showed the strongest association with patient survival. The derived prognostic model was validated in four independent MM data sets with a total of 1501 patients. Importantly, gene expression levels in the validation data sets were measured using microarray, which is different from the RNAseq used in training data sets, and supports the robustness of the TF activity model across different platforms (Figure 2). In addition, the TF activity model was able to predict clinical outcomes of patients that received bortezomib treatment, and it also revealed an association with metastasis in myeloma cell line xenograft experiments5 (Figure 3). Conclusion: We developed a highly robust prognostic model for predicting MM clinical outcomes. The inferred TF activity further improves our understanding of the underlying regulatory basis of MM progression. Reference 1. Rajkumar SV. Multiple myeloma: 2012 update on diagnosis, risk-stratification, and management. Am J Hematol. 2012;87(1):78-88. 2. l, a longitudinal study in multiple myeloma relating clinical outcomes to genomic and immunophenotypic profiles: Am Soc Hematology; 2013. 3. Consortium EP. The ENCODE (ENCyclopedia of DNA elements) project. Science 2004;306(5696):636-640. 4. Moerman T, Aibar Santos S, Bravo González-Blas C, et al. GRNBoost2 and Arboreto: efficient and scalable inference of gene regulatory networks. Bioinformatics. 2018;35(12):2159-2161. 5. Shen Y, Mishima Y, Shi J, et al. Deciphering Clonal Evolution and Dissemination of Multiple Myeloma Cells In Vivo: Am Soc Hematology; 2018. Disclosures Abonour: Celgene: Consultancy, Research Funding; BMS: Consultancy; Takeda: Consultancy, Research Funding; Janssen: Consultancy, Research Funding. Roodman:Amgen: Membership on an entity's Board of Directors or advisory committees.

Description: Treatment of acute myeloid leukemia (AML) has changed little over the last several decades and prognosis remains very poor. Allogeneic hematopoietic cell transplantation (allo-HCT) is one potentially curative option for relapsed or high-risk AML. The immunotherapeutic activity of allo-HCT is known as the graft-vs-leukemia (GVL) activity. However, GVL activity is often accompanied by T-cell reactivity to allo-antigens in normal host tissues, which leads to graft-versus-host disease (GVHD), another major cause of death after HCT with relapse. Therefore, there is a great unmet need to improve the current process of allo-HCT through increasing the GVL activity and decreasing GVHD. We have shown that an elevated plasma level of soluble (s)ST2 in HCT patients is a risk factor for severe GVHD (Vander Lugt et al, N Engl J Med, 2013) and that ST2 blockade reduces sST2-producing T cells while maintaining protective membrane (m)ST2-expressing T cells during GVHD (Zhang et al, Sci Transl Med, 2015). In addition, the interleukin (IL)-9-producing CD4 T helper (Th)9 and CD8 cytotoxic T (Tc)9 cell subsets (together T9 cells) have higher antitumor activity than Th1 and Tc1 cells in melanoma models (Lu et al, J Clin Invest, 2012 and Lu et al, ProcNatl Acad Sci, 2014). We hypothesized that activation of the ST2/IL-33 pathway in T9 cells will both alleviate GVHD and increase GVL. In our laboratory, we have shown that T9 cells express a high level of mST2 and that differentiation of total T cells into T9 cells in the presence of IL-33 (T9IL-33 cells) increases expression of mST2 (Figure 1A) and PU.1 (Figure 1B), a transcription factor that promotes IL-9 production on both CD4 and CD8 T cells. Adoptive transfer of T9IL-33 cells with bone marrow cells in a murine model of HCT resulted in less severe GVHD compared to transfer of T9IL33 cells generated from ST2-/- or IL-9-/- T cells (Figure 1C). Ex-vivo analysis of target organs such as gut showed a decrease in T9IL-33 interferon (IFN)g-producing T cells that was abolished in mice receiving T9IL-33 cells derived from ST2-/- or IL-9-/- T cells (not shown). Furthermore, T9IL-33 cells revealed higher anti-leukemic activity in vitro when cultured with a B cell lymphoma line (A20) or retrovirally transduced MLL-AF9 leukemic cells in cytolytic assays (not shown). In vivo GVL experiments with MLL-AF9 induced leukemia, and adoptive transfer of T9IL-33 cells resulted in increased survival compared to transfer of T9IL-33 cells generated from ST2-/- or IL-9-/- T cells (Figure 1D). Human T9 cells are poorly explored. We demonstrated that differentiation of human T9 cells in the presence of IL-33 enhanced IL-9 production by CD4 and CD8 T cells (Figure 2A). T9IL-33 cells also upregulated the expression of the cytolytic molecules granzymes A and B compared to T9 cells (Th9IL-33: 33.6%±4%, vs. Th9: 15.69%±2.53% p=0.021), (Tc9IL-33: 57.6%±4.7%, vs. Tc9: 34.61%±3.4% p=0.018), as well as demonstrated higher in vitro anti- leukemic cytolytic activity when incubated with MOLM14, an aggressive AML tumor cell line expressing FLT3/ITD mutations (Figure 2B). Transcriptome analysis of T9IL-33 cells from wild-type and ST2-/- T cells showed upregulation of molecules implicated in anti-leukemic activity (granzymes A and B, CD8α, IL-15, IL-15rα, IFNα, and IL-1α) on both CD4 and CD8 T cells (Figure 2C), and such upregulation was confirmed at the protein level (Figure 2D). Furthermore,investigations into the possible mechanism of activation using transwell assays revealed that both soluble factors and cell contact between Th9IL-33 and Tc9IL-33 T cells were required for maximum killing (not shown). We next investigated the possible mechanism of action and hypothesized that CD8α might be the contact-dependent component. CD8α blockade with neutralizing antibody during human T9IL-33 differentiation reduced the cytotoxicity of both murine T9IL-33 and human T9IL-33 cells (Figure 2E). Altogether, our observations suggest that adoptive transfer of T9IL-33 cells represents a promising cellular therapy following HCT. Figure 1. Figure 1. Figure 2. Figure 2. Disclosures Paczesny: Viracor laboratories: Patents & Royalties: "Methods of detection of graft-versus-host disease" (US- 13/573,766).