ALBERT

Beschreibung: Journal of Chemical & Engineering Data DOI: 10.1021/acs.jced.8b00845

Beschreibung: Debris flows have been always a serious problem in the mountain areas. Research on the assessment of debris flows susceptibility (DFS) is useful for preventing and mitigating debris flow risks. The main purpose of this work is to study the DFS in the Shigatse area of Tibet, by using machine learning methods, after assessing the main triggering factors of debris flows. Remote sensing and geographic information system (GIS) are used to obtain datasets of topography, vegetation, human activities and soil factors for local debris flows. The problem of debris flow susceptibility level imbalances in datasets is addressed by the Borderline-SMOTE method. Five machine learning methods, i.e., back propagation neural network (BPNN), one-dimensional convolutional neural network (1D-CNN), decision tree (DT), random forest (RF), and extreme gradient boosting (XGBoost) have been used to analyze and fit the relationship between debris flow triggering factors and occurrence, and to evaluate the weight of each triggering factor. The ANOVA and Tukey HSD tests have revealed that the XGBoost model exhibited the best mean accuracy (0.924) on ten-fold cross-validation and the performance was significantly better than that of the BPNN (0.871), DT (0.816), and RF (0.901). However, the performance of the XGBoost did not significantly differ from that of the 1D-CNN (0.914). This is also the first comparison experiment between XGBoost and 1D-CNN methods in the DFS study. The DFS maps have been verified by five evaluation methods: Precision, Recall, F1 score, Accuracy and area under the curve (AUC). Experiments show that the XGBoost has the best score, and the factors that have a greater impact on debris flows are aspect, annual average rainfall, profile curvature, and elevation.

Beschreibung: The snow-albedo feedback is a crucial component in high-altitude cryospheric change but is poorly quantified over the Third Pole, encompassing the Karakoram and Tibetan Plateau. Here we present an analysis of present-day and future spring snow-albedo feedback over the Third Pole, using a 28 year satellite-based albedo and the latest climate model simulations. We show that present-day spring snow-albedo feedback strength is primarily determined by the decrease in albedo due to snow metamorphosis, rather than that due to reduced snow cover in the Karakoram, but not found in Southeastern Tibet. We further demonstrate an emergent relationship between snow-albedo feedback from the seasonal cycle and that from climate change across models. Combined with contemporary satellite-based snow-albedo feedback from seasonal cycle, this relationship enables us to estimate that the feedback strength for the Karakoram with a relatively high glaciated area is −2.42 ± 0.48% K−1 under an unmitigated scenario, which is much stronger than that for Southeastern Tibet (−1.64 ± 0.48% K−1) and for the Third Pole (−0.89 ± 0.44% K−1), respectively. Moreover, it is noteworthy that the magnitude of the constrained strength is only half of the unconstrained model estimate for the Third Pole, suggesting that current climate models generally overestimate the feedback of spring snow change to temperature change based on the unmitigated scenario. ©2018. American Geophysical Union. All Rights Reserved.

Beschreibung: Abstract 3584 Chronic lymphocytic leukemia (CLL) is typically characterized by defects in programmed cell death rather than alterations in cell cycle regulation. Transforming growth factor β (TGFβ), a ubiquitously expressed growth factor, regulates multiple normal cellular responses including proliferation, differentiation, migration and apoptosis. Loss of growth inhibition by TGFβ is thought to contribute to the development and progression of a variety of tumors including CLL (DeCoteau et al., PNAS 1997). Approximately 40% of patients contain mutations in the signal sequence of TGFβ receptor 1 (TBR-1) in the form of substitutions or deletions (Schiemann et al., Cancer Detect Prev 2004). In the wild type form, the signal sequence contains a nine alanine stretch, which if truncated has been shown to impair signaling through the receptor and specifically, a truncated, six alanine form is associated with increased cancer risk (Pasche et al., Cancer Res 1999). TGFβ signaling can regulate expression of micoRNAs (miRNA), which are ~22 nucleotide-long RNA gene regulators. Deregulated miRNA expression has been implicated in tumorigenesis, including CLL. Several miRNAs have been shown to be over-expressed in CLL as compared to normal B cells (Fulci et al., Blood 2007). This includes miR-155, which is part of a 13-miRNA signature that has prognostic implications, including a shorter need-for-treatment interval (Calin et al., N Engl J Med 2005). Interestingly, miR-155 has been shown to be upregulated by TGFβ in murine mammary gland cells (Kong et al., Mol Cell Biol 2008). The goals of our study are to investigate the link between TGFβ signaling and miR-155 in CLL and to determine how the interaction between the two may contribute to the pathogenesis of CLL. Here we show that miR-155 is in fact upregulated by TGFβ in mouse splenic B cells and in human peripheral blood B cells. In CLL, miR-155 expression inversely correlates with the proportion of CLL cells harboring signal sequence mutation in TBR-1, consistent with miR155 regulation by TGFβ in vivo. To understand the role of TGFβ-induced miR-155 in CLL pathobiology, identification of specific target genes in the context of this disease is essential. To this end, we compared the gene (cDNA) expression profile between CLL with high miR-155 vs. low miR-155 expression and identified putative miR-155 target genes by selecting those genes that are differentially expressed in SAM analysis with lower expression in the high miR-155 group, and which harbor predicted miR-155 binding sites in their 3’ untranslated region (UTR). Based on this algorithm, we have identified casein kinase 1 gamma 2 (CSK1γ2) as a target for miR155 in CLL. CSK1γ2 is a negative modulator of the TGFβ signaling pathway by targeting the phosphorylated form of SMAD3 for degradation (Guo et al., Oncogene 2008). MiR-155 represses luciferase reporter gene expression by specific binding to the miR-155 site in the CSK1γ2 3’UTR. In addition, we found that CSK1γ2 itself is upregulated in B cells upon TGFβ stimulation, and treatment of human B cells with PNA miR-155 inhibitor (Fabani et al., Nucleic Acids Research 2010) further increases CSK1γ2 mRNA levels. Surprisingly, comparison of CSK1γ2 protein levels between CLLs with high or low miR-155 by Western blotting revealed higher CSK1γ2 protein expression despite lower CSK1γ2 mRNA levels, suggesting that miR-155 may enhance CSK1γ2 translation in CLL cells and implying an intriguing regulatory interaction between miR-155 and CSK1γ2. In summary, our data indicates that the variation of miR-155 seen in CLL is primarily a function of TGFβ signaling activity. Moreover, miR-155 is an important player in a complex auto-regulatory network in TGFβ signaling by fine-tuning the negative feedback mechanism on TGFβ signaling mediated by CSK1γ2. In CLL cells harboring TBR-1 with wild-type signal sequence, higher miR-155 levels may help modulate the TGFβ signaling activity to a level optimal for the survival or other pathobiological functions of CLL. Furthermore, since CLL cells are predominantly non-proliferating, our findings that miR-155 may enhance translation of CSK1γ2 provide support to the model of cell cycle dependence of microRNA functions (Vasudevan et al., Cell Cycle 2008). Disclosures: No relevant conflicts of interest to declare.

Beschreibung: Abstract 3471 PRDM1/Blimp1, a lymphoma tumor suppressor and master regulator of plasma cell differentiation, is inactivated by genomic mutation and deletion in a subset of activated B-cell-type diffuse large B-cell lymphomas (DLBCL), underscoring the importance of terminal differentiation impairment in lymphoma pathogenesis. The role of PRDM1 inactivation in Burkitt lymphoma (BL) is currently not known. Two findings, however, may suggest possible PRDM1 involvement in BL pathogenesis. First, although PRDM1 protein is consistently absent in BL, about 40% of BL cases are positive for IRF4/MUM1, suggesting asynchronous expression of these two proteins. Secondly, it has been hypothesized that microRNA-127 overexpression in EBV+ BL may promote BL development through down-regulation of PRDM1. Sequence analysis of the coding region of PRDM1 in BL cell lines and primary tumors did not demonstrate any mutational changes. To investigate whether epigenetic inactivation of PRDM1 may play a pathogenetic role in BL, BL cell lines and primary cases were bisulfite sequenced to assess the methylation status of 41 CG dinucleotides in a 601 base-pair region spanning the distal promoter (DP) and a CG island that extends from the proximal promoter to the first exon of PRDM1. These include (i) 11 BL cell lines, 9 EBV+ (4 latency I, 2 Wp-restricted, and 3 with latency III drift) and 2 EBV-; (ii) 7 EBV+ lymphoblastoid cell line (LCL); (iii) 62 primary BL cases, formalin-fixed, paraffin-embedded tissues (FFPE); (iv) 4 B cell lymphoma, unclassifiable, with features intermediate between DLBCL and BL (BCL-U), FFPE. Naïve, germinal center (GC) and memory B cells sorted by flow cytometry were also analyzed as controls. Hypermethylation in PRDM1 promoter and exon 1 was seen in all EBV+ BL cell lines except one with latency III, but not in EBV- BL lines. None of the LCL demonstrated PRDM1 hypermethylation, implying that EBV is unlikely to mediate PRDM1 hypermethylation directly. Methylation status of PRDM1 was successfully determined in 39 BLs (30 sporadic, 9 HIV-related) and 5 BCL-Us (4 sporadic, 1 HIV-related). Hypermethylation was seen in 11 BLs and 1 BCL-U (28.6% of total). All the methylated cases were EBV(+) (p=0.004). Overall, 12 of 28 (43%) EBV(+) BL and BCL-U cases exhibited hypermethylation. PRDM1 hypermethaylation was independent of BL subtype and MUM1/IRF4 expression status. To determine if PRDM1 hypermethylation can potentially repress PRDM1 transcription, BL cell lines Ramos (PRDM1 unmethylated) and Mutu1 (PRDM1 hypermethylated) are treated with IL-21 (50ng/mL), a cytokine that can mediate terminal differentiation by inducing PRDM1, for 2 to 5 days. While PRDM1 is induced by IL-21 in Ramos, levels of PRDM1 are not significantly increased in Mutu1, even though STAT3, a downstream mediator of IL-21, is activated in the latter. This finding suggests that PRDM1 hypermethylation has the potential to repress PRDM1 transcription in the presence of PRDM1-inducing signals. This function, however, may be pathogenetically more relevant in a BL precursor cell than in the final BL tumor cell. EBV+ BL is thought to derive from an EBV-infected late germinal center (GC) B cell, which harbors c-myc translocation and begins, though abortively, memory B cell differentiation and adopts an EBV latency I form. BL consistently express BCL6, a known PRDM1 transcription repressor, and harbor very low levels of PRDM1 mRNA. This suggests that high BCL6 expression, rather than PRDM1 hypermethylation, may be the primary determinant of its repressed transcription seen in BL. Indeed, PRDM1 mRNA expression is 25 to 1200 fold higher in LCLs which harbor very low levels of BCL. Furthermore, reducing expression of MTA3, which forms a complex with BCL6 and mediates PRDM1 transcription repression, in Daudi increases PRDM1 transcript levels. Therefore, we hypothesize that PRDM1 hypermethylation in BL more likely represents a memory of an epigenetic event that functions in the earlier stage of BL pathogenesis. PRDM1 hypermethylation may function to repress its transcription in a BL precursor cell as it is exposed to inducing signals, e.g. IL-21, during GC transit. Our study expands the spectrum of B cell lymphomas in which PRDM1 plays a tumor suppressor role, and supports the importance of impairment of terminal differentiation in the pathogenesis of a subset of aggressive B cell lymphomas. Disclosures: No relevant conflicts of interest to declare.