Exercise reprograms the inflammatory landscape of multiple stem cell compartments during mammalian aging.

Exercise has the ability to rejuvenate stem cells and improve tissue regeneration in aging animals. However, the cellular and molecular changes elicited by exercise have not been systematically studied across a broad range of cell types in stem cell compartments. We subjected young and old mice to aerobic exercise and generated a single-cell transcriptomic atlas of muscle, neural, and hematopoietic stem cells with their niche cells and progeny, complemented by whole transcriptome analysis of single myofibers. We found that exercise ameliorated the upregulation of a number of inflammatory pathways associated with old age and restored aspects of intercellular communication mediated by immune cells within these stem cell compartments. Exercise has a profound impact on the composition and transcriptomic landscape of circulating and tissue-resident immune cells. Our study provides a comprehensive view of the coordinated responses of multiple aged stem cells and niche cells to exercise at the transcriptomic level.

The disordered N-terminal domain of DNMT3A recognizes H2AK119ub and is required for postnatal development.

DNA methyltransferase 3a (DNMT3A) plays a crucial role during mammalian development. Two isoforms of DNMT3A are differentially expressed from stem cells to somatic tissues, but their individual functions remain largely uncharacterized. Here we report that the long isoform DNMT3A1, but not the short DNMT3A2, is essential for mouse postnatal development. DNMT3A1 binds to and regulates bivalent neurodevelopmental genes in the brain. Strikingly, Dnmt3a1 knockout perinatal lethality could be partially rescued by DNMT3A1 restoration in the nervous system. We further show that the intrinsically disordered N terminus of DNMT3A1 is required for normal development and DNA methylation at DNMT3A1-enriched regions. Mechanistically, a ubiquitin-interacting motif embedded in a putative α-helix within the N terminus binds to mono-ubiquitinated histone H2AK119, probably mediating recruitment of DNMT3A1 to Polycomb-regulated regions. These data demonstrate an isoform-specific role for DNMT3A1 in mouse postnatal development and reveal the N terminus as a necessary regulatory domain for DNMT3A1 chromatin occupancy and functions in the nervous system.

Energy Efficient Pupil Tracking Based on Rule Distillation of Cascade Regression Forest.

As the demand for human-friendly computing increases, research on pupil tracking to facilitate human-machine interactions (HCIs) is being actively conducted. Several successful pupil tracking approaches have been developed using images and a deep neural network (DNN). However, common DNN-based methods not only require tremendous computing power and energy consumption for learning and prediction; they also have a demerit in that an interpretation is impossible because a black-box model with an unknown prediction process is applied. In this study, we propose a lightweight pupil tracking algorithm for on-device machine learning (ML) using a fast and accurate cascade deep regression forest (RF) instead of a DNN. Pupil estimation is applied in a coarse-to-fine manner in a layer-by-layer RF structure, and each RF is simplified using the proposed rule distillation algorithm for removing unimportant rules constituting the RF. The goal of the proposed algorithm is to produce a more transparent and adoptable model for application to on-device ML systems, while maintaining a precise pupil tracking performance. Our proposed method experimentally achieves an outstanding speed, a reduction in the number of parameters, and a better pupil tracking performance compared to several other state-of-the-art methods using only a CPU.

Light-Weight Student LSTM for Real-Time Wildfire Smoke Detection.

As the need for wildfire detection increases, research on wildfire smoke detection combining low-cost cameras and deep learning technology is increasing. Camera-based wildfire smoke detection is inexpensive, allowing for a quick detection, and allows a smoke to be checked by the naked eye. However, because a surveillance system must rely only on visual characteristics, it often erroneously detects fog and clouds as smoke. In this study, a combination of a You-Only-Look-Once detector and a long short-term memory (LSTM) classifier is applied to improve the performance of wildfire smoke detection by reflecting on the spatial and temporal characteristics of wildfire smoke. However, because it is necessary to lighten the heavy LSTM model for real-time smoke detection, in this paper, we propose a new method for applying the teacher-student framework to deep LSTM. Through this method, a shallow student LSTM is designed to reduce the number of layers and cells constituting the LSTM model while maintaining the original deep LSTM performance. As the experimental results indicate, our proposed method achieves up to an 8.4-fold decrease in the number of parameters and a faster processing time than the teacher LSTM while maintaining a similar detection performance as deep LSTM using several state-of-the-art methods on a wildfire benchmark dataset.

Haploinsufficiency of cohesin protease, Separase, promotes regeneration of hematopoietic stem cells in mice.

Cohesin recently emerged as a new regulator of hematopoiesis and leukemia. In addition to cohesin, whether proteins that regulate cohesin's function have any direct role in hematopoiesis and hematologic diseases have not been fully examined. Separase, encoded by the ESPL1 gene, is an important regulator of cohesin's function. Canonically, protease activity of Separase resolves sister chromatid cohesion by cleaving cohesin subunit-Rad21 at the onset of anaphase. Using a Separase haploinsufficient mouse model, we have uncovered a novel role of Separase in hematopoiesis. We report that partial disruption of Separase distinctly alters the functional characteristics of hematopoietic stem/progenitor cells (HSPCs). Although analyses of peripheral blood and bone marrow of Espl1+/Hyp mice broadly displayed unperturbed hematopoietic parameters during normal hematopoiesis, further probing of the composition of early hematopoietic cells in Espl1+/Hyp bone marrow revealed a mild reduction in the frequencies of the Lin- Sca1+ Kit- (LSK) or LSK CD48+ CD150- multipotent hematopoietic progenitors population without a significant change in either long-term or short-term hematopoietic stem cells (HSCs) subsets at steady state. Surprisingly, however, we found that Separase haploinsufficiency promotes regeneration activity of HSCs in serial in vivo repopulation assays. In vitro colony formation assays also revealed an enhanced serial replating capacity of hematopoietic progenitors isolated from Espl1+/Hyp mice. Microarray analysis of differentially expressed genes showed that Separase haploinsufficiency in HSCs (SP-KSL) leads to enrichment of gene signatures that are upregulated in HSCs compared to committed progenitors and mature cells. Taken together, our findings demonstrate a key role of Separase in promoting hematopoietic regeneration of HSCs.

The safety and clinical effects of administering a multiantigen-targeted T cell therapy to patients with multiple myeloma.

Multiple myeloma (MM) is an almost always incurable malignancy of plasma cells. Despite the advent of new therapies, most patients eventually relapse or become treatment-refractory. Consequently, therapies with nonoverlapping mechanisms of action that are nontoxic and provide long-term benefit to patients with MM are greatly needed. To this end, we clinically tested an autologous multitumor-associated antigen (mTAA)-specific T cell product for the treatment of patients with high-risk, relapsed or refractory MM. In this study, we expanded polyclonal T cells from 23 patients with MM. T cells whose native T cell receptors were reactive toward five myeloma-expressed target TAAs (PRAME, SSX2, MAGEA4, Survivin, and NY-ESO-1) were enriched ex vivo. To date, we have administered escalating doses of these nonengineered mTAA-specific T cells (0.5 × 107 to 2 × 107 cells/m2) to 21 patients with MM, 9 of whom were at high risk of relapse after a median of 3 lines of prior therapy and 12 with active, relapsed or refractory disease after a median of 3.5 prior lines. The cells were well tolerated, with only two transient, grade III infusion-related adverse events. Furthermore, patients with active relapsed or refractory myeloma enjoyed a longer than expected progression-free survival and responders included three patients who achieved objective responses concomitant with detection of functional TAA-reactive T cell clonotypes derived from the infused mTAA product.

Large DNA Methylation Nadirs Anchor Chromatin Loops Maintaining Hematopoietic Stem Cell Identity.

Higher-order chromatin structure and DNA methylation are implicated in multiple developmental processes, but their relationship to cell state is unknown. Here, we find that large (>7.3 kb) DNA methylation nadirs (termed "grand canyons") can form long loops connecting anchor loci that may be dozens of megabases (Mb) apart, as well as inter-chromosomal links. The interacting loci cover a total of ∼3.5 Mb of the human genome. The strongest interactions are associated with repressive marks made by the Polycomb complex and are diminished upon EZH2 inhibitor treatment. The data are suggestive of the formation of these loops by interactions between repressive elements in the loci, forming a genomic subcompartment, rather than by cohesion/CTCF-mediated extrusion. Interestingly, unlike previously characterized subcompartments, these interactions are present only in particular cell types, such as stem and progenitor cells. Our work reveals that H3K27me3-marked large DNA methylation grand canyons represent a set of very-long-range loops associated with cellular identity.

Dnmt3a loss and Idh2 neomorphic mutations mutually potentiate malignant hematopoiesis.

Mutations in the epigenetic regulators DNMT3A and IDH1/2 co-occur in patients with acute myeloid leukemia and lymphoma. In this study, these 2 epigenetic mutations cooperated to induce leukemia. Leukemia-initiating cells from Dnmt3a-/- mice that express an IDH2 neomorphic mutant have a megakaryocyte-erythroid progenitor-like immunophenotype, activate a stem-cell-like gene signature, and repress differentiated progenitor genes. We observed an epigenomic dysregulation with the gain of repressive H3K9 trimethylation and loss of H3K9 acetylation in diseased mouse bone marrow hematopoietic stem and progenitor cells (HSPCs). HDAC inhibitors rapidly reversed the H3K9 methylation/acetylation imbalance in diseased mouse HSPCs while reducing the leukemia burden. In addition, using targeted metabolomic profiling for the first time in mouse leukemia models, we also showed that prostaglandin E2 is overproduced in double-mutant HSPCs, rendering them sensitive to prostaglandin synthesis inhibition. These data revealed that Dnmt3a and Idh2 mutations are synergistic events in leukemogenesis and that HSPCs carrying both mutations are sensitive to induced differentiation by the inhibition of both prostaglandin synthesis and HDAC, which may reveal new therapeutic opportunities for patients carrying IDH1/2 mutations.

Driver's Facial Expression Recognition in Real-Time for Safe Driving.

In recent years, researchers of deep neural networks (DNNs)-based facial expression recognition (FER) have reported results showing that these approaches overcome the limitations of conventional machine learning-based FER approaches. However, as DNN-based FER approaches require an excessive amount of memory and incur high processing costs, their application in various fields is very limited and depends on the hardware specifications. In this paper, we propose a fast FER algorithm for monitoring a driver's emotions that is capable of operating in low specification devices installed in vehicles. For this purpose, a hierarchical weighted random forest (WRF) classifier that is trained based on the similarity of sample data, in order to improve its accuracy, is employed. In the first step, facial landmarks are detected from input images and geometric features are extracted, considering the spatial position between landmarks. These feature vectors are then implemented in the proposed hierarchical WRF classifier to classify facial expressions. Our method was evaluated experimentally using three databases, extended Cohn-Kanade database (CK+), MMI and the Keimyung University Facial Expression of Drivers (KMU-FED) database, and its performance was compared with that of state-of-the-art methods. The results show that our proposed method yields a performance similar to that of deep learning FER methods as 92.6% for CK+ and 76.7% for MMI, with a significantly reduced processing cost approximately 3731 times less than that of the DNN method. These results confirm that the proposed method is optimized for real-time embedded applications having limited computing resources.

PPM1D Mutations Drive Clonal Hematopoiesis in Response to Cytotoxic Chemotherapy.

Clonal hematopoiesis (CH), in which stem cell clones dominate blood production, becomes increasingly common with age and can presage malignancy development. The conditions that promote ascendancy of particular clones are unclear. We found that mutations in PPM1D (protein phosphatase Mn2+/Mg2+-dependent 1D), a DNA damage response regulator that is frequently mutated in CH, were present in one-fifth of patients with therapy-related acute myeloid leukemia or myelodysplastic syndrome and strongly correlated with cisplatin exposure. Cell lines with hyperactive PPM1D mutations expand to outcompete normal cells after exposure to cytotoxic DNA damaging agents including cisplatin, and this effect was predominantly mediated by increased resistance to apoptosis. Moreover, heterozygous mutant Ppm1d hematopoietic cells outcompeted their wild-type counterparts in vivo after exposure to cisplatin and doxorubicin, but not during recovery from bone marrow transplantation. These findings establish the clinical relevance of PPM1D mutations in CH and the importance of studying mutation-treatment interactions. VIDEO ABSTRACT.

DNMT3A and TET1 cooperate to regulate promoter epigenetic landscapes in mouse embryonic stem cells.

BACKGROUND: DNA methylation is a heritable epigenetic mark, enabling stable but reversible gene repression. In mammalian cells, DNA methyltransferases (DNMTs) are responsible for modifying cytosine to 5-methylcytosine (5mC), which can be further oxidized by the TET dioxygenases to ultimately cause DNA demethylation. However, the genome-wide cooperation and functions of these two families of proteins, especially at large under-methylated regions, called canyons, remain largely unknown. RESULTS: Here we demonstrate that DNMT3A and TET1 function in a complementary and competitive manner in mouse embryonic stem cells to mediate proper epigenetic landscapes and gene expression. The longer isoform of DNMT3A, DNMT3A1, exhibits significant enrichment at distal promoters and canyon edges, but is excluded from proximal promoters and canyons where TET1 shows prominent binding. Deletion of Tet1 increases DNMT3A1 binding capacity at and around genes with wild-type TET1 binding. However, deletion of Dnmt3a has a minor effect on TET1 binding on chromatin, indicating that TET1 may limit DNA methylation partially by protecting its targets from DNMT3A and establishing boundaries for DNA methylation. Local CpG density may determine their complementary binding patterns and therefore that the methylation landscape is encoded in the DNA sequence. Furthermore, DNMT3A and TET1 impact histone modifications which in turn regulate gene expression. In particular, they regulate Polycomb Repressive Complex 2 (PRC2)-mediated H3K27me3 enrichment to constrain gene expression from bivalent promoters. CONCLUSIONS: We conclude that DNMT3A and TET1 regulate the epigenome and gene expression at specific targets via their functional interplay.

Loss of Dnmt3a Immortalizes Hematopoietic Stem Cells In Vivo.

Somatic mutations in DNMT3A are recurrent events across a range of blood cancers. Dnmt3a loss of function in hematopoietic stem cells (HSCs) skews divisions toward self-renewal at the expense of differentiation. Moreover, DNMT3A mutations can be detected in the blood of aging individuals, indicating that mutant cells outcompete normal HSCs over time. It is important to understand how these mutations provide a competitive advantage to HSCs. Here we show that Dnmt3a-null HSCs can regenerate over at least 12 transplant generations in mice, far exceeding the lifespan of normal HSCs. Molecular characterization reveals that this in vivo immortalization is associated with gradual and focal losses of DNA methylation at key regulatory regions associated with self-renewal genes, producing a highly stereotypical HSC phenotype in which epigenetic features are further buttressed. These findings lend insight into the preponderance of DNMT3A mutations in clonal hematopoiesis and the persistence of mutant clones after chemotherapy.

DNA epigenome editing using CRISPR-Cas SunTag-directed DNMT3A.

BACKGROUND: DNA methylation has widespread effects on gene expression during development. However, our ability to assign specific function to regions of DNA methylation is limited by the poor correlation between global patterns of DNA methylation and gene expression. RESULTS: Here, we utilize nuclease-deactivated Cas9 protein fused to repetitive peptide epitopes (SunTag) recruiting multiple copies of antibody-fused de novo DNA methyltransferase 3A (DNMT3A) (dCas9-SunTag-DNMT3A) to amplify the local DNMT3A concentration to methylate genomic sites of interest. We demonstrate that dCas9-SunTag-DNMT3A dramatically increases CpG methylation at the HOXA5 locus in human embryonic kidney (HEK293T) cells. Furthermore, using a single guide RNA, dCas9-SunTag-DNMT3A is able to methylate a 4.5-kb genomic region and repress HOXA5 gene expression. Reduced representation bisulfite sequencing and RNA-seq show that dCas9-SunTag-DNMT3A methylates regions of interest with minimal impact on the global DNA methylome and transcriptome. CONCLUSIONS: This effective and precise tool enables site-specific manipulation of DNA methylation and may be used to address the relationship between DNA methylation and gene expression.

Targeted DNA methylation in vivo using an engineered dCas9-MQ1 fusion protein.

Comprehensive studies have shown that DNA methylation plays vital roles in both loss of pluripotency and governance of the transcriptome during embryogenesis and subsequent developmental processes. Aberrant DNA methylation patterns have been widely observed in tumorigenesis, ageing and neurodegenerative diseases, highlighting the importance of a systematic understanding of DNA methylation and the dynamic changes of methylomes during disease onset and progression. Here we describe a facile and convenient approach for efficient targeted DNA methylation by fusing inactive Cas9 (dCas9) with an engineered prokaryotic DNA methyltransferase MQ1. Our study presents a rapid and efficient strategy to achieve locus-specific cytosine modifications in the genome without obvious impact on global methylation in 24 h. Finally, we demonstrate our tool can induce targeted CpG methylation in mice by zygote microinjection, thereby demonstrating its potential utility in early development.

Genome-Wide Analysis of DNA Methylation in Hematopoietic Cells: DNA Methylation Analysis by WGBS.

DNA methylation is a major epigenetic modification that regulates gene expression, genome imprinting, and development and has a role in diseases including cancer. There are various methods for whole-genome methylation profiling that differ in cost and resolution. Recent advances in high-throughput sequencing technologies coupled with bisulfite treatment enable absolute DNA methylation quantification and genome-wide single-nucleotide resolution analysis. In this chapter, we provide detailed protocols for whole-genome bisulfite sequencing (WGBS), which captures the complete methylome. Using WGBS, we are able to generate a reference DNA methylome for normal or malignant hematopoietic cells.

Inhibition of the B7-H3 immune checkpoint limits tumor growth by enhancing cytotoxic lymphocyte function.

The interaction between tumor and the immune system is still poorly understood. Significant clinical responses have been achieved in cancer patients treated with antibodies against the CTLA4 and PD-1/PD-L1 checkpoints; however, only a small portion of patients responded to the therapies, indicating a need to explore additional co-inhibitory molecules for cancer treatment. B7-H3, a member of the B7 superfamily, was previously shown by us to inhibit T-cell activation and autoimmunity. In this study, we have analyzed the function of B7-H3 in tumor immunity. Expression of B7-H3 was found in multiple tumor lines, tumor-infiltrating dendritic cells, and macrophages. B7-H3-deficient mice or mice treated with an antagonistic antibody to B7-H3 showed reduced growth of multiple tumors, which depended on NK and CD8+ T cells. With a putative receptor expressed by cytotoxic lymphocytes, B7-H3 inhibited their activation, and its deficiency resulted in increased cytotoxic lymphocyte function in tumor-bearing mice. Combining blockades of B7-H3 and PD-1 resulted in further enhanced therapeutic control of late-stage tumors. Taken together, our results indicate that the B7-H3 checkpoint may serve as a novel target for immunotherapy against cancer.

Chronic Infection Depletes Hematopoietic Stem Cells through Stress-Induced Terminal Differentiation.

Chronic infections affect a third of the world's population and can cause bone marrow suppression, a severe condition that increases mortality from infection. To uncover the basis for infection-associated bone marrow suppression, we conducted repeated infection of WT mice with Mycobacterium avium. After 4-6 months, mice became pancytopenic. Their hematopoietic stem and progenitor cells (HSPCs) were severely depleted and displayed interferon gamma (IFN-γ) signaling-dependent defects in self-renewal. There was no evidence of increased HSPC mobilization or apoptosis. However, consistent with known effects of IFN-γ, transcriptome analysis pointed toward increased myeloid differentiation of HSPCs and revealed the transcription factor Batf2 as a potential mediator of IFN-γ-induced HSPC differentiation. Gain- and loss-of-function studies uncovered a role for Batf2 in myeloid differentiation in both murine and human systems. We thus demonstrate that chronic infection can deplete HSPCs and identify BATF2 as a mediator of infection-induced HSPC terminal differentiation.

DNMT3A and TET2 compete and cooperate to repress lineage-specific transcription factors in hematopoietic stem cells.

Mutations in the epigenetic modifiers DNMT3A and TET2 non-randomly co-occur in lymphoma and leukemia despite their epistasis in the methylation-hydroxymethylation pathway. Using Dnmt3a and Tet2 double-knockout mice in which the development of malignancy is accelerated, we show that the double-knockout methylome reflects regions of independent, competitive and cooperative activity. Expression of lineage-specific transcription factors, including the erythroid regulators Klf1 and Epor, is upregulated in double-knockout hematopoietic stem cells (HSCs). DNMT3A and TET2 both repress Klf1, suggesting a model of cooperative inhibition by epigenetic modifiers. These data demonstrate a dual role for TET2 in promoting and inhibiting HSC differentiation, the loss of which, along with DNMT3A, obstructs differentiation, leading to transformation.

DNMT3A Loss Drives Enhancer Hypomethylation in FLT3-ITD-Associated Leukemias.

DNMT3A, the gene encoding the de novo DNA methyltransferase 3A, is among the most frequently mutated genes in hematologic malignancies. However, the mechanisms through which DNMT3A normally suppresses malignancy development are unknown. Here, we show that DNMT3A loss synergizes with the FLT3 internal tandem duplication in a dose-influenced fashion to generate rapid lethal lymphoid or myeloid leukemias similar to their human counterparts. Loss of DNMT3A leads to reduced DNA methylation, predominantly at hematopoietic enhancer regions in both mouse and human samples. Myeloid and lymphoid diseases arise from transformed murine hematopoietic stem cells. Broadly, our findings support a role for DNMT3A as a guardian of the epigenetic state at enhancer regions, critical for inhibition of leukemic transformation.