SWI/SNF-dependent genes are defined by their chromatin landscape.

SWI/SNF complexes are evolutionarily conserved, ATP-dependent chromatin remodeling machines. Here, we characterize the features of SWI/SNF-dependent genes using BRM014, an inhibitor of the ATPase activity of the complexes. We find that SWI/SNF activity is required to maintain chromatin accessibility and nucleosome occupancy for most enhancers but not for most promoters. SWI/SNF activity is needed for expression of genes with low to medium levels of expression that have promoters with (1) low chromatin accessibility, (2) low levels of active histone marks, (3) high H3K4me1/H3K4me3 ratio, (4) low nucleosomal phasing, and (5) enrichment in TATA-box motifs. These promoters are mostly occupied by the canonical Brahma-related gene 1/Brahma-associated factor (BAF) complex. These genes are surrounded by SWI/SNF-dependent enhancers and mainly encode signal transduction, developmental, and cell identity genes (with almost no housekeeping genes). Machine-learning models trained with different chromatin characteristics of promoters and their surrounding regulatory regions indicate that the chromatin landscape is a determinant for establishing SWI/SNF dependency.

Serpine1 mRNA confers mesenchymal characteristics to the cell and promotes CD8+ T cells exclusion from colon adenocarcinomas.

Serine protease inhibitor clade E member 1 (SERPINE1) inhibits extracellular matrix proteolysis and cell detachment. However, SERPINE1 expression also promotes tumor progression and plays a crucial role in metastasis. Here, we solve this apparent paradox and report that Serpine1 mRNA per se, independent of its protein-coding function, confers mesenchymal properties to the cell, promoting migration, invasiveness, and resistance to anoikis and increasing glycolytic activity by sequestering miRNAs. Expression of Serpine1 mRNA upregulates the expression of the TRA2B splicing factor without affecting its mRNA levels. Through transcriptional profiling, we found that Serpine1 mRNA expression downregulates through TRA2B the expression of genes involved in the immune response. Analysis of human colon tumor samples showed an inverse correlation between SERPINE1 mRNA expression and CD8+ T cell infiltration, unveiling the potential value of SERPINE1 mRNA as a promising therapeutic target for colon tumors.

TGF-ß-Upregulated Lnc-Nr6a1 Acts as a Reservoir of miR-181 and Mediates Assembly of a Glycolytic Complex.

Long non-coding RNAs (lncRNAs) have emerged as key regulators in a wide range of biological processes. Here, we identified a mouse miRNA-host gene lncRNA (lnc-Nr6a1) upregulated early during epithelial-to-mesenchymal transition (EMT). We show that when lncRNA is processed, it gives rise to two abundant polyadenylated isoforms, lnc-Nr6a1-1 and lnc-Nr6a1-2, and a longer non-polyadenylated microprocessor-driven lnc-pri-miRNA containing clustered pre-miR-181a2 and pre-miR-181b2 hairpins. Ectopic expression of the lnc-Nr6a1-1 or lnc-Nr6a1-2 isoform enhanced cell migration and the invasive capacity of the cells, whereas the expression of the isoforms and miR-181a2 and miR-181b2 conferred anoikis resistance. Lnc-Nr6a1 gene deletion resulted in cells with lower adhesion capacity and reduced glycolytic metabolism, which are restored by lnc-Nr6a1-1 isoform expression. We performed identification of direct RNA interacting proteins (iDRIP) to identify proteins interacting directly with the lnc-Nr6a1-1 isoform. We defined a network of interacting proteins, including glycolytic enzymes, desmosome proteins and chaperone proteins; and we demonstrated that the lnc-Nr6a1-1 isoform directly binds and acts as a scaffold molecule for the assembly of ENO1, ALDOA, GAPDH, and PKM glycolytic enzymes, along with LDHA, supporting substrate channeling for efficient glycolysis. Our results unveil a role of Lnc-Nr6a1 as a multifunctional lncRNA acting as a backbone for multiprotein complex formation and primary microRNAs.

The high mobility group protein HMG20A cooperates with the histone reader PHF14 to modulate TGFß and Hippo pathways.

High mobility group (HMG) proteins are chromatin regulators with essential functions in development, cell differentiation and cell proliferation. The protein HMG20A is predicted by the AlphaFold2 software to contain three distinct structural elements, which we have functionally characterized: i) an amino-terminal, intrinsically disordered domain with transactivation activity; ii) an HMG box with higher binding affinity for double-stranded, four-way-junction DNA than for linear DNA; and iii) a long coiled-coil domain. Our proteomic study followed by a deletion analysis and structural modeling demonstrates that HMG20A forms a complex with the histone reader PHF14, via the establishment of a two-stranded alpha-helical coiled-coil structure. siRNA-mediated knockdown of either PHF14 or HMG20A in MDA-MB-231 cells causes similar defects in cell migration, invasion and homotypic cell-cell adhesion ability, but neither affects proliferation. Transcriptomic analyses demonstrate that PHF14 and HMG20A share a large subset of targets. We show that the PHF14-HMG20A complex modulates the Hippo pathway through a direct interaction with the TEAD1 transcription factor. PHF14 or HMG20A deficiency increases epithelial markers, including E-cadherin and the epithelial master regulator TP63 and impaired normal TGFß-trigged epithelial-to-mesenchymal transition. Taken together, these data indicate that PHF14 and HMG20A cooperate in regulating several pathways involved in epithelial-mesenchymal plasticity.

Co-transcriptional splicing efficiency is a gene-specific feature that can be regulated by TGFß.

Differential splicing efficiency of specific introns is a mechanism that dramatically increases protein diversity, based on selection of alternative exons for the final mature mRNA. However, it is unclear whether splicing efficiency of introns within the same gene is coordinated and eventually regulated as a mechanism to control mature mRNA levels. Based on nascent chromatin-associated RNA-sequencing data, we now find that co-transcriptional splicing (CTS) efficiency tends to be similar between the different introns of a gene. We establish that two well-differentiated strategies for CTS efficiency exist, at the extremes of a gradient: short genes that produce high levels of pre-mRNA undergo inefficient splicing, while long genes with relatively low levels of pre-mRNA have an efficient splicing. Notably, we observe that genes with efficient CTS display a higher level of mature mRNA relative to their pre-mRNA levels. Further, we show that the TGFß signal transduction pathway regulates the general CTS efficiency, causing changes in mature mRNA levels. Taken together, our data indicate that CTS efficiency is a gene-specific characteristic that can be regulated to control gene expression.

The metabesity factor HMG20A potentiates astrocyte survival and reactive astrogliosis preserving neuronal integrity.

Rationale: We recently demonstrated that the 'Metabesity' factor HMG20A regulates islet beta-cell functional maturity and adaptation to physiological stress such as pregnancy and pre-diabetes. HMG20A also dictates central nervous system (CNS) development via inhibition of the LSD1-CoREST complex but its expression pattern and function in adult brain remains unknown. Herein we sought to determine whether HMG20A is expressed in the adult CNS, specifically in hypothalamic astrocytes that are key in glucose homeostasis and whether similar to islets, HMG20A potentiates astrocyte function in response to environmental cues. Methods: HMG20A expression profile was assessed by quantitative PCR (QT-PCR), Western blotting and/or immunofluorescence in: 1) the hypothalamus of mice exposed or not to either a high-fat diet or a high-fat high-sucrose regimen, 2) human blood leukocytes and adipose tissue obtained from healthy or diabetic individuals and 3) primary mouse hypothalamic astrocytes exposed to either high glucose or palmitate. RNA-seq and cell metabolic parameters were performed on astrocytes treated or not with a siHMG20A. Astrocyte-mediated neuronal survival was evaluated using conditioned media from siHMG20A-treated astrocytes. The impact of ORY1001, an inhibitor of the LSD1-CoREST complex, on HMG20A expression, reactive astrogliosis and glucose metabolism was evaluated in vitro and in vivo in high-fat high-sucrose fed mice. Results: We show that Hmg20a is predominantly expressed in hypothalamic astrocytes, the main nutrient-sensing cell type of the brain. HMG20A expression was upregulated in diet-induced obesity and glucose intolerant mice, correlating with increased transcript levels of Gfap and Il1b indicative of inflammation and reactive astrogliosis. Hmg20a transcript levels were also increased in adipose tissue of obese non-diabetic individuals as compared to obese diabetic patients. HMG20A silencing in astrocytes resulted in repression of inflammatory, cholesterol biogenesis and epithelial-to-mesenchymal transition pathways which are hallmarks of reactive astrogliosis. Accordingly, HMG20A depleted astrocytes exhibited reduced mitochondrial bioenergetics and increased susceptibility to apoptosis. Neuron viability was also hindered in HMG20A-depleted astrocyte-derived conditioned media. ORY1001 treatment rescued expression of reactive astrogliosis-linked genes in HMG20A ablated astrocytes while enhancing cell surface area, GFAP intensity and STAT3 expression in healthy astrocytes, mimicking the effect of HMG20A. Furthermore, ORY1001 treatment protected against obesity-associated glucose intolerance in mice correlating with a regression of hypothalamic HMG20A expression, indicative of reactive astrogliosis attenuation with improved health status. Conclusion: HMG20A coordinates the astrocyte polarization state. Under physiological pressure such as obesity and insulin resistance that induces low grade inflammation, HMG20A expression is increased to induce reactive astrogliosis in an attempt to preserve the neuronal network and re-establish glucose homeostasis. Nonetheless, a chronic metabesity state or functional mutations will result in lower levels of HMG20A, failure to promote reactive astrogliosis and increase susceptibility of neurons to stress-induced apoptosis. Such effects could be reversed by ORY1001 treatment both in vitro and in vivo, paving the way for a new therapeutic approach for Type 2 Diabetes Mellitus.

TGFß promotes widespread enhancer chromatin opening and operates on genomic regulatory domains.

The Transforming Growth Factor-ß (TGFß) signaling pathway controls transcription by regulating enhancer activity. How TGFß-regulated enhancers are selected and what chromatin changes are associated with TGFß-dependent enhancers regulation are still unclear. Here we report that TGFß treatment triggers fast and widespread increase in chromatin accessibility in about 80% of the enhancers of normal mouse mammary epithelial-gland cells, irrespective of whether they are activated, repressed or not regulated by TGFß. This enhancer opening depends on both the canonical and non-canonical TGFß pathways. Most TGFß-regulated genes are located around enhancers regulated in the same way, often creating domains of several co-regulated genes that we term TGFß regulatory domains (TRD). CRISPR-mediated inactivation of enhancers within TRDs impairs TGFß-dependent regulation of all co-regulated genes, demonstrating that enhancer targeting is more promiscuous than previously anticipated. The area of TRD influence is restricted by topologically associating domains (TADs) borders, causing a bias towards co-regulation within TADs.

The Cornelia de Lange Syndrome-associated factor NIPBL interacts with BRD4 ET domain for transcription control of a common set of genes.

Mutations in NIPBL are the major cause of Cornelia de Lange Syndrome (CdLS). NIPBL is the cohesin-loading factor and has recently been associated with the BET (bromodomains and extra-terminal (ET) domain) proteins BRD2 and BRD4. Related to this, a CdLS-like phenotype has been described associated to BRD4 mutations. Here, we show direct interaction of NIPBL with different BET members in yeast, and selective interaction with BRD4 in cells, being the ET domain involved in the interaction. To understand the relationship between NIPBL and BET proteins, we have performed RNA-Seq expression analysis following depletion of the different proteins. Results indicate that genes regulated by NIPBL largely overlap with those regulated by BRD4 but not with those regulated by BRD2. ChIP-Seq analysis indicates preferential NIPBL occupancy at promoters, and knockdown experiments show mutual stabilization of NIPBL and BRD4 on co-regulated promoters. Moreover, human fibroblasts from CdLS probands with mutations in NIPBL show reduced BRD4 at co-occupied promoters. Functional analysis in vivo, using mutants of Drosophila melanogaster, confirmed the genetic interaction between Nipped-B and fs(1)h, the orthologs of human NIPBL and BRD4, respectively. Thus, we provide evidence for NIPBL and BRD4 cooperation in transcriptional regulation, which should contribute to explain the recently observed CdLS-like phenotype associated with BRD4 mutations.

TBL1 is required for the mesenchymal phenotype of transformed breast cancer cells.

The epithelial-to-mesenchymal transition (EMT) and its reversion (MET) are related to tumor cell dissemination and migration, tumor circulating cell generation, cancer stem cells, chemoresistance, and metastasis formation. To identify chromatin and epigenetic factors possibly involved in the process of EMT, we compare the levels of expression of epigenetic genes in a transformed human breast epithelial cell line (HMEC-RAS) versus a stable clone of the same cell line expressing the EMT master regulator ZEB1 (HMEC-RAS-ZEB1). One of the factors strongly induced in the HMEC-RAS-ZEB1 cells was Transducin beta-like 1 (TBL1), a component of the NCoR complex, which has both corepressor and coactivator activities. We show that TBL1 interacts with ZEB1 and that both factors cooperate to repress the promoter of the epithelial gene E-cadherin (CDH1) and to autoactivate the ZEB1 promoter. Consistent with its central role, TBL1 is required for mesenchymal phenotypes of transformed breast epithelial and breast cancer cell lines of the claudin-low subtype. Importantly, a high expression of the TBL1 gene correlates with poor prognosis and increased proportion of metastasis in breast cancer patients, indicating that the level of TBL1 expression can be used as a prognostic marker.

High expression of SMARCA4 or SMARCA2 is frequently associated with an opposite prognosis in cancer.

The gene encoding the ATPase of the chromatin remodeling SWI/SNF complexes SMARCA4 (BRG1) is often mutated or silenced in tumors, suggesting a role as tumor suppressor. Nonetheless, recent reports show requirement of SMARCA4 for tumor cells growth. Here, we performed a computational meta-analysis using gene expression, prognosis, and clinicopathological data to clarify the role of SMARCA4 and the alternative SWI/SNF ATPase SMARCA2 (BRM) in cancer. We show that while the SMARCA4 gene is mostly overexpressed in tumors, SMARCA2 is almost invariably downexpressed in tumors. High SMARCA4 expression was associated with poor prognosis in many types of tumors, including liver hepatocellular carcinoma (LIHC), and kidney renal clear cell carcinoma (KIRC). In contrast, high SMARCA2 expression was associated with good prognosis. We compared tumors with high versus low expression of SMARCA4 or SMARCA2 in LIHC and KIRC cohorts from The Cancer Genome Atlas. While a high expression of SMARCA4 is associated with aggressive tumors, a high expression of SMARCA2 is associated with benign differentiated tumors, suggesting that SMARCA4 and SMARCA2 play opposite roles in cancer. Our results demonstrate that expression of SMARCA4 and SMARCA2 have a high prognostic value and challenge the broadly accepted general role of SMARCA4 as a tumor suppressor.

Analysis of the relationship between coexpression domains and chromatin 3D organization.

Gene order is not random in eukaryotic chromosomes, and co-regulated genes tend to be clustered. The mechanisms that determine co-regulation of large regions of the genome and its connection with chromatin three-dimensional (3D) organization are still unclear however. Here we have adapted a recently described method for identifying chromatin topologically associating domains (TADs) to identify coexpression domains (which we term "CODs"). Using human normal breast and breast cancer RNA-seq data, we have identified approximately 500 CODs. CODs in the normal and breast cancer genomes share similar characteristics but differ in their gene composition. COD genes have a greater tendency to be coexpressed with genes that reside in other CODs than with non-COD genes. Such inter-COD coexpression is maintained over large chromosomal distances in the normal genome but is partially lost in the cancer genome. Analyzing the relationship between CODs and chromatin 3D organization using Hi-C contact data, we find that CODs do not correspond to TADs. In fact, intra-TAD gene coexpression is the same as random for most chromosomes. However, the contact profile is similar between gene pairs that reside either in the same COD or in coexpressed CODs. These data indicate that co-regulated genes in the genome present similar patterns of contacts irrespective of the frequency of physical chromatin contacts between them.

Identification of morphological markers of sarcopenia at early stage of aging in skeletal muscle of mice.

The gastrocnemius muscle (GM) of young (3months) and aged (12months) female wild-type C57/BL6 mice was examined by light and electron microscopy, looking for the presence of structural changes at early stage of the aging process. Morphometrical parameters including body and gastrocnemius weights, number and type of muscle fibers, cross section area (CSA), perimeter, and Feret's diameter of single muscle fiber, were measured. Moreover, lengths of the sarcomere, A-band, I-band, H-zone, and number and CSA of intermyofibrillar mitochondria (IFM), were also determined. The results provide evidence that 12month-old mice had significant changes on skeletal muscle structure, beginning with the reduction of gastrocnemius weight to body weight ratio, compatible with an early loss of skeletal muscle function and strength. Moreover, light microscopy revealed increased muscle fibers size, with a significant increase on their CSA, perimeter, and diameter of both type I and type II muscle fibers, and a reduction in the percentage of muscle area occupied by type II fibers. Enhanced connective tissue infiltrations, and the presence of centrally nucleated muscle fibers, were also found in aged mice. These changes may underlie an attempt to compensate the loss of muscle mass and muscle fibers number. Furthermore, electron microscopy discovered a significant age-dependent increase in the length of sarcomeres, I and H bands, and reduction on the overlapped actin/myosin length, supporting contractile force loss with age. Electron microscopy also showed an increased number and CSA of IFM with age, which may reveal more endurance at 12months of age. Together, mice at early stage of aging already show significant changes in gastrocnemius muscle morphology and ultrastructure that are suggestive of the onset of sarcopenia.