ClusterSignificance: a bioconductor package facilitating statistical analysis of class cluster separations in dimensionality reduced data.

Description

The ClusterSignificance package provides tools to assess if class clusters in dimensionality reduced data representations have a separation different from permuted data. The term class clusters here refers to, clusters of points representing known classes in the data. This is particularly useful to determine if a subset of the variables, e.g. genes in a specific pathway, alone can separate samples into these established classes. ClusterSignificance accomplishes this by, projecting all points onto a one dimensional line. Cluster separations are then scored and the probability of the seen separation being due to chance is evaluated using a permutation method.

Developers / Authors

Jason T. Serviss. Author, maintainer

Jesper R. Gådin. Author

Per Eriksson. Author

Lasse Folkersen. Author

Dan Grandér†. Author

Publication

Oxford Bioinformatics

Project website

Reproduction of all publication figures.

Software

ClusterSignificance Website



CIM-seq

Description

Advances in single-cell biology has enabled us to investigate isolated single cells at an unprecedented scale and resolution. However, cells in multicellular organisms are largely defined by their spatial organization within organ structures, which means that general methods for studying direct cell interaction on a large scale are needed. We propose a novel method, Cell Interaction by Multiplet Sequencing (CIM-Seq) that uses computational deconvolution of RNA-seq data from partially dissociated tissue to create cell interaction maps. We applied CIM-seq to human fetal pancreas, demonstrating that it recapitulates known cell interactions such as acinar-ductal cell contacts. Furthermore, we discover a strong link between a mesenchymal cell subtype and endocrine progenitor cells, and identify the set of genes that distinguishes this mesenchymal subtype. Thus, CIM-Seq is a general method for cell interaction studies that can be used on cell types defined to an arbitrary resolution allowing identification of interacting sub-cell types or cell states.

Developers / Authors

Jason T. Serviss. Author, maintainer

Martin Enge. Author

Nathanael Johansson Andrews. Author

Project website

CIM-seq



An antisense RNA capable of modulating the expression of the tumor suppressor microRNA-34a.

Description

The microRNA-34a is a well-studied tumor suppressor microRNA (miRNA) and a direct downstream target of TP53 with roles in several pathways associated with oncogenesis, such as proliferation, cellular growth, and differentiation. Due to its broad tumor suppressive activity, it is not surprising that miR34a expression is altered in a wide variety of solid tumors and hematological malignancies. However, the mechanisms by which miR34a is regulated in these cancers is largely unknown. In this study, we find that a long non-coding RNA transcribed antisense to the miR34a host gene, is critical for miR34a expression and mediation of its cellular functions in multiple types of human cancer. We name this long non-coding RNA lncTAM34a, and characterize its ability to facilitate miR34a expression under different types of cellular stress in both TP53 deficient and wild type settings.

Developers / Authors

Jason T. Serviss. Author

Nathanael Johansson Andrews. Author

Jimmy Van den Eynden. Author

Felix Clemens Richter. Author

Miranda Houtman. Author

Mattias Vesterlund. Author

Laura Schwarzmueller. Author

Per Johnsson. Author

Erik Larsson. Author

Dan Grandér† Author

Katja Pokrovskaja Tamm Author

Project website

lncTAM34a

Publication

Biorxiv preprint

Cell Death and Disease



A drug screening assay on cancer cells chronically adapted to acidosis.

Description

Drug screening for the identification of compounds with anticancer activity is commonly performed using cell lines cultured under normal oxygen pressure and physiological pH. However, solid tumors are characterized by a microenvironment with limited access to nutrients, reduced oxygen supply and acidosis. Tumor hypoxia and acidosis have been identified as important drivers of malignant progression and contribute to multicellular resistance to different forms of therapy. Tumor acidosis represents an important mechanism mediating drug resistance thus the identification of drugs active on acid-adapted cells may improve the efficacy of cancer therapy.

Here, we characterized human colon carcinoma cells (HCT116) chronically adapted to grow at pH 6.8 and used them to screen the Prestwick drug library for cytotoxic compounds. Analysis of gene expression profiles in parental and low pH-adapted cells showed several differences relating to cell cycle, metabolism and autophagy.

The screen led to the identification of several compounds which were further selected for their preferential cytotoxicity towards acid-adapted cells. Amongst 11 confirmed hits, we primarily focused our investigation on the benzoporphyrin derivative Verteporfin (VP). VP significantly reduced viability in low pH-adapted HCT116 cells as compared to parental HCT116 cells and normal immortalized epithelial cells. The cytotoxic activity of VP was enhanced by light activation and acidic pH culture conditions, likely via increased acid-dependent drug uptake. VP displayed the unique property to cause light-dependent cross-linking of proteins and resulted in accumulation of polyubiquitinated proteins without inducing inhibition of the proteasome.

Our study provides an example and a tool to identify anticancer drugs targeting acid-adapted cancer cells.

Developers / Authors

Jason T. Serviss. Author

Angelo DeMilito. Author

Paola Pellegrini. Author

Thomas Lundbäck. Author

Nicolo Bancaro. Author

Magdalena Mazurkiewicz. Author

Iryna Kolosenko. Author

Di Yu. Author

Martin Haraldsson. Author

Padraig D’Arcy. Author

Stig Linder. Author

Project website

Acid-adapted screening

Publication

Cancer Cell International



STAT3 is activated in multicellular spheroids of colon carcinoma cells and mediates expression of IRF9 and interferon stimulated genes.

Description

Three-dimensional cell cultures, such as multicellular spheroids (MCS), reflect the in vivo architecture of solid tumours and multicellular drug resistance. We previously identified interferon regulatory factor 9 (IRF9) to be responsible for the up-regulation of a subset of interferon (IFN)-stimulated genes (ISGs) in MCS of colon carcinoma cells. This set of ISGs closely resembled a previously identified IFN-related DNA-damage resistance signature (IRDS) that was correlated to resistance to chemo- and radiotherapy. In this study we found that transcription factor STAT3 is activated upstream of IRF9 and binds to the IRF9 promoter in MCS of HCT116 colorectal carcinoma cells. Transferring conditioned media (CM) from high cell density conditions to non-confluent cells resulted in STAT3 activation and increased expression of IRF9 and a panel of IRDS genes, also observed in MCS, suggesting the involvement of a soluble factor. Furthermore, we identified gp130/JAK signalling to be responsible for STAT3 activation, IRF9, and IRDS gene expression in MCS and by CM. Our data suggests a novel mechanism where STAT3 is activated in high cell density conditions resulting in increased expression of IRF9 and, in turn, IRDS genes, underlining a mechanism by which drug resistance is regulated.

Developers / Authors

Jason T. Serviss. Author

Elin Edsbäcker. Author

Iryna Kolosenko. Author

Caroline Palm-Apergi. Author

Angelo DeMilito. Author

Katja Pokrovskaja Tamm Author

Publication

Scientific Reports



lncRNA signatures in pediatric ALL.

Description

Acute lymphoblastic leukemia (ALL) is the most common pediatric cancer and remains the leading cause of cancer morbidity among pediatric patients. ALL can be broadly divided into B- and T-cell leukemias and is further sub-classified by well-established recurring genetic abnormalities. These hallmark features of ALL are utilized clinically for treatment stratification but on their own cannot account for the observed prognostic heterogenecity within ALL or recapitulate the disease in in vivo models.

Comprehensive microarray-based studies have previously illustrated that mRNA expression patterns correlate well with specific genetic abnormalities observed in ALL and were also used to identify new subtypes based on gene expression profiling. In addition, global transcriptomic and epigenetic analysis have uncovered specific signatures related to ALL prognosis. Although largely unexplored, initial studies have also indicated that lncRNAs may be important biomarkers in some ALL subtypes and play specific roles in important clinical features, such as glucocorticoid resistance.

By performing strand-specific RNA-seq analysis on ALL samples, we will identify functional RNA genes that are critical in the oncogenic process and further to identify specific targets of these genes. Since extensive clinical data is available, this will allow us to both analyze how well previously defined patient subgroups, such as those with MLL and BCR-ABL translocations, are identified when only taking into account lncRNAs, and also to define whether new prognostic subgroups can be uncovered. lncRNA candidates showing promise will subsequently be subjected to in-depth bioinformatics-based analysis by incorporating additional publically available high-throughput datasets. Gene regulation, conservation, transcription factor binding, and “guilt-by-association” studies will be preformed uncovering characteristics and potential functions for the lncRNA candidates. In addition, mass spectrometry and GRO-seq will be performed on a subset of these samples. We will utilize this data to correlate lncRNA and protein expression, as well as, examining differential expression of actively transcribed enhancer RNAs within the ALL subtypes.