Short-lived AUF1 p42-binding mRNAs of RANKL and BCL6 have two distinct instability elements each

Regulation of mRNA stability by RNA-protein interactions contributes significantly to quantitative aspects of gene expression. We have identified potential mRNA targets of the AU-rich element binding protein AUF1. Myc-tagged AUF1 p42 was induced in mouse NIH/3T3 cells and RNA-protein complexes isolated using anti-myc tag antibody beads. Bound mRNAs were analyzed with Affymetrix microarrays. We have identified 508 potential target mRNAs that were at least 3-fold enriched compared to control cells without myc-AUF1. 22.3% of the enriched mRNAs had an AU-rich cluster in the ARED Organism database, against 16.3% of non-enriched control mRNAs. The enrichment towards AU-rich elements was also visible by AREScore with an average value of 5.2 in the enriched mRNAs versus 4.2 in the control group. Yet, numerous mRNAs were enriched without a high ARE score. The enrichment of tetrameric and pentameric sequences suggests a broad AUF1 p42-binding spectrum at short U-rich sequences flanked by A or G. Still, some enriched mRNAs were highly unstable, as those of TNFSF11 (known as RANKL), KLF10, HES1, CCNT2, SMAD6, and BCL6. We have mapped some of the instability determinants. HES1 mRNA appeared to have a coding region determinant. Detailed analysis of the RANKL and BCL6 3'UTR revealed for both that full instability required two elements, which are conserved in evolution. In RANKL mRNA both elements are AU-rich and separated by 30 bases, while in BCL6 mRNA one is AU-rich and 60 bases from a non AU-rich element that potentially forms a stem-loop structure.

Characterization of cancer stem cell subpopulations from patient-derived xenografts of advanced prostate cancer

Markus Germann

Androgen-deprivation therapy is the standard treatment for prostate cancer. Despite the initial response, a fraction of cases manifest progression to castration-resistant prostate cancer. PCa recurrence is possibly due to androgen-independent cancer stem cells that reinitiate tumor growth. The properties of cancer stem cells versus highly proliferative progenitor cells remain to be further characterized at the cellular and gene expression level. To address the molecular basis of CSC switch from androgen dependency to independence, we have employed patient-derived xenograft models (LAPC-9 and BM-18) that have different androgen sensitivity properties. We have performed microarray and proteomic analysis of BM-18 tumor tissues prior to and following castration as well as androgen replacement. To characterize the cancer stem cells in these models we have isolated different subpopulations by FACS cytometry, based on combination of selected markers (CD44 and ALDH activity measured by ALDELFUOR assay) aiming to assess the transcriptomic profile of these different populations. To assess response of these CD44+/-/ALDH+/- cells to androgen targeting compounds we have established an androgen-sensitive xenograft model (PNPCa) and tested it by ex vivo tissue culture and organoid viability assays. Proteomic and microarray analysis of bulk BM-18 tumors indicates enrichment of stem cell markers upon castration: CD44, NKX3.1 and ALDH1 isoforms. Readministration of testosterone induces decrease of these markers. Castration induces a rapid tumor volume decrease in the BM-18 as opposed to LAPC-9, reflecting different androgen-dependent cell states. CD44+/-/ALDHhigh/low subpopulations were isolated by flow cytometry and for transcriptomic analysis. Castration in the BM18 model induces an increase in the subpopulations of CD44+/ALDHlow and CD44+/ALDHhigh, while ALDHsingle and ALDHhigh total populations are decreased. Both CD44 expression and ALDH activity is increased in the LAPC-9 model upon castration. Treatment of PNPCa tumor parts ex vivo with enzalutamide inhibited expression of androgen receptor and keratin 8 expression. CD44+ organoids derived from the same tissue did not show altered viability after enzalutamide treatment. Different androgen-independent cancer stem cell populations may be distinguished by ALDH activity status and CD44. Androgen-independent cells in the BM-18 androgen-dependent model are reflected by enrichment of CD44 expression. The least abundant subpopulation CD44+/ALDHhigh is present in the LAPC-9 and expanded upon castration, while in the BM-18 model it is exclusively present after castration, reflecting different androgen sensitivity. Ongoing analysis may elucidate the molecular mechanisms controlling cancer stem cell fates, androgen sensitivity and drug resistance.

AMER ASSOC CANCER RESEARCH2018

The use of RNA interference to increase PEI-mediated transient gene expression in mammalian cells

Martin Bertschinger

DNA uptake by polyethylenimine (PEI)-mediated transfection was investigated in Chinese hamster ovary (CHO DG44) cells. Rapid DNA uptake was observed with the maximum occurring during the first 45-60 min after addition of PEI-DNA complexes to cells. With longer incubation times, the aggregation of PEI-DNA particles reduced the rate of DNA uptake. At early times after transfection, the kinetics of DNA uptake were constant, suggesting that the limiting factor to PEI transfection was the rate of endocytosis. The most efficient transfection with PEI was observed at a density of 2 x 106 cells/ml and at a PEI:DNA ratio of 2:1 (w/w). DNA uptake at higher PEI:DNA ratios was also efficient, but the toxicity of PEI decreased the recombinant protein yield. The optimal PEI:DNA ratio was found to be related to the number of cells in the culture. Once in the cell, PEI:DNA particles must be disassembled to allow transcription of the recombinant gene. Here their disassembly was investigated using a simple in vitro assay. The particles were formed with either a linear (25 kDa or JetPEITM) or a branched PEI (2, 25, or 750 kDa) and then mixed with varying amounts of a competitor (RNA, DNA, or heparin). The amount of plasmid DNA released from particles was determined by agarose gel electrophoresis or spectroscopy. The stability of the particles to changes in pH, osmolarity, and the PEI:DNA ratio was also determined. Either the presence of heparin or high salt concentrations (1-2 M) yielded complete particle disassembly for all PEIs tested. The addition of RNA to particles formed with linear PEIs or branched 2 kDa PEI resulted in the rapid release of all the plasmid DNA, even at an RNA concentration of 50 µg/ml. However, the presence of RNA at concentrations up to 400 µg/ml induced only partial disassembly of particles formed with branched PEIs of 25 or 750 kDa. In the presence of competitor DNA, slow particle disassembly was observed with particles made with linear 25 kDa PEI, JetPEI, and branched 2 kDa PEI, but DNA release was not seen for particles made with branched PEIs of higher molecular weight. The presence of BSA only resulted in partial disassembly of PEI-DNA particles, and DNA release was not observed after the exposure of particles to acidic pH as low as 3. For all the PEIs tested, their stability increased as the PEI:DNA ratio increased. It was concluded from these studies that branched PEIs have a higher affinity for DNA than linear PEIs and that the intracellular disassembly of PEI-DNA particles may involve interactions between PEI and cellular RNA. In the second part of the work, RNA interference (RNAi) was used to enhance transient gene expression by knockdown of two different mRNAs, the E1A mRNA in human embryonic kidney 293 EBNA (HEK293E) cells and the ube2i mRNA in CHO DG44 cells and HEK293 cells. HEK293 cells are transformed with the adenovirus E1A and E1B genes. Because the E1A proteins function as transcriptional activators or repressors, they may have a positive or negative effect on transient transgene expression in this cell line. Suspension cultures of HEK293 cells were co-transfected with a reporter plasmid expressing the GFP gene and a plasmid expressing a short hairpin RNA (shRNA) targeting the E1A mRNAs for degradation by RNAi. The presence of the shRNA in HEK293E cells reduced the steady state level of E1A mRNA up to 75% and increased transient GFP expression from either the elongation factor-1α (EF-1α) promoter or the human cytomegalovirus (HCMV) immediate early promoter up to 3-fold. E1A mRNA depletion also resulted in a 2-fold increase in transient expression of a recombinant IgG in suspension cultures when the IgG light and heavy chain genes were driven by the EF-1α promoter. These results demonstrated that E1A has a negative effect on transient gene expression in HEK293E cells. Efficient RNAi-mediated depletion of ube2i mRNA in CHO DG44 cells was accomplished by co-expression of 2 shRNAs targeting two distinct sides in the ube2i mRNA, whereas knockdown with only one of the shRNAs at a time was insufficient for reducing ube2i mRNA levels. The reduction of the ube2i mRNA level in CHO DG44 cells was shown to enhance transient GFP expression 2-fold and IgG expression 2.5-fold. For these experiments, the GFP gene was under the control of the CMV IE promoter and the IgG light and heavy chain genes were under the control of the human EF-1α promoter. Thus, the enhancement of transient gene expression resulting from ube2i mRNA depletion was not promoter-dependent. These results demonstrated that sumoylation has a negative effect on transient gene expression in CHO DG44 and HEK293E cells.

EPFL2006

Molecular heterogeneity of the ganglionic eminence and its relationship to the birth of neocortical interneurons

Susan Willi Monnerat

Neocortical function and malfunction depends critically on a spectrum of inhibitory interneurons. To gain novel insight in normal brain function and mechanisms leading to diseases, such as epilepsy or schizophrenia, the knowledge about the specific functions of each interneuron type appears crucial. To date, the mystery about the diversity of neocortical interneurons remains still unresolved, even though much effort was devoted to the characterization of mature interneurons in the neocortex. Ten years ago, the discovery of the ganglionic eminence (GE) as origin of neocortical interneurons allowed a novel approach to assess their various identities, and thus prompted researchers to follow the fate of cells accommodated in this region. Despite the knowledge of the source location and its subdivision into three GEs, their molecular identity remained unknown apart from few genes, and consequently unabled the specific isolation of precursor cells. To this purpose, we performed extensive microarray analysis of the three GEs at various developmental timepoints to reveal their molecular heterogeneity and identify novel precursor cell surface markers that would enable their prospective isolation by flow cytometry and subsequent fate assessment in vitro and in vivo. Surprisingly, molecular heterogeneity of the three GEs was restricted to few differentially expressed genes. However, the differences in gene expression became more apparent at late developmental stage that may reflect the increased specification during development. Among the three GEs, the LGE and CGE resembled molecularly each other most, while the MGE and CGE were most distantly related to each other. Furthermore, our genome-wide expression study suggested that the CGE, a structure of so far controversial nature, was rather a unique structure than a fusion of the MGE and LGE. Analysis across developmental timepoints revealed that the MGE underwent tremendous changes in gene expression when compared to the differential expression existing between the homochronic MGE and cerebral cortex. In our microarray screen, we could identify Boc as a novel marker of LGE/CGE-resident precursor cells. Taking advantage of Boc that encodes a cell surface molecule, we could isolate the Boc+ cells from the GE by flow cytometry, and demonstrate that in contrast to Boc– cells, they self-renewed and differentiated into oligodendrocytes, astrocytes, and neurons, and thus identify them as progenitor cells. Furthermore, immunohistochemical studies suggested that Boc+ GE cells may correspond to radial glia, the embryonic neural stem cells. The specific expression of Boc in the LGE and CGE, but not in the MGE during their entire lifetime suggested that Boc is a reliable marker for LGE/CGE-resident precursor cells. To assess the fate of Boc+ GE cells, we performed homotypic transplantations of Boc+ and Boc– cells onto organotypic slice cultures, and observed that in contrast to Boc– cells, Boc+ cells did not migrate from the GE to the neocortex suggesting that Boc+ precursor cells of the GE do not generate neocortical interneurons. In future, additional fate-mapping studies of Boc+ GE cells, such as in vivo transplantation, should shed more light on the ability of Boc+ LGE/CGE precursor cells to generate neocortical interneurons, and dependent on the outcome serve to construct a cell-lineage tree of neocortical interneurons. Thus, Boc will remain a useful tool to investigate developmental processes underlying the generation of neocortical and/or other neurons, such as striatal neurons.

EPFL2007