Death Induced by Survival gene Elimination (DISE) correlates with neurotoxicity in Alzheimer's disease and aging.

Alzheimer's disease (AD) is characterized by progressive neurodegeneration, but the specific events that cause cell death remain poorly understood. Death Induced by Survival gene Elimination (DISE) is a cell death mechanism mediated by short (s) RNAs acting through the RNA-induced silencing complex (RISC). DISE is thus a form of RNA interference, in which G-rich 6mer seed sequences in the sRNAs (position 2-7) target hundreds of C-rich 6mer seed matches in genes essential for cell survival, resulting in the activation of cell death pathways. Here, using Argonaute precipitation and RNAseq (Ago-RP-Seq), we analyze RISC-bound sRNAs to quantify 6mer seed toxicity in several model systems. In mouse AD models and aging brain, in induced pluripotent stem cell-derived neurons from AD patients, and in cells exposed to Aß42 oligomers, RISC-bound sRNAs show a shift to more toxic 6mer seeds compared to controls. In contrast, in brains of "SuperAgers", humans over age 80 who have superior memory performance, RISC-bound sRNAs are shifted to more nontoxic 6mer seeds. Cells depleted of nontoxic sRNAs are sensitized to Aß42-induced cell death, and reintroducing nontoxic RNAs is protective. Altogether, the correlation between DISE and Aß42 toxicity suggests that increasing the levels of nontoxic miRNAs in the brain or blocking the activity of toxic RISC-bound sRNAs could ameliorate neurodegeneration.

CD95/Fas ligand mRNA is toxic to cells through more than one mechanism.

CD95/Fas ligand (CD95L) induces apoptosis through protein binding to the CD95 receptor. However, CD95L mRNA also induces toxicity in the absence of CD95 through induction of DISE (Death Induced by Survival Gene Elimination), a form of cell death mediated by RNA interference (RNAi). We now report that CD95L mRNA processing generates a short (s)RNA nearly identical to shL3, a commercial CD95L-targeting shRNA that led to the discovery of DISE. Neither of the miRNA biogenesis proteins Drosha nor Dicer are required for this processing. Interestingly, CD95L toxicity depends on the core component of the RISC, Ago2, in some cell lines, but not in others. In the HCT116 colon cancer cell line, Ago 1-4 appear to function redundantly in RNAi. In fact, Ago 1/2/3 knockout cells retain sensitivity to CD95L mRNA toxicity. Toxicity was only blocked by mutation of all in-frame start codons in the CD95L ORF. Dying cells exhibited an enrichment of RISC bound (R)-sRNAs with toxic 6mer seed sequences, while expression of the non-toxic CD95L mutant enriched for loading of R-sRNAs with nontoxic 6mer seeds. However, CD95L is not the only source of these R-sRNAs. We find that CD95L mRNA may induce DISE directly and indirectly, and that alternate mechanisms may underlie CD95L mRNA processing and toxicity.

CD95/Fas ligand induced toxicity.

The role of CD95/Fas ligand (CD95L/FasL) in the induction of CD95-mediated extrinsic apoptosis is well characterized. Trimerized, membrane-bound CD95L ligates the CD95 receptor activating downstream signaling resulting in the execution of cells by caspase proteins. However, the expression of CD95L has been reported to induce cell death in contexts in which this pathway is unlikely to be activated, such as in cell autonomous activation induced cell death (AICD) and in CD95-resistant cancer cell lines. Recent data suggests that the CD95L mRNA exerts toxicity through death induced by survival gene elimination (DISE). DISE results from the targeting of networks of survival genes by toxic short RNA (sRNA)s in the RNA-induced silencing complex (RISC). CD95L mRNA contributes to this death directly, through the processing of its mRNA into toxic sRNAs that are loaded into the RISC, and indirectly, by promoting the loading of other toxic sRNAs. Interestingly, CD95L is not the only mRNA that is processed and loaded into the RISC. Protein-coding mRNAs involved in protein translation are also selectively loaded. We propose a model in which networks of mRNA-derived sRNAs modulate DISE, with networks of genes providing non-toxic RISC substrate sRNAs that protect against DISE, and opposing networks of stress-activated genes that produce toxic RISC substrate sRNAs that promote DISE.

SPOROS: A pipeline to analyze DISE/6mer seed toxicity.

microRNAs (miRNAs) are (18-22nt long) noncoding short (s)RNAs that suppress gene expression by targeting the 3' untranslated region of target mRNAs. This occurs through the seed sequence located in position 2-7/8 of the miRNA guide strand, once it is loaded into the RNA induced silencing complex (RISC). G-rich 6mer seed sequences can kill cells by targeting C-rich 6mer seed matches located in genes that are critical for cell survival. This results in induction of Death Induced by Survival gene Elimination (DISE), through a mechanism we have called 6mer seed toxicity. miRNAs are often quantified in cells by aligning the reads from small (sm)RNA sequencing to the genome. However, the analysis of any smRNA Seq data set for predicted 6mer seed toxicity requires an alternative workflow, solely based on the exact position 2-7 of any short (s)RNA that can enter the RISC. Therefore, we developed SPOROS, a semi-automated pipeline that produces multiple useful outputs to predict and compare 6mer seed toxicity of cellular sRNAs, regardless of their nature, between different samples. We provide two examples to illustrate the capabilities of SPOROS: Example one involves the analysis of RISC-bound sRNAs in a cancer cell line (either wild-type or two mutant lines unable to produce most miRNAs). Example two is based on a publicly available smRNA Seq data set from postmortem brains (either from normal or Alzheimer's patients). Our methods (found at https://github.com/ebartom/SPOROS and at Code Ocean: https://doi.org/10.24433/CO.1732496.v1) are designed to be used to analyze a variety of smRNA Seq data in various normal and disease settings.

DISE/6mer seed toxicity-a powerful anti-cancer mechanism with implications for other diseases.

micro(mi)RNAs are short noncoding RNAs that through their seed sequence (pos. 2-7/8 of the guide strand) regulate cell function by targeting complementary sequences (seed matches) located mostly in the 3' untranslated region (3' UTR) of mRNAs. Any short RNA that enters the RNA induced silencing complex (RISC) can kill cells through miRNA-like RNA interference when its 6mer seed sequence (pos. 2-7 of the guide strand) has a G-rich nucleotide composition. G-rich seeds mediate 6mer Seed Toxicity by targeting C-rich seed matches in the 3' UTR of genes critical for cell survival. The resulting Death Induced by Survival gene Elimination (DISE) predominantly affects cancer cells but may contribute to cell death in other disease contexts. This review summarizes recent findings on the role of DISE/6mer Seed Tox in cancer; its therapeutic potential; its contribution to therapy resistance; its selectivity, and why normal cells are protected. In addition, we explore the connection between 6mer Seed Toxicity and aging in relation to cancer and certain neurodegenerative diseases.

CD95/Fas ligand mRNA is toxic to cells.

CD95/Fas ligand binds to the death receptor CD95 to induce apoptosis in sensitive cells. We previously reported that CD95L mRNA is enriched in sequences that, when converted to si/shRNAs, kill all cancer cells by targeting critical survival genes (Putzbach et al., 2017). We now report expression of full-length CD95L mRNA itself is highly toxic to cells and induces a similar form of cell death. We demonstrate that small (s)RNAs derived from CD95L are loaded into the RNA induced silencing complex (RISC) which is required for the toxicity and processing of CD95L mRNA into sRNAs is independent of both Dicer and Drosha. We provide evidence that in addition to the CD95L transgene a number of endogenous protein coding genes involved in regulating protein translation, particularly under low miRNA conditions, can be processed to sRNAs and loaded into the RISC suggesting a new level of cell fate regulation involving RNAi.

DISE: A Seed-Dependent RNAi Off-Target Effect That Kills Cancer Cells.

Off-target effects (OTEs) represent a significant caveat for RNAi caused by substantial complementarity between siRNAs and unintended mRNAs. We now discuss the existence of three types of seed-dependent OTEs (sOTEs). Type I involves unintended targeting through the guide strand seed of an siRNA. Type II is caused by the activity of the seed on the designated siRNA passenger strand when loaded into the RNA-induced silencing complex (RISC). Both type I and II sOTEs will elicit unpredictable cellular responses. By contrast, in sOTE type III the guide strand seed preferentially targets essential survival genes resulting in death induced by survival gene elimination (DISE). In this Opinion article, we discuss DISE as a consequence of RNAi that may preferentially affect cancer cells.

Induction of DISE in ovarian cancer cells in vivo.

The death receptor CD95/Fas can be activated by immune cells to kill cancer cells. shRNAs and siRNAs derived from CD95 or CD95 ligand (CD95L) are highly toxic to most cancer cells. We recently found that these sh/siRNAs kill cancer cells in the absence of the target by targeting the 3'UTRs of critical survival genes through canonical RNAi. We have named this unique form of off-target effect DISE (for death induced by survival gene elimination). DISE preferentially kills transformed cells and cancer stem cells. We demonstrate that DISE induction occurs in cancer cells in vivo after introducing a lentiviral CD95L derived shRNA (shL3) into HeyA8 ovarian cancer cells grown as i.p. xenografts in mice, when compared to a scrambled shRNA. To demonstrate the possibility of therapeutically inducing DISE, we coupled siRNAs to templated lipoprotein nanoparticles (TLP). In vitro, TLPs loaded with a CD95L derived siRNA (siL3) selectively silenced a biosensor comprised of Venus and CD95L ORF and killed ovarian cancer cells. In vivo, two siRNA-TLPs (siL2-TLP and siL3-TLP) reduced tumor growth similarly as observed for cells expressing the shL3 vector. These data suggest that it is possible to kill ovarian cancer cells in vivo via DISE induction using siRNA-TLPs.

Many si/shRNAs can kill cancer cells by targeting multiple survival genes through an off-target mechanism.

Over 80% of multiple-tested siRNAs and shRNAs targeting CD95 or CD95 ligand (CD95L) induce a form of cell death characterized by simultaneous activation of multiple cell death pathways preferentially killing transformed and cancer stem cells. We now show these si/shRNAs kill cancer cells through canonical RNAi by targeting the 3'UTR of critical survival genes in a unique form of off-target effect we call DISE (death induced by survival gene elimination). Drosha and Dicer-deficient cells, devoid of most miRNAs, are hypersensitive to DISE, suggesting cellular miRNAs protect cells from this form of cell death. By testing 4666 shRNAs derived from the CD95 and CD95L mRNA sequences and an unrelated control gene, Venus, we have identified many toxic sequences - most of them located in the open reading frame of CD95L. We propose that specific toxic RNAi-active sequences present in the genome can kill cancer cells.