The long non-coding RNA ‘TRASH’ is essential for cell survival in MAPK-driven melanoma


 MAPK-pathway up-regulation is responsible for over 40% of human cancer cases. Finding effective therapeutic targets for melanoma therapy continues to be a challenge due to drug resistance. Using a computational and experimental pipeline, we discovered the nuclear enriched long non-coding RNA (lncRNA) TRASH, which is induced upon MAPK-pathway-activation in melanoma. LncRNA hold essential regulatory functions in many cancer types. TRASH-targeting Antisense Oligonucleotides (TRASH-ASOs) greatly reduced cell-growth of melanoma cells in culture and systemic TRASH-ASO treatment significantly inhibited tumor growth of melanoma cell-line and patient-derived tumor xenografts without apparent side effects. We found that TRASH is essential for protein stability of the MAPK-pathway-regulating and apoptosis-inhibiting oncogene hnRNPA2/B1. TRASH knockdown induced apoptosis by down-regulating anti-apoptotic kinase activity and pro-survival signaling pathways. Compared to the well-studied oncogenic lncRNA MALAT1, the unique feature of TRASH is that it governs tumor cell-survival through maintaining activity of anti-apoptotic kinases and pro-survival signaling pathways in melanoma cells. TRASH-ASO treatment can bypass MEK-inhibitor (MEKi) resistance, and unlike MEKi, it does not induce early-onset treatment resistance. Furthermore, TRASH-ASO and MEKi combinations confer synergistic effects in melanoma treatment. Our findings show that TRASH-ASO treatment presents promising features for long lasting RNA-targeting melanoma therapy.

more lncRNA transcripts such as MALAT1, PANDAR or PCA3 are getting identified and characterized as important tumor suppressing or oncogenic regulators. 17 In recent years, an increasing number of RNA-targeting Antisense Oligonucleotides (ASOs) have been brought to clinical trials and obtained FDA approval. 18,19 In particular, emerging findings in lncRNAtargeted gene silencing show promising results. 20 We explored a link of several lncRNAs to the MAPK-pathway and their relevance for melanoma cellsurvival and tumor progression. As a result, we identified the lncRNA TRASH (TRanscript ASsociated with HnRNPA2/B1) and its central role in melanoma biology. We found that TRASH is required for melanoma cell-survival and sustains melanoma cell-growth. TRASH knockdown mediated by Antisense Oligonucleotides (TRASH-ASOs) resulted in efficient suppression of anti-apoptotic kinase activity and promoted cell death in a broad panel of melanoma cell-lines, including melanoma resistant to MEKinhibition (MEKi). TRASH was found to bind to hnRNPA2/B1 and TRASH-ASO treatment led to concomitant down-regulation of hnRNPA2/B1 protein levels. In vitro evaluation of additional probable clinical features of TRASH-inhibition unveiled that TRASH-ASO treatment showed synergistic treatment effects in dual application with MEKi. In contrast to MEKi treatment, TRASH-ASO treatment did not lead to early-onset treatment resistance.
In mouse models of melanoma cell-line xenografts and patient-derived tumors, TRASH-ASO treatment induced apoptosis and strongly reduced tumor growth. In summary, these findings demonstrate the strong potential of TRASH-ASOs for the treatment of melanoma. 6

Identification of MAPK-pathway-activation responsive lncRNAs in melanoma
The oncogene NRAS is an upstream regulator of the MAPK-pathway. 21 NRAS-mutations are an early event in melanocytic tumorigenesis and NRAS-activation leads to activation of the downstream targets AKT and ERK. [22][23][24] To identify lncRNA transcripts that respond to MAPK-pathway-activation, we introduced an NRAS Q61 -mutant plasmid into primary human melanocytic cell-lines (PHM Q61 ) via lentiviral transduction. Alongside this, an empty vector was transduced into a control group of primary human melanocytic cell-lines (PHM E ). Successful transduction was confirmed through up-regulated levels of phosphorylated ERK and AKT (p-ERK and p-AKT) in PHM Q61 (Suppl. Fig. 1a). Activating NRAS-mutations like NRAS Q61 occur in benign nevi and additional transformations are needed to fully-unfold the malignant potential of melanocytes. 25 No significant differences in cell proliferation could be measured comparing PHM Q61 and PHM e , confirming that a sole NRAS Q61 -mutation is insufficient to equip melanocytic cell-lines with significant cell-growth characteristics (Suppl. Fig. 1a).
We present a schematic workflow overview of the combined in silico and in vitro processes to identify lncRNAs regulated in response to MAPK-pathway-activation which are essential for melanoma cellsurvival. (Fig. 1a) First, we compared paired-end non-poly A enriched 101-bp RNA-Seq data from PHM (no vector), PHM E , PHM Q61 , and two melanoma cell-lines (D04, MM415) harboring MAPK-pathway hyper-activating mutations. 197 genes were differently expressed (DE) and showed the same tendency in PHM Q61 , D04, and MM415 when compared to standard melanocytes (PHM Q61 ΔPHM E ; D04ΔPHM; M415ΔPHM) ( Fig. 1b-c, Suppl. Fig. 1b). 81 of the DE genes were up-regulated lncRNAs in all three comparisons. 24 of those lncRNAs were potentially clinically relevant as they were also expressed (FPKM-values > 0.2) in >90% of patient-derived NRAS mutant melanoma tumors from the Cancer Genome Atlas Skin Cutaneous Melanoma (TCGA-SKCM) dataset. (Fig. 1d, Suppl. Fig. 1c). We used small interfering RNA (siRNA) and esiRNA libraries (endoribonuclease prepared siRNA) to screen for lncRNAs essential for melanoma cell-survival. For 6 of the targets, RNAi led to significant cell-growth reduction in melanoma cell-lines, while no such impact could be observed in primary human melanocytic cell-lines. Fig. 1e shows RNAi screening effects for silencing of the lncRNA AC004540.4, the transcript we identified as a top candidate for further investigations. These findings unveiled MAPK-activation responsive lncRNAs that are essential for melanoma cell-survival.
The lncRNA AC004540.4 (TRASH) is a nuclear regulator of hnRNPA2/B1 AC004540.4 is a lncRNA located on chromosome 7, expressed in two isoforms, and not conserved in other species (Suppl. Fig. 2a). Typically, the regulatory functions of lncRNAs are closely related to their subcellular localization. 26 With lncRNAs primarily localized in the nucleus, the role of AC004540.4 in melanoma homeostasis was assessed by analysis of subcellular fractionations. We demonstrate that AC004540.4 is highly enriched in the nuclear compartment versus the cytoplasm in the melanoma cellline D04 (Fig. 2a). Nuclear enriched lncRNAs often exist in inefficiently spliced states. 27 Using four different primer pairs for comparison of relative quantification of different regions of AC004540.4 through qRT-PCR revealed that exonic, intronic, and exon/intron transition regions of AC004540.4 were detectable in different quantities, indicating that AC004540.4 transcripts may exist to a certain extent in inefficient spliced states (Fig. 2b). Nuclear enriched lncRNA functions typically involve cis-regulation and co-expression of protein-coding genes that are located in close proximity. 27 The closest coding gene to AC004540.4 on the same strand is the oncogene hnRNPA2/B1. HnRNPA2/B1 is part of the family of heterogeneous nuclear ribonucleoproteins (hnRNPs), a group of proteins that have at least one RNAbinding motif and regulate nucleic acid metabolism. 28 HnRNPA2/B1 interacts with lncRNAs and exerts regulatory functions in MAPK-pathway signaling. 6-10 Afterwards, we used the TCGA and The Genotype-Tissue Expression (GTEx) project databases to study whether AC004540.4 and hnRNPA2/B1 have any functional relationship in patient derived melanoma and/or non-malignant skin biopsies. We explored 8 the correlation of expression between both genes compared it to permutations of randomly chosen genes. Most notably, RNA-expression of AC004540.4 and hnRNPA2/B1 was significantly higher in melanoma (Fig 2c). The correlation of both genes was almost always significantly stronger in melanoma than the average correlation of each gene to 10 sets of random genes (p<0.05 10/10 for AC004540.4 and 8/10 for hnRNPA2/B1). In contrast, in non-malignant skin samples, there were no significant differences in AC004540.4-, and hnRNPA2/B1-expression in any of the 20 comparisons (Suppl. Fig. 2b-e).
Based on the gene expression correlation and the subsequently presented functional studies, we named AC004540.4 TRanscript ASsociated with HnRNPA2/B1 (TRASH).
Synthetic nucleic acids such as siRNA and antisense oligonucleotides (ASOs) are frequently used laboratory methods for silencing lncRNA-expression. Several have proven to bear strong clinical value and received FDA and/or EMA approval for therapeutic approaches. 20 Compared to siRNA, ASOmediated gene silencing is more effective for silencing nuclear localized lncRNAs and allows more chemical modification of synthetic nucleic acids to reduce undesirable side effects. 29,30 ASO improvement with locked nucleic acids (LNAs) leads to more specificity and less toxicity. 17 With the goal of clinical applicability in mind, we focused on LNA-improved ASOs for all further TRASH-knockdown experiments (TRASH-ASOs).
TRASH-ASO treatment did not significantly affect hnRNPA2/B1 transcript levels in D04 melanoma cells, indicating that hnRNPA2/B1 transcription does not depend on TRASH transcript levels (Fig. 2d). To investigate if TRASH expression regulates hnRNPA2/B1 protein expression, we examined protein levels of hnRNPA2/B1 1 and 2 days after TRASH-ASO treatment. By Immunoblot, hnRNPA2/B1 protein was found to be essentially undetected upon TRASH-inhibition (Fig. 2e). Although loss of hnRNPA2/B1 protein by TRASH-inhibition could be an indirect effect, our data suggests that TRASH either has an essential role for hnRNPA2/B1 protein translation or affects hnRNPA2/B1 protein stability by direct association in the nucleus. However, the former is unlikely as TRASH is highly enriched in the nucleus and hnRNPA2/B1 has a half-life of several days in primary human cells (Fig. 2a). 31 To investigate if hnRNPA2/B1 protein associates with TRASH in the nucleus, we performed RIP-assay by pulling down hnRNPA2/B1 from melanoma cell lysate with the use of antibodies against hnRNPA2/B1 and examined RNA associated with the immunoprecipitated material. TRASH was successfully immunoprecipitated with hnRNPA2/B1 (shown by immunoblot, Fig. 2f) and it was enriched >65-fold compared to negative control pulldown with normal IgG. Therefore, TRASH and the protein hnRNPA2/B1 physically interact (Fig. 2f).
These findings indicate that up-regulation of TRASH and hnRNPA2/B1 transcription is a melanoma specific event and TRASH and hnRNPA2/B1 protein physically interact with each other. Considering prior studies demonstrating that lncRNAs can stabilize proteins and inhibit their degradation, TRASHexpression may be essential for maintaining stable hnRNPA2/B1 protein levels in melanoma. 15 TRASH is an anti-apoptotic regulator in MAPK-driven melanoma.
NRAS-, BRAF-, and c-KIT-mutations are MAPK-pathway activating mutations that frequently occur in melanoma patients. [32][33][34] We investigated the effects of TRASH-inhibition on MAPK-driven melanoma by applying TRASH-ASO treatment to a repository of established and primary patient-derived melanoma cell-lines that harbor these mutations (see Methods). In comparison to non-targeting control ASO (Control-ASO) treatment, TRASH-ASOs significantly reduced cell-growth in all tested cell-lines (Fig. 3a).
Using clonogenic assays, we further investigated the colony growth potential in three melanoma celllines after TRASH-ASO treatment. TRASH-ASOs drastically reduced the capability of melanoma cells to produce colonies when compared to Control-ASO treatment (Fig. 3b). To test if these inhibitory effects of TRASH-ASO treatment may be partially related to the TRASH-hnRNPA2/B1 axis, we performed hnRNPA2/B1-ASO treatment in the D04 cell-line. Like TRASH-ASO treatment, hnRNPA2/B1-ASO treatment also significantly reduced cell-growth, but to a lesser extent. (Fig. 3b-c). Afterwards, we investigated if the observed ASO treatment effects are due to apoptotic cell death. As a readout for the induction of apoptosis, caspase-3 & -7 activity was significantly increased 3-fold after TRASH-, and 1.7fold after hnRNPA2/B1-ASO treatment (Fig. 3d).
To examine the biomolecular changes upon TRASH-inhibition in melanoma, D04 cells were treated with TRASH-ASOs and Control-ASOs and RNA was extracted and used for RNA-Seq. Differential expression (DE) analysis showed that TRASH-ASOs had a global effect on gene expression. We found that 574 genes were down-regulated, and 493 genes were up-regulated, when compared to Control-ASO treatment (Cut off was >1.5-fold change and FDR <0.05, Suppl. Table 1). GO term analysis revealed that the top enriched GO term clusters associated with the down-regulated genes were related to "ECM-receptor interaction" and "PI3K-AKT signaling pathway", while the top enriched GO term clusters associated with the up-regulated genes included the terms "protein tyrosine kinase activity" (GO: 0004713) and "Ras guanyl-nucleotide exchange factor activity" (GO0005088) (Suppl. Table 2). These GO terms consisted of genes encoding growth factors, tyrosine kinases, G protein coupled receptor subunits, and collagen subunits.
These findings suggest that TRASH governs melanoma cell survival and inhibits apoptosis to a stronger extent than its protein binding partner hnRNPA2/B1 and that TRASH may execute its anti-apoptotic functions as a regulator of the MAPK and PI3K-AKT signaling cascade.
Kinase activity profiling reveals unique anti-apoptotic features of TRASH-expression Kinases cover a wide range of apoptosis regulating functions in cancer. Given the findings that TRASH-ASO treatment strongly affects the transcriptional regulation of genes that are related to kinase signaling pathways, we aimed to perform functional profiling of kinase activity shifts triggered by TRASHinhibition. To do so, we used a kinase activity screening platform 35 (named High Throughput Kinase Activity Mapping -HT-KAM) that enables the simultaneous identification of kinase enzymes functional state in cancer cells across a broad range of kinase families (see Methods for details). We generated protein extracts of two versions of the D04 (D04treatment naïve; D04RMtrametinib resistant) and the MM415 melanoma cell-lines, treated with Control-ASOs or TRASH-ASOs. We tested these cell extracts on HT-KAM and performed unsupervised hierarchical clustering of peptide-associated phosphorylation profiles (Fig. 4a) and of kinase activity signatures (Fig 4b). The changes in kinases' activity upon TRASH-ASO treatment indicate conserved responses across cell-lines, whether kinases are up-regulated or down-regulated (respectively in yellow or blue in Fig. 4b).
Due to the effects of TRASH-ASO treatment on cell viability and apoptosis induction (Fig. 3a+d), we focused on kinases with anti-apoptotic functions. We found that the pro-survival/proto-oncogenic kinases AKT1, CDK1, LYN, YES1, CHEK1, PKCA, STK11, PKCa and PIM1 were significantly less active upon TRASH-inhibition ( Fig. 4c left panel). These kinases have been reported to regulate the state of caspases and pro-survival pathways including the RAF-MAPK and PI3K-AKT axes. [37][38][39][40][41][42][43][44] To further test if these observations are TRASH-ASO treatment specific, we generated MALAT1-ASO treated extracts from the same cell-line models. MALAT1 is a known oncogenic lncRNA in various types of cancer, including melanoma. 45,46 MALAT1-ASO treatment reduced cell-growth but displayed a significantly reduced effect on apoptosis induction in comparison to TRASH-inhibition (p=0.002 for 1.5fold versus 3.0-fold Caspase-3 & -7 activity increase respectively in Fig. 4d and Fig. 3d). Using the HT-KAM platform, we found that the activity of the kinases associated with cell-survival were not downregulated in MALAT1-ASO treated cells (Fig. 4c right panel), but significantly and specifically downregulated upon TRASH-ASO treatment (Fig. 4c, p < 0.00007; Fig. 4e, kinase signatures of TRASH-, versus MALAT1-ASO treatment). In summary, our data indicate that TRASH-ASO treatment specifically downregulates the activity of anti-apoptotic kinases and pro-survival signaling pathways in melanoma cells, supporting the potential therapeutic relevance of TRASH-ASO treatment (Fig. 4f).

TRASH-ASO treatment shows characteristics of potential clinical value
The MEK-inhibitor (MEKi) trametinib is a FDA approved drug for the treatment of melanoma as monoand combinatorial therapy and used in clinics worldwide. 47 Drug resistance is the main limiting factor in modern oncology. 48 Therefore, therapeutic applications that reduce growth of drug resistant tumors are urgently needed. TRASH-ASO treatment in a panel of cell-lines that are resistant to MEKi led to significant cell-growth reduction, comparable to the effect seen in their non-resistant treatment-naïve cell-line counterparts (Fig. 3a and 5a). Combinational application of drugs is a common strategy in clinical oncology to synergize drug effects and to hamper the development of drug resistance. 49,50 We show synergistic effects between TRASH-ASOs and MEKi in a broad panel of different concentrations and combinations in the melanoma cell-line D04 and in the primary patient-derived cell-line AV5.
Synergy strongly increased with higher concentrations of TRASH-ASOs. More importantly, no notable effects of drug antagonism in regards of one drug blocking the effectiveness of the other drug, could be observed (Fig. 5b).
Next, we rescued cells that survived initial TRASH-ASO and MEKi treatment and after a phase of regrowth in drug free media, we repeated the preceding drug treatment. D04 cells responded with increased vulnerability to TRASH-ASO treatment, implying that no drug resistance could be measured.
On the other hand, D04 cells that underwent MEKi responded with significantly less cell-growth inhibition to further MEKi, implying that these cells developed resistance mechanisms that decreased vulnerability to MEKi (Fig. 5c). To further evaluate the clinical potential of TRASH-inhibition in melanoma, we aimed to test the effects of TRASH-ASO treatment versus Control-ASO treatment in mouse-models of melanoma. We treated xenograft mouse-models harboring melanoma cell-line tumors (D04), primary melanoma cell tumors (AV5) and a patient-derived tumor xenograft (PDX). Systemic treatment with 60µg subcutaneous ASO injections twice a week for 21 days, co-applied with an in vivo transfection reagent, significantly reduced tumor growth in the TRASH-ASO treatment groups in all three 13 mouse models (Fig. 5d). The PDX tumor model TM01341 showed extremely high rates of tumor growth.
Tumor growth could be significantly hampered in the TRASH-ASO treatment group, but in accordance with institutional guidelines for maximum acceptable tumor sizes, Control-ASO group mice had to be euthanized before the desired endpoint of the experiment. To simulate the experiment to the desired endpoint, tumor growth in the PDX control group was forecasted using a regression model (Fig. 5d).
There were no significant differences in mice weight change between the TRASH-ASO and Control-ASO treatment group in any of the three xenograft tumor models (Fig. 5d). qRT-PCR of tumor tissue extracted after the conclusion of the treatment period showed that in vivo TRASH-ASO treatment strongly reduced TRASH-expression in D04 tumors (Fig. 5e). Under certain conditions ASO treatment can lead to toxic side effects, in particular hepatotoxicity. 51 To investigate potential toxicity, liver tissue of treated mice was collected for H+E staining after the end of the treatment period. Neither animals receiving TRASH-ASOs nor those receiving Control-ASOs had any detectable pathologic changes in liver tissue (Fig. 5f). Additionally, immunohistochemistry staining for the apoptosis marker cleaved-caspase-3 confirmed induction of apoptosis through systemic TRASH-ASO in vivo treatment compared to Control-ASO treatment in D04 tumors (Fig. 5f).
In summary, these findings show that TRASH-ASOs could help to bypass the resistance of melanoma cells to MEKi therapy. Additionally, TRASH-ASO treatment has the potential to amplify the efficiency of MEKi. Mechanisms of resistance to RNAi have been reported in mammalian cells. 52 However, to our knowledge, no data regarding resistance mechanisms against ASO mediated lncRNA depleting therapy in mammalian cells exists. Our findings highlight that in contrast to MEKi, no early-onset treatment resistance could be observed for TRASH-ASO treatment in melanoma. Most notably, systemic TRASH-

ASO treatment induces apoptosis and significantly reduces TRASH-expression and tumor growth in vivo
while showing no signs of toxicity.

Discussion
There has been major progress in the development of melanoma therapeutics in the past decade, but many patients have limited benefit from these advances due to the acquisition of resistance to treatment. Therefore, additional treatment options are urgently needed. Over 40% of human cancers are caused and characterized by MAPK-pathway hyper-activation, which propels expression of essential oncogenic regulators that can serve as therapeutic targets in cancer therapy. 2,11,53 Using a novel pipeline that is composed of a broad set of analytical in silico and in vitro steps, we identified a set of lncRNAs that are up-regulated through MAPK activation, including the nuclear enriched lncRNA TRASH.
Previous reports indicate that TRASH could play a role in non-alcoholic fatty liver disease, but to our knowledge, its oncogenic features have never been unveiled before. 54 Our findings indicate that TRASH- while having no such effects on non-malignant melanocytic cell-lines. Some of the oncogenic features of TRASH may rely on the stabilizing effect on its protein-binding partner, the product of the anti-apoptotic oncogene hnRNPA2/B1. Analyses of patient-derived melanoma and non-malignant skin tissues from TCGA and GTEx databases demonstrate that TRASH and hnRNPA2/B1 up-regulation is tightly correlated in melanoma. Given the melanoma cell-growth reducing effects of hnRNPA2/B1-ASOs in vitro, we propose that the functional axis of TRASH and hnRNPA2/B1 contributes to the oncogenic nature of TRASH and is a critical mechanism for melanoma cell homeostasis.
15 TRASH-ASO treatment followed by transcriptome analysis and GO term analysis provided additional indication that TRASH knockdown may down-regulate the pro-survival MAPK and PI3K-AKT signaling pathways.
ASOs that down-regulate transcription of apoptosis inhibiting coding genes have already been proven to lead to long remission rates in cancer in clinical trials. 18 Our findings indicate that TRASH is a novel and central lncRNA that can be targeted for apoptosis-inhibition. TRASH-ASO treatment induced apoptosis through specific down-regulation of the activity of the anti-apoptotic/pro-survival kinases AKT1, CDK1, LYN, YES1, CHEK1, PKCA, STK11, PKCa and PIM1. Kinases of the AKT family and CHEK1 inhibit caspase activity. Lyn is a caspase-activation dependent regulator of AKT. [37][38][39]41 PKCa stimulates AKT-dependent RAF-1-activation. 40 Yes1-inhibition has been reported to increase apoptotic effects of chemotherapy and it is positively regulated by the kinase CDK1. 43 Tumor cells lacking the upstream master kinase STK11 are hypersensitive to apoptosis. 42 PIM1 regulates cell-survival through down-regulation of the MAPKpathway kinase ASK1. 44 An increasing number of lncRNAs playing crucial and specific roles in cancer have been identified, with MALAT1 being one of the well-known examples. 17 Comparing the effects on kinase activity profiles of TRASH-, and MALAT1-inhibition demonstrated the unique pro-survival and anti-apoptotic features of TRASH. To our knowledge, such a pattern of anti-apoptotic kinase activity inhibition through ASO mediated gene silencing has never been reported.
Our research identified additional clinically significant features of TRASH-ASO treatment: MEKiresistance does not desensitize melanoma cells to their TRASH-dependency in vitro, indicating that TRASH-ASOs presents the prerequisite to serve as treatment for melanoma refractory to MEKi. Dual application of TRASH-ASOs and MEKi amplifies the effects of mono application in vitro, demonstrating the synergistic effects of multi-drug regimens that clinical dermato-oncologists strive for. In addition, melanoma cells display early-onset drug resistance to MEKi treatment, while no such response was observed with TRASH-ASO treatment. 16 For systemic in vivo application, we combined TRASH-ASOs with the transfection reagent JetPEI®, which has been tested in clinical trials, including melanoma gene therapy. [55][56][57][58] Interestingly, the combination of TRASH-ASOs and JetPEI® achieved tumor growth reduction using 10-20x reduced ASO dosage, when compared to other ASO treatment studies in rodents. 59,60 ASO based therapy has gained approval by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) and is currently tested in clinical trials for a large repertoire of diseases and with various application routes. 19 Therapeutic options that aim to lower the threshold for apoptosis are a central strategy in oncology and lncRNA-targeting therapy showed promising results in cancer treatment. 17 Hence, TRASH-ASO treatment could play an important role in RNA-targeting melanoma therapy and even be beneficiary for patients who suffer from other malignancies.

Reference Annotation
A custom reference annotation of total 75,506 transcripts, referring to 35,101 genes, of which 16,405 were classified as non-coding, was built by integrating 13,870 lncRNA genes from the GENCODE 61

Assembly and identification of previously unidentified lncRNAs
After alignment to the human genome with TopHat (version 2.0.11), the reads were assembled into transcripts with Cufflinks (version 2.1.1). To discover novel lncRNAs, we excluded all transcript IDs that overlapped with any gene IDs from our initial reference annotation. To filter out transcriptional noise, we kept only multi-exonic transcript IDs which were > 200bp and had at least one intron region > 10bp.
Next, isoforms were merged with Cuffcompare.

Coding Potential Assessment of Transcripts
To identify transcript IDs with a coding potential, we ran (a) the HMMER3 algorithm (considering all 6 open reading frames) to identify any protein family domain as noted in the Pfam database (release 27.0, Pfam-A and Pfam-B domains considered) and (b) the Coding Potential Assessment Tool (CPAT v1.2.1).

Filter for DE genes
Cuffdiff (v.2.1.1) was used to identify differential gene expression analysis between PHM E and PHM Q61 .

TCGA data extraction and processing
Raw .fastq files were obtained from The Cancer Genome Atlas (TCGA) from 86 NRAS mutant melanoma patients from the TCGA-SKCM dataset. Transcript de-novo assembly was performed as described above. Tumor samples were also placed in RNAlater™ Stabilization Solution (Thermo Fisher Scientific®) and stored at -20°C. Invitrogen™ TRIzol™ Solution (Thermo Fisher Scientific®) was used to extract RNA from tissue and qPCR was performed to analyze gene expression.

Immmunohistochemistry
Tumor tissues were extracted from mice immediately after euthanasia and fixed in 10% neutral buffered formalin for 24 hours, followed by storage in 70% EtOH. Histopathology was conducted by the UCSF Histology and Biomarker Core.

Viral transduction
NRAS Q61R cDNA was cloned into the Gateway entry vector pENTR/D-topo. pENTR/D-topo-NRAS Q61R was subjected to site-directed mutagenesis to generate mutants which were then validated by Sangersequencing. NRAS Q61R cDNA in pENTR was cloned into the Gateway cloning-enabled destination vector gFG12. After lentiviral transduction, cells were grown for two weeks followed by cell sorting facilitated by GFP expression intensity on a FACS Aria II cell sorter.

Sanger-Sequencing
RNA from PHM cell-lines was extracted using Purelink™ RNA extraction kit (Thermo Fisher Scientific®) and transcribed into cDNA. Sanger-Sequencing was performed by Quintarabio Inc. For NRAS amplification, the forward primer CGCACTGACAATCCAGCTAA and the reverse primer TCGCCTGTCCTCATGTATTG were used.

Protein extraction and immunoblotting
1 x 10^5 D04 cells were seeded in six well-plates one day prior to transfection. One day after seeding, cells were incubated in media with final oligonucleotide concentration. Total protein lysates were homogenized in 1x RIPA buffer and Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific®) followed by centrifugation at 14,000 RPM/minute at 4°C. Protein concentration was quantified using the Pierce™ BCA Assay Kit (ThermoFisher Scientific®). Linear absorbance was measured using the Synergy™ HT (Agilent Technologies Inc) plate reader. Total protein in 1× Laemmli buffer with

RNA extraction and quantitative real-time PCR (qRT-PCR)
TRIzol™ Solution (Thermo Fisher Scientific®), Phenol:chloroform:isoamyl alcohol (MilliporeSigma®) or NucleoSpin® RNA kit (Takara Bio USA, Inc.) were used for extracting Total RNA from cells and tissues according to the manufacturer's instructions. Total RNA was quantified by NanoDrop™ ND-1000 (Thermo Fisher Scientific®) or Quibit™ 4 (Thermo Fisher Scientific®). 50ng or RNA was reverse transcribed using the cDNA synthesis and gDNA removal QuantiTect® Reverse Transcription Kit (Thermo Fisher Scientific®). Real time PCR was performed using the iTaq TM Universal SYBR® Green Supermix (Bio-Rad Laboratories, Inc.), 10ng (20ng for RIP Assay) of cDNA and on a QuantStudio TM 5 Real-Time PCR System or a 7500 fast real time PCR system (both from Thermo Fisher Scientific®). Relative gene expression was calculated using the comparative Ct method, normalized to GAPDH or β-actin. Primers are listed in Supplementary Table 3. TRASH primer pair 4 was used for knockdown evaluation, splicing efficiency evaluation and RIP. TRASH primer pairs 1, 2, and 3 were used for splicing efficiency evaluation.
Primers were obtained from Integrated DNA Technologies, Inc.

Oligonucleotide transfection
EsiRNA was generated following standard protocol. 63 Primer sequences TCACTATAGGGAGAGACACTCAAAGCCTGAGTAACAGA and TCACTATAGGGAGACTGACTGAGATTTTATTGAGCTGTG were used to create TRASH-targeting esiRNA.
SiRNA was purchased from Horizon Discovery Biosciences Ltd and designed using the siDESIGN software.

Cell viability assay
Dependent on cell doubling time, 0.7-2 x 10^3 cells were seeded in 96 well-plates one day prior to transfection. One day after seeding, cells were incubated in media with final oligonucleotide concentration and/or MEKi and transfection reagent. Three (synergy experiments) or five (solely ASOs) days after transfection Total luminescence was measured on the Synergy™ HT (Agilent Technologies Inc) plate reader using Promega® CellTiter-Glo® and Gen5 software. Cell-growth is shown in relation to cells incubated with Control-ASOs.

Calculation of Combinational Index (CI)
The effects of drug combinations on cell growth were assessed by calculating the combination index (CI) =log2[Ea,b/(-EaxEb)], where Ea and Eb correspond to the effects of drugs A and B alone at a given concentration, and E a,b corresponds to the combined effects of drugs A and B at the same concentration, and a combination index of <0 indicates synergy while a combination index of > 0 relates to an antagonistic effect. The individual combination indices per drug combination were then averaged.

Caspase Glo 3 & 7 assay
Dependent on cell doubling time, 2-3 x 10^3 cells were seeded in 96 well-plates one day prior to transfection. One day after transfection Total luminescence was measured on the the Synergy™ HT (Agilent Technologies Inc) plate reader using The Promega® Caspase-Glo® 3/7 Assay and Gen5 software.
Experiments were performed in quadruplicates.

Colony formation Assay
Dependent on cell doubling time, 1-2 x 10^3 cells were seeded in six well-plates. One day after seeding, cells were incubated in media with 50nM oligonucleotide concentration and transfection reagent. Six days after transfection, cells were washed with PBS, fixed with 10% neutral buffered formalin, and stained with 0.1% crystal violet solution. Colonies were defined as cell conglomerates with >50 cells.
Digital images of plates were evaluated by two independent reviewers for colony counts. The final counts were calculated as the average count of both reviewers for all triplicates.

26
Luminescence data are inversely correlated with the amount of kinase activity. For a more detailed description of the peptide sensors design, sequence and connectivity between peptides and kinases, as well as data normalization steps and analysis, refer to: 35,65,66 . The activity of kinase enzymes is derived from their respective subset of biological peptide targets included in the assay.

RNA-sequencing
Total RNA was isolated using the RNeasy® mini-Kit (QIAGEN N.V.) following the manufacturer's protocol.

Analysis of TRASH-ASO induced DE gene expression
Differential expression (DE) analysis was done using DESeq2. Differentially expressed genes were defined by more than 1.5-fold changes (log2 >0.58 or <-0.58) in expression with FDR<0.05. GO term analysis was done using DAVID 67 Functional Annotation Clustering analysis (version 6.8). 67,68

Statistics and reproducibility
Error bars in all the plots indicate mean ± S.D. P-value < 0.05 was considered statistically significant. ***p-value < 0.001, **p-value < 0.01, *p-value <0.05 by one tailed Student's t-test. All experiments were performed at least three times, unless otherwise indicated. Statistics was calculated with Microsoft® Excel Version 2107.  Error-bars represent standard deviation, significance is shown as p-values calculated by Students t-test. *=p<0.05, **=p<0.01, ***=p<0.001. (cut-off for significance was adjusted p-value < 0.05). Data was obtained from RNA-Seq of D04 melanoma cells, treatment period was three days. In a+c ATP quantitation was used as marker for metabolically active cells. Significance is shown as p-values calculated by Students t-test. *=p<0.05, **=p<0.01, ***=p<0.001. Error bars represent standard deviation.  Cell-growth is relative to incubation with Control-ASOs (TRASH) or drug free media (MEKi). Drug-incubation time was five days (n=3). ATP quantitation was used as marker for metabolically active cells.