Mutant KRAS-driven selective mRNA translation reveals mechanisms and therapeutic vulnerabilities in cancer Authors: Ankita Shrivastava, Eric Nels Pederson, Trang Uyen Nguyen, Jacky Chuen, Prathibha Mohan, Shivangi Pande, Ana Ruiz Toribio, Nicolas Lecomte, Jerry Melchor, John C McAuliffe, Paul M Grandgenett, Vinagolu K Rajasekhar, Kaila Nishikawa, Anna Knörlein, Yael David, Ian M Willis, Christine A Iacobuzio-Donahue, Zhengqing Ouyang, Kamini Singh Published: 2026 ABSTRACT SUMMARYMutant KRAS-driven control of protein synthesis remains poorly defined. Here, we define KRAS-dependent translational programs and their acute remodeling upon KRAS inhibition. We find that mutant KRAS controls the translation of a subset of mRNAs and affects the production of proteins of the mRNA translation apparatus. Interestingly, these specific subsets of mRNAs have short, weakly folded 5′UTRs and harbor low folding energy consensus RNA sequences. We observe ribosome accumulation on selective mRNAs. Our findings clarify the indispensable role of mutant KRAS in regulating mRNA translation, setting it apart from the other previously known mechanisms that depend on mTOR and EIF4E-EIF4A signals. Our findings uncover a mechanism by which mutant KRAS selectively uncouples the translation of mRNAs for protein synthetic machinery from the broader mRNA pool, redefining our understanding of the oncogenic regulation of mRNA translation in cancer. INTRODUCTION mRNA translation is highly regulated and rapidly controls the functional proteome in response to physiological cues in health and disease.1–7 Upregulated mRNA translation and ribosome biogenesis are defining hallmarks of cancers and are required for elevating protein synthesis rates to sustain tumor growth and maintenance.1,2,8,9 Mutant KRAS signaling to ERKs is globally found across many cancer types and intersects with the mTOR and EIF4E-EIF4F complex, thereby influencing cap-dependent mRNA translation to promote tumorigenesis.10–12 Furthermore, mutant KRAS activates MAPK/ERK and PI3K/ AKT/mTOR pathways that converge on ribosomal protein S6 and the translation initiation factor eIF4E to enhance mRNA translation.13,14 With recent clinical advances in KRAS-targeted therapies,15–21 identifying the mutant KRAS inhibitor-dependent changes in the mRNA translational landscape and their role in cellular adaptations will provide a better mechanistic understanding of the therapeutic effects of mutant KRAS-targeted therapies. KRAS activates the oncogene MYC, which can enhance the transcription of ribosomal RNA and translation-related proteins such as EIF4E, further boosting protein synthetic capacity.22–25 Cap-dependent mRNA translation mediated by EIF4E and EIF4A controls the translation of specific subsets of mRNAs to promote oncogenesis.2,3,7,26–29 EIF4A mediates an mRNA translation program that supports oncogenic programs, redox, and central carbon metabolism in the mutant KRAS G12D-driven model of pancreatic ductal adenocarcinoma (PDAC).7,28,29 Many studies have highlighted that KRAS-AKT-PI3K-mTOR-EIF4E activation is central in fueling mRNA translation programs essential for oncogenesis.30 However, the mechanisms by which mutant KRAS selectively reprograms the mRNA translational landscape remain poorly defined, representing a critical gap in understanding how oncogenic KRAS drives tumor progression in the era of KRAS-targeted therapies.15–20 This underscores the need to further explore how mutant KRAS inhibitors induce changes in mRNA translation and the functional proteome in cancer. To study the direct early effects of mutant KRAS inhibition, independent of the secondary late changes that arise due to cellular adaptations, we utilized an acute treatment regimen of mutant KRAS inhibitors in cancer cells. Using this approach, we characterize the mutant KRAS-driven changes in the mRNA translation landscape and identify the hitherto unknown underlying mechanism of regulation. RESULTS Mutant KRAS-dependent changes in mRNA translation We investigated the changes in mRNA translation following acute inhibition of mutant KRAS. We utilized mutant KRAS G12C and KRAS G12D specific inhibitors, sotorasib and MRTX1133, respectively, that induce cell death in human PDAC cell lines harboring the respective KRAS mutations (Figures S1A and S1B). We confirmed that these inhibitors decreased the levels of p-ERK in the respective human cancer cell lines, including PDAC, lung adenocarcinoma, and colorectal cancer cell lines, as early as 2 h post-treatment (Figures S1C and S1D). To study mutant KRAS-dependent genome-wide changes in mRNA translation, we deep-sequenced total RNA (RNA-seq) and ribosome-protected RNA (ribosome footprinting = RF) in human PDAC cell lines, namely, MiaPaca-2 (KRAS G12C) and Panc10.05 (KRAS G12D), treated with mutant KRAS inhibitor sotorasib (10 μM; 2 h) and MRTX1133 (10 μM; 2 h), respectively (Figure 1A). We also performed ribosome footprinting with and without Harringtonine treatment to map the translation start site.31 A schematic of the ribosome footprinting strategy is shown in Figure 1A. We removed reads mapping to ribosomal RNAs, non-coding RNAs, library linkers, and incomplete alignments, thereby selecting for reads between 15 and 35 nucleotides that mapped uniquely to the protein-coding sequence (CDS). The total number of RF counts that mapped to exons was 678,778 in DMSO and 352,960 in sotorasib-treated MiaPaca-2 samples, which corresponded to 15,612 protein-coding genes. The total number of RF that mapped to exons was 6,722,242 in DMSO, and 6,702,263 in MRTX1133-treated Panc10.05 samples, which corresponded to 16,972 protein-coding genes. Complete read statistic tables for RNA and RF are provided in Tables S1A and S1B. While performing read count correlation analyses, we observed distinct clusters between DMSO vs. sotorasib (Figures S1E and S1F) or MRTX1133 (Figures S1G and S1H) treated samples for both total RNA and RF reads, suggesting a strong correlation between replicate samples. Despite relatively lower footprint depth in MiaPaCa-2 cells compared to other models, the Ribo-seq libraries meet established quality metrics, including clear 3-nt periodicity, enrichment of reads within annotated coding regions, appropriate footprint length distributions, and strong concordance between biological replicates. Ribosome footprinting identifies the KRAS-dependent mRNA translation in PDAC cells (A) Schematic of ribosome footprinting strategy from MiaPaca-2 (KRAS G12C) or Panc10.05 (KRAS G12D) cells treated with 10 μM sotorasib or MRTX1133 for 2 h, respectively. Comparison of ribosome-protected RNA (RF) and total mRNA abundance isolates the translational efficiency (TE) for each mRNA genome-wide. (B and C) Changes in TE following sotorasib treatment in MiaPaca-2 and MRTX1133 treatment in Panc10.05 cells. We identify mRNAs with decreased (TE down, red) and increased (TE up, blue) and unchanged translation (background, gray); three biological replicates; the most significantly affected genes are indicated on each side using a q value cutoff of 0.0001 and 0.1 for MiaPaca-2 and Panc10.05, respectively. (D and E) Metagene plot showing the reduction in ribosome coverage in TE down transcripts. KRAS G12C inhibition reduced overall ribosome coverage, while KRAS G12D inhibition significantly reduced the ribosome coverage in the 5′UTR region of the TE down transcripts. Paired two-sided Welch’s test across the coverage in 5′UTR, CDS, and 3′UTR showed a p value of = 7.50 × 10−9 and 4.05 × 10−4 for sotorasib and MRTX1133, respectively. Next, to determine the translation efficiency (TE) for each mRNA, we performed RiboDiff analysis, as previously reported,32 to determine the TE for each mRNA. We used stringent p value and q-value (FDR) cut-offs to define TE downregulated (TE down) and TE upregulated (TE up) mRNAs in both datasets. In order to maintain a similar number of significantly affected genes across both datasets for downstream comparison, we defined the q-value cut-off for the MiaPaca-2 and Panc10.05 datasets as <0.0001 and <0.1, respectively. In MiaPaca-2 cells, we identified 805 TE down and 211 TE up mRNAs following sotorasib treatment (Figure 1B; Tables S2A–S2C). In Panc10.05 cells, we identified 92 TE down and 463 TE up mRNAs following MRTX1133 treatment (Figure 1C; Tables S3A–S3C). Next, we performed metagene analysis to calculate the normalized ribosome coverage in TE down, TE up, and background mRNAs (background = expressed at a TE > 4 but not significantly altered upon drug treatment, q > 0.5). Both sotorasib and MRTX1133 treatments significantly reduced normalized ribosome coverage on TE down mRNAs (Figures 1D and 1E) and slightly, but significantly increased ribosome coverage on TE up mRNAs across the length of the mRNA (Figures S1I and S1J). The background group of mRNAs showed a significant decrease (p = 6.78 × 10−4) in the sotorasib-treated samples, suggesting a generalized global decrease in mRNA translation following sotorasib treatment, while MRTX1133-treated samples showed no significant change in the normalized ribosome coverage (Figures S1K and S1L). Puromycin incorporation of nascently synthesized polypeptides in MRTX1133 and sotorasib-treated samples revealed no significant difference in global protein translation (Figures S1M–S1P). To further investigate these nuances, we delved into understanding how mutant KRAS inhibition affects the translation of specific mRNAs at the transcript level. Targeting mutant KRAS resulted in differential ribosome recruitment to target mRNAs To gain a more holistic understanding of how mutant KRAS inhibition affects ribosome coverage across individual transcripts, we utilized ribosome footprinting data from Harringtonine-treated MiaPaca-2 and Panc10.05 cells to calculate the P-site offset. P-site metagene plots for ribosome coverage revealed that KRAS G12C and KRAS G12D inhibitors induced a buildup of ribosome occupancy at the start site (position 0) across all the open-reading frames (ORFs) detected, suggesting an initiation block following treatment (Figures 2A and 2B). We also observed an accumulation of ribosomes 70–90 nt downstream of the mRNA translation start site, specifically in sotorasib-treated MiaPaca-2 cells, consistent with early translation elongation pausing (Figure 2A). Ribosome coverage increased upstream of the mRNA translation start site in TE down targets following sotorasib and MRTX1133 treatment, suggesting increased mRNA translation from the 5′ untranslated regions (Figures 2C and S2A). Ribosomes also accumulated upstream of the stop site in TE down targets, indicating a slowdown or block in ribosome elongation following sotorasib and MRTX1133 treatment (Figures 2D and S2B). Ribosome coverage is affected by mutant KRAS inhibitors (A and B) Ribosome coverage on all ORFs showed increased ribosomes at the initiation site following sotorasib and MRTX1133 treatment in MiaPaca-2 and Panc10.05 cells, respectively (p = 2.73 × 10−2). No significant effect on the CDS in MRTX1133-treated Panc10.05 cells (p = 0.414). (C) Ribosome coverage on TE down transcripts showed the accumulation of ribosomes upstream and downstream of the start site following sotorasib treatment in MiaPaca-2 (p = 4.00 × 10𢈒2), suggesting a slowdown in translation. (D) Ribosome coverage on TE down transcripts showed the accumulation of ribosomes upstream of the stop site following sotorasib treatment in MiaPaca-2 (p = 1.66 × 10−4), suggesting a block in translation elongation. (E) An increase of ribosome coverage between 70 and 90 nt region was observed in TE up transcripts following sotorasib (blue) treatment over control (black) in the 5′UTR and CDS around the start site (p = 1.73 × 10−2). (F) Fraction of pause-site codons (RxK, KKK, PP, PD, and PG) across the detected ORFs for the first 50 codons. (G, H, and I) Sucrose gradient fractionation of RNAse I digested polysome showing the accumulation of disome in MiaPaca-2 cells treated with sotorasib and quantification of the disome peak relative to DMSO (p < 0.05). Immunoblot shows RPL5 and RPS10 in the indicated fractions collected from RNase I-digested polysome following DMSO vs. sotorasib treatment in MiaPaca-2. (J, K, and L) Immunoblot shows the induction of ZNF598, EDF1, p-ZAKα, ZAKα, RPS10-Ub following sotorasib treatment in MiaPaca-2. Anisomycin (ANS) treatment is included as a positive control for ribosome stalling and collision-induced stress signaling. β-actin is used as a loading control. (M) Mutant KRAS G12C inhibitor sotorasib affects translation initiation and elongation, resulting in ribosome stalling, while mutant KRAS G12D inhibitor MRTX1133 predominantly reduces translation initiation. Among the TE up transcripts, we observed an increased ribosome accumulation in both the 5′ untranslated regions upstream of the start site and at the +70–90 nt position only in sotorasib-treated MiaPaca-2 cells (Figure 2E). In contrast, MRTX1133-treated Panc10.05 cells did not show enrichment of ribosome coverage at either the +70–90 nt position or upstream of the start site in the TE up targets, suggesting that KRAS G12D inhibition primarily affects translation initiation without inducing translation elongation pausing (Figure S2C). Similar to the TE down targets, we observed a modest increase in ribosome accumulation upstream of the stop site following sotorasib, indicating a block in ribosome elongation (Figure S2D). This accumulation was absent in MRTX1133-treated samples, suggesting that KRAS G12D inhibition does not block ribosome elongation on TE up transcripts (Figure S2E). An increase in ribosome occupancy in TE up transcripts may be the consequence of an initiation block, increased ribosome stalling, slower elongation across the transcript, or delayed termination. To further delineate how mutant KRAS inhibitors increase ribosome occupancy among the TE up transcripts, we performed pause site codon analysis and found that ribosome accumulation strongly and positively correlates with PP, PG, PD, RXK, and KKK correlating with codons rich in GC and basic amino acids that may mediate an increased tendency for ribosome stalling (Spearman r = 0.36; p value = 0.0104) (Figure 2F). Consistent with these findings, polysome profiling of MiaPaca-2 cells treated with sotorasib following RNase I digestion revealed a marked accumulation of disome, indicative of ribosome stalling and/or collision events (Figure 2G). Quantification of the disome peak showed a significant increase compared to DMSO, further supporting sotorasib-induced ribosome stalling (Figure 2H). Immunoblot analysis of sucrose gradient fractions confirmed the presence of both small and large ribosomal subunit proteins (RPS10 and RPL5), demonstrating that intact ribosomes accumulate within these stalled fractions (Figure 2I). There was a non-significant increase in the 80s ribosome abundance (Figure S2F). Ribosomal proteins RPL5 and RPL11 remained unchanged in total protein lysates, supporting preserved levels of these ribosome proteins (Figure S2G). Next, polysome profiling of MiaPaca-2 cells treated with sotorasib in the absence of RNAse I digestion showed an overall decrease in ribosome abundance, indicating reduced mRNA translation (Figure S2H). Ribosome stalling is known to activate ribosome quality control pathways, including ZNF598 and EDF1, and the ribosome collision activates phosphorylation of stress sensor ZAKα.33,34 We observed increased protein levels of ZNF598 and EDF1 in MiaPaca-2 treated with sotorasib (Figure 2J). Importantly, sotorasib treatment also induced a robust and time-dependent increase in phosphorylated and total ZAKα (Figure 2K). In contrast, MRTX1133-treated Panc10.05 cells did not exhibit the activation of these ribosome quality control markers, suggesting that this stress response is specific to sotorasib treatment (Figure S2I). The RNase I-treated polysome profiles are consistent with increased disome formation following mutant KRAS G12C inhibition; however, we note that sucrose gradient analysis alone cannot definitively demonstrate ribosome collision, and therefore, we further utilized complementary biochemical evidence by investigating the ubiquitination of the small ribosomal subunit protein, RPS10. Sotorasib treatment induced an increase in ubiquitinated RPS10 in a time-dependent manner similar to the response observed following anisomycin (ANS) treatment, a known inducer of ribosome collision, indicating the activation of the ribosome collision pathway (Figure 2L). Together, these findings demonstrate that mutant KRAS inhibitors induce an initiation block in translation, while sotorasib uniquely promotes ribosome stalling and collision during translation elongation (Figure 2M). Neither of the mutant KRAS inhibitors utilized in these studies showed any effect on mRNA translation termination kinetics and 3′ UTR read-throughs. Which transcripts are affected by mutant KRAS? We compared the mRNA transcripts affected by sotorasib versus MRTX1133 and observed 152 transcripts that were commonly affected by both the above inhibitors (Figure S3A). Next, we performed pathway analysis using STRING (STRING: functional protein association networks). Sotorasib downregulates the mRNA translation of transcripts involved in RNA processing, metabolism, protein processing, and translation (Figure S3B). MRTX1133 targeted the mRNA translation of transcripts involved in ECM organization, collagen, glycoside hydrolase activity, pancreatitis, calcium regulation, and ubiquitin ligase activity (Figure S3C). Unique TE down transcripts affected by MRTX1133 showed pathways associated with pancreatic secretion, ECM organization, peptidase activity, keratin, and complement and coagulation pathways (Figure S3D). Both sotorasib and MRTX1133 inhibited EEF1A1 translation among other translation factors, while sotorasib has additional TE down targets involved in RNA binding, metabolism, and chromatin binding (Figures 3A–3C; S3E). Systematic analysis of mutant KRAS-dependent changes in mRNA translation (A and B) TE targets are downregulated in both sotorasib and MRTX1133 or only with sotorasib-treated MiaPaca-2 and Panc10.05 cells, respectively. (C) Translation efficiency of translation factors is affected by sotorasib, EEF1A1 being downregulated by the drug. (D–H) Immunoblot analyses show the downregulation of RPS6, EIF4E, EEF1A, and EEF2 following sotorasib treatment in MiaPaca-2 cells, SW837 cells, and RMC-7977 treatment in MiaPaca-2 cells and MRTX1133 treatment in Panc10.05 cells. β-actin is used as a loading control. (I) Intracellular immunostaining and FACS analysis show the downregulation of EEF1A in human primary PDAC-derived organoids treated with RMC-7977 (* p < 0.05, ** p < 0.01). (J–L) In vivo MiaPaca-2 xenograft studies show reduced tumor volume upon RMC-7977 treatment compared to vehicle, and immunoblots show the downregulation of RPS6, EIF4E, and EEF2 protein levels in xenograft tissues (* p < 0.05). (M and N) Western blot analysis and quantification of EEF1A1, EEF2, and RPS6 proteins show upregulation in KRAS G12D, or KRAS G12C compared to isogenic KRAS WT HEK293 cells (** p < 0.01 and *** p < 0.001). β-actin is used as a loading control. (O and P) Analysis of pancreatic, lung, and colorectal cancer cell lines with distinct KRAS genotypes [WT: H6C7], [G12D: Panc10.05, PANC-1, AsPC-1], [G12C: MiaPaca-2, H358, SW837] shows that KRAS G12D and KRAS G12C cells exhibit elevated protein expression of EEF1A1, EEF2, RPS6, and EIF4E compared to KRAS WT (** p < 0.01 and *** p < 0.001). β-actin is used as a loading control. (Q–S) Comparison of sotorasib TE targets with total proteomics data includes all eight trajectory clusters (C1-C8). Most of the TE down and TE up targets, which include cytoplasmic ribosomal proteins and translation factors, are downregulated at the protein level (clusters C1, C2, C5, C6, and C7). TE down mitochondrial ribosomal proteins are upregulated at the protein level (cluster C3, C4, C8). We validated that the total protein levels of translation initiation factors, RPS6 and EIF4E, and translation elongation factors, EEF1A1 and EEF2, are downregulated following treatment with sotorasib, MRTX1133, and RMC-7977 (a potent multi-selective RAS(ON) inhibitor that targets the GTP-bound form of RAS) (Figures 3D–3H). RMC-7977 treatment inhibited p-ERK levels in MiaPaca-2 cells (Figure S3F). Protein levels of RPS6 and EIF4E remained unchanged following MRTX1133 treatment in mutant KRAS G12D expressing PDAC cells (Figures S3G and S3H). The total mRNA expression of EEF1A1, EEF2, RPS6, EIF4E, ACADM, DLAT, ECHS1, and PDHB was not affected, suggesting that mutant KRAS inhibitors affect the protein levels due to changes in mRNA translation and not due to transcriptional changes (Figures S3I and S3J). Sotorasib treatment reduced total protein levels of DLAT, ECHS1, PDHB, and ACADM proteins involved in short and medium-chain fatty acid metabolism (Figure S3K). These proteins were not affected by MRTX1133 treatment in Panc10.05 cells (Figure S3L). Next, using human primary PDAC patient-derived organoids harboring the KRAS G12D mutation, we show that EEF1A1 protein is downregulated following RMC-7977 and MRTX1133 treatment in vitro (Figures 3I and S3M). Through colony formation assays, we observed that RMC-7977 and MRTX1133 treatment induced cell death in these organoids (Figures S3N and S3O). Next, we generated MiaPaca-2-derived xenografts in J:Nu mice and treated them with vehicle or RMC-7977 in vivo. Compared to the vehicle, RMC-7977 treatment reduced tumor growth significantly (Figure 3J). We observed that RMC-7977 reduced the total protein expression of RPS6, EIF4E, and EEF2 while the total mRNA levels remained unchanged (Figures 3K, 3L, and S3P). To confirm whether the gain-of-function of mutant KRAS would exhibit similar effects, we generated isogenic HEK-293T lines expressing KRAS wildtype (WT), KRAS G12D, or KRAS G12C. Immunoblot analyses confirmed the expression of the respective KRAS constructs in these isogenic models (Figure S3Q). Compared to KRAS WT, both KRAS G12D and G12C expressing cells showed increased protein levels of EEF1A1, EEF2, and RPS6 (Figures 3M and 3N). We next examined whether this effect extends across PDAC, lung adenocarcinoma, and colorectal cancer cell lines expressing distinct KRAS mutations. In comparison to the KRAS WT expressing cells, H6C7, we observed increased protein levels for EEF1A1, EEF2, RPS6, and EIF4E in KRAS G12C and KRAS G12D expressing cell lines, further supporting the role of mutant KRAS in controlling translational machinery components (Figures 3O and 3P). To determine whether mRNA translational changes correlate with changes in protein abundance, we examined a published proteomics dataset35 comparing total protein levels in MiaPaca-2 cells treated with the mutant KRAS inhibitor ARS-1620 over 1-, 3-, and 7- days. At the global levels, we observed weak or no correlation between TE and protein abundance in the two datasets (Figure S3R, correlation between sotorasib (2 h) treated MiaPaca-2 cells and ARS-1620 1 day treatment: Spearman r = −0.0822, p = 0.0145, ARS-1620 3-day treatment: Spearman r = 0.00338, p = 0.9213, and ARS-1620 7-day treatment: Spearman r = −0.0117, p = 0.7329, respectively). The disconnect between mRNA translation and protein expression at the global level could be due to several factors, including differences in model systems or a result of mRNA degradation kinetics between mRNA and protein levels. To specifically test whether the downregulation of TE results in the downregulation of protein levels, we narrow our analysis to only TE down targets. Within TE down targets, we observed a total of eight trajectory clusters (C1-8) for changes in protein levels. For each time point, a gene is either upregulated ( ^) or downregulated/unchanged (v) and this pattern determines the trajectory cluster where unbiased clustering defined eight clusters, C1 = vvv (1/3/7 days), C2 = vv^, C3 = v^^, C4 = v^v, C5 = ^v^, C6 = ^vv, C7 = ^^v, and C8 = ^^^ (Figure 3Q). The majority of TE down targets fall in C1, C2, and C6, reflecting the downregulation of protein levels, with the exception of a smaller fraction of TE down targets that showed the upregulation of protein abundance in C7 and C8 (Figure 3Q). TE down targets in clusters C2, C3, and C5 showed variable levels of total protein abundance at early and late time points, suggesting a more dynamic regulation at the protein level (Figure 3Q). Similar analyses performed on the TE up targets demonstrate that a majority of TE up targets fall under clusters C1, C6, and C7, indicating that increased ribosome occupancy is correlated with reduced protein levels (Figure 3R). A small portion of mRNAs with increased ribosome occupancy has a bona fide increased protein abundance across 1–7 days of ARS-1620 treatment, as indicated by cluster C7 and C8 (Figure 3R). These observations support our theory that most TE up targets induced through mutant KRAS inhibitor treatment have an increased ribosome occupancy as a result of ribosome stalling or delayed ribosome initiation and elongation kinetics that likely result in lower protein abundance. Next, when we focused on cytoplasmic translation factors (EIs, EEs, and RPs), we observed an overwhelming positive correlation between TE and protein abundance, as the majority of the TE down targets clustered in C1, suggesting that mutant KRAS inhibition downregulates translation factors both at the mRNA translation and at the protein level (Figure 3S). Interestingly, mitochondrial ribosomal proteins demonstrate an anti-correlative trend between the mRNA translation and the proteomic dataset, as observed in the prominence of clusters C7 and C8 (Figure 3S). In summary, we demonstrate that mutant KRAS inhibitors control the mRNA translation of key cytoplasmic translation initiation and elongation factors, which is corroborated by downregulated protein levels. Conserved features of mutant KRAS-dependent mRNA transcripts We investigated whether any conserved sequences and structures in mRNA transcripts could define the translational dependency on mutant KRAS activity. We compared the 5′UTR and 3′UTR sequence composition and structure of TE down, TE up, and background groups identified from MiaPaca-2 and Panc10.05 cells treated with sotorasib and MRTX1133, respectively. We observed that TE down mRNAs have significantly shorter 5′UTR than TE up and background groups (Figure 4A). 5′UTR folding energy was substantially smaller in magnitude for the TE down mRNAs as compared to the TE up and background groups (Figures 4B and 4C). This suggests that the translation of mRNA transcripts with shorter 5′ UTR and less stable secondary structures is likely to be affected following mutant KRAS inhibitor treatment. Next, with de novo sequence analysis, we identified two conserved sequence motifs in sotorasib and three conserved sequence motifs in MRTX1133 TE down targets (Figures S4A and S4B). These motifs are enriched in approximately 2% of the sotorasib TE down group and approximately 12–15% of the MRTX1133 TE down group (Figures S4C and S4D). 5′UTR structure folding analysis revealed enrichment of IRES motifs in sotorasib TE down targets (Figure S4E). We did not identify any conserved features in the 3′UTR of TE down, TE up, and background groups. Conserved features in 5′UTR define the translational dependency on mutant KRAS activity (A) TE down targets of sotorasib have significantly shorter 5′UTR (average length = 252 nt) compared to TE up (average length = 371 nt) and background transcripts (average length = 461 nt). (B and C) 5′UTR of TE down targets exhibit significantly less folding energy density compared to the TE up and background targets in both sotorasib and MRTX1133 datasets. (D) All Translation factors (EIs, EEs, and RPs) have shorter 5′UTR length (average length = 109 nt) compared to the rest of the mRNA (average length = 230 nt) in the transcriptome. (E–G) Relative cap-dependent Firefly luciferase activity and IRES-dependent Renilla luciferase in control versus indicated drug-treated (sotorasib 10 μM, RMC-7977 10 μM, and MRTX1133 10 μM, for 24 h) in MiaPaca-2 and Panc10.05 cells. Luciferase activity is normalized to total protein levels. (G) The IRES sequence and structure found in the BAG1 gene. (H and I) The predicted secondary structure of the wildtype 5′UTR sequence of canonical EEF1A1 transcript sequences contains putative IRES structurally resembling BAG1 IRES, and the predicted secondary structure of the mutated 5′UTR that lacks IRES-like secondary structure. (J) Firefly luciferase activity of full-length and mutated 5′UTR of EEF1A1 following sotorasib treatment (10 μM; 24 h) in MiaPaca-2 cells. Luciferase activity is normalized to total protein levels. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. Next, to investigate the mutant KRAS-mediated regulation of cap-dependent and IRES-dependent mRNA translation, we utilized the previously published dual luciferase reporter assay.7 Sotorasib, MRTX1133, and RMC-7977 treatment significantly reduced the cap and IRES-dependent mRNA translation, suggesting that mutant KRAS facilitates cap and IRES-mediated mRNA translation (Figures 4E and 4F). Based on the similarity between a known IRES containing 5′UTR of the BAG1 gene, we identified that the 5′UTR of EEF1A1 contains an IRES-like structure (Figure 4G). We mutated 9 nt (10% of total) to adenine in the wildtype 5′UTR of EEF1A1, resulting in a 5′UTR that no longer folds into an IRES-like structure (Figures 4H and 4I). Next, we cloned the full-length and mutated 5′UTR of EEF1A1 in the luciferase reporter system. Sotorasib and RMC-7977 treatment reduced the translation mediated by the full-length EEF1A1 5′UTR but did not affect the translation mediated by the mutant EEF1A1 5′UTR, suggesting that translational repression utilizing mutant KRAS inhibitors is dependent on the specific identified sequence and clover-leaf-like secondary structure of the 5′UTR (Figure 4J and S4F). In summary, our data suggest that the translation of mRNAs with shorter 5′UTR and less folded structures is preferentially downregulated by mutant KRAS inhibition. Comprehensive comparison of KRAS, mTOR, and EIF4A translational targets Mutant KRAS signaling cross-talks with mTOR and the EIF4E-EIF4A signaling axis.2,14 Therefore, we hypothesize that mRNAs that are differentially translated following mutant KRAS inhibitor treatment are expected to overlap with translational targets of mTOR and EIF4E-EIF4A. The translational landscape of mTOR and EIF4A is well-characterized in previous studies.7,28,36 Using these existing datasets, we compared the translational targets of sotorasib with the mTOR inhibitor, Torin1. Overall comparison of differentially translated transcripts between sotorasib and Torin1 revealed that ~62% of sotorasib targets are also controlled by Torin1, while only 16% of Torin1 targets are affected by sotorasib (Figure S5A). Comparison of sotorasib TE down targets with Torin1 TE targets revealed that only half of the sotorasib TE down targets are also downregulated by Torin1, while the rest are TE up following Torin1 treatment (Figure S5B). The majority (2/3) of sotorasib TE up targets are also upregulated following Torin1 treatment (Figure S5C). While it has been previously shown that mTOR controls the translation of ribosomal protein mRNAs,36 we were curious to understand whether mTOR and mutant KRAS regulate the same or different ribosomal proteins and translation factors. To elucidate this, we systematically compared the extent and the effect of mutant KRAS and the mTOR inhibitor on these specific ribosomal proteins. There are a total of 253 proteins in the transcriptome classified as eukaryotic initiation (EIs), eukaryotic elongation (EEs), ribosomal proteins small subunit (RPSs), ribosomal proteins large subunit (RPLs), mitochondrial ribosomal protein-small subunits (MRPSs), and mitochondrial large ribosomal protein large subunits (MRPLs) that constitute the translational machinery (Figure S5D). Sotorasib affected ~32% of these proteins, while Torin1 affected a larger fraction (63%) of translation factors, with an overlap of 25% affected by both sotorasib and Torin1 (Figures S5E–S5G). We observed that mutant KRAS inhibition predominantly affected the translation of RPLs, while mTOR inhibition affected MRPLs (Figures S5E and S5F). Within the common targets, RPLs, MRPSs, and MRPLs are the major fraction affected by Torin1 and sotorasib (Figure S5G). The majority of these EIs, EEs, and RPs are downregulated upon KRAS and mTOR inhibition, except for a small fraction of MRPSs/MRPLs that are TE up following mTOR inhibition (Figure S5H). This data suggested that mutant KRAS and mTOR overall downregulated the translation of translation factor mRNAs with specific effects on distinct classes of ribosomal proteins. Next, we compared the translational effect of mutant KRAS with EIF4A. We utilized a previously published dataset that described the EIF4A-dependent changes in PANC-1 cells.28 The overall comparison showed that ~34% of the sotorasib targets are also regulated by EIF4A inhibitor CR-1-31B, while only 3% of CR-1-31B are affected by sotorasib, which suggests that mutant KRAS regulates most of the EIF4A translational targets while EIF4A targets a broader subset of mRNAs that may not be directly dependent on mutant KRAS activity (Figure S5I). Within the sotorasib TE down targets, only a small fraction of mRNA transcripts is also downregulated by CR-1-31B, while a majority is TE up following EIF4A inhibition (Figure 5A). Within the sotorasib TE up targets, half of the mRNA transcripts are also TE up by CR-1-31B, and the other half is TE down (Figure 5B). This data suggests that mutant KRAS and EIF4A activity regulate the same transcripts in opposite directions. Thirty-five (~13%) of translation factors are commonly regulated by EIF4A and mutant KRAS (Figure S5J). Interestingly, while sotorasib downregulated the translation of these targets, CR-1-31B resulted in the TE upregulation of these same translation factors (Figure 5C). Following this same trend, we also noticed that the translation of mRNAs for translation factors, oxidative phosphorylation, and fatty acid metabolism are regulated by EIF4A and mutant KRAS in a reciprocal fashion (Figures S5K–S5M). Comparison of mutant KRAS-dependent mRNA translation with mTOR and EIF4A-dependent mRNA translation (A and B) TE fold change analysis of common transcripts affected by sotorasib and CR-1-31B shows that mutant KRAS and EIF4A oppositely regulate the translation of these mRNAs. Most of the sotorasib TE down targets are TE upregulated following EIF4A inhibition with CR-1-31B. (C) TE fold change analysis of common translation factors between sotorasib and CR-1-31B reveals the upregulation of all the translation factors following CR-1-31B treatment, except for four proteins that are downregulated by CR-1-31B. (D and E) Relative luciferase activity driven by G-quadruplex (GQ) and GC-rich motifs in response to sotorasib, RMC-7977, or MRTX1133 treatment demonstrates that sotorasib induces GQ and GC-rich mRNA translation ( * p < 0.05 and *** p < 0.001). (F and G) Luciferase reporter assays using the full-length or mutated 5′UTR of EEF1A1 show altered luciferase activity upon CR-1-31B and Torin1 treatment in MiaPaca-2 cells (* p < 0.05 and *** p < 0.001). (H) Combination treatment of CR-1-31B and Torin1 with mutant KRAS inhibitors in a panel of cancer cell lines harboring either G12C or G12D KRAS mutations leads to an additive enhancement of cell death, based on the zero interaction potency (ZIP) model. (I) Summary of our findings shows that the mutant KRAS-dependent mechanism of translational control uncouples the regulation of translation machinery from the regulation of the rest of the mRNA translation. Mutant KRAS-dependent mechanism is distinct from the mTOR and EIF4A-dependent mechanism that depends on longer and structured 5′UTR features and controls the translation of a distinct subset of mRNAs. mTOR positively regulates TOP-motif dependent translation while EIF4A facilitates RNA GQ dependent mRNA translation.36 To better understand whether mutant KRAS also mediates the translation of mRNAs with these specific motifs, we performed luciferase reporter assays using different motifs and observed that sotorasib treatment induced TOP, RNA GQ, and GC-rich dependent mRNA translation (Figures S5N, 5D, and 5E). In line with these findings, we observe that sotorasib treatment increases p-S6, suggesting a feedback activation of mTOR signaling (Figure S5O). MRTX1133 and RMC-7977 did not induce TOP, RNA GQ, and GC-rich dependent mRNA translation and reduced p-S6 protein expression, suggesting that these inhibitors may not result in the feedback activation of mTOR and EIF4A dependent mRNA translation (Figures S5N, S5O, 5D, and 5E). In our previous studies, we established that EIF4A regulates RNA G-quadruplex-mediated mRNA translation and utilizes IRES-mediated translation to escape EIF4A inhibitor-mediated repression.7,28 We observed that CR-1-31B treatment induced mRNA translation of the full-length 5′UTR of EEF1A1, which contains an IRES but had no effect on the truncated 5′UTR (Figure 5F). Torin1 did not show any effects on the translation mediated by either the full length or the truncated 5′UTR of EEF1A1, suggesting that IRES-dependent mRNA translation is not regulated through mTOR activity (Figure 5G). This study suggests that while mutant KRAS preferentially regulates the translation of transcripts with shorter, weakly folded, and IRES-containing 5′UTRs, the same transcripts may escape the EIF4A inhibitor-mediated repression of translation. Distinct yet complementary roles of mutant KRAS, mTOR, and EIF4A in mRNA translation Based on these findings, we speculated that combining CR-1-31B and/or Torin1 with a mutant KRAS inhibitor may enhance cell death in PDAC cells. We show that the combination of both CR-1-31B or Torin1 with mutant KRAS inhibitors exhibits an additive cytotoxic effect in a panel of PDAC cell lines, as indicated by a zero interaction potency (ZIP) score of −10 to 10 (Figures 5H and S5P). The additive effects suggest that mutant KRAS, EIF4A, and mTOR affect the translation of mRNAs that are in part independent of each treatment condition, as opposed to interdependent, which would be predicted to have synergistic combinational effects. Thus, we summarize that mutant KRAS, mTOR, and EIF4A regulate the mRNA translation of specific targets, and each node of signaling follows characteristically unique mechanisms that depend on the specific 5′UTR features of the target mRNA. Mutant KRAS activity controls the mRNA translation of the translation machinery that tends to have shorter and weakly folded 5′UTRs, conferring more accessibility for translation, while EIF4A is preferentially required for the mRNA translation of transcripts with longer and highly structured 5′UTRs (Figure 5I). Therefore, targeting both mutant KRAS with mTOR or EIF4A may have a profound effect on the global mRNA translation (Figure 5I). Our data suggests that mutant KRAS, mTOR, and EIF4A inhibition may all independently target similar mRNA transcripts, but in a similar and/or opposite direction. Moreover, co-targeting mutant KRAS, mTOR, and EIF4A may provide better translational repression and is worth investigating as a prospective functional therapy approach. DISCUSSION A mutant KRAS selective mechanism of translational control Mutant RAS and AKT signaling have been implicated in the differential recruitment of ribosomes to specific transcripts in glioblastoma.37 However, the underlying mechanisms of how mutant RAS regulates the translation of the specific mRNAs remain unknown. Our findings provide new insight into the mechanism of mutant KRAS-dependent control of mRNA translation. Using mutant KRAS inhibitors and ribosome footprinting in human PDAC cells, we define the mutant KRAS-dependent translational landscape in cancer. Mutant KRAS inhibition results in a reduction in ribosome occupancy and TE of a specific subset of TE down mRNAs, while also slowing down the mRNA translational kinetics of another subset of TE up mRNAs. Mutant KRAS inhibitors mediate translational repression in part by reducing the translational initiation of specific transcripts and in part by inducing ribosome stalling. Ribo-seq measures relative ribosome occupancy rather than absolute protein synthesis. Thus, the observation that more transcripts show increased ribosome density does not indicate increased global translation. Instead, consistent with reduced elongation factor expression and decreased protein output, our data support a model in which KRAS inhibition slows elongation, increasing ribosome residence time on subsets of transcripts. This translation elongation-limited state leads to increased ribosome density despite an overall reduction in global protein synthesis. Among the transcripts with reduced ribosome occupancy (TE down), we observe an enrichment of transcripts with short- and weakly structured 5′UTRs that contain specific RNA motifs that could fold into clover-leaf structures. These findings suggest a preferential downregulation mediated by mutant KRAS inhibition. Interestingly, a significant number of these transcripts include translation factors and ribosomal proteins that may further amplify translational repression. Among the transcripts with an increased ribosome occupancy (TE up), we observe an enrichment of ribosomes at the PP, PG, PD, RXK, and KKK peptides correlating with codons rich for GC and basic amino acids that may be responsible for conferring susceptibility to ribosome stalling. The distinct mechanism of mutant KRAS-dependent control of the mRNA translation machinery Mutant KRAS cooperates with mTOR-EIF4E signaling to promote oncogenesis, and therapeutics co-targeting these pathways have been established in cancer.2,13,14,37,38 Many transcripts encoding ribosomal proteins and translation factors possess short 5′UTRs. Therefore, reduced mTORC1 activity could potentially contribute to the observed translational phenotype. However, comparison of KRAS inhibitor-dependent translational changes with mTOR inhibition revealed only a partial overlap between these regulatory programs. These findings suggest that while some short 5′UTR transcripts may be sensitive to mTOR signaling, the translational program downstream of mutant KRAS inhibition is largely distinct from canonical mTOR-dependent TOP mRNA regulation. Studies showed that EIF4A/4F activity can confer resistance to BRAF inhibitors, and EIF4A inhibitors are highly effective in combination with KRAS inhibitors in lung cancer.30,39 Our data identify the specific mechanisms that uncouple the translational targets of mutant KRAS and EIF4A. These findings support the concept of combining EIF4A inhibitors with mutant KRAS inhibitors to enhance the therapeutic efficacy in cancer.39 We establish that mutant KRAS and EIF4A control mRNA translation of a distinct subset of mRNAs through unique mechanisms. While EIF4A inhibitor repressed the translation of transcripts with longer, structured, and RNA G-quadruplex containing 5′UTRs, mutant KRAS preferentially targets transcripts with short and weakly structured 5′UTRs. This mutant KRAS-dependent mechanism differs from mTOR and EIF4E-EIF4F regulation, which governs the cap-dependent translation of mRNAs with highly structured, GC-rich 5′ UTRs or TOP motifs.4,7,28,36,40 Consistent with our findings, mutant KRAS inhibitors in combination with mTOR inhibitors or EIF4A inhibitors show additive or synergistic effects, further supporting our findings that KRAS, mTOR, and EIF4A inhibition targets the translation of specific and distinct subsets of mRNA.41 Our study demonstrates unique mechanisms that uncouple the translation of translational machinery mRNAs from the rest of the mRNAs. Herein, we observe that the former is directly controlled by mutant KRAS activity, while the latter is regulated by EIF4A/4F activity. Uncoupling of regulation is mediated through 5′UTR length, sequence, and structure. Crucially, our findings reveal that mutant KRAS orchestrates the translation of a discrete subset of mRNAs independent of canonical mTOR and EIF4E-EIF4A signaling. Co-targeting these pathways offers a rational strategy to selectively inhibit tumor-promoting mRNA translation, minimize off-target effects, and mitigate therapy-induced resistance. Collectively, our work defines a mutant KRAS-dependent mechanism of translational control that reprograms the translational landscape in cancer and provides a compelling rationale for combinatorial therapeutic intervention. Limitations of the study While in this study, we focused on the immediate effects of mutant KRAS inhibition on mRNA translation, providing mechanistic insight into mutant KRAS-dependent regulation of mRNA translation, several limitations should be considered when interpreting the findings. First, our analyses primarily capture the acute effects of KRAS inhibition following 2-h treatment in vitro, which were intentionally designed to isolate direct translational consequences while minimizing secondary adaptive responses. However, this approach does not fully recapitulate the temporal complexity of tumor adaptation, where prolonged drug exposure can induce transcriptional rewiring, metabolic changes, and selection of resistant clones. Indeed, proteomic comparisons indicate a limited correlation between early TE changes and longer-term protein abundance, underscoring the need for longitudinal in vivo studies to define durable translational states. Characterizing mRNA translational changes in patients with PDAC, lung adenocarcinoma, and colorectal cancer receiving mutant KRAS inhibitors will provide important insights into how translational reprogramming contributes to therapeutic response and resistance. Second, although ribosome profiling provides high-resolution insights into ribosome occupancy at the single-nucleotide resolution, it does not directly measure protein synthesis rates or functional protein output. Increased ribosome density observed in TE-up transcripts may reflect ribosome stalling, altered elongation kinetics, or delayed termination, rather than productive translation. While we incorporated orthogonal approaches such as polysome profiling, RPS10 ubiquitination, and ZAKα activation, fully disentangling initiation versus elongation versus collision-driven effects remains technically challenging and may require complementary approaches such as ribosome run-off assays, nascent proteomics, or single-molecule translation imaging. Importantly, the relatively lower footprint depth in MiaPaCa-2 cells compared to other models should be considered as a technical limitation. Although the Ribo-seq libraries meet established quality metrics, including clear 3-nt periodicity, enrichment of reads within annotated coding regions, appropriate footprint length distributions, and strong concordance between biological replicates, supporting the robustness of the dataset, with the possibility of variation between different models. Third, our study relies on a defined set of cell line models and selected KRAS inhibitors (sotorasib, MRTX1133, RMC-7977). Although we include multiple cancer types and patient-derived organoids, these systems may not fully capture the heterogeneity of human tumors, including stromal interactions, immune components, and microenvironmental constraints that influence translational control. Moreover, the distinct pheno-types observed between KRAS G12C and G12D inhibition, such as ribosome stalling versus initiation block, highlight that findings may be allele- and drug-specific, limiting broad generalization across all KRAS-driven cancers. Fourth, while we identify 5′UTR features associated with KRAS-dependent translation, these associations are largely correlative. Functional validation was performed for select candidates, but a comprehensive dissection of causal RNA elements and RNA-binding proteins governing this selectivity remains incomplete. In addition, contributions from RNA modifications, codon optimality, tRNA availability, and ribosome heterogeneity were not systematically evaluated and may further refine translational specificity. Fifth, although we demonstrate partial overlap and functional interplay between KRAS, mTOR, and EIF4A pathways, the mechanistic basis of their crosstalk and feedback regulation remains incompletely defined. For example, sotorasib-induced activation of mTOR signaling and RNA GQ-dependent translation suggests compensatory translational rewiring, but the extent to which this contributes to therapeutic resistance in vivo requires further investigation. Importantly, the additive rather than strongly synergistic effects observed with combination treatments indicate that network-level redundancies may buffer translational inhibition. Finally, while our findings suggest that adaptive rewiring of translation via mTOR- and EIF4A-dependent pathways may represent a mechanism of resistance to KRAS-targeted therapies, this remains a predicted model that requires validation in patient samples and clinical settings. Comprehensive profiling of translational states in tumors from patients undergoing KRAS inhibitor therapy will be essential to establish the clinical relevance, durability, and therapeutic tractability of these mechanisms. RESOURCE AVAILABILITY Lead contact Further information and requests for resources and reagents should be directed to the lead contact, Kamini Singh (kamini.singh@einsteinmed.edu). Materials availability All materials generated in this study are included in this article and its supplemental information. Materials that can be shared will be made available to qualified researchers under a material transfer agreement for non-commercial research use. STAR★METHODS EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS Cell lines The following human cell lines MiaPaca-2 (ATCC CRL-1420), Panc10.05 (ATCC CRL-2547), PANC-1 (ATCC CRL-1469), 293T (ATCC CRL-3216), and AsPC-1 (ATCC CRL-1682), SW837 and SW1463 (gift from Dr. Chaoyuan Kuang, Department of Molecular Pharmacology, Albert Einstein College of Medicine), H358 (gift from Dr. Piro Lit, Department of Medicine, Memorial Sloan Kettering Cancer Center), and KPC-4622 (gift from Dr. Robert Vonderheide, Abramson Cancer Center, University of Pennsylvania). Cells were cultured according to ATCC or provider recommendations. MiaPaca-2 cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS), 2.5% horse serum, and 100 μg/mL Penicillin Streptomycin. Panc10.05 cells were cultured in RPMI 1640 supplemented with 15% FBS and 100 μg/mL Penicillin Streptomycin. The PANC-1, 293T, SW837, SW1463, and KPC cells were cultured in DMEM supplemented with 10% FBS and 100 μg/mL Penicillin Streptomycin. H358 and AsPC-1 cells All cells were regularly tested for Mycoplasma contamination using PCR-based detection and were confirmed negative. No sex or gender-based variables are applicable to cell line-based experiments. Human pancreatic cancer organoids Human pancreatic cancer organoids HPO-12 and HPO-46 were obtained as a generous gift from Dr. Nicolas Lecomte (Memorial Sloan Kettering Cancer Center). Organoids were derived from deidentified human pancreatic tumor specimens. Donor demographic information (including sex, age, race and ethnicity) was not available. Organoids were cultured following the Muthuswamy laboratory protocol in DMEM/F-12 (Gibco 11320033) supplemented with 500 μg/mL Anti-Anti (Gibco 15240-052), 10 mM 2-[4-(2-hydroxyethyl)piperazin-1-yl] ethanesulfonic acid (HEPES) (Gibco 15630-080), 2 mM Glutamax (Gibco 35050-061), 0.5 mM A83-01 (Tocris 9001799-5), 50 ng/mL human EGF (PeproTech AF-100-15), 100 ng/mL human FGF10 (PeproTech 100-26), 100 ng/mL human noggin (PeproTech 120-10C), 10 nM human gastrin-1 (Sigma Aldrich G9020), 1.25 mM N-acetylcysteine (Sigma Aldrich A0737), 10 nM nicotinamide (Sigma Aldrich N0636), 1× B27 supplement (Gibco 17504044), 10.5 μM Rho Kinase inhibitor (Y-27632) (Tocris 1254), and R-Spondin1 conditioned media 10% v/v from a R-spondin1 expressing 293T cell line (Sigma SLC111). Organoids were split once a week in a 2D culture onto a growth factor reduced Matrigel (Corning 354320) coated 12 well plates (Fisher FB012928). All human derived materials were obtained under appropriate institutional oversight at originating institution. No additional IRB approval was required for this study at the receiving institution due to de identification and prior approval status of derived materials. Tumor xenograft model All animal procedures were approved by the Albert Einstein College of Medicine IACUC in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals (Protocol Number: 00001545). Female J: NU mice (8 weeks old) were subcutaneously injected in the right flank with 10 × 10 ^6 MiaPaca-2 GFP+ luciferase cells, resuspended in 200 μL of 1 1:1 mixture of PBS and Matrigel. Tumor volume was measured using digital calipers. Once the tumors reached 100 mm3, mice were randomized to either the RMC-7977 treatment or the vehicle control group. RMC-7977 (Revolution Medicine) was prepared in a solvent mixture containing 10% DMSO, 40% PEG300, 5% Tween80, and 45% water for injection. Mice in the treatment group received 10 mg/kg RMC-7977 via oral gavage daily for a week. Tumor size was monitored every other day, and the experiment was concluded when the tumors exceeded 500 mm3 in volume. Only female mice were used, therefore, sex based biological variables could not be accessed. METHOD DETAILS Drugs and IC50 assays For ribosomal footprinting, Cycloheximide (Sigma Aldrich C4859) and Harringtonine (LKT Labs H0169) were used to arrest mRNA translation and map start site usage, respectively. Additionally, various KRAS inhibitors were utilized throughout this study. The KRAS G12C-specific inhibitor, Sotorasib, was purchased from MedChemExpress (HY-114277). The KRAS G12D specific inhibitor, MRTX1133, was purchased from MedChemExpress (HY-134813). RMC-7977, a potent multi-selective RAS(ON) inhibitor that targets the GTP-bound form of RAS, was provided by Revolution Medicines, Inc., [under a material transfer agreement]. Anisomycin was purchased from Sigma (A9789). IC50 assays were first performed utilizing the various KRAS inhibitors across different mutant KRAS-expressing cell lines. 10,000 cells per well were seeded in clear bottom 96-well plates (Invitrogen 165306) and were treated with each respective KRAS inhibitor for 72 h. ATP production was measured from viable cells using the CellTiter-Glo Luminescent Cell Viability Assay following the manufacturer’s instructions (Promega G7571). To investigate the potential effect of co-targeting mutant KRAS, EIF4A, and mTOR, PDAC cell lines were treated with increasing concentrations of mutant KRAS inhibitors (Sotorasib and MRTX1133) alone or in combination with EIF4A inhibitor CR-1-31B or mTOR inhibitor Torin1. Cell viability was assessed using CellTiter Glo 2.0 Luminescent Cell Viability Assay (Promega G9241) according to the manufacturer’s instructions. Dose-response heatmaps were generated for single and combination treatments to evaluate additive effects on cell viability across a panel of KRAS-mutant PDAC cell lines. To quantitatively assess drug interactions, we utilized SynergyFinder+, a web application for synergy analysis, and calculated the ZIP synergy score.54,55 Ribosome footprinting MiaPaca-2 and Panc10.05 pancreatic cancer cells were each treated with either DMSO or KRAS inhibitors, Sotorasib, and MRTX1133, respectively (10 μM, 2 h). A duplicate set of samples was treated with Harringtonine to map the translational start site of nascent translating mRNAs. Total RNA and ribosome-protected RNA fragments were isolated following a previously published protocol, and deep sequencing libraries were created from fragments using unique adapter sequences and subsequently sequenced on a HiSeq-2000 platform performed at the Genomics Core Facility, Albert Einstein College of Medicine.31 Sequence alignment Total RNA-sequencing (RNA-seq) and ribosome sequencing (Ribo-seq) reads were first adapter trimmed using TrimGalore.42,43 Trimmed reads that aligned to a house rRNA panel from 45S rRNA and rRNA from different Mycoplasma species were filtered out. Unfiltered reads were aligned to the GRCh38 (ENSEMBL build 105) human reference genome and transcriptome using an STAR aligner.44 Sequencing RNA-seq and Ribo-seq reads that mapped to CDS regions were extracted using a shell script with SAMtools.45 Unique Ribo-seq reads between 15 and 30nt (MiaPaca-2) or 25-35nt (Panc10.05) were defined as ribosome footprints (RF) and were used for further analysis. As an internal control, gene expression profiles between sample replicates were compared. Briefly, for both total RNA-seq and Ribo-seq datasets, an expression matrix was tabulated using the read counts for each gene and subsequently was used to compute Spearman r coefficients for the log10(n+1) transformed values. Translation efficiency analysis using RiboDiff Gene expression profiles were created from RSEM and were used to estimate the log2 fold change (log2 FC) of total RNA levels.46 To calculate changes in translational efficiency (TE), uniquely aligned reads from RNA-seq and Ribo-seq footprint counts were tabulated across the transcriptome and were used as the input counts matrix for RiboDiff.32 For both RNA-seq and Ribo-seq, the log2 FC of Drug Treated/DMSO was calculated using DESeq247 and RiboDiff, respectively. Genes with a significant difference in TE were defined using the adjusted p-value significance thresholds of 0.0001 and 0.1 for MiaPaca-2 treated with Sotorasib and Panc10.05 treated with MRTX1133, respectively. Metagene analysis of ribosome coverage Metagene plots were created to visualize large-scale changes in ribosome coverage. For a given gene or gene list, we use the metagene2 package to calculate the ribosome coverage for 10 bins each for the 5′ UTR and 3′ UTR regions and 20 bins for the CDS regions.49 Coverage is normalized such that the maximum value in a drug and DMSO-treated pair is 1. P-site metagene analysis of harringtonine footprints Harringtonine-treated MiaPaca-2 and Panc10.05 samples were used to identify translation start sites from ribosome coverage peaks. Read counts for each base were normalized by the respective library sizes of each replicate and summed, which is reported as normalized read coverage with the units of counts-per-million. P-site offsets for each read length were determined using the ORF-RATER algorithm incorporating maximum peak height and triplet periodicity to select the optimal offset.48 We use the same offset settings for their corresponding parent group to avoid artifacts that may arise from differences in library size and read counts. The significance of the difference in the read coverage of local maxima is determined by unpaired one-sided Welch’s test. Fraction of pause-site codons (RxK, KKK, PP, PD, and PG) across all of the detected ORFs for the first 50 codons.56 Polysome profiling and protein isolation from sucrose gradient fractions MiaPaca-2 cells were cultured to ~80% confluency and treated with Sotorasib (10 μM) for 2 h, followed by the addition of cycloheximide (CHX, final concentration 100 μg/mL) for 10 min to arrest ribosomes on mRNA. Cells were then washed once with PBS containing 100 μg/mL CHX, scraped, and collected into a 15 mL conical tube. Cell pellets were obtained by centrifugation at 1000 rpm for 5 min at 4°C and lysed in polysome extraction buffer (20mM Tris-HCl, 100 mM KCl, 5mM MgCl2, 0.5% Nonidet P-40) supplemented with protease inhibitors, RNase inhibitor, and 100 μg/mL CHX. Lysates containing equal amounts of total RNA were loaded onto a 10–50% sucrose gradient and centrifuged at 40,000 rpm for 3 h at 4°C using a Beckmann Coulter SW41 rotor. For RNAse I digested profiles, lysates containing equal total RNA were treated with RNAse A (Thermo Fisher Scientific, AM2294; 4 mg/L) for 45 min at room temperature, followed by quenching with 200 U of SUPERaseIn RNase inhibitor (Thermo Fisher Scientific, AM2694) before sucrose gradient centrifugation under identical conditions (gradients prepared in 200 mM Tris-HCl, 1 M NaCl, 50mM MgCl2, and 1mg/mL Heparin). Following ultracentrifugation, gradients were fractionated using a UA-6 UV/VIS Detector (ISCO 23252), and A260 absorbance was recorded to generate polysome profiles. Proteins were isolated from sucrose gradient fractions using TRIzol reagent (Thermo Fisher Scientific 15596026) according to the manufacturer’s protocol. Briefly, protein fractions were precipitated from the phenol-ethanol phase with isopropanol, washed twice with 0.3 M guanidine hydrochloride in 95% ethanol, and once with 100% ethanol. The protein pellets were air-dried, solubilized in 2× Laemmli dye at 95°C for 5 min before immunoblotting. Immunoblots Cells were harvested and washed in PBS (Invitrogen 10010023) before lysis in TNN lysis buffer (50 mM Tris-Cl, 250 mM NaCl, 5 mM EDTA, 0.5% NP-40, supplemented with protease inhibitor in nuclease-free water). After lysis, protein concentrations were quantified using the Bradford Protein Assay Kit (Genesee Scientific 18–443). 50 μg of protein lysate was prepared in 4× Laemmli buffer (Bio-Rad 1610747) and denatured for 8 min at 95°C before loading onto an SDS-PAGE gel (Invitrogen NW04120BOX). Separated proteins were transferred onto a nitrocellulose membrane (Invitrogen 88018) and incubated in a blocking buffer (Licor 927–80001) for 30 min before staining with the appropriate antibodies. The following primary antibodies were used as per the manufacturer’s recommended dilution: ZNF598 (Bethyl A305-108A), EDF1 (ProteinTech 12419-1-AP), ZAKα (Bethyl A301-993A), RPS6 (Cell Signaling Technology 2317), p-S6 (Cell Signaling Technology 2215), EIF4E (Cell Signaling Technology 9742), EEF1A1 (ProteinTech 11402-1-AP), EEF2 (Cell Signaling Technology 2332), p-ERK (Cell Signaling Technology 9101), ERK (Cell Signaling Technology 4695), RPL5 (Cell Signaling Technology 51345S), RPS10 (Abcam ab151550 and Invitrogen PA596460), DLAT (Cell Signaling Technology 12362), PDHB (Invitrogen MA5-32757), ECHS1 (ProteinTech 66117-1-LG), ACADM (ProteinTech 55210-1-AP), β-actin (Sigma-Aldrich A5316). Membrane blots were incubated at 4°C with gentle rocking overnight, washed, and incubated with the respective secondary antibodies at room temperature for 1 h (Licor 926–68071, 926–32210). The Licor Odyssey and Image Studio applications were used for the development and quantification of blots, respectively. Puromycin labeling to measure global protein synthesis Nascent protein synthesis from MiaPaca-2, H358, Panc10.05, and PANC-1 cells following KRAS inhibitors was assessed using Cy5-conjugated Puromycin labeling (synthesized at Memorial Sloan Kettering Cancer Center Organic Synthesis Core Facility). Before collection, cells were incubated in Cy-5 Puromycin at a final concentration of 10 μM for 1 h at 37°C. Cells were then washed and filtered through a 0.45 μM strainer. Single-cell suspensions were analyzed for changes in median fluorescent intensity using flow cytometric analysis on the Attune Nxt Flow Cytometer. Flow cytometric assays Organoid and cell lines were washed in FACS buffer (10% FBS in PBS) and resuspended in round-bottom 96-well plates (Greiner 650-185) for intracellular flow cytometric-based staining. Briefly, cells were first fixed and permeabilized using the BD Fixation/Permeabilization Kit (BD 554714) as per the manufacturer’s instructions. EEF1A1 intracellular staining was subsequently performed by incubating in a primary (ProteinTech 11402-1-AP) and secondary (Invitrogen A11012) antibody mixture at 4°C overnight. Cells were then washed and resuspended in FACS buffer before analyses of median fluorescent intensity on the Attune Nxt Autosampler. STRING and GSEA analysis STRING analysis was performed on different subsets of TE targets to identify enriched protein-protein interactions.52 GSEA analysis was performed using Hallmark and KEGG gene sets using the Molecular Signature Database (MSigDB) web server.53 Real-time quantitative PCR Cells were washed in PBS (Invitrogen 10010023) and harvested in QIAzol Lysis Reagent (Qiagen 79306) prior to RNA extraction using the miRNeasy micro kit (Qiagen 217004) as per manufacturer instructions. Total RNA was quantified on a Nanodrop spectrophotometer. cDNA synthesis was performed with 2 μg of total RNA using the SuperScript III First-Strand Synthesis kit (Invitrogen 1808051). RT-qPCR was performed on a QuantStudio 6 Flex using the TaqMan Universal Master Mix II (Invitrogen 444040) and the following TaqMan Gene Expression Assay probes: human EEF1A1 Hs05015684_g1, human EEF2 Hs00157330_m1, human RPS6 Hs04195024_g1, human EIF4E Hs00854166_g1, human ACADM Hs00936584_m1, human DLAT Hs00898876_m1, human ECHS1 Hs00187943_m1, human PDHB Hs00168650_m1, human ACTB Hs01060665_g1, and human GAPDH Hs02786624_g1. ΔΔCt values were used for further analysis. Generation of isogenic mutant KRAS expressing HEK293 cells Isogenic HEK293 cells expressing mutant KRAS were generated using viral transduction. Briefly, plasmids pBabe-KRAS G12C (Addgene Plasmid #58901) and pLEX-HA-BirA*-KRAS G12D-IRES-Puro (Addgene Plasmid #120562) were transfected into Phoenix A and HEK293 cells, respectively, using Fugene transfection reagent according to the manufacturer’s instructions. Fresh viral supernatant was collected 48–72 h post transfection. HEK293 cells were then transduced with the viral supernatants in the presence of polybrene (8 μg/mL) to enhance infection efficiency. After 48 h, transduced cells were selected with puromycin (10 μg/mL) for another 48 h. Following selection, KRAS genotype and expression were confirmed by western blot analysis using anti- HA antibody. Motif enrichment and structure analysis of 5′ UTRs 5′ UTR sequences were selected from the ENSMBL MANE canonical protein-coding transcripts for Sotorasib and MRTX1133 affected genes. Motif enrichment was then performed using the XSTREME pipeline in MEME-suite 5.5.4.50,57 The background sequences were selected from the pool of actively translated genes that were unaffected by drug treatment by setting minimum significance and untreated TE thresholds. For Sotorasib-affected genes, the significance and TE thresholds were 0.5 and 4.0, respectively, for a total of 2,758 genes. For MRTX1133-affected genes, the significance and TE thresholds were 0.1 and 0.5, respectively, for a total of 1,236 genes. Motifs between 8 nt and 18 nt that were statistically enriched in the TE up, or TE down genes relative to the background sequences were selected for downstream analysis. The RNAfold command in the Vienna RNA package was used to predict RNA secondary structure and folding energy.51 EEF1A1 5′ UTR cloning We utilized Applied Biological Materials Inc. to clone the full-length and mutated 5′UTR of the EEF1A1 transcript (EEF1A1-201; ENST00000309268) that is conserved among three other protein-coding EEF1A1 isoforms: EEF1A1-211, EEF1A1-220, and EEF1A1-222. The expression of ENST00000309268 in our cells was confirmed by RNA-seq. Next, 5,000,000 EEF1A1 5′ UTR variants were randomly generated using a 10% single-point mutation rate (9 nt). Their secondary structures were predicted, as reported earlier.51 The variant with the fewest predicted base pairs and weakest folding energy (−5 kcal/mol in variant vs. −14.4 kcal/mol) was selected for experimental validation. In this variant, all 9 positions were mutated to adenine. The full-length 5′ UTR and mutated 5′ UTR of EEF1A1 were then cloned onto a luciferase expressing, puromycin-resistant lentiviral reporter plasmid. Full-length sequence of 5′ UTR of EEF1A1 5′CTTTTTCGCAACGGGTTTGCCGCCAGAACACAGGTGTCGTGAAAACTACCCCTAAAAGCCAAAATGGGAAAGGAAAAGACTCATATCAAC. Mutated sequence of 5′ UTR of EEF1A1: Mutated positions are underlined 5′CTTTATCGCAACGGGTATGCCGCCAGAACACAGATATCGTGAAAACTACCACTAAAAGCCAAAATAGAAAAGGAAAAGACACATAACAAC. Luciferase assay We utilized previously established dual luciferase reporter constructs comprising cap-firefly luciferase and IRES-renilla luciferase to measure changes in both cap-dependent and IRES-mediated translation. MiaPaca-2 and Panc10.05 cell lines were transfected with respective luciferase plasmids and treated with Sotorasib or MRTX1133, respectively, for 24 h. To assess translation driven by specific 5′ UTR elements, we employed luciferase reporter constructs containing oligopyrimidine (TOP), G-quadruplex (GQ), and GC-rich motifs upstream of the firefly luciferase gene. Following treatment, luciferase activity was measured from equal amounts of protein lysates using the Dual-Luciferase Reporter Assay System (Promega E1960) following the manufacturer’s instructions. An empty firefly luciferase plasmid, pGL4.13, or an empty renilla luciferase plasmid, pGL4.73, was used as an internal control for normalization. For EEF1A1 5′ UTR luciferase reporter assay, we generated stable PDAC cells expressing either full-length, mutated, or blank luciferase reporter plasmid. Luciferase activity was measured following indicated drug treatment using the Luciferase Reporter Assay System (Promega E1501), following the manufacturer’s instructions. QUANTIFICATION AND STATISTICAL ANALYSIS Statistical analysis Statistical analysis was performed using Prism (v10.0, GraphPad) and R (v4.2.1).58,59 For all metagene and P-site metagene analyses between untreated and treated samples, a two-sided unpaired Welch’s t test was performed in R. All correlation analyses with significance testing were performed in R. For Cy5 Puromycin labeling and all other flow-based analyses, two-sided unpaired T-tests were performed on GraphPad Prism to compare changes in MFI between untreated samples and samples treated at different time points. Motif enrichment between differentially translated groups was performed using a hypergeometric test. Candidate motifs were selected based on the criteria that the ratio of true positive over false positive was greater than 1 and p < 0.05. 5′ UTR length distributions and folding energy differences were compared using a two-sided unpaired Welch’s t test in R59. Luciferase activity differences utilizing various reporter plasmids were statistically compared to untreated samples using a two-sided unpaired Student’s t test. For all significance testing, statistical significance is defined as p < 0.05. When appropriate, p-values were adjusted according to the Benjamini-Hochberg procedure. Supplementary Material Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2026.117520. KEY RESOURCES TABLEREAGENT or RESOURCESOURCEIDENTIFIERAntibodiesZNF598BethylCat# A305-108A; RRID: AB_2631503EDF1ProteinTechCat# 12419-1-AP; RRID: AB_2230895ZAKαBethylCat# A301-993A; RRID: AB_1576612RPS6Cell Signaling TechnologyCat# 2317; RRID: AB_2238583p-S6Cell Signaling TechnologyCat# 2215; RRID: AB_331682EIF4ECell Signaling TechnologyCat# 9742; RRID: AB_823488EEF1A1ProteinTechCat# 11402-1-AP; RRID: AB_2096966EEF2Cell Signaling TechnologyCat# 2332; RRID: AB_10693546p-ERKCell Signaling TechnologyCat# 9101; RRID: AB_331646ERKCell Signaling TechnologyCat# 4695; RRID: AB_390779RPL5Cell Signaling TechnologyCat# 51345; RRID: AB_2799391RPS10AbcamCat# ab151550; RRID: AB_2714147RPS10InvitrogenCat# PA596460DLATCell Signaling TechnologyCat# 12362; RRID: AB_2797893PDHBInvitrogenCat# MA5-32757ECHS1ProteinTechCat# 66117-1-IG; RRID: AB_2881516ACADMProteinTechCat# 55210-1-AP; RRID: AB_10837361β-actinSigma-AldrichCat# A5316; RRID: AB_476743IRDye 680RD Goat anti-Rabbit IgG Secondary AntibodyLI-CORCat# 926-68071IRDye 800CW Goat anti-Mouse IgG Secondary AntibodyLI-CORCat# 926-32210Cy-5 Conjugated PuromycinMemorial Sloan Kettering Cancer Center, Organic Synthesis CoreN/AGoat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 594InvitrogenCat# A11012Bacterial and virus strainsStable Competent E. coli (High Efficiency)NEBCat# C3040HBiological samplesHuman PDAC Organoids HPO-12Dr. Nicolas Lecomte, Dr. Jerry Melchor, Dr. Christine A. Iacobuzio-DonahueN/AHuman PDAC Organoids HPO-46Dr. Nicolas Lecomte, Dr. Jerry Melchor, Dr. Christine A. Iacobuzio-DonahueN/AChemicals, peptides, and recombinant proteinsTGFB (A83-01)TocrisCat# 9001799-5Human EGFPeproTechCat# AF-100-15Human FGF10PeproTechCat# 100-26Human NogginPeproTechCat# 120-10CHuman Gastrin-1Sigma-AldrichCat# G9020N-acetylcysteineSigma-AldrichCat# A0737NicotinamideSigma-AldrichCat# N0636Rho Kinase Inhibitor (Y-27632)TocrisCat# 1254CycloheximideSigma-AldrichCat# C4859HarringtonineLKT LabsCat# H0169SotorasibMedChemExpressCat# HY-114277MRTX1133MedChemExpressCat# HY-134813RMC-7977Wasko et al.20N/AAnisomycinSigma-AldrichCat# A9789Protease InhibitorSigma-AldrichCat# 11836170001RNase I InhibitorThermo Fisher ScientificCat# AM2294SUPERaseIn RNase inhibitorThermo Fisher ScientificCat# AM2694HeparinSigma-AldrichCat# H3149Guanidine HydrochlorideSigma-AldrichCat# G3272CR-1-31BMedChemExpressCat# HY-136453Torin1MedChemExpressCat# HY-13003Critical commercial assaysCellTiter-Glo Luminescent Cell Viability KitPromegaCat# G7571CellTiter-Glo 2.0 Luminescent Cell Viability KitPromegaCat# G9241Superscript II First-Strand Synthesis KitInvitrogenCat# 1808-051Taqman Universal Master Mix IIInvitrogenCat# 444040Dual Luciferase Reporter AssayPromegaCat# E1960Deposited dataIllumina sequencing: Ribosome footprinting and total RNAThis paperGEO: GSE285857Experimental models: Cell linesMiaPaca-2ATCCCat# CRL-1420Panc10.05ATCCCat# CRL-2547PANC-1ATCCCat# CRL-1469ASPC-1ATCCCat# CRL-1682SW837Dr. Chaoyuan KuangN/ASW1463Dr. Chaoyuan KuangN/AH358Dr. Piro LitoN/AKPC-4622Dr. Robert VonderheideN/AR-spondin1 expressing 293TSigma AldrichCat# SLC111Phoenix Amphotropic cellsATCCCat# CRL-3213HEK293TATCCCat# CRL-3216Experimental models: Organisms/strainsJ:NUJackson LaboratoriesCat# 002019OligonucleotidesEEF1A1 Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs05015684_g1EEF2 Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs00157330_m1,RPS6 Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs04195024_g1EIF4E Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs00854166_g1ACADM Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs00936584_m1DLAT Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs00898876_m1ECSH1 Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs00187943_m1PDHB Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs00168650_m1,ACTB Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs01060665_g1GAPDH Taqman Gene Expression AssayThermo Fisher ScientificCat# Hs02786624_g1Recombinant DNApBabe-KRAS G12CAddgeneCat# 58901pLEX-HA-BirA*-KRAS G12D-IRES-PuroAddgeneCat# 120562Cap-FF-luciferase IRES-Renilla-LuciferaseWolfe et al.7N/ATOP-Luciferase reporter plasmidWolfe et al.7N/AGQ-Luciferase reporter plasmidWolfe et al.7N/AGC-rich luciferase reporter plasmidWolfe et al.7N/ApGL4.13PromegaCat# E6681pGL4.73PromegaCat# E6911EEF1A1 Full length 5′UTR luciferase reporterThis paperN/AEEF1A1 Truncated 5′UTR luciferase reporterThis paperN/ASoftware and algorithmsRR Core Team https://cran.r-project.org/ PrismGraphPad Software https://www.graphpad.com/ CutadaptMartin et al.42 https://github.com/marcelm/cutadapt TrimGaloreKrueger et al.43 https://github.com/felixkrueger/trimgalore STARDobin et al.44 https://github.com/alexdobin/STAR SAMtoolsLi et al.45 https://github.com/samtools/samtools RSEMLi et al.46 https://github.com/deweylab/RSEM DESeq2Love et al.47 https://github.com/thelovelab/DESeq2 RiboDiffZhong et al.32 https://github.com/ratschlab/ribodiff ORF-RATERFields et al.42,48 https://github.com/alexfields/ORF-RATER Metagene2Fournier et al.49 https://bioconductor.org/packages//release/bioc/html/metagene2.html MEME suite 5.5.4 (XSTREME)Bailey et al.50 https://meme-suite.org/meme/tools/xstreme RNAfoldLorenz et al.51 https://github.com/ViennaRNA/ViennaRNA STRINGSzklarczyk et al.52 https://string-db.org/ MSigDBSubramanian et al.53 https://www.gsea-msigdb.org/gsea/msigdb SynergyFinder+Yadav et al.,54 Zheng et al.55 https://www.synergyfinderplus.org/ OtherDMEMThermo Fisher ScientificCat# 11965-092RPMI 1640Thermo Fisher ScientificCat# 11875-093Fetal Bovine SerumThermo Fisher ScientificCat# A6708-01Horse SerumSigma-AldrichCat# H1270Penicillin StreptomycinThermo Fisher ScientificCat# 15140-122DMEM/F-12Thermo Fisher ScientificCat# 11320033HEPESThermo Fisher ScientificCat# 156300-080GlutamaxThermo Fisher ScientificCat# 35050-061Growth Factor Reduced MatrigelCorningCat# 354320SucroseIBI ScientificCat# 21D5001TRIzolThermo Fisher ScientificCat# 15596026Qiazol Lysis ReagentQiagenCat#79306miRNAeasy micro kitQiagenCat# 217004LI-COR Intercept TBS Blocking BufferLI-CORCat# 927-80001LI-COR Intercept TBS Antibody DiluentLI-CORCat# 927-65001BD Fixation/Permeabilization KitBDCat# 554714Fugene Transfection ReagentPromegaCat# E591A KEY RESOURCES TABLE Highlights Mutant KRAS regulates the translation of mRNAs with short, weakly structured 5′UTRs Mutant KRAS regulates the translation of mRNAs with short, weakly structured 5′UTRs KRAS G12C inhibition induces ribosome stalling during translation KRAS G12C inhibition induces ribosome stalling during translation KRAS, mTOR, and EIF4A pathways show redundant control of mRNA translation KRAS, mTOR, and EIF4A pathways show redundant control of mRNA translation Co-targeting KRAS with mTOR or EIF4A enhances translational repression Co-targeting KRAS with mTOR or EIF4A enhances translational repression ACKNOWLEDGMENTS KS is supported by a start-up grant from Montefiore Einstein Comprehensive Cancer Center (MECCC) and Albert Einstein College of Medicine, Cancer Center Support Grant 5P30CA013330-53, a Seed grant from the Hirshberg Foundation for Pancreatic Cancer, NIH CTSA Grant 1UM1TR004400 from the Institute of Clinical and Translational Research, Albert Einstein College of Medicine, and a Department of Defense Peer-Reviewed Cancer Research Idea Award. PMG is supported by the Pancreatic Cancer Detection Consortium, U01CA210240, NCI Cancer Center Support Grant, P30CA36727, and NCI Research Specialist, R50CA211462. We thank Dr. Piro Lito at Memorial Sloan Kettering Cancer Center for providing Sotorasib. We thank Dr. Mallika Singh at Revolution Medicines for providing the RAS(ON) multi-selective inhibitor RMC-7977 and for her feedback on this manuscript. We acknowledge the Cancer Center Support Grant (P30 CA013330) that supports the Epigenomics core facility (RRID: SCR_023284) used as a shared resource at MECCC. Footnotes DECLARATION OF INTERESTS The authors declare no competing financial interests. Data and code availability Ribosome footprinting and total mRNA sequencing raw and processed data generated in this study have been deposited in the NCBI Gene Expression Omnibus (GEO). Database: Gene Expression Omnibus (GEO): GSE285857. The data are publicly available as of the date of publication. Ribosome footprinting and total mRNA sequencing raw and processed data generated in this study have been deposited in the NCBI Gene Expression Omnibus (GEO). Database: Gene Expression Omnibus (GEO): GSE285857. The data are publicly available as of the date of publication. Code: This paper does not report original code. Code: This paper does not report original code. Other Items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Other Items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. REFERENCES Associated Data Supplementary Materials Data Availability Statement Ribosome footprinting and total mRNA sequencing raw and processed data generated in this study have been deposited in the NCBI Gene Expression Omnibus (GEO). Database: Gene Expression Omnibus (GEO): GSE285857. The data are publicly available as of the date of publication. Ribosome footprinting and total mRNA sequencing raw and processed data generated in this study have been deposited in the NCBI Gene Expression Omnibus (GEO). Database: Gene Expression Omnibus (GEO): GSE285857. The data are publicly available as of the date of publication. Code: This paper does not report original code. Code: This paper does not report original code. Other Items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Other Items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.