Tumour progression towards dedifferentiated cell clusters plays a critical role in intratumour heterogeneity and therapy resistance. While tumour microenvironmental stress has been implicated, the underlying mechanisms remain poorly defined. Using mouse models of lung adenocarcinoma, we demonstrate that activation of the integrated stress response (ISR)-marked by phosphorylation of eIF2 (p-eIF2) and ATF4 induction-drives tumour heterogeneity. ISR activation facilitates the emergence of high-plasticity, undifferentiated and pre-epithelial-to-mesenchymal transition clusters characterized by elevated ATF4 and MYC activity. This process is MYC dependent and involves ISR-mediated repression of NKX2-1, a key determinant of alveolar identity, and induction of CHCHD10, a regulator of mitochondrial integrity and metabolic fitness. Disruption of the p-eIF2-ATF4 axis induces mitochondrial dysfunction, limits dedifferentiation and suppresses tumour growth. In human lung adenocarcinoma, ISR-driven dedifferentiation correlates with advanced disease and poor prognosis, identifying the ISR as a central driver of lineage reprogramming and metabolic fitness in tumour progression.

mRNA translation and stability are tightly regulated and functionally linked through cis-acting sequence elements and trans-acting factors, including RNA-binding proteins (RBPs). Here, we report that two chordate-specific paralogous RBPs, ZC3H7A and ZC3H7B, preferentially bind the coding region (CDS) and 3' untranslated region (3' UTR) of A/U-rich mRNAs, particularly those with enrichment of A/U at their wobble sites (A/U3 codons). Upon binding to target mRNAs, ZC3H7A/B promote mRNA degradation through recruitment of the CCR4-NOT deadenylase complex. Furthermore, these proteins engage ribosomes lacking elongation factors and repress translation initiation via the GIGYF2/4EHP translation repressor complex. Depletion of ZC3H7A/B or 4EHP impairs the translational repression of A/U3-rich mRNAs. Together, these findings reveal a mechanism in higher eukaryotes that links A/U-rich sequence content within the CDS and 3' UTR to the coordinated post-transcriptional regulation of mRNA stability and translation.

Abstract Activating mutations in KRAS occur at high frequency in colorectal, lung, and pancreatic cancers, which together account for a substantial proportion of global cancer mortality. Mutant KRAS is constitutively biased toward the GTP-bound state, driving persistent proliferative signaling, but simultaneously imposes oncogenic stresses that threaten cellular homeostasis. To sustain transformation, KRAS-mutant cells engage adaptive stress-response mechanisms, many of which converge on translational control mediated by the eIF2-eIF2B axis. While eIF2B is classically known as a guanine nucleotide exchange factor (GEF) for eIF2 during translation initiation, its potential role in directly regulating oncogenic signaling pathways has remained unexplored. Here, we identify a non-canonical function of eIF2B as a direct activator of mutant KRAS signaling. We demonstrate that eIF2B forms a tripartite complex with SOS and mutant KRAS at the plasma membrane (PM), thereby enhancing KRAS activation and tumorigenic signaling. Biochemical assays and structural modeling support an interaction between the catalytic ε subunit of eIF2B and the allosteric Ras-binding site of SOS, stabilizing SOS in an active conformation. This interaction potentiates SOS-mediated GDP/GTP exchange on mutant KRAS and promotes KRAS nanoclustering at the PM. Importantly, eIF2B exhibits marked specificity for mutant KRAS4B, but not KRAS4A, HRAS, or NRAS. This selectivity arises from KRAS4B’s unique polybasic membrane-anchoring domain and from eIF2B-dependent remodeling of plasma membrane lipid composition. eIF2B enhances glycosphingolipid (GSL) biosynthesis, particularly GM3 and SM4, through translational upregulation of B4GALT5, generating a membrane environment that preferentially supports mutant KRAS4B anchoring and signaling. Disruption of GSL synthesis impairs formation of the eIF2B:SOS:KRAS complex and selectively reduces mutant KRAS activation. Notably, eIF2B’s stimulation of mutant KRAS signaling occurs independently of eIF2α phosphorylation, separating its translational stress-response function from its oncogenic signaling role. Functionally, eIF2B promotes tumor growth specifically in KRAS-mutant cancer models, including human xenografts and an autochthonous KRAS G12C-driven lung adenocarcinoma model. Clinically, high expression of eIF2Bε correlates with poorer outcomes in patients with KRAS-mutant tumors. Our findings identify eIF2B as a previously unrecognized regulator of mutant KRAS-driven tumorigenesis that links translational control, membrane lipid remodeling, and oncogenic signaling. By coordinating SOS activation, KRAS membrane nanoclustering, and selective translation, eIF2B emerges as a central modulator of KRAS oncogenic output and a potential therapeutic and prognostic target. In vivo targeting of eIF2B supports its use as a combinatorial strategy with KRAS inhibition to broaden the therapeutic window in KRAS-mutant cancers. This abstract was edited and refined with the assistance of generative artificial intelligence to improve clarity and conciseness. Citation Format: Hyungdong Kim, Shiqi Diao, Kwang-Jin Cho, Hyun-Ro Lee, Junchen Liu, Pascal Egea, Tatu Pantsar, Milla Kurki, Nour Ghaddar, Shuo Wang, Jia Yi Zou, Mehdi Amiri, Ritchel Gannaban, John F. Hancock, Kylie M. Rice, Atsuo Sasaki, John Asara, Brajendra Tripathi, Douglas Lowy, Rosalie Lawrence, Maria Hatzoglou, Carlos R. Azpilcueta-Nicolas, Jean-Philip Lumb, John Columbus, Thomas J. Turbyville, Christopher B. Marshall, Mitsuhiko Ikura, Jay T. Groves, Nahum Sonenberg, Peter Walter, Antonis E. Koromilas. eIF2B Selectively Anchors and Activates Mutant KRAS4B [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: RAS Oncogenesis and Therapeutics; 2026 Mar 5-8; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(5_Suppl_1):Abstract nr B001.

Reduced SERCA2a expression impairs cardiomyocyte contraction and relaxation, contributing to the progression of congestive heart failure (CHF). The translation of specific mRNAs, including SERCA2a, is suppressed by 4E-binding proteins (4E-BPs) through preventing the assembly of an intact initiation complex. Here, we investigated the role of 4E-BPs in regulating cardiac SERCA2a expression and transverse aortic constriction (TAC)-induced CHF in mice via genetic deletion of both 4E-BP1 and 4E-BP2 (4E-BP1/2 DKO), as well as through cardiac-specific knockdown or overexpression of 4E-BP1. 4E-BP1/2 DKO markedly alleviated TAC-induced CHF and mortality without affecting left ventricular hypertrophy. RNA-seq and mitochondrial respiratory analyses showed that 4E-BP1/2 DKO mitigated TAC-induced mitochondrial dysfunction and oxidative stress. Similar protective effects were observed with 4E-BP1 knockdown, whereas 4E-BP1 overexpression worsened these pathological changes. Mechanistically, 4E-BP1/2 DKO increased SERCA2a mRNA binding to eIF4G, thereby enhancing SERCA2a protein translation in the myocardium. Cardiac-specific Serca2a knockdown in TAC-challenged 4E-BP1/2 DKO mice reversed these protective effects. Consistent findings were obtained in a cardiomyocyte cell line. In summary, our results demonstrated that 4E-BP1/2 DKO translationally increased myocardial SERCA2a expression and rescued mice from CHF development after TAC, indicating that interventions reducing the signaling of cardiac 4E-BPs may be a novel therapeutic approach for treating CHF.

The integrated stress response (ISR) plays a crucial role in cognition via bidirectional modulation of the two major forms of synaptic plasticity, long-term potentiation, and long-term depression (LTD). Specifically, inhibition of the ISR blocks metabotropic glutamate receptor-dependent LTD (mGluR-LTD), whereas its activation facilitates this form of synaptic depression. However, the contribution of activating transcription factor 4 (ATF4), the best studied downstream effector of the ISR, to mGluR-LTD remains unknown. Here, we show that pharmacological activation of group I mGluRs in mouse hippocampal slices increases ATF4 protein levels without altering its transcription and concurrently downregulates the expression of oxidative phosphorylation (OXPHOS) proteins. Selective deletion of ATF4 in excitatory neurons impairs mGluR-LTD and prevents the downregulation of OXPHOS proteins. Notably, administration of a small molecule inhibitor of OXPHOS rescues the impaired mGluR-LTD in ATF4-depleted hippocampal slices, indicating that ATF4 regulates this type of synaptic plasticity by modulating mitochondrial function. Moreover, ATF4 deletion in excitatory neurons disrupts object-place learning, an mGluR-LTD-dependent behavior paradigm. Together, these findings reveal a role of ATF4 as a key mediator of protein synthesis-regulated synaptic depression and related behaviors.

Sensitization of spinal nociceptive circuits plays a crucial role in neuropathic pain. This sensitization depends on new gene expression that is primarily regulated via transcriptional and translational control mechanisms. The relative roles of these mechanisms in regulating gene expression in the clinically relevant chronic phase of neuropathic pain are not well understood. Here, we show that, in mice, changes in gene expression in the spinal cord during the chronic phase of neuropathic pain are substantially regulated at the translational level. Downregulating spinal translation at the chronic phase alleviated pain hypersensitivity. Cell type-specific profiling revealed that spinal inhibitory and excitatory neurons exhibited substantial changes in translation after peripheral nerve injury. Notably, increasing translation selectively in all inhibitory neurons or parvalbumin-positive (PV+) interneurons, but not excitatory neurons, promoted mechanical pain hypersensitivity. Furthermore, increasing translation in PV+ neurons decreased their intrinsic excitability and spiking activity. Conversely, reducing translation in spinal PV+ neurons prevented the nerve injury-induced decrease in excitability but did not alleviate mechanical hypersensitivity. Together, these findings advance our understanding of translational control mechanisms in the spinal cord during neuropathic pain and highlight their cell type- and phase-specific contributions to gene expression and pain hypersensitivity.

mRNA translational repression by eukaryotic initiation factor 4E-binding proteins (4E-BPs), plays a critical role in synaptic plasticity and the formation of long-term memory (LTM). Among the three 4E-BP paralogs, 4E-BP2 is the predominant form expressed in neurons, and its full-body deletion in mice causes memory deficits. Mice lacking 4E-BP2 in GABAergic inhibitory interneurons, but not excitatory neurons, display autistic-like behaviors and deficits in object location and recognition. The specific mRNAs translationally regulated by 4E-BP2 in GABAergic interneurons, and how they contribute to spatial and associative memory, are unknown. Here, we show that conditional knockout (cKO) mice lacking 4E-BP2 selectively in GABAergic interneurons exhibit impairments in long-term spatial and contextual fear memory formation. We further demonstrate that 4E-BP2 deletion controls the translation of selective mRNAs in interneurons without increasing general protein synthesis. One of the mRNAs is Gal, which encodes a neuropeptide that modulates memory. Our findings provide evidence that 4E-BP2 selectively controls the translation of a subset of mRNAs in inhibitory neurons that are required for LTM formation.

Local protein synthesis is a crucial process that maintains local proteostasis in neurons. A large percentage of mRNAs translated in developing neurons are associated with stalled ribosomes. FMRP, the protein lost in Fragile X syndrome, is highly enriched in RNA granules that contain stalled ribosomes. Previous examination of ribosome-protected fragments (RPFs) from stalled neuronal ribosomes identified sequences that match those found in mRNAs associated with FMRP. To investigate whether FMRP recognition of these sequences is important for determining where ribosomes stall on mRNAs, we examined RPFs isolated from P5 mice of both sexes that lack the FMRP protein. The loss of FMRP had no significant effect on the proteins associated with neuronal stalled ribosomes, on ribosome structure, or the stalling sites (locations where RPFs accumulated). There was a small, but significant decrease in the number of RPFs from mRNAs previously shown to be associated with FMRP by CLIP. Additionally, the number of neuronal RNA granules containing stalled ribosomes, as assayed by ribopuromycylation, decreased. These results suggest a role of FMRP in neuronal RNA granules that contain stalled ribosomes, though loss of FMRP does not influence where ribosomes are stalled or the formation of stalled ribosome.

The anaphase promoting complex/cyclosome (APC/C) is a large, ubiquitin ligase and a central regulator of cell cycle progression. By targeting key substrates for degradation during mitosis and G1 phase, the APC/C coordinates metabolic fluctuations that occur during the cell cycle. A diverse range of viruses have convergently evolved mechanisms to bind and inhibit the APC/C; however, a molecular understanding for these interactions has never been demonstrated. Here, we use chicken anemia virus (CAV), a small single-stranded DNA virus encoding only three proteins, to demonstrate the importance of viral APC/C inhibition during replication. We show that the Vp3 protein of CAV inhibits the APC/C, causing a dramatic mitotic arrest during infection. The mutant virus lacking Vp3 is defective for replication and can be rescued by APC/C inhibition. Metabolomic profiling during CAV infection revealed that Vp3 expression mediates a broad increase in nucleotide pools. Moreover, viral inhibition of the APC/C resulted in stabilization of enzymes required for nucleotide biosynthesis. These findings suggest that the APC/C is a general target of many viruses to elevate nucleotide levels and facilitate viral genome replication.

Elevated expression of components of the translation initiation complex (eIF4F) is frequent in cancer and results in enhanced synthesis of oncogenic proteins. Given its essentiality in normal tissues, targeting eIF4F is challenging. Here, combining chemical and in silico screens, we identified a small molecule (M19 and its analog M19-6) that targets the MA3 domain of the eIF4F subunit eIF4G1, interferes with eIF4F assembly and alleviates melanoma resistance to BRAF and MEK inhibitors. Ribosome profiling revealed that the M19-6 selectively perturbs the melanoma translatome, limiting synthesis of factors that promote cell proliferation and neoplastic growth. Screens in melanoma models revealed that M19-6 synergizes with autophagy or HDAC inhibitors in cell culture and potentiates anti-neoplastic and anti-metastatic effects of doxorubicin in vivo. Overall, we describe a novel eIF4F complex inhibitor that offers a new therapeutic modality to target clinically challenging melanomas and could provide a molecular basis for combination with currently employed therapies.

Supplementary Table S9: mRNAs affected by tomivosertib.

GIGYF2 (growth factor receptor–bound protein 10 [GRB10]-interacting GYF [glycine–tyrosine–phenylalanine] protein 2) reduces mRNA stability and translation via microRNAs, ribosome quality control, and several RNA-binding proteins. GIGYF2 was first identified in mouse cell lines as an interacting partner with GRB10, which binds to the insulin receptor and the insulin-like growth factor receptor 1. Mutations in the human GIGYF2 gene were reported in autism. In mouse models, Gigyf2 mutations engender several diseases. It was therefore thought that the GIGYF2-associated disease in humans is caused by defective GRB10 signaling. We show here that GIGYF2 does not interact with GRB10 in human cell lines, as determined by co-immunoprecipitation and proximity ligation assays. The lack of interaction is explained by the absence of the critical GYF domain–binding PPGΦ sequence in the human GRB10 protein. These results contrast with the current understanding that a GIGYF2/GRB10 complex is associated with human disease via insulin receptor and insulin-like growth factor receptor 1 signaling and underscore alternative mechanisms responsible for the observed phenotypes associated with mutations in the human GIGYF2 gene.

Amyloid beta oligomers (AβO) are pivotal in Alzheimer’s Disease (AD), cleared by microglia cells, as immune cells in the brain. Microglia cells exposed to AβO are involved with migration, apoptosis, phagocytosis, and activated microglial receptors through AβO, impacting cellular mechanobiological characteristics such as microglial adhesion strength to the underlying substrate. Herein, a label-free microfluidic device was used to detect advancing AD conditions with increasing AβO concentrations on microglia BV2 cells by quantitatively comparing the cell-substrate adhesion. The microfluidic device, acting as an AD model, comprises a single channel, which functions as a cell adhesion assay. To assess cell-substrate adhesion under different AβO concentrations of 1 µM, 2.5 µM, and 5 µM, the number of the cells attached to the substrate was counted by real-time microscopy when the cells were under the flow shear stress of 3 Pa and 7.5 Pa corresponding to Reynolds number (Re) of 10 and 25, respectively. The data showed that quantifying the cell-substrate adhesion using the microfluidic device could successfully identify conditions of advancing AβO concentrations. Our findings indicated that the increased incubation time with AβO caused reduced cell-substrate adhesion strength. Additionally, increased AβO concentration was another factor that weakened microglial interaction with the substrate. The quantification of cell-substrate adhesion using 3 Pa compared to 7.5 Pa clearly demonstrated advancing AβO in AD. This study using the chip provides an AD model for a deeper understanding mechanobiological behaviors of microglia exposed to AβO corresponding to diagnosed AD conditions under an in vitro microenvironment.

Supplementary Fig. S4: Proteomic profiling. Heatmap showing the Normalized Enrichment Score (NES) for the main independent pathways enriched in at least 2 of the patients in the cohort.

Alzheimer's disease (AD) is characterized by progressive memory decline. Converging evidence indicates that hippocampal mRNA translation (protein synthesis) is defective in AD. Here, we show that genetic reduction of the translational repressors, Fragile X messenger ribonucleoprotein (FMRP) or eukaryotic initiation factor 4E (eIF4E)-binding protein 2 (4E-BP2), prevented the attenuation of hippocampal protein synthesis and memory impairment induced by AD-linked amyloid-β oligomers (AβOs) in mice. Moreover, genetic reduction of 4E-BP2 rescued memory deficits in aged APPswe/PS1dE9 (APP/PS1) transgenic mouse model of AD. Our findings demonstrate that strategies targeting repressors of mRNA translation correct hippocampal protein synthesis and memory deficits in AD models. Results suggest that modulating pathways controlling brain mRNA translation may confer memory benefits in AD.

Dual reporters encoding two distinct proteins within the same mRNA have had a crucial role in identifying and characterizing unconventional mechanisms of eukaryotic translation. These mechanisms include initiation via internal ribosomal entry sites (IRESs), ribosomal frameshifting, stop codon readthrough and reinitiation. This design enables the expression of one reporter to be influenced by the specific mechanism under investigation, while the other reporter serves as an internal control. However, challenges arise when intervening test sequences are placed between these two reporters. Such sequences can inadvertently impact the expression or function of either reporter, independent of translation-related changes, potentially biasing the results. These effects may occur due to cryptic regulatory elements inducing or affecting transcription initiation, splicing, polyadenylation and antisense transcription as well as unpredictable effects of the translated test sequences on the stability and activity of the reporters. Unfortunately, these unintended effects may lead to misinterpretation of data and the publication of incorrect conclusions in the scientific literature. To address this issue and to assist the scientific community in accurately interpreting dual-reporter experiments, we have developed comprehensive guidelines. These guidelines cover experimental design, interpretation and the minimal requirements for reporting results. They are designed to aid researchers conducting these experiments as well as reviewers, editors and other investigators who seek to evaluate published data.

Supplementary Fig. S9: Imaging mass cytometry. (A) Heatmap of signaling markers expression within cell clusters and (B) Neighbourhood analysis showing cell-to-cell interactions over all patients before and after treatment with tomivosertib.

Supplementary Fig. S5: Translatomic profiling. Scatter plots demonstrating translation efficiency (TE) (top), Ribo-Seq (middle), RNA-Seq) (bottom) the fold change difference between ON MNKi and Pre-MNKi conditions for three different patients, patient 13 (A), patient 12 (B), and patient 16 (C). Only reads aligning with CDS were considered for this analysis. The x-axis represents the normalized number of reads corresponding to the experiment/condition of minimal expression. Red and blue dots denote significantly up or down-regulated mRNAs, respectively (Z-score > 2 and log2 FC > 0.5 for up and Z-score < -2 and log2 FC < -0.5 for down).