Transcriptional induction of 4E-BP3 prolongs translation repression

The mammalian target of rapamycin (mTOR) is a multifaceted serine/threonine kinase implicated in a large number of physiological processes and pathological states including cancer.¹ mTOR forms 2 distinct complexes, mTOR complex 1 (mTORC1) and 2 (mTORC2), which differ in their composition, downstream targets, regulation, and sensitivity to the mTOR allosteric inhibitor rapamycin. mTORC1 is frequently hyperactivated in cancers due to aberrant upstream mutations in oncogenes and tumor suppressors (e.g. PI3K, TSC2, PTEN). Hyperactivation of this signaling pathway is central to cancer cell proliferation and survival, thus there is considerable interest in the development of therapeutic strategies targeting mTORC1 with pharmacological inhibitors.¹

One critical role of mTORC1 is to regulate translation, in part, by phosphorylating (inactivating) the eukaryotic translation initiation factor 4E (eIF4E)-binding proteins (4E-BPs). In mammals 3 4E-BPs exist (annotated as 4E-BP1, 4E-BP2, and 4E-BP3), each functioning as a translational repressor.^2,3 Under conditions of poor nutrition or pharmacological inhibition of mTORC1, 4E-BPs become hypophosphorylated (activated) and bind to the 5′-cap-binding protein eIF4E with high affinity. The binding of 4E-BPs to eIF4E prevents eIF4G association and eIF4F complex assembly, thereby limiting initiation of cap-dependent translation. Upon stimulation of mTORC1 by nutrients, insulin or growth factors, 4E-BPs are then hyperphosphorylated, resulting in their release from eIF4E and favoring of cap-dependent translation.¹ While these proteins have apparent identical molecular functions, 4E-BPs share approximately 60%– protein identity and have been reported to have different tissue distributions and expression, with publications showing 4E-BP1 highly present in skeletal muscle, pancreas, and adipose tissues, 4E-BP2 ubiquitously expressed but predominant in the brain and in lymphocyte cells, and 4E-BP3 expressed at lower levels but found to be present in most tissues.^2-5 4E-BP1 and 4E-BP2 are also frequently overexpressed and hyperphosphorylated in various tumors, and in particular, 4E-BP1 expression and phosphorylation are used as surrogate markers to predict patient outcome in several cancers.¹

Perhaps due to the availability of commercially available reagents, most studies to date have focused on the regulatory mechanisms and functions of 4E-BP1 and 4E-BP2 rather than 4E-BP3. We previously found that depletion of 4E-BP1 and 2 renders cells resistant to anti-tumor effects of mTOR inhibitors, due to sustained translation of eIF4E-sensitive mRNAs encoding pro-proliferative proteins.^6,7 While these studies were performed in transformed mouse embryonic fibroblasts with limited to no protein expression of 4E-BP3⁷, our latest study ² demonstrates that in several human cancer cell lines, 4E-BP3 can become strongly induced both transcriptionally and at protein levels upon prolonged mTORC1 inhibition. In addition to these in vitro experiments, we found that 4E-BP3 is transcriptionally increased in tumors of a chemically-induced liver cancer mouse model chronically treated with mTOR-targeting agents, and importantly, that 4E-BP3 mRNA expression in human breast cancer patients inversely correlates with activation of the mTORC1 pathway. These results indicate that 4E-BP3 is normally expressed at low levels in cells but is considerably increased after long-term mTORC1 inhibition. This is in sharp contrast to 4E-BP1 and 4E-BP2 proteins, which have been found to decrease in expression or be degraded during prolonged mTOR-inhibitory drug treatments.^2,6

Our recent studies show that the induction of 4E-BP3 is mediated by the MiTF (Microphthalmia-associated transcription factor) family transcription factor TFE3, which is activated upon mTORC1 inhibition (Fig. 1).² We found that the EIF4EBP3 promoter contains cis-elements for TFE3 in the immediate upstream region of the transcription start site and the mutated cis-elements impair the promoter activation in response to mTORC1 inhibition. Furthermore, knockdown of TFE3 was shown to suppress the induction of 4E-BP3 mRNA as well as other target genes. To address whether 4E-BP3 acts as an effector downstream of mTORC1, we generated human 4E-BP3 knockout cancer cells using CRISPR-Cas9. Ablation of 4E-BP3 rendered cancer cells more resistant to mTOR inhibitors, as prolonged translation repression of eIF4E-target mRNAs was impaired. Thus, when compared to 4E-BP1 or 4E-BP2, which contribute to limit eIF4E activity in the immediate to early phases of mTORC1 inhibition, succeeding transcriptional induction was shown to render 4E-BP3 an effective translation repressor when mTORC1 remains inhibited for long periods. Such multi-phase regulations of 4E-BPs may be required for rigorous control of target mRNA translation and of the protein synthesis machinery during conditions of poor nutrient bioavailability and of prolonged mTORC1 inhibition (Fig. 1).

Figure 1.

Model for translation repression during prolonged mTORC1 inhibition. In response to mTORC1 inhibitors, hypophosphorylated 4E-BP1 and 4E-BP2 rapidly suppress cap-dependent translation while TFE3 activates transcription of 4E-BP3 through binding to cis-elements in the EIF4EBP3 promoter. 4E-BP3 then becomes an important translation repressor at late phases of mTORC1 inhibition in proportion to its expression. This multi-staged successive regulation of 4E-BPs may represent an important regulatory function to rigorously control translational subsets of mRNAs during long-term mTORC1 inhibition.

Highly specific inhibitors of mTORC1, rapamycin and its analogs (rapalogs), are in the clinic for the treatment of advanced renal cell carcinoma, pancreatic neuroendocrine tumors, advanced breast cancers and mantle cell lymphomas. Active-site mTOR inhibitors also show promise as targeted anti-cancer therapies. Our recent findings establish 4E-BP3 as an important effector downstream of mTORC1 under prolonged inhibition, in a unique mechanism that differs from that of 4E-BP1 and 2. Interestingly, we observed that in a number of cancer cell lines, 4E-BP3 expression was not induced but rather suppressed due to gene silencing by DNA methylation.² Given that 4E-BP3 protein levels inversely reflected the activation status of mTORC1 in vitro, and in vivo from databases of cancer patients and targeted therapies in mouse models of cancer, the presence of 4E-BP3 may be a useful biomarker to predict tumor malignancy and therapeutic response to prolonged treatment with mTOR-targeting agents.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.