mRNA translation is a therapeutic vulnerability necessary for bladder epithelial transformation

Using genetically engineered mouse models, this work demonstrates that protein synthesis is essential for efficient urothelial cancer formation and growth but dispensable for bladder homeostasis. Through a candidate gene analysis for translation regulators implicated in this dependency, we discovered that phosphorylation of the translation initiation factor eIF4E at serine 209 is increased in both murine and human bladder cancer, and this phosphorylation corresponds with an increase in de novo protein synthesis. Employing an eIF4E serine 209 to alanine knock-in mutant mouse model, we show that this single posttranslational modification is critical for bladder cancer initiation and progression, despite having no impact on normal bladder tissue maintenance. Using murine and human models of advanced bladder cancer, we demonstrate that only tumors with high levels of eIF4E phosphorylation are therapeutically vulnerable to eFT508, the first clinical-grade inhibitor of MNK1 and MNK2, the upstream kinases of eIF4E. Our results show that phospho-eIF4E plays an important role in bladder cancer pathogenesis, and targeting its upstream kinases could be an effective therapeutic option for bladder cancer patients with high levels of eIF4E phosphorylation.


Introduction
Bladder cancer is predicted to have afflicted 81,400 individuals within the United States in 2020.Furthermore, it is estimated that 17,980 men and women will have died from the disease in the same year (1).While new therapeutics including checkpoint blockade agents (2), FGFR inhibitors (3), and antibody drug conjugates (4) have recently been approved for patients with highly aggressive urothelial carcinoma, metastatic disease remains fatal for the majority of patients, despite a growing wealth of genomic data, which has expanded our understanding of lethal disease (5)(6)(7)(8).Therefore, new biology uncovering previously unrecognized therapeutic vulnerabilities is needed.Protein synthesis is a fundamental molecular process that is critical for cancer initiation and progression (9)(10)(11)(12)(13).However, it is unknown whether intact protein synthesis is necessary for normal bladder development and maintenance or the natural history of bladder cancer initiation and progression.These are critical questions because precisely targeting the translation apparatus is becoming increasingly possible through the development of targeted therapeutics with physiologic efficacy in patients (14)(15)(16).As such, understanding the protein synthesis requirements of the normal bladder and how it is perturbed in urothelial cancers represents a potential new treatment paradigm.
Here, we demonstrate that mRNA translation is a requisite for efficient bladder urothelial cell transformation but is dispensable for normal bladder epithelial development and maintenance in vivo.Through a candidate gene analysis of key translation regulators, we determine that phosphorylation of the oncogene eukaryotic translation initiation factor 4E (eIF4E) is significantly upregulated in the context of bladder tumor formation.Utilizing an eIF4E S209-knock-in mouse model, we show that this posttranslational modification is not only associated with transformation, but is also necessary for efficient tumorigenesis.The sole kinases responsible for eIF4E S209 phosphorylation are MAPK interacting serine/threonine kinase 1 (MNK1) and MNK2 -and eFT508, a selective inhibitor of MNK1 and MNK2, was recently developed and is currently being evaluated in trials (14,17,18).Using this compound along with murine-and human-derived organoid models and Using genetically engineered mouse models, this work demonstrates that protein synthesis is essential for efficient urothelial cancer formation and growth but dispensable for bladder homeostasis.Through a candidate gene analysis for translation regulators implicated in this dependency, we discovered that phosphorylation of the translation initiation factor eIF4E at serine 209 is increased in both murine and human bladder cancer, and this phosphorylation corresponds with an increase in de novo protein synthesis.Employing an eIF4E serine 209 to alanine knock-in mutant mouse model, we show that this single posttranslational modification is critical for bladder cancer initiation and progression, despite having no impact on normal bladder tissue maintenance.Using murine and human models of advanced bladder cancer, we demonstrate that only tumors with high levels of eIF4E phosphorylation are therapeutically vulnerable to eFT508, the first clinical-grade inhibitor of MNK1 and MNK2, the upstream kinases of eIF4E.Our results show that phospho-eIF4E plays an important role in bladder cancer pathogenesis, and targeting its upstream kinases could be an effective therapeutic option for bladder cancer patients with high levels of eIF4E phosphorylation.

Results
Optimal protein synthesis is required for efficient urothelial transformation.In order to determine if robust protein synthesis is necessary for bladder tumor formation, we used a mouse model haploinsufficient for the ribosomal protein L24 (rpL24 +/-).rpL24 +/-mice have a pleiotropic phenotype that includes a white ventral midline, white hind feet, and a kinked tail.rpL24 +/-are 20% smaller than WT littermates (19).Using the puromycin incorporation assay to measure de novo protein synthesis in vivo (20), we observed that rpL24 +/-mice display a 60% decrease in protein synthesis within bladder urothelium, the tissue of origin for urothelial carcinoma, compared with WT mice (Figure 1A) (19,21,22).Importantly, we found that rpL24 +/-mice develop normal bladders as determined by histological analysis, demonstrating that a reduction in global protein synthesis does not impact the formation or maintenance of normal urothelium (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.144920DS1).Therefore, this model, which exhibits a decrease in protein synthesis, enabled us to generally decrease mRNA translation levels without affecting homeostatic functions, which is why we chose this model to study the impact of reducing protein synthesis on bladder cancer formation.We used N-butyl-N-(4-hydroxybutyl) nitrosamine (BBN), a potent bladder-tropic procarcinogen found in cigarette smoke (the most common cause of bladder cancer; ref. 23), to drive cellular transformation in WT and rpL24 +/-mice (24)(25)(26).Importantly, BBN treatment in mice leads exclusively to muscle-invasive urothelial carcinomas that have been shown by Fantini et al. to be similar at a genomic and histologic level to human bladder cancer (27).We found that BBN administered in the drinking water caused invasive urothelial carcinoma after just 15 weeks of treatment in WT mice.By 21 weeks, nearly all WT mice grew invasive bladder cancer characterized by hematuria (blood in the urine) and hydronephrosis (distended kidneys from obstruction), which severely compromised their survival (Supplemental Figure 2, A and B).Histologic analysis revealed that this model produced hyperplasia and invasive urothelial carcinoma (Supplemental Figure 2B).We conducted a survival study using BBN in the 2 murine cohorts and observed a significant delay in bladder cancer-related deaths in rpL24 +/-mice (Figure 1B).The first WT mouse died from bladder cancer after 69 days of starting BBN, while the first reported death in rpL24 +/-mice occurred after 140 days.Furthermore, the median survival of WT mice was 167 days compared with 191 days for rpL24 +/- mice (P = 0.02).Overall, rpL24 +/-mice had a 12% increase in total lifespan on BBN compared with WT mice.All mice that died had histologic evidence of bladder cancer.However, age-matched rpL24 +/-mice developed smaller tumors compared with WT mice (Supplemental Figure 3A).We also confirmed by mass spectrometry that rpL24 +/-mice concentrated high levels of BCPN -the carcinogenic metabolite of BBN (25) -in the urine, similar to WT mice (Supplemental Figure 3B).Therefore, the improvement in survival and small tumors was not a result of decreased exposure to carcinogen in rpL24 +/-mice.As such, robust protein synthesis is needed for efficient urothelial cancer initiation, and a decrease in protein synthesis significantly delayed the onset of death and ultimately extended the survival of mice from bladder cancer.
Urothelial carcinoma is associated with increased protein synthesis and phosphorylation of the translation initiation factor eIF4E.These findings demonstrate the requirement of optimal protein synthesis for bladder cancer pathogenesis and raise the question of how protein synthesis is deregulated in bladder cancer.As such, we first sought to determine how the process of cellular transformation alters de novo protein synthesis rates in the urothelium.To this end, we conducted an immunoblot-based puromycin incorporation assay to measure new protein synthesis in normal primary bladder urothelial organoids and organoids derived from BBN-transformed bladder tumors.We observed a significant 2-fold increase in new protein synthesis in the context of cancer, which we confirmed by [ 35 S]-methionine incorporation (Figure 1C and Supplemental Figure 3C).This led us to ask if a specific node of the protein synthesis apparatus was associated with transformation in the BBN urothelial carcinoma model.We conducted a candidate gene analysis of critical regulators of mRNA translation previously associated with cancer formation, including: the mTOR signaling pathway, the translation initiation inhibitor eIF4E binding protein 1 (4EBP1), the translation elongation factor eEF2, the integrated stress response target eukaryotic initiation factor 2α (eIF2α), and eIF4E (11,(28)(29)(30)(31).We measured the protein levels and activation status of each candidate by Western blot analysis of primary bladder cancer organoids developed from BBN-treated mice.From these candidates, we observed an increase in eIF2α phos-

R E S E A R C H A R T I C L E
JCI Insight 2021;6(11):e144920 https://doi.org/10.1172/jci.insight.144920phorylation, which is known to decrease global protein synthesis as a stress response (32), and well as a slight decrease in 4EBP1 phosphorylation (Figure 1D).However, we noticed that eIF4E phosphorylation at S209 was upregulated in cancerous compared with normal urothelial organoids (Figure 1D).We also found that phospho-eIF4E levels increased significantly, from normal urothelium to invasive carcinoma in vivo (Figure 1E).These findings demonstrate that the activation of a central regulator of mRNA translation initiation, eIF4E, is strongly associated with increased protein synthesis and bladder cancer.
Phosphorylation of eIF4E at S209 is dispensable for normal urothelial homeostasis and necessary for carcinogeninduced bladder tumor initiation.To further investigate the role of eIF4E phosphorylation in bladder cancer formation, we utilized the eIF4E S209 mutant mouse model (eIF4E S209A/+ , eIF4E S209A/S209A ) (31).In this knock-in model, the serine at position 209 has been replaced by an alanine that cannot be phosphorylated.This mouse model develops normally without any change in body size compared with WT mice (Supplemental Figure 4A).IHC analysis of the urothelium confirmed a progressive decrease in eIF4E phosphorylation comparing WT, heterozygous, and homozygous mice (Figure 2A).Furthermore, the bladder epithelium in eIF4E S209A/+ and eIF-4E S209A/S209A mice develop normally, suggesting that eIF4E phosphorylation is not required for normal bladder tissue formation and maintenance (Figure 2A).We next asked whether a single mutation to S209 was sufficient to impact de novo protein synthesis within the bladder.Using the puromycin incorporation assay, we found that eIF4E S209A/+ and eIF4E S209A/S209A urothelium displayed decreased levels of new protein synthesis compared with WT urothelium (8.31% reduction [P < 0.0001] and 24.97% reduction [P < 0.0001], respectively) (Figure 2B).As such, eIF4E phosphorylation is necessary to maintain steady state levels of protein synthesis in the bladder urothelium but is not necessary for bladder development and homeostasis.
To understand the importance of phospho-eIF4E in bladder cancer initiation, WT, eIF4E S209A/+ , and eIF4E S209A/S209A mice were treated with BBN and all were euthanized at 9 or 15 weeks, which corresponds to the initiation of precancerous lesions and tumors, respectively.We observed that abrogating eIF4E phosphorylation significantly reduced urothelial thickness of precancerous lesions in 9-week-treated eIF4E-S209A/+ and eIF4E S209A/S209A mice compared with WT mice by 18.6% and 29.3%, respectively (Figure 2C).Moreover, we observed, after 15 weeks of BBN treatment, that WT mice had tumors approximately 5 times larger than eIF4E S209A/+ and eIF4E S209A/S209A mice (Figure 2D).These findings demonstrate that deficient phosphorylation of eIF4E leads to reduced bladder cancer initiation, consistent with an impairment in cellular transformation.To understand the impact of this deficiency on bladder cancer survival, we treated an additional cohort of mice with BBN until the development of terminal bladder cancer, as marked by pathological confirmation of bladder tumors and > 20% weight loss after 3 consecutive measurements.Similar to our findings in the rpL24 +/-mouse model, eIF4E S209A/S209A mice displayed a longer time to first death, as well as a significantly extended lifespan on BBN compared with WT and eIF4E S209A/+ mice (median survival: WT = 207 days, eIF4E S209A/+ = 197 days, and eIF4E S209A/S209A = 223 days [P = 0.03 for WT versus eIF4E S209A/S209A ; P = 0.01 for eIF4E S209A/+ versus eIF4E S209A/S209A ]) (Figure 2E).Importantly, the 16-26 day increase in survival translates into a 10% increase in the median lifespan compared with the WT mice in this study.We also confirmed by mass spectrometry that eIF4E S209A/S209A mice concentrated high levels of BCPN in the urine, similar to WT mice (Supplemental Figure 4B).Together, these findings show that phospho-eIF4E is necessary for efficient carcinogen-induced tumor initiation.eIF4E S209 phosphorylation is necessary for bladder cancer progression.Given our observation that eIF4E phosphorylation is necessary for efficient bladder cancer formation (Figure 2, C-E), we next asked whether eIF4E phosphorylation is required for the maintenance of established tumors.The primary kinases responsible for the phosphorylation of eIF4E are MNK1 and MNK2 (17,33).It has also been shown that both kinases are necessary for PTEN loss-mediated tumorigenesis (34).Furthermore, a potent and highly selective MNK1

R E S E A R C H A R T I C L E
JCI Insight 2021;6(11):e144920 https://doi.org/10.1172/jci.insight.144920and MNK2 inhibitor called eFT508 was recently developed and is currently in clinical trials for patients with advanced cancers (14).To determine if eIF4E phosphorylation is necessary for bladder cancer progression, we treated WT primary bladder cancer organoids with eFT508.We observed that eIF4E phosphorylation-competent tumors are sensitive to the inhibitory effects of MNK1 and MNK2 inhibition (Figure 3A).This finding shows that MNK1 and MNK2 activity are critical for the growth of established bladder cancer.However, it raises the question of whether eIF4E is the substrate responsible for the therapeutic response to eFT508.To this end, we generated eIF4E S209A/S209A bladder cancer organoids from BBN-induced bladder cancer mice and treated them with eFT508.Remarkably, eIF4E phosphorylation-defective mutant organoids were completely impervious to the antitumor effects of the drug (Figure 3A).These findings show that the growth inhibitory effects of eFT508 are mediated by the on-target inhibition of the MNK1/2-eIF4E signaling pathway.Thus, eIF4E phosphorylation is required for bladder cancer progression and represents a potential biomarker for responsiveness to MNK1 and MNK2 inhibition in urothelial malignancies.eIF4E S209 phosphorylation is a requisite for a therapeutic response to eFT508 in bladder cancer.To determine if these findings can be extended to human models of urothelial cancer, we screened a series of bladder cancer PDX models.We observed that 5 of 9 PDX lines exhibited moderate to high levels of eIF4E S209 phosphorylation, while 4 exhibited negligible levels of phosphorylation (Figure 3B).This finding demonstrates the heterogeneity of eIF4E S209 phosphorylation across models of bladder cancer.We subsequently created organoid models from 1 low eIF4E phosphorylation model (CoCaB1) (35) and 2 high eIF4E phosphorylation models (CoCaB14.1 and TM01029), which recapitulated the in vivo models (Supplemental Figure 5).These low-passage organoid lines were treated with increasing concentrations of eFT508 (0.01-10 μM), validated for target inhibition (Figure 3C), and measured for cell viability.Both high phospho-eIF4E lines (CoCaB14.1 and TM01029) demonstrated a dose-dependent decrease in cell viability in response to eFT508 (Figure 3D).Importantly, eFT508 did not have any cytotoxic effects on the phospho-eIF4E-negative line (CoCaB1) at all concentrations tested (Figure 3D).Taken together, our data demonstrate that phospho-eIF4E levels in human bladder cancer organoids positively correlate with responsiveness to MNK1 and MNK2 inhibition.
Next, we sought to determine if these in vitro finding were applicable to in vivo models of human bladder cancer.To this end, we used the CoCaB1 (phospho-eIF4E low) and TM01029 (phospho-eIF4E high) PDX models (Figure 3B).Mice were randomized to either 10 mg/kg of eFT508 or vehicle daily by oral gavage.We observed that the phospho-eIF4E high line was exquisitely sensitive to MNK1 and MNK2 inhibition, which led to significantly smaller tumors and an improvement in survival (Figure 4, A and B).At a cellular level, we observed that eFT508 treatment led to an increase in apoptosis (Figure 4C).Mice that received eFT508 exhibited no overt toxicity (Supplemental Figure 6).Remarkably, the phospho-eIF4E low line was completely insensitive to eFT508, with both the vehicle and drug treatment groups perishing at the same rate with no observable change in apoptosis (Figure 4, D-F).These results show that eFT508 functions through the precise inhibition of the MNK1/2-eIF4E signaling pathway.Furthermore, phospho-eIF4E plays an important role in bladder cancer progression, and targeting its upstream kinases could be an effective therapeutic option for bladder cancer patients with elevated phospho-eIF4E.

Discussion
Protein synthesis represents a fundamental molecular process that is critical for cancer behavior (36).However, not all types of cancers share the same dependencies on protein synthesis.For example, it has been shown that, while MYC-induced lymphomas are sensitive to decreases in protein synthesis, p53-induced solid tumors are impervious to similar decreases (37).Moreover, not all aspects of the protein synthesis machinery, even when they are converged upon by the same upstream pathway, are required for tumor initiation and progression.It has been shown that, despite the fact that mTOR regulates protein synthesis through posttranslational control of the 4EBP1/eIF4E and the S6K/rpS6 axes, the former and not the latter is required for AKT-mediated tumorigenesis of the thymus (28).However, in a genetically engineered model of pancreatic insulinoma, the S6K-rpS6 pathway is essential for tumor development (38).As such, different

R E S E A R C H A R T I C L E
JCI Insight 2021;6(11):e144920 https://doi.org/10.1172/jci.insight.144920tissue types have varying quantitative and qualitative requirements for protein synthesis to drive the process of tumor formation.Therefore, in order to understand whether a specific tissue type requires aberrant protein synthesis for the development of tumors, it is imperative to evaluate the tissue of origin in question.
Using a confluence of carcinogen-based genetically engineered mouse models, we demonstrate that the bladder urothelium requires robust protein synthesis to promote the process of cellular transformation and tumor growth.This is mediated, in part, through hyperphosphorylation of the oncogene eIF4E, since genetic perturbation of this posttranslational modification delays tumor initiation and progression, and improves overall survival.Moreover, we found that eIF4E phosphorylation levels dictate the ability of bladder tumors to respond to the clinical-grade MNK1 and MNK2 inhibitor eFT508.Importantly, this work demonstrates that the functional target of eFT508 in bladder cancer is eIF4E and that inhibition of its phosphorylation is required for its therapeutic efficacy.Moreover, these findings show that inhibition of eIF4E phosphorylation does not inhibit the growth of all tumors indiscriminately but, rather, requires a state of oncogenic addiction to this prominent posttranslational modification.Importantly, eIF4E phosphorylation has been demonstrated to be important in other cancers, including malignancies of the breast, colon, and prostate (31,39,40).Currently, phase 2 clinical trials are being conducted using eFT508 in combination with checkpoint inhibitors and concurrently with chemotherapy in metastatic triple-negative breast cancer (NCT03616834 [https://clinicaltrials.gov/ct2/show/NCT03616834?ter-m=NCT+03616834&draw=2&rank=1] and NCT04261218 [https://clinicaltrials.gov/ct2/show/ NCT04261218?term=NCT+04261218&draw=2&rank=1]).This study and the prevalence of eIF4E hyperphosphorylation within muscle-invasive bladder cancer provide the preclinical rationale for conducting phase 2 studies in urothelial carcinoma patients.
Urine BCPN analysis.Urine was collected from either WT and rpL24 +/-mice or WT and eIF4E S209A/S209A mice after 9 days of 0.075% BBN treatment.A total of 5 uL of urine was analyzed on a Xevo QTof mass spectrometer.BCPN concentrations were calculated by comparing the urine BCPN measurements to standards of known BCPN concentrations also measured by mass spectrometry.