An ATM/TRIM37/NEMO Axis Counteracts Genotoxicity by Activating Nuclear-to-Cytoplasmic NF-κB Signaling Free

Chromosomal integrity of all living organisms is endlessly jeopardized by genotoxic stress generated from endogenous metabolic sources, such as reactive oxygen species (ROS), and environmental resources, such as ionizing radiation and genotoxic chemicals (1–4). Unrepaired or inappropriately repaired DNA damage consequently attribute to genetic variation, apoptosis, aging, degenerative diseases, inflammation, and cancer (2–5). Meanwhile, cells have evolved a precisely controlled network of DNA damage response (DDR) to respond to genotoxic stresses, includes sensing of damaged DNA, activation of cell cycle checkpoints, assembly of DNA repair machineries, and transactivation of DNA damage-responsive gene expression. For instance, ATM kinase and PARP-1 act as sensor proteins to detect DNA lesions and modify a variety of proteins, which initiate DNA repair and cell-cycle checkpoint control (6, 7).

On the other hand, the NF-κB signaling pathway was emerged as a vital mediator for cellular responses to genotoxic threats via inducing survival genes that allow cells to repair damaged DNA and promote survival (8, 9). However, unlike the well-established classical NF-κB pathway in which signals are initiated from cell surface receptors and transduced from the cytoplasm to the nucleus (10, 11), genotoxic agents triggered the “nuclear-to-cytoplasmic” NF-κB pathway that is initiated from the nucleus and transferred to the cytoplasm (8, 9). In response to DNA damage signals, PARP-1 is rapidly recruited to DNA damage sites and induces auto-poly(ADP-ribosyl)ation (PARylation), which assembles NEMO, PIASy, ATM, PIDD, and LRP16 into a nucleoplasmic signalosome (12–16). Furthermore, signalosome formation induces sumoylation of NEMO at K277/K309 in a PIDD/PARP1/PIASy-dependent manner and ATM-dependent phosphorylation of NEMO at S85 (12–17). Then phosphorylated-NEMO is monoubiquitinated and exported from the nucleus together with ATM and ELKS, which form a complex with IKK catalytic subunits in the cytoplasm, consequently resulting in activation of IKK/NF-κB signaling (18). Therefore, nuclear monoubiquitin of NEMO, which is essential for its nuclear export, is the key step in genotoxic agent-triggered “nuclear-to-cytoplasmic” NF-κB signaling. However, the E3 ligase responsible for NEMO nuclear monoubiquitination remains unknown.

Tripartite motif containing 37 (TRIM37) is a newly identified E3 ubiquitin ligase that comprises a tripartite motif (TRIM, RING-B-box-coiled-coil) domain, TRAF domain (TD), and polyacidic domain (19–21). It has been recently reported that TRIM37 plays vital roles in various biological processes depending on TRIM domain-dependent E3 ligase activity, such as promotion of peroxisomal matrix protein import via direct monoubiquitination of PEX5 at K464 and silencing of gene expression through monoubiquitination of histone H2A (22, 23). In this study, we unveiled a novel function of TRIM37 in regulating nuclear-to-cytoplasmic NF-κB signaling. We found that, upon genotoxic stimulation, TRIM37 rapidly translocated into the nucleus where it interacted directly with TRAF6 to catalyze monoubiquitination of NEMO at K309. Blocking TRIM37/TRAF6 interaction using a cell-penetrating TAT-TD peptide abrogated NEMO monoubiquitination-dependent NF-κB signaling, resulting in hypersensitivity of esophageal cancer cells to DNA-damaging chemotherapeutics. Hence our study reveals a crucial role of TRIM37 in genotoxic stress-induced NF-κB activation and sheds light on mechanisms of inducible chemotherapy resistance in esophageal cancer.

Informed consent was signed by all patients, and the investigation has been conducted in accordance with the ethical standards according to the Declaration of Helsinki, national and international guidelines, which has also been approved by the authors' Institutional Review Board.

All of the patients received standardized platinum-based chemotherapy. Informed consent was obtained from all patients and approvals from Sun Yat-sen University Cancer Center Institutional Research Ethics Committee were obtained for this study. A total of 441 paraffin-embedded, archived esophageal cancer specimens and freshly collected 1 esophageal cancer-adjacent normal tissue and 9 esophageal cancer tissues were histopathologically and clinically diagnosed at Sun Yat-sen University Cancer Center (Guangdong, China) between 2005 and 2016. The clinical information of the samples is shown in Supplementary Tables S1 to S3. Detailed description provided in Supplementary methods.

Primary cultures of normal esophageal epithelial cells were established from fresh specimens of the adjacent noncancerous esophageal tissue, which was over 5 cm from the cancerous tissue, according previously report (24). The Eca109 cells were kindly provided by Professors Tsao SW (The University of Hong Kong) and grown in DMEM medium (Invitrogen) supplemented with 10% FBS (HyClone). All the cell lines been tested for Mycoplasma contamination. All cell lines were authenticated by short tandem repeat (STR) fingerprinting at Medicine Lab of Forensic Medicine Department of Sun Yat-Sen University.

Human TRIM37, NEMO, and TRAF6 were amplified by PCR and cloned into the pSin-EF2 vector. Fragments of the human TRIM37 and TRAF6-coding sequence were amplified by PCR and cloned into the pSin-EF2 vector. The indicated mutants were created using primers and a Stratagene Mutagenesis Kit according to the protocol recommended by the manufacturer. pNF-κB-luc and control plasmids (Clontech) was used to quantitatively examine NF-κB activity. Transfection of siRNAs or plasmids was performed using the Lipofectamine 3000 reagent (Invitrogen) according to the manufacturer's instruction. Retroviral production and infection were performed as described previously (25). Stable cell lines expressing indicated genes were selected for 10 days with 0.5 μg/mL puromycin 48 hours after infection. The primers used were listed in Supplementary Table S4.

IHC analysis was performed to study altered protein expression in 441 human esophageal cancer tissues according previous report (26). Paraffin-embedded tissues were analyzed using IHC with anti-TRIM37 antibody (Abcam; 1:200), anti-NF-κB p65 antibody (Abcam; 1:200), anti-Bcl-XL antibody (Cell Signaling; 1:100), anti-XIAP (1007-1008) antibody (Proteintech; 1:100). All the antibodies used in this study has been listed in Supplementary Table S5. The degree of immunostaining of formalin-fixed, paraffin-embedded sections were reviewed and scored separately by two independent pathologists uninformed of the histopathologic features and patient data of the samples. The scores were determined by combining the proportion of positively-stained tumor cells and the intensity of staining. The scores given by the two independent pathologists were combined into a mean score for further comparative evaluation. Tumor cell proportions were scored as follows: 0, no positive tumor cells; 1, <10% positive tumor cells; 2, 10% to 35% positive tumor cells; 3, 35% to 75% positive tumor cells; 4, >75% positive tumor cells. Staining intensity was graded according to the following standard: 1, no staining; 2, weak staining (light yellow); 3, moderate staining (yellow brown); 4, strong staining (brown). The staining index (SI) was calculated as the product of the staining intensity score and the proportion of positive tumor cells. Using this method of assessment, we evaluated protein expression in benign esophageal epithelia and malignant lesions by determining the SI, with possible scores of 0, 2, 3, 4, 6, 8, 9, 12, and 16. Samples with an SI ≥ 8 were determined as high expression and samples with an SI < 8 were determined as low expression. Cutoff values were determined on the basis of a measure of heterogeneity using the log-rank test with respect to OS.

Cells grown in 100-mm culture dishes were lysed using 500 μL of lysis buffer [25 mmol/L HEPES (pH 7.4), 150 mmol/L NaCl, 1% NP-40, 1 mmol/L EDTA, 2% glycerol, 1 mmol/L phenylmethylsulfonyl fluoride (PMSF)]. After being maintained on ice for 30 minutes, the lysates were clarified by microcentrifugation at 12,000 rpm for 10 minutes. To preclear the supernatants, the lysates were incubated with 20 μL of agarose beads (Calbiochem) for 1 hours with rotation at 4°C. After centrifugation at 2,000 rpm for 1 minutes, the supernatants were incubated with 20 μL of antibody-cross-linked protein G-agarose beads overnight at 4°C. The agarose beads were then washed six times with wash buffer [25 mmol/L HEPES (pH 7.4), 150 mmol/L NaCl, 0.5% NP-40, 1 mmol/L EDTA, 2% glycerol, 1 mmol/L PMSF]. After removing all the liquid, the pelleted beads were resuspended in 30 μL of 1 M glycine (pH 3), after which, 10 μL of 4× sample buffer was added, the samples were denatured, and the sample components were electrophoretically separated on SDS-PAGE for immunoblot analysis.

Eca109 cells seeded on No. 118 mm round coverslips were washed twice with PBS, fixed with 4% PFA in PBS for 1 hour, permeabilized with the PBST for 10 minutes, and blocked with 4% BSA in PBS for 1 hour. The cells were then incubated with primary antibodies for 3 hours and with Alexa Fluor 647 conjugated goat anti-rabbit IgG (1:200 dilution) and pre-adsorbed Alexa Fluor 568 conjugated goat anti-mouse IgG (1:200 dilution; Abcam, ab175733) antibodies for 1.5 hours. The sample was kept in PBS until imaging.

Peptides were synthesized by the ChinaPeptides using standard HOBt/Fmoc chemistry and purified by reverse phase HPLC to >95% purity. The final amino acid compositions were verified utilizing amino acid analysis and MALDI TOF mass spectrometry. The cell permeation sequence used was previously demonstrated to have 10-fold increased intracellular concentration compared with the native TAT sequence from HIV-1 (27). We designed a negative control peptide by changing DFEVGE to AFAVGA. The amino acid sequences of these peptides are: TAT-Ctrl: RKKRRORNRRRAFAVGA; TAT-TBM: RKKRRORNRRRDFEVGE.

In the subcutaneous patient-derived xenografts (PDX) tumor model, freshly isolated clinical esophageal cancer patient tissues were subdivided into 2 to 3 mm³ pieces, coated with Matrigel (BD Biosciences) and media in a 1:1 ratio, and embedded within the subcutaneous space underneath the skin of female NOD/Shi-scid/IL-2Rγ^null (NOG) mouse (6–8 weeks old; CREA Japan Inc.). Tumor growth was monitored by measuring the tumor diameters, and the tumor volume was calculated using the equation (L × W²)/2. When the tumor became palpable, mice were intratumoral treated with cisplatin (CDDP; 5 mg/kg) and intratumoral injection of TAT-control peptide or TAT-TRIM37/TBM peptide three times per week (as per cycle) for up to 6 weeks. At the end of treatment, the mice were sacrificed and the tumors were excised and weighed, and confirmed by histology.

In the subcutaneous tumor model, the indicated luciferase expressing cells (1 × 10⁶) were injected subcutaneously into nude mice. When the luminescence signal reached 2 × 10⁷ p/sec/cm²/sr, mice were intratumoral treated with CDDP (5 mg/kg) three times per week (as per cycle) for up to 6 weeks. Mice were sacrificed when moribund as determined by an observer blinded to the treatment, and tumors were excised and paraffin-embedded. Serial 4.0 μm sections were cut and subjected to IHC and hemotoxylin and eosin (H&E) staining. After deparaffinization, sections were H&E-stained with Mayer's hematoxylin solution, or IHC-stained using antibodies of NF-κB p65 (1:100; Abcam), or stained with TUNEL (In Situ Cell Death Detection Kit, TMR red; Roche Applied Science), and counterstained with phalloidin (Alexa Fluor 488; Invitrogen) and DAPI (Sigma-Aldrich) according to manufacturer's protocols. The images were captured using the AxioVision Rel.4.6 computerized image analysis system (Carl Zeiss). The mice used in this study were sacrificed when the volume of control tumors reached to 1.5-cm diameter or the mice become moribund. All of the animal procedures were approved by the Sun Yat-sen University Animal Care Committee.

All statistical analyses were carried out using SPSS130.0 statistical software. A chi-squared test was used to analyze the relationship between TRIM37 expression and the clinicopathologic characteristics. Survival curves were plotted using Kaplan–Meier method and compared by log-rank test. Survival data were evaluated by univariate and multivariate Cox regression analyses. P < 0.05 was considered statistically significant.

To identify potential nuclear E3 enzyme for NEMO monoubiquitination, affinity purification and mass spectrometry (MS) was conducted using nuclear extracts from etoposide-treated Eca-109 esophageal cancer cells. In addition to previously reported interacting proteins, such as TRAF6, ATM, and PIASy (12–17), we found that E3 ligase TRIM37 may also be a potent nuclear NEMO-interacting protein (Fig. 1A). Prominently, compared with control cells, etoposide-induced NF-κB activation was rapidly elevated in TRIM37-overexpressing cells but decreased in TRIM37^−/− cells (Fig. 1B and C), suggesting that TRIM37 played a vital role in etoposide-induced NF-κB activation. This hypothesis was further confirmed by multiple assays, in which overexpression of TRIM37 significantly increased, but knockout or knockdown of TRIM37 decreased the IKK activity, the expression of NF-κB-regulated antiapoptotic genes Bcl-XL and XIAP, and the NF-κB transcriptional activity in the etoposide-treated cells (Fig. 1C; Supplementary Fig. S1A and S1B). Importantly, the elevated NF-κB activity induced by irradiation, camptothecin, or CDDP was also drastically decreased in TRIM37^−/− cells (Supplementary Fig. S1C and S1D), demonstrating that TRIM37 contributed to genotoxic stress-induced NF-κB activation.

Statistical analyses revealed that TRIM37 levels were positively correlated with level of activated NF-κB and expression of Bcl-xl and XIAP in 441 paraffin-embedded and nine freshly collected esophageal cancer specimens (Fig. 1D and E). IHC staining showed that TRIM37 protein was slightly expressed in normal esophageal tissues but showed markedly higher expression in esophageal cancer and was further elevated in relapsed esophageal cancer (Fig. 1F). Furthermore, we found that TRIM37 expression was positively associated with the clinical stage (P = 0.001), tumor-node-metastasis classification (P = 0.026; P = 0.004; P = 0.039) and inversely associated with overall and relapse-free survival in patients with esophageal cancer (Supplementary Table S1–S3 and Fig. 1G). Importantly, Kaplan–Meier plotter analysis revealed that expression of TRIM37 were significant correlated with shorter overall and relapse-free survival in patients with breast cancer, ovarian cancer, or liver cancer (Supplementary Fig. S1E). These results indicate that TRIM37 overexpression may contribute to development and progression of multiple types of human cancer.

NF-κB activation usually induces various antiapoptotic genes that allow cells to survive in the presence of DNA-damaging drugs (8, 9). We therefore examined the effect of TRIM37 on resistance to CDDP, a DNA-damaging drug that is commonly used in anticancer therapies. As shown in Fig. 2A and B, overexpressing TRIM37 decreased the CDDP-induced apoptotic death and increased the colony formation. The same results were also obtained using an in vivo tumor model in which TRIM37/tumors exhibited remarkable resistance to CDDP therapy, as indicated by rapid tumor progression, lower proportion of apoptotic cells, and increased NF-κB activation, consequently resulting in shorter survival of tumor-bearing mice (Fig. 2C–F). In contrast, CDDP treatment led to significant remission of TRIM37^−/−/tumors (Fig. 2C–F). These results demonstrate that TRIM37 overexpression confers resistance to CDDP via NF-κB activation.

Although we observed no alterations of etoposide-induced sumoylation and phosphorylation of NEMO between TRIM37-dysregulated and control cells, the expression of etoposide-induced monoubiquitinated NEMO was rapidly increased in TRIM37-transduced cells but decreased in TRIM37^−/− cells (Fig. 3A; Supplementary Fig. S2A). The promotive effect of TRIM37 on etoposide-induced NEMO monoubiquitination was drastically abrogated by silencing of ATM or PIASy but remained in IKKβ-, cIAP1-, or TAK1-silenced cells (Fig. 3B). These results suggest that TRIM37 may participate in NEMO monoubiquitination.

MS analysis revealed that Lys309 was the TRIM37-monoubiquitinated residue (Fig. 3C). Accordingly, overexpressing TRIM37 in the etoposide-treated NE1/NEMO^−/− cells had no impact on the monoubiquitinated level of Lys309 NEMO mutant (K309A) but significantly increased the monoubiquitinated level of NEMO/WT and other NEMO lysine mutants (K277A, K285A, and K399A; Fig. 3D). Meanwhile, TRIM37 overexpression in NE1/NEMO^−/− cells dramatically enhanced the recovery effect of NEMO/WT, but not NEMO/K309A, on the genotoxic stress-induced DNA-binding and transcriptional activities of NF-κB and IKK activity (Fig. 3E and F; Supplementary Fig. S2B–S2D). These results demonstrate that TRIM37-mediated NEMO monoubiquitination at K309 is vital for genotoxic NF-κB activation.

In agreement with previous findings that monoubiquitination of NEMO is crucial for its nuclear export (15), we found that overexpressing TRIM37 dramatically decreased the nuclear expression of NEMO/WT but had no impact on nuclear level of NEMO/K309A in etoposide-treated cells (Fig 3G), suggesting that TRIM37 promoted nuclear export of NEMO. This hypothesis was further supported by the observation that the duration of etoposide-induced nuclear NEMO in TRIM37-transduced cells was much shorter than that in control cells, while TRIM37^−/− cells showed sustained nuclear NEMO signal (Fig. 3H; Supplementary Fig. S2E). However, the effect of TRIM37 on nuclear export of NEMO was abolished by NEMO/K309A (Fig. 3H). Consistent with previous report (28), calcium chelation prevented the cytoplasmic expression of monoubiquitinated NEMO in the TRIM37-transduced cells (Fig. 3I and J), which provided further evidence that TRIM37 mediated monoubiqutination of NEMO in the nucleus. Remarkably, the etoposide-induced association of NEMO with Ran, an export receptor and the cargo protein for nuclear export of NEMO (29), was increased in TRIM37-transduced cells but was nearly undetectable in TRIM37^−/− cells, and no interaction of Ran and NEMO/K309A was observed (Fig. 3K). These results demonstrated that TRIM37- mediated monoubiquitination of NEMO promotes nuclear export of NEMO.

Coimmunoprecipitation (Co-IP) assays revealed the genotoxic stress-induced TRIM37/NEMO interaction only occurred in the nucleus but not in the cytoplasm (Fig. 4A–C; Supplementary Fig. S3A and S3B). We then asked whether genotoxic stress could induce translocation of TRIM37 into nuclear whereas TRIM37 interacted with and monoubiquitinatd NEMO. Interestingly, the nuclear expression of TRIM37 dramatically elevated at 30 min after genotoxic stress (Fig. 4D and E; Supplementary Fig. S3C). However, TRIM37 containing a mutant nuclear localization signal prevented genotoxic stress-induced nuclear translocation of TRIM37 and interaction with NEMO, as well as TRIM37-induced genotoxic NF-κB activation and NEMO monoubiquitination (Supplementary Fig. S3D–S3G).

Far-Western blot analysis that TRIM37 could not interact directly with NEMO, suggesting that TRIM37-mediated genotoxic NEMO monoubiquitination required other protein(s). By individually silencing all identified nuclear NEMO-interacting proteins in Figure 1A, we found that silencing TRAF6 in the etoposide-treated cells almost entirely abrogated genotoxic stress-induced TRIM37/NEMO interaction but knocking down NEMO did not reduce the binding affinity of TRIM37 for TRAF6, and that ablating TRIM37 had no obvious impact on NEMO/TRAF6 association, indicating that TRAF6 is required for formation of a TRIM37/TRAF6/NEMO complex (Fig. 4F). Consistently, the nuclear level of TRAF6 was also drastically elevated at 30 min after genotoxic stress (Supplementary Fig. S4A). The direct nuclear interaction of TRIM37 and TRAF6 was confirmed by far-western blot and STORM analyses (Fig. 4G and H). Interestingly, although the zoom-in STORM image showed TRAF6 (red) stained in close proximity with TRIM7(green) both in the cytoplasm (middle, top) and nucleus (middle, bottom), 3D render revealed that TRAF6 interacted directly with TRIM37 in the nucleus (right, bottom) but not in the cytoplasm (right, top; Fig. 4H). These results provided further evidence that TRIM37 forms complex with TRAF6 in the nucleus upon genotoxic treatment. Consistently, silencing TRAF6 abolished the effect of TRIM37 on genotoxic stress-induced NEMO monoubiquitination and NF-κB activity (Fig. 4I; Supplementary Fig. S4B). Moreover, co-IP assays revealed that with increasing degrees of genotoxic stress, more TRIM37 was found complexed with TRAF6 and NEMO, and that TRAF6 induced-interaction between TRIM37 and NEMO was in a TRAF6 dose-dependent manner (Fig. 4J and K; Supplementary Fig. S4C). Importantly, overexpressing TRAF6-C70A mutant or S13A/T330A failed to recover the level of NEMO monoubiquitination in TRAF6-silencing cells treated with etoposide, suggesting that nuclear translocation and E3 activity of TRAF6 contributes to TRIM37-mediated NEMO ubiquitination (Fig. 4L). In vitro protein binding and ubiquitination assays further demonstrated that TRAF6 was essential for TRIM37-mediated NEMO monoubiquitination via direct interaction with TRIM37 (Fig. 4M). Notably, NEMO S85A mutant also dramatically abrogated the effect of TRIM37 on NEMO monoubiquitination, indicating that ATM-mediated S85 phosphorylation of NEMO was early event of its monoubiquitination (Fig. 4M).

TRIM37 was previously reported to be localized in peroxisomal membranes (19–23) and ATM kinase was also found to be recruited by PEX5 to peroxisomal membranes (30, 31). Consistently, immunofluorescence and subcellular fractionation assays revealed that TRIM37 and ATM were colocalized in peroxisomal membranes (Fig. 5A and B). Interestingly, etoposide-induced TRIM37 nuclear translocation was drastically prevented by an inhibitor of ATM (Fig. 5C), suggesting that ATM was involved in TRIM37 nuclear translocation upon genotoxic stress. Meanwhile, we found that genotoxic stress induced the phosphorylation of TRIM37 at 196TQ/801SQ motifs (Fig. 5D) and mutating TRIM37 TQ/SQ sites T196 and S801 to alanine reduced genotoxic stress-induced phosphorylation and nuclear translocation of TRIM37 (Fig. 5E). Moreover, we observed that that the genotoxic stress-induced ATM/TRIM37 complex only formed in the cytoplasm but not in the nucleus (Fig. 5F and G). Co-IP assays using ATM and serially truncated TRIM37 fragments demonstrated that ATM interacted with the TD of TRIM37 (Fig. 5H). Importantly, treatment with antioxidants not only dramatically prevented genotoxic stress-induced TRIM37 nuclear translocation but also reduced genotoxic stress-induced ATM/TRIM37 interaction and NEMO monoubiquitination (Fig. 5I–L). Therefore, oxidative stress may be also involved in ATM-mediated phosphorylation and nuclear translocation of TRIM37 and NEMO monoubiquitination upon genotoxic stress.

Co-IP assays using serially truncated TRIM37 and TRAF6 fragments demonstrated that the TD was the interaction region of TRIM37 and TRAF6 (Fig. 6A and B). The TD of TRIM37 contains a sequence (DFEVGE; residues 366-371) with homology to a consensus TRAF6-binding motif (TBM), PXEXX (aromatic/acidic residue; ref. 32). Interestingly, a TRIM37 mutant containing a TBM deletion (TRIM37/ΔTBM) lost the capacity to directly bind to TRAF6, resulting in resistance to genotoxic stress-induced monoubiquitination and nuclear export of NEMO, as well as NF-κB activation (Fig. 6C–F; Supplementary Fig. S5A and B). Consistently, immunoprecipitation assays using anti-Ran antibody showed that TRIM37/ΔTBM and TRIM37/RF-mu abrogated the interaction between NEMO and Ran (Fig. 6G).

To further determine whether the E3 ligase activity of TRIM37 was essential for NEMO monoubiquitination, a TRIM37 derivative (TRIM37/RF-mu) bearing a point mutation in a conserved cysteine residue in the RING finger motif (C18R), which interferes with catalytic activity, was constructed. Although TRIM37/RF-mu could still form a complex with TRAF6/NEMO upon etoposide treatment, overexpressing TRIM37/RF-mu in the TRIM37^−/− cells could not recover the genotoxic stress-induced monoubiquitination and nuclear export of NEMO and NF-κB activation (Fig. 6C–G; Supplementary Fig. S5A and S5B), demonstrating that TRIM37-mediated NEMO monoubiquitination is TRIM domain-dependent.

The cell-penetrating HIV-1 transactivator of transcription (TAT)-peptide has been widely used as an anticancer molecular delivery system due to its high solubility and penetrability (33, 34). We then examined the inhibitory effect of TAT-conjugated TRIM37/TBM (TAT-TBM) peptide on genotoxic NF-κB activation and cancer progression. As shown in Fig. 7A and B, treatment with TAT-TBM peptide dramatically reduced genotoxic stress-induced TRIM37/TRAF6/NEMO complex formation, IKK/NF-κB activation, and NEMO monoubiquitination in the cells with higher TRIM37 expression, consequently resulting in enhanced effect of CDDP on cell death as indicated by increased caspase 3 positive cells and decreased colony formation (Fig. 7C and D). Furthermore, a PDX model, using two freshly collected clinical esophageal cancer tissues, showed that cotreatment with CDDP and TAT-TBM peptide resulted in significant remission of esophageal cancer tumor volume and mass (T#6: 34 mg vs. 657 mg; T#9: 65.8 mg vs. 983 mg) compared with control tumors (Fig. 7E), as well as higher percentage of TUNEL⁺ cells and reduced genotoxic NF-κB activity (Fig. 7F and G), in comparison with CDDP treatment alone. Taken together, these results further support the notion that TRIM37 overexpression enhanced genotoxic stress-induced NF-κB activation, consequently resulting in chemoresistance and poorer clinical outcomes in human cancer.

Unlike activation of the classical NF-κB signaling pathway that displays a fast response to typical signals initiated from cell surface receptors, “atypical” NF-κB activators, such as DNA damage or oxygen stress, trigger a slow NF-κB signal (with peak activities reached after 2–4 hours; refs. 8, 35). These delayed kinetics were previously thought to be at least partially attributable to the time required to transfer the nuclear damage signal to the cytoplasmic IKK complex. Multiple recent studies have provided insights that genotoxic threats triggering NF-κB activation require molecular trafficking between the nucleus and cytoplasm. For instance, genotoxic stress-induced nuclear translocation of IKK unbound-NEMO is the key event for DNA damage-dependent IKK/NF-κB signaling (36). Meanwhile, genotoxic stress-induced nuclear translocation of PIDD resulted in augmentation of sumoylation and ubiquitination of NEMO (14). However, monoubiquitinated NEMO, which complexes with ATM and ELKS, was exported from the nucleus and was shown to be essential for cytoplasmic IKK activation (18). Herein, we found that in response to genotoxic stress, it took nearly 30 minutes for peroxisomal E3 ligase TRIM37 to translocate into the nucleus where it associated TRAF6 and NEMO, which resulted in NEMO monoubiquitination at K309. Therefore, the DNA damage-triggered slow NF-κB signal may be caused by TRIM37 nuclear translocation. Importantly, we demonstrated that ATM kinase, a sensor of DNA damage, played an important role in genotoxic stress-induced nuclear translocation of TRIM37 via direct physical interaction with and phosphorylation of TRIM37. Interestingly, two ATM-mediated phosphorylation sites are present in TRIM37, in which T196Q is within the peroxisomal targeting signal of TRIM37, and S801Q is in proximity to the nuclear localization signal of TRIM37. These results suggested that ATM-mediated phosphorylation may lead to dissociation of TRIM37 from the peroxisome and nuclear transition via exposure of the TRIM37 nuclear localization signal. Although it has been reported that genotoxic stress-activated nuclear ATM translocated into cytosolic and membrane fractions within 10 minutes (36), we found that antioxidants prevented genotoxic stress-induced ATM/TRIM37 interaction and TRIM37 nuclear translocation, suggesting that oxidative stress-activated ATM, instead of DNA damage-activated ATM, contributed to phosphorylation of TRIM37. This was consistent with previous reports that oxidative stress could induce activation of peroxisomal ATM kinase (30, 31). Therefore, our study provided a novel mechanism and role for ATM in genotoxic stress-induced NF-κB activation.

The TRIM37 gene on chromosome 17q22–23 was originally found to be frequently mutated in patients with mulibrey nanism, a disease with dramatic growth impairment in several organs (19, 20). Further studies demonstrated that the roles of TRIM37 in various biological processes depend on TRIM domain-dependent E3 ligase activity. For instance, TRIM37 is involved in peroxisomal matrix protein import via monoubiquitination of PEX5 (22). Enforced expression of a TRIM37 mutant that lacks E3 ligase activity could not prevent the TRIM37 depletion-resulted supernumerary centrosomal-component foci (37). Although multiple studies reported that TRIM37 is mainly distributed in the cytoplasm, such as in peroxisomes (20–22), Bhatnagar and colleagues found that TRIM37 associated with PRC complex in the nucleus to establish a repressive chromatin structure (23), suggesting that TRIM37 could translocate into the nucleus. Herein, we found that in response to DNA damage, ATM-mediated phosphorylation of TRIM37 led to its rapid translocation into the nucleus, where it forms a TRIM37/TRAF6/NEMO complex that catalyzes NEMO monoubiquitination, ultimately leading to NF-κB-mediated anti-apoptotic transcription. Treatment with a cell-penetrating TAT-TBM peptide, which blocked interaction of TRIM37/TRAF6, abrogated genotoxic stress-induced NEMO monoubiquitination and NF-κB activation, resulting in hypersensitivity of cancer cells to genotoxic chemotherapy. Therefore, our findings not only reveal a crucial role of TRIM37 in genotoxic stress-induced NF-κB activation, but also have important translational implications for the mechanistic understanding of therapeutic TRIM37 inhibitors that can potentate the effect of chemotherapeutic drugs or ionizing radiation in cancer therapy.

In conclusion, preventing genotoxic stress-induced NF-κB activation, which results in development of chemotherapy resistance, will be beneficial for a large group of patients with cancer. Herein, we demonstrated that genotoxic stress-induced E3 ligase TRIM37 contributed to NEMO monoubiquitination and genotoxic IKK/NF-κB activation, consequently leading to genotoxic stress resistance of esophageal cancer. Therefore, further investigation into the role of TRIM37 in resistance of chemotherapy-induced genotoxic stress will not only provide valuable insights to better understand imitation and progression of cancers but also may eventually lead to the development of novel therapeutic strategies for treatment of human cancers.

No potential conflicts of interest were disclosed.

Conception and design: L. Song, J. Li

Development of methodology: G. Wu, J. Zhu,

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): G. Wu, J. Zhu, Y. Hu, L. Cao, Z. Tan, S. Zhang, Z. Li,

Writing, review, and/or revision of the manuscript: L. Song, J. Li

Study supervision: J. Li

This work was supported by Natural Science Foundation of China [No. 81830082, 91740119, 91529301, and 81621004 (all to J. Li); 91740118, 81773106, and 81530082 (all to L. Song)]; Guangzhou Science and Technology Plan Projects (201803010098 to J. Li); Guangdong Natural Science Foundation (2018B030311009 to J. Li; 2016A030308002 to L. Song); The Fundamental Research Funds for the Central Universities (No. 17ykjc02 to J. Li).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.