Aberrant regulation of nuclear factor kappa B (NF-κB) transcription factor is involved in cancer development, progression and resistance to chemotherapy. JG3, a marine-derived oligomannurarate sulfate, was reported as a heparanase and NF-κB inhibitor to significantly block tumor growth and metastasis in various animal models. However, the detailed functional mechanism remains unclear. Here, we report that JG3 inhibits NF-κB activation by specifically antagonizing the doxorubicin-triggered Ataxia-telangiectasia-mutated kinase (ATM) and the sequential MEK/ERK/p90Rsk/IKK signaling pathway but does not interfere with TNF-α-mediated NF-κB activation. This selective inactivation of the specific NF-κB cascade is attributed to the binding capacity of JG3 for Mre11, a major sensor of DNA double-strand breaks (DSB). Based on this selective mechanism, JG3 showed synergistic effect with doxorubicin in a panel of tumor cells and did not affect immune system function as shown in the in vivo delayed-type hypersensitivity (DTH) and hemolysis assays. All these highlight the clinical potential of JG3 as a favorable sensitizer in cancer therapy. In addition, identification of Mre11 as a potential target in the development of NF-κB inhibitors provides a platform for the further development of effective anticancer agents.
Nuclear factor kappa B (NF-κB) is a ubiquitously expressed family of transcription factors, which participates in a wide spectrum of cellular functions, such as the cell cycle, apoptosis, migration as well as immune and inflammatory responses.1–3 It is clear that aberrant activation of NF-κB and the signaling pathways regulating NF-κB activity are involved in cancer development and progression as well as resistance to chemotherapy.4–6
NF-κB activation is subject to complex regulation. Under basal conditions, NF-κB is sequestrated in the cytoplasm by the inhibitor of κB (IκB). Restricted NF-κB is released via three distinct signaling pathways, specifically, the canonical, alternative and atypical modes, which are induced by stimuli such as tumor necrosis factor α (TNF-α), CD40 ligand and genotoxic agents, respectively.7–9 The canonical and alternative ways play an important role in innate and adaptive immunity.10 Different from the classical or alternative pathway initiated by cytokines, the “atypical” pathway is triggered by DNA damage in the nucleus, which is a common feature of genotoxic agents, such as doxorubicin, VP16 and CPT.11–14
Several chemotherapeutic agents induce NF-κB activation, which, in turn, is used by tumor cells to achieve resistance to anticancer drugs. So, inhibiting NF-κB activation appears to be a promising option to improve the efficacy of conventional anticancer therapies.15–17 Accordingly, considerable efforts have been made to develop novel NF-κB inhibitors. To date, over 780 inhibitors targeting the NF-κB pathway have been identified, but few are clinically available.18 The major reason is that most inhibitors are general inhibitors targeting classical pathway, which leads to inevitable side-effects.18–20 The theory of selective inhibition NF-κB signaling pathway has attracted more and more attention, but still lack of successful examples.4, 21 The highlighted requirement for selectively targeting NF-κB signaling components is the driving force for developing novel NF-κB inhibitors that are distinct from conventional compounds.
JG3, a novel marine-derived semisynthetic oligosaccharide, was first reported to significantly inhibit angiogenesis and tumor metastasis in animal models by targeting heparanase.22 Previous experiments by our group further demonstrated that JG3 significantly inhibited in vivo tumor growth through blocking NF-κB activation, but the functional mechanism was not clear.23 In our study, we try to figure out the underlying mechanism of NF-κB inhibition and extend the potential clinical usage of JG3.
JG3 was obtained by semisynthesis after sulfate modification by reacting oligomannurarate with CLSO3H in formamide. The pH of products was adjusted to 7.0 with 4M NaOH and desalted using Sephadex G-10. The product peak was pooled and freeze-dried. The molecular weights of JG3 were analyzed with high performance gel permeation chromatography (HPGPC) using a G3000PWxl column (300 mm × 7.8 mm) (TOSOH, Japan). The structure of JG3 is depicted in Supplemental Figure 1a.
JG3 mAb was generated by hybridoma fusions of BALB/C mouse spleen cells and NS-1 myeloma cells. This mAb displays high affinity for JG3 with a KD value of 2.33 × 10−9 M, as determined using SPR. Analysis of the rate of cross-reactivity in an ELISA assay showed that the mAb did not react with the carrier proteins, BSA or OVA.
Electrophoretic mobility shift assay
BEL-7402 cells were pretreated with 50 μg/ml JG3 for 24 hr or left untreated and exposed to 1 μM doxorubicin for 12 hr or 10 ng/ml TNF-α for 30 min. Nuclear extracts were prepared as described previously.24 The supernatant was quantified using the BCA protein assay kit. Electrophoretic mobility shift assay (EMSA) was performed using the LightShift® Chemiluminescent EMSA kit (Pierce, Rockford, IL) with NF-κB biotin-labeled oligonucleotide.
NF-κB luciferase reporter assays
The NF-κB luciferase reporter assay was performed as described earlier.25 Briefly, cells were cotransfected with 3 × κBL and renilla luciferase expressing plasmid (internal control to normalize transfection efficiency) with Lipofectamine 2000, according to the manufacturer's instructions (Invitrogen, CA). After drug treatment for different time periods, firefly and renilla luciferase activities were assessed using a dual luciferase reporter gene assay kit (Beyotime, China). The following equation was applied: NF-κB transcriptional activity = (relative light units of firefly luciferase/relative light units of renilla luciferase) × 100.
Western blot analysis
Whole-cell protein lysates and cytoplasmic or nuclear extracts were electroblotted onto nitrocellulose membranes and probed with p-IKK (S176/180), p-IκB (S32), IκB, p-p65 (S536), p-ATM (S1981), p-MEK1/2 (S217/221), p-P44/42MAPK (T202/Y204), p-P90Rsk (S380), p-P38MAPK (T180/Y182) antibodies (1:1000 dilution, Cell signaling, Beverly, MA) as well as p65(RelA) and GAPDH antibodies (1:1000, Santa Cruz, CA).
Whole cell lysates of BEL-7402 were collected and subjected to affinity chromatography using beads bound to JG3. Columns were washed with 0.15 M NaCl solution to wash away unbound protein. A graded concentration of elution buffer was used to elute proteins bound to JG3.
DOCK 4.0 was used for conformational screening based on the X-ray crystal structure of the Mre11 protein reported in the Brookhaven protein database. Residues within 5 Å of the active center, Domains I and II of Mre11 were extracted separately as the binding pocket for docking. During docking simulation, different conformational isomers of trimannurarate were used to present JG3.
Surface plasmon resonance assay
The JG3 binding sequence on Mre11 was determined with the BIAcore X surface plasmon resonance apparatus. Briefly, based on the results of molecular modeling, we synthesized two wild-type binding and two mutant sequences (1, E227RWDFGDYEVRYE239WDGIKFK246ER248YG250; 2, ERWDFGDYEVRYAWDGIKFAEA YG; 3, I274DVK277IKGS281 and 4, IDVAIKGS). JG3 was immobilized on CM5 sensor chips, and unreacted surface moieties were blocked with ethanolamine. Changes in mass due to the binding response were recorded as resonance units. All binding experiments were performed at 25°C with a constant flow rate of 10 AL/min HBS-EP.
The Mre11 siRNA sequence used was 5′-GCC UCG AGU UAU UAA GAA ATT-3′ (225-245). Cells were transfected with siRNA using Oligofectamine 2000 reagent, according to the manufacturer's instructions (Invitrogen, CA).
Terminal transferase dUTP nick end labeling (TUNEL) assay
BEL-7402 cells (5 × 105 cells/ml were pretreated with 50 μg/ml JG3 for 12 hr or left untreated, followed by exposure to 1 μM doxorubicin for 48 hr. After treatment, the TUNEL assay was performed according to the manufacturer's instructions (Roche, Basel, Switzerland) and images were obtained using the Olympus BX51 UV fluorescence microscope.
Flow cytometric analysis
Cells were treated in a similar manner as above, fixed in 70% ice-cold ethanol, double labeled with annexin V FITC and propidium iodide and subjected to flow cytometry (FACSCalibur, Becton Dickinson, USA). FITC+/PI-, representing early stage apoptotic cells, was analyzed.
Delayed-type hypersensitivity (DTH) assay
Six- to eight-week old BALB/C mice were randomly divided into negative control (normal saline), positive control (cyclophosphamide) and treatment (JG3) groups, with 10 mice in each group. On the first day, the mice in positive control group were treated with 200 mg/kg cyclophosphamide once via i.p. and mice in JG3 group started treatment with 20 mg/kg/day JG3 via s.c. The mice in negative control group were treated with normal saline via s.c. On the third day, all the mice received an i.p. injection of 5% sheep red blood cells (SRBC). The mice in JG3 group continued treating with 20 mg/kg JG3 for 1 week. One week later, animals were challenged with 5% SRBC (0.05 ml) inoculated into the right footpad. A comparable volume of normal saline was injected into the left footpad as the control. The DTH reaction was recorded after 24 hr and footpad swelling was measured through comparing the thickness of right and left footpads with a digital caliper. After measurement, all the mice were euthanized. The protocol was approved by the IACUC of the institute.
Animal grouping and drug administration were similar to those described for the DTH assay. On the third day, mice received an i.p. injection of 5% SRBC(0.2 ml/mouse). One week later, serum samples were collected and react with SRBC. The hemolysis amount will be measured through plate reader with 413 nm as test wavelength. The OD value reflected the hemolysin level. After blood collection, all the mice were euthanized. The protocol was approved by the IACUC of the institute.
Data are presented as means ± SE, and differences considered significant at p < 0.05 were determined using the Student's t test.
Details of cell lines information and a number of assays are described in Supplementary Methods.
JG3 selectively inhibits doxorubicin-, but not TNF-α-, triggered NF-κB activation
Our previous study demonstrated that JG3 can inhibit NF-κB activation both in vitro and in vivo, but the underlying mechanism remains unclear. Given JG3 represses p65, a key member of NF-κB, nuclear translocation without affecting its expression, it is thus likely that JG3 blocks p65 translocation through direct interaction. Using Co-IP assay, we found that the monoclonal antibody raised against JG3 could not pull down the p65 protein in the cell lysate of BEL-7402 cells that were exposed to 50 μg/ml of JG3 for 24 hr, suggesting no direct interaction between p65 and JG3(Supplemental Fig. 1b).
Because JG3 does not bind to P65, we assumed that JG3 inhibits NF-κB activation by targeting upstream molecule(s). There are three distinct pathways, including classical, alternative and atypical pathways, which can stimulate NF-κB activation. Of them, the p65 activation mainly responds to the classical and atypical pathways. We, thus, selected TNF-α and doxorubicin,9, 13, 21, 26, 27 two typical stimuli to trigger the classical and atypical pathways, respectively. At first, we examined the influence of JG3 on classical pathway stimulated by TNF-α. We found that 10 ng/ml of TNF-α rapidly activated IKK and increased phosphorylation of p65 in 5 min and then followed by the IκB degradation in 10 min, indicative of the NF-κB activation. However, pretreatment with JG3 at 50 μg/ml for 24 hr did not reverse the activation of TNF-α-triggered upstream molecules (Fig. 1a), indicating that that JG3 does not affect classical pathway. For further confirmation, we examined the effect of JG3 on the interaction of NF-κB with DNA and transcriptional activity triggered by TNF-α. Pretreatment with JG3 at 50 μg/ml for 24 hr failed to interfere with TNF-α-triggered DNA-NF-κB binding event (Fig. 1c). Similar results were obtained in reporter assay, as evidenced by the findings that JG3 took little influence on the TNF-α-driven NF-κB transcriptional activity (Fig. 1d).
Having demonstrated that JG3 selectively inhibited NF-κB activation independent of the TNF-α-stimulated classical pathway, we focused on the effect of JG3 on atypical pathway driven by doxorubicin. Different from the classical one stimulated by TNF-α, we found that 1 μM of doxorubicin slowly led to the phosphorylation of IKK and p65 in 4 hr, followed by IκB degradation in 8 hr. Notably, pretreatment with 50 μg/ml of JG3 for 24 hr caused a marked reduction in phosphorylation of IKK and p65 and blocked IκB degradation accordingly (Fig. 1b). To fully substantiate this issue, we also measured the NF-κB-DNA binding and transcriptional activity under the same condition. Similarly, using EMSA, based on the densitometric analysis, we found that JG3 (50 μg/ml, 24 hr) attenuated approximately 35% DNA-NF-κB binding induced by doxorubicin, as compared to that in doxorubicin-stimulated group (Fig. 1c). Similarly, in reporter assay, JG3 significantly antagonized doxorubicin-triggered NF-κB transcriptional activity about one third as compared to that in doxorubicin-stimulated group (Fig. 1d).
Together these, we demonstrated that JG3 selectively inhibited doxorubicin-driven atypical NF-κB activation but not the concanical TNF-α-mediated NF-κB activation.
JG3 arrests doxorubicin-triggered ATM activation and sequential MEK/ERK/P90Rsk signaling pathway
Doxorubicin-driven DNA double strand breaks can be rapidly responded by ATM,28 which in one cascade initiates the activation of a MEK/ERK/IKK signaling pathway26 and in the other cascade recruits NEMO together with RIP into nucleus, in a more complicated manner, to active NF-κB.29 We first focused on the effects of JG3 on the ATM/MEK/ERK signaling pathway. Using western blotting, we found that 1 μM of doxorubicin triggered phosphorylation of ATM and activation of the MEK/ERK/P90Rsk signaling pathway. Intriguingly, pretreatment with 50 μg/ml of JG3 for 24 hr effectively inhibited doxorubicin-induced ATM phosphorylation as well as the MEK1/2/ERK1/2/P90Rsk signaling pathway (Fig. 2a). To further confirm the selective inhibition of JG3, we also examined the effect of JG3 on the same pathway stimulated by TNF-α. Consistent with our previous findings, JG3 had no effect on the activation of ERK1/2/P90Rsk induced by TNF-α (Fig. 2b).
We then examined the effect of JG3 on NEMO/RIP cascade. Stimulation with 1 μM of doxorubicin did not lead to a marked nuclear translocation of NEMO/RIP. Similarly, JG3 pretreatment also did not affect their nuclear translocation (Fig. 2c). Our results collectively indicated that JG3 failed to inhibit NEMO/RIP axis but selectively suppressed doxorubicin-driven NF-κB activation by specifically interfering with ATM/MEK/ERK/IKK pathway.
JG3 suppresses doxorubicin-induced NF-κB activation via interaction with Mre11
To determine the molecular basis of how JG3 inhibits doxorubicin-induced NF-κB activation, we focused on ATM first, because JG3 did not cause obvious cell growth inhibition (Fig. 4a) nor DNA damage, but effectively inhibited ATM phosphorylation. For this, we assumed that JG3 might physically interact with ATM. Unexpectedly, coimmunoprecipitation assay revealed that no directly interaction between JG3 and ATM (data not shown).
To find out the exact target of JG3, JG3-based affinity chromatography was used. Whole-cell lysates of BEL-7402 were collected and subjected to affinity chromatography using beads bound to JG3. The potential target proteins that bind to JG3 were eluted and identified by mass spectrum. Through screening, we found that JG3 can bind to Mre11. Mre11 is a member of MRN complex that plays an important role in both recruitment of ATM to the sites of DNA damage and efficient activation of ATM.30 Using JG3-based affinity chromatography combined with western blotting analysis, we found that JG3 exhibited high binding affinity for Mre11 (Fig. 3a, left) but failed to bind to the other two members of the MRN complex, Rad50 and NBS. To examine whether this setting can be replicated within cells, we performed a coimmunoprecipitation assay. The whole-cell lysates of BEL-7402 cells that incubated with 50 μg/ml JG3 for 24 hr were subjected to precipitation using monoclonal antibody raised against JG3. Consistent with JG3-based affinity chromatography data, Mre11 protein was pulled down by the JG3 antibody, further supporting that JG3 binds to MRE11. Consistently, JG3 failed to pull down Rad50 or NBS1 protein (supplementary Fig. 3), suggesting a specific interaction of JG3 with Mre11.
To further clarify the binding mode between Mre11 and JG3, computer molecular simulations were applied. A JG3 trimer was selected as the docking probe. Among more than 20 distinct docking runs using conformations of interest, the trimer docked strongly to amino acids on the β-sheet structure of domain II of Mre11, including R237, E239, K246, R248, K277 and R303 with hydrogen bond as a driving force (Fig. 3b). Based on the simulation prediction, two separate peptides, EG-24(07) (E227RWDFGDYEVRYE239WDGIKFK246ER248YG250) and IS-8 (I274DVK277IKGS281), were then synthesized. In parallel, two mutated peptides, EG-24(08) (E227RWDFGDYEVRYA239WDGIKFA246EA248YG250) and IS-8-2 (I274DVA277 IKGS281), were synthesized. The SPR assay revealed that both IS-8 and EG-24(07) bound strongly to JG3, yielding KD values of 2.99E-05 and 1.65E-06, respectively. By contrast, mutation at K277 (IS-8-2) led to a nearly complete loss of JG3 binding, whereas mutations at E239, K246 and R248 in EG-24(08) resulted in a 100-fold decrease in binding affinity (Fig. 3c). All these substantiate that E239, K246, R248 and K277 are critical for JG3-Mre11 binding.
To examine whether interactions between JG3 and Mre11 are functionally involved in doxorubicin-stimulated NF-κB activation, Mre11 expression was knocked down via siRNA manipulation. The status of NF-κB activation was measured using a reporter gene assay. On doxorubicin stimulation, NF-κB, in particular, was activated. Treatment with Mre11 siRNA potently inhibited doxorubicin-induced NF-κB activation compared to the mock control, indicating a critical involvement of Mre11 in doxorubicin-stimulated NF-κB activation. Similarly, JG3 also abrogated doxorubicin-triggered NF-κB activation as compared to the mock control, but failed to inhibit NF-κB activation on knock down of Mre11 (Fig. 3d). All these strongly suggested that Mre11 is involved in JG3-mediated NF-κB inactivation.
JG3 selectively sensitizes tumor cells to doxorubicin-driven events
NF-κB inhibitors were reported as favorable agents to sensitize chemotherapy.31, 32 The above results demonstrated that JG3 selectively inhibit doxorubicin-induced NF-κB activation. So, we extended to evaluate the sensitizing effects of JG3 on cell proliferation inhibition induced by doxorubicin. JG3 at 50 μg/ml didnot cause obvious cell growth arrest in a panel of human tumor cell lines, with an average cell growth inhibitory rate of 9.31% (Fig. 4a left). However, the pretreatment of JG3 at 50 μg/ml for 12 hr led to a marked increase in doxorubicin-triggered growth inhibition, with an average IC50 value of 0.202 μM, about 44% of the doxorubicin-alone group (with an average IC50 of 0.455 μM) (Fig. 4a right).
We then extended to examine the effects of JG3 on doxorubicin-induced apoptosis using BEL-7402 cells as a representative cell model. Data from the TUNEL assay showed that pretreatment with JG3 at 50 μg/ml enhanced doxorubicin-induced apoptosis, as compared to doxorubicin-treated alone (Fig. 4b left). Similar results were obtained in the FACS assay. As shown in Figure 4b right, a single treatment with JG3 did not induce apoptosis, whereas the application of 1 μM of doxorubicin for 12 hr resulted in 9.7% apoptosis. Strikingly, the combination treatment induced a significant apoptosis, with the apoptotic rate increasing to 17.6% (p < 0.05), nearly 2-fold increase than doxorubicin alone.
Next, we evaluated the correlation between NF-κB activation status and JG3-triggered sensitizing effects. We selected three kinds of cell lines: drug-sensitive liver carcinoma BEL-7402 cells, a less sensitive ovarian carcinoma SK-OV-3 cell line and an insensitive oral epidermoid carcinoma KB-3-1 cell line, as representative cell lines. We used p65 as a probe to reflect NF-κB activation status. Under basal conditions, p65 was localized mainly in the nucleus, suggesting the constitutive activation of NF-κB in BEL-7402 cells. However, in both SK-OV-3 and KB-3-1 cell lines, the majority of p65 was localized in the cytoplasm, indicating the inactive status of NF-κB. After exposure to 1 μM of doxorubicin for 12 hr, p65 was significantly activated and translocated to the nucleus in all three cell lines examined. However, the concomitant presence of JG3 induced a dramatic blockage on doxorubicin-induced NF-κB activation in BEL-7402 cells, a partial suppression in SK-OV-3 cells and limited inhibition in KB-3-1 cells (Fig. 4c). These data suggested that the sensitization to doxorubicin was closely correlated with NF-κB-inactivating potency by JG3.
To exclude out that JG3 enhances doxorubicin activity via increasing the intracellular concentration of doxorubicin, the intracellular concentrations of doxorubicin were examined in the presence or absence of JG3. Because doxorubicin exhibits self-fluorescence, the intensity of intracellular fluorescence was adopted to reflect its intracellular concentration through the FACS assay. We found that the fluorescence intensity in the doxorubicin group was significantly higher than that that of the blank group. However, the fluorescence intensity in the JG3 treatment group was identical to that of the doxorubicin group, indicating that JG3 does not influence intracellular concentration of doxorubicin (Fig. 4d).
JG3 does not interfere with immune function
Classical NF-κB activation plays an important role in innate and adaptive immune responses.10 Given JG3 can selectively inhibit doxorubicin-induced NF-κB activation bypassing the classical one mediated by TNF-α, we thus proposed that JG3 might not interfere with immune function. To test this hypothesis, we examined the effects of JG3 on immune system, using standard cell-mediated immunity model, delayed-type hypersensitivity (DTH) assay and humoral immunity model, hemolysis assay. From Figure 5a, we found that the increase in footpad thickness reach to 0.34 mm in negative control group, whereas reference drug cyclophosphamide significantly inhibited the paw swelling, which means the DTH model was successfully established. But the paw thickness in JG3-treated group (20 mg/kg/day) is almost the same with that of the negative control, indicating JG3 did not interfere with T cell function. Similarly, in hemolysis assay, dosing with 20 mg/kg/day JG3 for 9 days induced no decrease in hemolysin concentration compared to negative control group, indicating that JG3 does not result in B-cell function suppression in BALB/c mice (Fig. 5b). All these data supported that the selective inhibition of JG3 on NF-κB activation does not interfere with immune function.
NF-κB transcript factor has been well demonstrated to play an important role in innate and adaptive immune responses, tumor development, progression and resistance to therapeutics.1, 2, 33 Currently, expanding literatures have evidenced that NF-κB inhibitors can act as an adjuvant of conventional radiotherapy and chemotherapy, to enhance the treatment of many different malignancies in preclinical studies.17, 18, 20, 34–37
Although this strategy is very attractive, there are few clinical outcomes. One major obstacle is the mechanism-based immunosuppression.16, 20, 38 In view of the diverse structures of compounds targeting NF-κB, the majority of NF-κB inhibitors mainly target IKKβ that responses to classical and atypical pathways.10, 16, 18 As both classical and alternative NF-κB activation pathways play important roles in innate and adaptive immune responses,10 it should be considered the risk of severe immunodeficiency after general and prolonged NF-κB inhibition.4, 21 Based on this condition, many scholars suggested that an effective NF-κB inhibitor for cancer therapy should prevent NF-κB activation but have limited effects on other signaling pathways.16, 21 So, it was supposed to minimize systemic toxicity and avoid broad suppression of innate immunity by selectively targeting specific NF-κB signaling components. Because NF-κB activation pathway is very complex and lack of effective target, the development of selective inhibitor is quite slow.
In our study, we demonstrated that JG3 functions as a selective NF-κB inhibitor. JG3 specifically arrests atypical NF-κB activation induced by doxorubicin through inhibiting ATM activation and sequent MEK/ERK/IKK pathway, which bypasses NEMO/RIP pathway. This selective mechanism favors the attractive synergistic effect of JG3 with doxorubicin in a panel of tumor cells. The additional finding that JG3 also sensitized tumor cells to VP16 and other chemotherapeutics, substantiating that JG3 is a choice of combination with those chemotherapeutics that induce NF-κB activation (supplementary Fig. 2).
Attractively, JG3 had little influence on TNF-α-driven NF-κB transcriptional activity. This is evidenced by the fact that JG3 failed to reverse TNF-α-triggered either upstream activation or downstream DNA-NF-κB binding or eventual NF-κB transcription. Through both delayed-type hypersensitivity (DTH) and hemolysis assays, we further demonstrated that JG3 does not interfere with immune function in the animal models. All these data supported that JG3 not only sensitized the tumor cells to doxorubicin and other chemotherapeutics but also make low influence on the immune system.
This selective effect of JG3 was further found out basing on targeting of Mre11, a member of the Mre11-Rad50-Nbs1complex. The MRN complex can quickly sense genotoxic stress, then migrates quickly to the site of DNA damage, binds broken DNA ends and subsequently recruits ATM, facilitating ATM activation.28, 30 Our data revealed that JG3 bind to Mre11 through hydrogen bonds via four dominant residues, E239, K246, R248 and K277, which are spatially oriented away from the DNA-binding motifs. After eliminating Mre11with siRNA, we further demonstrated that Mre11 is involved in NF-κB activation induced by doxorubicin. Based on these findings, we speculate that interactions between JG3 and Mre11 disrupt MRN complex formation, in turn blocking the MRN complex response to DNA damage and leading to subsequent inactivation of ATM and downstream NF-κB. Because Mre11 locates in nucleus, it is no chance to involve in TNF-α stimulated classical pathway, which explains the selective effect of JG3 and also provide a clue to develop selective NF-κB inhibitors.
In summary, we show for the first time that JG3, a novel sulfated oligosaccharide, obviously sensitize tumor cells to doxorubicin's treatment through selectively blocking atypical NF-κB activation pathway. Our study highlights the importance of JG3 as a novel potential chemotherapy sensitizer in cancer therapy and also substantiates the hypothesis of selective inhibition on NF-κB activation. In addition, identification of Mre11 as a key mediator in atypical NF-κB activation aids in conceptualizing the potential value of Mre11 as a promising target in developing selective NF-κB inhibitors.
We are grateful to Dr. Spiros Linardopoulos (Institute for Cancer Research, Sutton, UK) for providing NF-κB-dependent firefly luciferase reporter plasmid 3×κBL and Renilla luciferase reporter plasmid.