Tris DBA Abrogates Tumor Progression in Hepatocellular Carcinoma and Multiple Myeloma Preclinical Models by Affecting Multiple Oncogenic Proteins

Background: STAT3 is an oncogenic transcription factor that controls the expression of genes associated with proliferation, apoptotic resistance, malignant transformation, and metastasis. Persistent activation of STAT3 is observed in many types of human malignancies including hepatocellular carcinoma (HCC) and multiple myeloma (MM). Methods: Here, we have investigated the action of Tris(dibenzylideneacetone) dipalladium 0 (Tris DBA), a palladium complex on STAT3 signaling cascade in HCC and MM. The cytotoxic and proapoptotic activity of Tris DBA was evaluated by various biochemical assays. The action of Tris DBA on cytokine-induced/constitutive activation of STAT3, non-receptor tyrosine kinases (NRTKs), and expression of STAT3 driven genes was evaluated. Nuclear translocation of STAT3 and its DNA interaction was also studied. The antitumor activity of Tris DBA was investigated in two different preclinical studies namely, xenograft MM and orthotopic HCC mice models. Results: Tris DBA decreased cell viability, increased the apoptosis, and inhibited the activation of STAT3 and NRTKs. Tris DBA downmodulated the nuclear translocation of STAT3 and reduced its DNA binding ability. It upregulated expression of SHP2 (protein and mRNA) to induce STAT3 dephosphorylation and inhibition of SHP2 reversed this effect. It downregulated the expression of STAT3-driven genes and suppressed cell motility. Tris DBA signicantly inhibited tumor growth in xenograft MM and orthotopic HCC mice models with reduction in the expression of various prosurvival biomarkers in MM tumor tissues without displaying any signicant toxicity. Conclusions: Tris DBA functions as a good inhibitor of STAT3 signaling in preclinical HCC and MM models. Anilkumar NC, Rangappa S, Shanmugam MK, Mishra S, Chinnathambi A, et al. Novel 1,3,4-Oxadiazole Induces Anticancer Activity by Targeting NF-κB in Hepatocellular Carcinoma Cells.


Background
Cancer is a prominent public health concern and the second major cause of death globally after cardiovascular disease and both its incidence and mortality rates are increasing every year (1)(2)(3). Hepatocellular carcinoma (HCC) is the leading type of liver cancer which affected 0.8 million people in 2012 throughout the globe (4) and the 5-year survival rate has been estimated to be between 5-14% which is critically low (5). Hepatitis B infection remains the leading risk factor of HCC followed by nonalcoholic fatty liver disease, alcoholic hepatitis, a atoxin B1 intake, and hemochromatosis (6). HCC may be often detected at metastatic stage, which drastically reduces the treatment e cacy and options (7,8).
Multiple myeloma (MM) stands second among hematological malignancies in the western countries contributing approximately 10% of hematological malignancies (9)(10)(11). Although precise risk factors are not clearly listed, factors such as age, gender, race, and family history are often believed to contribute to the development of MM (12).
Signal transducer and activator of transcription 3 (STAT3) can be aberrantly activated in solid and haematological cancers including HCC and MM respectively (13)(14)(15)(16). A number of studies have demonstrated the critical role of STAT3 in malignant transformation and progression. STAT family comprises of seven members-STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b, and STAT6 (17)(18)(19)(20). STAT1, STAT3, and STAT5 are widely studied in the cancer context, among which STAT3 stands in the forefront (21)(22)(23). STAT3 is predominantly activated by IL-6 family cytokines transiently in non-diseased conditions. In the canonical signaling, the binding of IL-6 to its receptor results in the activation of gp130, and other nonreceptor tyrosine kinases (NRTKs) such as JAK, and Src (24). The phosphorylated NRTKs or some of the activated receptor tyrosine kinases (such as EGFR) phosphorylates STAT3 Y705 and activates it (25)(26)(27). The phosphorylated monomer undergoes dimerization with another monomer to translocate into the nucleus through speci c importin (28,29). STAT3-binding sites have been identi ed in the promoter region of several genes that are involved in cell growth and proliferation, in ammation, prosurvival, antiapoptosis, proangiogenic, and metastasis (30)(31)(32). The overexpression of STAT3 targeted genes due to its persistent activation in cancers provide enormous growth potential to the cancer cells and encourage the advancement of the disease. Therefore, mitigation of the STAT3 cascade may provide a signi cant clinical bene t to patients.
Tris(dibenzylideneacetone)dipalladium(0) (Tris DBA) is an organometallic complex compound where two palladium atoms are bound to the alkene units of three dibenzylideneacetone (33). The compound is moderately soluble in organic solvents. The antitumor potential of Tris DBA in few cancer models have been analysed previously. Bhandarkar and colleagues demonstrated that Tris DBA can reduce melanoma cell proliferation by inhibiting the activation of MAPK, Akt, STAT3, phospho-S6 kinase, and downregulated the expression of N-myristoyltransferase-1 (34). Tris DBA decreased proliferation and showed an additive cytotoxic effect with proteasome inhibitors such as bortezomib and car lzomib towards MM cells (35). In addition, Tris DBA regulated the expression of the Bcl-2 family proteins by targeting ribosomal protein S6 in primary chronic lymphocytic leukemia B-cells (36). Tris DBA has been proposed to target Nmyristoyltransferase 1 to induce growth-inhibitory and antimetastatic activity in pancreatic cancer cells (37). These ndings have established Tris DBA as a promising anticancer agent, although its precise mechanism of action is yet to be explored. In the present study, we have evaluated the effect of Tri DBA on STAT3 signaling in HCC and MM cells and its antitumor e cacy in preclinical settings.

MTT assay
The cytotoxic effect of Tris DBA towards MM and HCC cells was evaluated by the in vitro cytotoxicity assay using MTT dye as reported earlier (38,39).

Annexin V assay
Flow cytometric analysis was performed to determine the extent of apoptosis in MM and HCC cells using Annexin V and PI staining as described previously (40).

Cell cycle analysis using ow cytometry
The alignment of tumor cells in different phases of the cell cycle was analyzed using ow cytometry by staining with PI as reported earlier (41,42). The cell pellet was suspended in propidium iodide (2.5 µg/ml) and RNAse (0.5 mg/ml RNase A in PBS) after treatment and analyzed using ow cytometry.

Western blotting
The cells were incubated with Tris DBA for given time points at given doses and lysed using MPer lysis buffer (Thermo Scienti c) as described previously (43,44).

DNA binding assay
The effect of Tris DBA on DNA interaction ability of STAT3 was assessed using the TransAM assay kit (Active Motif, Carlsbad, CA,). The experiment was done as per the directions from the manufacturer as described previously (45).

Immunocytochemistry for STAT3 distribution
The cells were seeded in Ibidi glass-bottom dishes in complete medium and treated with Tris DBA or vehicle control for 2 h and then xed with 10% buffered formalin after washing with PBS. These cells were treated with Triton-X 100 for permeabilization followed by blocking with goat serum (5%) for 1h.
Thereafter, the preparation is treated with monoclonal mouse anti-STAT3 antibody (1:100 dilution) overnight at 4ºC followed by rinsing with PBS and incubated in goat anti-mouse secondary antibody (Alexa-Fluor 568) for 1h. The cells were then washed and counterstained with DAPI to stain the nuclei. The cells were analyzed immediately under a confocal microscope (Nikon A1).

RNA isolation and reverse transcription polymerase chain reaction (RT-PCR)
The cells were incubated with Tris DBA, washed, and suspended in Trizol reagent (Invitrogen, Carlsbad, CA) and reverse transcribed as discussed earlier (46).

Real-time quantitative PCR
Total RNA was isolated from the cells using Trizol reagent and subjected to real-time quantitative PCR analysis as reported previously (47).

Transfection with SHP2 siRNA
We used the Neon™ Transfection System (Invitrogen, Carlsbad, CA) to transfect cells with SHP2 siRNA. Brie y, cells and SHP2 siRNA (100ng/μl) were taken into a clean, dry, sterile centrifuge tube and following aspiration of cell-siRNA cocktail using Neon Tip in Neon Pipette to restrain the formation of air bubbles.
Thereafter, Neon Pipette was placed into the Neon Tube which has Neon Electrolytic Buffer E. Next cells were subjected to a pulse of 1,150 voltage with a width of 30 for transfection. 48h post-transfection, cells were incubated with Tris DBA and the cell lysate was prepared and analyzed for various proteins.

In vitro migration assay
This experiment was performed to determine the action of Tris DBA on migration as described before (48,49).

Invasion assay
This assay was carried out to assess the anti-invasive property of Tris DBA on HCC and MM cells using the Bio-Coat Matrigel invasion assay system (BD Biosciences, San Jose, CA) as elaborated upon previously (50).
In vivo acute toxicity studies All the in vivo acute toxicity experiments were carried out as per the approved procedures by the SingHealth Institutional Animal Use and Care Committee (protocol number: 2013/SHS/870). Eight-weekold NCr nude male mice (Invivos, Singapore) were used to perform toxicity-related experiments. The experimental animals were administered intraperitoneally with 0.1% DMSO or 25, 50, 100, and 200 mg of Tris DBA. The experimental animals were monitored regularly for any changes in feed and water consumption, body weight, behavior, physical appearance, movement, and other criteria up to the eighth day. On the last day, animals were sacri ced by cardiac puncture and blood was collected for further analysis for biochemical analysis.

Xenograft MM mouse model
In vivo xenograft tumor experiments were carried out as per the approved procedures by KHU Institutional Animal Care and Use Committee [KHUASP(SE)-17-110]. The subcutaneous xenograft MM mouse model was established by injecting U266 cells to the right ank of six-week-old athymic nu/nu female mice (NARA Biotech, Korea). The mice were randomly distributed into three groups (n=6/group) upon tumor reaching the size of 0.25 cm. The rst group served as control which was intraperitoneally administered with PBS (200 μL, i.p. thrice/week). The second and third groups intraperitoneally received Tris DBA (50 mg/kg, thrice/week) and (100 mg/kg, thrice/week) respectively for four weeks. All mice were sacri ced after one week and primary tumors were collected for subsequent analysis.

Immunohistochemical analysis of MM tumor samples
The tumors collected from xenograft experiments were processed using phosphate-buffered formalin (10%) and impregnated in the para n blocks. Thereafter, they were subjected to cutting, depara nization using xylene and dehydration using graded alcohol. The slides were boiled in sodium citrate (10 mM, pH 6.0) for 30 min to retrieve antigens. This is followed by immunohistochemistry using the kit obtained from Vector Laboratories (ImmPRESS TM Reagent Kit). In brief, hydrogen peroxide (3%) and blocking reagent were used to quench tissue-derived peroxidases and to block non-speci c interactions. Further, sections were treated with various antibodies for overnight. The next day, slides were repeatedly rinsed with PBS and incubated with ImmPRESS TM reagent. 3, 3-diaminobenzidine tetrahydrochloride (DAB) reagent was used as a substrate to identify immunoreactive species. Next, Gill's hematoxylin was used to counterstain the sections followed by the photography of sections using the Olympus microscope (20x).
Positive cells appeared brown and quanti ed using the Image-Pro Plus 6.0 software package (Media Cybernetics, Inc.)

Orthotopic HCC tumor model
In vivo orthotopic tumor experiments were carried out as per the approved procedures by the SingHealth Institutional Animal Use and Care Committee (protocol number: 2013/SHS/870) and as reported earlier (51,52). The orthotopic HCC mouse model was established by placing a small piece of tumor-derived from HCCLM3-Luc cells. The mice were randomly distributed into three groups upon the bioluminescence signal reaching 10 6 . The rst group served as control which was intraperitoneally administered with 0.1% DMSO (thrice/week). The second and third groups intraperitoneally received Tris DBA (50 mg/kg, thrice/week) and (200 mg/kg, thrice/week) respectively for four weeks. The tumor progression/regression was recorded weekly twice by quantifying the bioluminescence signals. Experimental animals were sacri ced by CO 2 inhalation. Tumor tissues were harvested and stored at -80 o C for further processes.

Statistical analysis
The obtained numerical data are presented as the mean ± SD unless otherwise mentioned. Student's ttest or one-way ANOVA were employed to measure statistical signi cance.

Results
Tris DBA decreased the viability and increased the subG1 population of HCC and MM cells.
We deciphered the action of Tris DBA on the viability of HCC and MM cells by MTT assay. For this, we used HCC (HCCLM3, Huh7, HepG2), MM (U266), and bortezomib-resistant/sensitive MM (RPMI-8226) cell lines. All the cells responded to Tris DBA, which resulted in a signi cant decrease of viability thus indicating the cytotoxic effect of Tris DBA on both types of cancer cells (Fig. 1B). Flow cytometric analysis revealed that the treatment with Tris DBA signi cantly increased the subG1 cell population up to 80% in HCC and MM cells (Fig. 1C). However, the subG1 cell population is relatively less (30%) in bortezomib-resistant RPMI-8226 cells.
Tris DBA induced apoptosis of HCC and MM cells.
We performed FITC-Annexin V-PI staining to verify the action of Tris DBA is inducing apoptosis and observed ta dose-dependent elevation in the percentage of apoptotic cells in all the tested cell lines ( Fig. 2A). The percentage of the apoptotic population in bortezomib-resistant MM cells were relatively lower than the sensitive counterpart.
Tris DBA altered STAT3 phosphorylation in tumor cells.
Initially, we examined the effect of Tris DBA on activation of STAT3 in HCC and MM cells. Tris DBA signi cantly reduced the constitutively activated STAT3 in U266 and HCCLM3 cells (Fig. 1B). RPMI-8226 has a basal level of phospho-STAT3 Y705 expression and treatment of these cells with IL-6 triggered the phosphorylation of STAT3 Y705 in time-and dosage-dependent manner (Fig. 1C, upper panels). The treatment with Tris DBA resulted in suppression of IL-6 driven STAT3 activation without a change in the total STAT3 expression (Fig. 1C, lower panel).
Tris DBA decreased nuclear localization and ablated the binding of STAT3.
Next, we deciphered the action of Tris DBA on the cellular distribution of STAT3 in HCCLM3 cells. Tris DBA downregulated the nuclear pool of STAT3 in HCCLM3 cells compared with vehicle-treated samples as demonstrated by the results of immunocytochemistry (Fig. 3A). Further, we prepared the nuclear extract of DMSO/Tris DBA treated cells and analyzed whether Tris DBA also in uences the DNA interaction ability of STAT3. We observed a signi cant reduction in the DNA interaction ability of STAT3 in different tumor cells (Fig. 3B). We also observed an inhibition of IL-6 triggered DNA binding of STAT3 in RPMI-8226 cells (Fig. 3B), thus indicating that Tris DBA may interfere with STAT3 activation and subsequent gene transcription.
Tris DBA inhibited the phosphorylation of upstream kinases of STAT3 signaling.
Next, we measured the potential of Tris DBA treatment on the phosphorylation of upstream kinases regulating STAT3 signaling such as JAK1, JAK2, and Src in HCC and MM cells. We observed a signi cant decline in the constitutive activation of JAK1, JAK2, Src in U266 cells, and JAK2, Src in HCCLM3 cells (Fig. 4A). We also investigated the effect of Tris DBA on IL-6-induced phosphorylation of JAK1 and JAK2 in RPMI-8226 cells. We observed a substantial reduction in the IL-6 induced activation of JAK1 and JAK2 (Fig. 4B), thereby indicating that Tris DBA suppresses the activity of upstream kinases to modulate STAT3 functions.
Tris DBA altered the expression of SHP2.
We further determined the action of pervanadate on Tris DBA induced STAT3 inhibition in MM cells. The exposure to pervanadate reversed the Tris DBA induced STAT3 inhibition thereby indicating the involvement of protein tyrosine phosphatases (PTPs) (Fig. 4C). The action of Tris DBA on the cellular levels of major PTPs such as SHP1, SHP2, and PTP1B was measured. Interestingly, Tris DBA elevated the protein levels of SHP2 in a dosage-and time-dependent fashion (Fig. 4D). We also observed that treatment with Tris DBA increased the mRNA expression of SHP2, thus indicating that this protein may be regulated at the transcriptional level (Fig. 4E). However, a substantial change was not noted in the expression of SHP1 and PTP1B proteins (Fig. 4F).
Depletion of SHP2 reverses the Tris DBA induced STAT3 inhibition.
We transfected U266 cells with siRNA direct towards SHP2 to study the speci city of Tris DBA towards the STAT3 signaling pathway. The treatment of Tris DBA alone decreased STAT3 phosphorylation with a corresponding increase in SHP2. The deletion of SHP2 using siRNA resulted in the restoration of STAT3 phosphorylation and Tris DBA did not affect the STAT3 activation in SHP2-depleted cells (Fig. 4G). The cells treated with scrambled siRNA served as control. These results substantiated the role of SHP2 in mediating the STAT3 inhibitory actions of Tris DBA.
Tris DBA altered the expression oncogenic proteins.
We determined the action of Tris DBA on the expression of STAT3 driven proteins in MM (Fig. 5A) and HCC cells (Fig. 5B). Consistent with inhibition of STAT3 phosphorylation, Tris DBA downregulated the expression of oncogenic factors such as cyclin D1, Bcl-2, Mcl1, survivin with an upregulation of proapoptotic factor such as Bak protein in the tested cell lines. Furthermore, Tris DBA also promoted the cleavage of caspase-3 and PARP thus, indicating that cells are undergoing apoptosis (Figs. 5A and 5B).
Tris DBA signi cantly modulated the migratory and invasive potential.
Next, we measured the outcome of Tris DBA on cellular motility using invasion and migration assay systems. Tris DBA signi cantly suppressed the invasion of U266, RPMI-8226 (sensitive and resistant), and HCCLM3 cells in a concentration-dependent fashion (Fig. 6A). In addition, signi cant anti-migratory effects were observed in Tris DBA treated HCCLM3 cells (Fig. 6B).
Tris DBA did not display toxicity in preclinical studies.
We next determined the sub-lethal doses of Tris DBA to perform in vivo tumor studies by measuring its toxicity. For this, male NCr nude mice were injected with Tris DBA (25, 50, 100, and 200 mg/kg/day) for eight days and experimental animals were regularly monitored. The mice-group which received DMSO (0.1%) served as control. We did not observe signi cant changes in Tris DBA treated group of animals relative to the DMSO treated group in terms of body weight, water and feed consumption (Fig. 7). Notably, no signi cant alterations were observed in the activity of liver function marker enzymes such as alanine aminotransferase (ALT), and aspartate aminotransferase (AST) (Fig. 7), thereby suggesting that Tri DBA did not exhibit signi cant toxicity in tested animals.
Tris DBA imparted antitumor effect in xenograft MM and orthotopic HCC mice models.
We investigated the anticancer activity of Tris DBA in xenograft MM and the orthotopic HCC mice models. We subcutaneously implanted U266 cells to the right ank of six weeks old athymic nu/nu female mice to establish the xenograft tumor model. When tumor size attained 0.25 cm, the mice were randomly divided into three groups (n = 6). The rst group served as a control. The second and third groups received an intraperitoneal injection of Tris DBA (50 mg/kg, thrice a week; and 100 mg/kg, thrice a week, respectively). At the end of four weeks, we observed a signi cant reduction in the volume of tumors in the 100 mg/kg treated group (p < 0.01) (Fig. 8A).
To verify the antitumor effect of Tris DBA on solid tumors, we next established an orthotopic HCC mouse model as described in methods and randomly distributed into three groups. The rst group was used as control. The second and third group animals were intraperitoneally injected with Tris DBA (50 mg/kg, thrice a week; and 200 mg/kg, thrice a week, respectively) for four weeks and the tumor burden was measured by a non-invasive technique of bioluminescence. Tris DBA signi cantly inhibited the tumor growth in both the groups (p = 0.0007) (Fig. 8B).
Tris DBA reduced the number of Ki67 + cells in tumor tissues from the xenograft MM mouse model. We next processed the MM tumor tissue samples and performed immunohistochemical analysis to determine the proliferation index. Tumor tissues from Tris DBA treated mice displayed a marked reduction in the levels of Ki-67 + cells in a dose-dependent fashion (Fig. 9A). These results suggest that Tris DBA may also induce its antitumor activities by reducing the proliferation of cancer cells.
Tris DBA repressed the activation of STAT3 signaling proteins in tumor tissues.
We further analyzed the phosphorylation status of STAT3, JAK1, and JAK2 proteins in the tumor tissues of the xenograft MM mouse model. We observed a substantial reduction in the phosphorylation of STAT3, JAK1, and JAK2 in the tumor tissues of mice treated with 100 mg/kg of Tris DBA (Fig. 9B). These results establish that the lowering of tumor burden by Tris DBA could be due to the abrogation of the STAT3 signaling pathway.
Tris DBA modulated the expression of apoptosis-related proteins.
We pro led the apoptosis-related proteins (Bcl-2, Bcl-xL, Survivin, PARP, and procaspase 3) in tumor tissues of the xenograft MM mouse model. We observed a marked decrease in Bcl-2, Bcl-xL, Survivin, fulllength PARP and procaspase-3 (Fig. 9C), thus indicating that Tris DBA can interfere with tumor development by inducing apoptosis in tumor cells.

Discussion
Aberrant activation of STAT3 is often observed in numerous human malignancies including MM, leukemias, lymphomas, breast, ovary, melanoma, glioma, head and neck, liver, and kidney cancers (53)(54)(55). A few studies have demonstrated that deregulated STAT3 activation has not been observed in the normal tissues adjacent to tumors (56,57). Additionally, the expression of STAT3 in tumor tissue serves as a predictive marker of prognosis and elevated STAT3 expression is associated with reduced prognostic ability and worse 3-year overall survival rate of human solid tumors (58)(59)(60)(61). The constitutive activation of STAT3 can be controlled by various mechanisms including the loss of negative regulators (PTPs, SOCS, and PIAS), the formation of positive feedback mechanisms in the tumor microenvironment, elevated activity of upstream kinases, and somatic mutations in the STAT3 gene (62). For instance, elevated levels of IL-6 is produced due to positive feedback mechanism in the tumor microenvironment, which may regulate the persistent phosphorylation of STAT3 signaling in human cancers (63)(64)(65).
Here we have evaluated the action of Tris DBA on the STAT3 signaling in HCC and MM cell-based studies and preclinical settings. Tris DBA was initially described as an inhibitor of N-myristoyltransferase 1 and here we have comprehensively demonstrated its mode-of-action in two different tumor models. Tris DBA decreased the viability of both types of cancer cell lines (HCC and MM) and caused a pronounced inhibition of STAT3 activation. DNA undergoes fragmentation in the cells that are committed to apoptosis due to activation of caspases and thus they are known as hypodiploid cells, which are marked as subG1 cell population (66,67). We observed an increased amount of subG1 cell population on treatment with Tris DBA in the owcytometric analysis. The pro-apoptotic actions of drug were further con rmed by FITC conjugated Annexin-V-PI staining. To understand the underlying molecular mechanism of Tris DBA induced cytotoxicity, we evaluated the potential of Tris DBA to modify STAT3 phosphorylation in tumor cells and noticed a marked decline in the activation of STAT3 in all the tested cell lines.
An enormous number of studies have reported that STAT3 phosphorylation is essential for its translocation into the nucleus to express the target genes (68)(69)(70)(71). In contrast to our ndings, Kay and colleagues have demonstrated that Tris DBA did not signi cantly inhibit the activation of STAT3 in primary chronic lymphocytic leukemia B-cells (36). However, we hypothesized that the cytotoxic e cacy of Tris DBA could be due to the alteration of STAT3 signaling cascade. We next evaluated the effect of Tris DBA on IL-6 stimulated STAT3 phosphorylation. RPMI-8226 cells lack persistently activated form of STAT3 and IL-6 can function to promote the activation of STAT3 signaling (46,72). Bone marrow microenvironment contains several growth factors along with IL-6 released by bone matrix, and osteoblasts creating an ideal for promotion of oncogenesis in various cancers including MM (73). Therefore, it is apt to have an inhibitor that can counteract the effect of IL-6 induced signaling. We observed a signi cant decrease in IL-6 driven STAT3 phosphorylation in MM cells.
Phosphorylation of STAT3 Y705 is crucial for its translocation to the nucleus to transcribe the downstream genes. The substitution of tyrosine with phenylalanine using site-directed mutagenesis can cause a substantial decrease of PIAS3-STAT3 complex and their nuclear localization despite EGF-induction (74). Tris DBA downregulated the levels of nuclear STAT3 and suppressed the DNA binding ability of nuclear STAT3. This could be due to the inhibition of STAT3 phosphorylation, which is essential for nuclear localization and target gene expression. We next determined the action of Tris DBA on the activation status of STAT3 upstream NRTKs such as JAK1, JAK2, and Src in MM and HCC cells. Phosphorylation of these kinases was inhibited in both types of cancer cells. Moreover, cytokine-induced activation of these NRTKs was also downregulated indicating that abrogation of STAT3 phosphorylation may be caused through the inhibition of upstream signaling partner proteins.
The phosphorylation state of STAT3 signaling proteins can also be regulated by a class of protein tyrosine phosphatases (PTPs) including SHP1, SHP2, PTPεC, PTP1B, and CD45 (75)(76)(77). Therefore, we treated U266 cells with Tris DBA and pervanadate, a broad range phosphatase inhibitor and tested for the activation of STAT3. Interestingly, pervanadate treatment reversed the Tris DBA driven STAT3 inhibition.
This indicated that PTP is involved in the Tris DBA induced STAT3 signaling inhibition. We again treated U266 cells with pervanadate and Tris DBA and analyzed the expression of major PTPs such as SHP1, SHP2, and PTP1B. We observed a substantial enhancement in the levels of both SHP2 mRNA and protein. The treatment of U266 cells with SHP2-directed siRNA resulted in the loss of SHP2 protein and further treatment of these cells with Tris DBA displayed no effect on STAT3 phosphorylation. However, further investigation may be required to understand the relationship between the expression of SHP2 and Tris DBA in MM cells. Not surprisingly, SHP2 suppressed the proliferation of esophageal squamous cell cancer by inhibiting STAT3 phosphorylation and depletion of SHP2 resulted in the attenuation of cisplatin sensitivity (78).
Bortezomib is the rst proteasome inhibitor approved by the US FDA for the treatment of various malignancies (79). Li and colleagues reported that bortezomib can increase the expression of phosphorylated STAT3 and STAT3 in head and neck squamous cell carcinoma cells, which may contribute to a substantial reduction in its therapeutic e cacy, and co-treatment of bortezomib with a STAT3 inhibitor has been suggested for better anticancer effects (80). Guan  We observed a marked cleavage of procaspase-3 and downstream effector PARP indicating that cells are undergoing apoptosis on treatment with Tris DBA. STAT3 can also regulate the activation of key transcription factors such as Snail and Twist which mediate epithelial-mesenchymal transition (EMT) process, and in turn, metastasis of cancer cells to the distant organs (84). IL-6 has been reported to activate STAT3 to promote EMT through induction of Snail expression in cancers (85,86). Phosphorylated STAT3 also triggered the expression of the Twist gene to promote oncogenic functions (87). Next, we determined the e cacy of Tris DBA on invasion and migration of cancer cells. Tris DBA signi cantly blocked the cell motility in both the assay systems. The inhibition of migration and invasion could be due to suppression of the STAT3-dependent EMT process.
Interestingly, Tris DBA was observed to be non-toxic (up to 200 mg/kg body weight) in acute toxicity studies conducted in mice. Next, we examined its antitumor potential in a xenograft MM and an orthotopic HCC mouse models. Tris DBA displayed signi cant antitumor activities and it effectively abrogated the tumor progression in both the preclinical models. It has been also reported that Tris DBA did not induce local nor systemic toxicity at 40 mg/kg/day in the A375 melanoma mouse model (34). Importantly, Tris DBA downmodulated the expression of activated JAK1, JAK2 and STAT3, and Ki-67 in MM tumor tissues. These results are consistent with the outcome of our various cell-based assays as summarized in Fig. 10.

Conclusions
Tris DBA has been demonstrated as a potent inhibitor of STAT3 signaling cascade in MM and HCC cell lines and preclinical models for the rst time. It can be reconciled that Tris DBA is a broad-spectrum anticancer agent and further research may be warranted to completely understand its off-targets and its possible potential clinical application.   (B) U266 and HCCLM3 cells were exposed to Tris DBA at indicated concentrations and time points and cell lysates were used for SDS-PAGE and Western blotting to analyze phospho-STAT3/STAT3 expression.
(C) RPMI-8226 cells were incubated with Tris DBA (25 µM) at given time points and induced with IL-6 (10ng/ml). The cell lysates were used for SDS-PAGE and Western blotting to analyze phospho-STAT3/STAT3 expression.

Figure 4
Tris DBA suppresses upstream proteins and enhances the expression of SHP2. (A) U266 and HCCLM3 cells were exposed to Tris DBA (25 and 50 µM respectively) and cell lysates were used to analyze phosphorylated NRTKs and total proteins expression. (B) RPMI-8226 cells were exposed to Tris DBA (25 µM) at indicated time points and stimulated with IL-6 (10 ng/ml). The cell lysates were used to analyze phosphorylated NRTKs and total protein expression. (C) U266 cells were treated with Tris DBA (25 µM) and pervanadate at indicated doses to analyze phosphorylated STAT3 and total STAT3 expression.     were subcutaneously injected into xenograft tumors. After tumors reaching 0.25 cm in diameter, the mice were divided into three groups (n = 6/group) and administered with 50 mg or 100 mg/kg body weight.
Tumor volume is monitored throughout the study tenure. (B) HCCLM3-Luc cells-induced tumors are orthotopically implanted to the liver tissue followed by treatment with 0.1% DMSO (n=7) or Tris DBA (n=7) (administered 50mg/kg or 200mg/kg intraperitoneally, thrice a week, for four weeks). The tumor progression/regression was monitored twice a week by quantifying the bioluminescence intensity and the graph is plotted using these values. The scattered plot indicates the tumor burden that was quanti ed by measuring photon counts before the rst administration of Tris DBA and at the last dose. Unpaired t-test with Welch's correction. Bars indicate standard deviation; **p<0.01.

Figure 9
Tris DBA decreases the activation of STAT3 and its regulatory proteins as well as the levels of its