Blockade of Interleukin-6 (IL-6) Signaling in Dedifferentiated Liposarcoma (DDLPS) Decreases Mouse Double Minute 2 (MDM2) Oncogenicity via Alternative Splicing

Effective therapies for retroperitoneal (RP) dedifferentiated liposarcoma (DDLPS) remain unavailable. Loco-regional recurrence occurs in >80% of cases; 5-year disease-specific survival is only 20%. DDLPS is especially prevalent in the retroperitoneum and abdomen; evaluation of the DDLPS microenvironment in these high-fat compartments appears pertinent. Adipose is a main supplier of interleukin-6 (IL6); excessive activation of IL6 signal transducer glycoprotein 130 (GP130) underlies the development of some diseases. The role of GP130 pathway activation remains unstudied in DDLPS, so we examined the role of microenvironment fat cell activation of the IL6/GP130 signaling cascade in DDLPS. All DDLPS tumors and cell lines studied expressed elevated levels of the GP130-encoding gene IL6ST and GP130 protein compared to normal tissue and cell line controls. IL6 increased DDLPS cell growth and migration, possibly through increased signal transducer and activator of transcription 1 (STAT1) and 3 (STAT3) activation, and upregulated mouse double minute 2 (MDM2). GP130 loss conveyed opposite effects; pharmacological blockade of GP130 by SC144 produced the MDM2 splice variant MDM2-ALT1, known to inhibit full length MDM2 (MDM2-FL). Although genomic MDM2 amplification is pathognomonic for DDLPS, mechanisms driving MDM2 expression, regulation, and function beyond the MDM2:p53 negative feedback loop are poorly understood. Our findings suggest a novel preadipocyte DDLPS-promoting role due to IL6 release, via upregulation of DDLPS MDM2 expression. Pharmacological GP130 blockade reduced the IL6-induced increase in DDLPS MDM2 mRNA and protein levels, possibly through enhanced expression of MDM2-ALT1, a possibly targetable pathway with potential as future DDLPS patient therapy.


Introduction
Dedifferentiated liposarcoma (DDLPS) is a high grade tumor of unknown etiology, characterized by wildtype TP53 and genomic amplification of the p53 regulator MDM2.Surprisingly, targeted anti-MDM2 therapy displays modest impact with significant hematological toxicity (1,2), rendering radical surgery coupled with non-targeted toxic chemotherapies the mainstay treatment strategy, which often results in quality of life consequences (3,4).These treatment approaches have not improved 5-and 10-year disease-free survival rates of 20% and 10%, respectively, which have stagnated since the 1970s (5)(6)(7).Although little is known about the DDLPS tumor microenvironment (TME), we identified a tumor-promoting paracrine loop linking DDLPS cells and tumor-associated macrophages (TAMs) (8).DDLPS cells were observed to release microRNAs (miR) 25-3p and 92a-3p in extracellular vesicles which then interacted with macrophage toll-like receptors 7 and 8, inducing elevated IL6 secretion and consequently enhancing DDLPS growth, migration, and invasion (8).
Studies evaluating the impact of IL6 signaling on DDLPS MDM2 expression per se or MDM2 splice variant production are lacking.Although full length MDM2 is the most commonly expressed form (25), truncated variants produced by alternative mRNA splicing have been shown to be generated in response to genotoxic stress (26)(27)(28).MDM2-ALT1 splicing variant was found to be upregulated upon ultra-violet (UV) or cis-platinum treatment in osteosarcoma as well as breast and lung carcinomas bearing wild-type or deleted TP53, respectively, suggesting that MDM2-ALT1 production may be independent of TP53 status (26).To our knowledge, no studies have evaluated the mechanisms driving alternative DDLPS MDM2 splicing variant production.Given the genomic amplification of MDM2 in DDLPS as well as the general lack of knowledge regarding DDLPS MDM2 regulation and alternative splicing, we evaluated the impact of IL6 signaling and blockade on full length and truncated MDM2 expression.

Cell growth analysis
Cell growth endpoint analyses were performed using CellTiter96 Aqueous Non-Radioactive Cell Proliferation MTS Assay kit (Promega, Madison, WI, USA) per manufacturer specifications.To assess the impact of IL6 on DDLPS growth, tumor cells were seeded at a density of 2,000 cells per well in a 96-well plate in complete culture media The following day, cells were serum-starved by the removal of complete media and washing with PBS (Life Technologies, Waltham, MA, USA) prior to the addition of 100 μL plain DMEM to each well for 24h.On day 3, all media was replaced with fresh media containing plain DMEM with PBS (control) or 10 ng/mL IL6 .After 96h, 20 μL of MTS was added to each well and cells were incubated at 37˚C for 2h; absorbance was measured at 490 nm wavelength (Beckman Coulter, Indianapolis, IN, USA).
The effect of PreAdip-derived conditioned media (CM) on cancer cell growth was similarly evaluated.DDLPS cells were seeded at a density of 2,000 cells per well in a 96-well plate and 2.0 x10 6 PreAdip cells were seeded in a 10 cm dish in complete media.On day 2, the PreAdip media was removed, cells were washed twice with PBS, and fresh 5 mL DMEM supplemented with 0.5% bovine serum albumin (BSA, Amresco, OH, USA) was added to each plate.Also on day 2, the media was removed from the DDLPS wells, cells were washed with PBS and serum starved for 24h in plain DMEM.The following day, the PreAdip CM was collected and centrifuged (3000 rpm, 5 min, RT) .Media was removed from the well containing DDLPS cells, and was replaced with 100 μL PreAdip-derived CM.Cells within the control group received 100 μL plain DMEM supplemented with 0.5% BSA.After 96h, endpoint assessment was performed using MTS .
The effect of IL6 or anti-IL6 monoclonal antibody MAB206 on the pro-growth impact of PreAdip CM on DDLPS cells was assessed using 2,000 cells per well.After adherence, cells were serum starved for 24h prior to the addition of plain DMEM supplemented with 0.5% BSA in the absence or presence of CM, 10 ng/mL IL6, 0.6 μg/mL MAB206, CM with 0.6 μg/mL MAB206, or CM with 10 ng/mL IL6.After 96h, cellular growth was measured using MTS reagent .
To assess the anti-DDLPS effects of SC144, cells were seeded at a density of 4,200 cells per well and allowed to adhere overnight.Treatment commenced the following day, and cellular growth was assessed 96h after treatment with DMSO (control), IL6 (10ng/mL), or increasing doses of SC144 (Sigma-Aldrich, St. Louis, MO, USA) in the presence of IL6 (10ng/mL) by MTS

Gene knockdown
Cells seeded at a density of 1.75x10 5 cells per well in a 6-well plate were transfected with 40 nM of pooled anti-IL6ST (cat# s7317, s7318, s7319; Thermo Fisher Scientific) or non-targeting Silencer Select siRNA (Thermo Fisher Scientific) in Opti-MEM media (Life Technologies) with Xtreme Gene transfected reagent (Sigma-Aldrich).The control cells were incubated only with transfection reagent and Opti-MEM.

ELISA
Cells were seeded at a density of 1.75x10 5 cells per well of a 6-well dish and, on the following day were washed once with PBS and media was replaced with fresh DMEM supplemented with 0.5% BSA.After 72h, supernatant was collected and centrifuged (3000 rpm, 5 min, RT) and Secreted IL6 was measured using the Quantikine® ELISA kit (R&D Systems), following manufacturer's instructions.

Complementary DNA synthesis and quantitative real-time PCR
Complement DNA (cDNA) synthesis was performed using TaqMan® Reverse Transcription Reagents (Life Technologies) and analyzed by quantitative real-time PCR (qRT-PCR) using TaqMan® Fast Advanced Master Mix ( Thermo Fisher Scientific) and MDM2 TaqMan probe (Hs01066930_m1) or IL6ST TaqMan probe (Hs00174360_m1) (Life Technologies).Relative gene expression levels were normalized to GAPDH expression (TaqMan probe cat# Hs02758991_g1).cDNA was also generated using iScript™ Reverse Transcription Supermix (Bio-Rad Laboratories, Hercules, CA, USA) per manufacturer's specifications and assessed by qRT-PCR using SYBR® Green PCR Master Mix (Thermo Fisher Scientific).Primer sequences are listed below, relative gene expression was normalized to GAPDH.

IL6 (F: CACTCACCTCTTCAGAACCAAT
The StepOnePlus Real-Time PCR System (ThermoFisher Scientific) was used to execute the quantitative real-time PCR reactions.

Western blot
Protein extract was isolated from cells using 1x Cell Signaling Lysis Buffer (Cell Signaling, Danvers, MA, USA), supplemented with 1x Halt™ Protease Inhibitor Cocktail EDTA-free (Thermo Fisher Scientific).
For western blot analysis, 25 μg protein were subjected to electrophoretic run using a 4-15% Mini-PROTEAN® TGX™ Precast Protein Gels (Bio-Rad Laboratories), transferred to PVDF membranes using the Trans-Blot® Turbo™ Midi PVDF Transfer Packs (Bio-Rad Laboratories) and dried for at least 1h.
Membranes were incubated overnight at 4°C with the antibodies listed below.Primary antibody mixtures were removed, and membranes were washed and then incubated with secondary donkey anti-rabbit or donkey anti-mouse (IRDye 800CW) and/or donkey anti-goat, donkey anti-mouse, or donkey anti-rabbit (IRDye 680RD) (Li-Cor Biosciences, Lincoln, NE, USA) antibody for 1h.Membranes were washed and dried.Imaging and protein expression analyses were performed with the Odyssey CLx imaged (Li-Cor).

Migration assay
Cellular migration was assessed using 8μm trans-well ThinCerts™ migration chambers (Geiner Bio-One, Kremsmünster, Austria).DDLPS cells were seeded to 60% confluence in a 10 cm dish in complete culture media.The following day, cells were treated with PBS, 10 ng/mL IL6, DMSO, or 5 μM SC144 plus 10 ng/mL IL6 for 24h.The next day, treated cells were quantified and 300,000 viable cells were seeded into the upper chamber of the migration chamber in 200 μL plain DMEM.Antibiotic-free DMEM with 5% FBS served as chemoattractant.Endpoint staining was performed by fixing cells in 0.5% crystal violet in ethanol (Fisher Scientific) for 1h with gentle rotation, then washing the migration chambers gently three times with double-distilled water.Chambers were imaged after drying overnight or longer.

Indirect and direct co-culture analyses
The pro-DDLPS impact of culturing PreAdip cells with DDLPS cell lines was determined by both direct and indirect co-culture.Co-culture assays in which PreAdip cells were cultured directly with DDLPS cells were performed using red fluorescent protein (RFP)-tagged DDLPS cells.2,000 PreAdip cells were cultured into the basement well of a 96-well plate in 100 μL complete DMEM.The following day, DMEM + 0.5% BSA was added to each PreAdip well.2,000 RFP-tagged DDLPS cells were re-suspended in DMEM + 0.5% BSA and added to each well housing PreAdip cells (co-culture).Cells were observed for 5 days and viable cells were quantified using the IncuCyte Zoom live-imaging fluorescent microscope system.The total fluorescent signal per each well indicates the number of viable, fluorescently-tagged DDLPS cells in that respective well.Indirect co-culture experiments were performed by culturing DDLPS cells with PreAdip cells in 6-well Boyden 0.4 μm pores chambers (Corning Costar, Corning, NY, USA)).These experiments were performed as follows: On day 1, tumor cells were plated at a density of 2.0 x 10 5 cells/well into the basement well of a 6-well plate in 2 mL complete media (DMEM + 10% FBS + 1%P/S).Also on day 1, 10 5 PreAdip cells were seeded into the insert bearing the perforated membrane, and placed into a second 6-well plate containing complete PreAdip media.The following day, the wells with the DDLPS cells and the inserts with PreAdip cells were washed prior to the addition of plain DMEM supplemented with 0.5% BSA for 24h.On day 3, all media was aspirated and replaced with fresh DMEM plus 0.5% BSA, and the inserts bearing the PreAdip cells were placed into the wells with the adherent DDLPS cells.Cancer cells cultured in the absence of PreAdip cells in plain DMEM supplemented with 0.5% BSA served as the monoculture control for each respective DDLPS cell line.Endpoint analysis of DDLPS proliferation was performed at day 5, by quantifying viable cells count using trypan blue exclusion.

Cell viability
DDLPS cell viability in response to SC144 treatment was assessed using RFP-tagged DDLPS cells using the IncuCyte Zoom system detailed above.In short, 4,200 cells per well were plated and treated with increasing doses of SC144 in the presence of IL6 (10 ng/mL) and were monitored by live-imaging fluorescent microscopy over the course of 96h.Quantification of RFP-fluorescing cells indicated live DDLPS cells.

Flow cytometry
Apoptosis was assessed using the Apoptosis Detection Kit (BD Pharmagen, San Diego, CA, USA).

Nested PCR for MDM2 splicing variant assessment
RNA was extracted from Lipo246 and Lipo815 cell lines and Reverse transcription (RT) reactions were carried out using Transcriptor RT enzyme (Roche Diagnostics, Basel, Switzerland).Nested PCR was performed using a 2-step PCR reaction as previously described (31).

Immunohistochemistry analysis
Antibodies GP130 (cat# ab170257) and IL6 (cat# ab6672), Abcam, Cambridge, United Kingdom, were used for immunohistochemistry (IHC) assessment at a dilution of 1:250.IHC staining was conducted at Nationwide Children's Hospital (Columbus, OH, USA) and the Ohio State University College of Veterinary Medicine, and analysis was performed at the Polaris Innovation Center (The Ohio State University, Columbus, OH, USA).

Statistics
Mean ± SEM values were generated for all cell-based assays, and statistical analyses and EC50 values were computed using GraphPad Prism version 6.00 (GraphPad Software, La Jolla California USA, www.graphpad.com).All Student t tests performed were unpaired and two-sided, with Welch's correction applied to the analysis of patient IL6ST mRNA expression.Analyses across multiple cell lines were performed by one-way ANOVA.
During the investigation of the in vivo impact of SC144 on tumor growth, mean ± SEM was calculated for each treatment group, and end-point analyses determined using unpaired two-sided t test was utilized to determine variances.The average tumor volume (mm 3 ) and weight (g) for each study arm was measured and recorded.Significance was set at *p ≤ 0.05 for all in vitro and in vivo experimentation.

GP130 is overexpressed in DDLPS and requires an extracellular cytokine signal for activation
To determine if GP130 signaling facilitated DDLPS progression, expression of GP130 in DDLPS tumors (S1 Table ) and cell lines (S2 Table ) was examined.IL6ST (the gene that encodes GP130 protein) and GP130 measurements revealed consistently high DDLPS tumor and cell line expression (Figs 1A-D

GP130 signaling increases DDLPS MDM2 expression
Given the consistent overexpression of MDM2 in DDLPS tumors, we considered possible GP130 impacts on MDM2.DDLPS cells were cultured under serum-deprived conditions followed by IL6 treatment.

GP130 blockade by SC144 reduces DDLPS growth in vitro and in vivo
Our findings suggested potentially protumorigenic GP130 roles in DDLPS, so we investigated GP130 as a possible therapeutic target.MTS analysis of DDLPS treatment with anti-GP130 agent SC144 reduced

SC144 has GP130-specific effects on STAT1 and STAT3 activation, and reduces DDLPS MDM2 expression
Considering the GP130-specific effects of SC144 (24,38), we asked if SC144 would affect DDLPS GP130 signaling, perhaps via STAT1 and STAT3 activation.Interestingly, we observed a shift in GP130 expression from the higher to the lower molecular weight band, suggesting protein deglycosylation consistent with possible GP130 inactivation (Fig 7A) (24).Levels of pSTAT1 and pSTAT3 decreased

Discussion
To the best of our knowledge, we are the first to show a tumor-promoting paracrine loop in which DDLPS cells release microRNAs 25-3p and 92a-3p in extracellular vesicles which then interact with toll-like receptors 7 and 8 in macrophages, inducing elevated IL6 secretion.Consequently, we evaluated GP130 signaling in DDLPS, finding that IL6-rich TAM-derived conditioned media treatment enhances DDLPS growth, migration, and invasion (8).Here we demonstrate that PreAdip cells also release elevated levels of IL6, perhaps also acting to promote DDLPS.Signaling facilitated by IL6 is mediated by GP130, which is elevated in DDLPS patient samples and cell lines.Moreover, activation of GP130 by IL6 enhanced the oncogenic phenotype of DDLPS (i.e., increased growth and migration), perhaps by activating of downstream STAT1 and STAT3 effector proteins.Based on these findings, we then demonstrated that GP130 activation by IL6 increases pSTAT1 and pSTAT3 levels, leading to increased MDM2 expression.Detection of elevated DDLPS MDM2 copy number is a basis for disease diagnosis; however, mechanisms regulating its oncogenic expression are poorly understood.Our studies show that IL6 activation of GP130 results in elevated MDM2 expression whereas IL6ST KD results in reduced MDM2 levels, suggesting that DDLPS MDM2 expression is partially regulated by GP130 signaling.Upregulation of DDLPS MDM2 by GP130 signal transduction may be facilitated by STAT1 and STAT3 transcription factors, as seen in colorectal cancer (37) where GP130 activation by IL6 cytokine family member leukemia inhibitory factor led to increased MDM2 levels in a STAT3-dependent fashion (37).In CRCs harboring wildtype p53, this increase in MDM2 expression resulted in decreased apoptosis and increased chemoresistance.The MDM2 role in driving cancer drug resistance could underly the minimal DDLPS chemotherapeutic responses (39)(40)(41)(42)(43).
Synthetic anti-GP130 drugs have shown promising results.Xu et al. showed ovarian cancer tumor regression in vivo with the treatment of SC144 (24).SC144 treatment led to deglycosylation and subsequent phosphorylation and internalization of GP130, drastically reducing the level of surface-bound protein and IL6-induced GP130 signaling neutralization (24).Xiao similarly observed efficacy during the use of Bazedoxifene, an FDA-approved agent with known anti-IL6:GP130 axis activity currently used prevent postmenopausal osteoporosis (44).
Genotoxic stress induces production of alternative MDM2 variants and modulates the MDM2:p53 axis, thereby affecting p53 activity (26,28,45).Jacob et al. demonstrated that alternative splice product MDM2-ALT1 was able to bind MDM2-FL and MDMX, the conventional MDM2 binding partner involved in p53 negative regulation (28).As MDM2-ALT1 lacks the p53-specific binding site but maintains the RING domain required for MDM2 dimerization, it is believed to physically impair MDM2-FL activity via direct binding (28).Cytoplasmic binding of MDM2-ALT1 to MDM2 rendered MDM2 inert, which resulted in p53 stabilization, the subsequent upregulation of downstream p53 target p21, and an arrest in cell cycle progression.We similarly observed substantial increases in MDM2-ALT1 and TP53 levels with markedly elevated apoptotic levels and PUMA expression upon DDLPS treatment with SC144.These data suggest that p53 reactivation in response to GP130 blockade-induced MDM2-ALT1 production may impair DDLPS progression via apoptosis induction.Furthermore, the observed decreased MDM2-FL transcript and elevated MDM2-ALT1 expression upon GP130 blockade suggest that DDLPS MDM2 is negatively regulated on the genomic level as well as post-translationally.
Understanding the potential impact of TME non-cancer cells on DDLPS is critical, especially given the functional role of stromal cells in promoting cancer progression (46,47).Here, we have established a novel PreAdip:DDLPS paracrine loop in which PreAdip-derived IL6 enhances DDLPS growth and oncogenic phenotype, potentially via activation of downstream STAT1 and STAT3 effector proteins.Co-culture with DDLPS cells may also modify PreAdip biology.Others have described the transition of normal cells into cancer-promoting bodies in the presence of cancer cells (46,47).In a recent publication by Lee et al., cancer-associated adipocytes were shown to undergo dedifferentiation and express higher levels of IL6 when co-cultured with breast cancer cells.These modified fat cells were subsequently able to promote cancer cell progression (47).Therefore, genetic, proteomic, and morphological analyses comparing cancer-associated cells with their normal cell counterparts could be considered to better understand the PreAdip:DDLPS cell relationship and further define the tumor-promoting of PreAdip cells within the TME.Results of these studies will hopefully enable novel treatment strategies based on better understanding of the heterogeneous DDLPS tumor microenvironment.
External Sn: 5'-CTGGGGAGTCTTGAGGGACC-3' External ASn: 5'-CAGGTTGTCTAAATTCCTAG-3' Internal Sn: 5'-CGCGAAAACCCCGGGCAGGCAAATGTGCA-3' Internal ASn: 5'-CTCTTATAGACAGGTCAACTAG-3'In vivo xenograft modelSix week old female athymic nude (NCr-nu/nu) mice, outbred, were acquired from the athymic nude mouse colony maintained by the Target Validation Shared Resource at the Ohio State University; original breeders (strain #553 and #554) for the colony were received from the NCI Frederick facility, and were injected subcutaneously with 10 6 Lipo246 cells into the flank of 20 mice.Upon reaching 0.5 cm 3 tumor size, the 15 mice that displayed tumor formation were randomized into two arms and daily treated via intraperitoneal injections: Vehicle (0.9% NaCl/40% propylene glycol/59.1% ultra-pure distilled water; 8 mice), or SC144 (10 mg/kg diluted in vehicle solution; 7 mice).Mice were weighed and tumors were measured twice weekly.Mice were euthanized once tumors in the control group grew to ~1.5 cm 3 .Final tumor volumes and weights were measured; tumor volume was calculated using the formula (L x (W 2 )/2.Study approvalApproval for the proper handling of mice was obtained fromThe Ohio State University Office of Responsible Research Practices in adherence with the Institutional Animal Care and Use Committee protocols (approval #2014A00000085) in accordance with The Ohio State University's Animal Welfare Assurance (#A3261-01).DDLPS tumor samples were provided by patients of the James Cancer Center of The Ohio State University Wexner Medical Center.Written, informed consent to participate in the study was obtained from each donor, and approval for the handling of patient samples was attained from The Ohio State University Internal Review Board (approval #2014C0028).
), suggesting possible relevance to DDLPS development and/or progression.GP130 signal transduction activates downstream STAT1 and STAT3 effector proteins (17-19).STAT1 and STAT3 were inactive or marginally activated in DDLPS cells, but showed enhanced activation with exogenous IL6 (Fig 1E).Loss of DDLPS GP130 severely reduced active STAT1 and STAT3 in the presence of IL6, suggesting that their activation by IL6 occurs in a GP130-dependent manner (Fig 1F).AKT remained inactivated in DDLPS cells, and the addition of IL6 did not impact AKT or ERK1/2 activation (S1 and S2 Figs).Moreover, DDLPS cells express constitutively active ERK1/2 independent of IL6 GP130 activation (S1 and S2 Figs).
Patient sample immunohistochemistry (IHC) was performed to detect IL6 within DDLPS TME, revealing IL6 expression in DDLPS tumors (Fig 3A).Consistent with others(16), we observed an elevated expression and secretion of IL6 by PreAdip and DDLPS cells, albeit DDLPS expressed substantially lower IL6 levels (Fig 3B).DDLPS cells cultured directly or indirectly with PreAdip cells displayed markedly increased proliferation compared to DDLPS monoculture cells (Figs 3C and D).DDLPS cells cultured in PreAdip-derived conditioned media (CM) demonstrated similar growth increases that were not further augmented by exogenous IL6 (Fig 3E).Anti-IL6 neutralizing antibody (MAB206) reduced the DDLPS-promoting effect of PreAdip cells with simultaneous reduction of DDLPS GP130 activation (S3 and S4 Figs).However, treatment with MAB206 did not fully inhibit the pro-DDLPS growth effects of PreAdip, suggesting that these normal cells may also promote DDLPS growth by mechanisms other than IL6 release.

Fig 6 .
Fig 6.SC144-mediated GP130 blockade impairs DDLPS growth in vivo.Tumor volume (A) and concurrently with GP130 blockade (Fig 7B).Similarly, nuclear expression of activated STAT1 and STAT3 also decreased (Fig 7C).To assess the impact of GP130 blockade on MDM2 expression, Lipo246 DDLPS cells were treated for 48h with SC144 with or without IL6, after which GP130 and MDM2 levels were measured.Treatment with SC144 significantly decreased expression of MDM2 mRNA and protein with concurrently decreased GP130 levels (Fig 7D), indicating possible GP130 signaling in MDM2 regulation.As observed in vitro, IL6ST and MDM2 levels decreased in tumors of mice receiving SC144 treatment versus controls (Fig 8A).

Fig 9 .
Fig 9. Proposed mechanism of GP130 signaling-mediated upregulation of MDM2 and oncogene