Bone morphogenetic protein (BMP) receptor inhibitor JL5 synergizes with Ym155 to induce Apoptosis-Inducing Factor (AIF) caspase independent cell death

Background: Bone morphogenetic protein (BMP) is an evolutionarily conserved morphogen that is reactivated in lung carcinomas. BMP receptor inhibitors promote cell death of lung carcinomas by mechanisms not fully elucidated. The studies here reveal novel mechanisms by which the “survivin” inhibitor Ym155 in combination with the BMP receptor inhibitor JL5 synergistically induces death of lung cancer cells. Methods: This study examines the mechanism by which Ym155 in combination with JL5 downregulates BMP signaling and induces cell death of non-small cell lung carcinoma (NSCLC) cell lines. Validation experiments were performed on ve passage 0 primary NSCLC cell lines. Results: We found that Ym155, which is reported to be a survivin inhibitor, potently inhibits BMP signaling by causing BMPR2 mislocalization into the cytoplasm and its decreased expression. The combination of Ym155 and the BMP receptor inhibitor JL5 synergistically causes the downregulation of BMP Smad-1/5 dependent and independent signaling and the induction of cell death of lung cancer cell lines and primary lung tumors. Cell death involves the nuclear translocation of apoptosis inducing factor (AIF) from the mitochondria to the nucleus. This causes DNA double stranded breaks independent of caspase activation, which occur only when JL5 and Ym155 are used in combination. Knockdown of BMPR2 together with Ym155 also induced AIF localization to the nucleus. Conclusions: These studies suggest that inhibition of BMPR2 together with Ym155 can induce AIF caspase-independent cell death. AIF caspase-independent cell is an evolutionarily conserved cell death pathway that has never been targeted to induce cell death in cancer cells. These studies provide mechanistic insight of how to target AIF caspase-independent cell death using BMP inhibitors.

caspase-independent cell death. AIF caspase-independent cell is an evolutionarily conserved cell death pathway that has never been targeted to induce cell death in cancer cells. These studies provide mechanistic insight of how to target AIF caspase-independent cell death using BMP inhibitors.

Background
The bone morphogenetic proteins (BMP) are evolutionarily conserved cytokines that regulate a plethora of signaling events throughout development. Following the development of the fetal lung there is a decrease in BMP signaling, with little BMP activity seen in the mature lung [1,2]. BMP signaling is reactivated in lung and many other carcinomas [3]. In non-small cell lung carcinomas (NSCLC), the BMP-2 ligand is over-expressed 17-fold higher in comparison to normal lung tissue and benign lung tumors [4]. A number of studies have shown that BMP signaling has signi cant tumorigenic effects that involve the regulation of cell survival, migration, proliferation, stemness, and angiogenesis, and that high ligand expression correlates with a worse prognosis [3,[5][6][7][8][9][10].
Here we report that the BMP receptor inhibitor JL5 and the "survivin" inhibitor Ym155 synergistically promote cell death of lung cancer cell lines and primary tumors. We identi ed novel mechanisms by which Ym155 promotes cell death in cancer cells, which involves the inhibition of BMPR2 Smad-1/5 independent signaling. The combination of Ym155 and JL5 synergistically decreases BMP independent signaling by inhibiting BMPR2. Cell death induced by the combination of Ym155 and BMPR2 inhibition occurs independent of caspase activation and involves the translocation of AIF to the nucleus. These studies provide novel insights into how to target BMPR2 Smad-independent signaling to induce cell death of lung cancer cells.
Blots suggesting regulation of loading controls actin or GAPDH by Ym155 and/or JL5 were then probed with the spectrin, which is also a cytoskeletal protein.

Cell viability
Cells were plated in duplicate into 6-well plates and treated the next day for the designated period of time. Cell counts were determined using the automated cell counter Vi-CELL cell analyzer (Beckman Coulter).
Approximately 500 cells per sample were analyzed and trypan blue dye exclusion determined number of dead cells. The experiment was replicated three times in our laboratory.

Combined drug effects
The median-effect principle of Chou and Talalay was used to evaluate synergy between JL5 and Ym155 [23]. The combination index (CI) values were calculated using Compusyn software to determine mode of interaction with CI < 1.0 indicating synergism, 1.0 additive and > 1.0 antagonistic response.

Transient knockdown
Validated select siRNA was used to knockdown BMPR2 (Life Technologies). The ID numbers for the siRNA are: BMPR2 (s2044 and s2045). Silencer Select negative control siRNA (4390843) was used to evaluate selectivity. Transfections of the siRNA were performed in duplicate using Lipofectamine® RNAiMAX Reagent (Invitrogen, Carlsbad, CA, USA) according to manufacturer's protocol. Cells were transfected with 6 nM BMPR2 or 6 nM of siRNA control. The experiment was replicated three times in our laboratory.

Cytosol extraction
Cytosolic protein extraction was performed using Mitochondria/Cytosol fractionation kit as per manufacturer's instructions (Enzo Life Sciences, NY, USA). Cell pellets were resuspended in 100µl of icecold Cytosol Extraction Buffer Mix containing dithiothreitol (DTT) and Protease Inhibitors. After a 10minute (min) incubation on ice, cells were homogenized. The homogenates were collected to a fresh 1.5ml tube and centrifuged at 700 x g for 10 min at 4°C. The supernatant was collected as the cytosolic fraction and used for further experiments.
TUNEL assay DNA double strand breaks (DSB) after treatment were analyzed by using FlowTACS In Situ TUNEL-based apoptosis detection kit (Trevigen) according to the manufacturer's protocol. Cells were treated in duplicate. After treatment, cells were trypsinized and the cell pellet was xed with 4% formaldehyde and permeabilized with cytonin for 30 min. Cells were washed with labeling buffer and resuspended in reaction mix for 1 hour (h), then stained with strep-uorescein solution and analyzed using ow cytometry (LSRII, BD Biosciences). The experiment was replicated three times in our laboratory.
Immuno uorescence staining Cells were seeded for 24h onto microscope cover glasses in a 6-well plate then treated. Cells were xed with 4% formaldehyde and permeabilized with 0.5% triton-X. Cells were blocked with CAS-block for 1h; cells were stained with anti-BMPR2 antibody that recognizes an extracellular epitope (Sigma-Aldrich) or AIF for 1h at room temperature. To assess for BMPR2 on the plasma membrane the cells were xed but not permeabilized prior to staining. Cells were then washed with phosphate buffered saline (PBS) and stained with Alexa Flour 488 conjugated secondary antibody for 1h at room temperature. After washing with PBS, the nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) for 10 min. Fluorescent images were captured using a Nikon eclipse TE300 inverted epi uorescent microscope and a Cool Snap black and white digital camera. IP Lab imaging software was used to assign pseudocolor to each channel. All of the immuno uorescence experiments except the primary tumors cells were replicated three times in our laboratory. The AIF immuno uorescence study using primary tumor cell lines was performed once for each of the 5 tumors. Consecutive non-small cell lung carcinomas were used. The tissues were washed with PBS and then cut into small pieces. Human tissue dissociation kit (Miltenyl Biotec) was used to dissociate the tumor tissues according to the manufacturer's protocol. Dissociated cells were grown in Dulbecco's Modi ed Eagle Medium (DMEM) media supplemented with 20% Fetal Bovine Serum (FBS) and 1% antibioticantimycotic for approximately 10 days without splitting. At 70% con uence, cells were trypsinized and plated for all the experiments at the same time.

Tumor xenograft studies
Tumor xenografts from H1299 cells were established by injecting 2 million cells into the intradermal space of the anks of 6 week old female NOD-scid IL2Rgamma null (NSG) mice (bred in-house and randomly assigned to groups without blinding). When tumors reached approximately 4mm x 4mm, mice were treated with 3 or 10 mg/kg of JL5 twice daily (or DMSO control) for four days. Eight mice were used in these studies, 4 DMSO control and 4 JL5 treated. Treated mice were not blinded. At the end of the experiment mice were euthanized using CO2. Experiments were conducted in compliance with ethical regulations and approved by Rutgers IACUC.

Statistical Analysis
The mean of the control group was compared to the mean of each treated group using a paired student ttest assuming unequal variances. Differences with p values <0.05 were considered statistically signi cant.

Results
Ym155 decreases BMP signaling H1299 cells were injected into the intradermal space of the anks of NSG mice and were treated with DMSO or JL5 for 4 days. We previously reported that JL5 decreases the expression of Id1 but not XIAP in tumor xenografts [8]. Survivin belongs to the family of anti-apoptotic proteins that have been reported to be regulated by BMP signaling and are known to stabilize the expression of XIAP [24]. We found for the rst time that inhibition of BMP signaling with JL5 causes a signi cant increase in the expression of survivin in tumor xenografts (Fig. 1A).
We explored whether Ym155, a reported survivin inhibitor [25], would have synergy when used in combination with a BMP inhibitor. First, we looked at the effects of Ym155 on the expression of survivin and XIAP. Ym155 induced a signi cant decrease in the expression of XIAP in H1299 cells (Fig. 1B) and A549 cells (Fig. 1C) with little change in the expression of survivin. This is consistent with prior studies suggesting that Ym155 mediates its effects independent of survivin [26].
Since BMPR2 regulates XIAP through Smad-1/5 independent mechanisms, we examined whether Ym155 regulated BMP signaling. The effects of Ym155 on BMP Smad-1/5 dependent signaling was assessed by examining changes in the phosphorylation of the BMP transcription factor Smad-1/5 and its downstream transcriptional target Id1. Ym155 caused a signi cant decrease in the activity of Smad-1/5 and Id1 in both the H1299 (Fig. 1D) and A549 (Fig. 1E) cells at nanomolar concentrations. H1299 cells were stably transfected with the Id1 promoter driving the expression of the luciferase reporter. Ym155 caused a dose responsive decrease in the Id1 reporter (Fig. 1F). Transfection of constitutively active BMPR1A (caBMPR1A) activates Smad-1/5-Id1 signaling. Ym155 inhibited caBMPR1A activation of Smad-1/5 (Fig.  1G). These studies show that Ym155 regulates both Smad-1/5 dependent and independent signaling, raising the possibility it regulates BMPR2.
We examined the effects of Ym155 on cell survival and the induction of cell death. After 48 h the majority of A549 cells treated with 100 nM were dead (Fig. 1H). Examining the IC50 at 24 h with concentrations below 100 nM, we found the IC50 to be 89 nM and 19 nM for A549 and H1299, respectively (Fig. 1I). An increase in the percent dead cells was not seen in the A549 cells at these concentrations after 24 h however; a small increase in dead cells was seen in H1299 cells at 40 nM and above (Figs. S1A-B). In comparison, the BMP inhibitor JL5's IC50 was higher with an IC50 of 12 mM and 7 mM for A549 and H1299 cells, respectively ( Fig. 1J) with few dead cells seen except at high concentrations (10 mM) (Figs. S1C-D).
Ym155 synergizes with JL5 to decrease BMP signaling and induce cell death Since Ym155 does not target the BMP receptors, we examined whether it would induce synergy with JL5. To test for potential synergy, we used concentrations of Ym155 and JL5 that were below their IC50 and had no increase in cell death. JL5 caused signi cantly more cell death when used in combination with Ym155 compared to either compound alone in both H1299 and A549 cells ( Figs. 2A-B). The combination index (CI) [23] was calculated to be <1.0 (0.94 H1299 cells and 0.16 A549 cells) indicating the combined effect was synergistic. Morphologically, the nuclei were signi cantly smaller and demonstrated chromatin condensation when treated with both JL5 and Ym155 in comparison to either compound alone (Figs. S1E-F). By Western blot analysis, there was little change in BMP signaling with JL5 and Ym155 after 24 h. The combination of Ym155 and JL5 caused a greater decrease in the expression of Id1 and XIAP in comparison to each compound alone in both the A549 and H1299 cells (Figs. 2C-F). Ym155 alone induced apoptosis as demonstrated by the expression of activated caspase-3, with 17 kd and 19 kd fragments and cleavage of its downstream target poly ADP ribose polymerase (PARP) (Fig.   2G). We found no activation of caspase-3 or caspase-9 when examined at 1, 3, and 24 h (Fig. 2G and  data not shown). The pan-caspase inhibitor Z-VAD-FMK did not affect cell death induced by JL5 and Ym155 in combination (Fig. 2H). Necrostatin, which inhibits necrosis induced by receptor-associated adaptor kinase 1 (RIP1) [27], also had no effect on cell death induced by JL5 and Ym155 when used in combination (Fig. 2H). These studies suggest that Ym155 together with a BMP receptor inhibitor synergistically mediates cell death by mechanisms independent of the caspases or RIP1 induced necrosis.

Ym155 together with JL5 increases cytosolic AIF and Smac/DIABLO
Since apoptosis is not induced with JL5 and Ym155, we explored whether cell death involved the mitochondrial release of AIF [28]. In cells undergoing AIF induced cell death, AIF is cleaved producing 67 kd and 57 kd fragments, which are then rapidly transported to the nucleus. Nuclear AIF induces DNA fragmentation and chromatin condensation leading to cell death [28]. The combination of Ym155 and JL5 synergistically enhanced the release of Smac/DIABLO into the cytosol as early as 3h and persisted for at least 24h in A549 cells (Figs. 3A-B). In the H1299 cells, we did not identify cytosolic AIF after 3h but identi ed cytosolic AIF after 24h in cells treated with JL5 and Ym155 in combination BMP inhibitors with YM155 cause DNA double stranded breaks and nuclear localization of AIF A hallmark of AIF induced cell death is the induction of DNA double stranded breaks (DSB) and its localization to the nucleus. The TUNEL assay was used to determine DNA-DSB. Three hours following treatment with JL5 or Ym155 alone or in combination no DNA-DSB were found ( Figure 4A). After 24h, very few cells treated with JL5 or Ym155 alone demonstrated DNA-DSB. When JL5 was used in combination with Ym155, approximately 40% and 65% of cells demonstrated DNA-DSB in A549 and H1299 cells, respectively (Fig. 4B). DMH2 is similar to JL5, having potent inhibition of BMP type 1 receptors with some inhibition of BMPR2 [8]. DMH2 treated cells also demonstrated an increase in DNA-DSB when used in combination with Ym155 (Fig. 4C). To assess whether synergy also occurred by inhibiting TGFb signaling, we used the selective TGFb receptor inhibitor SB-505124 [29]. SB-505124 had no effect on DNA-DSB when used alone or in combination with Ym155 ( Figure 4C). Using immuno uorescent imagining, we examined whether AIF localized to the nucleus following treatment with JL5 and Ym155. JL5 and Ym155 alone did not cause the localization of AIF to the nucleus in either the H1299 or A549 cells but the combination of JL5 and Ym155 showed an increase in AIF localization to the nucleus (Figs. 4D-G).

Ym155 decreases BMPR2 expression and promotes BMPR2 mislocalization to the cytoplasm
Since Ym155 decreases the expression of XIAP and Id1, which are both regulated by BMPR2, we examined if Ym155 regulated the expression of BMPR2. Ym155 caused a dose-related decrease in the expression of BMPR2 that was associated with a corresponding decrease in the expression of XIAP in both the A549 and H1299 cells (Figs. 5A-B). At lower concentrations of Ym155 (20 nM), which has no effect on the level of BMPR2 expression, an enhanced decrease in expression of BMPR2 and XIAP was observed when Ym155 was used in combination with JL5 (Figs. 5C-D). As previously reported, JL5 caused the mislocalization of BMPR2 to the cytoplasm (Fig. 5E) [22]. We show for the rst time that Ym155 alone also caused an increase in the localization of BMPR2 to the cytosol (Fig. 5E), which was greatly increased when JL5 and Ym155 were used in combination (Fig. 5E).

To assess whether the increased nuclear localization of AIF in cells treated with JL5 in combination with
Ym155 was mediated by the inhibition of BMPR2, we knocked down the expression of BMPR2 with siRNA in H1299 cells then treated the cells with a low concentration of Ym155 (20 nM). The knockdown of BMPR2 together with Ym155 (20 nM) signi cantly enhanced AIF localization to the nucleus in comparison to siRNA control cells treated with Ym155 (20 nM) (Figs. 5F-G). These studies show that Ym155 decreases BMP signaling, which involves the downregulation of the expression of BMPR2 expression and its mislocalization to the cytoplasm. The data also suggests that the inhibition of BMPR2 is required for AIF localizing to the nucleus.

Synergy of Ym155 and JL5 in primary NSCLC
Primary NSCLC cells were obtained directly from 5 surgically resected lung tumors. Tumors 1, 2, 3 and 5 were adenocarcinomas and tumor 4 was a squamous carcinoma. Tumors were immediately gently digested and plated for cell culture. After approximately 10 days the cells were treated with JL5 and Ym155 alone and in combination for 48h. Primary cancer cells were examined for changes in cell survival, regulation of BMP signaling, and nuclear localization of AIF.
The combination of JL5 and Ym155 caused an increase in cell death compared to DMSO control in 5 of 5 tumors (Fig. S5A). In 3 of 5 tumors, the combination of JL5 and Ym155 induced cell death that was greater than either compound alone (Fig. 6A). XIAP was utilized as a marker of the downregulation of BMPR2 Smad-1/5 independent signaling. The combination of JL5 and Ym155 synergistically decreased the expression of XIAP in 3 of 5 tumors (tumors 1-3), which corresponded to the tumors that JL5 in combination with Ym155 resulted in enhanced cell death (Figs. 6A-C). In tumors 3 and 4, Ym155 alone caused a decrease in the expression of XIAP that was not enhanced further with JL5 (Figs. 6 A-C).
AIF nuclear localization was examined in primary lung tumors 2-5. AIF nuclear staining was not seen in primary tumors treated with DMSO or JL5 (Figs. 6 D-F). Only a small percentage of cells from the primary lung tumors treated with Ym155 demonstrated nuclear staining with AIF (0-10%). AIF nuclear staining occurred predominantly in primary tumors that demonstrated increased cell death with the combination of JL5 and Ym155 (tumors 2 and 3, 67% and 25% of cells, respectively) (Figs. 6D-E).

Discussion
Our prior studies suggested that targeted inhibition of BMPR2 induces cell death mechanisms not seen with the inhibition of BMP type 1 receptors [22]. Knockdown of BMPR2 decreases the expression of XIAP in lung cancer cell lines and increases cytosolic cytochrome c and Smac/DIABLO [22]. The BMP inhibitors JL5 and DMH2 decrease the expression of XIAP and cause more cell death than the BMP inhibitors DMH1 and LDN [7,8]. JL5 and DMH2 both demonstrate some inhibition of BMPR2 with in vitro phosphorylation kinase assays with an IC50 of approximately 8 mM [8]. Despite JL5 having a relatively low IC50 for BMPR2 it causes its mislocalization to the cytoplasm at 2.5 mM. This mislocalization does not occur with BMP inhibitors DMH1 or LDN, which have no activity for BMPR2 [22]. The over-expression of XIAP attenuates cell death induced by JL5 [22]. These studies suggest that cell death induced by JL5 and/or DMH2 involves the inhibition of BMPR2 and the downregulation of XIAP [22].
In tumor xenograft studies, JL5 caused some tumor regression, however when tumors were examined after 21 days there was little cell death [8]. Tumors examined after 4 days of treatment showed that Id1 expression was decreased but there was no decrease in the expression of XIAP [8]. This suggested that signaling pathways were induced that stabilized the expression of XIAP and/or BMPR2 was not su ciently inhibited. We show in the present study that the expression of survivin is signi cantly increased in tumor xenografts treated with JL5. Survivin expression is increased by BMP signaling [30] and is known to increase XIAP stability against ubiquitin-dependent degradation [31]. Surprisingly, Ym155, which is reported to be a survivin inhibitor [25], caused a signi cant decrease in Smad-dependent and independent signaling at nanomolar concentrations. We show that Ym155 causes a signi cant decrease in the expression of BMPR2 and its mislocalization to the cytoplasm, which are likely mechanisms leading to the decrease in BMP signaling. In lung cancer cells, the BMP type 1 receptors only regulate Smad-dependent signaling while BMPR2 regulates both Smad-dependent and independent BMP signaling. The mechanism by which Ym155 causes a decrease in BMPR2 expression is not known but may involve the tra cking of BMPR2 to the lysosome and subsequent degradation.
Our studies suggest that inhibition of BMPR2 together with Ym155 promotes AIF caspase-independent cell death. AIF is an evolutionary conserved protein that has two independent functions; biogenesis of the electron transport chain and cell death [32][33][34]. An increase in permeability of the outer mitochondrial membrane (OMM) is required for its release into the cytosol [35]. AIF is transported to the nucleus where it induces large-scale DNA fragmentation and cell death [28,34,35]. AIF caspase-independent cell death occurs in response to ischemic-reperfusion injury in neurons and myocardium and has only infrequently been reported in cancer cells.
Ym155 was originally reported to be a survivin inhibitor [25]. A subsequent study showed that Ym155 resulted in dose-dependent induction of yH2AX and pKAP1, which are markers of DNA damage, at a concentration lower than required to decrease the expression of survivin [26]. This report concluded that Ym155 is a DNA damaging agent and the suppression of survivin is a secondary event. The mechanism by which Ym155 initiates the downregulation of BMP signaling and promotes cell death needs further investigation. Although Ym155 has not shown much e cacy in human trials it causes signi cant cell death of cancer cells in vitro. Ym155 has a short half-life and its pharmacokinetic properties could hinder its activity in human tumors. The identi cation of the mechanism by which Ym155 regulates cell death may lead to the identi cation of other drugs that could be used in combination with BMPR2 inhibitors.

Conclusions
We demonstrate that Ym155 decreases BMP signaling that is associated with a downregulation of BMPR2 signaling in lung cancer cells. We show that a BMPR2 inhibitor and Ym155 can synergistically target BMPR2 signaling to induce cell death. Cell death induced by Ym155 and BMPR2 inhibition involves AIF caspase-independent cell death mechanisms in lung cancer cells. These studies provide further insight into how to target BMP signaling as a therapeutic in cancer and supports further drug development of more speci c and potent BMPR2 inhibitors.

Declarations
Ethics of approval and consent to participate Institutional review board of Rutgers University of New Jersey approved tissue to be obtained from patients' resected tumors. Consent was obtained from an honest broker who kept patient personal information anonymous and secure. Mice experiments were conducted in compliance with ethical regulations and approved by Rutgers IACUC.

Consent for publication
Our manuscript does not contain any individual person's data in any form.

Availability of data and materials
The datasets obtained and analyzed for this study will be made available from the corresponding author in a reasonable request.

Competing interests
A provisionary patent application was sent for BMP inhibitor JL5. The full patent application for JL5 has been withdrawn and is no longer being pursued. There are no active or pending nancial agreements regarding JL5 nor has any money been received or is pending.