Mechanistic Insights on Cytotoxicity of KOLR by targeting Signaling Complexes of phosphodiesterase 3B and Rap guanine nucleotide exchange factor 3


 Background

Protein signaling complexes play important roles in prevention of several cancer types and can be used for development of targeted therapy. The roles of signaling complexes of phosphodiesterase 3B (PDE3B) and Rap guanine nucleotide exchange factor 3 (RAPGEF3), which are two important enzymes of cyclic adenosine monophosphate (cAMP) metabolism, in cancer have not been fully explored.
Methods

The natural product Kaempferol-3-O-(3′′,4′′-di-E-p-coumaroyl)-α-L-rhamnopyranoside designated as KOLR was extracted from Cinnamomum pauciflorum Nees leaves using reverse phase chromatography, high-resolution mass spectrometry, and nuclear magnetic resonance spectroscopy. The antitumor effect of KOLR was analyzed by multiple cell proliferation and metastasis experiments. The PDE3B/RAPGEF3 complex was found to be the target of KOLR through mRNA sequencing, Co-Immunoprecipitation assay, gene knock-down, gene mutation of drug-resistance cell line, and molecular docking. In vivo studies have shown that KOLR has the same antitumor mechanism.
Results

KOLR exhibited cytotoxic effects against selected cancer cells, except for AsPC-1 pancreatic cancer cell line. KOLR stabilized PDE3B/RAPGEF3 signaling complex thus inhibiting AKT phosphorylation and Rap-1 activation. Notably, mutation of RAPGEF3 G557A inhibited effect of KOLR on stabilizing PDE3B/RAPGEF3 complex in AsPC-1 cells. Furthermore, downregulation of PDE3B expression inhibited cytotoxic effect of KOLR on tumor cells. Downregulation of RAPGEF3 and Rap-1 expression promoted apoptosis of tumor cells and inhibited tumor metastasis. PDE3B inhibits activity of RAPGEF3 and activation of downstream signaling pathway.
Conclusion

The findings of this study show that KOLR could stabilize PDE3B/RAPGEF3 signaling complex to play an anti-tumor role and the PDE3B/RAPGEF3 complex is a potential therapeutic cancer target.


Conclusion
The ndings of this study show that KOLR could stabilize PDE3B/RAPGEF3 signaling complex to play an anti-tumor role and the PDE3B/RAPGEF3 complex is a potential therapeutic cancer target.

Background
Approximately 19.3 million new cancer cases and almost 10.0 million cancer deaths were reported in 2020 globally [1]. Although targeted and immunomodulatory drugs are effective in a subset of cancer patients [2,3], most cancer patients do not respond to these treatments towing to drug resistance and severe side effects. Therefore, studies should explore new targets and discover effective drugs for the treatment of various cancers. Phenotypic screening for new anticancer agents against different cancer cell lines can be used to discover novel targets for cancer therapy and help in elucidating the speci c molecular mechanism [4,5]. Protein-protein interactions (PPIs) and protein complexes play important roles in regulation and execution of most biological processes and are regarded as phenotypic screening targets in cells [6,7].
Cyclic adenosine monophosphate (cAMP) is a second messenger and plays important role in most biological functions, including cardiac contraction, energy homeostasis, cell metabolism, and ion channel activation [8][9][10]. Intracellular cAMP concentrations are regulated by coordinated activities of adenylyl cyclases (ACs) and cyclic nucleotide phosphodiesterases (PDEs). G protein-coupled receptors (GPCRs) signaling activates ACs leading to generation of cAMP in cells [11]. PDEs control cellular content of cyclic nucleotides and recruit them into multiprotein signaling complexes by forming individual pockets or nanodomains of cyclic nucleotide signaling [12][13][14].
PDE3B inhibits activity of RAPGEF3 and activation of the downstream signaling pathway. The PDE3B/RAPGEF3 protein complex plays a key role in inhibiting cell growth and metastasis. The ndings of the current study show that PDE3B/RAPGEF3 complex is potential therapeutic cancer target.

Isolation and puri cation of KOLR from leaves of Cinnamomum pauci orum Nees
Cinnamomum pauci orum Nees leaves were crushed into powder using high-speed multifunctional mill. 50 g of the powder was weighed and added to ethanol at a mass to volume ratio of 1:10. The powder was soaked and extracted overnight in a shaker at room temperature, then centrifuged at 3000 rpm for 15 min. The supernatant was collected as ethanol extract (100 mg/mL). Different concentrations (30%, 50%, 70%, and 100%) of ethanol were used to separate the extracts through C18 trans separation solid-phase extraction cartridges InertSep (GL Sciences, Netherlands), and the washed components were collected. Analytical high performance liquid chromatograph (HPLC) was used for qualitative analysis of anti-tumor active components, and bioassays were performed to explore the anti-tumor activity. Active compounds were analyzed using high-performance liquid phase, C18 column, mobile phase comprising acetonitrile: water =55:45, and detection performed at a wavelength of 275 nm. The collected sample was concentrated to eliminate acetonitrile by pressure reduction and then freeze-dried to obtain a solid. Characterization of a single active ingredient was performed by vacuum drying at room temperature. The active compounds were identi ed by high-resolution mass spectrometry and nuclear magnetic resonance (NMR) spectroscopy.

Cancer cell lines culture
All cancer cells were purchased from Procell (Wuhan, China) and Cell Bank, Chinese Academy of Science (Shanghai, China). Cell lines were cultured in modi ed 1640 medium (Biological Industries) or Dulbecco Minimal Essential Medium (DMEM, Biological Industries) with 10% fetal bovine serum (FBS, Biological Industries) under a humidi ed incubator containing 5% CO 2 at 37℃. All cancer cells grew adherently.
Cancer cells in the logarithmic growth phase were selected for further experiments.

Cancer cell proliferation assays
Cells were seeded into 96-well plates at a density of 5×10 3 cells/100 μL in 1640/ DMEM medium containing 10% FBS. After culturing overnight, different concentrations of extracts or growth factors such as KOLR, IGF, and VEGF were added in 3 replicates. Cells in each well were treated with 10 μL CCK8 (ApexBio), and the absorbance was measured at 450 nm at different time points. Cell survival rate was calculated by GraphPad Prism 8.

Cancer cell colony formation assay
Cells were digested with trypsin, and the BxPC-3 cells were plated in a 6-well plate at a density of 1×10 3 cells/2 mL. Cells were cultured overnight in an incubator for 24h. After cells adhered to the wall, KOLR at a nal concentration of 2.5, 5, and 10 μg/mL was added and incubated for 15 days. Methanol was washed off using PBS and cells were stained with 0.4% crystal violet solution for 20 min. The crystal violet solution was washed off using PBS. Cells were observed and images collected under a microscope, and the number of BxPC-3 cells forming clones was determined.

Cancer cell EdU assay
Cells were digested with trypsin, and cultured into BxPC-3 cells in a 96-well plate with 1×10 4 cells/well.
After cell adherent, cells were treated with KOLR at a concentration of 2.5, 5, and 10 μg/mL for 24 h, and then 50 μL of the culture medium was extracted. 50 μL complete medium containing EdU ( nal concentration of 10 μm) it was added and incubated in an incubator for 2 h. Cells were observed and images collected under a uorescence microscope, and the number of newly proliferated cells was determined.

Apoptosis of cancer cells
Trypsin digested cells were seeded in a system of 3×10 5 cells/well into a six-well plate. After culturing overnight, the medium was changed and different concentrations of KOLR were added and incubated for 24 h. Trypsin digestion was followed by centrifugation at 1000 RPM for 5 min to collect dead and living cells. Cells were resuspended with 400 μL 1×assay buffer (prepared with deionized water) and transferred to ow tubes. Cells were stained with 5 μL PI and FITC for 15 min at room temperature. Flow cytometry was used to observe cell apoptosis. FlowJO 7.6.1 software was used to analyze data and calculate the cell apoptosis rate.

Cancer cell cycle assay
BxPC-3 cells were seeded into a 6-well plate at 3×10 5 cells /well, and cultured overnight in an incubator for 24 h. The culture medium (containing 10% FBS) was changed and KOLR was added for 12 h. Trypsin digestion was performed followed by PBS washing. Cells were centrifuged at 300 g for 5 min to collect the living cells. Cells were xed at 4℃ with 75% ethanol (prepared with deionized water) for 6 h and centrifuged at 700 g for 5 min to remove ethanol. Further, cells were washed twice with PBS, then 450 μL PI dye was added and transferred to a ow tube. Cells were stained at room temperature in dark for 15 min, and cell cycle was detected by ow cytometry.

Cancer cell invasion assay
Cancer cells were seeded in 6-well plate at 5×10 4 cells/well and placed into a chamber containing 50 μL matrix glue, and 650 μL complete medium containing 10% FBS was added to the lower chamber. KOLR was added and cells incubated for 48 h. The chambers were removed and xed with methanol for 20 min. After staining with 0.1% crystal violet (PBS) for 20 min, cells in the chambers were wiped off with cotton swabs. Cells were observed and images collected under an inverted microscope. Number of cells passing through the chambers was determined.

Cancer cell wound-healing assay
Cancer cells in the logarithmic growth phase were seeded into a six-well plate at a density of 4×10 5 /well and covered with wound healing experimental plug-ins. The plug-ins were placed in an incubator and cultured for 24 h. Plug-ins were removed, then the complete medium was replaced and KOLR was added. Images were obtained at 0 h and 24 h and area of the scratches was calculated with Image J.

RNAseq of BxPC-3 cells
BxPC-3 cells were seeded into a 6-well plate at a density of 3×10 5 cells/well, incubated for 12 h. 20 μg/mL YC was added for 12 h, cells were washed with PBS, and 800μL Trizol was added to each well to dissolve. Library preparation was performed following NEB common library construction method and sequencing was performed at Novogene company laboratories. HISAT2 was used to evaluate the quality of the original data and analyze the reference sequence. Correlation analysis between gene expression level and samples was performed by FPKM. Differential genes were screened by DEseq2. Clusterpro ler was used for KEGG pathway analysis of differential gene sets.

Western blot assay
Cancer cells were seeded in a 6-well plate at a density of 3×10 4 cells/well. After overnight incubation, cells were treated with different concentrations of KOLR for 24 h. Cell lysates were prepared by RIPA lysis and centrifuged at 12000 rpm/min for 15 min to separate the soluble components. Protein concentration was determined using a BCA protein detection kit following the manufacturer's instructions with slight modi cations. A solution of 30 μg/mL was prepared with RIPA lysate, and the protein fragments were boiled in a metal sampler for 5 min. Samples containing 20-50 μg total protein were isolated on 8%-15% SDS-PAGE gel. The protein was transferred from the gel to nitrocellulose membrane (PALL) at a constant current at 300 mA for 90 min. Protein sample was sealed with 3% BSA for 1 h at room temperature and incubated overnight with the corresponding primary antibody (diluted by 3% BSA) in an antibody incubator box. Protein sample was washed three times with TBST. Goat anti-rat and anti-rabbit IgG were used as secondary antibodies and incubated at room temperature for 1 h, then washed three times with TBST. Chemiluminescence kit and chemiluminescence imager were used to detect the imprinting region. Image J was used to determine area of the imprinting region.

Indirect immuno uorescence
Cancer cells and KOLR treated cancer cells were seeded in a 48-well plate at a density of 1.5×10 4 /well.
Cell samples were treated with KOLR after cell adhesion for 24 h. Cells were washed three times with PBS, xed with 4% paraformaldehyde for 15 min, and permeated with 0.5% TritonX-100 for 20 min. 3% BSA was added to samples and incubated for 1 h. Primary antibody was diluted at 1:5000 and incubated with cells in a shaker at room temperature for 1 h, then samples were washed with PBST. Diluted uorescent secondary antibody was added to cells and incubated in a shaker at room temperature for 1 h, then washed by PBST. 100 μL DAPI was added to each well, stained for 15 min, add anti-uorescence sealing tablets and nail oil sealing tablets were added. Cells were observed and images collected under a confocal microscope.

Co-Immunoprecipitation assay
After KOLR treatment, BxPC-3 cells were transferred to EP tubes at a density of 1×10 7 /EP tube. After IP lysis/washing buffer was used for cell lysis for 30 min, then the cell precipitate was removed by centrifugation at 10000 g for 10 min. Protein concentration was determined using a modi ed BCA protein quantitative kit. Protein concentration in the cells was diluted to 1.5 mg/mL with IP lysis/scrubbing buffer. 500 μl protein lysate was added into a centrifuge tube containing antibody coupled magnetic beads and incubated overnight on a rotator at low temperature (magnetic beads were suspended during incubation). Magnetic beads were collected with the magnetic rack to remove unbound samples. 500 μL immunoprecipitation (IP) lysis/scrubbing buffer was added to the centrifuge tube, then samples were washed twice and gently mixed well. Magnetic beads were collected and the supernatant removed. 500μl ultra-pure water was added to the centrifuge tube and samples were washed once to remove the supernatant. 100 μL elution buffer was added to the centrifuge tube and incubated on a rotator at room temperature for 5 min. Magnetic beads were separated by a magnetic frame and the supernatant containing the target antigen was retained. Samples obtained by immunoprecipitation were used for determination of expression levels of the target antigen by western blot.

Immunohistochemistry
For Immunohistochemistry (IHC) analysis, depara nized DRG sections were boiled in sodium citrate buffer and incubated with primary antibodies p-Akt and Ki67 at 4 ℃ overnight. Immunostaining was performed using 3,3'-diaminobenzidine-tetrahydrochloride-dihydrate and samples were counterstained with hematoxylin. Negative controls were processed without the primary antibody. Nuclei were blue after hematoxylin staining and the positive expression of DAB was brownish-yellow.  Table 4S.

RNA extraction and qRT-PCR
Total RNA was extracted using the Trizol method (Haoke, China), and cDNA was generated by reverse transcription kit (Thermo, USA). Gene expression analysis was performed by qRT-PCR using an SYBR Premix Kit (Apibixo). Relative gene expression was quanti ed using the comparative threshold cycle (2 −ΔΔCt ) method. Sequences of the primers used in qRT-PCR (Qingke Biological Technology Services Co., Ltd, China) are shown in Table 5S 2.17 Molecular docking Tertiary (3D) structure of RAPGEF3 was built through homology modeling using Modeller. Secondary and tertiary structures of PDE3B were modeled from scratch by RaptorX and Rosetta as the N-terminal tertiary structure of PDE3B was missing. RAPGEF3 and PDE3B protein-protein docking was performed using Rosetta software. Autodock was used for docking of KOLR ligands to proteins.

Xenograft mouse model
All animal studies were approved by the Animal Care Ethics Committee of Zhejiang Academy of Traditional Chinese Medicine. Male BALB/c nude mice (4 weeks old) were purchased from the Zhejiang Academy of Medical Science. BxPC-3 cells were injected subcutaneously on the left anks of each mouse at a density of 1.5×10 6 cells/100μL. After the tumor size was approximately 100 mm 3 , KOLR was administered at a dose of 25 and 50 mg/kg through intraperitoneal (i.p.) route for 5 days and tumor size was determined every day. Tumor size was determined using the equation V (in mm 3 ) = 0.52×length×width 2 . Nude mice were sacri ced by anesthesia cervical dislocation, and tumors were harvested for further analysis.

Statistical analysis
For all in vivo and in vitro experiments, a two-tailed students test was performed using GraphPad 8.0. All experimental data were reported as mean ± SD or mean ± SEM. Differences were considered statistically signi cant at P<0.05, and the statistical signi cance was shown as*p < 0.05 **p < 0.01; and ***p < 0.001.

KOLR inhibits cancer cell proliferation and tumor growth
Crude extracts of Cinnamomum pauci orum Nees showed anti-tumor effects against several cancer cell lines, with signi cant effects observed for pancreatic cancer BxPC-3 cell line (Figure 1 and S1A). Antitumor effect of 50% ethanol extract obtained through the C18 separation column showed signi cant effects ( Figure S1B). KOLR compound was the main active ingredient identi ed using high-resolution mass spectrometry and nuclear magnetic resonance (NMR) by comparing to a reference compound ( Figure S1 and S2, Table 1S and 2S). KOLR showed signi cant inhibition activity against proliferation and colony formation of BxPC-3 cell line in a dose-and time-dependent manner ( Figure 1). Notably, KOLR showed different activities against different cell lines, with an IC50 14.4 μg/mL (19.9 μM) for BxPC-3 cells (Table 3S). Structure of KOLR including Kaempferol, α-L-rhamnopyranoside, and p-hydroxy-cinnamic acid was elucidated (Figure 1 and S3). The IC50 of Kaempferol against BxPC-3 cells was approximately 100 μM, whereas trans-cinnamic acid, and p-hydroxy-cinnamic acid showed no signi cant anti-tumor effect (IC50>100 μM) ( Figure S3A-D). These ndings show that the different KOLR compounds apart from kaempferol had signi cant inhibitory effects against cancer cell proliferation in vivo. KOLR inhibited the effect of several pro-cancer factors such as arachidonic acid (AA), insulin-like growth factor (IGF-1), Platelet-activating factor (PAF), and vascular endothelial growth factor (VEGF) (Figure S3E-H).
To explore the mechanism of KOLR inhibition activity against cell proliferation, cell cycle and apoptosis of BxCP-3 cells treated with KOLR were explored. KOLR signi cantly promoted apoptosis of BxPC-3 cells in a dose-dependent manner and modulated cell cycle of BxPC-3 cells (Figure 2A and Figure S4A). KOLR modulated expression of apoptosis-related genes such as PARP1, Caspase 3, Caspase 9, and CDK4 in a dose-dependent manner (Figure 2A and Figure S4B). In addition, KOLR inhibited BxCP-3 cell invasion and wound healing ability ( Figure 2B-E). Moreover, KOLR modulated expression of metastasis-related genes such as E-cadherin and β-catenin ( Figure S4C). These ndings indicate that KOLR promotes cancer cell apoptosis and inhibits tumor metastasis.

KOLR targets RAPGEF3/PDE3B complex
RAPGEFs mainly act as guanine nucleotide exchange factors that activate the small G proteins, Rap1. Figure 4E-F and Figure 6SB). Rap-1, PI3K-Akt, and EKR1/2 signaling pathways were downregulated by low expression levels of RAPGEF3 ( Figure 4C). To further explore the target of KOLR, immunoprecipitation analysis was performed on KOLR treated BxCP-3 cells using RAPGEF3 antibody. The ndings showed a decrease in protein concentration of RAPGEF3, PI3K γ110, Akt, and RAP1A/B, however, concentration of PDE3B increased in BxCP-3 cells with an increase in KOLR concentration ( Figure 5A). Laser confocal analysis showed that KOLR induces aggregation of PDE3B and RAPGEF3 resulting in co-localization in the cytoplasm ( Figure 5B). These ndings indicated that KOLR promoted binding between RAPGEF3 and PDE3B, and inhibited binding between PI3Kγ110 and Akt. To further explore the function of PDE3B in the RAPGEF3/PDE3B complex, RNA interference of PDE3B was performed. The ndings showed no change in apoptosis and invasion of BxPC-3 cells after PDE3B RNA interference compared with the control ( Figure S7). Downregulation of PDE3B increases the resistance of BxCP-3 to KOLR ( Figure 5C-F). Molecular docking experiments showed that a stable complex was formed between KOLR and RAPGEF3/PDE3B ( Figure S8 and Table 4S). Notably, two different binding modes were observed depending on the order of binding ( Figure S8 and Table 4S). These ndings showed KOLR targets to RAPGEF3/PDE3B complex to inhibit Rap-1 and PI3K-Akt signaling pathways.

Mutation of RAPGEF3 G557A affects targeting of KOLR
The ndings of the current study showed that AsPC-1 cells were resistant to KOLR (Table 3S), whereas expression of RAPGEF3 and RAP1A affected invasion of AsPC-1 cells but not apoptosis of these cells ( Figure S8). Protein levels of RAPGEF3, Akt, and RAP1A/B did not show differences after treatment with KOLR. Notably, PI3Kγ110 expression level was weakly decreased, however, PDE3B protein level in AsPC-1 cells was slightly increased by treatment with KOLR but not in a dose-dependent manner ( Figure 5G). Further analysis showed a mutation of RAPGEF3 G557A in AsPC-1 ( Figure 6A). Furthermore, the threedimensional structure RAPGEF3 WT and RAPGEF3 G557A were built and molecular dynamics of RAPGEF3 WT and RAPGEF3 G557A were performed. After the mutation to A557, the hydrophobic -CH 3 of the alanine side chain formed an interaction and steric hindrance with the polar hydrophilic -C=O-NH 2 of the side chain of the adjacent amino acid N552 ( Figure 6B). The average structure of RAPGEF3 (G557) was signi cantly different from the structure of the mutant RAPGEF3 (A557), mainly in the vicinity of residue 557 ( Figure 6C). This change in structure resulted in a signi cant phase shift of the overall structure. Moreover, size of the binding pocket of PDE3B and KOLR was signi cantly decreased in the wild-type protein ( Figure 6C). The global energy for molecular docking was signi cantly different between KOLR and wild-type and mutant RAPGEF3 ( Figure 10S). Structure of RAPGEF/PDE3B/KOLR complex was changed by G557A mutation ( Figure 6D). The -OH groups on KOLR formed polar hydrogen bonds with surrounding amino acids (S478, A527, R514, P529, Q555, and D558,) whereas the benzene ring of KOLR formed hydrophobic interactions with the surrounding hydrophobic amino acids (A553, C552, L520, P529, A527, C517, L480, and V473) in the wild-type RAPGEF3 ( Figure 10SC). In addition, the -OH on KOLR formed polar hydrogen bonds with surrounding amino acids (Q151, F237, R798, R801, D821, and A848), and the benzene ring of KOLR formed hydrophobic interactions with the surrounding hydrophobic amino acids (L804, K820, M822, M851) with the mutant RAPGEF3 G557A ( Figure 10SD). These ndings showed that KOLR binding to RAPGEF3/PDE3B complex and RAPGEF3 G557A mutant leads to KOLR resistance.

KOLR inhibits growth of transplanted tumor in vivo
To further explore the anti-tumor effect of KOLR in vivo, the anti-tumor effects of KOLR were determined using a mouse tumor growth model xenotransplant. Nude mice were subcutaneously administered with BxPC-3 cells to achieve approximately 100mm 3 tumor size. Mice were then randomly divided into three groups including: control, 25mg/kg, and 50mg/kg. KOLR was administered intraperitoneally for ve consecutive days and tumor sizes were recorded each day. Mice were sacri ced on the sixth day and analysis showed that KOLR exhibits anti-tumor e cacy at a concentration of 50mg/kg ( Figure 7A and B).
Western blot analysis showed that antitumor mechanism of KOLR in vitro was consistent with in vivo mechanisms ( Figure 7C). Consistently, immunohistochemistry (IHC) assay showed that expression levels of p-Akt and the proliferation marker Ki67 were all signi cantly decreased in vivo after administration with 50mg/kg KOLR, compared with levels in the control group ( Figure 7D). In summary, these ndings show that KOLR targets RAPGEF3/PDE3B complex and promotes binding of the complex to exhibit antitumor effects in vitro and in vivo ( Figure 7E).

Discussion
It is important to nd new targets or biomarkers for cancer therapy owing to limited e cacy of conventional therapy approaches. The ndings of the current study showed that the natural product KOLR extracted from Cinnamomum pauci orum Nees leaves targets RAPGEF3/PDE3B complex to exert cytotoxic effects on cancer cells. cAMP is an important intracellular second messenger that plays an important role in tumor growth and progression. Previous studies report that RAPGEF is an important anticancer target owing to its role in downstream pathways such as Rap-1, PI3K/Akt, EKR1/2, and β-catenin implicated in carcinogenesis and cancer progression [19]. Rap-1 and PI3K-Akt signaling pathways were signi cantly downregulated in BxPC-3 cells after KOLR treatment (Figure 3). The ndings indicated that RAPGEF3 is a potential therapeutic target for KOLR. The RAPGEF speci c inhibitor ESI-09 inhibits cell proliferation and migration in different cell lines [31][32][33]. The ndings of the current study showed that apoptosis was promoted and metastasis was inhibited in BxCP-3 cells knockout of RAPGEF3, however, knockout of RAPGEF3 in AsPC-1 cells only inhibited metastasis but did not affect apoptosis. In addition, AsPC-1 cells showed higher resistance rate to ESI-09 and KOLR compared with BxPC-3 cells. The ndings showed a difference in the role of RAPGEF3 in AsPC-1 compared with BxPC-3 cells. Further, RAPGEF3 G557A mutant was identi ed in AsPC-1 cells, and the mutation affected binding of the PDE3B/RAPGEF3 complex. The role of this mutation in promoting the function of AsPC-1 cells should be further explored.
Protein-protein interactions (PPIs) and protein complexes are important in regulation and execution of most biological processes. A study on cell adhesion reported a signaling complex containing PDE3 activity and RAPGEF3 in HEK 293T cells, and immunoblot analysis showed that PDE3 activity was attributed to the presence of PDE3B [25]. The N terminal fragment of PDE3B binds to RAPGEF3, and heterologous overexpression of the N terminal fragment of PDE3B antagonizes PDE3B/RAPGEF3 complex formation in cells [25,34]. In addition, PDE3B/RAPGEF3 signalosome modulates functions of human arterial endothelial cells (HAECs) implicated in angiogenesis by integrating cAMP-and PI3Kγ-encoded signals [34]. Current family-speci c PDE3 inhibitors such as cilostamide and milrinone inhibit activity of PDE3, irrespective of their intracellular signaling network and enzyme inhibition alone is not su cient for cell death [35].