Evaluation of Phosphodiesterase 3A as a Biomarker and Target for a Novel Cytotoxic Agent
Miryam S. Saad
Luc de Waal Broad Institute of Harvard and M.I.T. Dana-Farber Cancer Institute
Donald E. Elmore Biological Chemistry, Wellesley College
Wellesley College May 2014
Submitted as Partial Fulfillment of a B.A. Degree with Honors in Biological Chemistry
© 2014 Miryam S. Saad and Donald E. Elmore Acknowledgments
I would like to express my gratitude to Dr. Matthew Meyerson and Luc de Waal for the opportunity to participate in a very exciting and interesting research project and for mentoring me throughout the process. My most sincere thanks go to my adviser and mentor, Luc, for his guidance, support, and encouragement throughout the development of my thesis work.
I wish to extend my deepest appreciation to Prof. Don Elmore, and my thesis committee, Prof. Elizabeth Oakes and Prof. Mala Radhakrishnan, for their unconditional support and encouragement throughout the year.
Muchísimas gracias a mi querida amiga, Elizabeth Torres, por brindarme su apoyo, consejos, y amistad que hicieron de esta experiencia una aventura inolvidable. Gracias por ayudarme a sobrevivir este reto. Amigas como tu dejan rastros en el corazón.
Le estoy eternamente agradecida a mi familia por su amor y apoyo incondicional. A mi padre por darme la oportunidad de realizarme profesionalmente y demonstrarme que todo en esta vida es posible. A mi madre por su amor, orientación, y enseñanzas que me han formado en la mujer que soy hoy en día. A mis hermanos por siempre estar presentes en mi vida apoyandome mientras nado contra la corriente y sigo mi camino. Gracias por creer en mí. No se que sería de mí sin ustedes.
Y como decimos con ganas, pa’ arriba, pa’ abajo, pal centro y pa’ adentro.
! 2! Table of Contents
Abstract 5 Introduction 6 Materials and Methods 14 Cell Line and Culture Condition 14 Cell-growth Parameters 14 Compound 1B and Analogue Treatment 15 Compound 1B and PDE 3 Inhibitor Co-Treatment 16 Results 18 Discussion 24 Literature Cited 28
! 3! Tables and Figures
Figure 1: Compound 1B exhibits remarkable potency and selectivity against TP53- deficient cancer cell lines.
Figure 2: PDE3A is a putative target for Compound 1B inhibition.
Figure 3: Cyclic Nucleotide Phosphodiesterases (PDEs) catalyze the hydrolysis of cAMP and/or cGMP and regulate cyclic nucleotide-mediated cell signaling mechanisms.
Figure 4: Structural organization of the PDE superfamily and PDE3 family.
Figure 5: Spatial distribution of Milrinone and Cilostazol bound to the PDE3A catalytic site.
Figure 6: Cancer cell lines with varying PDE3A expression levels revealed varying sensitivity to Compound 1B in high-throughput screening.
Figure 7: Cancer cell lines with varying PDE3A expression levels exhibit varying sensitivity to Compound 1B and validate high-throughput screening results.
Figure 8: Compound 1B sensitive cell lines were rescued in co-treatment with non-lethal PDE3 inhibitors.
Table 1: Compound 1B and analogue potency in sensitive cell lines.
! 4! Abstract
Synthetic lethal screening of a molecular library discovered Compound 1B, a phosphodiesterase 3 (PDE3) inhibitor, to exhibit remarkable potency and selectivity against a subset of tumor suppressor TP53-deficient cancer cell lines. High PDE3A expression was correlated to Compound 1B sensitivity, thereby suggesting PDE3A as a biomarker and putative target for Compound 1B cytotoxicity. High-throughput screening results revealed subsets of cancer cell lines that supported the proposed mechanism or suggested alternative mechanisms for Compound 1B cytotoxicity. A representative panel of cancer cell lines was subjected to Compound 1B treatment to validate previous high- throughput screening results. Cell lines that were found to be sensitive were subjected to analogue treatment in order to evaluate structure-activity relationships. Finally, rescue experiments were performed using non-lethal PDE3 inhibitors in order to validate PDE3A on target cytotoxicity. Our results validated previous high-throughput screening results, suggested Compound 1B and analogues exhibit the same cytotoxic mechanism of action, and revealed that competition for PDE3A binding rescued cell death. These findings support PDE3A is the target for Compound 1B cytotoxicity, but reveal PDE3A is a relatively sensitive and specific biomarker for Compound 1B sensitivity. The implications of these findings suggest an unknown dependency exposed through PDE3A modulation, and suggest the prospects of targeting PDE3A for cancer treatment.
! 5! Introduction
Cancer is the leading cause of death worldwide, with 8.2 million deaths reported in 2012 1. The multistep development of cancer is characterized by its biological hallmarks: genomic instability, inflammation, sustained proliferative signaling, evading growth suppressors, resisting cell death, enabling replicative immortality, inducing angiogenesis, activating invasion and metastasis, reprogramming energy metabolism, evading immune destruction, and creating a tumor microenvironment which fosters acquisition of hallmarks traits 2. The development of cancer therapies has evolved from general cytotoxic chemotherapies to molecule-specific targeted therapies, which interfere with specific oncogenic pathways characteristic to a tumor genotype and provide a personalized approach to cancer therapy 3. However, the complexity of cancer development presents a challenge for the development of personalized cancer therapies, as each oncogenic phenotype exhibits particular biological characteristics that differentiate in clinical behavior and response to treatment 4–7. Challenges include the identification of molecular targets essential for cancer maintenance. Additional limitations include the identification of clinically useful biomarkers to diagnose the disease state, select an appropriate therapeutic regimen, and monitor the progression of the disease. The continuous investigation and understanding of cancer pathogenesis has allowed for the discovery and development of targeted drugs that exploit the genetic dependencies of cancer, and provide personalized cancer therapies with improved treatment efficacy and reduced toxicity 8,9.
Synthetic lethal screening of a molecular library discovered Compound 1B to
exhibit remarkable potency (EC50 < 100 nM) and selectivity against a subset of tumor
! 6! suppressor TP53 deficient cancer cells lines, while being inactive against other cell lines,
including immortalized but nontumorigenic cell lines (Figure 1). A Pearson correlation
revealed a strong positive correlation for Phosphodiesterase 3A (PDE3A) and suggested
PDE3A is a biomarker for Compound 1B sensitivity (Figure 2A) 10. Screening for
phosphodiesterase inhibitory activity revealed Compound 1B is a selective
Phosphodiesterase 3 (PDE3) inhibitor, thereby suggesting PDE3A as the putative target
for Compound 1B cytotoxicity (Figure 2B). Phosphodiesterases (PDEs) are enzymes that
regulate intracellular concentrations of signal transduction molecules, cyclic adenosine
and guanine 3',5' monophosphate (cAMP and cGMP), and mediates nucleotide-dependent
cell signaling mechanisms 11,12. The therapeutic success of PDE inhibition has been
recognized, as observed in the treatment of erectile dysfunction, pulmonary hypertension,
acute refractory cardiac failure, and intermittent claudication 13–16. However, the potential
of PDE inhibition for cancer treatment remains under investigation. The lack of isoform-
selective PDE inhibitors and the limited knowledge on the roles of PDEs in cancer
development reveal some of the challenges in targeting PDEs for therapeutic purposes
17,18. The discovery of PDE3A as a biomarker and putative target for Compound 1B A549 H1355 H1563* H1395 Onebinib 48H BEAS-2B H1437 IMR-90 H1819 150 WI-38 H2172 H1734* HCC827
% H2228 MCF-7 100 y t i
l PC-9 PC-3 i b
a A427 A375 i
V 50 HCC33 HepG2 H82 HT29 0 BL-2087 H596 BL-2122 HCC2218 -3 -2 -1 0 1 H441 MKN45 Log uM concentration H1623 Hs746t H2087 MFE319 A" B" H2122* HeLaHeLa** HCT-116 Figure' 1.' Compound' 1B'exhibits' remarkable' potency' and' selectivity' against' TP53 A" B" Figure' 2.' Phosphodiesterase' 3A' is' a' putative' target' for' Compound' 1B' inhibition.' (A)' A' Pearson' correlation' revealed' a' strong' positive' correlation' for' Phosphodiesterase' 3A' (PDE3A)' for' Compound' 1B' sensivitiy.' (B)'A'screen'for'PDE'inhibitory'activity'revealed'Compound'1B'almost'completely'inhibits'PDE3A' (95%)'and'PDE3B'(97%)'at'a'concentration'of'100'nM. ' ' !cytotoxicity is significant, as it reveals an unknown role of PDE3A in cancer development and suggests the prospects of targeting PDE3A for cancer treatment. PDEs are enzymes that catalyze the hydrolytic cleavage of the cyclic nucleotide 3’,5’ phosphodiester bond to form the corresponding inactive 5’-monophosphate, such that PDE activity decreases intracellular cyclic nucleotide levels, and regulate the amplitude and duration of the nucleotide-dependent signaling pathways (Figure 3A) 11,12. PDEs regulate cyclic nucleotide gradients within signaling complexes that compartmentalize unique combinations of PDEs with signaling effectors, such as Protein Kinase A (PKA), Protein Kinase G (PKG), cAMP-activated exchange factors (EPAC), and cyclic nucleotide-gated ion channels (CNG), and mediate protein-protein interactions for nucleotide-dependent signaling cascades (Figure 3B) 19–22. Examples of cell signaling responses include cellular proliferation, differentiation, apoptosis, inflammation, and vasodilation (Figure 3B) 23–25. The PDE family is comprised of 11 members that share similar structural characteristics but exhibit distinct physiological roles, tissue expression ! 8! X N N 5' N O O N R 4' 1' 3' 2' O P O H O PDE 2 OH Mg2+ X N N OH 5' N O P O O N R OH 4' 1' + H 3' 2' OHOH R X Adenine = H NH2 Guanine = NH2 0 A " B" Figure'3.'Cyclic'Nucleotide'Phosphodiesterases'(PDEs)'catalyze'the'hydrolysis'of'cAMP'and/or'cGMP'and' regulate'cyclic'nucleotide PDE family is comprised of 21 genes and multiple transcriptional products that result from alternative splicing or transcription from different start sites 21. The structural organizations of PDEs share the following characteristics: a C-terminal HD domain that shows high affinity for cAMP and/or cGMP; a highly conserved catalytic domain; and lastly, highly variable N-terminal regulatory protein domains that regulate PDE enzymatic activity and subcellular localization characteristic to each PDE family (Figure 4A) 18,21. The PDE catalytic domains constitute 16 α-helices that form an active-site pocket characterized by highly conserved hydrophobic residues and two conserved metal- binding sites 27,28. The two metal ions, Zn2+ and Mg2+, have been proposed to coordinate ! 9! with a water molecule or hydroxide ion that acts as a nucleophile in the hydrolytic cleavage of the cyclic nucleotide phosphodiester bond by PDE 29,30. In addition to cyclic nucleotide hydrolysis, the functional roles of PDEs as scaffold proteins or exhibiting allosteric regulation to alter protein complex formation and mediate signal transduction pathways have been previously investigated 17. The PDE3 family is characterized by its biochemical ability to hydrolyze both 31,32 cAMP and cGMP with high affinities (Km cAMP <0.4 μM; Km cGMP <0.3 μM) . A unique characteristic of the PDE3 family is the presence of a 44-amino acid insert in the catalytic domain, which constitutes a flexible loop on the surface of the enzyme that undergoes conformational change to regulate substrate binding and catalytic activity 33. A " B" Figure'4.'Structural'organization'of'the'PDE'superfamily'and'PDE3'family.'(A)'The'PDE'families'share'the' following'structural'characteristics :' a' C4terminal' HD'domain'(shown'in' blue)' with'high'affinity'for'cAMP' and/or' cGMP;' a' highly' conserved' catalytic' domain' (shown' in' teal);' and,' highly' variable' N4terminal' regulatory'protein'domains'(Lugnier,'2006).'(B)'The'PDE3'family'is'comprised'of'three'PDE3A'that'differ'in' their'N 4terminal'portions:'PDE3A1'contains'membrane'association'domains'NHR1'and'NHR2,'and'PKB'(P1),' PKA' (P2)' and' PKC' (P3)' phosphorylation' sites;' PDE3A2' contains' NHR2' and' PKA' (P2)' and' PKC' (P3)' phosphorylation'sites;'and,'PDE3A3'lacks'membrane'association'domains'and'phosphorylation'sites.'Only' 1'PDE3B'transcript'has'been'reported'from'the'PDE3B'gene.'' ! ! 10! The PDE3 family consists of two genes: PDE3A, which is highly expressed in platelets and cardiac tissue, and PDE3B, which is expressed in hepatocytes and adipose tissue 34. PDE3A and PDE3B share a similar structural organization but exhibit divergent N- terminal regulatory protein domains, including membrane-association domains NHR1 and NHR2, which target for subcellular localization, and phosphorylation sites for Protein Kinase A (PKA), Protein Kinase B (PKB) and Protein Kinase C (PKC), which regulate PDE enzymatic activity 35,36. Three PDE3A isoforms that result from alternative translation initiation sites within a single PDE3A gene have been identified 37. The three PDE3A variants differ in their N-terminal regulatory sequences: PDE3A1, which is localized to the intracellular membranes, contains membrane-association domains NHR1 and NHR2, and three sites of phosphorylation by PKA, PKB, and PKC; PDE3A2, which is localized to both membranes and cytosol, contains NHR2 and two sites of phosphorylation by PKA and PKC; and lastly, PDE3A3, which is localized to the cytosol, lacks both NHR1 and NHR2, and all three phosphorylation sites (Figure 4B) 36,38,39. The three PDE3A variants exhibit very similar catalytic activities and sensitivity to PDE3 inhibitors 39,40 . The development of PDE3 inhibitors has facilitated the investigation of the roles of PDE3 in regulating important cyclic nucleotide-mediated processes, including myocardial contractility, platelet aggregation, vascular and airway muscle relaxation, oocyte maturation, and inflammation 17,21,41. PDE3 selective inhibitors include Trequinsin, Milrinone, Cilostazol, and Cilostamide (Supplementary Appendix, Table S1). Milrinone (Primacor ®) and Cilostazol (Pletal ®) are well characterized PDE3 inhibitors and are therapeutically used in the treatment of heart failure and intermittent claudication due to ! 11! their antiaggregant and vasolidator properties 16,42–44. The binding of Milrinone or Cilostazol to a PDE3A model indicates that specific hydrogen bond interactions between the inhibitor and an invariant glutamine are required for substrate recognition and selectivity, while a planar ring structure for the inhibitor to be held tightly within the active site by the hydrophobic clamp is required for substrate binding (Figure 5A for Milrinone; Figure 5B for Cilostazol) 27,28,45. A comparison of the spatial distributions of cAMP and Cilostazol bound to the PDE3A active site suggests a portion of the Cilostazol molecule is exposed towards the exterior of the active site pocket, which may introduce a new interface to the surface of the enzyme (Figure 5C). The mechanism for PDE inhibition has been suggested to include other functional possibilities, beyond the inhibition of nucleotide hydrolysis, such as: obstruction of enzyme dimerization, inhibition of phosphorylation, or interference with allosteric activation or regulation, among others 46. A" B" C" Figure' 5.' Spatial'distribution'of'Milrinone'and'Cilostazol'bound'to'the'PDE3A'catalytic'site.'Molecular' modeling' of' the' PDE3A' active' site' illustrates' the' binding' of' PDE3' selective' inhibitors,' Milrinone' and' Cilostazol,'to'the'PDE3A'catalytic'site'(Zhang' et'al.,'2001).'PDE3A'residues'involved'in'binding'are'labeled.' (A)' Milrinone' and'(B)' Cilostazol' are' represented'by'large'balls;'green,'carbon' atom;' red,' oxygen'atom;' blue,'nitrogen'atom.'(C)'A'comparison' of'the'spatial' distributions'of'bound'cAMP,'represented' as'small' green' balls,' and' bound' Cilostazol,' represented' as' purple' balls,' illustrates' the' substrate' and' inhibitor' binding'positions'but'suggests'a'portion'of'Cilostazol'may'introduce'a'new'interface'to'the'surface'of'the' enzyme.''''' ' ! 12! The molecular mechanism involved in the potency and selectivity for Compound 1B cytotoxicity remains unknown. HTS results revealed subsets of cancer cell lines that supported contradictory mechanisms: high PDE3A sensitive cell lines support PDE3A as the target for cytotoxicity; low PDE3A sensitive cell lines suggest a non-specific cytotoxic mechanism; and, high PDE3A insensitive cell lines suggest PDE3A is required but not sufficient for Compound 1B cytotoxicity (Figure 6). Furthermore, the distinction of non-lethal PDE3 inhibitors, Trequinsin and Cilostamide, and lethal PDE3 inhibitors structurally related to Compound 1B, Zardaverine and Anagrelide, indicates Compound 1B cytotoxicity is not dependent on its inhibition of PDE3 cyclic nucleotide hydrolysis and suggests a specific structural characteristic of Compound 1B may be responsible for mediating the cytotoxic effects. A representative panel of high and low PDE3A expressing, Compound 1B sensitive and insensitive cancer cell lines were used in this study to validate HTS results, characterize the sensitivity and specificity of PDE3A as a biomarker, and further investigate PDE3A on target cytotoxicity. """"""A" """"""B" """"""C" """"""D" Figure'6.'Cancer'cell'lines'with'varying'PDE3A'expression'levels'revealed'varying'sensitivity'to'Compound' 1B'in'high Cell Line and Culture Condition. Lu-65 (HSRRB, Lu-65 JCRB0079), HuT-78 (ATCC, HuT-78 CRM-TIB-161), RVH-421 (RIKEN, RVH-421 ACC127), ML-1 (DSMZ, ML-1 ACC464) and MHH-NB-11 (DSMZ, MHH-NB-11 ACC157) were cultured in RPMI- 1640 with L-glutamine (Corning, 10-040-CV). COV-318 (ECACC, COV-318 07071803), CAS-1 (ICLC, CAS-1 HTL97009), and L3.3 (Academic Laboratory) were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, 11995-065). JHOM-1 cells (RIKEN, JHOM-1 RCB1676) were cultured in a 1:1 nutrient mixture of DMEM and Ham’s F12 (Gibco, 11765-047) and 1% HyClone non-essential amino acids (NEAA; Fisher Scientific, SH40003-12). G-631 cells (ATCC, G-361 CRL-1424) were cultured in McCoy’s 5A (modified) medium (Gibco, 16600-082). 8505C cells (HSRRB, 8505C JCRB0826) were cultured in Eagle’s Minimal Essential Medium α (EMEM; Gibco, A10490-01) supplemented with 1% HyClone NEAA. All media were supplied with 10% Fetal Bovine Serum (FBS; Sigma F4135) and 1% Penicillin-Streptomycin-Glutamine (Pen-Strep-Glut; Gibco, 10378-016). All cell lines were cultured in tissue culture flasks of 75 cm2 with their corresponding media and supplements, and were maintained at 37°C in a 5% CO2 humidified incubator. Cell-growth Parameters. All cell lines were grown to 90% confluence and counted in a Beckman Coulter Vi-CELL cell viability analyzer with a trypan blue dye exclusion method. The cells were plated at densities ranging from 500 to 5,500 cells per well, with 500 cells per well increments, in a Corning® 384-well plate (Sigma, CLS3570BC) in 40 µL of the corresponding growth medium. The 384-well plate was incubated for 72 hours at 37°C in a 5% CO2 humidified incubator. After 72 hours of incubation, the 384-well ! 14! plate was removed from the incubator and allowed to cool to room temperature for 30 minutes. Cell viability was assessed by adding a 1:1 mixture of 25% CellTiter-Glo® (Promega, G7570), diluted with room temperature Phosphate Buffer Saline (PBS; Corning, 21-040-CV), to each well in the 384-well plate, incubating for ten minutes at room temperature, and measuring the luminescence signal on a PerkinElmer EnVision® with US LUM settings for 0.1 second per well. Cell-growth data was subjected to non- linear regression curve analysis using the GraphPad™ Prism ® v6.0 Software (GraphPad Software, La Jolla, USA) in order to determine the optimal cell density for cytotoxic assessment. Compound 1B and Analogues Treatment. All cell lines were grown to 90% confluence and counted in a Beckman Coulter Vi-CELL cell viability analyzer with a trypan blue dye exclusion method. Cytotoxicity assays were performed at the optimal cell density for each cell line, that is, the cell density at the peak of the exponential growth curve but prior to pre-plateau growth phase. The following cell densities were used for the cytotoxic screening: 2000 8505C cells, 1500 CAS-1 cells, 3500 COV-318 cells, 2500 G- 361 cells, 3000 HuT-78 cells, 4000 JHOM-1 cells, 1500 L3.3 cells, 3000 Lu-65 cells, 3500 MHH-NB-11 cells, 3000 ML-1 cells, and 2500 RVH-421 cells per well. The specific cell density per each cell line was plated in a 384-well plate in 40 µL of the corresponding growth medium and incubated for 24 hours at 37°C in a 5% CO2 humidified incubator. Compound 1B (Enamine, T0503-7331), or synthesized analogues, TP7, TP26, or TP51, were added with final concentrations ranging from 300 pM to 10 µM using a medium intermediate dilution. The 384-well plate was incubated for 48 hours at 37°C in a 5% CO2 humidified incubator. After 48 hours of incubation, the 384-well ! 15! plate was removed from the incubator and allowed to cool to room temperature for 30 minutes. Cell viability was assessed by adding a 1:1 mixture of 25% CellTiter-Glo®, diluted 1:3 with room temperature PBS, to each well in the 384-well plate using a Rainin multichannel pipette, incubating for ten minutes at room temperature, and measuring the luminescence signal on a PerkinElmer EnVision® with US LUM settings for 0.1 second per well. Luminescence units were normalized to percent of maximum proliferation. Dose-response data was subjected to non-linear regression curve analysis using the GraphPad™ Prism ® v6.0 Software. The mean EC50 was calculated for each compound in each cell line using GraphPad™ Prism ® v6.0 Software. EC50 is defined as the drug concentration required for 50 percent of maximum effect, that is, the halfway point between the baseline and the maximum response observed for the dose-response curves. Refer to Supplementary Appendix, Methods S1 for the detailed cytotoxicity assay protocols. Compound 1B and PDE3 Inhibitor Co-Treatment. Compound 1B sensitive cell lines, 8505C, CAS-1, HuT-78, JHOM-1, L3.3, and RVH-421, were subjected to Compound 1B and non-lethal PDE3 inhibitor co-treatment. Inhibitor co-treatment was performed at the optimal cell density previously determined for each cell line, that is, the cell density at the peak of the exponential growth curve but prior to pre-plateau growth phase. The following cell densities were used for the inhibitor co-treatment: 2000 8505C cells, 1500 CAS-1 cells, 3000 HuT-78 cells, 4000 JHOM-1 cells, 1500 L3.3 cells, and 2500 RVH- 421 cells per well. The specific cell density per each cell line was plated in a 384-well plate in 40 µL of the corresponding growth medium and was allowed to incubate for 24 hours at 37°C in a 5% CO2 humidified incubator. After 24 hours of incubation, ! 16! Trequinsin (VWR, 80058-202) or Cilostamide (Sigma Aldrich, C7971) were added in concentrations ranging from 0.3 nM to 10 µM using a medium intermediate dilution containing the EC90 of Compound 1B observed in the cytotoxic screening per each sensitive cell line. EC90 is defined as the drug concentration required for 90 percent of maximum effect, as observed for dose-response curves. The 384-well plate was allowed to incubate for 48 hours at 37°C in a 5% CO2 humidified incubator. After 48 hours of incubation, the 384-well plate was removed from the incubator and allowed to cool to room temperature for 30 minutes. Cell viability was assessed by adding a 1:1 mixture of 25% CellTiter-Glo®, diluted 1:3 with room temperature PBS, to each well in the 384- well plate using a Rainin multichannel pipette, incubating for ten minutes at room temperature, and measuring the luminescence signal on PerkinElmer EnVision® with US LUM settings for 0.1 second per well. Luminescence units were normalized to percent of maximum proliferation and cell line sensitivity was determined by a non-linear regression curve analysis. Refer to Supplementary Appendix, Methods S2 for the detailed co-treatment assay protocols. ! 17! Results Cancer cell lines with varying PDE3A expression levels and sensitivity to Compound 1B were studied in order to evaluate PDE3A as a biomarker and target for Compound 1B cytotoxicity. A strong correlation for PDE3A expression in Compound 1B sensitive cell lines and the identification of Compound 1B as a selective PDE3 inhibitor suggested PDE3A as the putative target for Compound 1B cytotoxicity (Figure 2). HTS for Compound 1B cytotoxicity revealed high PDE3A expressing sensitive cell lines that support the proposed mechanism, as well as high PDE3A insensitive cell lines and low PDE3A sensitive cell lines that suggest alternative mechanisms (Figure 6). A representative panel of cell lines were investigated in this study: high PDE3A expressing, Compound 1B sensitive cell lines JHOM-1 and L3.3; high PDE3A expressing, Compound 1B insensitive cell lines ML-1, Lu-65, MHH-NB-11, and G-361; and, low PDE3A expressing, Compound 1B sensitive cell lines 8505C, CAS-1, HuT-78, RVH- 421, and COV-318. All cell lines used in this study are characterized in the Cancer Cell Line Encyclopedia (CCLE) 10. Cell-growth parameters for each cell line were evaluated using non-linear regression curve analysis in order to determine the optimal cell density for cytotoxic assessment (Supplementary Appendix, Figure S1). The optimal cell densities used for this study are indicative of the cell culture density at which the exponential cell growth curve peaks, prior to the pre-plateau growth phase, for each cell line. Drug treatments were performed at the optimized cell density in order to maximize the dynamic range for the cytotoxicity assay and accurately evaluate drug-induced effects from cell viability dose-response curves. ! 18! All cell lines were subjected to Compound 1B treatment in order to validate previous results obtained in HTS results for Compound 1B cytotoxicity and evaluate PDE3A as a biomarker for sensitivity. High PDE3A expressing cell lines, JHOM-1 and L3.3, revealed sensitivity to Compound 1B in HTS and supported the proposed mechanism of PDE3A as the target for Compound 1B cytotoxicity (Figure 7A). JHOM-1 and L3.3 exhibited sensitivity to Compound 1B treatment and validated HTS results (Figure 7B). JHOM-1 and L3.3 support PDE3A as the target for Compound 1B cytotoxicity. Low PDE3A expressing cell lines, 8505C, CAS-1, HuT-78, RVH-421 and COV-318, revealed sensitivity to Compound 1B in HTS, thereby suggesting an alternative mechanism for Compound 1B cytotoxicity (Figure 7A). 8505C, CAS-1, HuT- 78, and RVH-421 exhibited sensitivity to Compound 1B treatment and validated HTS results. COV-318, however, did not exhibit sensitivity to Compound 1B treatment and was the only cell line to contradict HTS results (Figure 7C). High PDE3A expressing cell lines, ML-1, Lu-65, MHH-NB-11, and G-361, did not reveal sensitivity to Compound 1B in HTS, thereby suggesting an alternative mechanism for Compound 1B cytotoxicity (Figure 7A). ML-1, Lu-65, MHH-NB-11, and G-361 did not exhibit sensitivity to Compound 1B treatment, thereby validating HTS results (Figure 7D). All cancer cell lines subjected to Compound 1B treatment supported HTS results, with the exception of COV-318, thereby indicating the validity of HTS results. An evaluation of PDE3A as a biomarker reveals PDE3A is relatively sensitive and specific biomarker for Compound 1B sensitivity. Sensitivity is defined as the percentage of true-positive results, that is, the percentage of Compound 1B sensitive cell lines that exhibit high PDE3A expression, whereas specificity is defined as the percentage of true-negative results, that is, the ! 19! percentage of low PDE3A expressing cell lines that did not exhibit Compound 1B sensitivity. PDE3A exhibits 23.5% sensitivity and 98.7% specificity as a biomarker, as 23.5% of sensitive cell lines exhibited high PDE3A expression levels while 98.7% of insensitive cell lines exhibited low PDE3A expression levels (Figure 7A). However, PDE3A sensitivity is relative based on the PDE3A expression levels that serve as the standards to predict for Compound 1B sensitivity. Our findings indicate PDE3A is not a good predictive biomarker for Compound 1B sensitivity. A" B" Compound 1B 48H 150 JHOM-1 L3.3 100 Viability % 50 0 -4 -3 -2 -1 0 1 2 Log µM Concentration C" Compound 1B 48H D" Compound 1B 48H 150 150 RVH-421 ML-1 CAS-1 Lu-65 100 HuT-78 100 G-361 8505C MHH-NB-11 COV-318 Viability % Viability % 50 50 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration Figure' 7.' A' representative' panel' of' cancer' cell' lines' with' varying' PDE3A' expression' and' sensitivity' to' Compound'1B'validate'high ! 20! In order to investigate cytotoxic activity, Compound 1B sensitive cell lines were subjected to analogue treatment to evaluate structure-activity relationships. Structure- activity analysis was performed using three analogues with stronger, intermediate, and weak potencies in order to validate their activities and compare their cytotoxic behaviors across all cancer cell lines. Potency is inversely related to the EC50 value, such that a lower EC50, or drug concentration required for 50 percent of maximum effect, is indicative of a stronger potency. The following analogues were used, in their order of potency as observed for HeLa cells: TP51 > Compound 1B > TP26 > TP7 (Table 1). High PDE3A expressing cell lines, JHOM-1 and L3.3, and low PDE3A expressing cell lines, 8505C, CAS-1, HuT-78, and RVH-421, were subjected to TP51, TP26, and TP7 treatment and exhibited sensitivity (Supplementary Appendix, Figure S2). Structure- activity analysis indicated Compound 1B and analogues exhibit the same potency or EC50 trend for all cell lines and did not reveal off-target activities (Table 1). These findings suggest Compound 1B and analogues exhibit a similar mechanism of action for cytotoxicity across all sensitive cell lines. Table'1.'Compound'1B'and'analogue'TP51,'PT26,'and'TP6'activity'in'sensitive'cell'lines.'*'' ' O N O N O Cl O N O N NH N N N NH N NH ' ' F ' O2N ' Cell'Line' ''''''''''''''TP51'EC50'''''''>''''''''Compound'1B'EC50'''''''>'''''''TP26'EC50'''''''''''''''''>'''''''''''''''''TP7'EC50' HeLa' 1'nM' 10'nM' 30'nM' >'10'µM' JHOM41' 8'nM' 40'nM' 102'nM' >'1'µM' L3.3' 9'nM' 42'nM' 63'nM' >'1'µM' RVH4421' 5'nM' 30'nM' 32'nM' >'1'µM' CAS41' 12'nM' 35'nM' 143'nM' >'1'µM' HuT478' 24'nM' 47'nM' 134'nM' >'1'µM' 8505C' 30'nM' 57'nM' 154'nM' >'1'µM' *'Activity'refers'to'the'EC50'value,'the'drug'concentration'required'for'50'percent'of'maximum'effect.'' ! 21! Lastly, to evaluate PDE3A on target cytotoxicity, a rescue experiment was performed using non-lethal PDE3 inhibitors, Trequinsin and Cilostamide, to compete against Compound 1B for binding at the PDE3A catalytic site. Previous experiments revealed PDE3 inhibitors Trequinsin and Cilostamide did not exhibit lethal effects against HeLa cells, and thus are referred to as “non-lethal” PDE3 inhibitors, whereas Zardaverine, which is structurally related to Compound 1B, exhibited lethal effects (Data not shown). Compound 1B sensitive cell lines were subjected to a drug co-treatment using the EC90 of Compound 1B, as determined for each cell line, and increasing concentrations of Trequinsin or Cilostamide in order to evaluate cell viability effects resulting from competition for binding at the PDE3A catalytic site. EC90 is defined as the Compound 1B concentration required for 90 percent of maximum effect, as determined from dose-response curves for each cell line. High PDE3A expressing cell lines, JHOM-1 and L3.3, and low PDE3A expressing cell lines, 8505C, CAS-1, HuT-78, and RVH-421, were rescued in co-treatment with Trequinsin or Cilostamide (Figure 8; Supplementary Appendix, Figure S3). These findings support PDE3A is the target for Compound 1B Trequinsin 48H Cilostamide 48H 150 150 Cell'Line' EC90'Cell' JHOM-1 JHOM-1 Viability'%' L3.3 L3.3 JHOM41' 73'%' RVH-421 100 100 RVH-421 L3.3' 66'%' CAS-1 CAS-1 RVH4421' 79'%' 8505C HuT-78 CAS41' 65'%' Viability % 50 HuT-78 Viability % 50 8505C 8505C' 56'%' HuT478' 45'&' 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration A" "B" C" C" Figure'8 .'Compound'1B'sensitive'cell'lines'were'rescued'in'co Compound 1B cytotoxicity is not dependent upon its inhibition of PDE3 cyclic nucleotide hydrolysis, as other PDE3 inhibitors do not exhibit cytotoxic effects. The validation of HTS results revealed PDE3A exhibits relative sensitivity and specificity as a biomarker. Our findings indicate PDE3A exhibits 23.5% sensitivity and 98.7% specificity as a biomarker and suggest PDE3A is not a good predictive biomarker for Compound 1B sensitivity. Despite its limitations, PDE3A is the only known biomarker for Compound 1B cytotoxicity. Structure-activity analysis indicated Compound 1B and analogues exhibit similar cytotoxic activities, thereby suggesting Compound 1B and analogues exhibit the same cytotoxic mechanism of action in sensitive cell lines. Rescue experiments with non-lethal PDE3 inhibitors revealed that competition against Compound 1B for binding at the PDE3A catalytic site rescues cell death, thereby revealing PDE3A is the target for Compound 1B cytotoxicity. In summary, these findings support PDE3A is the target for Compound 1B cytotoxicity, but indicate PDE3A is relatively sensitive and specific biomarker for Compound 1B sensitivity. ! 23! Discussion An evaluation of PDE3A as a biomarker for Compound 1B sensitivity revealed PDE3A is relatively sensitivity and is specific as a biomarker for Compound 1B sensitivity. The proposed mechanism for Compound 1B cytotoxicity suggested PDE3A as the target; however, the validation of HTS results suggested the possibility of alternative mechanisms, based on PDE3A expression and sensitivity to Compound 1B. Structure-activity analysis and competition for Competition for PDE3A binding confirmed PDE3A on target cytotoxicity and revealed Compound 1B cytotoxicity is not dependent on the inhibition of PDE3A catalytic activity. These results suggest PDE3A is a potential therapeutic target for anticancer therapy, but suggest PDE3A is not a good predictive biomarker for sensitivity and may not be clinically useful, on its own, for biomarker-based patient selection. PDE3A is the only identified biomarker for Compound 1B sensitivity and exhibits relative sensitivity and specificity to predict for Compound 1B sensitivity. Even though PDE3A is not an ideal biomarker, PDE3A is a biomarker that can be further evaluated for drug development and clinical trials. The significance of predictive biomarkers as companions of targeted drugs is illustrated in the cases of targeted drugs Imatinib (Gleevec®) and Erlotinib (Tarceva®). Imatinib, a tyrosine kinase inhibitor, was specifically designed to inhibit the kinase activity of BCR-ABL, a translocation protein found in 95% of chronic myeloid leukemia (CML) 47,48. The discovery of BCR-ABL as a predictive biomarker allowed for biomarker-based patient selection in clinical trials, which facilitated the proof of imatinib clinical benefit for CML 49. The case of imatinib reflects the prospects of PDE3A as a relevant biomarker for drug development. The case ! 24! of erlotinib, however, reflects limitations of biomarkers. Erlotinib is an inhibitor of tyrosine kinase activity in epidermal growth factor receptor (EGFR) used for the treatment of non-small cell lung cancer (NSCLC) 50. Even though EGFR overexpression was initially a candidate predictive biomarker, EGFR overexpression was not a consistent biomarker and, overall, did not correlate erlotinib sensitivity 51. A recent study indicated EGFR overexpression exhibited 63.4% sensitivity and 76.8% specificity as a biomarker for erlotinib sensitivity 52. Even though PDE3A is more specific as a biomarker than EGFR overexpression, PDE3A is not a very sensitive biomarker and may not be useful, on its own, for clinical trials. PDE3A should be further investigated to evaluate the cytotoxic mechanism attributed to PDE3A inhibition and characterize downstream biomarkers to predict clinical response or with the potential to serve as therapeutic targets. PDE3A was identified as the target for Compound 1B cytotoxicity. However, the specific PDE3A isoform responsible for mediating the cytotoxic effects of Compound 1B remains unknown. Compound 1B and analogues are presumed to target all PDE3A isoforms, which exhibit indistinguishable sensitivity to PDE3 inhibitors 39. The three PDE3A isoforms, PDE3A1, PDE3A2, and PDE3A, are identical except for different lengths of their N-terminal regulatory sequences and distribution of membrane- association domains, which target for intracellular location, and phosphorylation sites, which regulate protein-protein interactions and catalytic activities. Recent studies indicated PDE3A1 and PDE3A2 are phosphorylated at alternative sites via different signaling pathways and exhibit distinct protein-protein interactions, thereby suggesting PDE3A1 and PDE3A2 exhibit distinct roles in regulating cyclic nucleotide signaling 36. ! 25! Even though Compound 1B and analogues are proposed to target all three PDE3A variants, the PDE3A isoform responsible for mediating the cytotoxic effects remains unknown. Even though Compound 1B is proposed to target all three PDE3A variants, our findings indicate that Compound 1B does not exhibit its cytotoxic effects via its inhibition of PDE3A cyclic nucleotide hydrolysis, as non-lethal PDE 3 inhibitors competed for PDE3A binding and rescued cell death. The discovery of PDE3A as a biomarker and target for Compound 1B cytotoxicity is significant as it reveals an unknown role for PDE3A in cancer development and suggests the prospects of targeting PDE3A for cancer treatment. Compound 1B sensitivity is not lineage-specific, thereby suggesting Compound 1B exhibits its cytotoxic effects by interfering with a specific pathway necessary for cancer development across all sensitive cell lines. The identification of non-lethal PDE3 inhibitors and lethal PDE3 inhibitors structurally related to Compound 1B suggests cytotoxicity is independent of PDE3 inhibition of catalytic activity and may dependent upon a specific structural characteristic of Compound 1B. A recent study indicated Zardaverine, a dual PDE3/4 inhibitor structurally related to Compound 1B, revealed potent and selective antitumor activity against hepatocellular carcinoma (HCC) independently of PDE3/4 inhibition 53. These findings support Zardaverine as a lethal PDE3 inhibitor and suggest cytotoxicity may be dependent on a structural characteristic of Zardaverine, as proposed for Compound 1B. The possibilities for Compound 1B cytotoxicity include alterations in protein-protein interactions that affect signaling cascades and disruption of PDE cross-talk or pharmacodynamics, such that the cytotoxic effects of PDE3A inhibition may mediated via the alteration of regulation or indirect ! 26! inhibition of another PDE within the same compartment 26,54. The implications of these findings support an unknown role for PDE3A in cancer development and indicate PDE3A is a potential target for anticancer therapies. The mechanisms for Compound 1B cytotoxicity should be further investigated to characterize the effects of Compound 1B in PDE3A regulation of protein interactions or identify downstream predictive biomarkers or therapeutic targets for anticancer activity. ! 27! Literature Cited 1. World Health Organization. GLOBOCAN 2012: Estimated Cancer Incidence, Mortality and Prevalence Worldwide in 2012. Int. Agency Res. Cancer (2012). at 2. Weinberg, R. A. & Hanahan, D. Hallmarks of Cancer: The Next Generation. Cell 144, 646–674 (2011). 3. Gerber, D. E. Targeted therapies: a new generation of cancer treatments. Am. Fam. Physician 77, 311–319 (2008). 4. Peppercorn, J., Perou, C. M. & Carey, L. A. Molecular subtypes in breast cancer evaluation and management: divide and conquer. Cancer Invest. 26, 1–10 (2008). 5. Van de Vijver, M. et al. A Gene-Expression Signature as a Predictor of Survival in Breast Cancer. N. Engl. J. Med. 347, 1999–2009 (2002). 6. Rosenwald, A. et al. The Use of Molecular Profiling to Predict Survival after Chemotherapy for Diffuse Large-B-Cell Lymphoma. N. Engl. J. Med. 346, 1937–1947 (2002). 7. Potti, A. et al. A Genomic Strategy to Refine Prognosis in Early-Stage Non–Small-Cell Lung Cancer. N. Engl. J. Med. 355, 570–580 (2006). 8. Hoelder, S., Clarke, P. A. & Workman, P. Discovery of small molecule cancer drugs: Successes, challenges and opportunities. Mol. Oncol. 6, 155–176 (2012). 9. Schilsky, R. L. Personalized medicine in oncology: the future is now. Nat. Rev. Drug Discov. 9, 363–366 (2010). 10. Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 483, 603–7 (2012). 11. Essayan, D. M. Cyclic nucleotide phosphodiesterases. J. Allergy Clin. Immunol. 108, 671–680 (2001). 12. Rall, T. & Sutherland, E. Formation of a cyclic adenine ribonucleotide by tissue particles. J. Biol. Chem. 232, 1065–1076 (1958). 13. Corbin, J. D. Mechanism of action of PDE5 inhibiton in erectile dysfunction. Int. J. Impot. Res. 16, S4–7 (2004). 14. Wilkins, M. R., Wharton, J., Grimminger, F. & Ghofrani, H. A. Phosphodiesterase inhibitors for the treatment of pulmonary hypertension. Eur. Respir. J. 32, 198–209 (2008). 15. Guazzi, M. Clinical Use of Phosphodiesterase-5 Inhibitors in Chronic Heart Failure. Circ. Hear. Fail. 1, 272–280 (2008). 16. Dobesh, P. P., Stacy, Z. A. & Persson, E. L. Pharmacologic therapy for intermittent claudication. Pharmacotherapy 29, 526–53 (2009). ! 28! 17. Bender, A. T. & Beavo, J. A. Cyclic nucleotide phosphodiesterases: molecular regulation to clinical use. Pharmacol. Rev. 58, 488–520 (2006). 18. Lugnier, C. Cyclic nucleotide phosphodiesterase (PDE) superfamily: a new target for the development of specific therapeutic agents. Pharmacol. Ther. 109, 366–398 (2006). 19. Houslay, M. D., Baillie, G. S. & Maurice, D. H. cAMP-Specific phosphodiesterase-4 enzymes in the cardiovascular system: a molecular toolbox for generating compartmentalized cAMP signaling. Circ. Res. 100, 950–966 (2007). 20. Kritzer, M. D., Li, J., Dodge-Kafka, K. & Kapiloff, M. S. AKAPs: the architectural underpinnings of local cAMP signaling. Curr. Top. Med. Chem. 52, 531–358 (2012). 21. Kotera, J. & Omori, K. Overview of PDEs and Their Regulation. Circ. Res. 100, 209–327 (2007). 22. Baillie, G. S. Compartmentalized signalling: spatial regulation of cAMP by the action of compartmentalized phosphodiesterases. FEBS J. 276, 1790–9 (2009). 23. Bos, J. L. Epac Proteins: Multi-purpose cAMP Targets. Trends. Biochem. Sci. 31, 680–6 (2006). 24. Conti, M. & Beavo, J. Biochemistry and physiology of cyclic nucleotide phosphodiesterases: essential components in cyclic nucleotide signaling. Annu. Rev. Biochem. 76, 481–511 (2007). 25. Francis, S., Blount, M. & Corbin, J. Mammalian cylic nucleotide phosphodiesterases: molecular mechanisms and physiological functions. Physiol. Rev. 91, 651–690 (2011). 26. Keravis, T. & Lugnier, C. Cyclic nucleotide phosphodiesterase (PDE) isozymes as targets of the intracellular signalling network: benefits of PDE inhibitors in various diseases and perspectives for future therapeutic developments. Br. J. Pharmacol. 165, 1288–305 (2012). 27. Beavo, J. A., Francis, S. H. & Houslay, M. D. Cyclic Nucleotide Phosphodiesterases in Health and Disease. 728 (CRC Press, 2010). 28. Zhang, K. Y. J. et al. A Glutamine Switch Mechanism for Nucleotide Selectivity by Phosphodiesterases. Mol. Cell. 15, 279–286 (2004). 29. Xu, R. X. et al. Atomic Structure of PDE4: Insights into Phosphodiesterase Mechanism and Specificity. Science (80-. ). 288, 1822–1825 (2000). 30. Huai, W., Colicelli, J. & Ke, H. The crystal structure of AMP-bound PDE4 suggests a mechanism for phosphodiesterase catalysis. Biochemistry 42, 13220–6 (2003). 31. Degerman, E., Belfrage, P., Newman, A. H., Rice, K. C. & Manganiello, V. C. Purification of the putative hormone-sensitive cyclic AMP phosphodiesterase from rat adipose tissue using a derivative of cilostamide as a novel affinity ligand. J. Biol. Chem. 262, 5797–5807 (1987). 32. Grant, P. G. & Colman, R. W. Purification and characterization of a human platelet cyclic nucleotide phosphodiesterase. Biochemistry 23, 1801–1807 (1984). 33. Hung, S. H. et al. New Insights from the Structure-Function Analysis of the Catalytic Region of Human Platelet Phosphodiesterase 3A: A Role for the Unique 44-Amino Acid Insert. J. Biol. Chem. 281, (2006). ! 29! 34. Shakur, Y. et al. Regulation and function of the cyclic nucleotide phosphodiesterase (PDE3) gene family. Prog. Nucleic Acid Res. Mol. Biol. 66, 241–77 (2000). 35. Ahmad, F., Manganiello, V. C. & Degerman, E. Cyclic Nucleotide Phosphodiesterase 3 Signaling Complexes. Horm. Metab. Res. 44, 776–785 (2012). 36. Vandeput, F. et al. Selective regulation of cyclic nucleotide phosphodiesterase PDE3A isoforms. Proc. Natl. Acad. Sci. U. S. A. 110, 19778–19783 (2013). 37. Bender, A. T. & Beavo, J. A. Cyclic nucleotide phosphodiesterases: molecular regulation to clinical use. Pharmacol. Rev. 58, 488–520 (2006). 38. Yan, C., Miller, C. L. & Abe, J. Regulation of Phosphodiesterase 3 and Inducible cAMP Early Repressor in the Heart. Circ. Res. 100, 489–501 (2007). 39. Hambleton, R. et al. Isoforms of cyclic nucleotide phosphodiesterase PDE3 and their contribution to cAMP hydrolytic activity in subcellular fractions of human myocardium. J. Biol. Chem. 280, 39168–39174 (2005). 40. Kenan, Y., Murata, T., Shakur, Y., Degerman, E. & Manganiello, V. C. Functions of the N- terminal region of cyclic nucleotide phosphodiesterase 3 (PDE 3) isoforms. J. Biol. Chem. 275, 12331–12338 (2000). 41. Thompson, P. E., Manganiello, V. & Degerman, E. Re-discovering PDE3 inhibitors: new opportunities for a long neglected target. Curr. Top. MCdicinal Chem. 7, 421–436 (2007). 42. Kambayashi, J. et al. Cilostazol as a unique antithrombotic agent. Curr. Pharm. Des. 9, 2289–302 (2003). 43. Cruickshank, J. M. Phosphodiesterase III inhibitors: long-term risks and short-term benefits. Cardiovasc. Drugs. Ther. 7, 655–60 (1993). 44. Sircar, I., Weishaar, R. E., Kobylarz, D., Moos, W. H. & Bristol, J. A. Cardiotonic agents. 7. Inhibition of separated forms of cyclic nucleotide phosphodiesterase from guinea pig cardiac muscle by 4,5-dihydro-6-[4-(1H-imidazol-1-yl)phenyl]-3(2H)-pyridazinones and related compounds. Structure-activity relationships and correl. J Med Chem 30, 1955–1962 (1987). 45. Zhang, W. E. I., Ke, H., Tretiakova, A. P., Jameson, B. & Colman, R. W. Identification of overlapping but distinct cAMP and cGMP interaction sites with cyclic nucleotide phosphodiesterase 3A by site-directed mutagenesis and molecular modeling based on crystalline PDE4B. 1481–1489 (2001). doi:10.1101/ps.6601.inhibition 46. DeNinno, M. P. Future directions in phosphodiesterase drug discovery. Bioorg. Med. Chem. Lett. 22, 6794–800 (2012). 47. Druker, B. J. Imatinib alone and in combination for chronic myeloid leukemia. Semin Hematol 40, 50–8 (2003). 48. Park, J. W. et al. Rationale for Biomarkers and Surrogate End Points in Mechanism-Driven Oncology Drug Development. Clin. Cancer Res. 10, 3885 (2004). ! 30! 49. Chau, C. H., Rixe, O., McLeod, H. & Figg, W. D. Validation of Analytical Methods for Biomarkers Employed in Drug Development. Clin Cancer Res 14, 5967–5976 (2008). 50. Rocha-Lima, C. M. & Raez, L. E. Erlotinib (Tarceva) for the treatment of Non-Small-Cell Lung Cancer and Pancreatic Cancer. Pharm. Ther. 34, 554–556, 559–564 (2009). 51. Lenz, H.-J. Biomarkers in Oncology: Prediction and Prognosis. 447 (Springer, 2012). 52. Masica, D. L. & Karchin, R. Collections of simultaneously altered genes as biomarkers of cancer cell drug response. Cancer Res. 73, 1699 (2013). 53. Sun, L. et al. Phosphodiesterase 3/4 Inhibitor Zardaverine Exhibits Potent and Selective Antitumor Activity against Hepatocellular Carcinoma Both In Vitro and In Vivo Independently of Phosphodiesterase Inhibition. PLoS One 9, e90627 (2014). 54. Bischoff, E. Potency, selectivity, and consequences of nonselectivity of PDE inhibition. Int. Jounral Impot. Res. 16, S11–S14 (2004). ! 31! Supplementary Appendix Table of Contents Page Supplementary Appendix, Methods S1: Detailed Cytotoxicity Assay Protocols. 2 Supplementary Appendix, Methods S2: Detailed Co-Treatment Assay Protocols. 7 Supplementary Appendix, Figure S1: Cell-growth curves. 10 Supplementary Appendix, Figure S2: Compound 1B and analogue treatment cell 11 viability curves. Supplementary Appendix, Figure S3: Compound 1B and PDE3 inhibitor co- 12 treatment cell viability curves. Supplementary Appendix, Table S1: PDE3 selective inhibitors. 13 References 14 Supplementary Appendix, Methods S1: Detailed Cytotoxicity Assay Protocols 8505C cytotoxicity assay using CellTiter-Glo® Day 0: 8505C cells (HSRRB, 8505C JCRB0826) are grown to 90% confluence in EMEM, 1% NEAA, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (2000 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. CAS-1 cytotoxicity assay using CellTiter-Glo® Day 0: CAS-1 cells (ICLC, CAS-1 HTL97009) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (1500 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. COV-318 cytotoxicity assay using CellTiter-Glo® Day 0: COV-318 cells (ECACC, COV-318 07071803) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (3500 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing ! 2! media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. G-361 cytotoxicity assay using CellTiter-Glo® Day 0: G-631 cells (ATCC, G-361 CRL-1424) are grown to 90% confluence in McCoy’s 5A, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (2500 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. HuT 78 cytotoxicity assay using CellTiter-Glo® Day 0: HuT 78 cells (ATCC, HuT 78 CRM-TIB-161) are grown to 90% confluence in RPMI-1640, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (3000 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. ! 3! JHOM-1 cytotoxicity assay using CellTiter-Glo® Day 0: JHOM-1 cells (RIKEN, JHOM-1 RCB1676) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (4000 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. L3.3 cytotoxicity assay using CellTiter-Glo® Day 0: L3.3 cells (Academic Laboratory) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (1500 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with a Rainin multichannel pipette and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. LU65 cytotoxicity assay using CellTiter-Glo® Day 0: LU65 cells (HSRRB, LU65 JCRB0079) are grown to 90% confluence in 1:1 DMEM and Ham’s F12, 1% NEAA, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (3000 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated ! 4! well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. MHH-NB-11 cytotoxicity assay using CellTiter-Glo® Day 0: MHH-NB-11 cells (DSMZ, MHH-NB-11 ACC157) are grown to 90% confluence in RPMI-1640, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (3500 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. ML-1 cytotoxicity assay using CellTiter-Glo® Day 0: ML-1 cells (DSMZ, ML-1 ACC464) are grown to 90% confluence in RPMI- 1640, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (3000 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. ! 5! RVH-421 cytotoxicity assay using CellTiter-Glo® Day 0: RVH-421 cells (RIKEN, RVH-421 ACC127) are grown to 90% confluence in RPMI-1640, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (2500 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Compound 1B, TP7, TP26, and TP51 using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media, and further dilute 1:5 by adding 10 μL of drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution will be 1000x (10 μM – 0.3 nM). Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. ! 6! Supplementary Appendix, Methods S2: Detailed Co-Treatment Assay Protocols 8505C co-treatment cytotoxicity assay using CellTiter-Glo® Day 0: 8505C cells (HSRRB, 8505C JCRB0826) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (2000 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Trequinsin and Cilostamide using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media containing Compound 1B EC90 of 100 nM, observed in the preliminary cytotoxic screening. Further dilute 1:5 by adding 10 μL of the drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution of Trequinsin and Cilostamide will be in the range of 10 μM – 0.3 nM, while the final concentration of Compound 1B will be a constant 100 nM. Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. CAS-1 co-treatment cytotoxicity assay using CellTiter-Glo® Day 0: CAS-1 cells (ICLC, CAS-1 HTL97009) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (1400 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Trequinsin and Cilostamide using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media containing Compound 1B EC90 of 100 nM, observed in the preliminary cytotoxic screening. Further dilute 1:5 by adding 10 μL of the drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution of Trequinsin and Cilostamide will be in the range of 10 μM – 0.3 nM, while the final concentration of Compound 1B will be a constant 100 nM. Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. ! 7! HuT-78 co-treatment cytotoxicity assay using CellTiter-Glo® Day 0: HuT-78 cells (ATCC, HuT 78 CRM-TIB-161) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (3000 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Trequinsin and Cilostamide using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media containing Compound 1B EC90 of 300 nM, observed in the preliminary cytotoxic screening. Further dilute 1:5 by adding 10 μL of the drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution of Trequinsin and Cilostamide will be in the range of 10 μM – 0.3 nM, while the final concentration of Compound 1B will be a constant 300 nM. Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. JHOM-1 co-treatment cytotoxicity assay using CellTiter-Glo® Day 0: JHOM-1 cells (RIKEN, JHOM-1 RCB1676) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (4000 cells per well) in 40 μl of culturing media using Corning® 384- well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Trequinsin and Cilostamide using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media containing Compound 1B EC90 of 300 nM, observed in the preliminary cytotoxic screening. Further dilute 1:5 by adding 10 μL of the drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution of Trequinsin and Cilostamide will be in the range of 10 μM – 0.3 nM, while the final concentration of Compound 1B will be a constant 300 nM. Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. L3.3 co-treatment cytotoxicity assay using CellTiter-Glo® Day 0: L3.3 cells (Academic Laboratory) are grown to 90% confluence in DMEM, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (1500 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. ! 8! Day 2: Dilute the stock Trequinsin and Cilostamide using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media containing Compound 1B EC90 of 300 nM, observed in the preliminary cytotoxic screening. Further dilute 1:5 by adding 10 μL of the drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution of Trequinsin and Cilostamide will be in the range of 10 μM – 0.3 nM, while the final concentration of Compound 1B will be a constant 300 nM. Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. RVH-421 co-treatment cytotoxicity assay using CellTiter-Glo® Day 0: RVH-421 cells (RIKEN, RVH-421 ACC127) are grown to 90% confluence in RPMI-1640, 10% FBS, 1% Pen-Strep-Glut. Day 1: Plate cells (2500 per well) in 40 μl of culturing media using Corning® 384-well plate and incubate for 24 hours at 37°C in a 5% CO2 and 95% humidity incubator. Day 2: Dilute the stock Trequinsin and Cilostamide using a serial dilution to obtain concentrations ranging from 10 mM to 1 μM of drug in DMSO. Dilute each concentration of a log and half-log serial dilution of compound by 1:200 in culturing media containing Compound 1B EC90 of 30 nM, observed in the preliminary cytotoxic screening. Further dilute 1:5 by adding 10 μL of the drug into the designated 40 μL plated well (50 μl total volume). Perform at least 8 replicates per concentration per drug. The final dilution of Trequinsin and Cilostamide will be in the range of 10 μM – 0.3 nM, while the final concentration of Compound 1B will be a constant 30 nM. Day 4: Remove the 384-well plate from the incubator and allow to cool for 30 minutes to room temperature. Add 40 μl of a 25% Promega CellTiter-Glo® solution (diluted 1:3 with room temperature PBS) with Rainin multichannel and incubate for 10 minutes at room temperature. Read on Perkin-Elmer EnVision with US LUM settings for 0.1 second per well. ! 9! Supplementary Appendix, Figure S1: Cell-growth curves. A" B C" 8505C 72H CAS-1 72H COV-318 72H 1500000 200000 800000 150000 600000 1000000 100000 400000 500000 50000 200000 Luminescence Units Luminescence Luminescence Units Luminescence Luminescence Units Luminescence 0 0 0 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Cells/well Cells/well Cells/well D" E F" G-361 72H HuT-78 72H JHOM-1 72H 800000 1500000 500000 400000 600000 1000000 300000 400000 200000 500000 200000 100000 Luminescence Units Luminescence Luminescence Units Luminescence Units Luminescence 0 0 0 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Cells/well Cells/well Cells/well G" H" I" L3.3 72H Lu-65 72H MHH-NB-11 72H 200000 2000000 800000 150000 1500000 600000 100000 1000000 400000 50000 500000 200000 Luminescence Units Luminescence Luminescence Units Luminescence Units Luminescence 0 0 0 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Cells/well Cells/well Cells/well J" K" ML-1 72H RVH-421 72H 150000 1500000 100000 1000000 50000 500000 Luminescene Unita Luminescene Luminescence Units Luminescence 0 0 0 2000 4000 6000 0 2000 4000 6000 Cells/well Cells/well Supplementary,Appendix,,Figure,S1:!Cell7growth,curves.,8505C,!CAS)1,!COV)318,!G)361,!HuT)78,!JHOM) 1,!L3.3,!Lu)65,!MHH)NB)11,!ML)1,!and!RVH)421,were!plated!at!densities!ranging!from!500!to!5,500!cells! per!well,!with!500!cells!per!well!increments,!and!incubated!for!72!hours!at!37°C.,Error!bars!represent!S.D.! of!mean,!n!=!32!replicates.,(A)!8505C,,(B),CAS)1,,(C),COV)318,,(D),G)361,,(E),HuT)78,,(F),JHOM)1,,(G),L3.3,, (H), Lu)65,, (I), MHH)NB)11,, (J)!ML)1,!and, (K), RVH)421!cell!growth!parameters!were!determined!by!non) linear!regression!curve!analysis.!, ! 10! Supplementary Appendix, Figure S2: Compound 1B and analogue treatment cell viability curves. A" 8505C 48H B" CAS-1 48H C" HuT-78 48H 150 150 150 1B 1B-PAR 1B-PAR TP7 TP7 TP7 100 100 100 TP51 TP51 TP51 TP26 TP26 TP26 Viability % 50 Viability % 50 Viability % 50 0 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration Log µM Concentration D" JHOM-1 48H E" L3.3 48H F" RVH-421 48H 150 150 150 1B-PAR 1B-PAR 1B TP7 TP7 TP7 100 100 100 TP51 TP51 TP51 TP26 TP26 TP26 Viability % 50 Viability % 50 Viability % 50 0 -4 -3 -2 -1 0 1 2 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration Log µM Concentration G" COV-318 48H H" G-361 48H I" Lu-65 48H 150 150 150 1B-PAR 1B 1B-PAR TP7 TP7 100 TP7 TP51 100 100 TP51 TP51 TP26 TP26 TP26 Viability % 50 Viability % 50 Viability % 50 0 0 -4 -3 -2 -1 0 1 2 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration Log µM Concentration J" MHH-NB-11 48H K" ML-1 48H 150 150 1B-PAR TP7 TP7 100 100 TP51 TP51 TP26 TP26 Viability % 50 1B-PAR Viability % 50 0 -4 -3 -2 -1 0 1 2 0 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration Supplementary,Appendix,,Figure,S2:!Compound,1B,and,analogue,treatment,cell,viability,curves.,8505C,, CAS)1,,HuT)78,,JHOM)1,,L3.3,!RVH)421,!COV)318,,G)361,,Lu)65,,MHH)NB)11,!and,ML)1!were!subjected!to! Compound!1B,!TP7,!TP26,!or!TP51!(0.3!nM!)!10!mM)!for!48H., Error!bars!represent!S.D.!of!mean,!n!=!8! replicates.,(A)!8505C,,(B),CAS)1,,(C),HuT)78,,(D),JHOM)1,,(E),L3.3,!and,(F),RVH)421!exhibited!sensitivity!to! Compound!1B!or!analogue!TP7,!TP26,!and!TP51!treatment.,(G),COV)318,,(H),G)361,,(I),Lu)65,,(J)!MHH)NB) 11,!and,(K),ML)1!did!not!exhibit!sensitivity!to!Compound!1B!or!analogue!treatment.!! ! 11! Supplementary Appendix, Figure S3: Compound 1B and PDE3 inhibitor co- treatment cell viability curves. A" 8505C 48H B" CAS-1 48H 150 150 Trequinsin 100 100 Trequinsin Cilostamide Cilostamide Viability % 50 Viability % 50 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration C" HuT-78 48H D" JHOM-1 48H 150 150 Trequinsin Trequinsin 100 100 Cilostamide Cilostamide Viability % Viability % 50 50 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration E" L3.3 48H F" RVH-421 48H 150 150 Trequinsin 100 100 Trequinsin Cilostamide Cilostamide Viability % 50 Viability % 50 0 0 -4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2 Log µM Concentration Log µM Concentration Supplementary, Appendix,, Figure, S3:! Compound, 1B, and, PDE3, inhibitor, co7treatment, cell, viability, curves., 8505C,, CAS)1,, HuT)78,, JHOM)1,, L3.3,! and! RVH)421! were! subjected! to! Compound! 1B! and! non) lethal!PDE3!inhibitors,!Trequinsin!or!Cilostamide!(0.3!nM!)!10!mM),!for!48H.,Error!bars!represent!S.D.!of! mean,!n!=!16!replicates., (A)!8505C,, (B), CAS)1,, (C),HuT)78,, (D),JHOM)1,, (E),L3.3,!and, (F),RVH)421!were! rescued!in!co)treatment!with!Compound!1B!and!Trequinsin!or!Cilostamide.!! ! 12! Supplementary Appendix, Table S1: PDE 3 selective inhibitors. ! Name, Structure, Reference, ! EC50, ! Trequinsin, 0.25!nM!! (1)!! O 65! ! O ! N ! N N ! O ! ! OPC733540, 0.32!nM! (2)! O ! 67! O N N ! H OH ! N O ! H ! ! Levosimendan, !! 2.4!nM!! (3)! N N NH H 69! ! N O ! N ! !! N ! Cilostamide, O 27!nM!71! (4)! ! O ! N N O ! H ! ! Anagrelide, Cl 36!nM!73! (5)! Cl N ! O ! N N H ! ! Milrinone! ! 56!nM!66! (6)! ! N ! CN ! H3C N O ! H ! ! Siguazodan! O 117!nM! (7)! ! NH 68! ! NH N NC ! N N H ! ! Cilostazol! 200!nM! (8)! ! 70! O N ! N N N ! O N H ! ! Zardaverine! ! 2.5!µM!72! (9)! ! F2HCO O ! N NH ! MeO ! Amrinone! H 4.8!µM!74! (10)! ! O N ! ! H2N N ! ! ! ! 13! References 1. Ruppert, D. & Weithmann, K. U. HL 725, an extremely potent inhibitor of platelet phosphodiesterase and induced platelet aggregation in vitro. Life Sci 31, 2037–2043 (1982). 2. Sudo, T. et al. Potent effects of novel anti-platelet aggregatory cilostamide analogues on recombinant cyclic nucleotide phosphodiesterase isozyme activity. Biochem Pharmacol 59, 347– 56 (2000). 3. Szilagyi, S. et al. The effects of levosimendan and OR-1896 on isolated hearts, myocyte-sized preparations and phosphodiesterase enzymes of the guinea pig. Eur J Pharmacol. 486, 67–74 (2004). 4. Hidaka, H. et al. Selective inhibitor of platelet cyclic adenosine monophosphate phosphodiesterase, cilostamide, inhibits platelet aggregation. J Pharmacol Exp Ther 211, 26–30 (1979). 5. Gillespie, E. Anagrelide: a potent and selective inhibitor of platelet cyclic AMp phosphodiesterase enzyme activity. Biochem Pharmacol 37, 2866–8 (1988). 6. Alousi, A. A., Stankus, G. P., Stuart, J. C. & Walton, L. H. Characterization of the cardiotonic effects of milrinone, a new and potent cardiac bipyridine, on isolated tissues from several animal species. J Cardiovasc Pharmacol 5, 804–11 (1983). 7. Christensen, S. B. & Torphy, T. J. Isozyme-selective phosphodiesterase inhibitors as antiasthmatic agents. J. Cardiovasc. Pharmacol. 21, 983–995 (1993). 8. Tanaka, T. et al. Effects of cilostazol, a selective cAMP phosphodiesterase inhibitor on the contraction of vascular smooth muscle. Pharmacology 36, 313–320 (1988). 9. Schudt, C., Winder, S., Muller, B. & Ukena, D. Zardaverine as a selective inhibitor of phosphodiesterase isozymes. Biochem. Pharmacol. 42, 153–162 (1991). 10. Bottorff, M. B., Rutledge, D. R. & Pieper, J. A. Evaluation of intravenous amrinone: the first of a new class of positive inotropic agents with vasodilator properties. Pharmacotherapy 5, 227 (1985). ! 14!