Methods for identifying compounds for inhibition of neoplastic lesions, and pharmaceutical compositions containing such compounds

IS 2 174 573 T3

DESCRIPTION

Procedures for the identification of substances for the inhibition of neoplastic lesions.

Background of the invention

This invention relates to the use of one or more forms of phosphodiesterase type 2 ("PDE2") and phosphodiesterase type 5 ("PDE5") and / or protein kinase G to identify useful compounds for the treatment and they prevented precancerous and cancerous lesions in mammals.

Currently, nonsurgical treatment of cancer involves administering one or more highly toxic chemotherapeutic products or hormonal therapies to the patient after cancer has progressed to a point where the therapeutic benefits of chemotherapy / hormone therapy outweigh its side effects. very serious. Such side effects are very well known to any oncoologist and vary from drug to drug. However, standard chemotherapy products are used topically only for short periods of time, often alternating chemotherapy with periods of no treatment, in order not to overwhelm the patient with the side effects of the drugs. Therefore, given the risk-benefit compensation, the collateral effects make it impossible to topically start chemotherapy when patients have precancerous lesions, or to continue chemotherapy or hormonal therapy chronically once the overt cancer has been eliminated. in an attempt to prevent its recurrence.

Starting about a decade ago, a ray of hope began to appear from an unexpected source: non-steroidal anti-inflammatory drugs (“NSAIDs”). Research on cancer and pre-cancer is replete with publications describing various biochemical molecules that are overexpressed in neoplastic tissue, leading group after group to investigate whether specific overexpressed molecules are responsible for the disease, and the fact that if such overexpression were inhibited, the neoplasm could be cured. For example, in familial adenomatous polyposis ("FAP"), Waddell in 1983 (Waddell, WR et al., "Sulindac for Polyposis of the Colon", Journal of Surgical Oncology, 24: 83-87, 1983) hypothesized that as prostaglandins were overexpressed in such polyps, non-steroidal anti-inflammatory drugs ("NSAIDs") should alleviate the condition. because NSAIDs inhibited prostaglandin synthetase (PGE2) activity. Therefore, he administered the non-steroidal anti-inflammatory drug ("NSAID") sulindac (a PGE2 inhibitor) to several patients with FAP. Waddell found that the polyps remitted and did not recur after such therapy. Inhibition of PGE2 results from inhibition of cyclooxygenase (COX) by NSAIDs. Waddell's success with sulindac and the PGE2 / COX relationship apparently confirmed the role of two other biochemical targets - PGE2 and COX - in carcinogenesis, and subsequent literature reinforced these criteria.

The ray of hope for patients suffering from neoplasia was that sulindac certainly had far fewer side effects than conventional chemotherapeutic or hormonal products, and it opened the possibility of treating cancer in the earlier stages of the disease and for longer periods of time. in comparison with conventional chemotherapeutic products. However, such hope had to be mitigated by the open question of whether a compound such as sulindac could be used to treat overt cancer, since Waddell had administered sulindac only to patients with a precancerous condition, FAP.

That hope was also mitigated by the NSAIDs' own set of side effects. Sulindac and other NSAIDs when administered chronically irritate the digestive tract in which PGE2 has a protective role. Furthermore, when taken chronically, they present collateral effects involving the kidney and interference with normal blood clotting. As Waddell unfortunately experienced, some of his sulindac patients stopped taking the drug due to side effects (see Waddell, WR et al., "Sulindac for Polyposis of the Colon," The American Journal of Surgery, 157: 175-79 , 1989), most likely returning to additional surgical interventions to control the formation of polyps. Therefore, for patients with neoplasia these drugs are not a practical chronic treatment, for example, for FAP, sporadic polyps or for men after a prostatectomy with increasing PSAs (an increasing PSA in such individuals indicates the recurrence of the disease, which may not present yet as a manifest, visible cancer). These side effects also limit the use of NSAIDs for any other indication in neoplasms that require long-term drug administration. More recently, some have suggested the use of COX-2 specific NSAIDs such as celecoxib. However, the renal and other side effects of such compounds are believed to limit the dosage and duration of treatment with such compounds for long-term antineoplastic indications. In addition, recently published data indicate that very high doses are needed for drugs such as celecoxib to reach

ES 2 174 573 T3 a marginal effect on colon polyps in only predefined regions of the colon-rectum. Perhaps more significant for the treatment of colon cancer is the fact that certain colonic neoplasms (e.g., HCT-116) have been reported not to express COX-2, and that such inhibitors are ineffective against such neoplasms (see, Sheng and col., "Inhibition of Human Colon Cancer Cell Growth By Selective Inhibition of Cyclooxygenase-2", J. Clin. Invest., 99 (9): 2254-9, 1997).

Recent discoveries have alienated scientists from COX / PEG targets₂as these targets may not be the primary (or perhaps even secondary) targets for successfully treating patients with chronic malignancies. Pamukcu et al., In US Patent No. 5,401,774, describe that sulfonyl compounds, which have been reported to lack virtually PEG inhibitory activity₂ and COX (and thus were not NSAIDs or anti-inflammatory compounds), unexpectedly inhibited the growth of a variety of neoplastic cells, including cells of colon polyps. These sulfonyl derivatives have been found to be effective in rat colon carcinogenesis models, and a variant (currently referred to as exisulind) has been shown to be effective in human clinical trials with PAF patients, and even more notably has shown an effect on overt cancer: the own prostate cancer, in a controlled clinical study presented later. Furthermore, very recent research has convincingly established that COX-I and / or COX-2 are not substantially expressed in all neoplasms, lowering the hope that a specific COX-I or COX-2 inhibitor will be broadly speaking. Therapeutically useful in the treatment of neoplasms (see, Lim et al., "Sulindac Derivatives Inhibit Growth and Induce Apoptosis in Human Prostate Cancer Cell Lines", Biochem. Pharmacology, Vol. 58, pp. 1097-1107 (1999) in press ).

Thus, like many other proteans overexpressed in neoplasms, overexpression of PGE2 / COX may not be the cause of some neoplasms, but rather a consequence of some of them. But the combination of such discoveries, however, has raised the question: how do compounds such as exisulind (which have a range of activity against COX-expressing and non-COX-expressing neoplasms) act? What do such compounds do to neoplastic cells?

Piazza et al. (in US Patent Applications No.^you 08 / 866.027 and 09 / 046.739) found that compounds (such as exisulind) inhibit calcium GMP-specific phosphodiesterase (for example, PDE5), and that other such compounds could be screened using that enzyme, which could lead to discovery of other compounds that could be developed and formulated in antineoplastic pharmaceutical compositions. Such pharmaceutical compositions can be highly antineoplastic and can be practically devoid of the side effects associated with conventional chemotherapeutic products, or even the side effects of COX or PGE2 inhibition, if such side effects are to be avoided. Furthermore, antineoplastic compounds that inhibit cGMP-specific PDE can induce apoptosis (a form of programmed cell death or suicide) in neoplastic cells, but not in normal cells. Thus, such new compounds have been referred to as a new class of antineoplastics known as selective apoptotic antineoplastic drugs ("SAANDs"). Consequently, SAANDs have challenged several issues of conventional wisdom: (1) that antineoplastic compounds cannot be effective without destroying normal cells as well; (2) that COXs are responsible for neoplasms and (3) that the prevention of colonic neoplasia by NSAIDs is probably mediated by the inhibition of one or both types of COX.

The new research presented below has shown, however, that not all compounds that exhibit classical PDE5 inhibition induce apoptosis in neoplastic cells. For example, the well-known PDE5 inhibitors, zaprinast and sildenafil, do not individually induce apoptosis and do not even inhibit the growth of neoplastic cells in our hands. However, since pro-apoptosis PDE5 inhibitors selectively induce apoptosis (that is, in neoplastic cells but not in normal cells), and can do so without substantial COX inhibition, the usefulness of PDE5 as a screening tool for antineoplastic compounds is not in question. desirable.

However, an improvement of the PDE5 selection method is desirable in order to find pro-apoptosis but safe antineoplastic compounds, such that new pharmaceutical compositions can be formulated for therapeutic use in the treatment of neoplasms, including pre-cancer and cancer.

Summary of the invention

In the course of investigating why some PDE5 inhibitors individually induce apoptosis while others did not, we discovered a form of calcium GMP-specific phosphodiesterase activity that had not been previously described. This new phosphodiesterase activity had not been previously characterized. Without being limited to a specific theory, we believe

ES 2 174 573 T3 that this new PDE activity may be a new conformation of PDE2 that substantially lacks cAMP hydrolyzing activity, that is, it is cGMP specific. Classic PDE2 is not cGMP specific (it also hydrolyzes in cAMP); classic PDE2 is also found in neoplastic cells. This new PDE and PDE2 are useful in selecting pharmaceutical compounds for their desirable antineoplastic properties. Basically, in neoplastic cells when PDE5 activity and PDE2 activity (in their new conformation and in their conventional conformation) are inhibited by a PDE5 inhibitor antineoplastic compound, the result is apoptosis. When only PDE5 is inhibited (but not the different forms of PDE2), apoptosis does not take place.

In its broadest aspects, this new PDE conformation has an activity characterized by:

(a) specificity of cGMP for cAMP;

(b) positive helper kinetic behavior in the presence of the cGMP substrate;

(c) submicromolar affinity for cGMP; and (d) insensitivity to incubation with purified cGMP-dependent protein kinase.

Other characteristics of this new PDE include: it has a reduced sensitivity to inhibition by zaprinast and E4021, it can be separated from the classical PDE5 activity by anioon exchange chromatography, it is not activated by calcium / calmodulin and it is sensitive to rolipram, vinpocetine and indolidane. .

Another embodiment of this invention includes evaluating whether a compound produces an increase in cGMP-dependent protein kinase G ("PKG") activity and / or a decrease in β-catenin in neoplastic cells. Unexpected features of SAANDs have been found to include elevated PKG activity and decreased β-catenin in neoplasm cells exposed to SAAND. We believe that the elevation of PKG activity is due, at least in part, to the increase in cGMP caused by the inhibition of appropriate PDEs by SAANDs, as previously described. The other characteristics of SAANDs are (1) inhibition of PDE5 as described in the prior '694 patent; (2) inhibition of the new cGMP-specific PDE conformation; (3) PDE2 inhibition; (4) the fact that they increase intracellular cGMP in neoplastic cells and (5) the fact that they decrease cAMP levels in some types of neoplastic cells.

Therefore, one embodiment of the new method of this invention is the evaluation of whether a compound causes PKG activity to rise in neoplastic cells and whether that compound inhibits PDE5. Another embodiment of the novel selection method of this invention is the evaluation of whether a compound increases PKG activity in neoplastic cells and whether that compound inhibits the new cGMP-specific PDE described above and / or PDE2. Still a third embodiment is the evaluation of whether a compound causes PKG activity to rise in neoplastic cells and whether that compound causes cGMP to rise in neoplastic cells and / or causes cAMP levels to decrease. Compounds successfully evaluated in such ways have applications as SAANDs.

Among other things, this invention relates to new in vitro and in vivo methods for screening compounds for their ability to safely treat and prevent neoplasms, especially precancerous lesions. In particular, the present invention is a method for selecting compounds that can be used to treat and prevent neoplasms, including precancerous lesions. The compounds thus identified may have mononymous side effects attributable to COX inhibition and other nonspecific interactions associated with conventional chemotherapeutic agents. Compounds of interest can be tested by exposing the new PDE described above to the compounds, and if a compound inhibits this new PDE, the compound is subsequently further evaluated (for example, in in vitro or in vivo animal or human models or analytical assays. ) to determine its antineoplastic properties.

One aspect of this invention therefore involves a screening / selection method to identify an effective compound for the treatment of neoplasms that includes the determination of the inhibition by the compound of this new PDE and / or PDE2 and its inhibition of COX. . Preferably, the screening and selection methods of this invention further include determining whether the compound inhibits the growth of tumor cells in vitro or in vivo.

By selecting compounds in this way, potentially beneficial and improved compounds can be identified to treat neoplasms more rapidly and with greater precision than was possible in the past to develop pharmacoeutical compositions and treat neoplasms therapeutically. More benefits will be obvious from the detailed description below.

IS 2 174 573 T3

Brief description of the figures

Figure 1 is a graph of cGMP activities of cGMP phosphodiesterases obtained from SW480 neopiasic cells, assayed from the eluent of a DEAE-Trisacryl M column.

Figure 2 is a graph of cGMP activities of recharged cGMP phosphodiesterases obtained from SW480 neoplastic cells, assayed from the eluent of a DEAETrisacryl M column.

Figure 3 is a graph of the kinetic behavior of the new PDE of this invention.

Figure 4 illustrates the effect of sulindac sulfide derivative and sulindac sulfone derivative (also called exisulind) on purified cyclooxygenase activity.

Figure 5 illustrates the effects of test compounds B and E on COX inhibition.

Figure 6 illustrates the inhibitory effects of sulindac sulfide and exisulind on PDE4 and PDE5 purified from cultured tumor cells.

Figure 7 illustrates the effects of sulindac sulfide on cyclic nucleotide levels in HT-29 cells.

Figure 8 illustrates the phosphodiesterase inhibitory activity of compound B.

Figure 9 illustrates the phosphodiesterase inhibitory activity of compound E.

Figure 10 illustrates the effects of sulindac sulfide and exisulind on apoptosis and necrosis of HT-29 cells.

Figure 11 illustrates the effects of sulindac sulfide and exisulind on the inhibition of HT-29 cell growth and the induction of apoptosis determined by DNA fragmentation.

Figure 12 illustrates the apoptosis-inducing properties of compound E.

Figure 13 illustrates the apoptosis-inducing properties of compound B.

Figure 14 illustrates the effects of sulindac sulfide and exisulind on tumor cell growth.

Figure 15 illustrates the growth inhibitory activity and the apoptosis inducing activity of sulindac sulfide and the control (DMSO).

Figure 16 illustrates the growth inhibitory activity of compound E.

Figure 17 illustrates the inhibition of premalignant neoplastic lesions in culture of rat mammary gland organisms by sulindac metabolites.

Figure 18A is an SDS gel of SW480 cell lysates proteins from drug-treated cell lysates in the absence of added cGMP, in which cells were cultured for 48 hours with DMSO (0.03%; lanes 1 and 2), exisulind (200, 400 and 600 μΜ; lanes 3, 4, 5) and E4021 (0.1, 1 and 10 μΜ; lanes 6, 7, 8).

Figure 18B is an SDS gel PKG assay (X-ray film challenge) of SW480 cell lysates from drug-treated cell lysates in the presence of added cGMP, in which cells were cultured for 48 hours with DMSO (0.03%; lanes 1 and 2), exisulind (200, 400 and 600 μΜ; lanes 3, 4, 5) and E4021 (0.1, 1 and 10 μΜ; lanes 6, 7, 8 ).

Figure 19 is a bar graph of the results of Western blot experiments of the effects of exisulind on β-catenin and PKG levels in neoplastic cells relative to control.

Figure 20 is a graph of cGMP activities of cGMP phosphodiesterases obtained from HTB-26 neoplastic cells, assayed from the eluent of a DEAE-Trisacryl M column.

IS 2 174 573 T3

Figure 21 is a graph of cGMP activities of cGMP phosphodiesterases obtained from HTB-26 neopiasic cells, assayed from the eluent of a DEAE-Trisacryl M column with low and high substrate concentration.

Figure 22 is a graph of cGMP activities of cGMP phosphodiesterases obtained from LNCaP neoplastic cells, assayed from the eluent of a DEAE-Trisacryl M column.

Figure 23 is a graph of cGMP activities of cGMP phosphodiesterases obtained from LNCaP neoplastic cells, assayed from the eluent of a DEAE-Trisacryl M column with low and high substrate concentration.

Figure 24 is a bar graph illustrating the unioan specificity of non-catalytic cGMP binding sites of PDE5 for cyclic nucleatide analogs and selected inhibitors of PDE5.

Figure 25 is a graph of cGMP activities of cGMP phosphodiesterases obtained from SW480 neoplastic cells, assayed from the eluent of a DEAE-Trisacryl M column using ethylan glycol in the buffer.

Figure 26 is a graph of cGMP activities of cGMP phosphodiesterases obtained from SW480 neoplastic cells grown in roller bottles, assayed from the eluent of a DEAE-Trisacryl M column.

Figure 27A shows a time-dependent increase in the amount of histone-associated fragmented DNA in LNCaP cell cultures after treatment with Compound I 50 µΗ.

Figure 27B shows the course of treatment of PrEC prostatic cells with Compound I (50 µM) which does not affect DNA fragmentation during a treatment of up to 4 days.

Detailed description of the preferred embodiments

I. The new neoplastic cell PDE2 and cGMP specific phosphodiesterase

A. Isolation of the new PDE conformation

Isolated cGMP specific phosphodiesterase (which appears to be a new conformation of PDE2) was first prepared from the human carcinoma cell line commonly referred to as SW480, available from the American Tissue Type Collection in Rockville, Maryland, USA. SW480 is a human colon cancer cell line originating from a moderately differentiated epithelial adenocarcinoma. As discussed later, a similar conformation has also been isolated from breast (ie, the HTB-26 cell line) and prostate (ie, LNCaP cell line) neoplasms.

By "isolated" we mean (it is still understood in the art) not only isolated from neoplastic cells but also produced by recombinant methods (for example, expressed in vector cell lines of a bacterial host or other non-human host). However, we currently believe that the isolation of the human neoplastic cell line is preferable, since we believe that the target protein thus isolated has a structure (that is, a conformation or topography) that is more proximal, if not identical. , to one of the native conformations in the neoplastic cell. This conformation aids in the selection of antineoplastic compounds that will inhibit the target enzyme (s) in vivo.

The new PDE activity will be found for the first time in SW480 colon cancer cell lines. To isolate the novel SW480 phosphodiesterase, approximately four hundred million SW480 cells were grown to confluence and collected by scraping from 150 cm tissue culture plates.² after two washes with 10 ml of cold PBS and agglutinated by centrifugation. The cells were resuspended in homogenization buffer (20 ml of TMPI-EDTA-Triton pH 7.4: 20 mM Tris-HOAc, MgAc₂ 5 mM, 0.1 mM EDTA, 0.8% Triton-100, 10 μM benzamidine, 10 μM TLCK, 2000 U / ml aprotinin, 2 μM leupeptin, 2 μM pepstatin A) and homogenized in an ice bath using a Polytron tissue homogenizer (three times, 20 seconds / pulse). The homogenate was centrifuged at 105,000 g for 60 minutes at 4 ° C in a Beckman L8 ultracentrifuge, and the supernatant was diluted with TMPI-EDTA (60 ml) and applied to a pre-equilibrated 10-milliliter DEAE-Trisacryl M column. with TMPI-EDTA buffer. The loaded column was washed with 60 ml

ES 2 174 573 T3 of TM-EDTA and PDE activities were eluted with a linear gradient of 120 ml of NaOAC (0-0.5

M) in TM-EDTA, at a flow rate of 0.95 ml / minute, 1.4 ml / fraction. Eighty fractions were collected and analyzed for cGMP hydrolysis immediately (ie, in minutes). Figure 1 shows the elution profile of the column, revealing two initial peaks of cGMP PDE activity, Peaks A and B, which were eluted by 40-50 mM and 70-80 mM NaOAC, respectively. As explained later, Peak A is PDE5, while Peak B is a new specific phosphodiesterase activity of cGMP.

The coclic nucleootide PDE activity of each fraction was determined using the modified two-step radioisotope method of Thompson et al. (Thompson, WJ et al., Adv. Cyclic Nucleotide Res., 10: 69-92, 1979), as further described below. The reaction was carried out in 400 µ! containing Tris-HCl (40 mM; pH 8.0), MgCl₂ (5 mM), 2-mercaptoethanol (4 mM), bovine serum albumin (30 µg), cGMP (0.25 µM-5 µM) with constant tritiated substrate (200,000 cpm). The incubation time was adjusted to give less than 15% hydrolysis. The mixture was incubated at 30 ° C followed by boiling for 45 seconds to stop the reaction. Subsequently, the mixture was cooled, snake venom (50 µg) was added and the mixture was incubated at 30 C for 10 minutes. MeOH (1 ml) was added to stop the reaction, and the mixture was transferred to an anion exchange column (Dowex

1-X8, 0.25 ml resin). The eluent was combined with a second ml of MeOH, applied to the resin, and after adding 6 ml of scintillation fluid, the tritium activity was measured using a Beckman LS 6500 counter for one minute.

In order to fractionate the cGMP hydrolotic activities of Peaks A and B later, fractions 15 to 30 of the original 80 were reloaded onto the DEAE-Trisacryl M column and eluted with a linear gradient of NaOAC (0 -0.5 M) in TM-EDTA. The fractions were again immediately analyzed for cGMP hydrolysis (using the procedure described above with 0.2, 2.5 μM substrate), the results of which are graphically presented in Figure 2. An observation on Peak B illustrated in Figure 2 is that the increase in the concentration of cGMP substrate dramatically increased the activity when compared to Peak A. Although this observation is consistent with the fact that it was a PDE2, the fact that the enzyme characterized in Figure 2 is specific for cGMP (see further below) suggests that it has a new conformation compared to the classic PDE2 described in the literature. . The activity of Peak A shows an apparent substrate saturation of the high affinity catalytic sites.

B. Isolation of PDE2 classics from SW480

Two methods were found that allowed the isolation of "Peak B" of SW480 so that the enzyme had classic PDE2 activity (that is, it was not specific for cGMP but was stimulated by cGMP). The first method involved growing SW480 in Corning 850 cm rotary bottles.² instead of in 150 cm tissue culture flasks². The SW480s were grown in rotary bottles at 0.5 rpm, each bottle containing 200 ml of RPMI 1640, 2 mM glutamine and 25 mM HEPES. Cells were collected by the following procedure. PBS medium was heated at 37 C for at least 15 minutes. 200 ml of RPMI 1640 complete medium / 5% FBS were prepared and 5 ml of glutamine was added. 5 ml of antibiotic / antifungal were also added.

70 ml of the PBS solution was added to 10 ml of 4X Pancreatin. The mixture was kept at room temperature. The medium was removed and the bottle was washed with 4 ml of PBS, making sure that the bottom of the bottle was covered. All the solution was removed with a pipette. 4 ml of diluted Pancreatin was added to the flask and the flask was shaken to cover its bottom. The flask was incubated at 37 ° C for 8-10 minutes. After incubation, the flask was quickly checked under an inverted microscope to ensure that all cells were rounded. The jar was carefully tapped on the side several times to help detach the cells. 10 ml of cold complete medium was added to the flask to stop pancreatin proteolysis. The solution was shaken on the bottom to collect the cells. The medium was removed using a 25 ml pipet and the cells were placed in 50 ml centrifuge tubes on ice. The tubes were centrifuged at 1000 rpm at 4 ° C for 5 minutes in a clonal centrifuge to agglutinate the cells. The supernatant was poured off and each pellet was frozen in liquid nitrogen for 15 seconds. The collected cells can be stored in a freezer at -70^°C.

The PDEs from harvested SW480 cells were isolated using an FPLC procedure. A Pharmacia AKTA FPLC was used to monitor sample loading and elution on an 18 ml DEAE-Trisacryl M column. Approximately 600 million SW480 cells were used for the profiles. After resuspending the cells in homogenization buffer (20 ml of TMPI-EDTA7

IS 2 174 573 T3

Triton pH 7.4: 20 mM Tris-HOAc, 5 mM MgAc2, 0.1 mM EDTA, 0.8% Triton-100, 10 μΜ Benzamidine, 10 μΜ TLCK, 2000 U / ml Aprotinin, 2 μΜ Leupeptin, Pepstatin At 2 μΜ), the samples were homogenized manually. FPLC buffer A was 8 mM TRIS-acetate, 5 mM Mg acetate, 0.1 mM EDTA, pH 7.5 and buffer B was 8 mM Tris-acetate, 5 mM Mg acetate, 0.1 mM EDTA , 1M Na acetate, pH 7.5. The supernatants were loaded onto the column at 1 ml per minute, followed by washing with 60 ml of buffer A at 1 ml per minute. A gradient of 0-15% buffer B in 60 ml, 15-50% buffer B in 60 ml and 50-100% buffer B in 16 ml was passed. During the gradient 1.5 ml fractions were collected.

The profile obtained was similar (Figure 26) to the profile of the new PDE activity (see, for example, Figure 1) obtained previously, except that Peak B isolated in this way presented hydrolytic activity of cAMP at 0.25 μM of substrate that could be activated 2-3 times by 5 µM cGMP.

A second method was performed to isolate classic PDE2 from SW480 using a non-FPLC DEAE column procedure as described above (see Section IA) with the modification that the buffers contained 30% ethyl glycol, 10 mM TLCK and 3.6 mM β-mercaptoethanol. The addition of these reagents to the buffers produces a displacement of the elution profile (see Figure 25) from the low to the high concentration of sodium acetate, in such a way that Peak A moves from 40 to 150 mM, Peak B from 75 to 280 mM and Peak C from 200 to 500 mM Na acetate (see Figure 25). Peak B of Figure 25 was assayed with 2 µM cAMP substrate and showed a two-fold activation by 5 µM cGMP (see Figure -Y). The selective PDE2 inhibitor NASH inhibited 2 μM cGMP PDE activity at this Peak B with an IC₅₀ 1.6 μM, and inhibited 2.0 μM cAMP PDE activity at Peak B with an IC₅₀ 3.8 μM (and with an IC₅₀ 2.5 µM with the addition of 10 µM rolipram).

C. Specificity of cGMP of PDE Peak A and the new activity of Peak B

Each fraction of the DEAE column of Section IA was also analyzed to determine the hydrolytic activity of cGMP (0.25 μM cGMP) in the presence or absence of Ca ++, or Ca ++ - CaM and / or EGTA, and to determine the activity hydrolytic cAMP (0.25 µM cAMP) in the presence or absence of 5 µM cGMP. Neither Peak A nor Peak B of PDE (fractions 5-22; see Fig. 1) significantly hydrolyzed cAMP, establishing that none had the activity of a classical PDE family that hydrolyzed cAMP (ie, a PDE 1, 2, 3).

AC⁺⁺ (with or without calmodulin) did not activate the hydrolytic activity of cAMP or cGMP of Peak A or B, and cGMP did not activate or inhibit cAMP hydrolysis. Such results establish that Peaks A and B constitute cGMP-specific PDE activities, but not classical or previously known PDE1, PDE2, PDE3 or PDE4 activities.

For the new PDE Peak B, as discussed below, cyclic GMP activated the cGMP hydrolytic activity of the enzyme, but did not activate any cAMP hydrolytic activity (in contrast to Peak B of Section IB above). This reveals that the new PDE Peak B - the novel phosphodiesterase of this invention - is not a cGMP-stimulated cAMP ("cGS") hydroylysis nor is it among the activities of the classical or previously known PDE2 family, because the known isoforms PDE2 hydrolyzes cGMP and cAMP.

D. Pico A is a classic PDE5, but the new Pico B -a new GMP-specific PDE- is not.

To characterize any PDE isoform, kinetic behavior and substrate preference must be determined.

Peak A exhibited typical "PDE5" characteristics. For example, the enzyme Km for cGMP was 1.07 μM, and the V_max it was 0.16 nmol / minute / mg. Also, as discussed below, zaprinast (CI₅₀= 1.37 μM) and E4021 (CI₅₀= 3 nM) and sildenafil inhibited the activity of Peak A. In addition, zaprinast showed inhibition of the hydrolytic activity of cGMP of Peak A, consistent with the results described in the literature.

PDE Peak B of Section IA exhibited considerably different kinetic properties when compared to PDE Peak A. For example, in Eadie-Hofstee plots of Peak A, cyclic GMP hydrolysis exhibits a single negative slope with increasing substrate concentrations, indicative of Michaelis-Menten kinetic behavior. Peak B, however, shows the new hydrolytic property of cGMP in the absence of cAMP with a slope

ES 2 174 573 T3 that decreases (apparent Km = 8.4) and subsequently increases (Km <1) as in the Eadie-Hofstee graphs with increasing cGMP substrate (see, Fig. 3). This establishes a submicromolar affinity of Peak B for cGMP (that is, where Km <1).

Consistent with the kinetic studies (ie, Fig. 3) and with the cooperative positive kinetic behavior in the presence of the cGMP substrate, was the hydrolytic activity of cGMP increased in the presence of increasing concentrations of the cGMP substrate. This was discovered by comparing cGMP concentrations of 0.25 μΜ, 2 μΜ and 5 μΜ in the presence of PDE Peak B after a second separation in DEAE to exclude hydroalysis of cAMP and to rule out that this new enzyme was a previously identified PDE5. . Higher cGMP concentrations produced disproportionately greater cGMP hydrolysis with PDE Peak B, as shown in Fig. 2.

These observations suggest that the binding of cGMP to the Peak B enzyme produces a conformational change in the enzyme. This confirms the advantage of using the native enzyme from neoplastic cells, but this invention is not limited to the native form of the enzyme having the characteristics set forth above.

E. Insensitivity to zaprinast and sildenafil of PDE Peak B relative to Peak A, and their effects on other PDE inhibitors

Different PDE inhibitors were studied using twelve drug concentrations from 0.01 to 100 μΜ and a substrate concentration ³0.25 μΜ H-cGMP. IC values ​​were calculated₅₀ with variable slope sigmoid curve fits, using Prism 2.01 (GraphPad). The results are shown in Table 1. While compounds E4021 and zaprinast inhibited Peak A (with high affinities), the IC50 values ​​calculated against the new PDE activity at Peak B (Section IA) are significantly increased ( > 50 times). This confirms that Peak A is a PDE5. These data further illustrate that the novel PDE activity of this invention is, for all practical purposes, insensitive to zaprinast and insensitive to E4021.

TABLE 1

Comparison of PDE inhibitors against Peak A and Peak B of section IA (cGMP hydrolysis)

<td>Compound</td><td>PDE Family Inhibitor</td><td>IC50 Peak A (μΜ)</td><td>IC50 Peak B (μΜ)</td><td>Ratio (CI₅₀Peak A / Peak B)</td>

<td>E4021</td><td> 5</td><td> 0,003</td><td> 8,4</td><td> 0,0004</td>

<td>Zaprinast</td><td> 5</td><td> 1,4</td><td> &gt;30</td><td> &lt;0,05</td>

<td>Compound E</td><td>5 and others</td><td> 0,38</td><td> 0,37</td><td> 1,0</td>

<td>Sulindac sulfide</td><td>5 and others</td><td> 50</td><td> 50</td><td> 10</td>

<td>Vinpocetine</td><td> 1</td><td> &gt;100</td><td> &gt;100</td><td></td>

<td>NASH</td><td> 2,5</td><td> &gt;100</td><td> 3,7</td><td></td>

<td>Indolidano</td><td> 3</td><td> 31</td><td> &gt;100</td><td> &lt;0,31</td>

<td>Rolipram</td><td> 4</td><td> &gt;100</td><td> &gt;100</td><td></td>

<td>Sildenafil</td><td> 5</td><td> 0,0003</td><td> &gt;10</td><td> &lt;0,00003</td>

In contrast, sulindac sulfide and Compound E competitively inhibited phosphodiesterases at Peak A and B with the same potency (CI₅₀= 0.38 μΜ for PDE Peak A; 0.37 μΜ for Peak BdePDE).

Significant for the treatment of neoplasms and for the selection of useful compounds for such treatment is the fact that Peak B (any form thereof) is insensitive to zaprinast, while Peak A and B are both sensitive to sulindac sulfide and to Compound E. We have tested zaprinast, E4021 and sildenafil to determine if they induce apoptosis or inhibit the growth of neoplastic cells, and we have done the same for Compound E. As explained later, zaprinast does not by itself have significant apoptosis-inducing or growth-inhibiting properties, whereas sulindac sulfide and Compound E are precisely the opposite. In other words, the ability of a compound to inhibit both PDE Peaks A and B is correlated with its ability to induce apoptosis in neoplastic cells, whereas if a compound (for example, zaprinast) has specificity only for PDE Peak A , that compound will not induce apoptosis by itself.

IS 2 174 573 T3

F. Insensitivity of the new PDE Peak B to incubation with cGMP-dependent protein kinase G

Additional differences between PDE Peak A and new Peak B (Section IA) were observed in their respective cGMP hydrolytic activities in the presence of varying concentrations of cGMP-dependent protein kinase G (which phosphorylates topical PDE5). Specifically, the Peak A and Peak B fractions from Section IA were incubated with different concentrations of protein kinase G at 30^°C for 30 minutes. The calcium GMP hydroalysis of both peaks was assayed after attempting phosphorylation. Consistent with previously published information on PDE5, Peak A showed increased cGMP hydrolytic activity in response to incubation with protein kinase G, indicating that Peak A was phosphorylated. However, Peak B was unaltered (that is, it was not phosphorylated and was insensitive to incubation with cGMP-dependent protein kinase G). These data are consistent with the fact that Peak A is an isoform consistent with the known PDE5 family and Peak B of Section IA is a new cGMP-specific PDE activity.

G. The new Peak B in prostate and breast cancer cell lines

The new Peak B was also isolated from two other neoplastic cell lines, a breast cancer cell line, HTB-26, and a prostate cancer cell line, LNCaP, by a procedure similar to that used previously to isolate it from SW480. . The protocol was modified in several respects. In order to provide even greater reproducibility to allow comparison of different cell lines, a Pharmacia AKTA FPLC will be used to monitor sample loading and elution on an 18 ml DEAE-Trisacryl M column. SW480 will be run by this same procedure multiple times to provide a Peak B reference. 200-400 million SW480 cells were used for the profiles. 70 million LNCaP cells were used for a profile (see Figures 22 and 23), and in a separate experiment 32 million HTB-26 cells were used for a profile (see Figures 20 and 21). After resuspending the cells in homogenization buffer, the samples were homogenized manually. FPLC Buffer A was 8 mM TRIS-acetate, 5 mM Mg acetate, 0.1 mM EDTA, pH 7.5 and Buffer B was 8 mM Tris-acetate, 5 mM Mg acetate, 0.1 mM EDTA , 1M Na acetate, pH 7.5. The supernatants were loaded onto the column at 1 ml per minute, followed by washing with 60 ml of buffer A at 1 ml per minute. A gradient of 0-15% buffer B in 60 ml, 15-50% buffer B in 60 ml and 50-100% buffer B in 16 ml was passed. During the gradient 1.5 ml fractions were collected. The cGMP PDE activity peaks eluted around fraction 65 which was Na acetate at 400 mM (see Figures 20-23). This activity was measured at 0.25 µΜ cGMP (indicating submicromolar affinity for cGMP). Rolipram, a PDE4-specific drug, inhibited most of the PDE activity of cAMP (ie, the activity on cAMP was due to PDE4), indicating that the activity on cAMP of Peak B was specific for cAMP on cAMP. The three B peaks (from SW480, HTB-26 and LNCaP) did not show any stimulation with calcium / calmodulin and were resistant to 100 nM E4021, a specific inhibitor of PDE5 like zaprinast (see Figures 20 and 22). Peak B also showed a dramatic increase in activity when the cGMP substrate was increased from 0.25 µΜ to 5 µΜ (suggesting positively cooperative kinetics) (see Figures 21 and 23). Furthermore, all three peaks show similar inhibition by exisulind and by Compound I, subsequently.

II. Involvement of β-catenin- Gy protein kinase in general

A series of experiments were performed to determine what effect, if any, an antineoplastic cGMP-specific PDE inhibitor such as exisulind had on cGMP-dependent protein kinase G ("PKG") in neoplastic cells containing the defect. of the adenomatous polyposis coli gene ("APC gene") or a defect in the gene encoding β-catenin. As explained below, such an inhibitor produces an elevation of PKG activity in such neoplastic cells. That increased activity was not due only to increased activation of PKG in cells containing either defect, but also to increased expression of PKG in cells containing the APC defect. Furthermore, when PKG is immunoprecipitated from neoplastic cells with any of the defects, it precipitates with β-catenin.

Beta-catenin has been implicated in a variety of different cancers because researchers have found elevated levels of it in patients with neoplasms that contain mutations in the APC tumor suppressor gene. People with birth mutations in this gene often develop thousands of small tumors in the inner lining of their colon. When functioning properly, the APC gene encodes a normal APC protein that is believed to bind to, and regulate, β-catenin. Therefore, the discovery that PKG in neoplastic cells containing the APC gene defect

ES 2 174 573 T3 or the defect of β-catenin binds to β-catenin, in effect strongly implicates PKG in one of the main cellular pathways leading to cancer. Furthermore, the relationship between specific inhibition of cGMP and elevation of PKG after treatment with SAANDs binds to cGMP with PKG / β-catenin / APC defect in such cells.

This latter relationship is further reinforced by the observation that β-catenin itself is reduced when neoplastic cells containing the APC defect or the β-catenin defect are exposed to a SAAND. This reduction in β-catenin is initiated by PKG itself. PKG phosphorylates βcatenin - which is another new observation associated with this invention. Phosphorylation of β-catenin allows β-catenin to be degraded by the proteasome-ubiquitin system.

This phosphorylation of β-catenin by PKG is important in neoplastic cells because it avoids the effect of APC and β-catenin mutations. The mutated APC protein affects the binding of β-catenin bound to the mutant APC proteon, a change in binding that has heretofore been believed to prevent phosphorylation of β-catenin by GSK-3b kinase. In the case of mutant β-catenin, an elevation of PKG activity also allows mutant β-catenin to be phosphorylated. The elevation of PKG activity in neoplasms together with the inhibition of PDE-cGMP allows the phosphorylation of β-catenin (leading to its degradation) in neoplastic cells that contain any of the types of mutations.

In summary, these findings not only lead to new pharmaceutical screening methods to identify additional SAAND candidate compounds, but also reinforce the role of cGMP-specific PDE inhibition in therapeutic approaches to neoplasms. This observation may also explain the unexpectedly wide range of neoplasms that SAANDs can inhibit, since neoplasms can be treated with and without the APC defect, as explained above.

III. Selection of pharmaceutical compositions using PDEs

A. General

The new PDE of this invention and PDE2 are useful with or without PDE5 to identify compounds that can be used to treat or prevent neoplasms, and that are not characterized by serious side effects.

Cancer and pre-cancer can be considered as diseases that involve unregulated cell growth. Cell growth involves several different factors. One factor is how quickly the cells proliferate, and another involves how quickly the cells die. Coells can die from necrosis or apoptosis, depending on the type of environmental stimuli. Cell differentiation is another factor that influences tumor growth kinetics. The resolution of which of the many aspects of cell growth is affected by a compound is important for the discovery of a relevant target for drug therapy. Screening assays based on this technology can be combined with other analyzes to select compounds that have growth inhibitory and pro-apoptotic activity.

This invention is the product of several important discoveries. First, the present inventors discovered that desirable tumor cell growth inhibitors induce premature death of cancer cells by apoptosis (see, Piazza, GA et al., Cancer Research, 55 (14), 3110-16, 1995). Second, several of the present inventors unexpectedly discovered that compounds that selectively induce apoptosis without substantial inhibition of COX also inhibit PDE5. In particular, and contrary to prominent scientific studies, desirable compounds for treating neoplastic lesions inhibit PDE5 (EC 3.1.4.17). PDE5 is one of at least ten phosphodiesterase gene families. PDE5 and the novel PDE of this invention are unique in that they selectively degrade colic and non-cAMP GMP, whereas the other PDE families selectively degrade / hydrolyze cAMP and non-cAMP, or non-selectively degrade cGMP and cAMP. Preferably, the desirable compounds used to treat neoplasms do not substantially inhibit the types of phosphodiesterase that non-selectively degrade cGMP or cAMP.

B. Selection of COX

A preferred embodiment of the present invention involves the determination of the cyclooxygenase inhibitory activity of a given compound and the determination of the cGMP-specific PDE inhibitory activity of the compound. The ability of test compounds to treat lesions is determined

ES 2 174 573 T3 neoplastic lesions directly or indirectly, comparing their activities against known compounds useful for treating neoplastic lesions. A standard compound known to be effective in treating neoplastic lesions without causing gaystric irritation is 5-fluoro-2-methyl-1- (p-methylsulfonylbenzylidene) 3-indenylacetic acid ("exisulind"). Other useful compounds for comparative purposes here include those known to inhibit COX, such as indomethacin and the sulindac sulfide metabolite: 5-fluoro-2-methyl-1- (p-methylsulfinylbenzylidene) -3-indenylacetic acid ("sulindac sulfide"). . Other useful compounds for comparison purposes here include those known to inhibit cGMP-specific PDEs, such as 1- (3-chloroanilino) -4-phenyphthalazine ("MY5445").

Seguin as used herein, the term "precancerous lesion" includes syndromes represented by abnormal neoplastic tissue changes, including dysplasia changes. Examples include dysplastic growths in colon, breast, priostate, or lung tissues, or conditions such as dysplastic nevus syndrome, a precursor to malignant melanoma of the skin. Examples also include, in addition to dysplastic nevus syndromes, polyposis syndromes, colloin polyps, precancerous lesions of the cervix (i.e., cervical dysplasia), esophagus, lung, prostatic dysplasia, prostatic intraneoplasia, breast and / or skin and related conditions ( for example, actinic keratosis), whether or not the lesions are clinically identifiable.

Following are used herein, the terms "carcinoma" or "cancer" refer to lesions that are cancerous. Examples include malignant melanomas, breast cancer, proist cancer, and colon cancer. Seguín are used herein, the terms "neoplasia" and "neoplasms" refer to cancerous and precancerous lesions.

Seguín is used herein, the abbreviation PG stands for prostaglandin; PS stands for prostaglandin synthetase; PGE2 represents prostaglandin E2; PDE represents phosphodiesterase; COX represents cyclooxygenase; cyclic nucleotide; RIA stands for radioimmunoassay.

COX inhibition by a compound can be determined by either of two methods. One method involves the measurement of PGE2 secretion by intact HL-60 cells after exposure to the compound being screened. The other method involves the measurement of the activity of purified cyclooxygenases (COXs) in the presence of the compound. Both methods involve protocols previously described in the literature, but the preferred protocols are described below.

Compounds can be tested to determine if they inhibit prostaglandin E2 ("PGE2") production by measuring PGE2. Using an enzyme immunoassay (EIA) kit for PGE2, such as that commercially available from Amersham, Arlington Heights, IL, USA Suitable cells here include those that produce an abundance of PG, such as HL-60 cells. HL-60 cells are human promyelocytes that differentiate with DMSO to mature granulocytes (see, Collins, SJ, Ruscetti, FW, Gallagher, RE and Gallo, RC, “Normal Functional Characteristics of Cultured Human Promyelocytic Leukemia Cells (HL-60) After Induction of Differentiation by Dimethylsulfoxide ”, J. Exp. Med., 149: 969-974, 1979). These differentiated cells produce PGE₂ after stimulation with a calcium ionophore, A23187 (see, Kargman, S., Prasit, P. and Evans, JF, "Translocation of HL-60 Cell 5-Lipoxygenase", J. Biol. Chem., 266: 23745 -23752, 1991). The hL-60 are available from the ATCC (ATCC: CCL240). They can be cultured in RPMI 1640 medium supplemented with 20% heat inactivated fetal bovine serum, 50 U / ml penicillin and 50 μg / ml streptomycin in an atmosphere with 5% CO2 a37^°C. To induce myeloid differentiation, cells are exposed to 1.3% DMSO for 9 days and then washed and resuspended in Dulbecco's phosphate buffered saline at a concentration of 3 x 10.⁶ cells / ml.

The differentiated HL-60 cells (3 x 10⁶ cells / ml) are incubated for 15 minutes at 37^°In the presence of the compounds tested at the desired concentration. Cells are subsequently stimulated with A23187 (5 x 10^-6 M) for 15 minutes. The PGE2 secreted to the external environment is measured as described previously.

As noted above, a second method for determining COX inhibition by a compound is to measure COX activity in the presence of a test compound. Two different forms of cyclooxygenase (COX-I and COX-2) that regulate prostaglandin synthesis have been described in the literature. COX-2 represents the inducible form of COX, while COX-I represents a constitutive form. COX-I activity can be measured using the method described by Mitchell et al. ("Selectivity of Nonsteroidal Anti-inflammatory Drugs as Inhibitors of Constitutive and Inducible Cyclooxigenase", Proc. Natl. Acad. Sci. USA, 90: 11693-11697, 1993, incorporated herein by reference) using purified COX-I of ram seminal vesicles following this described by Boopa12

ES 2 174 573 T3 thy and Balasubramanian, "Purification and Characterization of Sheep Platelet Cyclooxygenase" (Biochem. J., 239: 371-377, 1988). COX-2 activity can be measured using COX-2 purified from sheep placenta as described by Mitchell et al., 1993, supra.

The cyclooxygenase inhibitory activity of a drug can be determined by methods known in the art. For example, Boopathy and Balasubramanian, 1988, supra, described a procedure in which prostaglandin H synthase 1 (Cayman Chemical, Ann Arbor, Michigan) is incubated at 37 ° C for 20 minutes with 100 μΜ arachidonic acid (Sigma Chemical Co.) , cofactors (such as 1.0 mM glutathione, 1.0 mM hydroquinone, 0.625 μΜ hemoglobin, and CaCl₂ 1.25 mM in 100 mM Tris-HCl, pH 7.4) and the drug to be tested. After incubation, the reaction can be terminated with trifluoroacetic acid. After stopping the reaction by adding thiobarbituoric acid and malonaldehyde, the enzymatic activity can then be measured spectrophotometrically at 530 nm.

Obviously, a compound exhibiting more low COX-I or COX-2 inhibitory activity relative to its combined higher PDE5 / new PDE / PDE2 inhibitory activities may be a desirable compound.

The level of COX inhibition is determined by comparing the cyclooxygenase activity in the presence and absence of the test compound. A residual COX inhibitory activity (ie, less than about 25%) or no COX inhibitory activity at a concentration of about 100 µM indicates that the compound should be further evaluated for its usefulness in treating neoplasms.

C. Determination of phosphodiesterase inhibitory activity

Compounds can be screened for their inhibitory effect on the activity of the new phosphodiesterase of this invention using the enzyme isolated as described above, a recombinant version, or using the new PDE and / or PDE2 together with PDE5. Alternatively, the levels of coclic nucleotides in whole cells are measured by RIA and compared with those in untreated cells and with those in cells treated with zaprinast.

Phosphodiesterase activity can be determined using methods known in the art, such as a method that uses ³Radioactive colic H-GMP (cGMP) (3 ', 5'-guanosine monophosphate) as substrate for the PDE enzyme (Thompson, WJ, Teraski, WL, Epstein, PM, Strada, SJ, Advances in Cyclic Nucleotide Research, 10: 69-92, 1979, which is incorporated herein by reference). Briefly, a substrate solution³H-cGMP of defined specific activity (0.2 μM; 100,000 cpm; containing 40 mM Tris-HCl (pH 8.0), 5 mM MgCl2 and 1 mg / ml of BSA) is mixed with the drug to be tested in a volume total of 400 μ !. The mixture is incubated at 30 ° C for 10 minutes with the isolated PDE of this invention. The reactions are terminated, for example, by boiling the reaction mixture for 75 seconds. After cooling on ice, 100 μl of 0.5 mg / ml snake venom (O. hannah available from Sigma) and incubated for 10 minutes at 30 ° C. This reaction is subsequently terminated by the addition of an alcohol, for example 1 ml of 100% methanol. The test samples are applied to a 1 ml Dowex 1-X8 column and washed with 1 ml of 100% methanol. The amount of radioactivity from the front running through the column and from the column wash are combined and measured with a scintillation counter. The degree of phosphodiesterase inhibition is determined by calculating the amount of radioactivity in the drug-treated reactions and comparing it to a control sample (a reaction mixture that lacks the test compound but has the drug's solvent).

Alternatively, the ability of the desirable compounds to inhibit the phosphodiesterases of this invention was reflected by an increase in cGMP in neoplastic cells exposed to a compound that was being screened. The level of PDE activity can be determined by an assay to evaluate the amount of colic GPM in the extract from the treated cells using radioimmunoassay (RIA). In this procedure, HT-29 or SW480 cells are plated and cultured to confluence. As indicated above, SW480 contains PDE5 and the new PDE of this invention, so when PDE activity was evaluated in this way, a combined cGMP hydrolytic activity is simultaneously tested. The test compound is then incubated with the cell culture at a compound concentration between about 200 µM and about 200 µM. About 24 to 48 later, the culture medium is removed from the cells and the cells are solubilized. The reaction was stopped using 0.2N HCl / 50% MeOH. A sample is removed for the proteon assay. The colic GMP is purified from the acid / alcoholic extracts of the cells using anion exchange chromatography, such as a Dowex column. The cGMP is dried and acetylated according to procedures

ES 2 174 573 T3 published, such as by the use of acetic anhydride in triethylamine (Steiner, AL, Parker, CW, Kipnis, DM, J. Biol. Chem., 247 (4): 1106-13, 1971). Acetylated cGMP is quantified using radioimmunoassay procedures (Harper, J., Brooker, G., Advances in Nucleotide Research, 10: 1-33, 1979). Iodinated ligands (methyl tyrosine ester) of derivatized cyclic GMP are incubated with standards or with unknown samples in the presence of appropriate antisera and buffers. The antiserum can be produced using targeted cyclic nucleotide-hapten techniques. The antiserum comes from sheep injected with succinyl-cGMP-albumin conjugates and is diluted 1 / 20,000. A dose interpolation and an error analysis are applied from the standard curves as previously described (Seibert, AF, Thompson, WJ, Taylor, A., Wilboum, WH, Barnard, J. and Haynes, J., J Applied Physiol., 72: 389-395, 1992).

In addition, the culture medium can be acidified, frozen (-70^°C) and also analyzed for cGMP and cAMP.

In addition to observing increases in cGMP content in neoplastic cells caused by desirable compounds, decreases in cAMP content have also been observed. A particularly desirable compound (i.e., one that selectively induces apoptosis in neoplastic cells, but not substantially in normal cells) has been observed to follow a variation over time consistent with cGMP-specific PDE inhibition as an initial action that results in increased cGMP content in minutes. Secondarily, treatment of neoplastic cells with a desirable antineoplastic compound leads to decreased cAMP content within 24 hours. Intracellular targets of drug actions are being studied further, but current data support the concept that the initial rise in cGMP content and the subsequent drop in cAMP content precede apoptosis in neoplastic cells exposed to desirable compounds. .

Changing the ratio between the two cyclic nucleoitides may be a more accurate tool to evaluate the desirable cGMP-specific phosphodiesterase inhibitory activity of the test compounds, rather than the measurement of only the absolute value of cGMP, of only the inhibition of the specific phosphodiesterase of cGMP or of only the level of hydrolysis of cGMP. In neoplastic cells not treated with antineoplastic compounds, the ratio of cGMP content / cAMP content was in the range of 0.03-0.05 (that is, a cGMP content of 300-500 fmoles / mg of protein on a of cAMP of 6000-8000 fmoles / mg protein). After exposure to desirable antineoplastic compounds, the ratio increases several times (preferably an increase of at least about three times) as a result of the initial increase in cyclic GMP and the subsequent decrease in cyclic AMP.

Specifically, particularly desirable compounds have been observed to achieve an initial increase in cGMP content in treated neoplastic cells to a cGMP level greater than about 500 fmoles / mg protein. Furthermore, particularly desirable compounds cause the subsequent decrease in cAMP content in treated neoplastic cells to a cAMP level of less than about 4000 fmoles / mg protein.

To determine cyclic AMP content, radioimmunoassay techniques similar to those described above for cGMP are used. Basically, cyclic nucleotides are purified from acid / alcoholic cell extracts using anion exchange chromatography, dried, acetylated according to published procedures and quantified using radioimmunoassay procedures. Derivatized cyclic AMP and cyclic GMP iodinated ligands are incubated with standards or unknown samples in the presence of specific antisera and appropriate buffers.

Verification of cyclic nucleotide content can be obtained by determining the turnover or accumulation of cyclic nucleotides in intact cells. To measure the cAMP of intact cells, a pre-label with³H-adenine according to published procedures (Whalin, ME, Garrett Jr., RL, Thompson, WJ and Strada, SJ “Correlation of cell-free brain cyclic nucleotide phosphodiesterase activities to cyclic AMP decay in intact brain slices”, Sec. Mess . and Phos. Protein Research, 12: 311325, 1989). The procedure measures the flux of labeled ATP to cyclic AMP and can be used to calculate adenylate cyclase activity or cyclic nucleoitide phosphodiesterase activity in intact cells depending on the specific protocol. Cyclic GMP accumulation was too low to be studied with intact cell prelabeling according to published procedures (Reynolds, PE, SJ Strada and WJ Thompson, "Cyclic GMP Accumulation In Pulmonary Microvascular Endothelial Cells Measured By Intact Cell Prelabeling", Life Sci., 60: 909-918, 1997).

IS 2 174 573 T3

The effect of the PDE inhibitory activity of a compound can also be determined from a tissue sample. Human tissue biopsies or anesthetized animal tissues are collected from subjects exposed to the test compound. Briefly, a tissue sample is homogenized at 500 μ! 6% TCA. A known quantity of the homogenate is removed for protein analysis. The remaining homogenate is allowed to stand on ice for 20 minutes to allow protein to precipitate. The homogenate is then centrifuged for 30 minutes at 15,000 g at 4 ° C. The supernatant is recovered and the pellet is recovered. The supernatant is washed four times with five volumes of water-saturated diethyl ether. The top layer of Ether is discarded between each wash. The aqueous ether extract is dried in a "speed vac". Once dried, the sample can be frozen for future use, or used immediately. The dried extract is dissolved in 500 μ! of Assay Buffer. The level of specific inhibition of cGMP is determined by analyzing the amount of coclic nucleootides using RIA procedures as described above.

The degree of inhibition is determined by comparing the activity of the new PDE (or PDE2) in the presence and absence of the compound. Inhibition of the new PDE (or PDE2) activity is indicative that the compound is useful for treating neoplasms. A significant inhibitory activity greater than that of the exisulind reference compound, preferably greater than 50% at a concentration of 10 µM or less, is indicative that a compound should be further evaluated for its antineoplastic properties. Preferably, the IC50 value for inhibition of the new PDE should be less than 50 µM for the compound to be considered later for potential use.

D. Determination of whether a compound reduces tumor cell growth

In an alternative embodiment, the method of the present invention involves further determination of whether the compound reduces the growth of tumor cells. Various cell lines can be used in the sample depending on the tissue to be tested. For example, these cell lines include: SW480 - colon adenocarcinoma; HT-29 -denocarcinoma of the colon; A-427 - lung adenocarcinoma carcinoma; MCF-7 - breast adenocarcinoma; UACC-375 - melanoma lónea and DU145 - prostate carcinoma. The cytotoxicity data obtained using these cell lines are indicative of an inhibitory effect on neoplastic lesions. These cell lines are well characterized and are used by the US National Cancer Institute in its program for the selection of new anticancer drugs.

The ability of a compound to inhibit the growth of tumor cells can be measured using the human colon carcinoma cell line HT-29 obtained from ATCC. HT-29 cells have been previously characterized as a relevant model of colon tumor cell culture (Fogh, J. and Trempe, G. In: Human Tumor Cells in Vitro, J. Fogh (eds.), Plenum Press, New York , pp. 115-159, 1975). HT-29 cells are maintained in RPMI medium supplemented with 5% fetal calf serum (Gemini Bioproducts, Inc., Carlsbad, CA) and 2 mM glutamine and 1% antimycotic antibiotic in a humidified atmosphere with 95% air and 5% CO₂ at 37 ° C. Briefly, HT-29 cells are plated at a density of 500 cells / well in 96-well microtiter plates and incubated for 24 hours at 37 ° C prior to compound addition. Each determination of the number of cells involved six replicates. After six days in culture, the cells are fixed by the addition of cold trichloroacoetic acid to a final concentration of 10% and the levels of proteon are measured using the colorimetric test for staining of proteones with sulforhodamine B (SRB) as described previously. by Skehan, P., Storeng, R., Scudiero, D., Monks, A., McMahon, J., Vistica, D., Warren, JT, Bokesch, H., Kenney, S. and Boyd, MR, " New Colorimetric Assay For Anticancer-Drug Screening ”, J. Natl. Cancer Inst., 82: 1107-1112, 1990.

In addition to the SRB assay, various other methods are available to measure growth inhibition and could replace the SRB assay. These methods include counting viable cells after Trypan Blue staining, labeling cells capable of synthesizing DNA with BrdU or radiolabeled thymidine, staining of viable cells with neutral red, or staining of viable cells with MTT.

A significant inhibition of tumor cell growth greater than about 50% at a dose of 100 µM or less is further indicative that the compound is useful in treating neoplastic lesions. Preferably, the IC50 value is determined and used for comparative purposes. This value is the drug concentration necessary to inhibit the growth of tumor cells by 50% in relation to the control. Preferably, the IC value₅₀ it must be less than 100 μM for the compound to be considered later for potential use in the treatment of neoplastic lesions.

IS 2 174 573 T3

E. Determination of whether a compound induces apoptosis

In a second alternative embodiment, the selection method of the present invention further involves the determination of whether the compound induces apoptosis in tumor cell cultures.

Two different forms of cell death can be described by morphological and biochemical criteria: necrosis and apoptosis. The necrosis is accompanied by an increased permeability of the plasmaitic membrane; cells swell and the plasmid membrane ruptures within minutes. Apoptosis is characterized by membrane bubble formation, condensation of the cytoplasm, and activation of endogenous endonucleases.

Apoptosis occurs naturally during normal tissue turnover and during embryonic development of the organs and extremities. Apoptosis is also induced by cytotoxic T lymphocytes and by natural killer cells, by ionizing radiation and by certain chemotherapeutic drugs. It is believed that an incorrect regulation of apoptosis plays an important role in many pathological conditions, including cancer, AIDS or Alzheimer's disease, etc. Compounds can be screened for the induction of apoptosis using tumor cell cultures maintained under conditions as described above. Treatment of cells with test compounds involves pre or post-confluent cultures and treatment for two to seven days at various concentrations. Apoptoitic cells are measured in the attached compartment and in the "floating" compartment of the cultures. Both compartments are collected by extracting the supernatant, trypsinizing the adhered cells and combining both preparations after a centrifugal washing step (10 minutes, 2000 rpm). The protocol for treating tumor cell cultures with sulindac and related compounds to obtain a significant level of apoptosis has been described in the literature. (See, Piazza, GA et al., Cancer Research, 55: 3110-16, 1995). New features include the collection of adherent and floating cells, the identification of optimal treatment times and dose ranges to observe apoptosis, and the identification of optimal cell culture conditions.

After treatment with a compound, cultures can be analyzed for apoptosis and necrosis by fluorescence microscopy after labeling with acridine orange and ethidium bromide. The method for measuring apoptotic cell numbers has been previously described by Duke and Cohen, "Morphological And Biochemical Assays Of Apoptosis", Current Protocols In Immunology, Coligan et al., Eds., 3.17.1-3.17.16 (1992) .

For example, adherent, floating cells can be harvested by trypsinization and washed three times with PBS. Aliquots of the cells can be centrifuged. The pellet can then be resuspended in medium and a dye mixture containing acridine orange and ethidium bromide can be prepared in PBS and mixed gently. The mixture can then be placed on a microscope slide and examined to visualize morphological features of apoptosis.

Apoptosis can also be quantified by measuring an increase in DNA fragmentation in cells that have been treated with the test compounds. Commercial photometric EIAs are available for the quantitative in vitro determination of histone-associated cytoplasmic DNA fragments (mono- and oligonucleosomes) (Cell Death Detection ELISA^okys, Catalog No. 1,774,425, Boehringer Mannheim). The Boehringer Mannheim assay is based on the sandwich enzyme immunoassay principle using mouse monoclonal antibodies directed against DNA and histones, respectively. This allows the specific detection of mono and oligonucleosomes in the cytoplasmic fraction of cell lysates.

According to the provider, apoptosis is measured as follows. The sample (cell lysate) is placed on a streptavidin-coated micro-titer plate ("MTP"). Subsequently, a mixture of anti-histone-biotin and conjugate of anti-DNA and peroxidase is added and incubated for two hours. During the incubation period, the anti-histone antibody binds to the histone component of the nucleosomes and simultaneously binds the immune complex to streptavidin-coated MTP through its biotinylation. Additionally, the anti-DNA peroxidase antibody reacts with the DNA component of nucleosomes. After the elimination of unbound antibodies by a washing step, the number of nucleosomes is quantified by the peroxidase retained in the immune complex. Peroxidase is determined photometrically with ABTS7 (2,2'-azido- [3-ethylbenzothiazoline sulfonate]) as substrate.

For example, SW480 colon adenocarcinoma cells are plated in a 96-well MTP at a density of 10,000 cells per well. Cells are post-treated with the test compound and allowed to incubate for 48 hours at 37^°C. After incubation, the MTP is centrifuged

ES 2 174 573 T3 and the supernatant is removed. The cell pellet from each well is then resuspended in lysis buffer for 30 minutes. The lysates are then centrifuged and aliquots of the supernatant (that is, the cytoplasmic fraction) are transferred to a streptavidin-coated MTP. Care must be taken not to agitate the lysed pellets (that is, cell nuclei containing high molecular weight, non-fragmented DNA) in the MTP. The samples are subsequently analyzed.

The number of times of stimulation is determined for each compound tested at a given concentration (FS = OD_max/DO_veh), an indicator of apoptotic response. EC50 values ​​can also be determined by evaluating a series of concentrations of the test compound.

Statistically significant increases in apoptosis (that is, a stimulation greater than 2 times at a concentration of 100 μΜ) are also indicative that the compound is useful for treating neoplastic lesions. Preferably, the EC50 value for apoptosis activity should be less than 100 µΜ for the compound to be further considered for potential use in the treatment of neoplastic lesions. The EC50 is defined herein as the concentration that produces 50% induction of apoptosis relative to vehicle treatment.

F. Assays in the mammary gland organ culture model

Test compounds identified by the above methods can be tested for antineoplastic activity for their ability to inhibit the incidence of preneoplastic lesions in a mammary gland organ culture system. This technique for growing organism from rat mammary glands has been used with success by other researchers to study the effects of known antineoplastic agents such as certain NSAIDs, retinoids, tamoxifen, selenium and certain natural products, and is useful for the validation of the method. selection of the present invention.

For example, female BALB / c mice can be treated daily with a combination of estradiol and progesterone in order to stimulate the glands to be hormone sensitive in vitro. The animals are sacrificed and the thoracic mammary glands are aseptically excised and incubated for ten days in culture medium supplemented with insulin, prolactin, hydrocortisone and aldosterone. DMBA (7,12-dimethylbenz (a) anthracene) is added to the medium to induce the formation of premalignant lesions. Fully developed glands are then devoid of prolactin, hydrocortisone, and aldosterone, resulting in regress of glands but not premalignant lesions.

The test compound is dissolved in DMSO and added to the culture medium for the duration of the culture period. At the end of the culture period, the glands are fixed in 10% formalin, stained with carman alum and mounted on the glass slide. The incidence of formation of mammary lesions is the proportion of glands with mammary lesions compared to glands without lesions. The incidence of mammary lesions in the glands treated with the test compounds is compared with that of the untreated glands.

The extent of the area occupied by breast lesions can be quantified by projecting an image of the gland on a digitizing support. The area covered by the gland is traced on the support and considered as 100% of the area. The space covered by each of the structures that had not been submitted is also drawn on the digitizing support and quantified by the computer.

Various compounds were examined in the different protocols and were screened for their potential use in the treatment of neoplasms. The results of these tests are described below. Test compounds are designated hereinafter by a letter code corresponding to the following:

A- rac-threo- (E) -1- (N, N'-diethylaminoethanethio) -1- (butane-1 ', 4'-olido) - [3', 4 ': 1,2] -6-fluoro -2-methyl-3- (p-methylsulfonylbenzylidene) -indan;

B- (Z) -5-fluoro-2-methyl-1- (3,4,5-trimethoxybenzylidene) -3-acetic acid;

C-(Z) -5-fluoro-2-methyl-1- (p-chlorobenzylidene) -3-acetic acid;

D- rac- (E) -1- (butane-1 ', 4'-olido) - [3', 4 ': 1,2] -6-fluoro-2-methyl-3- (p-methylsulfonylbenzylidene) - 1S-indanyl-Nacetylcystean;

E- (Z) -5-fluoro-2-methyl-1- (3,4,5-trimethoxybenzylidene) -3-indenylacetamide, N-benzyl;

F- (Z) -5-fluoro-2-methyl-1- (p-methylsulfonylbenzylidene) -3-indenylacetamide, N, N'-dicyclohexyl;

IS 2 174 573 T3

G- ribo- (E) -1-triazolo- [2 ', 3': 1 ", 3"] - 1- (butan-1 ', 4'-olido) - [3', 4 ': 1,2 ] -6-fluoro-2-methyl-3- (p-methylsulfonylbenzylidene) -indane; Y

H- rac- (E) -1- (butane-1 ', 4'-olido) - [3', 4 ': 1,2] -6-fluoro-2-methyl-3- (p-methylsulfonylbenzylidene) - 1S-indanylglutation).

Example 1

COX inhibitory assay

Reference compounds and test compounds were analyzed for their COX inhibitory activity according to the protocol for the COX assay, supra. Figure 4 shows the effect of various concentrations of sulindac or exisulind sulfide on purified cyclooxygenase (Type 1) activity. Cyclooxygenase activity was determined using purified cyclooxygenase from sheep seminal vesicles as previously described (Mitchell et al., Supra). It was calculated that the CI value₅₀ for sulindac sulfide it was approximately 1.76 µΜ, while for exisulind it was greater than 10,000 µΜ. These data show that sulindac sulfide, but not exisulind, is a COX-I inhibitor. Similar data were obtained for the COX-2 isoenzyme (Thompson et al., Journal of the National Cancer Institute, 87: 1259-1260, 1995).

Figure 5 shows the effect of test compounds B and E on COX inhibition. The COX activity was determined the same as for the compounds shown in Figure 4. The data shows that neither of the test compounds B and E significantly inhibits COX-I.

TABLE 2

Cyclooxygenase Inhibitory Activity for a Series of Compounds

<td>Reference compounds</td><td>% Inhibition at 100 μΜ</td>

<td>Indomethacin</td><td> 95</td>

<td>5445</td><td> 94</td>

<td>Sulindac sulfide</td><td> 97</td>

<td>Exisulind</td><td> &lt;25</td>

<td>Test compounds</td><td>% Inhibition at 100 μΜ</td>

<td>TO</td><td> &lt;25</td>

<td>B</td><td> &lt;25</td>

<td>C</td><td> 87</td>

<td>D</td><td> &lt;25</td>

<td>AND</td><td> &lt;25</td>

According to the protocol, supra, compounds A to E were evaluated for their COX inhibitory activity as described in Table 2 above. Compound C was found to inhibit mine COX by 25% at a dose of 100 μΜ and therefore would not be selected for a further screening study.

Example 2

CGMP PDE inhibition assay

Reference compounds and test compounds were analyzed for their cGMP PDE inhibitory activity according to the protocol for the assay described supra. Figure 6 shows the effect of various concentrations of sulindac and exisulind sulfide on purified cGMP PDE4 or PDE activity from cultured HT-29 human colon tumor cells, as previously described (WJ Thompson et al., Supra). The IC50 value of sulindac sulfide for PDE4 inhibition was 41 µΜ, and for cGMP PDE inhibition it was 17 µΜ. The value of CI₅₀ of exisulind for PDE4 inhibition was 181 μΜ, and for cGMP PDE inhibition it was 56 μΜ. These data show that sulindac sulfide and exisulind both inhibit phosphodiesterase activity. Both of them

ES 2 174 573 T3 compounds show selectivity for cGMP PDE isoenzyme forms over PDE4 isoforms.

Figure 7 shows the effects of sulindac sulfide on cGMP or cAMP production determined in HT-29 cells grown according to the assay described above. HT-29 cells were treated with sulindac sulfide for 30 minutes and cGMP or cAMP was measured by the conventional radioimmunoassay method. As indicated, sulindac sulfide increased cGMP levels by more than 50% with an EC value₅₀ 7.3 μΜ (Figure 7A). CAMP levels were not affected by treatment, although a known PDE4 inhibitor, rolipram, increased cAMP (Figure 7B). The data demonstrate the pharmacological significance of cGMP PDE inhibition, relative to PDE4.

Figure 8 shows the effect of the indicated dose of test compound B on the phosphodiesterase isozymes PDE of cGMP or PDE4. The calculated value of the CI₅₀ it was 18 μΜ for the cGMP PDE and it was 58 μΜ for PDE4.

Figure 9 shows the effect of the indicated dose of test compound E on PDE4 or on cGMP PDE. The calculated value of the CI₅₀ it was 0.08 µM for cGMP PDE and greater than 25 µM for PDE4.

TABLE 3

PDE inhibitory activity of cGMP among a number of compounds

<td>Reference compounds</td><td>% Inhibition at 10 μM</td>

<td>Indomethacin</td><td> 34</td>

<td>MY5445</td><td> 86</td>

<td>Sulindac sulfide</td><td> 97</td>

<td>Exisulind</td><td> 39</td>

<td>Test compounds</td><td>% Inhibition at 10 μM</td>

<td>TO</td><td> &lt;25</td>

<td>B</td><td> &lt;25</td>

<td>C</td><td> &lt;25</td>

<td>D</td><td> 36</td>

<td>AND</td><td> 75</td>

The above compounds in Table 3 were evaluated for their PDE inhibitory activity as described in the protocol supra. Of the compounds that did not inhibit COX, only compound E was found to produce more than 50% inhibition at 10 µΝ. As seen in Figure 8, compound B showed greater than 50% inhibition at a dose of 20 µΗ. Therefore, depending on the dose level used in a single dose trial, some compounds that might otherwise be active at slightly higher doses could be ruled out. The dose used is subjective and can be lowered once active compounds are found at certain levels in order to identify even more potent compounds.

Example 3

Apoptosis assay

Reference compounds and test compounds were analyzed for their inhibitory activity against the new PDE according to the protocols for the assay supra. According to these protocols, Figure 10 shows the effects of sulindac sulfide and exisulind on apoptootic and necrotic cell death. HT-29 cells were treated for six days with the indicated dose of sulindac or exisulind sulfide. Apoptotic and necrotic cell death was previously determined (Duke and Cohen, In: Current Protocols in Immunology, 3.17.1-3.17.16, New York, John Wiley and Sons, 1992). The data show that sulindac sulfide and exisulind are both capable of producing apoptotic cell death without inducing necrosis. All data were collected from the same experiment.

IS 2 174 573 T3

Figure 11 shows the effect of sulindac sulfide and exisulind on the inhibition of tumor growth and on the induction of apoptosis determined by DNA fragmentation. Upper figure (11A); growth inhibition (unfilled symbols, left axis) and DNA fragmentation (filled symbols, right axis) by exisulind. Lower figure (11B); growth inhibition (unfilled symbols) and DNA fragmentation (filled symbols) by sulindac sulfide. Growth inhibition was determined by the SRB test after six days of treatment. DNA fragmentation was determined after 48 hours of treatment. All data were collected from the same experiment.

Figure 12 shows the apoptosis-inducing properties of compound E. HT-29 cells of colon adenocarcinoma were treated with the indicated concentration of compound E for 48 hours and apoptosis was determined by DNA fragmentation assay. The calculated EC50 value was 0.05 µM.

Figure 13 shows the apoptosis-inducing properties of compound B. HT-29 cells of colon adenocarcinoma were treated with the indicated concentration of compound B for 48 hours and apoptosis was determined by DNA fragmentation assay. The calculated EC50 value was approximately 175 µM.

TABLE 4

Apoptosis inducing activity among a series of compounds

<td>Reference compounds</td><td>N ° of induction times at 100 μM</td>

<td>Indomethacin</td><td> &lt;2,0</td>

<td>MY5445</td><td> 4,7</td>

<td>Sulindac sulfide</td><td> 7,9</td>

<td>Exisulind</td><td> &lt;2,0</td>

<td>E4021</td><td> &lt;2,0</td>

<td>Zaprinast</td><td> &lt;2,0</td>

<td>Sildenafil</td><td> &lt;2,0</td>

<td>NASH</td><td> &lt;2,0</td>

<td>Test compounds</td><td>N ° of times of Induction at 100 μM</td>

<td>TO</td><td> &lt;2,0</td>

<td>B</td><td> 3,4</td>

<td>C</td><td> 5,6</td>

<td>D</td><td> &lt;2,0</td>

<td>AND</td><td> 4,6</td>

According to the protocol of the induction fold number, supra, compounds A to E were analyzed for their apoptosis inducing activity, as described in Table 4 above. Compounds B, C and E showed significant apoptosis inducing activity, more than 2.0 times, at a dose of 100 µM. Of these three compounds, at this dose only B and E did not inhibit COX but did inhibit cGMP-specific PDE.

The apoptosis inducing activity of a series of phosphodiesterase inhibitors was determined. The data is presented in Table 5 below. HT-29 cells were treated for 6 days with various phosphodiesterase inhibitors. Apoptosis and necrosis were determined morphologically after marking with acridine orange and ethidium bromide according to the test described supra. The data show that the new cGMP-specific PDE is useful for the selection of compounds that induce apoptosis of HT-29 cells.

IS 2 174 573 T3

TABLE 5

Apoptosis induction data for PDE inhibitors

<td>Inhibitor</td><td>Selectivity described</td><td>% Apoptosis</td><td>% Necrosis</td>

<td>Vehicle</td><td></td><td> 8</td><td> 6</td>

<td>8-methoxy-IBMX</td><td>PDE1</td><td> 2</td><td> 1</td>

<td>Milrinone</td><td>PDE3</td><td> 18</td><td> 0</td>

<td>RO-20-1724</td><td>PDE4</td><td> 11</td><td> 2</td>

<td>MY5445</td><td>PDE5</td><td> 80</td><td> 5</td>

<td>IBMX</td><td>Not selective</td><td> 4</td><td> 13</td>

Example 4

Growth inhibition assay

Reference compounds and test compounds were analyzed for their PDE5 inhibitory activity according to the protocol for the assay supra. Figure 14 shows the inhibitory effect of various concentrations of sulindac and exisulind sulfide on the growth of HT-29 cells. HT-29 cells were treated for six days with various doses of exisulind (tri-angles) or sulindac sulfide (squares) as indicated. The number of cells was measured by the sulforhodamine assay as previously described (Piazza et al., Cancer Research, 55: 3110-3116, 1995). The value of CI₅₀for sulindac sulfide it was approximately 45 µM and 200 µM for exisulind. The data show that sulindac sulfide and exisulind are both capable of inhibiting the growth of tumor cells.

Figure 15 shows the growth inhibitory and apoptosis inducing activity of sulindac sulfide. An experiment of variation over time involving HT-29 cells treated with vehicle, 0.1% DMSO (unfilled symbols), or with 120 µM sulindac sulfide (filled symbols) is shown. Growth inhibition (15A, higher) was measured by counting viable cells after staining with Trypan Blue. Apoptosis (15B, lower) was measured by morphological determination after staining with acridine orange and ethidium bromide as previously described (Duke and Cohen, In: Current Protocols in Immunology, 3.17.1-3.17.16, New York, John Wiley and Sons, 1992). The data show that sulindac sulfide is capable of inhibiting the growth of tumor cells and that the effect was accompanied by an increase in apoptosis. All data were collected from the same experiment.

Figure 16 shows the growth inhibitory activity of test compound E. Colon adenocarcinoma HT-29 cells were treated with the indicated concentration of compound E for six days and the number of cells was determined by the SRB assay. The calculated value of the CI₅₀ it was 0.04 µM.

(See Table 6 on the next page)

IS 2 174 573 T3

TABLE 6

Growth inhibitory activity among a number of compounds

<td>Reference compounds</td><td>% Inhibition at 100 μM</td>

<td>Indomethacin</td><td> 75</td>

<td>MY5445</td><td> 88</td>

<td>Sulindac sulfide</td><td> 88</td>

<td>Exisulind</td><td> &lt;50</td>

<td>E4021</td><td> &lt;50</td>

<td>Sildenafil</td><td> &lt;50</td>

<td>Zaprinast</td><td> &lt;50</td>

<td>Test compounds</td><td>% Inhibition at 100 μM</td>

<td>TO</td><td> 68</td>

<td>B</td><td> 77</td>

<td>C</td><td> 80</td>

<td>D</td><td> 78</td>

<td>AND</td><td> 62</td>

According to the selection protocol in the section supra, compounds A to E were analyzed for their growth inhibitory activity, as described in Table 6 above. All test compounds showed greater than 50% activity in the 100 mM single dose assay.

The growth inhibitory activity of a series of phosphodiesterase inhibitors was determined. The data is shown in Table 7 below. HT-29 cells were treated for six days with various phosphodiesterase inhibitors. Cell growth was determined by the SRB assay described supra. The following data taken in conjunction with the above shows that the new PDE inhibitors were effective in inhibiting the growth of tumor cells.

TABLE 7

Growth Inhibition Data for PDE Inhibitors

<td>Inhibitor</td><td>Selectivity described</td><td>Growth inhibitor (IC50, mM)</td>

<td>8-methoxy-IBMX</td><td>PDE1</td><td>&gt; 200 mM</td>

<td>Milrinone</td><td>PDE3</td><td>&gt; 200 mM</td>

<td>RO-20-1724</td><td>PDE4</td><td>&gt; 200 mM</td>

<td>MY5445</td><td>PDE5</td><td>5 mM</td>

<td>IBMX</td><td>Not selective</td><td>&gt; 100 mM</td>

<td>Zaprinast</td><td>PDE5</td><td>&gt; 100 mM</td>

<td>Sildenafil</td><td>PDE5</td><td>&gt; 100 mM</td>

<td>E4021</td><td>PDE5</td><td>&gt; 100 mM</td>

In order to show the efficacy of this selection method on various forms of neoplasia, the compounds were tested on numerous cell lines. The effects of sulindac sulfide and exisulind were determined on various cell lines. The data is shown in Table 8 below. IC50 values ​​were determined by the SRB test. The data show the broad efficacy of these compounds on a wide range of neoplasms, with effectiveness at a comparable dose range. Therefore, the compounds identified and selected by this invention should be useful for the treatment of multiple forms of neoplasia.

IS 2 174 573 T3

TABLE 8

Data on growth inhibition of various cell lines

<td rowspan="2">Cell type / tissue specificity</td><td colspan="3">IC50 ^ M)</td>

<td>Sulindac sulfide</td><td>Exisulind</td><td>Compound AND*</td>

<td>HT-29, Colon</td><td> 60</td><td> 120</td><td> 0,10</td>

<td>HCT116, Colon</td><td> 45</td><td> 90</td><td></td>

<td>MCF7 / S, Mom</td><td> 30</td><td> 90</td><td></td>

<td>UACC375, Melanoma</td><td> 50</td><td> 100</td><td></td>

<td>A-427, Lung</td><td> 90</td><td> 130</td><td></td>

<td>Bronchial Epithelial Cells</td><td> 30</td><td> 90</td><td></td>

<td>NRK, Kidney (not flush)</td><td> 50</td><td> 180</td><td></td>

<td>KNRK, Rinon (ras transformed)</td><td> 60</td><td> 240</td><td></td>

<td>Human Prostate Carcinoma PC3</td><td> 82</td><td> 0,90</td><td></td>

<td>Colo 205</td><td></td><td></td><td> 1,62</td>

<td>DU-145</td><td></td><td></td><td> 0,10</td>

<td>HCT-15</td><td></td><td></td><td> 0,60</td>

<td>MDA-MB-231</td><td></td><td></td><td> 0,08</td>

<td>MDA-MB-435</td><td></td><td></td><td> 0,04</td>

* Determined by the neutral red test as described by Schmid et al., In Proc. ACCR, Vol. 39, p. 195 (1998).

Example 5

Activity in a mammary glandular organ culture model

Figure 17 shows the inhibition of premalignant lesions in a mammary gland organ culture by sulindac metabolites. The mammary gland organ culture experiment was performed as previously described (Mehta and Moon, Cancer Research, 46: 5832-5835, 1986). The results show that sulindac and exisulind effectively inhibited the formation of premalignant lesions, while sulindac sulfide was inactive. The data support the hypothesis that cyclooxygenase inhibition is not necessary for the antineoplaosis properties of the desired compounds.

Analysis

In order to select compounds for the treatment of neoplasms, this invention provides a basis for comparing experimental data of test compounds from various protocols. Within the framework of this invention, test compounds can be classified according to their potential use to treat neoplasms in humans. Those compounds that have desirable effects can be selected for further analysis and later use in humans.

Qualitative data for various test compounds and the different protocols are shown in Table 9 below. The data show that exisulind, Compound B and Compound E exhibit the appropriate activity to pass the selection of four tests: absence of COX inhibition and presence of effective cGMP-specific PDE inhibition, growth inhibition, and apoptosis induction. The activity of these compounds in the culture of mammary glandular organs validates the efficacy of this invention. The qualitative evaluations of the selection protocols classify compound E as the best, followed by compound B and then exisulind.

IS 2 174 573 T3

TABLE 9

Activity profile of various compounds

<td>Compound</td><td>COX inhibition</td><td>PDE inhibition</td><td>Inhibition of increase</td><td>Apoptosis</td><td>Mammary gland organ culture</td>

<td>Exisulind</td><td>_-</td><td> ++</td><td> ++</td><td> ++</td><td> +++</td>

<td>Sulfide</td><td></td><td></td><td></td><td></td><td></td>

<td>sulindac</td><td> ++++</td><td> +++</td><td> +++</td><td> +++</td><td> -</td>

<td>5445</td><td> ++++</td><td> +++</td><td> +++</td><td> +++</td><td> +</td>

<td>TO</td><td> -</td><td> -</td><td> +++</td><td> ++</td><td> ++</td>

<td>B</td><td> -</td><td> +++</td><td> +++</td><td> +++</td><td> ++</td>

<td>D</td><td> -</td><td> -</td><td> ++</td><td> -</td><td> -</td>

<td>AND</td><td> -</td><td> ++++</td><td> ++++</td><td> ++++</td><td> ++++</td>

<td>F</td><td> -</td><td> -</td><td> ++</td><td> +</td><td> -</td>

<td>G</td><td> -</td><td> -</td><td> +++</td><td> ++</td><td> +++</td>

<td>H</td><td> -</td><td> -</td><td> ++</td><td> -</td><td> -</td>

Codes in Table 9: Activity of compounds based on evaluation of a series of experiments involving assays to determine maximum activity and potency.

-Noactive + Slightly active ++ Moderately active +++ Powerfully active ++++ Highly active.

A new assay to determine PKG activity is also described that is used in the selection methods of this invention, but which also has a more general utility to analyze PKG activity for other purposes (for example, for the study of the role of PKG in normal cellular function). For purposes of explanation, it is useful to first describe the PKG assay, before describing how PKG activity can be useful in evaluating drugs to determine if a compound is potentially useful in the treatment of neoplasms.

The new PKG test

The novel PKG assay of this invention involves the binding to a solid phase of several amino acid sequences, each of which contains at least the cGMP binding domain and the phosphodiesterase type 5 phosphorylationin site ("PDE5"). ). This sequence is known and is described in the following literature. Preferably, the linked PDE5 sequence does not include the catalytic domain of PDE5, as described below. One way to join PDE5 sequences to a solid phase is to express those sequences as a fusion protein of the PDE5 sequence and one member of an amino acid binding pair, and chemically join the other member of that amino acid binding pair. to a solid phase (eg, spheres). A binding pair that can be used is glutathione S-transferase ("GST") and glutathione ("GSH"), with GST being expressed as a fusion protein with the PDE5 sequence described above, and GSH covalently bound to the phase. solid. In this way, the PDE5 / GST fusion protein sequence can be bound to a solid phase by simply passing a solution containing the fusion protein over the solid phase, as described below.

An RT-PCR method is used to obtain the cGB domain of PDE5 with forward and reverse primers designed from a cDNA sequence of bovine PDE5A ^ cAl ^ ter-Lucas, β.Μ. et al., J. Biol. Chem., 268: 22863-22873, 1993) and selection among PDE 1-10 families. 5'-3 'kits, Inc. are used to determine total RNA followed by oligo (dT) column purification of mRNA with HT-29 cells. A forward primer (GAA-TTC-TGT-TAG-AAA-AGC-CAC-CAG-AGA-AATG, 203-227) and a reverse primer (CTC-GAG-CTC-TCT-TGT-TTC-TTC-CTC) are used. -TGC-TG, 1664-1686) to synthesize the 1484 bp fragment that encodes the phosphorylation site and cGMP binding sites

ES 2 174 573 T3 high and low affinity of human PDE5A (203-1686 bp, cGB-PDE5). The synthesized cGB-PDE5 nucleoitide fragment encodes 494 amino acids with 97% similarity to bovine PDE5A. It is then cloned into the glutathione S-transferase (GST) fusion vector pGEX-5X-3 (Pharmacia Biotech) with the tac promoter and EcoRI and XhoI cleavage sites. The fusion vector is subsequently transfected into E. coli BL21 (DE3) bacteria (Invitrogen). Transfected BL21 bacteria are grown to the log phase and IPTG is subsequently added as an inducer. Induction is carried out at 20 ° C for 24 hours. The bacteria are collected and lysed. The soluble cell lysate is incubated with GSH conjugated to Sepharose 4B (GSH-Sepharose 4B). The GST-cGB-PDE5 fusion protein can bind to the GSH-Sepharose beads and the other proteins are removed from the beads by washing with excess cold PBS.

The expressed GST-cGB-PDE5 fusion protein is presented in a 7.5% SDS-PAGE gel as an 85 Kd protein. It is characterized by its binding to cGMP and phosphorylation by protein kinases G and A. It has two binding sites to cGMP and K_d is 1.6 ± 0.2 μM, which is close to K_d= 1.3 µM of native bovine PDE5. GST-cGB-PDE5 on GSH-conjugated sepharose spheres can be phosphorylated in vitro by cGMP-dependent protein kinase and cAMP-dependent protein kinase A. The K_m of GST-cGB-PDE5 phosphorylation by PKG is 2.7 μM and the V_max is 2.8 μM, while the K_m of BPDEtide phosphorylation is 68 µM. Phosphorylation by PKG shows a proportion of a molecular phosphate incorporated into a GST-cGB-PDE5 protein.

To analyze a liquid sample believed to contain PKG using the PDE5-bound solid phase described above, the sample and solid phase are mixed with phosphorylation buffer containing ³²Py-ATP. The solution is incubated for 30 minutes at 30 ° C to allow phosphorylation of the PDE5 sequence by PKG to take place, if PKG is present. The solid phase is then separated from the solution (for example, by centrifugation or filtration) and washed with phosphate buffered saline ("PBS") to remove any remaining solution and remove the³²Unreacted Py-ATP.

The solid phase can then be directly analyzed (for example, in a liquid scintillation counter) to determine if the ³²Q. If incorporated, it indicates that the sample contained PKG, as PKG phosphorylates PDE5. If PDE5 is bound through a fusion protein, as described above, the fusion protein containing PDE5 can be eluted from the solid phase with SDS buffer, and the eluent can be analyzed to determine the incorporation of³²Q. This is particularly advantageous if there is a possibility that other proteins are present, as the eluent can be processed (for example, by gel separation) to separate several proteins from each other so that the fraction of the fusion protein can be analyzed to determine the incorporation of ³²P. The phosphorylated fusion protein can be eluted from the solid phase with SDS buffer and subsequently resolved by electrophoresis. If gel separation is performed, proteins can be held to see the position (s) of the protein, and phosphorylation can be measured with³²P of the PDE5 portion of the fusion protein by PKG by exposing the gel to an X-ray film. ³²P becomes visible on the X-ray film, this indicates that PKG was present in the original sample containing PKG, which phosphorylated the PDE5 portion of the fusion protein eluted from the solid phase.

Preferably in the assay, an excess (eg, 100-fold) of a protein kinase inhibitor ("PKI") that specifically and potently inhibits protein kinase A ("PKA") without inhibiting PKG should be added to the test buffer. Inhibition of PKA is desirable as it can contribute to phosphorylation of the PKG substrate (eg, PDE5). By adding PKI, any contribution to phosphorylation by PKA will be eliminated, and any phosphorylation detected would most likely be due to PKG alone.

An assay kit of this invention can be prepared which contains the following reagents prepackaged in separate containers.

1. Cell lysis buffer: 50 mM Tris-HCl, 1% NP-40, 150 mM NaCl, 1 mM EDTA, 1mM Na3VO4, 1 mM NaF, 500 µM IBMX, proteinase inhibitors.

two. Solid phase substrate of protein kinase G: recombinant GST-cGB-PDE5 bound to Sepharose 4B (50% suspension).

3. 2X phosphorylation buffer: ³²Py-ATP (3000 mCi / mmol, 5 --- 10 μCi / assay), KH₂PO₄ 10 mM, K₂HPO₄ 10 mM, 200 μM ATP, MgCl₂ 5 mM.

Four. Inhibitor I of the protein kinase PKA.

Disposable containers and the like in which to carry out the above reactions can also be supplied in the kit.

IS 2 174 573 T3

From all of the above, a person skilled in analytical techniques will easily imagine several ways to adapt the test formats described to other formats. In summary, using at least one portion of PDE5 (or any other protein that can be selectively phosphorylated by PKG), the presence and relative amount (compared to a control) of PKG can be determined by evaluating the phosphorylation of the phosphorylatable protein, using a labeled phosphorylation agent.

SAANDs increase PKG activity in neoplastic cells

Using the PKG assay described above, the following experiments were performed to establish that SAANDs increase PKG activity due to an increase in PKG expression or an increase in cGMP levels (or both) in neoplastic cells treated with a SAAND.

Test procedures

Two different types of PDE inhibitors were evaluated for their effects on PKG in neoplastic cells. A SAAND was evaluated, exisulind, since it is antineoplastic. A class PDE5 inhibitor other than SAAND, E4021, will also be evaluated to determine if the PKG elevation was simply due to class PDE5 inhibition, or if the PKG elevation was involved in the pro-apoptotic effect of the inhibition. by SAANDs of the PDE5 and the new PDE described in US Patent Application No. 09 / 173,375 to Liu et al., filed October 15, 1998.

To analyze the effect of cGMP-specific PDE inhibition on a neoplasm containing the APC mutation, SW480 colon cancer cells were used. SW480 is known to contain the APC mutation. About 5 million SW480 cells in RMPI with 5% serum are added to each of 8 plates:

10 cm plates - 30 µΣ of control DMSO vehicle (no drug), 10 cm plates - 200 µΜ, 400 µΜ, 600 µΜ exisulind in DMSO, and 3 10 cm plates - E4021; 0.1 μΜ, 1 μΜ, and 10 μΜ in DMSO.

Plates are incubated for 48 hours at 37^°C in a 5% CO2 incubator.

The liquid medium is aspirated from the plates (the cells will adhere to the plates). Adhered cells are washed on each plate with cold PBS and 200 μΣ of cell lysis buffer (i.e., 50 mM Tris-HCl, 1% NP-40, 150 mM NaCl, 1 mM EDTA, Na₃VO₄ 1 mM, 1 mM NaF, IBMX 500 µΜ with proteinase inhibitors). Immediately after adding the cell lysis buffer, the lysed cells are collected by peeling off the cells from each plate by scraping. The cell lysate from each plate is then transferred to a microfuge tube, and the microfuge tubes are incubated at 4^°C for 15 minutes while gently shaking the microfuge tubes to allow the cells to fully lyse. Once lysis is complete, the microfuge tubes are centrifuged at maximum speed (14,000 rpm) for 15 minutes. The supernatant from each microfuge tube is transferred to a new microfuge tube.

A protein assay is subsequently performed on the content of each microfuge tube since the amount of total protein will be greater in the control than in the drug-treated samples, if the drug inhibits cell growth. Obviously, if the drug works, the total protein in the drug-treated samples should be virtually the same as in the control. In the above situation, the microfuge tubes with the control and with E4021 needed a dilution to normalize them with respect to the samples treated with the high dose of exisulind (the groups of the lower doses of exisulind had to be normalized with respect to the exisulind sample with the highest dose). Therefore, after performing the protean assays, the total protein concentration of the different samples must be normalized (eg by dilution).

For each drug concentration and for the control, two PKG assays are performed, one with added cGMP and the other without added cGMP, as described in detail below. The reason for performing these two different PKG assays is that cGMP specifically activates PKG. When analyzing PKG activity using the new PKG assay of this invention, it cannot be determined whether an increase in PKG activity is due to increased cGMP in cells (which may be caused by inhibition of cGMP-specific PDE), or if the level of PKG activity is due to increased expression of the PKG protein. Determining PKG activity in the same sample with and without cGMP

ES 2 174 573 T3 added, it can be determined whether the increase in PKG activity, if any, is due to increased expression of PKG. Thus, if an antineoplaosis drug increases the PKG activity relative to the control, it can be established whether the increase induced by the drug is due to an increased expression of the PKG proteon (instead of an activation) in the sample treated with the drug. drug if (1) the sample treated with the drug with extra cGMP shows a higher PKG activity compared to the control sample with extra cGMP, and (2) the sample treated with the drug without extra cGMP shows a higher PKG activity in relation to the control.

Subsequently, parallel samples are prepared with and without added cGMP, 50 μ! of each cell lysate to 20 µl of the PDE5 / GST solid phase substrate suspension described above. For each control or cell lysate sample with the drug to be evaluated, the reaction is started by adding phosphorylation buffer containing a solution with 10 μθϊ of³²Py-ATP (200 µM ATP; 4.5 mM MgCl; 5 mM KH2PO4; 5 mM K2HPO) to each mix. The resulting mixtures are incubated at 30 C for 30 minutes. The mixtures are then centrifuged to separate the solid phase and the supernatant is discarded. The solid phase of each tube is washed with 700 μl of cold PBS. Laemmli sample buffer (Bio-Rad) (30 μ ^) is added to the solid phase. The mixtures are boiled for 5 minutes and loaded in 7.5% SDS-PAGE. The gel is processed at 150 V for one hour. The bands obtained are stained with Coomassie Blue to visualize the bands of the 85 Kd GST-PDE5 fusion proteon, if present. The gel is dried and placed on an X-ray film which, if the PDE5 is phosphorylated, the film will show a corresponding darkened band. The darkness of each band is related to the degree of phosphorylation.

As shown in Figures 18A and 18B, SAAND exisulind causes PKG activity to increase in a dose-dependent manner in samples with added cGMP and in samples without added cGMP, relative to control samples with and without extra cGMP . This is evidenced by the darker appearance of the 85 Kd bands in each of the drug-treated samples. Furthermore, SW480 samples treated with exisulind show a higher phosphorylation activity of PKG with added cGMP in the assay relative to samples treated with exisulind alone (that is, without added cGMP). Therefore, the increase in PKG activity in the samples treated with the drug is not only due to the activation of PKG by the increase in cellular cGMP when SAAND inhibits the specific PDE of cGMP, the increase in PKG activity in neoplasia harboring the APC mutation is also due to increased expression of PKG.

Also, in the fact that the SW480 samples treated with E4021 do not present PKG activation in relation to the control (see Figures 18A and 18B), it shows that the increased activation of PKG caused by SAANDs in the neoplasia containing the APC mutation is not simply due to the inhibition of classic PDE5.

As an analytical technique to evaluate the activation of PKG, instead of exposure to an X-ray film as described above, the 85 Kd band of SDS-PAGE can be evaluated to determine the degree of phosphorylation by cutting the band from the gel. , and any ³²P incorporated into the extraode band can be counted in a scintillation counter (beta) in the window of the ³²P.

To analyze the effect of cGMP-specific PDE inhibition on neoplasms containing the β-catenin mutation, HCT116 colon cancer cells were used. HCT116 is known to contain the β-catenin mutation, but is known not to contain the APC mutation.

The same procedure is used to culture the HCT116 cells as used in the SW480 procedure described above. In this experiment, only exisulind and controls were used. Cells treated with exisulind produced PKG that was phosphorylated to a greater degree than corresponding controls, indicating that PKG activation had occurred in cells treated with the drug that was independent of the APC mutation.

Therefore, for the purposes of the present invention, we mention "β-catenin reduction" in the claims to refer to the wild-type and / or mutant forms of that proteon.

Confirmation of increased PKG expression and decreased β-catenin in SW480 by Western blot

As previously shown, SAANDs produce an increase in PKG expression and an increase in cGMP level, both of which produce an increase in PKG activity in neoplastic cells treated with SAANDs. This increase in the expression of the PKG proteon was true27

ES 2 174 573 T3 subsequently fied by a relatively quantitative Western blot, as described below.

SW480 cells treated with exisulind as previously described are collected from the microfuge tubes by washing once with ice cold PBS. Cells are lysed with a modified RIPA buffer for 15 minutes with shaking. The cell lysate is pelleted by centrifugation in a cold chamber. The supernatants are transferred to new microfuge tubes immediately after centrifugation. A Bio-Rad (Temecula, CA) DC Proteon Assay is performed to determine proteon concentrations in samples. The samples are normalized by the concentration of proteones, as previously described.

50 µg of each sample is loaded onto a 10% SDS gel. SDS-PAGE is performed and proteones are transferred to a nitrocellulose membrane. The transferred nitrocellulose membrane is blocked in fresh TBST containing 5% nonfat dried milk for one hour at room temperature with constant agitation.

A primary goat anti-PKG antibody is diluted to the recommended concentration / dilution in TBST / 5% fresh non-fat dried milk. The nitrocellulose membrane is placed in the primary antibody solution and incubated for one hour at room temperature with shaking. The nitrocellulose membrane is washed three times for ten minutes each with TBST. The nitrocellulose membrane is incubated in a solution containing a POD-conjugated rabbit anti-goat secondary antibody for 1 hour at room temperature with shaking. The nitrocellulose membrane is washed three times for ten minutes each with TBST. Detection is performed using Boehringer Mannheim's BM Blue POD substrate.

As graphically illustrated in Figure 19, exisulind produces a decrease in β-catenin and an increase in PKG, data that were obtained by Western blotting. SW480 cells were treated with exisulind or vehicle (0.1% DMSO) for 48 hours. 50 μg of the supernatant from each of the cell lysates was loaded onto a 10% SDS gel and transferred to a nitrocellulose membrane, and the membrane was probed with rabbit anti-β-catenin and rabbit anti-PKG antibodies. . SAANDs reduce β-catenin levels in neoplastic cells

This observation was made by culturing SW480 cells with 200, 400 or 600 µM exisulind or with vehicle (0.1% DMSO). Cells are collected 48 hours after treatment and processed for immunoblotting. Immunoreactive proteon can be detected by Western blotting. Western blot analysis showed that β-catenin expression was reduced by 50% in exisulind treated cells compared to control. These results indicate that β-catenin has been reduced by treatment with SAANDs. Together with the previous results that establish that PKG activity increases with such treatment and with the later results that establish that β-catenin is phosphorylated by PKG, these results indicate that the reduction of β-catenin in neoplastic cells is initiated by activation. by PKG. Therefore, the use of PKG activity in neoplasms as a selection tool to select antineoplastic compounds is useful.

Phosphorylation of β-catenin by PKG

PKG phosphorylates β-catenin in vitro. The experiment that established this involves immunoprecipitation of the β-catenin-containing complex from SW480 cells (not treated with any drug) in the manner described below under the heading "β-catenin immunoprecipitation". The immunoprecipitated complex, while still trapped in the solid phase (i.e., spheres) is mixed with³²PY-ATP and pure PKG (100 units). Corresponding controls are prepared without added PKG.

The proteon is released from the solid phase by means of an SDS buffer, and the mixture containing the proteon is processed on a 7.5% SDS-PAGE gel. Processing the mixture into the gel removes excess³²Py-ATP from the mixture. Any³²Py-ATP detected in the 93 Kd β-catenin band is therefore due to the phosphorylation of β-catenin. Any increase in³²Py-ATP detected in the 93 Kd β-catenin band treated with extra PKG relative to the control without extra PKG, is due to the phosphorylation of β-catenin in the band treated with extra PKG.

The results we obtained were that there was a notable increase in phosphorylation in the PKG-treated band compared to the control, which had a minimal phosphorylation, virtually undetectable. This result indicates that β-catenin can be phosphorylated by PKG.

IS 2 174 573 T3

Phosphorylation of mutant β-catenin by PKG

The same procedure described in the immediately preceding section was performed with HCT116 cells, which do not contain the APC mutation, but do contain a β-catenin mutation. The results of those experiments also indicate that the mutant β-catenin is phosphorylated by PKG.

Therefore, for the purposes of the present invention, we mention the phosphorylation of β-catenin in the claims to refer to the phosphorylation of wild-type and / or mutant forms of that protein. β-catenin precipitates with PKG

SW480 and HCT116 cell lysate supernatants are prepared in the same manner as described above in Western blot experiments. Cell lysates are pre-clarified by adding 150 ml of a suspension of protein A Sepharose spheres (50%) for every 500 mg of cell lysate and incubation at 4^°C for 10 minutes on a tube shaker. Protein A spheres are removed by centrifugation at 14,000 xg at 4^°C for 10 minutes. The supernatant is transferred to a new centrifuge tube. 10 Mg of rabbit polyclonal anti-β-catenin antibody (Upstate Biotechnology, Lake Placid, New York) is added to 500 Mg of cell lysate. The cell lysate / antibody mixture is gently mixed for 2 hours to 4^°C on a tube shaker. The immune complex is captured by adding 150 μ1 of protein A Sepharose bead suspension (75 μ1 of packed beads) and gently rocking the mixture on a tube shaker overnight at 4^°C. Sepharose beads are collected by pulse centrifugation (5 seconds in the microcentrifuge at 14,000 rpm). The supernatant fraction is discarded and the spheres are washed 3 times with 800 ml of ice cold PBS buffer. The Sepharose beads are resuspended in 150 ml of 2X sample buffer and mixed gently. The Sepharose spheres are boiled for 5 minutes to dissociate the immune complexes from the spheres. The spheres are collected by centrifugation and SDS-PAGE is performed on the supernatant.

A Western blot is performed on the supernatant and the membrane is then probed with a rabbit anti-β-catenin antibody. Subsequently, the membrane is washed 3 times, for 10 minutes each time, with TBST to remove excess anti-β-catenin antibody. Horseradish peroxidase conjugated goat anti-rabbit antibody is added, followed by a 1 hour incubation at room temperature. When this has been done, the presence of βcatenin can be visualized with an HRPO substrate. In this experiment, we were able to clearly visualize the presence of β-catenin.

To detect PKG on the same membrane, the anti-β-catenin antibody conjugate is first detached from the membrane with a 62 mM Tris-HCl buffer (pH 7.6) with 2% SDS and 100 mM 2β-mercaptoethanol in a 55 ° C water bath for 0.5 hours. The bare membrane is then blocked in TBST with 5% nonfat dried milk for one hour at room temperature while the membrane is shaken. The blocked bare membrane is then probed with rabbit anti-PKG polyclonal antibody (Calbiochem, LaJolla, CA), which is detected with a second HRPO-conjugated goat anti-rabbit antibody. The presence of PKG in the transferred membrane is visualized with an HRPO substrate. In this experiment, the PKG was, in effect, visualized. Since the only proteins on the membrane are those that immunoprecipitated with β-catenin in cell supernatants, this result clearly establishes that PKG was physically bound to the protein complex containing β-catenin in cell supernatants.

The same Western blot membrane was also probed after being stripped with anti-GS ^^ antibody to determine if it also co-precipitated with β-catenin. In that experiment, we also detected GSKS-β on the membrane, indicating that GSKB-β precipitated with β-catenin and PKG, suggesting that all three proteins may be part of the same complex. Since GSKS-β and β-catenin are part of the APC complex in normal cells, this PKG can be part of the same complex and may be involved in the phosphorylation of β-catenin as part of that complex. Antineoplastic Pharmaceutical Compositions Containing cGMP PDE Inhibitors

Seguín explained previously, exisulind is a compound that has desirable antineoplastic properties. Its efficacy and use as an antineoplastic was discovered before it was known that the compound acted by inhibiting the specific PDE activity of cGMP in neoplastic cells.

IS 2 174 573 T3

Among other things, the verification that the selection process of this invention could be used in order to select compounds for the treatment of humans was obtained in human clone trials in patients with neoplasms. Knowing later the fact that exisulind was antineoplastic (in vitro), that it had the profile of a desirable compound that fulfilled the selection criteria of this invention, the success of the compound in two human clone trials establishes that other compounds can be selected that meet the selection criteria of this invention.

As indicated above, several neoplasms harbor the APC mutation. Among other things, verification of the selection process of this invention was established in human clone trials in patients with neoplasms harboring the APC mutation.

The APC mutation was first discovered in patients with the inherited neoplasia adematous polyposis coli ("APC"). APC disease is characterized by the appearance in adolescence of hundreds to thousands of polyps in the colon, and the common therapy is surgical removal of the colon before the age of 20 years.

The first clone trial involved APC patients using exisulind. In that study, each patient's colon was already removed, except for a small section of the colon adjacent to the rectum (to which the small intestine was attached) to preserve rectal function. However, in such patients, polyps commonly form in the small section of the colon that remains, polyps that require periodic removal (eg, by electrocautery).

That trial in which exisulind was selected was a prevention trial designed to evaluate the antineoplastic characteristics of the drug by comparing the cumulative number of new polyps formed over twelve months in the drug-treated group and in the placebo-treated group. Eligible patients were those who formed between 9 and 44 polyps per year. Patients underwent total ablation (all polyps were removed) at the beginning of the study, at the end of 6 months, and at the end of 12 months. Thirty-four eligible patients were enrolled in the study. Based on the estimated mean number of polyps formed over a year in APC patients who had historically produced 9 to 44 polyps per year, exisulind was significantly better, clonically and statistically, than placebo in decreasing the rate of formation of polyps. Based on the mean number of polyps produced in the first six months of the study, exisulind-treated patients developed approximately one-third of the number of polyps produced by placebo-treated patients (mean values: 9 polyps / year and 26 polyps / year , respectively; p = 0.013). Based on the mean number of polyps produced over the total 12 months of the study, exisulind-treated patients produced approximately half the number of polyps as placebo-treated patients (mean values: 18 polyps / year and 38 polyps / year, respectively; p = 0.020).

A separate clone study was also carried out with male patients who had prostate cancer and as a result of this the prostate had been removed. The study was conducted in patients with detectable levels of PSA (prostate specific antigen) that were increasing after radical prostatectomy, indicating a recurrence of prostate cancer.

96 patients were enrolled in the evaluation of prostate cancer: a multicoentric, double-blind, placebo-controlled trial, which involved the administration of exisulind to patients receiving the drug at 500 mg / day. As presented later, the data show a statistically significant difference in PSA levels between the exisulind-treated group and the placebo-treated group. PSA levels in the exisulind group were significantly reduced compared to PSA levels in the placebo group. Although a rising PSA level is not by itself a disease condition, it is widely considered in the medical community as a surrogate marker indicative of the presence of a recurrence of prostate cancer in such men.

In addition to performing an assessment based on differences in mean PSA levels between the exisulind and placebo groups as a whole, the interim analysis included a subgroup analysis. Study patients were classified into high-risk, intermediate-risk, and low-risk groups in terms of their risk of developing metastatic disease. This classification was carried out using the methodology published in the Journal of the American Medical Association (JAMA, May 5, 1999, pp. 1591-97). To determine which study patients fell into each of the risk groups, medical records were provided to an investigator who was blinded as to whether the patients were in the drug or placebo group; He assigned the study patients to the appropriate risk groups according to the previously referenced published methodology. The

ES 2 174 573 T3 Statistical analysis reveals statistically significant differences in mean PSA levels between exisulind and placebo patients in the high and intermediate risk groups.

The data from the prostate study are as follows:

TABLE 10

Effect of exisulind on the mean PSA level in post-prostatectomy men with increasing PSA

<td>Group</td><td>Placebo</td><td>Exisulind</td><td>Value of "p"</td>

<td>Total</td><td> 4,49</td><td> 2,85</td><td> 0,0004</td>

<td>High risk</td><td> 4,98</td><td> 2,91</td><td> 0,0002</td>

<td>Intermediate Risk</td><td> 6,24</td><td> 2,95</td><td> 0,0053</td>

In these trials with exisulind and in several others that involved the drug in other indications, safety will be assessed by monitoring adverse events (AEs), clinical laboratory tests (hematology, serum chemistry and urianalysis), vital signs (blood pressure, blood speed). pulse, respiratory rate, temperature and weight), phasic examination and upper endoscopy.

No notable safety concerns have been demonstrated in clinical trials of exisulind to date in more than 400 patients. Exisulind will not demonstrate blood dyscrasia, dose-limiting vomiting, or neurolagic or renal toxicity associated with conventional chemotherapeutic agents. It also did not produce any clinically significant change in vital signs. In fact, in paired biopsies of colonic tissues with polyps and normals in patients with APC, it was found that exisulind increased apoptosis rates in colonic tissues with polyps but not in normal ones, suggesting minimal effects on normal tissues.

At doses above the maximum tolerated dose (MTD = 600 mg in patients with subtotal colectoma; 400 mg in patients with intact colon; 350 mg in pediatric patients), the only dose-limiting adverse events found were elevations in the trials. liver function tests (LFTs) observed early in treatment. When they did occur, elevations in LFTs were rapidly reversible and did not recur once the dose was lowered. Other events (eg, occasional abdominal pain) were typically short-lived and mild to moderate in intensity, and did not require interruption or reduction of the exisulind dose.

In summary, these trials demonstrated that exisulind is an effective, well-tolerated cranial therapy for the treatment of cancer of the skin. Thus, these results illustrate that selection of an additional compound that inter alia inhibits cGMP specific PDE activity (in addition to meeting the other selection criteria of this invention) may result in a therapeutically effective drug in vivo.

A second drug, which was also invented before its mechanism of action was found to involve cGMP inhibition and before it was known to meet the selection criteria for this invention, is (Z) -5-fluoro-2- methyl- (4-pyridylidene) -3- (N-benzyl) indenylacetamide hydrochloride (Compound I). It has been shown in in vitro and in vivo evaluations to be antineoplastic with activity against a wide range of neoplasms. It is also safe in animal studies and in one human study with increasing doses.

As a person skilled in the art will recognize from the data presented below, Compound I can be administered safely to animals at doses much higher than the tolerable (and in many cases toxic) doses of conventional chemotherapeutic products or chemotherapeutic agents. Antineoplastic NSAIDs. For example, in an acute toxicity study in rats, oral uanic doses of Compound I administered (in a 0.5% carboxymethylcellulose carrier) at doses up to and including 2000 mg / kg resulted in no observable signs of toxicity. . At 4000 mg / kg, body weight gains were slightly reduced. A single dose of 1000 mg / kg administered intraperitoneally resulted in a reduction in body weight gain, with mesenteric adhesions observed in some animals in this group at the time of necropsy.

In dogs, administration of Compound I in capsules at 1000 mg / kg did not result in signs of toxicity in the single group of two male and two female dogs. Due to the nature of the capsules with Compound I, this dose required the use of at least 13 capsules for each

ES 2 174 573 T3 animal, which was considered to be the maximum number without subjecting the animals to striae. Therefore, these dogs were subsequently administered seven consecutive doses of 1000 mg / kg / day. At no time during any of the dosing phases were obvious signs of drug-related effects observed.

Therefore, based on a single dose, Compound I is not acutely toxic. Based on the findings of these studies, the oral LD50 of Compound I was considered to be greater than 1000 mg / kg in dogs and 4000 mg / kg in rats, and the intraperitoneal LD50 was considered to be greater than 1000 mg / kg in rats.

A seven-day study to search for a range of doses in rats, in which Compound I was evaluated by administering it at doses of 0, 50, 500 or 2000 mg / kg / day, resulted in no observable signs of toxicity at 50 mg. / kg / day. At 500 mg / kg / day, treatment-related effects were limited to an increase in absolute and relative liver weights in female rats. At 2000 mg / kg / day, the effects included labored breathing and / or abnormal breath sounds, decreased weight gains and food intake in male rats, and increased liver weights in female rats. No haematological or blood chemistry changes, nor microscopic pathological changes, were observed at any dose level.

A 28-day study was also carried out in rats at 0, 50, 500 and 2000 mg / kg / day. There were no abnormal clinical observations attributed to CP-461, and changes in body weight, ophthalmoscope tests, hematology and blood chemistry values, and urine tests were normal. No gross tissue changes were observed at necropsy. Organ weight data revealed a statistically significant increase in liver weights at 2000 mg / kg / day, and statistically significant increases in thyroid weights in the 2000 mg / kg / day group. The slight increases in my lower doses were not statistically significant. Histopathological evaluation of the tissues indicated the presence of traces of follicular cell hypertrophy, an increased number of mitotic figures (suggesting possible cell proliferation) in the thyroid glandula, and mild centrilobular hypertrophy in the liver. These changes were generally limited to a small number of animals at the 2000 mg / kg / day dose, although a female at the 500 mg / kg / day dose had increased thyroid gland figures. Findings in the liver may be indicative of very mild stimulation of microsomal enzymes, resulting in increased metabolism of thyroid hormones, which in turn would result in stimulation of the thyroid. Thus, a person skilled in the art would recognize that these effects are extremely minimal compared to what would be expected from conventional chemotherapeutics or NSAIDs at similar doses.

In order to further establish the safety profile of Compound I, a study was carried out to evaluate whether Compound I-induced apoptosis in proist tumor cell lines was comparable with its effects on normal tissue-derived prostate epithelial cells. The androgen-sensitive priostate tumor cell line, LNCaP (from ATCC (Rockville, MD)) was propagated under standard conditions using RPMI 1640 medium containing 5% fetal calf serum and 2 mM glutamine. Primary proist epithelial cell (PrEC) cultures derived from normal proist (from Clonetics Inc. (San Diego, CA)) were cultured under the same conditions as the tumor cell line except that an optimized serum-free medium was used for the growth of such cultures (Clonetics Inc.). For the experiments, LNCaP or PrEC cells were seeded in 96-well plates at a density of 10,000 cells per well. After 24 hours, the cells were treated with vehicle (DMSO 0.1%) or with Compound I (free base) 50 MM solubilized in DMSO. After several times of drug treatment (4, 24, 48, 72 or 99 hours) the cells were lysed and processed to measure DNA associated with histones as an indicator of apoptite cell death (see, Piazza et al., Cancer Research , 57: 2452-2459, 1997).

Figure 27 shows a time-dependent increase in the amount of histone-associated fragmented DNA in LNCaP cell cultures after treatment with 50 MM Compound I (free base). A significant increase in fragmented DNA was detected after 24 treatment, and the induction was maintained for 4 days of continuous treatment. In contrast, treatment of PrEC cells (from "normal" priostate) with Compound I (50 MM) did not affect DNA fragmentation for 4 days of treatment. These results demonstrate a selective induction of apoptosis in neoplastic cells, rather than in normal cells. This is in marked contrast to conventional chemotherapeutic products that induce apoptosis or necrosis in the same way in normal cells and in rapidly growing neoplastic cells.

IS 2 174 573 T3

Finally, in terms of safety, in a single human clinical trial with increasing doses, in patients of the human safety study who took the drug orally, Compound I did not produce significant side effects at any dose, including doses above the level predicted necessary to produce anticancer effects.

Following previously indicated, Compound I also has potent antineoplastic properties. The value of CI₅₀ for growth inhibition obtained for Compound I it was 0.7 µΜ in the SW480 cell line. This result has been confirmed by the evaluation of Compound I in rodents using aberrant crypt foci ("ACF") as indicators of carcinogenesis (see, Bird, Cancer Lett., 37: 147-151, 1987). This established rodent model of azoxymethane ("ΆΘΜ") -induced carcinogenesis was used to determine the effects of Compound I (free base and salt) on the development of colon cancer in vivo. ACFs are precursors of colon tumors and ACF inhibition is predictive of chemopreventive efficacy.

In the rats in this experiment, the establishment of ACF was achieved by two consecutive weekly injections of the carcinogen. Compound I was administered one week prior to establishment of ACF and for the duration of the experiment. ACFs were assessed after 5 weeks of treatment. In Compound I 344 male Fisher rats were orally administered in the rat chow. Daily food intake (mg / kg body weight) varied during the course of the study and therefore the dose of Compound I was expressed as grams per kg of diet in order to provide a basis for comparison between doses. To determine whether Compound I had an adverse effect on growth and / or eating behavior, body weight was determined throughout the course of the experiment. The experimental groups gained less weight than the controls, which was indicative of bioavailability. However, the differences in weight were less than 10% and it was considered that they did not affect the formation of SCA.

The free base of Compound I inhibited ACF formation as measured by a reduction of crypts per colon. The data are summarized in Table 11. With the exception of the low dose group (only 0.5 g / kg of diet), the differences between the treatment and control groups were substantial, and statistically significant in the case of the 1.0 and 2.0 g / kg diet groups.

TABLE 11

Inhibition of aberrant crypt foci by Compound I

<td>Compound dosage</td><td>n</td><td>Mean ACF / colon</td><td>% Control</td><td>p (t test)</td>

<td>(g / kg of diet)</td><td></td><td>(± ES)</td><td></td><td>vs. control</td>

<td>Control</td><td> 10</td><td> 149 ± 9</td><td> -</td><td> -</td>

<td> 0,5</td><td> 7</td><td> 149 ± 14</td><td> 100</td><td> 0,992</td>

<td> 1,0</td><td> 10</td><td> 111 ± 9</td><td> 75</td><td> 0,008</td>

<td> 1,5</td><td> 10</td><td> 132 ± 4</td><td> 89</td><td> 0,101</td>

<td> 2,0</td><td> 10</td><td> 107 ± 15</td><td> 72</td><td> 0,029</td>

Furthermore, Compound I retrospectively met the selection criteria of this invention and was one of the compounds used to establish the validity of this selection criteria. For example, using previously described protocols, Compound I has a cGMP-specific PDE IC50 value of 0.68 µΜ using cGMP-specific PDE from HT-29 cell extracts. His COX-I inhibition (at 100 μΜ) was less than 25%.

As for being pro-apoptoitic, the EC50 of DNA fragmentation of Compound I was 15 µΜ. In addition, Table 12 shows the percentage of apoptosis for Compound I in SW480 at various drug concentrations.

IS 2 174 573 T3

TABLE 12

Induction of apoptosis in colon adenocarcinoma SW480 cells by Compound I determined by morphology

<td>Treatment</td><td>Dose</td><td>% Apoptosis</td>

<td>Vehicle (DMSO 0.1%)</td><td>_-</td><td> 1</td>

<td>Compound I</td><td>0.35 μM</td><td> 16</td>

<td>Compound I</td><td>0.7 μM</td><td> 27</td>

<td>Compound I</td><td>1.5 μM</td><td> 88</td>

The activity of Compound I is not limited to activity against colon cancer cell lines or in colon cancer animal models. It has a wide range of antineoplastic effects on various neoplastic cell lines. Various types of human cancer cell lines were propagated under sterile conditions in RPMI 1640 medium with 10% fetal bovine serum, 2 mM L-glutamine and sodium bicarbonate. To determine the growth inhibitory effects of Compound I, cells were seeded in 96-well plates at a density of 1000 cells per well. Twenty-four hours after plating, various concentrations of the free base of Compound I solubilized in DMSO (0.1% final concentration) were dosed to the cells. The effect of the drug on the growth of tumor cells was determined using the neutral red cytotoxicity test after five days of continuous treatment. Neutral red is a dye that is selectively taken up by viable cells through an ATP-dependent transport mechanism.

As summarized in Table 13, Compound I (free base) exhibited potent growth inhibitory activity when evaluated against a panel of cultured human cell lines derived from various tissue origins. Compound I exhibited comparable growth inhibitory effects regardless of the tumor histogenesis from which the cell lines were derived. The GI50 value (concentration of the drug that inhibits growth by 50% relative to the vehicle control) calculated for all cell lines was 1-2 µM.

In addition to the data in the following table, we observed a sensitivity comparable to Compound I (HCl salt) of human leukemia cell lines (CCRF-CEM, K562 and Molt-4), of a myeloma cell line (RPMI-8226), from a pancreatic tumor cell line (PANC-1) and from an ovarian tumor cell line (OVCAR-3).

TABLE 13

Inhibition of the growth of various lones of human tumor cells by Compound I

<td>Cellular lone</td><td>Origin of the tumor</td><td>GI50 © M)</td><td>GI₉o ^ M)</td>

<td>Colo 205</td><td>Colon</td><td> 1,6</td><td> 2,4</td>

<td>HCT-15</td><td>Colon</td><td> 1,7</td><td> 3,0</td>

<td>HT-29</td><td>Colon</td><td> 2,1</td><td> 8,0</td>

<td>SW620</td><td>Colon</td><td> 1,7</td><td> 2,5</td>

<td>DU145</td><td>Prostate</td><td> 1,6</td><td> 2,8</td>

<td>PC-3</td><td>Prostate</td><td> 1,7</td><td> 82,5</td>

<td>NCI-H23</td><td>Lung</td><td> 1,7</td><td> 2,5</td>

<td>NCI-H322M</td><td>Lung</td><td> 2,1</td><td> 13,2</td>

<td>NCI-H460</td><td>Lung</td><td> 1,9</td><td> 30,0</td>

<td>NCI-H82</td><td>Lung</td><td> 1,7</td><td> 5,8</td>

<td>MDA-MB-231</td><td>Mommy</td><td> 1,8</td><td> 77,6</td>

<td>MDA-MB-435</td><td>Mommy</td><td> 1,6</td><td> 2,3</td>

IS 2 174 573 T3

TABLE 13 (continued)

<td>Cellular lone</td><td>Origin of the tumor</td><td>GI50 ^ M)</td><td>GI90LM)</td>

<td>UISO-BCA-1</td><td>Mommy</td><td> 1,5</td><td> 4,7</td>

<td>Molt-4 *</td><td>Leukemia</td><td> 1,6</td><td>ND</td>

<td>CCRF-CEM *</td><td>Leukemia</td><td> 1,4</td><td>ND</td>

<td>K-562 *</td><td>Leukemia</td><td> 1,8</td><td>ND</td>

<td>RPMI-8226 *</td><td>Myeloma</td><td> 1,2</td><td>ND</td>

<td>OVCAR *</td><td>Ovary</td><td> 1,2</td><td>ND</td>

<td>PANC-1 *</td><td>Pancreas</td><td> 2,2</td><td>ND</td>

* The analysis was performed with the free base of the compound unless otherwise indicated with an asterisk, in which case the analysis was performed with the HCl salt.

Given the safety characteristics in animals and humans, and the efficacy in animals and in a very wide range of cell cultures of Compound I, it is clear that compounds that meet the selection criteria of this invention (including inhibition of specific PDE of CGMP) are useful antineoplastic therapeutic products.

Regarding the identification of structurally additional cGMP-specific PDE inhibitor compounds that may be therapeutically effective as antineoplastic agents, Several useful model compounds described herein (as well as their anaologs incorporated by reference) are available to a person skilled in the art that can be used as the basis for computer modeling of additional compounds that have the same conformations but are chemically different. For example, a computer program such as that sold by Molecular Simulations Inc., WebLab edition^&lt;RR ViewerPro^TM, includes molecular visualization and chemical communication capabilities. Such a program includes functionality, including 3D visualization of known active compounds to validate the accuracy of drawn or imported chemical structures. Furthermore, the program allows structures to be superimposed on the basis of characteristics defined by the user, and the user can measure distances, angles or dihedrons.

In this situation, as the structures of other active compounds are described above, group analysis and 2D and 3D similarity search techniques can be applied with such a program in order to identify new potential additional compounds that can be subsequently screened. and selected according to the selection criteria of this invention. These computer program methods are based on the principle that compounds that appear similar or have similar properties are very likely to also have similar activity, which can be confirmed using the selection criteria of this invention.

Similarly, when such additional compounds are modeled by computer, many such compounds and variants thereof can be synthesized using known combinatorial chemistry techniques that are commonly used by those of ordinary skill in the pharmaceutical industry. Examples of a few combinatorial chemistry services that can be contracted include those offered by New Chemical Entities, Inc. from Bothell Washington; Protogene Laboratories, Inc. of Palo Alto, California; Axys, Inc. of South San Francisco, California; Nanosyn, Inc. of Tucson, Arizona; Trega, Inc. of San Diego, California and RBI, Inc. of Natick, Mass. There are several other recruiting companies as well. Several large drug companies have a similar, if not superior, capacity of their own. In summary, a person skilled in the art can easily produce many compounds to be screened to select promising compounds for the treatment of neoplasms having the attributes of the compounds described herein. To help further identify compounds that can be screened and then selected using the criteria of this invention, it is interesting to know the binding of selected antineoplaosis compounds to the PDE5 proteon. By the procedures discussed below, desirable, preferable compounds meeting the selection criteria of this invention were found to bind to the catalytic cGMP region of PDE5.

To establish this a PDE5 sequence was used that did not include the catalytic domain. One way to produce such a sequence is to express that sequence as a fusion proteon, preferably with glutathione S-transferase ("GST"), for reasons that will be obvious.

IS 2 174 573 T3

The RT-PCR method is used to obtain the cGB domain of PDE5 with a forward primer and a reverse primer designed from a bovine PDE5A cDNA sequence (McAllister-Lucas, LM et al., J. Biol. Chem. ., 268: 22863-22873, 1993) and selection among PDE families 1-10. 5'-3 'kits, Inc. are used to determine total RNA followed by oligo (dT) column purification of the mRNA with HT-29 cells. A forward primer (GAA-TTC-TGT-TAG-AAA-AGC-CAC-CAGAGA-AAT-G, 203-227) and a reverse primer (CTC-GAG-CTC-TCT-TGT-TTC-TTC-CTC) are used. -TGC-TG, 1664-1686) to synthesize the 1484 bp fragment encoding the phosphorylation site and the high and low affinity cGMP binding sites of human PDE5A (203-1686 bp, cGB-PDE5). The synthesized cGB-PDE5 nucleaotide fragment encodes 494 amino acids with 97% similarity to bovine PDE5A. It is subsequently cloned into the glutathione S-transferase (GST) fusioan vector pGEX-5X-3 (Pharmacia Biotech) with the tac promoter and EcoRI and XhoI cleavage sites. The fusioan vector is subsequently transfected into E. coli BL21 (DE3) bacteria (Invitrogen). The transfected BL21 bacteria are cultured to the log phase and IPTG is subsequently added as an inducer. Induction is carried out at 20^°C for 24 hours. The bacteria are collected and lysed. The soluble cell lysate is incubated with GSH conjugated to Sepharose 4B (GSH-Sepharose 4B). The fusioan protein GST-cGB-PDE5 can bind to the GSH-Sepharose beads and the other proteins are removed from the beads by washing with an excess of cold PBS.

The expressed GST-cGB-PDE5 fusioan protein is presented on a 7.5% SDS-PAGE gel as an 85 Kd protein. It is characterized by its binding to cGMP and phosphorylation by protein kinases G and A. It has two binding sites to cGMP and K_d is 1.6 ± 0.2 μΜ, which is close to K_d= 1.3 μΜ of the native bovine PDE5. GST-cGB-PDE5 on GSH-conjugated sepharose beads can be phosphorylated in vitro by cGMP-dependent protein kinase and cAMP-dependent protein kinase A. The K_m of the phosphorylation of GST-cGB-PDE5 by PKG is 2.7 μΜ and the V_max is 2.8 μΜ, while the K_m of BPDEtide phosphorylation is 68 µΜ. Phosphorylation by PKG presents a molecular phosphate incorporated into a GST-cGB-PDE5 protein in a one-to-one ratio.

A cGMP binding assay of the compounds of interest (Francis SH et al., J. Biol. Chem., 255: 620-626, 1980) is performed in a total volume of 100 containing 5 mM sodium phosphate buffer (pH = 6.8), 1 mM EDTA, 0.25 mg / ml BSA, ³H-cGMP (2 μΜ, NEN) and the fusion protein GST-cGBPDE5 (30 μg / assay). Each compound to be analyzed is added at the same time as the substrate.³H-cGMP, and the mixture is incubated at 22^°C for 1 hour. Subsequently, the mixture is transferred to a Brandel MB-24 cell collector with GF / B as filter membrane followed by 2 washes with 10 ml of cold 5 mM potassium buffer (pH 6.8). The membranes are subsequently cut and transferred to scintillation vials, followed by adding to each vial 1 ml of H2O and 6 ml of Ready Safe liquid scintillation cocktail.^TM. The vials are counted on a Beckman LS 6500 scintillation counter.

For calculations, blank samples are prepared by boiling the unioan protein for 5 minutes, and the unioan counts are &lt;1% when compared to the unboiled protein. The damping caused by the filter membrane or other debris is also calibrated.

The inhibitors of PDE5, sulfur, exisulind, Compound B, Compound E, E4021 and zaprinast, and the analogs of caclic nucleoatides, cAMP, caclic IMP, 8-bromo-cGMP, calcium UMP, caclic CMP, 8-bromo-cAMP were selected. , 2'-O-butyl-cGMP and 2'-O-butyl-cAMP in order to analyze whether they could bind competitively to the cGMP binding sites of the GST-cGB-PDE5 protein. The results are shown in Figure 24. cGMP specifically binds to the GST-cGB-PDE5 protein. Calclic AMP, cAMP, CMP, 8-bromo-cAMP, 2'-O-butyl-cAMP and 2'-O-butyl-cGMP do not compete with cGMP at the junction. Cyclic IMP and 8-bromo-cGMP at high concentration (100 µΜ) may partially compete for binding to cGMP (2 µΜ). None of the PDE5 inhibitors showed competition with cGMP to bind GST-cGB-PDE5. Therefore, they do not bind to the sites of unioánaGMPcdelaPDE5.

However, compound E binds competitively (with cGMP) to PDE5 (that is, Peak A). (Compound E also binds competitively (with cGMP) to Peak B of PDE). Since Compound E does not bind to the cGMP binding site of PDE5, the fact that there is competitive binding between Compound E and cGMP means that desirable compounds such as Compound E bind to the catalytic cGMP site on PDE5, information that is readily obtainable by a person skilled in the art (with conventional competitive bonding experiments) but can help a person skilled in the art to more easily model other compounds. Therefore, with the chemical structures of the desirable compounds presented herein and the information on the cGMP binding site, a person skilled in the art can model, identify and select (using the selection criteria of this invention) other compounds. Chemicals to be used as therapeutic agents.