PARP INHIBITORS
Field of the invention
This invention relates to the fields of molecular biology and medicinal chemistry. More particularly, the invention relates to compounds and methods for modulating the activity of PARP.
Background of the Invention
Living organisms possess mechanisms to regulate cell cycle progression and to preserve genomic integrity. Failure of these mechanisms in multicellular organisms results in disorders ranging from the unregulated cell proliferation associated with cancer, to massive cell death following the fall of tissue oxygen and glucose levels in cardiac or brain ischemia (D.W. Choi, Nat Med (1997) 3(10.1:1073-74) .
A key cellular response to genomic damage is the post-translational modification of nuclear proteins in response to DNA strand breaks (PA. Jeggo, Curr Biol (1998)
8(2):R49-51; A.A. Pieper et al, Trends Pharmacol Sci (1999) 20£4):171-81). One known modification is the addition to specific proteins of up to 200 residues of ADP-ribose to form branched polymers. These polymers act as binding sites for repair proteins that play a central role in DNA metabolism (D. D'Amours et al, Biochem J (1999) 342(Pt 2):249- 68). The enzyme responsible for the addition of these polymers is poly (ADP-ribose) polymerase, or PARPl. PARPl associates with DNA and with chromatin binding proteins such as histones, transcription factors and key DNA repair proteins. Although a number of nuclear proteins such as histones are substrates for PARPl, a major substrate is PARPl itself, via auto-modification of the BRCA1 C-terminal homology (BRCT ) region. Regulation of auto-modification of PARP 1 is twofold: through PARP 1 -DNA interactions as well as PARPl-PARPl dimerization (H. Mendoza-Alvarez et al., J Biol Chem (1993) 268(30 22575-80).
PARPl acts together with the DNA damage repair system to regulate DNA base excision repair, apoptosis and necrosis (D. D'Amours et al., supra). Studies of mouse strains lacking the PARPl gene have identified two roles for this protein depending on the extent of DNA damage (V. Burkart et al., Nat Med (1999) 50}:314-19; M. Masutani
et al., Proc Natl Acad Sci USA (1999) 96(5):2301-04: M. Masutani et al., Mutat Res (2000) 462(2-3): 159-66). Moderate damage elicits a protection response similar to that observed for checkpoint genes, leaving PARPl knockout mice vulnerable to gamma irradiation and alkylating reagents. In cases of extensive DNA damage, PARPl activity depletes cellular energy pools, which eventually leads to cell death (V. Burkart et al., supra; A.A. Pieper, supra).
PARPl also has a putative role in signaling DNA damage and in recruiting proteins to sites of double strand breaks. This hypothesis was based on the ability of proteins, such as p53 and other repair enzymes, to bind to the poly (ADP) polymers present on PARPl (T. Lindahl et al, Trends Biochem Sci (1995) 20(10):405-11: D. D'Amours et al., supra). PARPl inhibitors exaggerate the cytotoxic effects of DNA damage by limiting the ability of cells to regulate DNA base excision repair. In this role inhibitors are being tested as chemosensitizing agents during cancer chemotherapy (P. Decker et al, Clin Cancer Res (1999) 5(5): 1169-72). Another response to more extensive DNA damage mediated by PARPl is the promotion of cell death as in cases of ischemic injury (C.a.D. Szabo, Trends Pharmacol Sci (1998) 19(7):287-98). This process can occur when PARPl activation is highly stimulated, and thus consumes large amounts of NAD, the source of ADP-ribose. This condition depletes the cellular energy stores (K. Takahashi et al., Brain Res (1999) 829(1- 2):46-54; A.A. Pieper et al., supra). PARPl knockout mice are highly resistant to ischemia during steptozocin induced type I diabetes, myocardial infarction, stroke, and neurodegeneration (Burkart, Wang et al. 1999; Shall and de Murcia 2000) (V. Burkart et al., supra). In support of a role for PARPl in cell death in various inflammation processes, several studies have shown protection against cellular injury in numerous target cells by using known PARPl inhibitors (C.a.D. Szabo, supra).
For many years PARPl has been the only known poly(ADP-ribose) polymerase. However, modification of cellular proteins with ADP-ribose polymers still occurs in PARPl knockout mice, suggesting the presence of other proteins with PARPl activity (M.K. Jacobson et al, Trends Biochem Sci (1999) 24(11):415-17). Indeed, new mem- bers of the PARP family have been identified based on the presence of domains that share considerable sequence similarity with the catalytic domain of PARPl (S. Smith et
al., Science (1998) 282(5393): 1484-87: J. Ame, N. Rolli et al., J Biol Chem (1999) 274(25): 17860-68; M. Johansson, Genomics (1999) 57(3):442-45: N.A. Kicl hoefer et al., J Cell Biol (1999) 146(5):917-28: F.R. Sallmann et al, J Biol Chem (2000) 275(20):15504-11) (see Fig. 1). PARPl has three functional domains; a Zn2+finger DΝA binding sequence, an automodification domain, and a C-terminal catalytic domain. All members of the PARPl family share a conserved catalytic domain, but a few also share a DΝA binding domain or in the case of Tankyrase and NPARP, have functional domains that are not present in PARPl. Although some members of the PARP family do not possess a well-defined Zn2+ finger DΝA binding motif or an automodification domain like that described for PARP 1 , they still catalyze the formation of ADP-ribose polymers in a DΝA dependent manner, and are capable of automodification (Ame & Rolli et al., 1999, supra; M. Johansson, supra).
Two additional members of the PARP family are tankyrase and NPARP (S. Smith et al., 1998, supra; N.A. Kickhoefer, 1998, supra). Tankyrase is associated with the telo- merase complex that is involved in aging by acting to regulate telomere length and
NPARP is a component of a multisubunit complex referred to as a "Vault" (F. d'Adda di Fagagna et al., Nat Genet (1999) 23(l):76-80; L.B. Kong et al., Structure Fold Des (1999) 7(4):371-79; E. Pennisi, Science (1999) 282:1395-96). The name "Vault" is based on its observed structure by scanning electron microscopy (L.B. Kong et al., supra). The cellular location of NPARP is mostly cytoplasmic, however, there is a small fraction associated with the mitotic spindle (N.A. Kickhoefer, 1998, supra). Unlike PARPl, tankyrase and NPARP are not activated by DΝA damage (S. Smith et al., 1998, supra; V.A. Kickhoefer, 1998, supra). Tankyrase modifies the telomere binding protein TRF1 in vitro (M.K. Jacobson et al., supra). TRF1 stabilizes the ends of chromosomes, and it has been proposed that modification of TRF1 with ADP-ribose polymers serves to regulate its ability to form a loop structure at chromosome ends (M.K. Jacobson et al, supra; E. Pennisi, supra). In other studies, t-mkyrase has been shown to promote telomere elongation in human cells (S. Smith et al., Curr Biol (2000) 10:1299-302). The primary substrate of VPA P is the major vault protein, MVP (it is also capable of automodifica- tion); these complexes are upregulated in multidrug resistant cancer cell lines (V.A.
Kiclchoefer et al, J Biol Chem (1998) 273(15): 8971-74; A.B. Schroeijers et al., Cancer
Res (2000) 60(4): 1104-10). The various cellular locations and domain structures of the PARP family members strongly suggest that they have distinct cellular roles. Identification of selective inhibitors might help elucidate the function of these enzymes. Poly- (ADP-ribose) polymers can be removed by poly(ADP-ribose) glycohydrolase or PARG; a member of a large family of related enzymes (W. Lin et al., J Biol Chem (1997)
272(18):! 1895-901; M.K. Jacobson et al, supra; J.C. Ame and E . Jacobson et al, Mol Cell Biochem (1999) 193(l-2):75-81; G. Pacheco-Rodriguez et al., Mol Cell Biochem (1999) 193(1-2):13-18; T. Shimokawa et al., J Biochem (Tokyo) (1999) 126(4):748-55). This enzyme is thought to regulate the cellular function of PARP family members by removing ADP-ribose units, which results in changes in the branching pattern of the polymers (Ame & Jacobson et al., supra; M.K. Jacobson et al., supra). There is some evidence to support the hypothesis that polymers synthesized by different PARP ortho- logs might be hydrolyzed by specific PARGs. Although a complete understanding of the physiological activities of PARPs remains unclear, inhibitors of the activity of PARPl and related proteins could provide new therapeutic approaches to both cancer and ischemia caused by reperfusion injury and inflammatory processes (C.a.D. Szabo, supra).
Kaiser et al. found that the constitutive expression of PARPl in Saccharomyces cerevisiae is only possible with simultaneous inhibition of ADP-ribosylation activity through the addition of the known inhibitor 3-methoxybenzamide to the growth media (P. Kaiser et al, Mol Gen Genet (1992) 232(2):231-39). Induction of fully active PARPl under the conditional galactose promoter led to growth arrest (Kaiser et al., supra; M.A. Collinge et al., Mol Gen Genet (1994) 245(6):686-93). The growth inhibition was relieved by removal of the N-terminal region of PARPl that contains the DNA binding domain. Yeast does not possess endogenous PARPl activity so the cause of the growth arrest is unknown. Antibodies raised against polymers of poly (ADP-ribose), were used to determine that a protein of approximately 116,000 D, the size of PARPl itself, was ADP ribosylated in the yeast cells expressing PARPl (Kaiser et al., supra). This lead to the proposal that growth inhibition in yeast might occur through sequestration of chromosomal proteins by the ADP-ribose decorated PARPl protein rather than toxicity due to direct modification of a yeast protein by PARP.
Summary of the Invention
We have now demonstrated that PARP2, a protein closely related to PARPl, also causes growth inhibition when conditionally expressed in yeast. The reduced growth phenotype was used as an assay to screen for novel inhibitors of these proteins by select- ing those compounds that can restore growth to yeast expressing PARPl or PARP2. The inhibitors identified in our screen also inhibit recombinant PARPl activity in vitro. A few inhibitors show selectivity for PARPl or PARP2. These compounds are the first that specifically recognize PARP orthologs in cells. Our data demonstrate the utility of yeast as a screening system to identify inhibitors of human genes and gene families, even when the protein is not highly conserved with any yeast proteins.
One aspect of the invention is a compound of formula 1:
where Ai is C(-E
t) or N; A
2 is C(R
5) or S;
Ri is H, lower alkyl, halo, or a carbonyl; R2 is H, lower alkyl, acyl, or forms a double bond with an adjacent ring atom;
R3 is H, lower alkyl, halo, aryl, aralkyl, acyl, lower alkenyl, or a radical of the form -(CH2)nC(O)-Ra, where Ra is lower alkyl, OH, NH2, lower alkoxy, lower alkylamino, di(lower alkyl)amino, aryl, or heterocyclyl, or forms a double bond with an adjacent ring atom; i is H, lower alkyl, or forms a double bond with an adjacent ring atom (preferably, R3 and R4 do not simultaneously form double bonds with adjacent ring atoms);
R5 is H, lower alkyl, OH, halo, lower alkoxy, lower alkyl-thio, aryl-thio, or heterocyclyl-thio;
R6 and R7 are each independently H, lower alkyl, OH, lower alkoxy, halo, nitro, amino, thio, acyl, lower alkylamino, acyloxy, acylamino, sulfmyl, sulfonyl, alkylsulfinyl, alkylsulfonyl, arylsulfonyl, aryl, heterocyclyl, aralkyl, or heterocyclyl-alkyl.
Another aspect of the invention is a compound of formula 2,
wherein R
10 is H, lower alkyl, lower alkenyl, aryl, heterocyclyl, aryl-lower alkyl, or heterocyclyl-lower alkyl; and Rπ, Rι
2, and R
13 are each independently halo, nitro, OH, NH
2, or lower alkyl, and pharmaceutically acceptable salts thereof.
Another aspect of the invention is a compound of formula 3:
wherein Rι
0, R
6 and R
7 are as defined above, and pharmaceutically acceptable salts thereof. Another aspect of the invention is a method for inhibiting the activity of a PARP enzyme, comprising contacting said enzyme with an effective amount of a compound of formula 1, 2, or 3.
Another aspect of the invention is a composition for treating a disorder modulated by a PARP enzyme, comprising an effective amount of a compound of formula 1, 2, or 3, and a pharmaceutically acceptable carrier.
Another aspect of the invention is an assay for compounds that modulate PARP activity, comprising a eukaryotic cell that expresses a heterologous PARP and exhibits a detectable phenotype as a result of heterologous PARP expression.
Another aspect of the invention is a method for identifying compounds that mod- ulate PARP activity, comprising providing a eukaryotic cell that expresses a heterologous PARP enzyme and exhibits a detectable phenotype as a result of said PARP expression, contacting the cell with a candidate PARP modulating compound, and determining whether the compound reverses the detectable phenotype.
Brief Description of the Figures
FIG. 1 is a diagram illustrating similarities between PARP enzymes and related proteins, by domain. The white bars represent the coding region of the genes and their length reflects the relative difference in size of the gene products. The family members each contain domains that are marked by square boxes filled in the following manner: diagonal lines (DNA binding domains), cross hatches (nuclear localization sequence), vertical lines (automodification domain) and black filled (domains not present in PARPl).
FIG. 2 is a graph depicting the results of an experiment in which S. cerevisiae expressing either wildtype PARPl or PARP2 (large or small diagonal lines) or a cat- alytically inactive mutant (solid black or gray) was grown in 2% glucose (uninduced) or 2% galactose media (induced).
FIG. 3 is a graph depicting the results of an experiment in which yeast cells carrying integrated copies of PARPl or PARP2 under control of the inducible GAL1 promoter (large and small diagonal lines, respectively) were grown in glucose (labeled repressed) and 2% galactose (labeled induced). Their growth was analyzed in the presence of epi- somally expressed PARG also under control of the GAL1 promoter (diagonal, solid black or horizontal lines), or a vector control (solid gray).
FIG. 4 is a graph depicting the results of an experiment in which cells expressing PARPl or PARP2 were exposed to varying concentrations of the inhibitor 6(5H)-phen- anthridinone the percentage of growth restoration is analyzed. The EC50 is 10.2 μM for PARPl and 36.3 μM for PARP2 expressing cells.
FIG. 5 is a graph depicting the results of an experiment in which normal yeast strains or strains carrying a deletion of the two efflux pumps PDR5 and SNQ2, were exposed to increasing concentration of the known inhibitor 6(5H)-phenanthridinone and two analogs, ICX56242099 and ICX56209576. The wild type strain is marked with solid circles while the pdr5Ωsnq2D strain is marked by solid squares.
FIG. 6 is a graph of the dose response curve for two hits identified by screening which showed selectivity for either PARPl or PARP2. The EC50 value for yeast expres- sing PARPl by extrapolation is approximately 60 μM. The PARPl hits were tested on
yeast expressing PARP2 and ICX56258231 showed a higher activity against PARP2 than PARPl, while ICX56259537 showed higher efficacy for PARPl.
FIG. 7 is a graph of the data obtained from an experiment in which recombinant PARPl was assayed by determining the level of incorporation of radiolabelled NAD in a TCA-precipitable polymer composed of ADP-ribose units. All the compounds identified by screening the yeast cell-based assay inhibited the activity of PARPl in vitro. An inactive analog of 6(5H)-phenanthridinone did not inhibit the activity of PARPl. Phen- anthridinone (solid circles), ICX56225328 (plus signs), ICX56304405 (open rhomboids), ICX56290675 (crosses), and ICX56258231 (solid squares). FIG. 8 is a graph of the data obtained from an experiment in which PARPl and
PARP2 activity from yeast cell extracts was assayed by determining the level of incorporation of radiolabelled NAD in a TCA-precipitable polymer composed of ADP-ribose units. No significant PARPl or PARP2 activity was detected in glucose medium (PARP expression repressed), but when cells were grown in galactose medium (expression of PARPl and PARP2 induced) activity was observed.
FIG. 9 depicts data evidencing that inhibitors of PARPl and PARP2 identified in screening that inhibit PARPl activity in vitro also inhibited the activity of PARPl and PARP2 in yeast cell extracts. An inactive analog of 6(5H)-phenanthridinone did not inhibit the activity of either PARPl or PARP2. Solid bars represent PARPl activity and gray bars represent PARP2 activity. The names of the different compounds tested are displayed on the x-axis of the histogram.
Detailed Description Definitions: "Compound of formula 1" refers to a compound having the structure:
wherein A
\ is C(R- or N; A
2 is C(R
5) or S; R
*. is H, lower alkyl, halo, or a carbonyl; R is H, lower alkyl, acyl, or forms a double bond with an adjacent ring atom;
R
3 is H, lower alkyl, halo, aryl, aralkyl, acyl, lower alkenyl, or a radical of the form - (CH
2)
nC(O)-R
a, where R
a is lower alkyl, OH, NH
2, lower alkoxy, lower alkylamino, di(lower alkyl)amino, aryl, or heterocyclyl, or forms a double bond with an adjacent ring atom; R
4 is H, lower alkyl, or forms a double bond with an adjacent ring atom; R
5 is H, lower alkyl, OH, halo, lower alkoxy, lower alkyl-thio, aryl-tliio, or heterocyclyl-thio; R
6 and R
7 are each independently H, lower alkyl, OH, lower alkoxy, halo, nitro, amino, thio, acyl, lower alkylamino, acyloxy, acylamino, sulfmyl, sulfonyl, alkylsulfinyl, alkylsulf- onyl, arylsulfonyl, aryl, heterocyclyl, aralkyl, or heterocyclyl-alkyl.
"Compound of formula 2" refers to a compound having the structure:
wherein R
10 is H, lower alkyl, lower alkenyl, aryl, heterocyclyl, aryl-lower alkyl, or heterocyclyl-lower alkyl; and Rπ, R
1 , and R
13 are each independently halo, nitro, OH, NH
2, or lower alkyl, and pharmaceutically acceptable salts thereof.
"Compound of formula 3" refers to a compound having the structure:
wherein R
10, R
6 and R
7 are as defined above, and pharmaceutically acceptable salts thereof.
The term "lower alkyl" refers to radicals containing carbon and hydrogen, without unsaturation, having from one to six carbon atoms, inclusive. Lower alkyl radicals can be straight or branched. Exemplary lower alkyl radicals include, without limitation, methyl, ethyl, propyl, isopropyl, hexyl, t-butyl, and the like. The term "lower alkenyl" refers to a hydrocarbon radical having 2-6 carbon atoms, and at least one double bond. Exemplary lower alkenyl radicals include, without limitation, vinyl, propenyl, butenyl, and the like. The term "lower alkoxy" refers to a radical of the form RO-, where R is
lower alkyl. Similarly, "lower alkylamino" refers to a radical of the form RNH-, and "di(lower alkyl)amino" and "dialkylamino" refer to radicals of the form RxRyN- where Rx and Ry are each independently lower alkyl. "Lower alkylthio" refers to a radical of the form RS-, alkylsulfinyl refers to a radical of the form RS(O)-, and alkylsulfonyl refers to a radical of the form RS(O2)-.
The term "aryl" refers to an aromatic carbocyclic or heterocyclic moiety, having one, two, or three rings. Exemplary aryl radicals include, without limitation, phenyl, naphthyl, pyridyl, pyrimidyl, triazyl, quinazolinyl, pyranyl, thiazolyl, and the like. The terms "aralkyl" and "aryl-lower alkyl" refer to an aryl moiety joined to a lower alkyl moiety, for example benzyl, phenethyl, 2-phenylpropyl, naphthylmethyl, and the like.
The term "heterocyclyl" refers to a cyclic organic radical containing one or more heteroatoms, such as oxygen, nitrogen, sulfur, or phosphorus. Heterocyclyl radicals within the scope of the invention can be aromatic or non-aromatic, and can have one, two, or three rings. Suitable heterocyclyl radicals include, without limitation, furyl, tefrahydrofuryl, pyranyl, tefrahydropyranyl, pyrrolyl, imidazolyl, pyrimidyl, piperidyl, pyridyl, pyridyl N-oxide, thiofuryl, oxazolyl, thiazolyl, oxazipinyl, indolyl, benzofuranyl, benzimidazolyl, quinolyl, carbazolyl, phenothiazinyl, and the like. The term "halo" refers to fluoro, chloro, bromo, and iodo. The term "leaving group" refers to a radical that is easily displaced by a nucleo- phile in an SN2 displacement reaction. Exemplary leaving groups include, without limitation, sulfonates such as tosylate and mesylate, silanes such as t-butyl-dimethylsilane, halogens such as bromo and chloro, and the like.
The term "pharmaceutically acceptable salts and esters" refers to derivatives of compounds of formula 1 obtained by addition of an acid or base to the compound, or con- densation with an alcohol or carboxylic acid to form an ester. In either case, the acid, base, alcohol, or carboxylic acid must not be unacceptably toxic at the concentrations at which the compound is administered. Suitable acids include, without limitation, inorganic acids such as hydrochloric acid, hydrobromic acid, sulfuric acid, nitric acid or phosphoric acid; organic acids such as acetic acid, propionic acid, glycolic acid, pyruvic acid, malonic acid, succinic acid, malic acid, maleic acid, fumaric acid, tartaric acid,
citric acid, benzoic acid, cimiamic acid, mandelic acid, methanesulfonic acid, ethanesul- fonic acid, p-toluenesulfonic acid and the like.
The term "modulate" as used herein refers to an alteration in PARP activity, and includes both increases and decreases in activity. Modulation of activity can occur as the result of direct interaction of a compound with PARP, interaction with another compound or protein that affects PARP activity directly or indirectly, or by altering the expression of PARP or of a protein that interacts with PARP directly or indirectly.
The terms "PARP" and "PARP enzyme" refer to PARPl and PARP2, and homo- logs that are related to either PARPl or PARP2 to at least the same degree as PARPl and PARP2 are related to each other, derived from any mammalian species.
The tenn "active amount" refers to a quantity of compound of the invention necessary in order to create an observable change in a system. The observable change will preferably be a difference of at least 5%, more preferably at least 10%, of a measurable parameter, whether positive or negative. For example, in a screening assay that measures survival of a host cell, an active amount of compound would increase the survival rate of host cells by at least 5% over controls.
The term "effective amount" refers to the quantity of a compound of the invention necessary to produce a therapeutic effect. A therapeutic effect can be prophylaxis (prevention of symptoms from occurring in a subject at risk for developing such symptoms), treatment (amelioration or reduction of symptoms and/or the underlying cause of disease), or potentiating (increasing the susceptibility of a target cell to other agents). The precise quantity of the effective amount depends on several factors, including the condition to be treated, the condition and size of the subject, and co-therapeutic agents, and the like, but can generally be determined by one of ordinary skill in the art by routine methods.
General Method:
Compounds of the invention are prepared by standard synthetic methods. The selection of synthetic methodology depends on the particular compound selected, and is within the skill of the average practitioner in the art. In general, compounds of the invention are generated by ring-closing reactions of suitably substituted benzene derivatives.
For example, l-((pyridine N-oxide)thio)carbonyl-2-(benzylcarbonyl)benzene is heated with hydrazine in ethanol under acid catalysis to provide l-benzyl-4-(N-oxide pyridyl- thio)phthalazine (R\ = benzyl, A\ = N, A2 = C(R5), R5 = N-oxide pyridylthio). Similarly, reactants such as l-(3-amino-l,3-dioxopropylthio)-2-(chlorocarboxy)-benzene can be
cyclized under suitable conditions to yield compounds such as compound ICX56259537: Other compounds of the invention are prepared by similar methods.
Compounds of the invention can be tested for biological activity using standard methodology, such as, for example, direct biochemical assays measuring inhibition of PARP directly, cellular assays measuring inhibition of PARP activity, functional assays measuring the reversal of a phenotype caused by PARP over-expression or under- expression, and the like. A suitable assay is described in the examples below.
Compounds of the invention can be administered to a subject, or can be applied directly to cells, for example in a cell culture. If administered to a cell culture, the compound is preferably first suspended or dissolved in a suitable carrier. Suitable carriers include, without limitation, water, saline solution, dimethylsulfoxide (DMSO) and solutions thereof, cell culture media, and the like.
Useful pharmaceutical carriers for the preparation of the pharmaceutical compositions hereof can be solids or liquids. Thus, the compositions can take the form of tablets, pills, capsules, powders, sustained release formulations, solutions, suspensions, elixirs, aerosols, and the like. Carriers can be selected from the various oils, including those of petroleum, animal, vegetable or synthetic origin, for example, peanut oil, soybean oil, mineral oil, sesame oil, and the like. Water, saline, aqueous dextrose, and glycols are preferred liquid carriers, particularly for injectable solutions. Suitable pharmaceutical excip- ients include starch, cellulose, talc, glucose, lactose, sucrose, gelatin, malt, rice, flour, chalk, silica gel, magnesium stearate, sodium stearate, glycerol monostearate, sodium chloride, dried skim milk, glycerol, propylene glycol, water, ethanol, and the like. Other
suitable pharmaceutical carriers and their formulations are described in "Remington's Pharmaceutical Sciences" by E. W. Martin.
A compound of formula 1 or a pharmaceutical composition containing same is administered via any of the usual and acceptable methods known in the art, either singly or in combination with another compound or compounds of the present invention or other pharmaceutical agents. These compounds or compositions can thus be administered orally, systemically (e.g., tra sdermally, intranasally or by suppository) or parenterally (e.g., intramuscularly, subcutaneously and intravenously), and can be administered either in the form of solid or liquid dosages including tablets, solutions, suspensions, aerosols, and the like, as discussed in more detail above. It is preferred to administer compounds of formula 1 orally. The formulation can be' administered in a single unit dosage form for continuous treatment or in a single unit dosage form ad libitum when relief of symptoms is specifically required.
PARP enzymes assist in the repair and maintenance of DNA: thus, appropriate inhibition of PARP can serve as cytotoxins, and as sensitizing agents that render target cells more susceptible to DNA-damaging therapies (such as radiotherapy and chemo- therapeutics). Disorders caused by overexpression of PARP can be treated directly with PARP inhibitors.
Examples
The following examples are provided as a guide for the practitioner of ordinary skill in the art. Nothing in the examples is intended to limit the claimed invention. Unless otherwise specified, all reagents are used in accordance with the manufacturer's recommendations, and all reactions are performed at standard temperature and pressure.
Example 1 (Materials and Methods) (A) Yeast strains, media and methods. Isogenic derivatives of the W303 S. cerevisiae background were used in this study (B.J. Thomas et al., Cell (1989) 56(4):619- -30; B.J. Thomas et al., Genetics (1989) 123(4):725-38) and listed below. Yeast strains lacking the major efflux pumps PDR5 and SNQ2 were constructed from a strain YM4 (a
gift of T. Lila, Microcide Pharmaceuticals) deleted for both via a two-step gene disruption process (T.L. Orr- Weaver et al., Meth Enzymol (1983) 101:228-45). ThepdrSU allele was deleted at base pair 101 through 4149 of the PDR5 open reading frame and snq2U allele was deleted for base pair 235 through 4105 of SNQ2 resulting in strain YPB63. To construct strain EIS20-2B, YPB63 was backcrossed to W303-lα and retention of the pdr 50 snq2U alleles was confirmed by whole cell PCR and increased sensitivity to cyclohexamide; a toxic pump substrate. YAN100 is an EIS20-2B derivative that contains a complete deletion of the HIS3 gene. The yeast media protocols and genetic molecular biology techniques used for these studies are standard protocols (F. Sherman, Meth Enzvmol (1991) 194:3-21).
The genotype of the strains is as follows: EIS20-2B and YPB63: MATa ade2-l his3-1115 leu2-3,112 trpl-1, ura3-l, canl-100, pdr5U, snq2U; YAN100: MATa ade2-l Ms3δ::KanMx, leu2-3,112 trpl-1, ura3-l, canrl-100, pdr5U, snq2U; W303α (= W303- 1BJ: MATa ade2-l his3-ll,15 leu2-3,112 trpl-1, ura3-l, canl-100, pdr5D, snq2U; YM4: MATa ade2-101 his3ό200 leu2-3,112 trpl-1 ura3-52.
(B) Cloning and analysis of PARPl. PARP2 and PARG. The cDNAs corresponding to the complete open reading frames of PARPl and PARP2 were PCR amplified from pooled total cDNAs initially synthesized from placental, fetal brain and fetal liver poly(A)+ mRNAs (Clontech Laboratories, Inc.). First and second strand cDNA synthesis was performed using Superscript II reverse transcriptase (Life Technologies, Inc.) as previously described (EX. Perkins et al., Proc Natl Acad Sci USA (1999) 96(5):2204-09). The 5' and 3' oligonucleotides for PARPl PCR amplification were, respectively, YS5PRP (5 ' GTTAATATACCTCTATACTTT-AACGTCAAGGAGAAAAAACGGGAGGATGG- CGGAGTCTTCGGATAAG) and YS3PRP (5 ' TGAATGTAAGCGTGACATAACTAATTACA- TGATGCGGCCCTCCTCTCCCAATTACCACAGGGAGGTC) and for PARP2 were
5YSADP21ac (5 ' GTTAATATACCTCTATACTTTAACGTCAAGGAGAAAAAACGGAAT- TGTGAGCGGATAACAATGGCTCCAAAGCCGAAGCCCTGGGTAC) and 3YSADP2 (5 ' T- GAATGTAAGCGTGACATAACTAATTACATGATGCGGCCCTCGGGCACTCAGAGGTGGAC- CTCCAGC). The oligonucleotide 5YSADP21ac contains the lacO operator site. PCR amplification was carried out using either Bio-X-ACT (Bioline USA Inc., Kenilworth, NJ) or pfuTurbo (Stratagene Inc., La Jolla, CA) high fidelity thermostable DNA
polymerases according to the manufacturer instructions. All oligonucleotides were designed to amplify their target cDNAs and carry approximately 40 base pairs of homology at their 5' ends with the yeast expression vector pYES2. The pYES2 vector (Invitrogen Inc., Carlsbad, CA) contains the yeast GALl promoter, CYC1 transcription terminator, 2μ replication origin and the URA3 gene.
After PCR, the amplified target cDNAs were agarose gel purified. Approximately 50 ng of target cDNA was cotransformed with approximately 200 ng of EcoRI digested pYΕS2 into W303-lα and transformants were selected for uracil prototrophy. Homologous recombination of the target cDNAs into pYES2 was confirmed by yeast whole cell PCR. After the synthetic lethal phenotype of the expressed PARPl and
PARP2 was verified, plasmids from at least three independent transformants were subsequently rescued into the E. coli strain, XL10-GOLD (Stratagene Inc., La Jolla, CA), via electroporation, and further characterized by restriction enzyme analysis and DNA sequencing of the 5' and 3' cloning junctions. These studies verified the presence of the desired structures.
During the course of this work it was noted that the pYES2-PARP2 plasmid in E. coli was toxic (data not shown), thus the primer 5 YSADP21ac incorporated a lacO repres- sor binding site; this established the plasmid during its propagation in XL10-GOLD (which carries the laclq gene). This vector construction still exhibited instability during growth in yeast. Thus the GALl-lac-PARP2-CYCl terminator cassette was recombina- tionally cloned into the yeast URA3 2μ vector YEP24 (D. Botstein et al, Gene (1979) 8(1): 17-24) using the oligonucleotides Y24smacyct (5 ' TCACAAATTAGAGCTTCAAT- TTAATTATATCAGTTATTACCCGGGGGCGCGCCGCAAATTAAAGCCTTCGAGC) and Y24pvu2gall (5 ' GGGCGAGCCGCCGAAGATTAGGCAAATTTGGTCGACGGAGCGCGC- AAAGCCACTACTGCCACTTTTGGAGACTGT) effectively putting the cassettes between the Smαl and PvuIJ sites of YEP24, and eliminating the tetracycline resistance gene.
For compound screening and subsequent genetic testing, the PARPl and PARP2 expression cassettes were integrated into the indicated strains using the pARC series of dual episomal/integrative plasmid constructs. A brief summary will be presented in lieu of a detailed description of the steps required to construct the pARC series of vectors. The vectors are derived from the pRS series of vectors (R.S. Sikorski et al., Genetics
(1989) 122(1): 19-27; T.W. Christianson et al, Gene (1992) 110(1): 119-22). The basic elements of the pARC plasmids include a 713 base pair Sphl/BamEI fragment containing the S. cerevisiae GAL1 promoter and a 243 base pair BgUllHindlll fragment containing the GAL4 terminator region. This region surrounds a 45 base pair polylinker that con- tains unique sites for -P-ytl, Sail, Spel, Xhol and --4vαl and replaces the pRS polylinker: the base vector and polylinker of the pRS vectors is pBLUESCRIPT II. Both CEN and 2μ based vectors were generated with 44 base pairs of the 5' end of the LYS2 gene (base pairs 8 -52 of the LYS2 open reading frame) and 43 base pairs of the 3' end of the LYS2 gene (base pairs 4133-4176 of the LYS2 ORF) flanking the CEN-ARSH4 element or the 2μ replication origin. In addition, Sfil sites flank the CEN-ARSH4 or 2μ elements such that digesting with this restriction enzyme liberates the elements and allows the subsequent integration into the endogenous LYS2 locus via a "gamma" deletion mechanism (R.S. Sikorski et al, supra). For this work, integrants were selected for alpha-amino- adipate resistance and by marker prototrophy, e.g. Leu+ or Ura+, integration was con- firmed by PCR.
For integration of PARPl, a Spell Mlul (the Mlul site made blunt with Klenow) GALl-PARPl-CYCl terminator fragment from pYES2-PARPl was ligated into aSpel/EcoPl cut pARC35A, a CΕN-ARSH4 LEU2 vector containing the original pBluescript II polylinker, these steps resulted in plasmid pARC35APARPl. Digestion with Sfil allowed integration of the PARPl expression cassette into the LYS2 locus. For integration of PARP2, the GAL1-PARP2-CYC1 terminator fragment from pYES2- PARP2 was amplified by PCR using Pfu Turbo with primers YSPRC (5 ' GATGTATAA- ATGAAAGAAATTGAGATGGTGCACGATGCACAGTTGTGAATGTAAGCGTGACATAACTA- ATTAC; a primer containing homology at the 5' end to the GAL4 terminator) and GLUAS 1 (5 ' TGAAGTACGGATTAGAAGCCGCCG; a primer with homology to the
GALl UAS element). This fragment was co-transformed into yeast with. 7zoI digested pARC25B (2μ EEL/--? GALl promoter and GAL4 terminator). As before, the recombin- ants were verified by phenotypic analysis and PCR prior to rescue into Ε. coli. The resulting pARC25BPARP2 plasmid was digested with Sfil and integrated into the LYS2 locus.
Phenotypic expression of the PARPl and PARP2 clones was confirmed by constructing GFP fusions to the carboxy-terminal end of the proteins using GFP-kanMX cassettes (A.A. Wach et al, Yeast (1994) 10(13): 1793-808; A.A. Wach et al., Yeast (1997) 13(l l):1065-75; M.S. Longtine et al., Yeast (1998) 14(10):953-61). Expression via fluorescence microscopy and Western Blot analysis was assessed. All active site mutations were confirmed by sequencing.
PARG was isolated and cloned using primers 5YSPARG (5 ' GTTAATATACCT- CTATACTTTAACGTCAAGGAGAAAAAACATGAATGCGGGCCCCGGCTGTGAACC) and 3YSPARG (5 ' TGAATGTAAGCGTGACATAACTAATTACATGATGCGGCCCTCTCAGG- TCCCTGTCCTTTGCCCTGAATGGTC) were used to amplify the complete open reading frame from a testes cDNA library (EX. Perkins et al., supra). The purified PCR product was co-transformed with digested pYES2 plasmid into EIS20-2B and the resulting recombinants were confirmed using whole cell PCR. Transformed isolates were subsequently reisolated and transformed into E. coli XL10-GOLD and confirmed by restriction enzyme analysis resulting in plasmid pYES2-PARG. To assess whether expressed PARG could suppress the synthetic lethality caused by PARPl and PARP2, plasmid pYES2-PARG was transformed into EIS20-2B cells containing an integrated chromosomal copy of PARPl or PARP2. Purified transformants were subsequently inoculated into galactose containing media and their growth assayed. (C) PARP Activity Assays. A PARP Activity Assay Kit (Trevigen, Inc.,
Gaithersburg, MD) was used to measure the functional activity of human PARPl and PARP2. Human PARPl supplied in the Activity Assay Kit from Trevigen, Inc., or crude cell extracts of yeast expressing a cDNA encoding the full-length human PARPl or PARP2 gene were used in the assay. The activity is measured by determining the level of incorporation of radiolabelled NAD in a TCA precipitable polymer composed of ADP- ribose units.
Crude cell lysates were prepared as described by M.A. Collinge et al., supra. Protein concentrations were determined using the Bio-Rad Protein Assay Kit (Bio-Rad Inc., Hercules, CA). Quantitative values for PARP activity were determined by scintillation counting of the acid-insoluble counts using 20 μg total protein from crude cell lysates in a reaction containing 1 mM NAD, 1 mg/ml histones, 2 mCi 32P-NAD, and 10 μg sheared
salmon testes DNA. The timed enzymatic reaction was initiated by the addition of the cell lysates, incubated at room temperature, and stopped by the addition of 20% TCA to precipitate ribosylated proteins. The protein precipitate was suspended in liquid scintillation fluid and analyzed using a scintillation counter. Each reaction was done in triplicate and the results expressed as the mean counts per minute.
(D) Compound screening of PARPl and PARP2 inhibitors. Strain YPB63 containing integrated PARPl was screened against a pilot library of 16,000 small molecular weight organic compounds as well as 22 compounds chosen based on computed 3D structural similarity to the known inhibitor 6(5H)-phenanthridinone. The 22 compu- tationally selected compounds were chosen from the Iconix corporate library based on morphological similarity. The algorithms defining this methodology have been previously described and they form the basis of a rapid chemical structural similarity search method (A.N. Jain, J Comput Aided Mol Des (2000) 14(2): 199-213). The pilot library represents a small but diverse collection derived from multiple commercial sources. All screening was done in duplicate at initial compound concentrations of 10 μg/ml. Cells were grown overnight in synthetic media with 2% glucose (repressed) to late logarithmic/early stationary phase. The next day cells were washed once with synthetic media without a carbon source and diluted to a final OD60o of 0.04 in synthetic media containing 2% galactose (induced). The diluted cells (90 μl) were added immediately to 96-well plates containing the test compound. The final volume in each well was 100 μl, and contained DMSO at a final concentration of 1%. As a control, cells containing vector (YPB63 with integrated vector plasmid) were similarly grown, washed and diluted to the same OD and then inoculated.
The plates were incubated at 30°C for 40-42 l rs, and the OD600 was read with a microtiter plate reader (Molecular Device, Menlo Park, CA) after shaking. The effect of compounds was measured as percent of growth restoration using the following equation: Percent Growth Restoration = (TEST - MEDarc)/(MEDvec - MEDarc) x 100, where TEST is the OD60o of the well with test compound, MEDarc is the median value of OD60o of the cells without compound, and MEDvec is the median value of OD600 of vector-con- taining cells. Compounds showing D 10% of growth restoration in duplicate tests were scored as hits.
(E) Hit confirmation: Hit compounds identified from the primary screening were confirmed by generating a dose response curve using YAN100 cells expressing either PARPl or PARP2. For the confirmation test the compounds were solubilized from powder and serially diluted (usually from 128 μM to 0.125 μM). The compounds were tested against a vector control strain (YAN100 carrying the integrated vector) and other isogenic strains expressing different cDNAs that also elicit synthetic lethal phenotypes but which are not related to PARPl .
Example 2 (Results)
(A) Growth inhibition of yeast cells expressing the human genes PARPl or PARP2 is relieved by mutations in the active site. As shown in Fig. 2, yeast harboring an expression vector carrying the cDNA for wild-type PARPl or PARP2 grew in 2% glucose (uninduced), but PARPl or PARP2 expression caused growth inhibition when the cells were grown in 2% galactose (induced). To ascertain whether the inhibitory effect of PARPl and PARP2 expression on yeast cell growth depends on their catalytic activity, a single mutation was generated in the conserved active site of the two enzymes. For PARPl, a glutamic acid was changed to an alaniiie at residue 988 (E988A) and the analogous change was engineered at residue 509 (E509A) in PARP2. Yeast expressing the mutant constructs grew equally well in 2% glucose or 2% galactose. Thus, the growth inhibition observed in yeast requires the catalytically active form of PARPl or PARP2. Fluorescence microscopy revealed that expressed mammalian PARPl and PARP2 accumulated in the nucleus.
Yeast cell growth was severely inhibited whether PARPl and PARP2 were expressed episomally, or from a cliromosomal locus (see Fig. 2 and Fig. 3). In our initial characterization of the growth phenotype caused by PARPl or PARP2, the proteins were episomally expressed, whereas yeast carrying a chromosomal copy of PARPl or PARP2 were used for compound screening.
(B) Expression of the human Poly (ADP-ribose) glycohydrolase (PARG) in yeast reverses the growth inhibition caused by PARPl or PARP2 expression. To determine whether PARG reverses the growth defect caused by human PARPl or PARP2
expressed in yeast, we conditionally co-expressed PARG from a high copy plasmid in yeast cells carrying an integrated copy of PARPl or PARP2 also under control of the inducible galactose promoter. Yeast cells were grown in glucose medium (uninduced) and then switched to 2% galactose (induced) for 40 hours in synthetic media lacking uracil. As a control, yeast carrying an integrated copy of PARPl or PARP2 and harboring a high copy plasmid with no cDNA was also tested for growth in glucose and galactose media. Fig. 3 shows the growth of yeast in the presence and absence of PARG expression. Yeast expressing PARPl or PARP2 alone only grew at approximately 1% of the normal wild type growth levels. However, when PARG was expressed in the pres- ence of PARPl and PARP2 expression, growth was restored to approximately 70% of wild-type levels. As shown in Fig. 3, PARG expression alone had little effect on the growth rate of yeast (compare growth in yeast with no integrated cDNA grown in glucose and Gal medium). It is likely that PARG expression reverses the deleterious effects of PARP expression by catabolizing the ADP-ribose polymers created by PARP. The inability of PARG expression to completely counteract the effect of PARPl expression may be due to the residual ADP-ribose monomer that is removed in mammalian cells by an ADP-ribosyl protein lyase (D.S. D'Amours et al., supra).
(C) A known inhibitor of PARPl. phenanthridinone, reversed the growth inhibition caused by the expression of PARPl and PARP2. To screen for small molecule inhibitors of PARPl and PARP2, yeast strains expressing these genes from a chromosomal locus were employed. The chromosomal expression of the genes increases the robustness of the screen by not allowing variation of plasmid copy number.
To further validate that the phenotype of growth inhibition observed in the integrated strains was due to the activity of PARPl and PARP2, and that this phenotype would be amenable to compound screening (i.e., reversed by compounds) we determined whether the known chemical inhibitor 6(5H)-phenanthridinone would reverse the growth inhibition (Fig. 4). Yeast carrying a chromosomal copy of PARPl, PARP2 or no human cDNA were exposed to varying concentrations of 6(5H)-phenanthridinone (0 μM to 128 μM, in triplicate). Cell growth was restored when yeast were exposed to increasing concentrations of 6(5H)-phenanthridinone: the EC5o value for PARPl was 10.2 μM and 36.3 μM for PARP2.
(D) Analogs of 6(5H)-phenanthridinone do not discriminate between PARPl and PARP2. The ability of computationally chosen analogs of 6(5H)-phenanthridinone to reverse the growth inhibition caused by PARPl and PARP2 expression was measured. Six of the twenty-two chosen analogs were active and restored growth of yeast strains expressing PARPl (three are shown in Table 1: ICX56316703, ICX56259835, and ICX56274004). Not surprisingly, as in the case of the known PARPl inhibitor compound, 6(5H)-phenanthridinone, the three active analogs that were tested on PARPl and PARP2 restored the growth of yeast to a similar degree.
TABLE 1
(E) Sensitivity of Yeast to analogs of 6(5H)-phenanthridinone is increased in strains lacking the two major efflux pumps, PDR5 and SNQ2. To optimize the sensitivity of the yeast strain to the screening compounds, mutations were made in the two efflux pumps, PDR5 and SNQ2 (M.A. Kolaczkowski et al., Microb Drug Resist (1998) 4(3): 143-58). Pdr5p and Snq2p are the major efflux pumps in S. cerevisiae that confer resistance to several unrelated fungal growth inhibitors. To examine whether yeast strains lacking these two major transporters were more sensitive to the active analogs of 6(5H)- phenanthridinone, the activity of the analogs were evaluated mpdr5U and snq2U strains (YPB63; lacking the two pumps) or (W303; wild type for PDR5 and SNQ2) (see Fig. 5).
While four analogs showed a similar effect regardless of the presence or absence of the two efflux pumps, yeast lacking the efflux pumps were more sensitive to 6(5H)- phenanthridinone, ICX56209576 and ICX56242099 than wild-type yeast (Fig. 5). The
results underscore that the use of yeast strains lacking efflux pumps for screening compounds is beneficial in some cases.
(F) Selectivity is observed among the compounds identified as "hits" in the yeast cell-based screen. To identify new classes of inhibitors, we screened the yeast strain carrying integrated PARPl against a 16,000 member pilot library of small organic "drug like" compounds for their ability to restore growth while PARPl was expressed (in the presence of galactose). Ten hits showed dose dependent growth restoration on cells expressing PARPl (see Fig. 6 and Table 1: the computationally chosen analog ICX56259835 was also identified as a screening hit). We then tested the ability of the confirmed hits to restore growth to yeast expressing PARP2. Unlike the analogs of 6(5H)-phenanthridinone, some of the ten compounds identified by screening showed patterns of selectivity for PARPl or PARP2 (Table 1).
One inhibitor class showed modest potency, ICX56225770, ICX56244215, ICX56280834, ICX56290675, whereas one compound in this group, ICX56290675, showed some selectivity for PARPl. ICX56222404 and ICX56259537 are novel inhibitors with more pronounced selectivity for PARPl. Finally, ICX56258231 showed selectivity for PARP2 (Fig. 6). Although a clear difference in the rate of growth restoration was observed for PARPl versus PARP2 in the presence of ICX56258231, an accurate EC50 value for PARPl could not be determined. Normally this value is derived from a curve fit to a 4-parameter logistic equation, but in this case at the highest concentration of soluble compound, a plateau in the growth restoration curve was not reached.
(G) Compounds that restored growth in the yeast cell-based assay inhibit purified PARPl in vitro. To clarify whether the small molecules identified in screening directly inhibited PARPl activity, the hits were tested for their ability to inhibit human recombinant PARPl in a biochemical assay. This assay measures the incorporation of radiolabelled ribose derived from NAD into PARPl. Baseline activity was established by measuring incorporation from the radiolabelled ribose in the absence of the inhibitors.
Dose response curves were constructed for three of the compound hits (ICX56304405, IC290675, ICX56258231 and included 6(5H)-phenanthridinone and its inactive analog, ICX56 225328; Fig. 7). These results show clearly that the compounds identified in the yeast screen are direct inhibitors of PARPl.
(H) Compounds that restored growth activity also inhibit the activity of PARPl and PARP2 from yeast cell extracts. The activity of PARPl or PARP2 in extracts from cells grown to log phase over 8 hours in 2% glucose (uninduced) was negligible. However, cells grown under similar conditions but with 2% galactose (induced) showed an increase of 20x for PARPl and lOx for PARP2 activity (Fig. 8). The screening compounds were then tested for their ability to inhibit the activity of PARPl and PARP2 observed in the yeast extracts. All the compounds inliibited PARP activity as effectively as the known inhibitor with the exception of the inactive analog, ICX56225328 (Fig. 9).