WO1999037802A1 - Methods to identify myelostimulants - Google Patents

Abstract

Compounds that interfere with the inhibition of JNK by GST isoenzymes especially of the π class are useful as myelostimulants and as sensitizers of tumor cells to treatment under toxic conditions.

Description

- 1 -

METHODS TO IDENTIFY MYELOSTIMULANTS

Technical Field

The invention relates to assays to identify compounds that have the metabolic effects of glutathione analogs interactive with the glutathione S-transferase π class . These effects include modulation of hematopoiesis in bone marrow or blood and chemosensitization of tumors.

Background Art The side effects of chemotherapeutic agents used in the treatment of malignancy and other indications are well known. Among these side effects are alterations in the levels of various blood cells, including neutrophils, platelets and lymphocytes. The results of these effects can be neutropenia, thrombocytopenia and immune suppression generally. These side effects are not only unpleasant, but they also restrict the efficacy of cancer therapy and place the subject at serious risk of infection and uncontrolled bleeding.

At the present time, there appears to be little practical remediation for these effects. Some approaches are merely palliative, such as supportive care. Others have their own side effects, such as large doses of antibiotics. Still others are expensive and invasive such as transfusions. Still another approach, the administration of growth factors, such as granulocyte colony-stimulating factor (GCSF), granulocyte macrophage colony-stimulating factor (GMCSF), and more newly developed factors such as megakaryocyte growth and development factor (MGDF) and thrombopoietin (TPO) are costly and must be administered by injection. They also have their own associated negative side effects.

Clearly there is a need for a simpler approach, for example a small molecule drug, preferably administrable by mouth, that can protect and restore bone marrow and also stimulate the production of neutrophils, platelets and lymphocytes both in conjunction with chemotherapeutic protocols and in response to other factors which result in hematopoietic suppression such as cyclic and idiopathic neutropenias, thrombocytopenia, and the effects of allograft transplants. In addition, in the context - 2 -

of chemotherapy, such side-effects would be mitigated if the dosage of the chemotherapeutic agent could be lowered. By providing chemosensitization selectively to tumor cells, such modification of dosage becomes possible. A single molecule which both chemosensitizes tumor cells and mitigates side-effects by effecting hematopoiesis is possessed of extremely useful characteristics for administration in conjunction with chemotherapeutic agents.

The problems related to current approaches for managing the side effects of chemotherapy and otherwise dealing with suppression of hematopoiesis are solved at least in part by the biological activity of certain simple tripeptide compounds which are inhibitors of the various isoenzymes of glutathione S-transferase.

PCT application WO95/08563 published 30 March 1995, discloses certain tripeptide compounds which are analogs of glutathione, and which show diverse specificities with respect to ability to inhibit glutathione S-transferase isoenzymes. One of these tripeptides, which is γ-glutamyl-S(benzyl) cysteinyl-phenylglycine including its amides, esters and hybrid amide/esters, has been shown to have the desired properties set forth above. The active component is believed to be the free acid, but typically, the esters or amides are administered as they are able to enter the cell; the diester is designated TER199. The stereoisomer wherein the γ-glutamyl and cysteinyl residues are in their native L-form, and phenylglycine is the R(-) isomer, is preferred. Thus, the active form of this compound has the ability to modulate hematopoiesis in bone marrow and in peripheral blood and therefore exert protective effects when chemotherapeutic agents destructive to the hematopoietic system are administered. It also potentiates the desired effects of chemotherapeutic agents. The diacid form of this compound shows inhibition of the π class of glutathione S-transferase (GST) at levels below 0.5 μM in vitro and has more than a 50-fold specificity for π with respect to other GSTs. The dual nature of the effects of TER199 highlights the key role of GST π in the overall cellular response to stress, with the detailed effects being cell type-specific. - 3 -

Disclosure of the Invention

The invention provides methods to identify compounds that have properties similar to those of the active form of TER199 by taking advantage of the mechanism whereby TER199 has its beneficial effects. A specific and particularly relevant ability of the isoenzyme for which TER199 is specific is inhibition of the activity of a major gene regulating factor, jun kinase (JNK). Thus, the invention method identifies compounds which are useful in modulating hematopoiesis generally and as aids to chemotherapeutic treatment of tumors by virtue of their ability to exert a positive effect on the hematopoietic system, and in chemosensitizing tumors. The compounds may be orally active and can be used not only in chemotherapy but in any context where it is desirable to modulate the hematopoietic processes in bone marrow or peripheral blood or to modulate other bone marrow processes.

Thus, in one aspect, the invention is directed to a method to identify a compound that has myelostimulation activity which method comprises measuring the activity of JNK on an appropriate substrate such as jun or fragments thereof (or other appropriate substrate such as ATF2, GADD45, or p53) in the presence of a kinase- inhibiting amount of glutathione-S-transferase especially of the π subclass (GST π) in the absence of a candidate compound; measuring the kinase activity of JNK in the presence of a kinase inhibiting amount of GST in the presence of said candidate compound; comparing the kinase activity in the presence of said compound with the kinase activity in the absence of said compound; whereby a compound whose presence results in an increase in said kinase activity is identified as a candidate myelostimulant.

In other aspects, the invention is directed to a method to identify a compound that will chemosensitize tumor cells, especially those tumor cells which contain high levels of GST π, i.e., wherein this particular isoenzyme is the predominant form of GST isoenzyme present. The method is as set forth above, as compounds that interfere with the normal function of GST, especially GST π, including but not limited to catalytic activity and including interaction of GST with JNK, also have this chemosensitizing effect. - 4 -

Both of the foregoing assays may be conducted extracellularly using the components in an in vitro system. Alternatively, the assay may be conducted in mammalian cells or in yeast using any of a variety of recombinant constructs which permit the JNK/JNK substrate complex and GST to interact intracellularly. These constructs include, for example, the yeast-2-hybrid system with AP-1 (activated by JNK action on jun) driven promoters linked to β-galactosidase. All of the foregoing assays can be conducted by testing the compound to be evaluated in competition with the active form of TER199. If the assay is conducted extracellularly in vitro, the free acid form is preferred. If the assay is performed in cell culture, the diester, TER199, or an amide or ester amide would be required.

The invention is also directed to compounds identified by the described methods, to pharmaceutical compositions thereof, and to methods to effect myelostimulation, including chemoprotection, or chemosensitization using the compounds identified by the invention method.

Brief Description of the Drawings

Figure la shows the effect of TER199 on the survival of tumor cells treated with various concentrations of chlorambucil.

Figure lb shows the toxic effect of TER199 in contrast to its unesterified form on HT4-1 cells.

Figure 2 is a graph showing the effectiveness of i.v. versus i.p. versus oral administration of TER199 on stimulation of GM-CFU measured 24 hours after administration.

Figure 3 shows the time course of TER199 stimulation of GM-CFU administered i.p..

Figure 4 is a graph comparing the effects on GM-CFU suppression in mice of i.p. administration of TER199 at 24 hours after administration of two different doses of 5-FU.

Figure 5 shows the effect of TER199 on differentiation of CD34^*** cells with respect to CFU-GEMM and BFU-E. - 5 -

Figure 6 is a series of graphs which show that TER199 accelerates recovery of and diminishes toxicity to myeloid and lymphoid lineages in rats following treatment with 5-FU.

Figure 7 is a photocopy of a radiogram showing the effect of ultraviolet radiation on phosphorylation of the j un- JNK complex.

Figure 8 shows that increasing the concentration of the jun- JNK complex overcomes inhibition by GST π.

Figure 9 shows the results determined with a radioimaging blot analyzer of fractions eluted from a Superdex 75 column tested for the ability to inhibit JNK phosphorylation activity. Most of the activity is in fractions 16-24.

Figure 10 is a graph of the effect of increasing concentrations of GST π on inhibition of c-jun phosphorylation presented as either GST-jun^5"89"JNK (5 μg/reaction) or _hisc-jun-JNK (7 μg/reaction) prior to addition of γ-labeled ATP. The inserts show the autoradiographs of the respective reactions. The GST on this construct is Schistosome GST and does not have the activity of GST-π.

Figure 11 is a graph showing the inhibition of c-jun phosphorylation can be overcome by immunodepletion of GST-π. Reaction mixtures for JNK activity were treated with whole-cell extracts from nonstressed cells (at μg amounts indicated in panel A) or after immunodepleting GST π from the extracts (panel B and bar "antiGST serum" in the graph). A control reaction with normal serum and protein A/G is shown in panel C (bar "normal serum" in graph)..

Figure 12 shows the inhibitory action of GST-π on JNK activity in mouse fibroblasts with or without treatment with ultraviolet light can be overcome by TER 199. Figure 13 shows the percent inhibition of JNK activity in mouse fibroblasts transfected with varying amounts of an expression system for GST π.

Figure 14 shows, diagrammatically, the preferred assay format of the invention. - 6 -

Modes of Carrying Out the Invention

The methods of the invention to identify compounds having desired hematopoietic activities or the ability to chemosensitize (or radiosensitize) tumor cells is based on the nexus between these properties and the ability of a compound to interfere with the interaction of GST π with the JNK/JNK substrate complex. The Examples below show that GST π, which is inhibited by TER199, is an effective inhibitor, and indeed the natural inhibitor of the kinase activity of JNK. GST π apparently binds to the complex formed by JNK with its substrate. JNK is a participant in a stress-activated signaling pathway resulting ultimately in the activation of transcription factors, including jun. Phosphorylated forms of jun are transcription factors in a class known as AP-1, which are active in the form of dimers. Interference with, or inhibition of, JNK would prevent the stress response pathway from effecting transcription of genes that generate cytokines necessary for hematopoiesis. Among the factors which can activate this signaling pathway are TNFα, UN radiation, and oxidative stress.

In appropriate cells, apoptosis is induced by clustering of fas cell surface receptor; Yang, X., et al, Cell (1997) 89:1067-1076. In other cells, fas stimulation can result in proliferation. Two pathways have been identified which mediate apoptotic effects, one which is sensitive to bcl-2 and one which is not. The former is mediated by a protein called Daxx which regulates JΝK activation. See Yang, X. et at, ibid.

Complex feedback loops are characteristic of this kind of signaling, and the active transcription factor that includes jun, AP-1, stimulates GST π transcription. The inflammatory cytokine IL-1 also increases GST π while decreasing several other GST isoenzymes. Thus, it is consistent with the known activity of TER199 that it interferes with interactions of GST π and the JΝK/JΝK substrate complex, thus permitting the activation of JΝK.

As used herein, "myelostimulation" refers generally to the activation of the proliferation and differentiation of bone marrow cells. As such, it includes hematopoiesis as well as other beneficial results of such stimulation. The protective effect of TER199 with respect to the effects of chemotherapy is consistent with this - 7 -

mode of action. The ability of TER199 to potentiate the effects of chemotherapeutic agents on tumor cells is also consistent with the nexus between JNK and apoptosis.

It has now been demonstrated that GST π specifically interacts with the complex formed by JNK with its substrate and inhibits its phosphorylation of jun. Thus, any compound which interferes with this interaction will, in effect, prevent the inhibition of JNK by this isoenzyme and permit the desired transcription factors to effect the desired hematopoietic or myelopoietic response to stress. The interaction appears particularly strong for GST π, as it is known that Schistosome GST does not effect this inhibition. Other GST isoenzymes such as α and μ are substantially less effective as inhibitors of JNK in vitro and do not appear to associate with JNK in vivo. These results are consistent with the specificity of the active form of TER199 in inhibiting GST π, but not other GSTs. A comparison of the crystal structure of Schistosome GST with that of GST π complexed to the active form of TER199 shows that the binding pocket which provides specificity of the compound for the π family of GSTs is absent in Schistosome GST (as well as from the α and μ families).

The substrate most useful in assay systems for assessing the kinase activity of JNK is either the c-jun protein or a fragment thereof. Commonly used forms are jun extended at the N-terminus with histidine or other tags, such as HA and FLAG, for ease of purification, and particular fragments of jun such as the 5-89 fragment. The kinase activity assay is typically conducted using γ-labeled ATP and measuring the incorporation of labeled phosphate into the complex. The effect of GST π in inhibiting the activity is associated with the ability of GST π to bind the complex rather then either component individually.

Directly testing for interference by a candidate compound with the inhibition of JNK by GST π or with another isoenzyme of GST as appropriate, is the preferred assay for discovery of compounds like TER199, but other straightforward analyses are possible such as assay of binding to GST π directly or in competition with TER199 analogs or the assay of the candidate compound's ability to inhibit the activity of GST π. There is apparently not an exact correspondence between the ability of a compound to inhibit GST π and to inhibit the interaction of this isoenzyme with JNK - 8 -

however. This has been verified by showing that some GST π inhibitors are inactive as myelostimulants, although poor cell permeability may have contributed to this result.

The interaction of GST π with JNK in its complexed form has been demonstrated in several ways.

First, direct protein-to-protein association of the JNK complex with GST π has been demonstrated by their coimmunoprecipitation from 3T3-4A cells using antibodies to either GST π or JNK, or to jun implying a ternary complex exists in unstressed cells. Second, it has been demonstrated that overexpression of GST π decreases the level of JNK activity. Third, administering TER199 has the effect of enhancing JNK activity.

The sequelae in the discovery of the association of GST π with the JNK complex further validates the finding. It was known that an inhibitor of JNK was present in unstressed cells and that it was decreased in cells stressed, e.g., with UN light. Purification of the inhibitor over 7 fractionation steps led to a homogeneous protein; partial amino acid sequencing showed identity with GST π.

The known properties of TER199 in its active form as a myelostimulant and as a chemosensitizer in tumor cells are consistent with the demonstration that TER199 reduces the inhibition of JΝK. JΝK, as set forth above, is a known participant in signaling pathways which results, ultimately, in myelostimulation as well as apoptosis depending on cell type. Activation of JΝK by preventing its association with GST π, thus is consistent with these desired effects.

In the method of the invention, a candidate compound is tested for its ability to inhibit the interaction of GST, especially that of the π subclass, with the JΝK/JΝK substrate complex. In its most straightforward form, the method is conducted by measuring the kinase activity either on junker se or on a different or on an artificial substrate in the presence of a kinase-inhibiting amount of GST, most effectively GST π, in the presence and absence of the compound. If the compound is successful in enhancing the activity of JΝK under these conditions, it can be presumed to interfere with the inhibition of JΝK by GST π and thus to exhibit the effects of the active forms of TER199. The assays can be conducted in vitro using purified forms of - 9 -

GST and JNK along with a suitable substrate or can be conducted in cell culture by measuring the sequelae of JNK activation. These sequelae could include a variety of reporter genes activated naturally or artificially by AP-1. Natural responding genes such as IL-2 can be assayed at mRNA or protein level by well known methods. Alternatively, a reporter such as β-gal or Green Fluorescent Protein can be used. The kinase activity can be measured by assessing the phosphorylation of a suitable substrate. Suitable substrates include, of course, the jun protein per se, as well as, for example, GADD45, p53 or the artificial substrates such as the amino acid 5-89 fragment of jun. The level of phosphorylation can be assessed, for example, by adding to the reaction mixture an antibody or fragment thereof that immunoreacts specifically with the phosphorylated form of the substrate, to the exclusion or effective exclusion of its unphosphorylated form. The presence of the antibody complexed to the phosphorylated substrate can then be detected using conventional methods. For example, the antibody itself may be coupled to a fluorescent moiety, an enzyme, or a radioactive label. Alternatively, a second specific reagent can be used bearing the appropriate label. Such methods for labeling immunoreagents and methods for detecting complex formations between specific binding partners, such as antibodies and their targets are well known in the art. Alternatively, labeled phosphorylating agents such as γ-labeled ATP may be used and the amount of ³²P associated with the substrate may be measured.

A preferred assay is shown in Figure 14 using his-tagged jun to immobilize the jun-JNK-GST π complex on a Ni-chelate 96-well microplate. Incubation with compound and ³²P-γ-ATP results in increasing incorporation of radiolabel as GST π is dissociated. The foregoing assays are most preferably preferred using GST π. However, due to overlap in some properties, it may be possible to use other isoenzymes of the GST class in such assays.

The assays for kinase activity or other assay formats may be performed intracellularly or in extracellular environments. Indeed, the assays can be performed using isolated forms of JNK/JNK substrate complex and isolated GST. Such components can be packaged into kits for convenient performance of these assays. - 10 -

Additional assay formats may also be employed. For example, the affinity of a candidate compound for the JNK/JNK substrate complex may be measured directly; the ability of the candidate compound to form an association with GST, especially GST π, may also be determined directly using known methods of labeling or in competition with TERl 99. If performed as a competition assay, the TERl 99 may be labeled, thus avoiding the necessity for labeling the candidate compound.

Besides those set forth above, additional assay formats can readily be envisioned by those of ordinary skill.

The compounds identified as successful in interfering with the GST π/JNK interaction will be useful in modulating hematopoiesis in bone marrow or peripheral blood and in sensitizing tumor cells carrying a burden of GST π to radiation or chemotherapy. Suitable candidate compounds include peptides derived from JNK or GST, especially GST π. The compounds are formulated according to their nature and administered as described below.

Administration and Use

By "modulating hematopoiesis in bone marrow or peripheral blood" is meant altering the rate of blood cell formation as measured by the capacity to form colonies or differentiated cells. Differentiated cells include neutrophils, platelets, red blood cells, lymphocytes, macrophage, granulocytes, granulocyte-macrophage and the like. It is unclear what the mechanism of this modulation is; the cells themselves may or may not be directly stimulated by the compounds identified by the method of the invention; rather, a change in number and/or size of colonies of differentiated cells may be due to preferential survival, inhibition of apoptosis, or any one of a number of factors. As used in the present application, "modulating hematopoiesis in bone marrow or peripheral blood" refers to the ability of bone marrow or blood treated with the compounds identified by the method of the invention to exhibit colony formation or generation of differentiated cells at a level different from that of untreated bone marrow. Similarly, fractions of bone marrow or peripheral blood which contain suitable progenitors will exhibit this effect. It should be noted, that as used herein, - 11 -

"peripheral blood" specifically includes cord blood. "Myelostimulation" includes at least these phenomena.

In addition to modulating hematopoiesis, the compounds identified by the method of the invention affect bone marrow cells directly and may exert a beneficial effect on bone marrow cells other than those of hematopoietic origin. For example, these compounds may also enhance the formation of osteoblasts so as to aid in bone regeneration. Thus, their beneficial effects on bone marrow are not limited to modulation of hematopoiesis per se; and "myelostimulation" includes these effects. In general, when agents are employed which typically have destructive effects on bone marrow or on hematopoiesis in blood, the compounds of the invention exert a protective effect. By "protective effect" is meant that the resultant damage to the bone marrow or blood is less when the compound is administered than when it is not. The net decrease in damage may be due to protection er se — i.e., preventing the destructive effects that would normally occur or may result from accelerating recovery from such destruction. Thus, "protective effect" includes the effect of achieving this desirable result regardless of the mechanism by which it is achieved.

There are a number of situations in which the protective effect of the compounds of the invention are useful. These include instances where irradiation has resulted, or may result prospectively, in negative effects, instances where a subject is immunocompromised for any reason, instances wherein a subject exhibits damage to the kidneys, as well as instances wherein the subject has been subjected to chemotherapy. In addition, the compounds of the invention may be used in transplantation settings to increase the number of cells in the bone marrow of a donor; typically, in this case the compound may be administered in vivo or ex vivo. In this setting also, the compounds identified by the method of the invention promote the movement of progenitor cells into the peripheral blood of the donor which thus improves the recovery of peripheral blood white cell numbers in this donor; similarly, the compounds of the invention may improve the recovery of peripheral white blood cell numbers in the recipient. In general, the compounds will improve expansion and promote the eventual engraftment of transplanted cells after exposure to the - 12 -

compounds of the invention in vivo or ex vivo. The compounds identified by the method of the invention can be used directly in the recipient to hasten recovery.

In addition, patients subjected to kidney dialysis are aided by these compounds in reconstituting blood. The compounds are also useful in encouraging bone growth generally.

The compounds identified by the method of the invention can be used either in vitro or in vivo. For example, these compounds can be employed to expand or otherwise modulate hematopoietic cells in bone marrow prior to allogeneic or xenogeneic transplants. Treatment of subjects using ex vivo techniques whereby expansion of relatively undifferentiated cells from the blood stream may also be employed. The compounds can also be formulated for in vivo administration.

When ex vivo administration is employed, either bone marrow or peripheral blood (including cord blood) or both can be directly contacted with the invention compounds or fractions of these materials may be treated so long as the fractions contain suitable target progenitor cells. Preferred target progenitor cells include CD34⁺ cells, GEMM, and BFU-E.

Formulations for in vivo administration will employ standard methods such as those described in Remington's Pharmaceutical Sciences, latest edition, Mack Publishing Company, Easton, PA. The compounds may be formulated for injection, for oral administration, or for alternative methods of administration such as transmucosal or transdermal administration. Injection can be intravenous, intraperitoneal, intramuscular, or by any other conventional route. As shown hereinbelow, the compounds of the invention are effective when administered orally as well as when introduced directly into the blood stream or when administered i.p. Since oral administration is particularly convenient, and since the compounds of the invention are active when administered orally, formulations suitable for administration by mouth are particularly preferred. Such formulations include, as is well understood, pills, tablets, capsules, syrups, powders, or flavored liquids. The various formulations can be prepared in unit dosage form and can, if desired, be self- administered by the subject. The percentage of active ingredient compound (or mixture of compounds) in the formulation may vary over a wide range from about - 13 -

0.5% w/w to about 95% w/w. The preferred percentage of active ingredient will be dependent on the nature of the formulation per se. Suitable excipients included in these formulations include fillers, buffering agents, stabilizers and the like.

For administration, if desired, by injection, preferred formulations include balanced physiological solutions and liposomal compositions.

Suitable subjects who will benefit from administering the compounds of the invention, either a single compound or mixtures thereof, include vertebrate subjects, particularly mammalian or human subjects whose bone marrow progenitor cells are inadequate in number or physiological status to sustain differentiation or differentiate inappropriately. Failure of progenitor cells to result in required numbers of effector cells occurs, in particular, when the subject has been exposed to bone marrow destructive agents, such as chemotherapeutic agents, radiation, exposure to toxins in the environment and the like. Also included are those with bone marrow degenerative diseases and conditions. Thus, appropriate subjects for administration of the invention compounds include patients undergoing chemotherapy; immunocompromised patients, patients showing symptoms of anemia, neutropenia, thrombocytopenia, or lack of adequate platelet levels, and prospective subjects for treatment with cytotoxic agents. As the compounds of the invention also potentiate the cytotoxicity of chemotherapeutic agents with respect to malignant cells specifically, subjects may benefit from treatment with the compounds of the invention even though the hematopoietic system is not necessarily compromised by the chemotherapeutic treatment.

As stated above, a single compound of the invention may be included as active ingredient or the treatment may comprise use of mixtures of these compounds. In addition, the compounds identified by the method of the invention may be mixed with or used in addition to other beneficial agents such as immunostimulants or growth factors.

The dosage required depends on the nature of the subject, the nature of the condition, the manner of administration, and the judgment of the attending physician or veterinarian. Suitable dosage ranges are adjusted according to these parameters. In general, typical doses per patient will be in the range of 0.1-100 mg/kg per day for 10- - 14 -

40 days, more preferably 1-10 mg/kg per day for 14-28 days. These ranges are merely illustrative and the correct dosage optimization can be determined by routine methods.

If the invention compounds are administered as protective agents with regard to chemotherapeutic treatment, the timing of administration may also be relevant. The timing will, however, depend on the nature of the chemotherapeutic agent used. As shown below, for example, when 5 FU is used for chemotherapy, administration seems advantageous about 24 hours subsequent to administration of the 5 FU; on the other hand, although this timing of administration is also effective when cisplatin is the chemotherapeutic agent, administration about 24 hours prior to cisplatin dosing is more effective. It is clearly within routine skill to determine appropriate timing for the specific chemotherapeutic agent employed.

The following examples are intended to illustrate, but not to limit, the invention.

The examples below demonstrate that TERl 99, a specific inhibitor of GST π, potentiates the effect of cytotoxic agents in tumor cells and is also an effective myelostimulant. The results also demonstrate that jun- JNK complexes are associated with GST π and that GST π inhibits the kinase reaction catalyzed by JNK. The results in these examples, correlated with the ability of TERl 99 to interfere with the inhibition of JNK by GST π, validate the method of the invention as an assay for compounds having similar therapeutic uses.

Example 1 Potentiation of Cytotoxic Agents in Tumor Cells This example describes: 1) potentiation in human tumor cells of a cytotoxic agent currently used in cancer chemotherapy by GST inhibitors, as well as

2) enhanced intracellular efficacy of TERl 99.

HT-29 (human colon adenocarcinoma) cells were obtained from Dr. Roberto Ceriani (Cancer Research Fund of Contra Costa County, Walnut Creek, C A) and were used in log phase of growth unless otherwise specified. Chlorambucil (CMB) was obtained from Sigma (St. Louis, MO) and was dissolved in 100% ethanol. All GST - 15 -

inhibitors were dissolved in ethanol, DMSO, or water just prior to use. The same amount of solvent added to culture medium served as the vehicle control.

In a modified clonogenic assay for cytotoxicity, cells were suspended at 2 x 10⁵ cells/ml in serum- free medium in the presence of vehicle or inhibitor. Inhibitors were used at concentrations that resulted in >90% survival in the presence of inhibitor alone, when compared to vehicle treated cells. Cells were incubated for 2 hours, then varying doses of CMB were added. At the end of a second 2-hour incubation, cells were diluted to 7.5-10 x lOVml in serum-containing medium and plated in quadruplicate at 200 ml/well in Microtest III microtiter plates. Plates were incubated for 6 days and assayed by a modified methylene blue method. Briefly, cells were fixed with 1.25% glutaraldehyde in PBS then stained with 0.05% methylene blue in distilled water. Plates were washed several times in distilled water to remove unretained dye and retained dye was resolubilized in 0.03 N HC1. Plates were read at 650 nm in a Molecular Devices Vmax plate reader (Molecular Devices, Redwood City, CA). IC₅₀ values (inhibitor concentration causing 50% reduction in cell viability) were determined for the drug in the presence or absence of inhibitor from dose-response curves. A dose modification factor (DMF), a measure of potentiation of cytotoxicity, was calculated for each inhibitor by dividing the IC₅₀ value of CMB without inhibitor treatment by the IC₅₀ value for CMB with inhibitor treatment.

The results in Tables 1-3 show that several GSH analogs found to be inhibitors of GSH also potentiate killing of human rumor cells in culture by CMB which is a substrate for various GSTs. Results of potentiation tests with several GST inhibitors in HT29 cell cultures are summarized in Table 1.

- 16 -

Table 1

Potentiation of Chlorambucil Cytotoxicity in Human Cells by

GST Inhibitors and Their Esters

Parent Compound Diethyl ester

GST Inhibitor Dose tested³ DMF^b Dose tested⁰ DMF^D

(μM) (μM)

γE-C(octyl)-G N.D. - 5 0.86±0.02 γE-C(Hx)-φG 100 1.1±0.02 12.5 1.27±0.02 γE-C(Bz)-φG 100 1.08±0.01 12.5 1.65±0.04 γE-C(naphthyl)-G 200 12.5 1.21-tO.Ol

Test dose was determined from toxicity curve and analogs were used at the dose at which >90% survival occurred in the presence of the analog alone. ^bDose modification factor. Values are mean ±S.D. of 2-3 experiments.

As shown in Table 1 , this potentiation is greatly enhanced by esterification which is designed to enhance uptake of the GST inhibitors. Thus, γE-C(Bz)-φG (TERl 17) at 100 μM did not enhance cell killing by CMB, reducing the concentration CMB needed for 50% cell killing by a DMF of 1.08. In contrast the diethyl ester of γE-C(Bz)-φG (TER 199) at only 12.5 μM enhanced CMB cytotoxicity by a factor of 1.65.

The effect of esterification or amidation of the tested compounds on their potentiation of chlorambucil cytotoxicity in HT-29 cells was also determined. The

DMF was determined for the diethyl ester, the diamide, and the ester/amide of γE-C(Bz)-φG at relevant concentrations. The diester showed a DMF of 1.65 ± 0.04 for chlorambucil toxicity at 12.5 μM; the diamide showed a DMF of 1.0 in a single experiment at 200 μM; the ester/amide hybrid showed a DMF of 1.45 ± 0.16 at 50 μM concentration. The results for the diethyl ester and the ester/amide hybrid are given as the mean ±SD of three experiments.

Preferential expression of GST π isoenzyme has been reported in a range of human tumors. In the present study the efficacy of CMB potentiation of the several GST inhibitors tested correlated directly with their potencies as inhibitors of the human π class GST as shown in Table 2. - 17 -

Table 2

Rank Correlation of Chlorambucil Dose Modification Factors (DMFs) of GST

Inhibitors with Kj value for Inhibition of Human GST π

Rank R Reellaattiivvee KKii vvaalluuee R Raannkk DMF^a

Inhibitor ooff p paarreenntt ccoommppoouunndd oorrddeerr of DEE

γE-C(Bz)-φG 11 11 1.651 γE-C(Hx)-φG 22..11 22 1.272 γE-C(naphthyl)-G 33 33 1.213 γE-C(octyl)-G

44..88 44 0.864 ^aDose modification factor of diethyl ester. Values are mean ±S.D. of 2-3 experiments.

Diethyl esters of γE-C(octyl)-G (TERl 83) and γE-C(Bz)-φG (TERl 99) were tested in a standard clonogenic assay using three cell lines: HT4-1, a subclone of HT-29; SKON-3 an ovarian carcinoma, and VLB, a vinblastine-resistant variant of SKON-3. Four chemotherapeutic drugs, chlorambucil, adriamycin, mitomycin C and doxorubicin were used as the toxic agents. In these assays, the cells were seeded at 300 cells/well in 2 ml of medium in 6-well plates in the presence of the above compounds as the diethyl esters. The compounds were used at concentrations that resulted in more than 85% survival when compared to controls. After incubation for

1-2 hours to permit cells to attach, varying doses of the chemotherapeutic agents were added. At least three replicate wells were plated for each test condition and the plates were incubated for two weeks. Colonies were fixed in 95% ethanol and stained with crystal violet for colony counting. IC₅₀ values were determined for the chemotherapeutic agent in the presence or absence of the test compound and dose modification factors were calculated by dividing the IC₅₀ value of drug without the test compound by the IC₅₀ value of the drug with the test compound. The modification factors obtained in each protocol are shown in Table 3. 18

Table 3

Ability of selected GSH analogs to potentiate drug toxicity as demonstrated in a clonogenic assay

DMF^a for:

Cell GSH

Line Analog Chlorambucil Adriamycin Mitomycin C

Doxorubicin

HT4-1 TERl 99 2.39 1.2 1.03 1.20

TERl 83 1.74 1.13 1.56 n.d.^d

SKON-3 TERl 99 1.24 1.14 1.03 1.14

TERl 83 1.03 1.24 n.d.^b n.d.^d (@5 uM)^c

VLB TERl 99 Ν.D.^d 2.50 0.82 2.50

(5 μM TER199)^c

TERl 83 N.D. 1.06 1.63 n.d.^{d a}Dose modification factor.

•No data due to toxicity of analog.

Test dose was different from listed at the left. "Not determined.

As shown in Table 3, significant modification was obtained when chlorambucil was used as the drug versus HT4-1 cells in the presence of 25 μM of TERl 99. Significant modification was also achieved in NLB cells when treated with adriamycin or doxorubicin in the presence of 25 μM of the same compound.

Figure la illustrates the results for varying dosages of chlorambucil and the modifying effect of 25 μM of the diethyl ester of γE-C(Bz)-φG (TERl 99). The open squares (D) represent chlorambucil alone, the closed circles (•) chlorambucil in the presence of the invention compound. As seen in Figure la, the survival rate is markedly diminished when the invention compound is added. Figure lb confirms that the diethyl ester is necessary to penetrate the cells. HT4-1 cells were tested for survival in the presence of either γE-C(Bz)-φG (TERl 17) (closed squares, ■) or its diethyl ester (TER199) (closed circles, •). The unesterified form, TERl 17, has substantially no effect on these cells while the diethyl ester (TERl 99) is clearly toxic. - 19 -

Example 2 Stimulation of Bone Marrow Granulocyte Macrophage (GM) Progenitors TERl 99 also stimulates the production of GM progenitors in bone marrow when administered to mammalian subjects. In an illustrative assay three B6D2F, mice were treated with various doses of TERl 99 intraperitoneally. Femoral bone marrows were harvested 24 hours later and assayed for GM-CFU by the method of East, C.J. et. al. Cancer Chemother Pharmacol (1992) 31_^:123-126. An increase in the number of colonies in a dose-dependent manner up to a dosage of 90 mg/kg of TER199 was obtained. At 90 mg/kg, approximately 275 colonies/10⁴ nucleated cells were obtained compared to about 140 colonies/10⁴ nucleated cells for controls.

In an additional experiment, male B6D2F, mice, five weeks old, 20-24 grams were divided into groups of three mice and administered various dosages of TERl 99 either orally or intraperitoneally. The TERl 99 was prepared in sterile nanopore water and administered orally using a gavage tube and a 1 cc syringe or intraperitoneally in saline using a 1 cc syringe with a 28 gauge needle. Mice in the control group were injected with water or saline. Bone marrow cells were harvested 24 hours after drug treatment and added to alpha minimum essential medium (alpha MEM) supplemented with methylcellulose (0.8% w/v), fetal bovine serum (20% v/v), deionized BSA (1% w/v), Pokeweed mitogen-stimulated spleen-cell conditioned medium (PWM-SCCM)¹ (10% v/v) and gentamycin (50 μg/ml). One ml aliquots were plated (four replicate plates) and incubated for seven days at 370C. A dissecting microscope was used to count the granulocyte/macrophage colonies having more than 50 cells per colony (GM-CFU).

Pokeweed mitogen-stimulated spleen cell condition medium (PWM-SCCM) was prepared according to the procedure of Gringeri et al, 1988. Spleens were removed aseptically from four male B6D2F, mice enforced through a 200 μm wire mesh screen to obtain a single cell suspension. Ten ml of the suspension (2-4 X 10⁷ cell/ml was added to 90 ml alpha-MEM supplemented with 1% deionized BSA, 50 μg/ml gentamycin, 0.3% freshly reconstituted pokeweed antigen, 10 μM 2-mercaptoethanol. The mixture was incubated for 5 days at 370°C in a 5% C0₂ atmosphere and the resulting conditioned medium was centrifuged at 800g for ten minutes and filtered through a 0.22 μm filter. Aliquots were kept frozen at -200°C until use. - 20 -

Figure 2 shows the effect of oral versus i.p. vs i.v. administration of TERl 99 on bone marrow GM-CFU in a single treatment. The data are mean ± SEM for three mice per group. The asterisk indicates that the value is statistically significant from the control p<0.05. As shown in Figure 2, i.v. administration (closed squares, H) is most effective at 40-80 mg/kg; i.p. administration (closed circles, •) is most effective at at least 80 mg/kg oral administration (closed triangle -A-) is shown as requiring somewhat higher dosage. The results show that TERl 99 may be administered orally as well as i.p or i.v., although higher dosage levels may be required for oral administration.

Example 3 Time Course of TERl 99 Stimulation of Bone Marrow Macrophage (GM) Progenitors

The procedures of Example 2 were repeated using a single 60 mg/kg dose of TERl 99 administered i.p. on day 0 and harvesting bone marrow cells at various times after administration. The GM-CFU for the mice administered TERl 99 was compared to controls, and the results are shown as a function of day after administration in Figure 3. Maximum stimulation appeared to occur at day 2 and day 5.

Example 4 TERl 99 Effect on Mouse and Human Bone Marrow Colony Formation

The effect of TERl 99 on colony formation by granulocyte-macrophage (CFU- GM), erythroid (BFU-E), and multipotential (CFU-GEMM) progenitor cells was evaluated. TERl 99 enhances the proliferation of human and murine myeloid progenitor cells in vitro. The effects are dose-dependent, usually in the range of 1.0 to 10.0 μM, and in most cases for cells stimulated by GM-CSF, G-CSF, M-CSF,

Flt3/Flk-2 and Steel factor (stem cell factov/c-kit ligand). Of particular interest was the finding that TERl 99 enhances colony formation stimulated by combinations of cytokines. Additionally, the enhancing effect is more pronounced in human than in murine bone marrow. These results suggest that TERl 99 has enhancing effects on multiple lineages of myeloid stem cells and progenitors. That there is a greater effect on human marrow is consistent with the specificity of TERl 99 for the human GST 21

isozyme Pl-1. Results from a representative set of these experiments are presented in Tables 4-9.

Table 4

Influence of TERl 99 on colony formation by normal human bone marrow GM- progenitor cells stimulated by single cytokines.

Colony Number (% Change)* Colony & Cluster Number (% Change)*

Growth Control TER199 TER199 TER199 Control TER199 TER199 TER199 Factor Medium (0.1μM) d_/t/M) (10/M) Medium (0.1μM) (1 M) (10vM) (Per ml)

None 0 O(-) θ(-) 1 ± 1 (-) 22 ±1 22 ±2 (0) 49±2 51 ±5 (132) (123)

GM-CSF 29 ±1 28 ±2 (-3) 35±2 39 ±3 (34)* 54±1 54 ± 1 (0) 75±3(39) 81 ±6 (50) (10U) (21)*

GM-CSF 56±3 53 ±1 (-5) 60 ±1 (7) 70±2 (25)* 80±2 73 ±2 (-9) 84±2 (5) 90 + 3 (13)* (100U)

G-CSF 14±2 14±1 (0) 20 ±1 23 ±2 (64)* 26 ±1 28 ±2 (8) 39±2 42 ±4 (62)* <10U) (43)* (50)*

G-CSF 19±2 17±1 (- 17±2 (-10) 25 ±1 (32)* 33±2 29±2 (-12) 31 ±2 (-6) 42 ±1 (27)* (100U) 10)

IL-3 12±1 13±1 (8) 21 ±2 26 ±1 37 ±1 35±2 (-5) 62±8 59±5 (59)* (10U) (75)* (117)* (68)*

IL-3 39 ±5 38 ±3 (-2) 37 ±2 (-5) 52±1 (33)* 63±6 58 ±1 (-8) 64±4 (2) 81 ±2 (29)* (100U)

M-CSF 2±0.3 3±1 (50) 3 ±0.3 (50) 5±0.6 19±3 26±4(37) 37±0.3 49 ±3 (100U) (150)* (95)* (158)*

M-CSF 4±0.3 8±1 10±1{150) 11 ±1 41 ±4 43±3(5) 52±3 65±6 (59)* (1000U) (100)* (175)* (27)*

Flt3-L 11 ±3 19±3 20 ±1 23±3 28±4 48 ±1 55±3 57±4 OOng) (73)* (82)* (109)* (71)* (96)* (104)*

SLF 28±2 44 ±1 41 ±2 43 ±5 (54)* 45±2 72±2 65±4 67 ±5 (49)*

50ng) (57)* (46)* (60)* (44)*

^•Statistically significant

-22

Table 5:

Influence of TERl 99 on colony formation by normal human bone marrow

GM-progenitor cells stimulated by combinations of cytokines.

Colony Number (% Change)*

Growth Factors (per ml) Control TER199 TER199 TER199 Medium (0.1_/uM) (1|/M) (10/M)

Flt-3 dOOng) + 100U GM-CSF 77 ±1 95±5 (23)* 114±7 (48)* 115±5 (49)*

Flt-3 (100ng) + 100U G-CSF 32±4 41 ±0.6 (28)* 51 +3 (59)* 48 ±3 (50)*

Flt-3 (100ng) + 100U IL-3 55±3 55 ±2 (0) 77 ±3 (40)* 77 + 2 (40)*

Flt-3 (100ng) + 50ng SLF 38 ±4 62±2 (63)* 77±3 (103)* 77 + 4 (103)*

SLF(50ng) + 100U GM-CSF 92 + 5 92 + 5 (0) 121 ±7 (32)* 125 + 5 (136)*

SLF(50ng) + 100U G-CSF 40 + 3 41+2 (3) 55±5 (38)* 58 + 5 (45)*

SLF (50ng) + 100U IL-3 60±2 77±4 (28)* 103±10 (72)* 109±4 (82)*

•Statistically significant

Only colonies formed when Flt3-L or SLF were added together or with GM- CSF, G-CSF, or IL-3.

Table 6

Influence of TERl 99 on colony formation by normal human bone marrow erythroid

(BFU-E) and multipotential (CFU-GEMM) progenitor cells.

Colony Number (% Change from Control)

Growth Factors Added Control TER199 TER199 TER199 (per ml) Medium (0.1/vM) (I.O M) (10/YM)

BFU-E

None 0 0 0 0

Epo (1U) 36 ±6 35 ±3 (6) 60±2 (82)* 57±4 (73)

Epo (1U) + 100U IL-3 48 ±5 47 ±3 (-2) 62 + 4 (29)* 65 + 7 (35)*

Epo (1U) + 50ng SLF 88 + 4 92±4 (5) 107±7 (22)* 109±2 (24)*

CFU-GEMM

None - - - -

Epo (1U) - - - -

Epo (1U) + 100U IL-3 - - - -

Epo (1U) + 50ng SLF 22 ±2 19±2 (-14) 23 + 2 (5) 30 ±1 (36)*

^Statistically significant 23

Table 7

Influence of TERl 99 on colony and cluster formation by normal BDF, mouse bone marrow granulocyte-macrophage (CFU-GM) progenitor cells.

Colony Number (% Change)* Colony & Cluster Number (% Change)*

Growth

Control TER199 TER199 Factor TER199 Control TER199 TER199 TER199 Medium (0.1 M) (1/vM) (10//M) (Per ml) Medium (0.1 M) HμM) (10μM)

None o o 0 0 0 0 0 0

GM-CSF 9±0.6 10±2(11) 13±0.3(11) 13 ±0.7(44) 72 ±1 72 ±4(0) 80±2(11)* 81 ±2 (13)* (10U)

GM-CSF 65±4 62±4(-5) 80 ±1(23)* 71 ±0.3(9) 74±5 71 ±3(-4) 91 ±0.3(23)* 84±0.5(14)* (100U)

M-CSFdOU) o o 0 o 9±1 9±1(0) 24±0.6(167)* 26±2(189)*

M-CSF 44±2 43 ±1 (-2) 67±5(52)* 63 ±6(43)* 247 ±6 247 ±3(0) 304 ±17(23)^* 259±8(5) (100U)

PWMSCM t 72±2 74±5(3) 118±3(64)* 110±4(53)* 117 ± 1 115±9(-2) 172±2(47)* 157 ±1(34)*

(10% v/v)

*S ^sttaattiissttiiccaallllyy ssiiggnnimficcaanntt tPWMSCM = Pokeweed mitogen stimulated spleen cell conditioned medium

Tables 8 and 9 show the results of an experiment designed to compare the results obtained when TERl 99 was contacted with human bone marrow erythroid and multipotential progenitor cells as opposed to their murine counterparts. As shown in these tables, the effects ex vivo in humans (Table 8) are substantially greater than those exhibited in their murine counterparts (Table 9).

Table 8 Influence of TERl 99 on colony formation by normal human bone marrow erythroid (BFU-E) and multipotential (CFU-GEMM) progenitor cells.

Colony number (% change from control)

TER199 {μM) 0 0.1 1.0 10

BFU-E:

EPO 1 U/ml 33±6 35 ±3 (6) 60 ±2 (82)* 57±4 (73)

42±4 38 ±3 (-10) 58±1 (38)* 59±2 (3D*

EPO + 50 ng/ml SLF 88±4 92 + 4 (5) 107±7 (22)* 109±2 (24)*

66 ±3 80 + 5 (21)* 80±3 (35)* 85 + 3 (29)*

CFU-GEMM:

22 ±2 19±2 (-14) 23±2 (5) 30 ±1 (36)*

EPO + 50 ng/ml SLF 10±2 12±1 (20) 17±2 (70)* 16±1 (60)*

'Significant increase compared to control, p<0.05 -24

Table 9

Influence of TERl 99 on colony formation by normal BDF, mouse bone marrow erythroid (BFU-E) and multipotential (CFU-GEMM) progenitor cells.

Colony number (% change from control)

TER199 (μM) 0 0.1 1.0 10

BFU-E:

EPO 1 U/ml 2±1 2 + 1 (0) 2±1 (0) 2±1 (0)

4+1 4±1 (0) 3±1 (-25) 4±1 (0)

EPO + 50 ng/ml SLF 7±1 8±1 (14) 8±1 (14) 8±1 (14)

9±1 9±1 (0) 9±1 (0) 9 + 1 (0)

CFU-GEMM:

2±1 2±1 (0) 2±1 (0) 2±1 (0)

EPO + 50 ng/ml SLF 2±1 2±1 (0) 1 + 1 (-50) 2±1 (0)

Example 5

Effect of TERl 99 on Peripheral Blood Cells The effect of TERl 99 (90 mg/kg/day x 5, i.p.) on peripheral blood counts was evaluated in Sprague-Dawley derived rats. Rats were divided into two groups and each group was bled on alternating days. Mean total leukocyte, absolute lymphocyte and absolute neutrophil counts increased over the study period. TERl 99 causes a twofold increase in the levels of circulating white blood cells in rats. There was no significant change in red blood cell or platelet counts with the exception of a mean decrease in platelet count on day 9 (data not shown). In addition, TERl 99 did not appear to have any deleterious effects on these animals.

Example 6 TERl 99 Amelioration of the Effect of Chemotherapeutic Agents a) Effect of a single i.p. dose of TERl 99 on GM-CFU suppression caused by 5-fluorouracil.

The male B62F, mice described in Example 2 were administered 75 or 150 mg/kg of 5-fluorouracil (5-FU) prepared in 0.9% sterile saline and administered i.p. Mice in groups of three were injected i.p. with 60 mg/kg TERl 99 in sterile water either simultaneously with 5-FU administration, 24 hours before, 1 hour before or 24 hours after 5-FU administration. The control group was not treated with either drug. - 25 -

Bone marrows were harvested and GM-CFUs were determined 24 hours after the final injection. TER199 @-24hr.; @-lhr; and @+24 hr means TER199 was given 24 hours before, 1 hour before or 24 hours after 5-FU, respectively. 5-FU treatment alone reduces the GM-CFU to 15% of control mice. TERl 99 significantly decreases the 5- FU-induced GM-CFU suppression. Simultaneous injection of TERl 99 with fluorouracil results in a fourfold increase in the number of GM-CFUs per femur as compared with injection of fluorouracil alone. Injection of TER199, 24 hours after fluorouracil, results in GM-CFUs greater than values of GM-CFU counts per femur compared to 5-FU alone. These values were 70-95% of control, depending on the dosage of 5-FU. See Figure 4.

Administration of TERl 99 as described above 24 hours after administration of 5-FU hastened the recovery of bone marrow cells and resulted ultimately in stimulation of this capability above controls not administered 5-FU. By day 4 after 5-FU administration, mice administered 5-FU only showed GM-CFU approximately equal to control while those which had received TERl 99 in addition to 5-FU showed GM-CFU about twice that of control. Similar experiments but administering TERl 99 24 hours prior to 5-FU had essentially no effect on GM-CFU. b) Effect of a single oral dose of TERl 99 on GM-CFU suppression caused by 5-fluorouracil. The effects of TER199 administered 24 hours after injection of 5-FU by an i.p. route were also obtainable when the TERl 99 was administered orally. Bone marrow was harvested 48 hours after administering 75 or 150 mg/kg 5-FU by i.p. When administered 24 hours after 5-FU (75 or 150 mg/kg i.p.), TERl 99 (150 mg/kg p.o.) causes a two-fold increase in GM-CFU at the lower dose of 5-FU (90% vs 47% of control), and a nine-fold increase with the higher dose (71% vs 8%); see Figure 4.

Values are the mean + SE of three mice per point. c) Effect of TERl 99 on GM-CFU suppression caused by cisplatin.

The effect of a single p.o. or i.p. dose of TERl 99 was evaluated for its ability to reduce cisplatin-induced GM-CFU suppression in mice. TERl 99 (60 mg/kg i.p.) was administered 24 hours before, one hour before, or simultaneously with cisplatin

(15 mg/kg i.p.). Bone marrows were harvested 24 hours after cisplatin administration. - 26 -

GM-CFU values are the mean +SE of three mice per point. Prior administration of TERl 99 increases GM-CFUs compared to administration of cisplatin alone. Injection of TERl 99 24 hours before cisplatin results in a twofold increase in the number of GM-CFUs per femur as compared with injection of cisplatin alone (62% vs 31% of control).

The effect of oral administration of TERl 99 24 hours pretreatment or 24 hours posttreatment on cisplatin induced GM-CFU suppression was also tested. Bone marrows were harvested 24 hours after administration of the second drug. When administered orally 24 hours before cisplatin (20 mg/kg i.p.), TERl 99 (150 mg/kg p.o.) results in nearly a fourfold increase in GM-CFU (52% vs 14% of control). Administration of TER199 24 hours after cisplatin results in a 2.5-fold increase in GM-CFU (40% vs 14%). These results indicate TERl 99 may be useful in the prevention and treatment of cisplatin-induced neutropenia. d) Effect of TERl 99 on carbop latin-induced GM-CFU suppression in mice.

The effect of TERl 99 on reducing carboplatin-induced GM-CFU suppression was determined in experiments similar to those described above. TERl 99 (120mg/kg, i.p.) was administered 24 hours before, 24 hours after or simultaneously with carboplatin (90 mg/kg, i.p.). Bone marrows were harvested 24 hours after administration of the second drug. TERl 99 reduces carboplatin-induced GM-CFU suppression in mice. Oral administration of TERl 99 (150 mg/kg p.o.) is even more effective. e) Effect of TERl 99 on cyclophosphamide-induced GM-CFU suppression in mice. Administration of TER199 (120 mg/kg, i.p.) 24 hours after cyclophosphamide

(200 mg/kg, i.p.) reduces GM-CFU suppression in mice. Oral administration of TERl 99 (150 mg/kg, p.o.) is similarly effective. f) Effect of TERl 99 on melphalan-induced GM-CFU suppression in mice. The effect of TERl 99 on reducing melphalan-induced GM-CFU suppression was determined in experiments similar to those described above. Injection with - 27 -

melphalan (10 mg/kg i.p.) alone results in only 2% of GM-CFU remaining. The addition of TERl 99 (90 mg/kg i.p.) given 1 hour prior to melphalan increases the GM-CFU fourfold to 8% of control value.

Example 7

Stimulation of Cytokine Production Human stromal cell cultures were established from freshly obtained human bone marrow as described by East, C.J. et al, Blood (1992) 5:1172. On day 2, the cells were exposed for one hour to 100 μM TERl 99; culture medium was removed and replaced with fresh medium, and at 24 and 48 hours later, culture supematants were collected and tested for the presence of interleukin-1 (IL-1). The results are shown in Table 10. IL-1 levels were more than twice those of controls at both 24 and 48 hour time points.

Table 10 IL-1 levels in human bone marrow stromal cells in response to TER199

IL-1 concentration (% control)

Treatment 24 Hours 48 Hours

None 1 1 pg/ml (100) 97 pg/ml (100)

TER1 99 (100μM) 323 pg/ml (283) 245 pg/ml (253)

Example 8 Effect of TERl 99 on CD34⁺⁺⁺ Differentiation in the Presence of Various Cytokines Highly purified CD34⁺⁺⁺ cells from human cord blood or bone marrow plated at 300 cells/ml were treated with various concentrations of TERl 99 in the presence of various cytokines. Figure 5 shows the effect of concentrations of 0.1 μM-10 μM TERl 99 on granulocyte-erythrocyte-macrophage-megakaryocyte colony formation (CFU-GEMM) in the presence of 1 unit/ml of recombinant erythropoietin, 100 unit/ml of recombinant IL-3, and 50 ng/ml of recombinant steel factor. Figure 5 also shows the effect of these concentrations of TERl 99 on erythrocyte progenitor cells (BFU-E) in the presence of 1 unit/ml recombinant erythropoietin and 100 unit/ml of recombinant IL-3. As shown, these concentrations have modest positive effects on 28 -

both CFU-GEMM and BFU-E at even the lowest concentration (0.1 μM) of TERl 99. These results appear consistent as regards two individual donors.

Example 9 Peripheral Blood Response to 5-FU Treatment ± TERl 99 a) 5-FU Treatment + i.p. administration of TERl 99. The effect of TERl 99 was evaluated for its ability to lessen the degree and shorten the duration of hemato logical suppression caused by 5-FU. Sprague-Dawley derived rats were treated according to the schedule below (Table 11). The results of this study are presented in Figure 6.

Table 1 1 TER1 99 Peripheral Blood Effects Treatment Schedule

Group n = Day One Injection Day 2 — 10 Injection

I 1 2 sterile water sterile water

II 12 fluorouracil ( 1 50 mg/kg i.p.) sterile water

III 1 2 fluorouracil ( 1 50 mg/kg i.p.) TER199 (60 mg/kg b.i.d. i.p.)

IV 1 2 fluorouracil ( 1 50 mg/kg i.p.) TER1 99 (1 20 mg/kg q.d. i.p.)

The response in white blood cell, neutrophil, and lymphocyte levels in the TER199-treated groups reached pretest levels sooner than the 5-FU-treated group and at Day 12 exceeded pretest levels. The pattern differences in this response for each of these cell populations for the TER199-treated groups were significantly different from the 5-FU-treated control group (p < 0.05). These data demonstrate that, in rats, population levels of white blood cells, neutrophils, and lymphocytes in the peripheral blood supply suppressed by 5-FU, recovered and reached pretest levels more quickly following treatment with TERl 99 in comparison to placebo-treated animals.

In TER199-treated animals, platelet levels recovered to normal levels by study Day 12. In contrast, the 5-FU control animals platelet levels remained severely suppressed. This response for platelets in the TER199-treated groups was significantly different from the 5-FU-treated control group (p < 0.05). Red blood cell counts continually decreased in all groups during the course of this study. Although the observed decrease is reduced in TER199-treated animals - 29 -

compared to the 5-FU control animals, the study was terminated too early to determine if the reduced decline is a delay or an actual reduction in the nadir. b) 5-FU treatment + oral administration of TERl 99.

The treatment protocol of administering 150 mg/kg 5-FU i.p. followed 24 hours later by an oral dose of 150 mg/kg TERl 99 or vehicle in controls, followed 48 hours after 5-FU administration was repeated with additional groups of six mice each. The mice were bled through the retroorbital plexus and the blood samples were analyzed for changes in blood counts. Essentially no significant difference in total white cell counts was found. A statistically significant difference for neutrophils was obtained only on day 9. No differences were found for lymphocytes. There was a statistically significant difference for monocytes only on day 9.

The foregoing examples show that TERl 99, known to bind to and inhibit GST π, has myelostimulatory effects which are of value for themselves and further which offset the negative effects of chemotherapeutic agents. Furthermore, the foregoing examples demonstrate that TERl 99 sensitizes tumor cells to chemotherapeutic agents. The following examples demonstrate that these properties may be based on interference with the inhibition of JNK by GST π. In other words, the desirable biological properties of TERl 99 may be explained by virtue of its ability to interfere with the jun-JNK complex/GST π interaction. Therefore, the ability of a compound to interfere with the interaction of GST π with the jun-JNK complex is predictive of its ability to show myelostimulatory effects and chemosensitization of tumor cells.

Example 10 Activation of JNK in the Presence of UN Light

Whole-cell extracts were prepared from mouse 3T3-4A cells 30 minutes after sham irradiation or UN irradiation at various dosages ranging from 10-40 J/m². Preformed Jun- JΝK complex was incubated with said whole-cell extracts prior to the addition of γ-P labeled ATP. The results, shown in Figure 7, demonstrate that UN light activates the kinase activity. - 30 -

For protein kinase assays, the c-jun amino-terminal protein kinase (JNK) was purified from UV-treated 3T3-4A cells as described in Coso, O., et al., Cell (1995) 81: 1137. The purity of substrate JNK was confirmed via silver stained SDS-PAGE. The substrate was supplied as fusion protein, GST-jun (amino acids 5-89; Coso, O. et al. (supra) or _hisc-jun (full length; Treier, L.M., et al. Cell (1994) 78:787).

To avoid the effect of active JNK presence in the UN-treated cells themselves, the kinase reaction mixture contained an excess both of jun and JΝK.

In addition, UN-mediated JΝK activation is inhibited by proteins obtained from cells maintained under normal growth conditions. Figure 8 shows these results. Lane 1 shows phosphorylation of GST-jun^5"89 by JΝK purified from UV-treated mouse 3T3-4A cells. Lane 2 contains the kinase mixture after 2 μg of whole-cell extract proteins prepared from untreated cells were added before adding labeled ATP to the reaction mixture containing 5 μg of the preformed complex. Lanes 3-6 show the results of incubating the proteins from untreated cells with increasing amounts of GST-jun- JΝK complex so as to outcompete the concentration of inhibitor.

Example 11 Purification of Inhibitor from Untreated Cells Whole-cell extracts from nonstressed 3T3 cells were first tested for ability to inhibit GST-jun phosphorylation based on molecular weight. Protein extracts

(4 mg/ml) were precipitated with ammonium sulfate followed by fractionation on small concentrators with cut-off membranes of 100 kD and 30 kD; the filtrate was further concentrated to 50 μl using a 3 kD membrane. An initial determination of the inhibitor activity of proteins of various sizes showed that most of the inhibitor activity was present in fractions of <30 kD.

The fractions of <30 kD were concentrated to 50 μl as described above and loaded onto a Superdex 75 column (SMART system; Pharmacia) using a flow rate of 40 μl/min of kinase buffer. A sample of 20 μl of each fraction was tested for inhibitor activity with the results shown in Figure 9. The inhibitor activity was quantified using a radioimaging blot analyzer. Most of the activity was in fraction 16-24. - 31 -

The active fractions were pulled and applied to a Mono Q column with a gradient of 20-500 mM NaCl in kinase buffer at a flow rate of 75 μl/min. Fractions that showed inhibitory activity were pooled and loaded onto a phenyl sepharose column in a buffer (20 mM K₂HPO₄ pH 7.5, 10% glycerol, 0.6 M ammonium sulfate). The inhibitory factor eluted at 0.2 M ammonium sulfate. This was concentrated on a 3-kD column and visualized on silver-stained SDS-PAGE. When added to preformed jun-JNK complex, this fraction inhibited JNK activity. This fraction was then purified using GST-jun- JNK or _hisjun-JNK complex as an affinity reagent and when visualized on silver stained SDS-PAGE showed a single band. The identity of a single band to GST π was confirmed through microsequencing and immunoblotting with antibodies specific for GST-π. Similar results were obtained whether GST-jun- JNK or _hijun-JNK were used in the final purification step.

Example 12 Association of GST π with JNK in vivo

It was further shown that GST π is associated with jun-JNK in vivo in unstressed 3T3-4A cells. Whole-cell extracts obtained from these cells prior to or after UV radiation were immunoprecipitated either with antibodies to JNK (clone 333; PharMingen) or to c-jun (FC-45; Santa Cruz). The protein A/G bead-bound complex was analyzed on immunoblot using antibodies to GST π. These antibodies bind only to the complex isolated from the unstressed cells.

Purified GST π was added at various concentrations ranging from 25- 400 μg/μg of jun substrate either to GST-jun^5"89- JNK (2 μg/reaction) or _hisc-jun-JNK (7 μg/reaction) prior to addition of labeled ATP. The results are shown in Figure 10. As indicated, increasing concentrations of GST π show enhanced inhibition of phosphorylation.

Example 13 Characteristics of the Inhibition JNK purified from UV-treated cells was used in the kinase reaction described above to phosphorylate c-jun. Either whole-cell extract from nonstressed cells - 32 -

(putatively containing inhibitor) or such extract after immunodepletion of GST π with antibodies was added to the reaction. The results are shown in Figure 11. Bars 1 , 4 and 10 (gel Lanes A) show the results obtained when 1, 4 or 10 μg of whole-cell extract were added. Significant inhibition of phosphorylation was observed. Lane B (bar labeled "anti GST serum") shows the results when GST π was immunodepleted from the 10 μg extracts. Lane C (bar labeled "normal serum") is a control reaction with normal serum which provides an inhibition comparable to whole-cell extracts. In a second experiment, mouse fibroblasts were treated with TERl 99 for 2 hours followed either by sham treatment or UV treatment at 50 J/m². After 45 minutes, proteins were prepared from whole-cell extracts and assessed for JNK activity in the kinase reaction. As shown in Figure 12, phosphorylation was inhibited in control cells although a lower degree of inhibition was seen in the presence of TERl 99. In UV-treated cells, the kinase reaction took place even in the absence of TERl 99, but in its presence, the kinase reaction occurred at a markedly higher level. The effect of enhancing the amount of endogenous GST π through recombinant production was also explored. GST π cDNA in varying amounts was cotransfected with the β-gal (5 μg) construct into 3T3 cells via lipofection. Control vector was added to reach a constant amount of transfected DNA (15 μg). Forty-eight hours after transfection, the cells were subject to 50 J/m² UV treatment and whole-cell extracts were prepared 45 minutes later. An increase in expression of GST π corresponding to the amount of cDNA inserted was confirmed in each case. The cell extracts were then assessed in the kinase reaction using 2 μg of _hisc-jun substrate. Figure 13 shows the level of inhibition of phosphorylation in the presence of various amounts of transfected cDNA. As shown, the degree of inhibition increases in a dose- dependent manner.

Claims

- 33 -Claims

1. A method to identify a compound that has myelostimulation activity which method comprises determining the ability of said compound to interfere with the interaction of the JNK JNK substrate complex with glutathione-S-transferase (GST) wherein a compound with said ability to interfere is identified as a myelostimulant.

2. The method of claim 1 wherein said determining: a) is by a method that comprises measuring the activity of jun kinase (JNK) with respect to its complexed substrate in the presence of a kinase inhibiting amount of GST in the absence of a candidate compound; measuring the kinase activity of JNK with respect to its complexed substrate in the presence of a kinase inhibiting amount of GST in the presence of said candidate compound; comparing the kinase activity in the presence of said compound with the kinase activity in the absence of said compound; whereby a compound whose presence results in an increase in said kinase activity is identified as a myelostimulant; or b) is by a method that comprises measuring the affinity of said compound for the JNK JNK substrate complex or for GST, wherein a high affinity identifies said compound as a myelostimulant.

3. The method of claim 1 or 2 wherein said measuring the activity in a) is performed by assessing the rate of phosphorylation of a substrate for JNK; and/or wherein the substrate in a) is a jun protein or fragment, which is optionally his- tagged; and said measuring of affinity in b) is performed by detecting the formation of an association between said compound and said JNK/JNK substrate complex or said

GST in an in vitro assay, or - 34 -

wherein said determining in b) is performed by measuring the association between TER 199 and GST in the presence and absence of said compound, whereby a decrease of association of TER 199 in the presence as opposed to the absence of said compound identifies said compound as a myelostimulant; or wherein said determining in b) is by a method that comprises determining the interaction of the JNK/JNK substrate complex with GST in a two-hybrid system using an API controlled expression system as a reporter gene, wherein an increase in expression of the reporter gene identifies said compound as a myelostimulant.

4. The method of claim 3 wherein said assessing comprises complexing said substrate with an antibody that immunoreacts with the phosphorylated form of said substrate, or wherein said assessing comprises measuring the radioactivity of phosphorylated substrate phosphorylated with ³²P.

5. The method of any of claims 1-4 which is performed in an extracellular environment, and optionally wherein said GST and the JNK/JNK substrate complex are in isolated form, or wherein said GST is GSTπ.

6. A method to identify a compound that sensitizes tumor cells containing elevated levels of GST π enzymes to conditions of toxicity which method comprises determining the ability of said compound to interfere with the interaction of the JNK/JNK substrate complex with GST π wherein a compound with said ability to interfere is identified as sensitizing tumor cells.

7. The method of claim 6 wherein a) said determining is by a method that comprises measuring the activity of jun kinase (JNK) in the presence of a kinase inhibiting amount of glutathione-S-transferase in the absence of a candidate compound; - 35 -

measuring the kinase activity of JNK in the presence of a kinase inhibiting amount of GST in the presence of said candidate compound; comparing the kinase activity in the presence of said compound with the kinase activity in the absence of said compound; whereby a compound whose presence results in an increase in said kinase activity is identified as a sensitizing tumor cells; or b) wherein said determining is by a method that comprises measuring the affinity of said compound for the JNK/JNK substrate complex or for GST π, wherein a high affinity identifies said compound as sensitizing tumor cells.

8. The method of claim 7 wherein said measuring of activity in a) is performed by assessing the rate of phosphorylation of a substrate for JNK; or wherein the substrate in a) is a jun protein or fragment, which is optionally his- tagged; or wherein said determining comprises complexing said substrate with an antibody that immunoreacts with the phosphorylated form of said substrate; or wherein said determining comprises measuring the radioactivity of phosphorylated substrate phosphorylated with ³²P; or wherein said measuring of affinity in b) is performed by detecting the formation of an association between said compound and said JNK/JNK substrate complex or said GST ╧Ç in an in vitro assay; or wherein said measuring of affinity in b) is performed by measuring the association between TER 199 and GST ╧Ç in the presence and absence of said compound, whereby a decrease of association of TER 199 in the presence as opposed to the absence of said compound identifies said compound as sensitizing tumor cells; or wherein said measuring of affinity in b) is by a method that comprises determining the interaction of the JNK/JNK substrate complex with GST ╧Ç in a two-hybrid system using an API controlled expression system as a reporter gene, wherein an increase in expression of the reporter gene identifies said compound as sensitizing tumor cells. - 36 -

9. The method of any of claims 6-8 which is performed in an extracellular environment, optionally wherein said GST π and the JNK/JNK substrate complex are in isolated form; and/or wherein GST is GST π.

10. A compound that effects myelostimulation identified by the method of any of claims 1-5.

11. A compound that sensitizes tumor cells to toxic conditions identified by the method of any of claims 6-9.

12. A kit for performing the method of claim 2 or 7 which comprises a ternary complex of JNK, a JNK substrate and GST, along with a kinase buffer and ATP.