EP0822811A1 - Inhibition of apoptosis in cells expressing fas-ligand following activation - Google Patents

Abstract

Method for mediating the expression by mammalian cells of Fas-Ligand following activation using retinoid compounds, and specifically retinoid compounds which activate both retinoid X receptors and retinoic acid receptors. Examples of such retinoid compounds are 9-cis retinoic acid and related compounds. The method provides a means for inhibiting apoptosis in cells which express Fas-Ligand following activation.

Description

SPECIFICATION

Inhibition of apoptosis in cells expressing Fas-Ligand following activation

Field of the Invention

This invention relates to means for modulating the function of intracellular receptors using ligands therefor. More specifically, this invention relates to the use of retinoid compounds having both RXR and RAR activity to regulate the expression of Fas-Ligand (FasL) following cell activation, which can be utilized to mediate the inhibition of apoptotic cell death in certain cells, e.g., T cells.

Background of the Invention

Abbreviations used herein include: 9-cis RA, 9-cis retinoic acid; AIDS, acquired immunodeficiency syndrome; ATRA, all-trans retinoic acid; Fas, the Fas antigen or CD95; FasL, Fas-Ligand; FCS, fetal calf (bovine) serum; FITC, fluorescein iso-thiocyanate; FL1, fluorescence channel 1; FL2, fluorescence channel 2; FL3, fluorescence channel 3; IFN-γ, interferon gamma; IL2, interleukin 2; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate- buffered saline; RAR, retinoic acid receptor; RAR , β or γ, retinoic acid receptor alpha, beta or gamma; RARE, retinoic acid receptor response element; RNase, ribonuclease; RXR, retinoid X receptor; RXR , β or γ, retinoid X receptors alpha, beta or gamma; RXRE, retinoid X receptor response element; RT/PCR, reverse transcription/polymerase chain reaction; TCR, T cell antigen receptor; TdR, Thymidine deoxyribonucleotide; TdT, terminal deoxytransferase; TTNPB, (E) -4- [2- (5, 5, 8, 8- tetramethyl -5,6,7,8 -tetrahydro-2 -naphthalenyl ) - 1 - propenyl]benzoic acid; TUNEL, terminal deoxytransferase UTP nick end labeling; UTP, Uridine triphosphate. Normally in cells there is a fine balance between the cellular processes of proliferation, differentiation, and cell death (apoptosis) . Apoptosis, or active cell death, plays a number of essential roles in normal development, homeostasis and pathology.

In the broadest terms, apoptosis is defined as a form of cell death in which the cell plays an active role in its own demise, via synthesis of necessary macromolecular mediators (for example Fas and FasL) and the expenditure of energy. Herein, the terms "apoptosis," "active cell death" and "activation-induced cell death" are used interchangeably and refer specifically to the induction of apoptotic death in T-cell hybridomas and/or T cells by activation of said cells via the T-cell receptor (TCR) and by means which mimic that activation.

Apoptosis, sometimes referred to as physiological cell death, is distinguished from another form of cell death, i.e., necrosis, in which the cell plays a passive role, dying when membrane damage proceeds to an irreparable point of no return. Necrosis thus describes injury-type death, and it is characterized by swelling and rupture of plasma and organelle membranes and the loss of cytoplasmic organization, which is later followed by disruption of the nucleus. Necrotic tissues often become infiltrated by neutrophils. Apoptosis is a morphologically distinct form of cell death, characterized by condensation of the chromatin and loss of nuclear structure, and the formation of plasma membrane blebs. In apoptosis there is often a fragmentation of genomic DNA into an oligonucleosomal ladder, a phenomenon which is usually not seen in necrotic cell death. In some cases, apoptosis, unlike necrosis, depends upon continued RNA and protein synthesis by the dying cell, leading to the concept of cell "suicide". Cells that die by apoptosis are rapidly phagocytosed and cleared from the system without an inflammatory consequence. Apoptosis is functionally defined by its morphological features (especially chromatin condensation) , and this has persisted as the most reliable method of characterization. When DNA fragmentation occurs as it often does during active cell death, it is a useful and highly quantifiable feature (see below) ; however, apoptosis can occur in the absence of such DNA fragmentation (Tomei, L.D., et al . , Proc . Na tl . Acad . Sci . U. S . A . , 90:853-857 (1993)) . It remains possible that single-stranded DNA fragmentation (Tomei, L.D., et al. , Proc. Natl . Acad. Sci . U. S. A . , 90:853-857 (1993)) and/or cleavage into very large (approx. 50 KB) fragments (Brown, D.G., et al. , J. Biol . Chem. , 268:3037-3039 (1993)) are always associated with apoptosis, but this has not been confirmed. Similarly, active cell death can depend upon expression of specific genes in some cases (Colotta, F., et al. , J. Biol . Chem . , 267:18278-18283 (1992) ; Ellis, H.M., et al. , Cell , 44:817-829 (1986) ; Ellis, R.E., et al . , Genetics, 129: (1991)) , but again, apoptosis is not defined by this requirement, unless genes whose functions are universally required for apoptosis are identified. One interesting form of apoptosis is seen in lymphocytes (e.g., B-cells, T-cells) following stimulation through antigen receptors (e.g., immunoglobulin, T-cell receptor) , a signal that is usually associated with lymphoid activation and proliferation. Such activation-induced apoptosis is thought to be the mechanism of negative selection in thymocytes, peripheral deletion in mature T-cells, and loss of non-infected T-cells in patients having AIDS (e.g., Meyaard, L. , et al. , Science, 257:217-219 (1992)) . The induction of apoptosis in response to a receptor- transduced signal is by no means restricted to the T cell, or even hemopoietic tissue compartment. An example is the induction of apoptosis in normal hepatocytes following interaction between Fas antigen expressed on the surface of the liver cells and an antibody specific for the Fas antigen (Ogasawara, J., et al . , Nature, 364:806-809 (1993) ) .

T-cell hybridomas, thymocytes, and T cells can be induced to undergo apoptotic cell death by activation through the T-cell receptor (TCR) (Cohen, J.J., et al . , J. Immunol . , 132:38-42 (1984) ; Shi, Y., et al . , J. Immunol . , 144:3326-3333 (1990) ; Odaka, C, et al . , J. Immunol . , 144:2096-2101 (1990) ; Ucker, D.S., et al . , J. Immunol . , 143:3461-3469 (1989)) . This process requires macromole- cular synthesis and thus gene expression, and has been shown to be influenced by factors regulating gene transcription (Shi, Y., et al. , Science , 257:212-214 (1992) ; Bissonnette, R.P., et al . , J. Exp. Med. , 180:2413- 2418 (1994) ; Woronicz, J.D., et al. , Nature, 367:277-281 (1994) ; Liu, Z.G., et al . , Nature, 367:281-284 (1994)) . Recently, it has been shown that following activation, T-cell hybridomas rapidly express the Fas/CD95 receptor and its ligand, Fas-Ligand (FasL) , which interact together to transduce the death signal in the activated cell (Dhein, J., et al . , Na ture, 373:438-441 (1995)) . Both Fas and FasL molecules are induced within 4 hours of activation, and competitive inhibition of their inter¬ action blocks the induction of apoptosis (Brunner, T., et al. , Nature, 373:441-444 (1995)) . Moreover, the evidence suggests that in the mouse model of activation-induced cell death, this interaction can take place on the surface of a single cell, with no requirement for cell:cell contact between two separate cells (Brunner, T., et al . , Nature, 373:441-444 (1995)) . In the human version of this model, FasL is not displayed on the cell surface, but rather is secreted, where it forms trimers and prior to interacting with the target cell expressing the Fas antigen (Dhein, J., et al. , Nature, 373:438-441 (1995)) . Regardless of the model, activation-induced apoptosis in T cells proceeds via a cell-autonomous (single cell) Fas/FasL interaction. Outside of the Fas/FasL interaction, multiple other signaling pathways are also known to be involved in regulating cell death. Furthermore, activation-induced apoptosis is inhibited by several different agents, including the immunosuppressive drugs cyclosporine (Shi, Y., et al., Nature, 339:625-626 (1989) ; Shibuya, H.,_. et al., Cell , 70:57-67 (1992) ; Dowd, D.R., et al. , J. Biol . Chem . , 266:18423-18426 (1991)) and FK506 (Bierer, B.E., et al., Proc . Na tl . Acad . Sci . U. S . A . , 87:9231-9235 (1990) ) , glucocorticoids (Zacharchuk, CM., et al . , J. I_nmu__ol . , 145:4037-4045 (1990)) and retinoids (Yang, Y.L., et al . , Proc . Na tl . Acad . Sci . U. S. A . , 90:6170-6174 (1993)) , among others. Because activation-induced apoptosis in T cells and other cells (for example B cells) is known to play a fundamental role in the control of cell numbers (homeostasis) , in development (e.g., the elimination of autoreactive cells or negative selection) , in the regulation of immune reactivity systemically (clonal deletion or peripheral tolerance) , as well as contribute to the pathology of many diseases (e.g., AIDS, chronic hepatitis) , the ability to modulate this basic biological pathway has the potential to be of great benefit.

Retinoids modulate the specific expression of target genes by binding to and activating two distinct subfamilies of intracellular receptors, the retinoic acid receptors (RARs) and the retinoid X receptors (RXRs) (reviewed in Mangelsdorf, D.J., et al. , The Retinoids: Biology. Chemistry and Medicine 2nd ed. , pp. 319-349 (1994) . Retinoids are derivatives and/or metabolites of vitamin A. Examples of retinoids include all- rans retinoic acid (ATRA) , 9-cis retinoic acid (9-cis RA) , and (E) -4- [2- (5,5,8, 8tetramethyl-5, 6,7, 8-tetrahydro-2 - naphthalenyl) -1-propenyl] benzoic acid (TTΝPB) . Like other members of the nuclear receptor superfamily, RARs (α,(β and γ) and RXRs α , β and 7) become transcriptionally active upon ligand binding and transactivate their target genes by binding to hormone response elements that generally consist of two "direct repeat" (DR) half-sites of the consensus sequence AGGTCA. All-trans retinoic acid (ATRA) is a high affinity ligand for RAR (a , β and 7) only, whereas 9-cis retinoic acid (9-cis RA) , the active stereoisomer of ATRA, is a high affinity ligand for both RARs and RXRs (Allegretto, E.A., et al. , J. Biol . Chem. , 268:26625-26633 (1993)) . RAR/RXR heterodimers activate transcription in response to ATRA or 9-cis RA, while RXR/RXR homodimers transactivate in response to 9-cis RA. The observation that 9-cis RA is 10-fold more potent than

ATRA in blocking activation-induced apoptosis suggests that this inhibition involves retinoic X receptors (Yang,

Y.L., et al., Proc . Na tl . Acad . Sci . U. S . A . , 90:6170-6174

(1993) ) . Retinoids are potent modulators of apoptosis in a number of experimental models. However in most of the experimental models where the effects of retinoids have been investigated, retinoids have been found to induce either differentiation or apoptosis, or in many cases both (for extensive reviews, see Moon, R.C., et al. , The Retinoids: Biology. Chemistry, and Medicine 2nd ed. , 573- 595 (1994) ; ibid., Hong, W.K., et al. , 597-630) . Interestingly, retinoids have also been shown to inhibit activation-induced apoptosis in T-cell hybridomas and developing T-cells (thymocytes) (Yang, Y.L., et al. , Proc . Natl . Acad. Sci . U. S. A. , 90:6170-6174 (1993) ; Iwata, M. , et al., J. Immunol . , 149:3302-3308 (1992)) . However, the specific pathway by which retinoids act to inhibit activation-induced apoptosis has not been previously identified or demonstrated. We believe the present dis¬ closure represents the first instance where the mechanism whereby retinoids inhibit activation-induced apoptosis has been characterized and identified, and demonstrate that the mechanism of retinoid-mediated inhibition of that process is by the inhibition of expression of Fas-Ligand. The entire disclosures of the publications and references referred to above and hereafter in this specification are incorporated herein by reference.

Summary of the Invention In this invention, we have determined and demonstrated that certain retinoid compounds inhibit T-cell receptor mediated apoptosis in T-cell hybridomas by specifically blocking the expression of Fas-Ligand (FasL) on T cells following activation, and in particular, retinoid compounds having high affinity and activating potency for both RARs and RXRs.

Description of the Figures

Figure 1 shows the inhibition of activation-induced apoptosis in anti-CD3 stimulated 2B4 T-cell hybridomas cells by 9-cis RA, expressed as % cells undergoing apoptosis. The T-cell hybridoma 2B4 was activated in anti-CD3-coated 6-well microtiter plates for 12 hours at 37°C, in the presence or absence of 1 μM 9-cis RA. The cells were collected, fixed in a 1% parafor aldehyde in phosphate-buffered saline (PBS) , assayed for DNA fragmentation by terminal deoxytransferase UTP nick end-labeling (TUNED, according to the kit manufacturer's instructions, and analyzed by flow cytometry. Cells undergoing apoptosis with accompanying DNA fragmentation incorporate label (UTP-digoxigenin/digoxigenin-FITC) and can be detected by flow cytometry as a peak shifted to the right of the non-apoptotic cells, indicating an increase in DNA fragmentation and apoptosis. The data shown represents the percentage of apoptotic cells (%TdT positive) in each sample (inset) .

Figure 2 shows the ability of 9-cis RA to inhibit the expression of FasL functional activity, expressed as % DNA fragmentation induced in Fas^* target cells. Under normal conditions, i.e., tissue culture, 2B4 do not express FasL, and therefore do not induce apoptosis in Fas⁺ target cells via a Fas/FasL interaction. Following activation by antibody against the TCR-CD3 complex (anti-CD3) , 2B4 cells do express. FasL functional activity was measured as the cytotoxic activity of 2B4 cells against Fas" target cells. L1210 and L1210.fas target cells were labeled with [³H] -TdR, then plated with either 2B4 cells, at the ratios indicated, into 96-well plates coated with anti-CD3 antibody. The cultures were incubated in the presence or absence of 9-cis RA for 10 hours at 37°C, extracted, and the radioactivity retained on glass fiber mats was counted in a scintillation counter. The results shown are expressed as % DNA fragmentation relative to control wells (absence of 2B4 effector cells) and represent the mean of 6 replicates with standard deviation (bars) . Symbols: D unstimulated effector cells (2B4) , o anti-CD3-stimulated, ■ unstimulated effector cells plus 1 μM 9-cis RA, • anti-CD3 stimulated effector cells plus 1 μM 9-cis RA.

Figure 3 demonstrates that the ability of 9-cis RA to inhibit the expression of FasL functional activity is specifically via modulation of FasL expression and not by interference with the ability of the Fas antigen to transduce a death signal. In this example, this is shown where the induction of the death signal is mediated by the interaction between Fas and FasL directly, and is expressed as the % DNA fragmentation induced in Fas+ target cells. 2B4 cells were plated at time 0 into 96-well plates coated with anti-CD3 antibodies, in the presence or absence of 9-cis RA. The L1210 and L1210.fas target cells were not added at this time. The T hybridoma cells were then cultured for 6 hours in the presence of the activating antibody to allow the expression of FasL, at which time the [³H] -TdR-labeled L1210 and L1210.fas target cells (as in Figure 1) were added. A parallel set of cultures which had not received retinoid at time 0, did so at time +6 hours. All plates were cultured a further 6 hours and extracted as described above in Figure 1. The results shown are expressed as % DNA fragmentation relative to control wells (absence of effector cells) and represent the mean of 6 replicates with standard deviation (bars) . Symbols: D unstimulated effector cells (2B4) mixed with L1210 cells, ■ unstimulated effector cells mixed with L1210.fas cells, o anti-CD3 stimulated effector cells plus L1210, • anti-CD3 stimulated effector cells plus L1210.fas.

Figure 4 demonstrates that the ability of 9-cis RA to inhibit the expression of FasL functional activity is specifically via the modulation of FasL expression and not by interfering with the ability of the Fas antigen to transduce a death signal. However, in this example this is shown where the induction of the death signal is mediated by a monoclonal antibody specific for Fas, expressed as the % DNA fragmentation induced in Fas^'* targets cells. Fas^* Jurkat cells were labeled with

[³H] -TdR, then plated with into 96-well plates together with anti-Human Fas at the indicated concentration and 1 μM 9-cis RA for 10 hours at 37°C. The results shown are expressed as % DNA fragmentation relative to control wells (absence of effector cells) and represent the mean of 6 replicates with standard deviation (bars) . Symbols: D Untreated control cultures, ♦ cells plus 1 μM 9-cis RA.

Figure 5 shows the ability of 9-cis RA to modulate the expression of FasL at the mRNA level, detected qualitatively by the reverse transcription of polyA mRNA and specific oligodeoxynucleotide pri er/Taq DNA polymerase amplification of FasL mRNA (RT/PCR) . RT/PCR analysis was performed on total RNA obtained from 2B4 cells activated by anti-CD3 in the presence or absence of 1 μM 9-cis RA. RNA was reverse transcribed using an oligo dT primer and the resulting cDNA phenol :chloroform extracted. PCR amplification was performed on 10-fold serially diluted cDNA using the primer pairs designed for each mRNA. The products were electrophoresed on a 1% agarose in TAE buffer (40 mM Tris acetate, pH 8.5, 2 mM EDTA) gel, stained with ethidium bromide and photographed. Figure 6 shows the ability of 9-cis RA to modulate the expression of FasL at the mRNA level, detected quantitatively by RNase protection assay using FasL-specific probes, expressed as fold induction above controls (control = 1) . RNase protection analysis was performed using a [³²P] -UTP labeled RNA probes generated by T7 RNA polymerase and the same total RNA preparations used for RT/PCR analysis in Figure 5. The RNase protected fragments were resolved on a 6% urea/PAGE sequencing gel. The gel was dried, exposed to X-ray film, and the resulting autoradiograph analyzed by densitometry scan. The data shown in the graph represents the amount of FasL mRNA expression normalized to the g-actin in each sample and is expressed as fold induction FasL signal relative to unactivated 2B4 controls.

Figure 7 shows the correlation between the ability of 9-cis RA to block FasL expression at the mRNA level and the ability of 9-cis RA to prevent the expression of FasL protein, detected on the cell surface by flow cytometry. 2B4 cells were incubated for 6 hours at 37°C on 6-well plates coated with anti-CD3, in the presence or absence of 1 μM 9-cis RA. The cells were stained with the chimeric protein FasFc, which comprises the N-terminal portion of Fas, minus the transmembrane and cytoplasmic regions, with the Fc portion of human IgGl. FasFc binds to FasL, which is then detected by staining with anti-human Fc/biotin, and streptavidin PerCP.

Figure 8 shows the structures of the receptor- selective retinoid compounds used to demonstrate the bifunctional (RAR and RXR) receptor requirements for the inhibition of FasL expression by retinoids.

Figure 9 demonstrates the retinoid receptor activation selectivity of the receptor-selective compounds illustrated in Figure 8, expressed as the fold induction of retinoic acid receptor (RARE) -driven and retinoid X receptor (RXRE) -driven gene expression. Ligand was added to CV-1 cells transiently co-transfected with receptor 11 expression vectors (RAR or RXR) , together with either of the two reporter constructs containing response elements for RAR (RARE, DRl-tk-Luc) or RXR (RXRE, DR5-tk-Luc) . Transfection efficiency was monitored by co-transfection with a b-galactosidase expression vector. The data is expressed as fold induction relative to controls without ligand, and were normalized relative to the b-galactosidase signal. Symbols: □ TTNPB, ■ LG100268, • 9-cis RA. Figure 10 shows the relative ability of the receptor-selective compounds to inhibit activation-induced apoptosis, expressed as % apoptotic cells. 2B4 cells were cultured in anti-CD3 coated 6-well microtiter plates in the presence or absence of the indicated compounds, added at a concentration of lμM. After 12 hours the cells were collected and assayed for DNA fragmentation by TUNEL, with propidium iodide counterstaining. Plotted as FL3 (PI, DNA content) on the Y-axis and FL1 (FITC, TdT positive) on the X-axis, apoptotic cells appear low and to the right (TdT positive and hypodiploid) . The numbers (inset) shown represent the percentage of apoptotic cells in each sample using the region marker shown in the dot plots.

Figure 11 shows the relative ability of the receptor-selective compounds to inhibit the expression of the functional activity of FasL on 2B4 effector cells, expressed as the % DNA fragmentation induced in Fas* targets cells. The experimental procedures used here are the same as those described for Figure 2. [³H] -TdR-labeled L1210 and L1210.fas target cells were exposed to anti-CD3 activated 2B4 cells at the ratios indicated in the presence or absence of 1 μM ligand. The data shown are the results obtained from an effector (2B4) to target

(L1210, L1210.fas) ratio of 2:1. They are expressed as %

DNA fragmentation relative to control wells (absence of 2B4 effector cells) and represent the mean of 6 replicates with standard deviation (bars) . Symbols: ■ L1210, D L1210.fas. Figure 12 shows the effect of adding increasing concentrations of the indicated retinoid compounds on the induction of apoptosis in 2B4 cells following activation. The figure demonstrates that the requirement for both RAR- and RXR-selective compounds for effective inhibition of FasL expression is not simply due to the presence of more retinoid, but to bifunctional activation of both retinoid receptor (RAR, RXR) pathways, expressed as % DNA fragmentation induced in activated 2B4 cells. Functional expression of FasL was determined by the ability of activated 2B4 cells to express FasL-mediated cytotoxicity on Fas+ L1210 cells, as described for Figure 2. Symbols: □ unstimulated controls, +/- retinoid, ■ anti-CD3- activated, +/- retinoid. Figure 13 demonstrates a) the requirement for both RAR- and RXR-selective compounds for effective inhibition of FasL expression at the mRNA level, and b) the specificity of that inhibition, by demonstrating the lack of similar inhibition of Fas expression, expressed as fold induction of Fas and FasL mRNA relative to controls. Fas and FasL mRNAs were quantitated by RNase protection as described in Figure 3. Included in the experiment was a 7-actin internal control. The T hybridoma cells were activated for 2 hours in 6-well plates coated with anti-CD3, the RNA extracted for RNase protection. Retinoids were added at 1 μM each. The data shown in the graph represents the amount of FasL mRNA expression normalized to the 7-actin in each sample and is expressed as fold induction of Fas or FasL signal relative to unactivated controls (=1) . Symbols: π unstimulated controls, ■ anti-CD3 activated.

Figure 14 shows the effect of retinoids on the production of the cytokine interleukin-2 (IL2) by activated 2B4 cells, expressed as the incorporation of [³H] -TdR label by IL2-responsive CTLL-2 cells, and as units of IL2 produced (in parentheses) . 2B4 cells were cultured 16 hours in 96-well plates coated with anti-CD3 antibody, with the indicated retinoids added at 1 μM. The culture supernatants were collected, centrifuged, and IL2 unit concentrations determined using the IL2-dependent cell line CTLL-2 and a recombinant IL2 standard. The CTLL-2 cells were cultured with the supernatants for 16 hours and pulsed 6 hours with 5 μCi/ml [³H]-TdR. Units of IL2 per ml were calculated using a standard curve and linear regression analysis. The data shown represents the mean of 6 determinations (DPM) with standard deviation. Symbols: O unstimulated controls, ■ anti-CD3 activated, o activated plus TTNPB, • activated plus LG100268, Δ activated plus 9-cis RA, + standard.

Detailed Description of the Invention

Activation-induced apoptosis in T-cell hybridomas occurs via multiple signaling pathways, but requires and is dependent upon the expression of two molecules, Fas (CD95) and Fas-Ligand (FasL) , which interact to transduce the apoptotic signal (Ju, S.T., et al. , Nature, 373:444-448 (1995) ; Brunner, T., et al . , Na ture, 373:441-444 (1995) ; Dhein, J. , et al . , Nature, 373:438-441 (1995)) . Using several different T-cell hybridomas, it has been found that both molecules are induced within 4 hours of activation, and that competitive inhibition of their interaction blocks the induction of apoptosis, and that this interaction can occur even on a single cell

(Dhein, J., et al., Nature, 373:438-441 (1995) ; Brunner,

T., et al., Nature, 373:441-444 (1995) ; Ju, S.T., et al. ,

Na ture, 373:444-448 (1995)) . Thus, activation-induced apoptosis in T-cell hybridomas proceeds via a cell autonomous Fas/FasL interaction.

We have now demonstrated, as shown in the following example, that retinoids such as 9-cis retinoic acid (9-cis RA) specifically inhibit expression of FasL following activation, and by that specific pathway block activation-induced apoptosis. We have also demonstrated that retinoids which have high affinity and activation potency for both RXRs and RARs (pan-agonists) are most effective in blocking FasL expression and thus activation-induced apoptosis, indicating that this inhibition involves retinoid X receptors. The following examples are presented to illustrate, but jiot limit, the invention.

Example 1

The inhibition and detection of apoptosis in T-cell hybridomas activated by antibodies to the T-cell receptor and treated with 9-cis RA:

All cells were maintained in RPMI 1640 or IMDM media supplemented with 10% FCS, glutamine and b-ME, using standard tissue culture techniques. For the assay, the mouse T-cell hybridoma line 2B4 was used and has been described in detail elsewhere. The 2B4 cell line expresses the T-cell receptor-CD3 (TCR-CD3) surface receptor complex and undergoes activation-induced apoptosis when exposed to anti-TCR antibodies (Shi, Y., et al., Nature, 339:625-626 (1989) ; Mercep, M. , et al. , J. Immunol . , 142:4085-4092 (1989)) .

The pan-RAR/RXR agonist 9-cis retinoic acid (9-cis RA) was synthesized at Ligand Pharmaceuticals. The synthesis of 9-cis RA has been described in detail elsewhere (Boehm, M.F., et al., J. Med. Chem., 37:2930-2941 (1994)) . 9-cis RA was solvated to 1 mM in DMSO:ethanol.

2B4 cells used in the assay were maintained in log phase growth. To activate the cells to undergo apoptosis, the cells were incubated in the presence of a monoclonal anti-CD3 antibody (clone 145.2C11) The B cell hybridoma clone 145.2C11 secreting a monoclonal hamster anti-mouse CD3 antibody (Leo, 0., et al . , Proc. Natl . Acad. Sci . U. S.A . , 84:1374-1378 (1987)) was obtained from the ATCC. 145.2C11 culture supernatants were clarified and the antibody was affinity purified using protein A-Sepharose chromatography (Pharmacia, Piscataway, ΝJ) . Tissue culture plates used in the activation-induced cell death experiments were pre-coated with the above purified antibody, which had been diluted in to 0.5 μg/ml in a Tris-HCl buffer, (0.05 M pH 9.0) . The plates/dishes were coated for 2 hours @ 37°C, following which they were stored at 4°C for a period of no more than 2 weeks before use.

Just prior to being used, the plates/dishes were washed extensively with a phosphate-buffered saline solution (PBS) to remove unbound antibody. The cells were seeded in to plates, together with retinoid, for a period of not less than 8 hours. Typically, the experiments were allowed to proceed for a period of 16 hours before assays of apoptosis would be performed. By 16 hours, activated but otherwise untreated 2B4 cells undergo apoptosis in excess of 50%.

In this example (Figure 1) , the quantitation of apoptosis was performed by terminal deoxytransferase UTP nick-end labeling of fragmented DNA (TUNEL, ref. (Gorczyca, W. , et al. , Jnt. J^". One, 1:639-648 (1992)) , Oncor ApopTag™, Gaithersburg MD) . Briefly, the cells were fixed in paraformaldehyde, washed several times and then incubated with the enzyme terminal deoxytransferase (TdT) in the presence of labeled (digoxygenin) UTP, which results in the incorporation of label at points of DNA strand breakage. The cells were then incubated with an

FITC-labeled anti-digoxygenin antibody, to detect the amount of UTP-label incorporated and thus the amount of DNA fragmentation present. The cells were then counterstained with propidium iodide for DNA content, and analyzed by flow cytometry on a FACScan using

Macintosh-based CELLQuest™ software (Becton-Dickinson, San Jose, CA) .

The results of a typical experiment are shown in

Figure 1. The data is depicted by a histogram, where the area under the histogram at any one point along the x

(horizontal) axis represents the number of cells (y-axis) having a certain amount of fluorescence (x-axis) , which here is direct quantitation of DNA fragmentation and thus apoptosis.

As shown in Figure 1, 9-cis RA effectively inhibited activation-induced apoptosis in anti-CD3 stimulated 2B4 (from 77.0% to 28.6%) cells. The concentration of 9-cis RA required to produce a 50% inhibition of this response was 150 nM (data not shown) . Similar results were obtained using other methods (not shown) , and these observations are consistent with those reported by others (Yang, Y.L., et al. , Proc . Natl . Acad. Sci . U. S. A. , 90:6170-6174 (1993)) , demonstrating that the RAR/RXR pan-agonist 9-cis RA is a potent modulator of activation-induced apoptosis.

Example 2 The inhibition of apoptosis in activated T-cell hybridomas via the negative regulation, bv 9-cis RA, of Fas-Ligand expression:

It was recently demonstrated that activation-induced apoptosis in T cell hybridomas proceeds via the interaction between the Fas antigen and its physiological ligand, Fas-L. Therefore, we envisioned three non¬ exclusive possibilities for the mechanism of inhibition by 9-cis RA of activation-induced apoptosis: inhibition of FasL expression, inhibition of Fas expression, and/or inhibition of Fas-transduced apoptosis. These possibilities were examined functionally in this example, by the experiment shown in Figure 2. In this example, L1210 mouse T leukemia cells transfected to constitutively express human Fas (L1210.fas) were used to measure functional activity, and therefore the presence of the Fas-L protein.

Functional activity of FasL was determined by the ability of cells expressing FasL to induce apoptosis in the Fas^* L1210 target cell line transfected to constitutively express mouse Fas (L1210.fas, (Rouvier, E., et al., J. Exp . Med . , 177:195-200 (1993))) . To determine the extent of DNA fragmentation, and thus apoptosis in this example, a well-established assay which detects the amount of DNA fragmentation as a loss of genomic DNA, was utilized. DNA fragmentation was determined as follows: L1210 and L1210.fas cells were labeled for 2 hours at 37°C with 5 μCi/ml [^**__] -Thymidine deoxyribonuleotide ( [³H] -TdR) in RPMI 1640/10% FCS. The cells (50xl0³ per well) were washed twice in warm HBSS, resuspended in IMDM/10% FCS, and seeded into 96-well plates uncoated or coated anti-CD3 antibody. Twelve hours later, the plates were extracted with a 96-well plate harvester (Skatron, Sterling, VA) onto glass fiber filter mats (Skatron, Sterling, VA) and the unfragmented, high molecular weight [³H] -TdR-labeled DNA retained was counted in a liquid scintillation counter

(Pharmacia LKB) . The difference between the amount of radioactivity recovered in experimental cultures relative to control cultures is a measure of the loss of low molecular weight, fragmented DNA, and thus apoptosis. The data are expressed as percent DNA fragmentation, i.e., 100 X (1-cpm in experimental group over cpm in unstimulated cells) . Each data point represents the mean of 6 wells with standard deviation.

L1210 wild type and Fas+ L1210.fas cells were labeled for 2 hours at 37°C with 5 μCi/ml t³H] -TdR in RPMI 1640/10% FCS. The cells were washed twice in warm HBSS and resuspended in IMDM/10% FCS. The target cells (25 x 10³ per well) were added to 96-well plates containing T hybridoma cells which had been seeded into uncoated or anti-CD3 coated wells at densities yielding the indicated final ratio of effector to target cells.

The cultures were incubated for 8 hours at 37°C, and the unfragmented, high molecular weight DNA was extracted and counted as described above. Determinations of incorporated label from control (without effector cells) cultures of L1210 and L1210.fas target cells at the beginning and end of the assay reveal very little significant loss (plus or minus 5%) of incorporated [³H] -TdR label (59,922 DPM vs 62,446 DPM for L1210, and 55,643 DPM vs 54,936 DPM for L1210.fas cells) . The data are expressed as percent DNA fragmentation, i.e., 100 X (1-cpm in experimental group/cpm in targets alone) . Each data point represents the mean of 6 wells with standard deviation.

L1210 T cell leukemia cells normally express neither FasL nor CD3, and only barely detectable levels of Fas, and are therefore resistant to FasL-induced cell death. L1210 cells transfected with mouse fas (L1210.fas) cells express high levels of Fas constitutively and are sensitive to the induction of apoptosis either by cells expressing FasL, or anti-Fas antibody. Therefore, they can be used as target cells to detect FasL expression when mixed with 2B4 effector cells activated via the TCR to express FasL (and Fas) . Activation of the T-cell hybridoma with anti-CD3 antibody resulted in apoptosis in L1210.fas target cells, as measured by DNA fragmentation (Figure 2) . The induction of apoptosis in this well-characterized reaction is mediated by Fas/FasL interactions, since no apoptosis is seen in L1210 targets lacking Fas expression, and since the apoptosis is blocked with a soluble Fas-Fc chimeric protein which acts as a competitive inhibitor of the FasL (Brunner, T., et al. , Nature, 373:441-444 (1995)) . Addition of 9-cis RA significantly inhibited the cytotoxic activity of the T-cell hybridomas activated with anti-CD3 antibody (Figure 2) , thus implicating a regulation of the expression of Fas-L by 9-cis RA.

Example 3

The selectivity of the inhibition of FasL expression bv

9-cis RA:

When the Fas antigen interacts with Fas-Ligand, either in solution or on the surface of the same or other cells, that interaction between these two molecules results in the transduction of one-way death signal into the Fas-expressing cell. Utilizing the methods used in Example 2, the selectivity of the 9-cis RA-mediated inhibition of FasL expression was evaluated. The data shown in Figure 3 demonstrates that the ability of the retinoid to inhibit the activity, i.e., the expression.of FasL and the subsequent cytotoxicity against the Fas+ L1210.fas target cells, was clearly not due to a simple blockade of Fas receptor signaling per se (i.e., that it did not interfere with the ability of Fas to transduce a death signal once it had interacted with FasL) . To make this determination, we simply delayed the addition of either L1210.fas target cells or retinoid for 6 hours, thus allowing for the expression of FasL (Brunner, T., et al., Nature, 373:441-444 (1995)) .

Control cultures received retinoid only at time 0. After the delay, both retinoid and target cells were added, and the cultures incubated a further 6 hours to allow for the induction of apoptosis in the L1210.fas targets, as measured by DΝA fragmentation. If retinoids inhibit activation-induced cell death via a block in Fas signal transduction, delaying the addition of retinoid should nevertheless produce comparable levels of inhibition. As shown in Figure 3, when 9-cis RA was added to cultures of 2B4 cells at the same time as exposure to anti-CD3, the ability to induce apoptosis in L1210.fas was completely inhibited. Delaying the addition of the target cells had no effect on the outcome. More importantly, delaying the addition of retinoid completely removed the inhibitory effect normally seen when retinoid is added at time 0, suggesting that retinoids do not inhibit activation-induced apoptosis via a block Fas signal transduction but rather some event (s) taking place immediately following activation. The data shown supports the interpretation that 9-cis RA does not interfere with Fas signaling, but inhibits FasL expression. The above data demonstrating that the specificity of 9- cis RA-mediated inhibition of apoptosis is at the level of FasL expression and not due to a block in Fas signal transduction is supported and confirmed by the data shown in Figure 4. Jurkat cells express Fas, and can be induced to undergo apoptosis by either FasL or by monoclonal antibodies specific for Fas. Thus the interaction between Fas and antibody binding Fas can be thought to mimic the interaction between Fas and FasL. Figure 4 shows that the result of that interaction between Fas and antibody specific for Fas causes apoptosis, and is not blocked by 9-cis RA. The data confirms that 9-cis RA blocks not Fas signal transduction or the interaction between Fas and FasL if FasL is already expressed, but rather prevents the expression of FasL.

Example 4

Inhibition of the induction of FasL mRNA expression, in

2B4 cells following activation, in the presence of 9-cis

RA: mRNAs expressed following activation were detected by Reverse transcriptase-Polymerase Chain reaction (RT/PCR) . Total RNA was obtained from cells, activated as described in detail in examples 1-3, by acid-guanidium thiocyanate-phenol extraction (Xie, W.Q., et al. , Biotechniques, 11:324-327 (1991)) . Reverse transcriptase synthesis of cDNAs was performed using oligo dT primers and a commercially obtained kit (cDNA Cycle Kit, Invitrogen, San Diego, CA) . The resulting cDNA:RNA hybrids were phenol :chloroform extracted, precipitated, and serially diluted for PCR amplification with Taq polymerase (Gibco/BRL, Grand Island, NY) and the following primer pairs: GAPDH, sense (5' GTGAAGGTCGGAGTCAACG) and antisense (5' TGAAGACGCCAGTGGACTC) ; nur77a, sense (5' GGAACACCAGCAACGAGC) and antisense (5' CATCTGGAGGCTGCTTGG) ; FasL, sense (5' TTCTCTGGAGCAGTCAGCGT) and antisense (5' TAAGGACCACTCCATGGACC) . _τ_ _™τ_,__

PCT/US96/06090

21

The products of the PCR reaction were resolved on 1.5% agarose in TAE (40 mM Tris acetate, pH 8.5, 2 mM EDTA) gels, stained with ethidium bromide and photographed. The developed films were scanned and the bands representing the resolved PCR products corresponding to the indicated mRNAs were analyzed using a PDI (Huntington Station, NY) scanner and Quantity One™ software.

As shown in Figure 5, FasL expression in 2B4 cells as measured by RT/PCR was induced as a result of anti-CD3 stimulation. This expression was inhibited by the presence of 9-cis RA. Simultaneously we monitored the expression of nur77a, an orphan intracellular receptor known to be induced in T-cell hybridomas following activation (Woronicz, J.D., et al. , Nature, 367:277-281 (1994) ; Liu, Z.G., et al. , Nature, 367:281-284 (1994)) . Νur77a expression was induced following activation. However, it was not altered by the addition of 9-cis RA to the cultures.

The data shown indicate that activation-induced expression of FasL is clearly and selectively blocked by the addition of 9-cis RA.

Example 5

Inhibition of the induction of FasL mRNA expression in

9-cis RA-treated 2B4 cells, as measured by the guantitative RNase protection method

The RNA used in this example was prepared as described for Example 4. However, in this example, the very sensitive and quantitative technique of RNase protection was used to determine the levels of FasL mRNA, following activation, in the presence of 9-cis RA.

FasL mRNA expression was determined by RNase protection using RT/PCR-generated DNA probe templates corresponding to the sequences of murine FasL and human g-actin, which has 100% homology with the corresponding murine g-actin sequence. For the construction of the FasL probe, RNA obtained from activated 2B4 cells was 22 reverse transcribed and amplified with primer pairs s p e c i f i c f o r F a s L ( s e n s e 5 ' GGCCGAATTCAGATGGAAGGAGGTCTGTGA and antisense 5' GGCCAAGCTTAACGGCCTCTGTGAGGTAGT) . By design, the resulting cDNAs possessed convenient EcoRI and Hindlll linker ends, which were used for insertion into pGEM4Z (Promega) . A similar procedure was employed to construct the control g-actin probe (sense 5' TTGATGCTGCAGGTCACCAACTGGGACGACATG and antisense 5' AACCCTAAGCTTCGCAGCTCGTTGTAGAAGG) . The probes were sequenced to ensure identity with target sequences. a-³²P-UTP-labeled RNA transcripts were prepared with a kit (Ambion, MAXIscript™) using T7 RNA polymerase .

RNase protections were performed on total RNA samples according to the kit (RPA II™, Ambion) manufacturer's instructions. The samples were resolved on 8 M urea/6% polyacrylamide sequencing gels, the gels were dried and exposed to Kodak Biomax AR™ film. The developed films were scanned and analyzed using a PDI (Huntington Station, NY) scanner and Quantity One™ software. Measured by RNase protection (Figure 6) , FasL expression in 2B4 cells was induced greater than 50-fold following anti-CD3 stimulation, and this expression was profoundly inhibited by the presence of 9-cis RA (to 20% of control) .

Example 6

The demonstration of a correlation between the observed inhibition of FasL mRNA expression in 9-cis RA-treated cells with a eguallv profound inhibition of FasL protein expression on the cell surface: For the detection of FasL expression, 2B4 cells were activated with anti-CD3 antibody for 6 hours, which is sufficient time for the expression of FasL (Brunner, T. , et al., Nature, 373:441-444 (1995)) , were removed from culture and washed twice with PBS/1% FCS/0.1% sodium azide. The pellets, containing approximately 1x10^s cells, were resuspended in PBS containing 50 μg/ml normal mouse „_--_.-_._,

PCT/US96/06090

23

IgG, 0.1% sodium azide and 1% FCS, and were incubated for 15 minutes at room temperature to saturate non-specific binding sites.

To detect FasL, a chimeric Fas-Fc protein recognizing FasL expressed on cells following activation was utilized

(Brunner, T., et al . , Nature, 373:441-444 (1995)) . The

Fas-Fc chimeric protein was produced in a baculovirus expression system (Crowe, P.D., et al . , J. Immunol .

Methods , 168:79-89 (1994)) . The details of the construction of the FasFc expression vector have been described elsewhere (Brunner, T., et al . , Nature, 373:441-444 (1995)) . Briefly, the Fas-Fc insert was constructed from the cDΝA encoding the extracellular domain of Fas, ligated to the cDΝA encoding the hinge, CH2 and CH3 domains of human IgGl (Brunner, T., et al . , Nature, 373:441-444 (1995)) . Tn5Bl-4 cells were infected with a baculovirus transfer vector containing the Fas-Fc cDΝA, and the Fas-Fc protein was protein-G purified from cells grown in serum-free medium (Brunner, T., et al . , Nature, 373:441-444 (1995)) .

15 μg of purified Fas-Fc was added as above to the cells suspended in PBS/0.1% sodium azide/l% FCS/normal mouse IgG and incubated for 30 minutes on ice. The cells were washed once in PBS/0.1% sodium azide/l% FCS, incubated 30 minutes with anti-human IgG-biotin, washed, and stained with Streptavidin TRICOLOR (Caltag, San Francisco, CA) or Streptavidin-PerCP (Becton-Dickinson, San Jose, CA) for 30 minutes on ice. The cells were fixed and analyzed by flow cytometry. As shown in Figure 7, 2B4 cells express significant levels of surface FasL following activation, and 9-cis RA inhibits this surface expression.

Thus, activation-induced FasL expression in T cell hybridomas, measured functionally and at the RΝA and protein expression levels, is clearly inhibited by 9-cis RA. Example 7

The use of retinoid-receptor selective retinoids to determine the bifunctional Nature of the inhibition of FasL expression by 9-cis RA. Retinoids are known to mediate their effects through two separate classes of retinoid receptors; the RARs and the RXRs. In the presence of ATRA or 9-cis RA, RAR/RXR heterodimeric complexes regulate gene transcription by binding to retinoic acid response elements (RAREs) (Glass, C.K., Endocr. Rev., 15:391-407 (1994)) . Additionally, in the presence of 9-cis RA, RXRs will form homodimers and activate transcription by binding to retinoid X response elements (RXREs) . Earlier studies showed that 9-cis RA induces activation of both RAREs and RXREs, and is a more potent inhibitor of activation-induced cell death in T-cell hybridomas and thymocytes than is ATRA (which binds RAR only) . This suggested a role for the RXR-mediated pathway in this process. To further explore the involvement of the different retinoid receptor subtypes mediating this process we examined the ability of synthetic retinoids with receptor-selective binding and activation properties to inhibit activation-induced apoptosis and FasL expression, either alone or in combination. All retinoids were synthesized at Ligand Pharmaceuticals. The syntheses of the pan-agonist 9-cis retinoic acid (9-cis RA) and the RAR-selective retinoid TTΝPB ((E)-4-[2-(5,5, 8, 8-tetramethyl-5, 6, 7, 8-tetrahydro- 2-naphthalenyl) -1-propenyl]benzoic acid) have been described in detail elsewhere (Boehm, M.F., et al. , J. Med . Chem . , 37:2930-2941 (1994)) . The synthesis of the RXR-selective retinoid LG100268 is described in copending application U.S. Serial No. 08/141,496, whose disclosure is incorporated herein by reference. All compounds were solvated to 1 mM in DMSO:ethanol.

The relative selectivity of the compounds synthesized are routinely determined using retinoid receptor binding assays and co-transfection assays designed to detect the activation of the retinoid receptor's ability to regulate gene transcription.

Ligand competition-binding assays were performed using receptors prepared with a baculovirus expression system. Receptors were obtained from lysis-extracts (Allegretto, E.A., et al., J. Biol . Chem. , 268:26625-26633 (1993)) of SF21 cells infected with baculovirus transfer vectors expressing hRAR , β and 7 or hRXR a , β and 7. The receptor extracts were incubated for 2 hours at 0°C with

[³H] -labeled 9-cis RA (Boehm, M.F., et al . , J. Med. Chem. ,

37:408-414 (1994)) in the presence or absence of 200-fold excess unlabeled ligand. Specific ligand binding to receptor was determined by a hydroxyapatite assay according to Wecksler and Norman (Wecksler, W.R., et al . , Anal . Biochem. , 92:314-323 (1979)) .

The co-transfection assay to determine ligand activation of receptor transcriptional activity has been described in detail elsewhere (Heyman, R.A. , et al . , Cell, 68:397-406 (1992)) . The retinoid receptor expression vectors used in this example were pRS-hRAR 7 (human RAR 7) (Giguere, V., et al . , Nature, 330:624-629 (1987)) and pRS-hRXR a, (human RXR a) (Mangelsdorf, D.J., et al. , Nature, 345:224-229 (1990)) . The RAR reporter construct used, ΔMTV-TRE-Luc, contains two copies of the TRE-palindromic response element inserted into the basal reporter construct ΔMTV-Luciferase. The RXR reporter construct TK-CRBPII-Luc contains one copy of the DR1 response element from CRBPII (cellular retinol binding protein II) , linked to the herpes simplex virus thymidine kinase (tk) minimal promoter upstream of the luciferase gene.

Briefly, CV-1 cells were seeded into 96-well plates and transiently co-transfected using the calcium phosphate method with 10 ng of either receptor-expression vector, 50 ng of reporter plasmid, and 50 ng of pRS-/3-GAL

(/3-galactosidase) internal control for 6 hours. The cells 26 were washed, incubated in the presence of ligand for an additional 36 hours, and assayed for luciferase and 3-GAL activity as described by Heyman, R.A., et al. , Cell , 68:397-406 (1992) . All determinations were done in triplicate and normalized for transfection efficiency with the /.-GAL internal control.

Table 1 shows the comparative abilities (Kd) of the RAR-selective retinoid agonist, TTNPB, the RXR-selective LG100268 and the RAR/RXR pan-agonist 9-cis RA to bind the different baculovirus-expressed retinoid receptors. The structures of these compounds are shown in Figure 8. Using a competitive binding assay, we have determined that LG100268 binds strongly to RXR α, β and 7, but not to RAR CK, β and 7. Conversely, TTNPB binds with high affinity to RAR α, β and 7, but not at all to RXR a, β and 7. The pan-agonist 9-cis RA binds with high affinity to both RARs and RXRs in this assay. We also determined the ability of the synthetic receptor-selective retinoids to activate RARs and RXRs to mediate retinoid-dependent transcription employing a co-transfection assay with either RAR 7 or RXR α and the appropriate reporter constructs. As shown in Figure 9, the RAR-selective retinoid TTNPB effectively activated RAR-dependent transcription but not RXR-dependent transcriptional responses. Similarly, the RXR-selective LG100268 activated only an RXR-dependent response. 9-cis RA, which binds both RARs and RXRs, activated both retinoid-dependent transcriptional responses. In the experiment shown, expression vectors for RAR 7 and RXR were employed; however, similar results were obtained for the other RAR and RXR subtypes (not shown) , and the data obtained from all receptor subtypes correlates well with the binding data shown in Table 1. TABLE 1. Kd values (nM) of receptor-selective ligands

RAR a RAR β RAR 7 RXR a RXR β RXR 7

9-cis-RA 93 97 148 8 15 14

(pan)

TTNPB 20 39 51 8113 4093 2566

(RAR)

LG100268 10000 10000 10000 3 3 3

(RXR)

**When Kd value > 10,000, default value = 10,000. Table 1 shows the receptor binding selectivity of the compounds used on the RAR and RXR retinoid receptor subfamilies. Ligand receptor binding (Kd, nM) was determined for each ligand-receptor combination in a competitive binding assay using [³H] -labeled 9-cis RA and baculovirus expressed retinoid receptor. Each data point represents the mean of three independent determinations, performed in triplicate.

In this example, the ability of the receptor-selective retinoids to inhibit activation-induced apoptosis was measured directly by TUNEL assay (Figure 10) , as described in Example 1.

2B4 cells were cultured in anti-CD3 coated 6-well microtiter plates in the presence or absence of the indicated compounds . After 12 hours the cells were collected and assayed for DNA fragmentation by TUNEL, with propidium iodide counterstaining. Propidium iodide counterstaining is widely used to facilitate the quantitation of the cell's DNA content, and is very useful as a measure of DNA fragmentation. Cells which have undergone DNA fragmentation contain less than normal amounts of DNA (are said to be hypodiploid) due the loss of the soluble DNA fragments, which can been observed by flow cytometry as a shift to the left (reduction) in 28 fluorescence intensity. Thus, the application of both methods, i.e., staining with propidium iodide together with the TUNEL assay to detect strand breakage, permits a more definitive and sensitive determination of the amount of apoptosis which has occurred. Plotted as FL3 (PI, DNA content) on the Y-axis and FL1 (FITC, TdT positive) on the X-axis, apoptotic cells appear low and to the right (TdT positive and hypodiploid) .

As shown in Figure 10, the RXR-selective compound LG100268 was found to be significantly less potent than 9-cis RA. The numbers (inset) shown represent the percentage of apoptotic cells in each sample using the region marker shown in the dot plots. Similarly, the RAR-selective compound TTNPB was much less effective than the pan agonist 9-cis RA.

Since RARs and RXRs can form functional heterodimers which can be activated by either TTNPB or 9-cis RA, and since TTNPB alone was insufficient, these data suggested that the engagement of both RAR/RXR heterodimer and RXR/RXR homodimer receptor pathways was important in mediating the inhibition. To specifically address this issue, the two receptor-selective ligands were added together. Shown in Figure 10, the combination of the RAR- and RXR selective retinoids together was as effective as 9-cis RA, inhibiting anti-CD3-induced apoptosis in 2B4 cells.

Example 8

The use of retinoid-receptor selective retinoids to demonstrate the requirement for both RAR- and RXR-selective ligands for the inhibition of FasL functional activity by 9-cis RA.

In this example, the relative ability of the two receptor-selective ligands to block the expression of FasL functional activity on 2B4 cells was tested. The assay for FasL functional activity was as described in Example

2. All compounds were added at a concentration of 1 μM final. As shown in Figure 11, neither the RAR-selective compound TTNPB nor the RXR-selective compound LG100268 were effective inhibitors of the expression of FasL functional activity. When mixed together, however, they were as effective as the pan agonist 9-cis RA, showing that both RAR and RXR receptor pathways must be engaged for complete and effective inhibition of FasL expression.

Example 9

In the mixing experiments described above, 1 μM TTPNB was added with 1 μM LG100268 for a final retinoid concentration in the cultures of 2 μM. To unequivocally demonstrate the requirement in this system for the presence of both RAR- and RXR- ligand binding, we performed a series of ligand titrations. DNA fragmentation and apoptosis were determined using the radioactive label method described in Example 2. As shown in Figure 12, the addition of progressively higher

(10-fold) concentrations of either TTNPB or LG100268 alone produced only minor decreases in DNA fragment tion. Adding both ligands together, however, resulted in an inhibition of DNA fragmentation equal to or greater than the pan-agonist 9-cis RA.

These data demonstrate clearly that both RAR- and RXR-selective ligands must be present to effectively block activation-induced cell death in 2B4 cells.

Example 10

The demonstration of a correlation between the observed bifunctional Nature of the inhibition of FasL expression by 9-cis RA with the inhibition of FasL mRNA expression: The detection of FasL mRNA expression by RNase protection was performed as described in Example 4.

The data shown in Figures 10, 11 and 12 indicate that both RAR and RXR receptor pathways must be engaged in order for inhibition of FasL function to be effective. RNase protection was performed to determine the relative levels of both FasL mRNA, and Fas mRNA, which does not appear to be regulated by retinoids (not shown) . As for

Figure 13, the data shown is normalized relative to a

7-actin internal control. The data shown correlates with the direct (Figures 10 and 12) functional assays (Figure

11) and demonstrates that maximum inhibition of FasL expression requires the presence of both RAR-selective

(TTNPB) and RXR-selective (LG100268) ligands.

Furthermore, retinoids do not block the expression of Fas mRNA, as detected by RNase protection, which correlates well with flow cytometric data showing no significant change in Fas surface protein levels (not shown) .

Example 11

Demonstration of the non-immunosuppressive Nature of the regulation of FasL expression by 9-cis RA:

An important aspect of any method or treatment designed to modify responses in T cells, as well as other cellular components of the immune system, is that the described method or treatment not be globally suppressive. In addition to undergoing Gl/S cell cycle arrest and apoptosis, T-cell hybridomas activated via the TCR produce IL2 (Achier, D.S., et al . , J. Immunol . , 143:3461-3469 (1989) ; Ashwell, J.D., et al. , J. Exp . Med. , 165:173-194 (1987) ; Mercep, M. , et al . , J. I-nmunol. , 142:4085-4092 (1989) ; Mercep, M. , et al. , J. Immunol., 140:324-335

(1988) ) , and the production of this lymphokine is often used as an indicator that activation has taken indeed place (Shi, Y. , et al . , Science, 257:212-214 (1992) ;

Bissonnette, R.P., et al. , J. Exp . Med. , 180:2413-2418 (1994)) . To demonstrate the non-immunosuppressive Nature of the ability of 9-cis RA to block FasL expression and as a result, activation-induced cell death, we performed determinations of this particular measure of T cell activation, i.e., the production and secretion of the cytokine interleukin-2 (IL2) . Supernatants were assayed for IL2 activity as described by Gillis, S., et al . , J. Immunol . , 120:2027-2032 (1978) . Briefly, tissue culture super¬ natants were obtained from 2B4 cells which had been stimulated to undergo apoptosis in the presence or absence of 9-cis RA, as well as the two receptor-selective ligands TTNPB and LG100268. The collected supernatants were titrated into cultures of the IL2-dependent cell line CTLL-2, which had been starved of IL2 for 4 hours prior to the addition of the IL2-containing supernatants. The cells were cultured 16 hours and then pulsed for 6 hours with 5 μCi/ml of [³H] -TdR. The labeled DNA was harvested and the amount of label incorporated determined. Units of IL2 per ml were calculated using a standard curve obtained with recombinant IL2 and linear regression analysis.

Previous reports have suggested that while retinoids block activation-induced apoptosis, they do not interfere with the corresponding cell cycle arrest or production of IL2. As shown in Figure 14, the presence of the retinoids had little if any effect, only partially inhibiting IL2 production if at all. This is consistent with the notion that retinoids are not globally immunosuppressive, having only minor inhibitory effects on various measures of T cell activation. Therefore, it has now been determined and shown that 9-cis RA, which is known to block the process of activation-induced apoptosis in T-cell hybridomas (Figure 1) does so by the mechanism of inhibition of FasL expression following activation of the cell. This was demonstrated by functional assay (Figures 2, 3 and 4) , analysis of FasL mRNA expression (Figures 5 and 6) , and by the detection of the cell-surface FasL protein (Figure 7) .

It has also been determined and demonstrated that ligand activation of both the RAR and RXR pathways is required to achieve a maximal response. This is further supported by the demonstration that the addition of both a RAR- and a RXR-selective compound added simultaneously can mimic the pan-agonist 9-cis RA (Figures 10, 11, and 12) .

Activation-induced apoptosis in normal T lymphocytes is known to occur under various conditions in vivo . Stimulation of immature thymocytes via the T-cell receptor results in apoptosis and this is an important mechanism of negative selection during T-cell development. Another form of activation-induced activation-induced apoptosis occurs in mature T-cells in vivo and in vi tro, following expansion of antigen-specific T-cells, and this homeostatic process has been called peripheral deletion. This process is mediated by Fas/FasL interaction (Russell, J.H., et al., Proc . Na tl . Acad . Sci . U. S . A . , 90:4409-4413 (1993) . Activation-induced apoptosis in T-cells also occurs under pathological conditions, and it has been implicated as a mechanism for the loss of both CD4+ and CD8+ T cells in AIDS. Finally, there are several pathological conditions where Fas and FasL have been implicated. One example of such a condition is the disappearance of hepatocytes during chronic Hepatitis C infections. It has been shown hepatocytes express Fas in both normal (Ogasawara, J., et al . , Nature 364:806-809

(1993)) and pathological (Hiromatsu, Ν., et al . ,

Hepatology, 19:1354-1359 (1994)) settings. It has also been demonstrated that lymphocytes infiltrating the liver of such patients express FasL (Mita, E., et al . , Biocheπi. Biophys . Res . Commun . , 204:468-474 (1994)) . The implication here is that T-cells express FasL as a result of chronic stimulation due to persistent infection, thus interact with Fas expressed on hepatocytes to induce apoptosis in the liver cells, thus causing the massive destruction of liver cells associated with this disease.

Another example of the potential role for FasL and thus its regulation in disease can be found in several inflammatory conditions where there are skin lesions which have been characterized by apoptosis in keratinocytes. The common factors associated with these conditions are 33 lesions, the expression of both Fas and the cell:cell adhesion molecule ICAM-1 on keratinocytes associated with these lesions, and a lymphoid component in the pathology of the condition. These include but are not restricted to lesions caused by Lichenoid drug eruption, Herpes zoster virus (varicella zoster virus) , Erythema Multiforme, and Contact dermatitis (Sayama, K. , et al . , J. Invest . Derma tol . , 103:330-334 (1994)) . It is therefore concluded that infiltrating lymphocytes responding to the inflammation secrete Interferon-gamma (IFN-g) , and express both FasL and the cell:cell adhesion molecule LFA-1, which binds to ICAM-1. Keratinocytes respond by expressing ICAM-1 and Fas. The ICAM-1:LFA-1 interaction facilitates contact which then results in a Fas/FasL interaction leading to apoptosis.

It has been found that following hypoxia, cardiomyocytes express Fas and undergo apoptosis (Tanaka, M., et al., Circ. Res . , 75:426-433 (1994)) . Although a lymphoid (i.e., FasL) component has not been identified, most cells expressing Fas become receptors for FasL and thus targets for Fas/FasL induced apoptosis. Another example involves the normal, physiological cell death

(apoptosis) of ovarian follicles and oocytes, a phenomenon known as atresia. Apoptotic cell death of follicles and hyperovulated oocytes correlated with the expression of Fas, and again, although a lymphoid (i.e., FasL) component has not been identified, cells expressing Fas are potential targets for Fas/FasL induced apoptosis, and thus a another circumstance where retinoid regulation of FasL expression might conceivably be useful. (Guo, M.W. , et al . , Biochem . Biophys . Res . Commun . , 203:1438-1446 (1994) ) .

Accordingly, the selective blocking of the expression of Fas-Ligand following activation by use of the retinoid compounds of the type discussed herein can be expected to be useful in inhibiting T-cell receptor-induced Fas/FasL-mediated apoptosis and its effects in patients. 34

While the preferred embodiments have been described and illustrated above, various substitutions and modifications may be made thereto without departing from the scope of the invention. Accordingly, it is to be understood that the present invention has been described by way of illustration and not limitation.

Claims

35We claim:

1. A method for inhibiting the expression by mammalian cells of Fas-Ligand by contacting said cells with a retinoid compound which modulates both Retinoid X Receptors and Retinoic Acid Receptors.

2. The method of claim 1 wherein said retinoid compound is 9-cis retinoic acid.

3. A method for inhibiting the expression by mammalian cells of Fas-Ligand by contacting said cells with a retinoid compound which modulates Retinoid Receptors.

4. A method for inhibiting the expression by mammalian cells of Fas-Ligand by contacting said cells with a retinoid compound which activates both Retinoid X Receptors and Retinoic Acid Receptors.

5. A method for inhibiting apoptosis in T-cell hybridomas by contacting said T-cell hybridomas with a retinoid compound capable of blocking the expression of Fas-Ligand following activation.

6. A method for inhibiting apoptosis in mammalian cells where said apoptosis is mediated by the interaction between Fas expressed on said cells and FasL expressed by T cells by contacting said T cells with a retinoid compound capable of blocking the expression of Fas-Ligand following activation.

7. A method of screening compounds to select compounds having the ability to inhibit activation-induced apoptosis in T cells comprising the identification of retinoid compounds capable of blocking the expression of Fas-Ligand in T-cell hybridoma following activation. 36

8. A method of screening compounds to select compounds having the ability to inhibit the induction of apoptosis, wherein the apoptosis is the result of cells expressing Fas contacting Fas-Ligand expressed by the same or other cells, comprising the identification of retinoid compounds capable of blocking the expression of Fas-Ligand in said cells expressing Fas-Ligand either idiopathically or following activation.