AGN193109 is a highly effective antagonist of retinoid action in human ectocervical epithelial cells.

Retinoids are important physiological agents that regulate epithelial cell differentiation and proliferation. The importance of these agents in regulating growth, development, and differentiation has led to a search for new retinoid agonists and antagonists. In the present manuscript we show that AGN193109, a retinoid analog, is an efficient antagonist of retinoid action in human cervical epithelial cells. Treatment of ECE16-1 cells with natural or synthetic retinoids reduces cytokeratin K5, K6, K14, K16, and K17 levels, increases cytokeratin K7, K8, and K19 levels, increases retinoic acid receptor-β (RARβ) mRNA levels, suppresses proliferation, and alters cell morphology. Co-treatment with AGN193109 prevents these responses. Half-maximal and maximal antagonism is observed at a molar ratio of AGN193109:retinoid agonist of 1:1 and 10:1, respectively. When administered alone AGN193109 has no agonist activity. Thus, AGN193109, which binds to RARα, RARβ, and RAR with K values = 2, 2, and 3 nM, respectively, but is unable to bind to the retinoid X receptors, is a highly active antagonist of retinoid action in ECE16-1 cells.

In the present study we describe the effects of a new retinoid, AGN193109, a specific and highly effective antagonist of RAR function (40), on the differentiation and proliferation of the ECE16-1 human cervical cancer cell line. ECE16-1 cells are an human papillomavirus type 16-immortalized cell line that was generated by transfection of normal human cervical cells with the molecularly cloned human papillomavirus type 16 virus (22). They are a model for study of the early stages in cervical dysplasia and retinoid effects on cervical cell differentiation. These cells respond to retinoid stimulation with characteristic changes in proliferation rate, morphology, and marker gene expression (13,(22)(23)(24)(25). The predominant receptor forms expressed in ECE16-1 cells are RAR␣, RXR␣, and RAR␥; however, RAR␤ is induced by retinoid treatment (37).
In the present study we demonstrate that AGN193109 is a highly active antagonist of retinoid action in ECE16-1 cells. We show that it is an efficient antagonist of retinoid-dependent changes in ECE16-1 cell morphology, gene expression, and proliferation, but has no intrinsic agonist activity.
Cell Proliferation Studies-ECE16-1 cell proliferation studies were performed exactly as described previously (23,25,26). Cells (10,000/ cm 2 ) were seeded in complete medium and allowed to attach overnight. The cells were then shifted to DM, allowed to equilibrate for 24 h, and treatment was initiated by addition of fresh DM or DM containing EGF or retinoid as indicated. After 3 days of daily treatment with retinoid, the cells were harvested with 0.025% trypsin, 1 mM EDTA, fixed in isotonic buffer containing 4% formaldehyde, and counted using a Coulter counter (26).

Retinoid Effects on Cell Morphology-
AGN193109 Antagonizes Retinoid-dependent Growth Suppression-ECE16-1 cells remain quiescent in defined medium ( Fig. 2A, open circle) (25,26). However, they proliferate rapidly in the presence of epidermal growth factor (filled square) ( Fig.  2A) (25,26,37) and addition of 10 nM TTNPB suppresses this proliferation. Addition of increasing concentrations of AGN193109 in the presence of 10 nM TTNPB prevents the TTNPB-dependent suppression of proliferation (filled circles). The growth suppression is half-reversed by 10 nM and completely reversed by 100 nM AGN193109. Although it can antagonize the effects of TTNPB, 1 uM AGN193109 given alone has no effect on cell proliferation (filled triangle).
The ability of AGN193109 to antagonize the effects of TT-NPB could be related to the fact that TTNPB is an RAR-specific ligand that does not interact with RXR receptors (Table I). We therefore determined whether AGN193109 could reverse the effects of tRA, 13-cRA, and 9-cRA. These retinoids interact with RAR or both RAR and RXR receptors ( Table I). As with TTNPB, AGN193109, when present at a 10-fold molar excess, completely reverses the growth suppressive effects of tRA, 13-cRA, and 9-cRA (Fig. 2B).
AGN193109 Effects on Biochemical Markers of Retinoid Action-We next determined whether AGN193109 could antagonize retinoid effects on expression of retinoid-responsive marker genes. Cytokeratin proteins form the intermediate filaments in epithelial cells and are important markers of retinoid action in cervical epithelial cells (22,(33)(34)(35). As shown in Fig. 3A, ECE16-1 cells grown in retinoid-free medium express cytokeratins K5, K6, K7, K8, K13, K14, K16, K17, and K19 (A) (22). Treatment with 10 nM TTNPB suppresses K5, K6, K14, K16, and K17 levels and increases K7, K8, and K19 levels (Fig.  3B). Simultaneous treatment with 100 nM AGN193109 eliminates the effects of 10 nM TTNPB (Fig. 3C). No changes in keratin expression are observed when cells were incubated with 0.01-1000 nM AGN193109 alone (not shown). We have shown previously that changes in keratin mRNA level mediate changes in keratin protein level (13,22,35). As a representative keratin, we show the effects of each treatment on the level of mRNA encoding K6 (Fig. 3B). AGN193109 (100 nM) has no effect and 10 nM TTNPB suppresses K6 mRNA levels. The TTNPB-dependent suppression is completely inhibited when cells are incubated with 10 nM TTNPB plus a 10-fold molar excess of AGN193109. AGN193109 also inhibits the retinoiddependent change in the level of mRNA encoding the other keratins (not shown).
Regulation of RAR␤ mRNA Levels-We have shown previously that RAR␤ mRNA levels are increased by retinoids (37). The results in Fig. 3B show that TTNPB increases RAR␤ mRNA levels, that the increase is completely antagonized by AGN193109 and that AGN193109 has no intrinsic retinoid activity in regulating this marker. within and among cell lines (29,37). In ECE16-1 cells, RARspecific ligands are more active retinoids than RXR-specific ligands (37). In the present study, we examined the effects of a novel synthetic retinoid, AGN193109, on a variety of retinoiddependent responses in ECE16-1 cells. We found that AGN193109 is able to eliminate retinoid-dependent changes in cell morphology and prevent retinoid-dependent suppression of proliferation. Moreover, at the biochemical level, AGN193109 prevents retinoid-dependent changes in marker gene expression. We examined genes that are increased (K7, K8, K19, RAR␤), as well as those that are suppressed (K5, K6, K14, K16, K17) by retinoid treatment (22). In each case, the retinoid-dependent response was antagonized by AGN193109. Thus, AGN193109 prevents both stimulatory and suppressive responses. Taken together these results suggest that AGN193109 may be a global antagonist of retinoid action in ECE16-1 cells (i.e. it is likely to antagonize most, and perhaps all, retinoiddependent responses).
AGN193109 Is an Efficient Antagonist with No Agonist Activity-Several problems can be encountered with receptor antagonists. First, they may have partial agonist activity. This can complicate interpretation of their effects in in vivo systems. To determine whether AGN193109 has intrinsic agonist or antagonist activity, we evaluated its effects on ECE16-1 cell function in the absence of added retinoid agonist. Concentrations of AGN193109 ranging from 0.01 to 1000 nM produced no detectable changes in cell morphology, cell proliferation, or retinoid-responsive marker gene expression. Based on these results we conclude (i) that AGN193109 is a highly effective antagonist of retinoid action in these cells and (ii) that it has not intrinsic agonist activity.
Second, antagonists may have a low affinity for the target receptor. Such ligands must be used at extremely high levels to accomplish the antagonism. AGN193109, however, has an affinity for the RARs of 2-3 nM. Considering that the receptor affinities of tRA, 13-cRA, and 9-cRA range from 7 to 141 nM and TTNPB has affinities ranging from 20 to 51 nM, AGN193109 is a high affinity ligand. This is reflected in its potency at preventing biological end responses. The compound is maximally effective at inhibiting retinoid agonist activity when present at a 10-fold molar excess and is half-maximally effective when present in equal molar amounts. Thus, AGN193109 is a high affinity antagonist.
Natural Ligands That Interact with RAR and RXR Receptors-Our initial studies showed that the synthetic retinoid, AGN193109, which interacts solely with RAR receptors, is effective at antagonizing the effects of the RAR-specific retinoid agonist, TTNPB. However, it was unclear whether the activity of natural ligands would be inhibited by AGN193109. We therefore tested the ability of AGN193109 to antagonize the effects of tRA, 13-cRA, and 9-cRA. 13-cRA, 9-cRA, and tRA interact with RARs, while 9-cRA also has a high affinity for RXRs. The results show that AGN193109 is an effective antagonist of the activity of each of these ligands. Thus, although AGN193109 only interacts with the RAR forms, it can also antagonize the effects of an agent (9-cRA) that interacts with relatively high affinity with both RARs and RXRs. Since AGN193109 does not interact with the RXRs, it is not likely that it directly prevents 9-cRA from interacting with RXR receptor. Moreover, RXR-specific ligands are relatively inactive in ECE16-1 cells (37). Since 9-cRA can efficiently bind to RARs, it is likely that AGN193109 is inhibiting 9-cRA effects mediated via RARs.
Receptor antagonists can be extremely useful ligands, both for understanding mechanism of action and for designing new therapies. In the present manuscript, we show in ECE16-1 cells, a human ectocervical epithelial cell line, that AGN193109 FIG. 3. AGN193109 regulation of retinoid-responsive marker gene expression. In A, ECE16-1 cells were grown in DM containing 10 ng/ml EGF (DM/EGF) (A) or DM/EGF supplemented with 10 nM TT-NPB (B) or 10 nM TTNPB ϩ 100 nM AGN193109 (C) for 3 days. The cells were labeled with [ 35 S]methionine for 24 h prior to harvest for preparation of cytokeratins. Equal numbers of radioactive counts from each sample were electrophoresed in nonequilibrium pH gradient electrophoresis gels as described previously (22). The second dimension was a 6% acrylamide gel. In B, poly(A) ϩ RNA was prepared from cells treated for 3 days with the indicated agent under the same conditions as in A. The RNA was transferred to Biodyne A membrane and hybridized with 32 P-labeled cDNAs encoding cytokeratin K6, RAR␤, or glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Glyceraldehyde-3-phosphate dehydrogenase is included as a control, as its level is not regulated by retinoids.  (26,44). Asterisks indicate a K d for receptor of Ͼ10,000 nM and double asterisks indicate K d for receptor of Ͼ1,000 nM. is an extremely active high affinity antagonist of retinoid action that appears to have no intrinsic retinoid agonist activity.