Applying Cell Based Screening to Telomerase

Telomerase presents a number of opportunities for cell-based assay development. At the level of transcription, signaling pathways feeding either the hTERC or hTERT promoters, or the transcription factors that bind them, are excellent candidates for inhibition. Reporter assays of promoter activity are the most frequently used method to rapidly test candidate factors that influence gene promoter regions and are readily scalable (Kiss-Toth et al. 2000, Kiss-Toth et al. 2006). In a reporter screen for de-repressors of hTERT promoter activity in normal cells, Won and co-workers identified the molecule CGK1026 from a library of 20,000 compounds (Won et al. 2004). Promoter deletion and mutagenesis identified a CGK1026-responsive non-canonical E2F binding site in the hTERT promoter. CGK1026 was shown by chromatin immunoprecipitation (ChIP) to inhibit recruitment of histone deacetylases 1 and 2 to E2F-pocket protein complexes at the hTERT promoter. In the case of hTERC, the JNK inhibitor SP600125 was shown to increase hTERC-reporter activity and endogenous hTERC levels in several cell lines. Mutation of Sp1 sites in the hTERC promoter attenuated induction of reporter activity by the drug, whereas co-transfection of MEKK1 with Sp3 synergistically repressed promoter activity. It was shown by ChIP that SP600125 caused a switch in the ratio of Sp1/Sp3 binding to the endogenous promoter (Bilsland et al. 2006).

Such data validate the reporter assay approach to cell-based screening for novel small molecules with potential to regulate telomerase. Benefits of this approach are the speed and ease of the assay systems, as well as the ability to dissect pathways at the level of transcription factor binding via promoter mutagenesis. It may also be possible to tentatively identify the candidate target pathways and place novel inhibitors within them using complementary approaches such as RNAi or overexpression if some upstream regulatory factors are known. Importantly, however, no prior knowledge of the exact target pathway is required, since hits scored by this approach will target pathways that are intrinsically draggable (Fig. 13.4). Further, given the close relationship between telomerase expression and prosurvival signaling, the telomerase promoters can be regarded as probes for both known and unknown cell-survival and cell-death pathways, in the sense that it may not even be necessary to use leads identified in this way as classic telomerase inhibitors since they may have more rapidly acting anticancer properties mediated through inhibition of key survival pathways (Horikawa and Barrett 2003). However, target identification is not necessarily easy, and to demonstrate specificity it will be essential to perform adequate secondary assays, ideally employing normal control cells.

Although core telomerase activity can be reconstituted in vitro with the hTERC and hTERT components alone, telomerase is a multiprotein complex in

Fig. 13.4 Screening for novel inhibitors of telomerase gene expression. Several small molecule inhibitors of EGFR/MAPK signaling at various stages of development have been demonstrated to inhibit hTERT expression in cell lines. Although several pro- and anti-survival pathways are known to regulate the hTERC and hTERT promoters, the mechanisms underlying transcrip-tional regulation of the genes is incomplete. Therefore, a cell-based assay approach using the cloned promoters as probes could identify novel small molecule inhibitors of both known and unknown pathways

Fig. 13.4 Screening for novel inhibitors of telomerase gene expression. Several small molecule inhibitors of EGFR/MAPK signaling at various stages of development have been demonstrated to inhibit hTERT expression in cell lines. Although several pro- and anti-survival pathways are known to regulate the hTERC and hTERT promoters, the mechanisms underlying transcrip-tional regulation of the genes is incomplete. Therefore, a cell-based assay approach using the cloned promoters as probes could identify novel small molecule inhibitors of both known and unknown pathways vivo, and its activity is modulated by interaction partners such as the chaperone protein hsp90 (Akalin et al. 2001, Forsythe et al. 2001, Holt et al. 1999). It is possible that hsp90 may be required for telomerase stability or for mediating conformational changes allowing the holoenzyme to translocate after synthesis of each individual telomere repeat. Interfering with hsp90 activity using the inhibitors geldanamycin, 17-AAG, and novobiocin inhibits telomerase activity and increases chemosensitivity in melanoma and head and neck cancer cells, providing experimental support for the functional importance of the interaction (Chang et al.

2006, Villa et al. 2003). 17-AAG has been tested in several phase I clinical trials in patients with advanced cancers refractory to standard therapies (Goetz et al. 2005, Grem et al. 2005, Nowakowski et al. 2006). However, hsp90 has multiple client proteins in both normal and cancer cells, and blanket inhibition of hsp90 may give rise to toxicity in normal cells. In contrast, selectively targeting the hTERT-hsp90 interaction or other functionally important hTERT interactions would be predicted to have a highly specific effect. Although protein-protein interactions have conventionally been considered difficult to drug with small molecules, there are noteworthy exceptions such as the nutlins, which inhibit the p53-MDM2 interaction, or FJ9, which inhibits the interaction between the PDZ domain of dishevelled and the Frizzled-7 receptor on the canonical Wnt pathway (Fujii et al. 2007, Vassilev et al. 2004). Protein-protein interaction assays such as the mammalian-2-hybrid or reporter complementation assays are amenable to up-scaling for screening purposes and may present a useful platform to screen for molecules that interfere with the integrity of the telomerase holoenzyme (Park et al.

2007, Paulmurugan and Gambhir 2003, Zhao et al. 2004).

Subcellular targeting has also recently emerged as a mechanism regulating telomerase activity (Akiyama et al. 2003b, Seimiya et al. 2000, Wong et al. 2002). Fluorescence tagging, either with antibodies, small molecules, or creation of fusion proteins, is central in studies of subcellular localization. Assay development is therefore dependent on the availability of tagging reagents that will not disrupt natural behavior of the target protein. In an elegant study employing a GFP-hTERT fusion protein, Wong and co-workers demonstrated transformation and DNA damage induced subnuclear shuttling of hTERT. In primary cells, ectopically expressed nonsaturating concentrations of GFP-hTERT localized primarily in nucleoli during G1/S but were excluded from nucleoli in late S/G2. In contrast, cancer cells showed constitutive nucleolar exclusion of the fusion, except in response to DNA damage, whereupon transient relocalization was observed (Wong et al. 2002). The authors speculate that nucleolar localization may sequester functional telomerase holoen-zyme from telomere ends. Importantly, the fusion was demonstrated to be functional, as shown by reconstitution of telomerase activity and telomere maintenance in hTERT-negative cells, providing a potential screening platform to identify small molecules that interfere with pathways involved in hTERT nucleolar exclusion in cancer cells.

Screens based on fluorescent imaging of subcellular relocalization have previously been successful (Kau et al. 2003). Fluorescent imaging screens have the advantage of being intrinsically "high-content" in that multiple phenotypic parameters can simultaneously be assessed (Eggert and Mitchison 2006). However, visual interpretation of data has sometimes been necessary, and automated quantitative image analysis and data extraction have conventionally been important bottlenecks for high-throughput application of imaging screens. Although a number of commercial image processing packages have been used, availability of versatile and open-source image cytometry platforms such as CellProfiler and ImageJ is an important development (Carpenter et al. 2006, Rajwa et al. 2004). The ability to rapidly extract multiple diverse quantitative measurements such as cell number, shape, size, cell-cycle status, and the levels or location of fluorescent tag, is essential to support a variety of phenotypic screens.

Tankyrase also presents an intriguing target for inhibitor development. Tankyrase is a poly(ADP-ribose) polymerase (PARP) present in the nucleus and at other sites in the cell which associates with the telomere-binding protein TRF1 and mediates its poly(ADP-ribosyl)ation. TRF1 is a specific negative regulator of telomere length and its poly(ADP-ribosyl)ation by tankyrase leads to loss of telomere association and subsequent degradation by the ubiquitin/proteasome pathway (Chang et al. 2003, Smith et al. 1998). Therefore, tankyrase in the nucleus acts as a positive regulator of telomere length. Overexpression of nuclear localization signal-tagged tankyrase is able to confer resistance to telomerase inhibitors by blocking telomere shortening in several cancer cell lines (Seimiya et al. 2005). Following long-term treatment with the synthetic telomerase inhibitor MST-312, cells overexpressing tankyrase showed stabilized shortened telomere lengths, whereas cells treated with both the telomerase inhibitor and a conventional PARP inhibitor overcame the block and reached telomere-dependent crisis at significantly earlier time points than those treated with the telomerase inhibitor alone. Interestingly, fluorescent immunostaining showed that TRF1 expression was not observed in NLS-tankyrase expressing cells, but co-localization was observed on administration of a PARP inhibitor (Seimiya et al. 2005). Since cells expressing NLS-tagged tankyrase show significantly reduced TRF1 expression, a screening assay to identify selective tankyrase inhibitors based on TRF1 levels using the "in-cell western" technique can be readily envisaged (Wong 2004). Alternatively, co-localization of tankyrase and TRF1 in the presence of a PARP inhibitor suggests that a two-color cell imaging or FRET approach could be applied.

0 0

Post a comment