Use of Transcription Factors as Agents and Targets for Drug Development

Cancer is a genetic disease wherein mutations of growth regulatorygenes result in abnormal proliferative capacity, recognized clinicallyas the occurrence of a malignant tumor. Transcription factorsgovern the expression of genes, be they "housekeeping"or regulatory. These factors organize the first crucial step inestablishing the function of the gene, namely, the transcriptionof information in DNA into messenger RNA (mRNA). Translation ofmRNA results in the synthesis of the oncogenic protein. Hence,the design of therapeutic agents targeted at transcription factorsregulating the initial flow of "bad" information from"damaged" genes should be the ultimate goal of effortsto develop new weapons in the therapeutic armamentarium of theoncologist and, indeed, the general internist.

Ideally, recognition of abnormal transcription factor activityleading to the expression of "proneoplastic" genes couldallow introduction of specific therapy targeted at the abnormaltranscription factor. The action of such a therapy would quietdown the unruly gene, and therefore, avert the neoplastic catastrophelying in wait for the patient. Alternatively, a critical transcriptionfactor related to sustaining the actual growth of the tumor couldbe counteracted, with the resultant death of the established tumor.

Are these goals "pie in the sky"? On one level, no.Every oncologist who prescribes tamoxifen (Nolvadex) for metastaticbreast cancer, uses leuprolide (Lupron) to treat metastatic prostatecancer, or considers retinoid therapy for patients who have survivedan initial upper aerodigestive tract cancer is manipulating transcriptionfactors as the initial step in the efficacy of those respectivetherapies. Indeed, measurement of the estrogen-receptor contentin a particular patient's tumor is a recognized predictor of theefficacy of tamoxifen treatment,[1] and the estrogen receptoris a member of but one of the families of therapeutically relevanttranscription factors reviewed by Smith and Birrer. The criticalquestions are, how to generalize from these examples to the varietyof transcription factors affecting other neoplasms, and how toincrease the efficacy of that approach? And these questions leadto a consideration of a number of problems surrounding the developmentof such agents.

Issues in Developing Anti-Transcription Factor Agents

One major unknown is how "neoplastic cell" transcriptionfactors differ from "normal cell" transcription factors.As Smith and Birrer point out, the sharing of a tumor's growth-regulatorypathways with those of normal cells from the tissue of originraises the concern that the ultimate therapeutic index for suchan approach may be narrow. However, this issue has not been clearlyor accurately investigated in the vast majority of human tumors.In fact, there is hope from recent studies in hematopoietic neoplasmsthat there may, indeed, be very tumor-specific rearrangementsof transcription factors.

For example, as reviewed by Gauwerky and Croce,[2] the t(1:19)(q23;p13)chromosomal translocation present in 30% of pediatric B-cell acutelymphocytic leukemia (B-ALL) apparently encodes a chimeric transcriptionfactor that joins the effector domains of a transcription factorregulating immunoglobulin gene expression with the DNA-bindingmotif of a homeotic gene. The product resulting from this "monstergene" is clearly like nothing else found in the organismand would be a potentially useful point of departure for consideringagents to interdict its function. This circumstance thereforehighlights the need to better understand how transcription factorsand their complexes in epithelial neoplastic cells actually differfrom "normal." Workers investigating the common epithelialneoplasms have not paid attention to this issue with as much clarityas might be useful.

A second general problem is the utility of current models in supportingthe development of anti-transcription-factor-related therapies.In general, the pharmaceutical industry is not very interestedin antineoplastic agents that do not cause overt shrinkage ofthe usual repertoire of established human tumor xenograft models.Given our understanding of the biology of how transcription factorswould be expected to promote neoplastic growth, these models probablyare not at all relevant to the development of anti-transcriptionfactor-directed therapeutics.

At best, anti-transcription factor therapies may be cytostaticor differentiating, rather than overtly cytotoxic. At worst, ifany useful effect is to emerge, such agents may need to be usedearly in a tumor's ontogeny. This circumstance would call fora means to detect the presence of a "protoneoplastic"population of cells to identify patients in whom the efficacyof such therapies could be demonstrated. This state of affairswould entail challenges to the normal developmental process forsuch drugs, both preclinically and in clinical trials. Therefore,efforts to develop anti-transcription factor-directed therapiesfor the common adult neoplasms must go hand in glove with effortsto define the role of transcription factors in tumor pathogenesis.Only then will models be defined that will clearly delineate thepotential of such agents.

How to Discover Anti-Transcription Factor Drugs

Smith and Birrer describe several potential modes of screeningfor anti-transcription factor drugs, including: (1) examinationof the interdictors of "marker gene" activity, suchas chloramphenical acetyl transferase (CAT) assays in engineeredcell lines; (2) modulation of the kinases and phosphatases thatwould be expected to regulate the activity of the transcriptionfactor complex; and (3) efforts to design molecules related tothe structure of the transcription factor. In my opinion, it isthe last generic type of effort that should be focused on withintensity. We must define agents that can interact with high affinitywith the structural domains of transcription factors importantin promoting complexing with either other transcription factorsor specific DNA sequences.

The authors gloss over the realities of delivering large molecules,such as transcription factors or even fragments of transcriptionfactors, into cells. Rather, the true value of the "dominant-negative"technology described in the article and practiced so elegantlyby Dr. Birrer's laboratory is to define the important moleculesand thus help find small molecules that may act like the dominant-negativemutants. In this way, the dominant-negative approach describedby Drs. Smith and Birrer will not only prove to be the Rosettastone for understanding the basis of the action of transcriptionfactors but also will lead to the building of Trojan horses toabrogate their tumor-promoting action.

References:

1. Allegra J, Lippman M, Thompson E, et al: Estrogen receptorstatus: An important variable in predicting response to endocrinetherapy in metastatic breast cancer. Eur J Cancer 16:323-331,1980.

2. Gauwerky CE, Croce CM: Molecular genetics and cytogeneticsof hematopoietic malignancies, in Mendelsohn J, Howley P, IsraelM, et al (eds): The Molecular Basis of Cancer, pp 18-37. Philadelphia,WB Saunders, 1995.