Norman E. Sharpless | Department of Genetics

Research Interests

The major focus of my lab is to understand the role and regulation of the products of the INK4a/ARF locus, p16INK4a and p14ARF, in human cancer. Tumorigenesis is a problem specific to long-lived metazoans. From this point of view, however, mammals and many lower vertebrate species such as Xenopus appear to contain a serious design flaw. Namely, two of the most important tumor suppressor proteins, p16INK4a and p14ARF, are encoded in overlapping open reading frames of the same 3x104 bp stretch of DNA in the 3x109 bp genome. These ORFs share second and third exons, but have different first exons (1a and 1b) which splice into the common second exon in alternate reading frames (Fig 1). Therefore, p16INK4a and p14ARF are not isoforms, and share no amino acid homology. While the encoding of ORFs in overlapping stretches of the genome is common in viruses and bacteria, this genomic arrangement is almost unique in vertebrate species. Moreover, it is seemingly exactly the wrong arrangement for tumor suppressor genes, as cancer is a disease characterized by the wholesale amplification and deletion of large stretches of the genome. Therefore, two of the most important tumor suppressor proteins can be inactivated by a common genetic lesion of the INK4a/ARF locus. Indeed, this does happen with alarming frequency, as loss of 9p21 (harboring INK4a/ARF) is one of, if not the, most common cytogenetic events detected in human cancer. How p16INK4a and p14ARF became arranged in this way, and why this arrangement has persisted for hundreds of millions of years, is one of the great mysteries of cancer genetics.

These proteins function in distinct anti-cancer pathways: p16INK4a regulates Rb and p14ARF regulates p53. Loss of p16INK4a and/or p14ARF is rivaled only by p53 mutation in frequency in human cancer, and is seen in a wide variety of tumor types including melanoma, glioblastoma, certain leukemias, and carcinoma of the lung, oropharynx, bladder, pancreas, colon and breast. Furthermore, in addition to point mutation and deletion, loss of expression of these tumor suppressor proteins is also seen in cancers by a variety of epigenetic mechanisms, for example promoter methylation or transcriptional repression by oncogenic transcription factors such as Twist or BMI. Loss of p16INK4a is often a very early event in neoplastic progression, and humans with a germline inactivation of a single INK4a/ARF allele are predisposed to melanoma, astrocytoma and pancreatic adenocarcinoma. Our previous work in the mouse using targeted alleles specifically lacking only p16INK4a or p19ARF, but with normal function of the remaining Ink4a/Arf product, has established that both proteins play important and distinct roles in mammalian tumor suppression (see Sharpless et al. Nature 2001 and Sharpless et al., Canc. Res 2002). Furthermore, we have begun to assess the relative roles of these proteins in certain specific tumor types such as melanoma, glioblastoma, or lung cancer (See Sharpless Nature 2001, Bachoo, Cancer Cell 2002; also work in press and submission). To date, we have learned that there is potent synergy between inactivation of the p16INK4a—Rb and p14ARF—p53 pathways in tumor induction (for example, Fig 2 shows the effects of p16INK4a and/or p19ARF loss in a mouse melanoma model), as would have been predicted from human cancer genetics.|

Continuing to use the mouse as a model system, my lab is studying three specific areas regarding the INK4a/ARF locus:

Gene regulation: Despite the importance of these tumor suppressors in cancer, little is known about their control in normal tissues and tumors. Furthermore, p16INK4a is inactivated by promoter methylation with high frequency in cancer, yet the molecular biology of this occurrence is poorly understood. Through the use of high-throughput, low background genetic screens, my lab is interested in addressing the regulation of the INK4A/ARF locus under both normal and malignant circumstances.

INK4a/ARF function in normal physiologic processes: While mice lacking either p16INK4a or p19ARF are predominantly normal except for their increased tumor susceptibility, by further characterizing mice with specific deficiencies of either p16INK4a or p19ARF, we are uncovering unexpected functions of these proteins in normal tissues. We are specifically interested in the roles of these proteins in normal physiologic processes such as differentiation, organ self-renewal, metabolism and aging. We are also interested in how these proteins relate to other members of the Rb and p53 pathways, and the tumor phenotypes of compound Rb and p53 pathway mutants.

Tumor modeling to identify cooperating oncogenic events: Concomitant perturbation of the Rb and p53 pathways is not sufficient for tumor formation in mice. Therefore, we are combining oncogenic lesions in tissue-specific models with p16INK4a and/or p19ARF loss in an effort to develop faithful mouse tumor models of lung and colon cancer. Tumors from these models are then further analyzed by genome-wide methods such as RNA expression profiling and array comparative genomic hybridization (aCGH) to identify cooperating events in tumorigenesis.

The goal of these projects is to better understand how these two canonical tumor suppressor proteins function to prevent cancer, with the hope that this increased understanding will translate into improved diagnosis or therapy of this important human disease.