HHMI-NIH Research Scholars
Learn about the HHMI-NIH Research Scholars Program, also known as the Cloister Program. More

Janelia Farm Research Campus
Learn about the new HHMI research campus located in Virginia. More

Summary: Charles Sherr studies tumor-suppressor–dependent signaling networks that prevent progression through the mammalian cell division cycle in response to activated oncogenes and are inactivated in virtually all forms of human cancer.

Cancers arise due to mutations that impinge upon two general families of molecular regulators that govern cell division. Gain-of-function mutations that activate cellular oncogenes result in the production of deregulated proteins that stimulate uncontrolled cell proliferation. A second class of tumor-suppressor genes opposes the actions of oncogenes, and their loss of function predisposes to cancer. The realization that oncogenes and tumor-suppressor genes encode proteins that govern processes such as gene expression, cell differentiation, tissue development, and responses to environmental stress, and that mutations affecting their functions constitutively deregulate signaling networks important for cell proliferation have provided key mechanistic insights into tumor formation.

A Hallmark of Cancer: Inactivation of the RB and p53 Network
Transformation of normal cells to cancer cells commonly involves the disruption of signaling networks regulated by two canonical tumor-suppressor genes encoding the retinoblastoma protein (RB) and the p53 transcription factor. Although the functions of RB and p53 are not normally required for cell division, both proteins govern protective "checkpoint" responses that are triggered by DNA damage or by oncogene stimulation. Activation of RB and p53 induces global changes in gene expression, which either arrest cell division or induce suicide (apoptosis) of irreversibly damaged cells. Disruption of RB and p53 compromises these protective programs, allowing incipient cancer cells to continue to divide, to consolidate mutational damage, and, ultimately, to form tumors.

	The Arf-GFP indicator mouse...

Lesions affecting other genes governing the RB and p53 signaling network can mimic RB and p53 loss of function. Among these are mutations that disrupt INK4/ARF, a compact genetic locus that specifies three additional tumor-suppressor proteins, p16^INK4a, p15^INK4b, and p14^ARF (p19^Arf in the mouse). Expression of the p16^INK4a and p15^INK4b proteins in response to oncogenic stress leads to RB activation, whereas synthesis of p14^ARF induces p53. Therefore, chromosomal deletions that encompass the INK4/ARF locus simultaneously compromise the activities of both RB and p53 and represent one of the most frequent genetic events in cancer.

p53-Independent Functions of Arf
Although the prevailing view is that the tumor-suppressive functions of Arf are mediated through p53, mice lacking both Arf and p53 develop a broader spectrum of cancers than animals lacking either gene alone, implying that the Arf protein also has p53-independent functions. In this regard, Arf can induce the sumoylation of many cellular proteins by inhibiting the activity of Senp3, a desumoylating protease. Given that the modification of target proteins by sumoylation influences many cellular processes, including protein transport, ribosomal biogenesis, DNA replication and repair, and mitosis, Arf-induced sumoylation is likely to underlie its p53-independent tumor-suppressor activity. Pinpointing the most relevant targets is now a major challenge.

ARF Tumor Suppression in Philadelphia Chromosome–Positive Acute Lymphocytic Leukemia
The ARF gene selectively responds to abnormally sustained and elevated thresholds of proliferative signals generated by oncogene activation. In this manner, ARF acts as a fuse that monitors potentially harmful mitogenic "current." Inappropriate growth-promoting signals trigger the ARF-p53 circuit, resulting in the arrest or elimination of cells that have sustained oncogene mutations. Conversely, cells lacking ARF, much like those lacking functional p53, fail to undergo cell cycle arrest or apoptosis in response to oncogenic stimulation and are more highly tumorigenic.

Most human hematological malignancies are initiated by chromosomal translocations that inappropriately activate cellular oncogenes, and investigations of leukemias and lymphomas have provided fertile ground for gaining insights into when ARF is induced in the course of disease. The Philadelphia chromosome (Ph), the first tumor-specific cytogenetic anomaly identified, results from a reciprocal translocation that adjoins the ABL oncogene on chromosome 9 to a breakpoint cluster region (BCR) on chromosome 22 to produce the BCR-ABL fusion protein. This translocation is the founding lesion of two distinct hematological malignancies—chronic myelogenous leukemia (CML) and Ph+ acute lymphoblastic leukemia (ALL)—in which BCR-ABL encodes a constitutively active protein tyrosine kinase. CML affects immature hematopoietic progenitor cells in which the ARF checkpoint is normally silenced and INK4-ARF is seldom deleted. In contrast, in more mature lymphoblastic cells that arise in Ph+ ALL, ARF is induced and likely plays a protective role in the early stages of disease. However, in a recent survey of patients with newly diagnosed Ph+ ALL, we determined that almost two-thirds of them had already sustained biallelic INK4-ARF deletions in their leukemic blasts before therapy could be initiated, suggesting that ARF inactivation could contribute to aggressive disease and poor therapeutic responsiveness.

Despite the dramatic efficacy of FDA-approved BCR-ABL kinase inhibitors (imatinib, dasatinib, and nilotinib) in treating CML patients, these drugs fail to provide durable therapeutic benefit in those with Ph+ ALL. We established a mouse model of Ph+ ALL that relies on rapid generation of BCR-ABL–expressing Arf^–/– pre–B cells and their introduction into healthy syngeneic recipient mice. Essentially every BCR-ABL+, Arf^–/– donor cell has leukemia-initiating cell capacity, whereas many more BCR-ABL+, Arf^+/+ donor cells fail to induce disease in this setting. By adapting this model for in vivo luminescent monitoring of leukemia progression, we observed that although continuous, daily dasatinib administration rapidly kills established leukemic cells and induces transient remissions, the treated mice have residual disease and ultimately relapse on therapy. As in human Ph+ ALL, most of these drug-resistant leukemias sustain clinically relevant mutations in the BCR-ABL kinase that render them insensitive to targeted therapy. Our investigations indicate that Arf inactivation enhances the biological "fitness" of Ph+ ALL cells in the hematopoietic microenvironment, diminishes the efficiency with which targeted therapy can eradicate trace levels of disease, and facilitates the subsequent emergence of tumor cell-intrinsic drug resistance, most frequently manifested by BCR-ABL kinase mutations.

Visualizing Arf Expression in Living Mice
By replacing p19^Arf-coding sequences in the mouse genome with a DNA cassette specifying green fluorescent protein (GFP), we were able to visualize Arf promoter activity in homozygous Arf ^GFP/GFP animals. Because these mice lack functional p19^Arf activity, they are highly prone to tumor development. The spontaneously occurring cancers arising in Arf^GFP/GFP mice, as well as more rapidly appearing tumors in animals engineered to coexpress activated oncogenes, exhibited vivid green fluorescence. By contrast, green fluorescent cells were not detected in most normal tissues of these animals. These results highlight a key feature of Arf regulation—namely, the Arf promoter is insulated from responding to physiologic signals conveyed by proteins that drive normal cell proliferation; yet, Arf is expressed when abnormally elevated signaling thresholds are triggered by mutationally activated oncogenes.

Nonetheless, the fact that inactivation of Arf leads to spontaneous tumor formation implies that the protein must be at least transiently expressed in normal tissues in order to eliminate rare cells that have sustained oncogenic mutations. With this in mind, we generated a new reporter mouse in which the cDNA encoding Cre recombinase was introduced in place of GFP downstream of the Arf promoter. We reasoned that interbreeding Arf-Cre mice to an indicator strain engineered to synthesize Cre-induced β-galactosidase (β-Gal) would not only allow histochemical identification of β-Gal+ cells in which the Arf promoter had been transiently activated but would also define their cellular progeny, thereby giving rise to marked, enzyme-positive "clones." These experiments have now indicated that Arf is transiently expressed in cell populations that arise from certain tissue stem cells, consistent with the idea that Arf acts as part of a molecular switch that limits stem cell self-renewal. Arf silencing may therefore be required for immortalization of both normal stem cells and abnormally self-renewing cancer cells.

Last updated June 10, 2008