For the General Public


Cancer is caused by increased activity of growth-promoting genes and decreased activity of growth-inhibitory genes. These genes produce proteins that can be targeted by specific drugs. Our lab works on the molecular mechanisms of the cancer-specific changes in the regulation of cell growth and proliferation. We then apply the new knowledge to the identification of small chemical inhibitor compounds and of novel biological molecules with the goal of developing these into targeted cancer therapies.

For Scientists


Background and History


Cancer is a disease initiated by genetic changes that affect cell proliferation and metabolism. We now have a wealth of basic information on the cancer cell and its unique properties. New therapeutic targets have been defined:  oncogenes that through their products, the oncoproteins, promote  cell proliferation and survival and tumor suppressor genes and proteins that attenuate cell growth and proliferation. In cancer, oncoproteins show a gain of function and tumor suppressor proteins suffer a loss of function. Much of the excitement in cancer research today comes from the opportunities to translate these basic scientific insights into clinical applications. Targeted therapy, inhibiting the activity of an oncoprotein, can lead to spectacular clinical success.


Research in our lab originated in the study of cancer-inducing retroviruses in animals and in cell culture. Tumor virology has laid much of the foundation for our understanding of cancer. Early milestones were the discovery of the first retroviral oncogenes, genes that are responsible for the tumor inducing activity of a virus. All retroviral oncogenes have been acquired from the genome of the host cell and as cellular genes encode important growth-regulatory factors in the cell, including protein and lipid kinases, transcription factors and adaptor proteins. We have participated in several of these pioneering discoveries (Toyoshima and Vogt, 1969; Duesberg and Vogt, 1970; Vogt, 1971a; Vogt, 1971b; Stehelin et al., 1976; Duesberg et al., 1977; Bohmann et al., 1987).


Current Research


Current activities in our lab focus on basic genetic and molecular mechanisms of oncogenesis. But increasing efforts are also devoted to applied and translational problems. This work is stimulated and supported by collaborations with world class chemists at the institute. Collaborations between the lab and pharmaceutical and biotech companies are also ongoing.


  1. PI 3-kinase: viral oncoprotein and human cancer target

PI 3-kinase is a lipid kinase with oncogenic potential. An avian retrovirus discovered in our lab carries a homolog of the gene encoding the catalytic subunit of PI 3-kinase as its oncogene (Chang et al., 1997). This virus causes aggressive tumors in chickens. PI 3-kinase also plays an important role in human cancer. The alpha isoform of PI 3-kinase is mutated in many tumors and most mutations map to two hotspots in the gene. These mutations increase enzymatic activity, deregulate PI 3-kinase signaling and make the protein oncogenic (Kang et al., 2005). The mutant proteins are ideal cancer targets: they occur only in tumor tissue, and because they are enzymes and show increased function, they could be readily controlled by small molecular inhibitors. Mutant-specific drugs would leave the life-sustaining activities of the wild-type PI 3-kinase untouched.


Besides the hot spot mutations in p110α, other mutations occur in cancer tissue. They are rare and map over much of the PI 3-kinase gene. Virtually all of them show oncogenic activity, but this activity is lower than that of the hot spot mutants (Gymnopoulos et al., 2007).


The excitement about PI 3-kinase and cancer has focused mainly on p110α, because of the cancer-specific gain-of-function mutations that occur only in that isoform. Class I PI3K includes three other isoforms, beta, gamma and delta. Surprisingly, the wild-type protein of these non-alpha isoforms can induce oncogenic transformation in cell culture when they are overexpressed (Kang et al., 2006). In certain human cancer types, there is recurrent overexpression of specific non-alpha isoforms. Thus, p110δ is frequently overexpressed in hematopoietic malignancies, and a p110δ-specific inhibitor, Cal101, is clinically effective in B-cell lymphomas, is FDA-approved and marketed under the name Zydelig by Gilead Sciences (Lannutti et al., 2011; Forcello and Saraiya, 2014).


Cancer-specific mutations have also been identified in the PI3K regulatory subunit p85. These mutations affect the interactions with p110 and reduce or eliminate the inhibitory effect of p85 on p110. Several of these mutations have oncogenic potential and increase signaling activity of PI3K. They work preferentially through interaction with the p110α isoform (Sun et al., 2010).


In a separate development, we have discovered a novel link that connects PI3K activity with the activation of Stat3. This connection is mediated by the Tec kinase Bmx. Bmx is a non-receptor tyrosine kinase that by virtue of its PH domain can be activated by PI3K. The activation of Stat3 is necessary for p110α-induced oncogenic transformation of chicken embryo fibroblasts. It is also seen in several human cancer cell lines (Hart et al., 2011).


In a recent study, we explored the molecular consequences of a single point mutation in one allele of PIK3CA, the gene encoding p110α. This smallest possible genetic change can lead to extensive molecular remodeling of the cell. These changes reach far beyond the established activities and roles that can be attributed to PI3K. We have referred to this cellular reprogramming as the "butterfly effect" in cancer (Hart et al., 2015), borrowing a concept from meteorology (Lorenz, 1963). In this particular case, we used the breast epithelial cell line MCF-10A and its knock-in mutant counterpart. The PI3K mutation induces a gene signature that resembles that of mature basal breast cancer, but not that of other histologic types of breast cancer. It appears that MCF-10A cells are preprogrammed to respond to a PI3K gain-of-function mutation by activating a definitive pattern of phenotypic changes that are characteristic of a specific histologic type of cancer.


This background information defines several currently active projects. These include


  1. investigating the butterfly effect with cells derived from different tissues and with mutations in other oncogenes

  2. determining the molecular mechanisms for the gain of function in mutant p110 and p85

  3. defining functions of down- and upstream components of PI 3-kinase signaling in oncogenicity

  4. explaining the oncogenic activity of the wild-type non-alpha isoforms of PI 3-kinase

  5. finding novel cellular regulators of PI 3-kinase activity

  6. determining the significance of the PI3K-Stat3 connection in human cancer


  1. MYC, the emperor of oncoproteins


MYC and the non-coding transcriptome: As part of a characterization of a novel MYC inhibitor, we have analyzed the MYC-regulated transcriptome. We found that most, and perhaps all, of the noncoding transcriptome is regulated by MYC (Hart et al., 2014; Weinberg et al., 2015). Our data are in agreement with the interpretation of MYC as a general rheostat of transcription. The controlling effect of MYC on the noncoding transcriptome is a huge and exciting expansion of the sphere of MYC action. Some of the non-coding transcripts affected by MYC are bound to be important in generating the oncogenic cellular phenotype. The important task now is to identify these transcripts and determine their functions.


Inhibiting protein-protein interactions: Synthetic organic chemistry is strongly represented at The Scripps Research Institute. A wealth of non-redundant libraries of chemical compounds is available for testing and for screening. We take advantage of these unique resources to obtain inhibitors of oncoproteins.


All cellular activities depend on specific protein interactions. These would be prime targets for intervention were it not for the fact that the interacting surfaces are large and often shallow. There are usually no obvious docking sites for small molecules in these surfaces. Yet some protein-protein interactions have been interrupted by small molecules. Compounds referred to as Nutlins inhibit the interaction of the p53 protein with its regulator MDM2. In collaboration with Dale Boger’s and Kim Janda’s labs in the Department of Chemistry at The Scripps Research Institute, we have identified small molecules that interfere with the dimerization of the Myc and Max proteins (Berg et al., 2002, Xu et al., 2006). This dimerization is essential for the oncogenic activity of the Myc protein. At present, the best candidate inhibitor of MYC is KJ-Pyr-9 (Hart et al., 2014). It binds to the MYC-MAX dimer and to MYC, but not to MAX, inhibits MYC-MAX dimerization and downregulates the MYC-dependent transcriptome.


Our current focus is to:


  1. carry out a systemic analysis of the MYC-regulated non-coding transcripts and their functions in cancer, using CRISPR technologies

  2. generate derivatives of KJ-Pyr-9 with improved solubility and efficacy

  3. identify the exact binding site of KJ-Pyr-9 and its effect on the conformation of MYC

  4. develop target-specific high-throughput screens for inhibitors of protein-protein interactions

  5. identify inhibitors of other cancer-relevant protein-protein interactions: RAS/p110, Rheb/FKBP38 and bcl9/β-catenin interaction