Scientific Accomplishments

Mapped the oncogenic potential of retroviruses to the LTR and proposed the insertional mutagenesis model for oncogenesis.

In the course of my training in Medicine and Hematology I developed a strong interest in the Biology of Cancer. Given that high throughput and other technologies that allow the study of human cancer at the molecular level were not available at that time, I searched for animal models that would permit an in depth investigation of the cancer process. This led me to retroviruses, which were known to cause hematologic and epithelial neoplasms in a variety of species. Working with John Coffin, we embarked in studies aiming to genetically map the oncogenic potential of retroviruses. This was done by testing the oncogenic potential of recombinants between RAV0 (a non-oncogenic avian retrovirus) and td-PrRSVB, an oncogenic, transformation defective mutant of Rous Sarcoma Virus (RSV). Prior to addressing the oncogenic potential of the recombinants, we mapped their genome by oligonucleotide fingerprinting. These experiments allowed us to map the oncogenic potential of retroviruses in the LTR, which contains no coding sequences and is involved in the integration of the proviral DNA into the cellular genome. Based on these observations, we formulated the insertional mutagenesis model for oncogenesis, which we proposed in the XLIV Cold Spring Harbor Symposium of Quantitative Biology in 1979 (Tsichlis and Coffin, 1979 and Tsichlis and Coffin, 1980). The importance of this work was that it provided the first concrete molecular evidence for the genetic nature of cancer. In addition, it led to a new unbiased technology that was used extensively by us and others to identify a host of novel cancer genes.

In the early eighties, several laboratories embarked in studies aiming to validate the insertional mutagenesis model of oncogenesis. My laboratory was the first to validate this model in retrovirus-induced lymphomas in rodents (Tsichlis et al, 1983). In the course of these studies we were also the first to provide definitive proof that provirus integration may transcriptionally activate genes in cis from a long distance (Lazo et al. 1990).

Pioneered the use of insertional mutagenesis as a tool for the discovery and characterization of cancer genes.

Once the insertional mutagenesis model of oncogenesis was validated, we (and others) used it as a tool to identify genes involved in tumor induction and progression. Particularly informative were studies focusing on retrovirus-induced rat T cell lymphoma cell lines we established. Using these cell lines, we observed changes in the pattern of provirus integration upon selection for different phenotypes in culture and in animals. More important, new provirus insertions observed after the selection, were shown to target genes whose activation reproducibly contributed to the selection. This observation allowed us to identify genes whose activation was responsible for the changing phenotype of the tumor cells (Bear et al., 1989; Patriotis et al.1993 and Gilks et al., 1993).

Introduced a host of new players to the genetic and epigenetic regulation of innate and adaptive immunity, inflammation, hematopoiesis and cancer.

The goal of the insertional mutagenesis screens we performed over the years was to identify novel cancer genes and to use them as probes to investigate biological function from the molecular to the animal level, both under normal conditions and in disease states. These studies introduced a number of novel players to the genetic and epigenetic regulation of innate and adaptive immunity, inflammation, hematopoiesis and cancer. The papers listed in this section present some of our earlier findings that established the biological role and raised interest in these molecules. Specifically, the paper by Grimes et al established the role of Gfi1 in transcriptional regulation and facilitated studies which established this molecule as a critical regulator of hematopoiesis. In addition, it identified the SNAG domain, an important module involved in the epigenetic regulation of transcription. The papers by Dumitru et al and Hatziapostolou et al generated significant interest on the role of the Tpl2 kinase in immunity and cancer. These and other studies that followed allowed us and others to gain novel insights into these processes. Finally, the paper by Kottakis et al was one of the first to provide a link between signaling and epigenetic regulation of gene expression.

Identified the Akt proto-oncogene and demonstrated that it is regulated by the PI-3 Kinase, through the interaction of PI-3K-generated phosphoinositides with the Akt PH domain.

Our interest in oncogenesis by retroviruses led us to the AKT8 virus, a transforming virus isolated from an AKR mouse thymoma. Cloning and characterization of AKT8 revealed that it was the product of recombination between the ecotropic murine retrovirus of the AKR mouse and a cell-encoded serine-threonine protein kinase with significant homology to PKC (Akt). In the original publication, we also pointed out that this kinase was distinct from other PKC isoforms because it had a unique N-terminal domain with distant homology to the SH2 domain and we predicted that this domain (named later the pleckstrin homology-PH domain) plays an important role in the regulation of the kinase. (Bellacosa et al.1991). Finally, we showed that the oncogenic activation of the Akt kinase was due to a retrovirus-derived myristoylation signal, which was introduced to the N-terminus of the kinase and promoted its translocation to the plasma membrane (Ahmed et al 1992). Following the original publications showing that Akt functions as an oncogene, we focused on the regulation of the Akt-encoded kinase. We observed that in the absence of external signals, the Akt kinase was inactive, but it was activated by tyrosine kinase receptor (and other) signals. These studies led to the observation that Akt activation depends on the PI-3 Kinase and that the PH domain of Akt interacts with PI-3K-generated phosphoinositides. More important, in the course of these studies we generated point mutants of the PH domain and we showed that they could not be activated by the PI-3K-tranduced signals. Ultimately, these mutants proved critical in confirming the specificity of the phosphoinositide-PH domain interaction and in showing that the PH domain was functionally indispensable for Akt activation (Franke et al. 1995). These observations opened up the field of PI3K/Akt signaling because they linked Akt to the PI-3 Kinase, which was known to play an important role in cell physiology, but had no other known targets at the time of this discovery. In addition, they provided the framework for studies on the regulation of phosphoinositide metabolism and its role in the regulation of other PI-3K-dependent kinases (Chan et al. 1999).  With more than 50,000 publications linking the PI-3K/Akt pathway to cell biology, the pathway we discovered 20 years ago, is clearly critical to our basic understanding of the regulation of cell function. More important, molecules involved in this pathway represent some of the most promising targets for cancer and other human diseases to-date.

Confirmed that the three Akt isoforms are functionally distinct and showed that signaling differences are responsible for at least some of the functional differences between isoforms.

Early on, after the discovery of Akt, we became interested in the functional differences between the three Akt isoforms. The rationale for this interest was provided by studies on Akt knockout mice, which showed that although the three isoforms share common functions they also have properties that are unique to each of them (Maroulakou et al 2007). Our studies were focused on whether the functional differences between Akt isoforms are due to signaling differences. To address this question, we employed a set of immortalized lung fibroblasts from Akt1fl/fl/ Akt2-/-/Akt3-/- mice, which were engineered to express different Akt isoforms but were otherwise identical. Using these cell lines we were able to show major Akt isoform-dependent differences in microRNA gene expression in response to external signals. Along these lines, the differential regulation of the microRNAs of the miR-200 family was found to be responsible for the selective ability of Akt2 to promote EMT in response to TGFβ stimulation (Iliopoulos et al 2009). Furthermore, the preferential upregulation of the microRNA miR-21 by Akt2 in response to hypoxia was found to confer hypoxia resistance to Akt2-expressing tumor cells (Polytarchou et al 2011). More important, these Akt2-dependent effects on microRNA regulation were shown to play an important role in breast and ovarian cancer.

A phosphoproteomics screen of these cells identified 606 proteins that are phosphorylated on Akt phosphorylation consensus sites in at least one of the Akt isoform-expressing cells. Analysis of the phosphoproteome data identified signaling pathways and cellular functions preferentially regulated by one or another isoform. One of the differentially-regulated functions was the post-transcriptional processing of mRNA (RNA metabolism) and one of the differentially-phosphorylated proteins involved in the regulation of RNA metabolism was IWS1. Our further studies revealed the IWS1 phosphorylation is required for the recruitment of the histone methylatransferase SetD2 to a transcriptional elongation complex assembling on the CTD of RNA Pol II and containing, in addition to IWS1 and SetD2, Spt6 and Aly. SetD2 recruitment to this complex directs the trimethylation of histone H3 at K36 during transcription and promotes transcriptional elongation and alternative RNA splicing of target genes. More important, these events appear to contribute to the oncogenic activity of Akt isoforms in human cancer (Sanidas et al 2014).