Cancer can be described as a cellular disease, which is thought to arise from misbehaving cells that divide uncontrollably in vivo. Our basic understanding of why this occurs is because these “cancer cells” have lost its ability to respond appropriately to endogenous stop signals that normally work to maintain the structural integrity of normal tissues. The resulting uncontrolled cell division results in the formation of a tumor.

However, it is becoming clear that cancer doesn’t kill by just simply growing out of control. Other deadly features of cancer include their uncanny ability to spread, to encourage vascular ingrowth feeding the tumor- allowing the tumor to take root indefinitely in the host’s body. These properties enable the tumor to not only persist and take root in its primary location, but to also spread into many organs throughout the body where secondary tumors can take root and grow.

As tumor growth in these secondary locations progresses, they will start disrupting normal tissue integrity, which at later stages can result in full-blown organ failure leading to death. With the quest to understand precisely how cancer cells from a single tumor can be lead to widespread destruction, researchers have uncovered many molecular mechanisms (ranging from genomic, transcriptomic and proteomic changes) that can explain how and why cancer cells behave this way.

The most recent hypothesis suggests that cancer’s destructive behavior is attributed to a population of stem-like cancer cells that are dubbed the “cancer stem cells”. While this is often debated, researchers reveal evidence that cancer stem cells (CSCs) are contributors to the invasive behavior of cancers, but also express vascular growth factors to encourage tumor blood vessel growth to continue feeding the tumor.

An interesting property of CSCs is their ability to undergo multi-lineage differentiation, creating a jumbled array of progeny lineages that make up the diverse tumor mass. In brain tumors for example, cancer stem cells would not only generate simply glial-like progeny population that make up the bulk of the tumor, but also endothelial-like cells expressing markers that are normally expressed in endothelial cells making up the tumor’s vasculature (Ricci-Vitiani et al., 2007).

Based on this observation, the De Maria group in 2007 hypothesized that CSCs might play an active role in building tumor blood vessels in brain tumors, specifically by differentiating into endothelial progeny to be building blocks of the tumor vasculature. Although an intriguing hypothesis, further evidence was needed to confirm that these CSC-derived endothelial progeny are actually members of the tumor vasculature. To that end, the De Maria group used an ingenious Tie2-tk system to functionally demonstrate the CSC-derived endothelial cells do indeed make up the blood vessels feeding brain tumors (Ricci-Vitiani et al., 2010).

Tie2 is a prominent endothelial cell marker that is shown to be expressed in CSC-derived endothelial cells. One way to show that these Tie2+ CSC-derived endothelial cells are important for the tumor vasculature is to selectively ablate them in vivo, and then show whether that can impact tumor growth and survival. A classical ablation system uses the basic properties of the herpes simplex virus thymidine kinase (tk) gene, whose expression would render cells vulnerable to the cell-killing effects of ganciclovir. Using this system to selectively ablate CSC-derived endothelial cells, the authors generated a lentivirus that expresses tk under the control of the endothelial Tie2 promoter.

CSCs infected with these lentiviruses will express tk whenever they undergo endothelial differentiation (when Tie2 is expressed), rendering all CSC-derived endothelial cells susceptible to gancilovir. Using this system, the authors demonstrated solid evidence that CSCs contribute to the tumor vasculature via endothelial differentiation. Specifically, CSCs were shown to express tk whenever they undergo endothelial differentiation in vitro. The authors further observed the presence of tk+ CSC-derived endothelial cells in the tumor vasculature of subcutaneous brain tumor xenografts in vivo.

Moreover, selective ablation of these tk+ endothelial cells with ganciclovir ultimately destroyed the tumor’s vasculature, causing profound tumor shrinkage (Ricci-Vitiani et al., 2010). From this study, the De Maria group could confidently conclude that CSCs play an active role in the growth of tumor blood vessels, allowing the tumor to continually feed its own growth and to take root in the host.

By unraveling the cell biology underlying this deadly process, researchers can now develop better treatment strategies to stop tumor spread and its deadly consequences in the host.


1) Ricci-Vitiani et al. Nature. 2007 Jan 4;445(7123):111-5
2) Ricci-Vitiani et al. Nature. 2010 Dec 9;468(7325):824-8.