As every homeowner with a lawn knows, you have to remove both the leafy part of a weed and its roots to stop the wretched plant from growing back.

Harvard oncologist and scientist D. Gary Gilliland says a similar principle applies in cancer: It is not enough to kill cancer cells; to cure the disease, you have to annihilate the root cells that generate the cancer cells.

Cancer is thought to originate from a cancer stem cell, a cell that renews itself and also produces the cancer cells in different tissues, he says. Current cancer treatments do not cure disease because cancer stem cells have molecular machinery that resists chemotherapy and radiation.

Understanding cancer stem cells' vulnerabilities may enable their destruction with new treatments, and is a major research focus of Gilliland's laboratory. His group finds genes that cause cancer and "stemness," or cancer stem cell properties, in blood-derived cancers and other tumors; determines how the genes exert damage; and then tries to develop treatments based on understanding the disease-causing gene pathways.

Gilliland, director of the Leukemia Program at the Dana-Farber/Harvard Cancer Center and of the Cancer Stem Cell Program of the Harvard Stem Cell Institute, has been interested in cancer biology and finding rational approaches to its treatment since graduate school.

For his Ph.D., which he received in 1980, Gilliland studied ways to exploit natural poisons, such as diphtheria toxin, to kill cancer cells. He and other scientists at the time aimed to fuse a poison with an antibody specific for a cancer cell, which would take up the hybrid molecule and then die. They developed proof of principle for this approach, but it was nearly 20 years until the FDA approved a cancer drug based on the fusion concept.

The difficulty in translating laboratory results into patient care frustrated Gilliland as a graduate student. "I was struck by how little clinicians knew about basic science and how little I knew about clinical science," Gilliland recalls.

To help bridge the gap, Gilliland went to medical school at the University of California, San Francisco, after finishing his Ph.D. He then specialized in hematology and medical oncology at Brigham and Women's Hospital and the Dana-Farber Cancer Institute at Harvard Medical School, and established his laboratory in Boston in 1990.

Thankfully, Gilliland says, the interface between laboratory bench and patient bedside has rapidly expanded since the 1980s. For example, the clinical response to the tyrosine kinase inhibitor imatinib (Gleevec) led Gilliland and his colleagues to the discovery in 2003 that the genetic basis of the blood cancer hypereosinophilic syndrome (HES) is a mutated tyrosine kinase, called PDGFRalpha.

Tyrosine kinases are enzymes that control cell growth and proliferation. Mutations in tyrosine kinase genes, of which 90 exist in the human genome, result in uncontrolled growth and proliferation of cells seen in cancer.

In 2004, Gilliland used a high-throughput DNA-sequencing approach in the human genome to identify another tyrosine kinase mutation, JAK2V617F, responsible for blood cancers known as myeloproliferative diseases. Gilliland now is a coinvestigator on clinical trials, begun in 2007, for drugs that inhibit the JAK2 tyrosine kinase in these patients.

Although this tyrosine kinase gene work has been fruitful, Gilliland points out that patients are not cured with imatinib and related treatments, apparently because these drugs are unable to kill the cancer stem cell.

In studying cancer stem cells, Gilliland has found that genes conferring stemness fall into two types: transcription factor genes that activate or repress other genes; or chromatin-remodeling genes that shape chromosomal structure to allow or prevent a gene's activity.

Gilliland believes alterations in transcription factor genes turn on a cell pathway, normally off, that enables a cell to become a stem cell. Chromatin remodelers, he suggests, may revert cells to an embryonic state, when cells in the body grow rapidly and form organs. Cancer stem cells with remodeling mutations lead to uncontrollable cell growth, not to form a new organ as their normal counterparts do during development, but to create a tumor in an organ.

Besides providing important clues about cancer, cancer stem cells, Gilliland believes, should reveal insights about healthy adult stem cells. Understanding both is important because drugs that target cancer stem cells have to be safe for adult stem cells. Also, knowledge gleaned from cancer stem cells may be applied to adult stem cell organ regeneration work, another area of active research.

Much remains to be learned about cancer stem cells, Gilliland says. But seeing cancer patients continue to die despite our best treatments, he says, is a powerful motivator for him to return to the laboratory after being in the clinic. "Very few people have not been touched by cancer," he says. Gilliland hopes his research will soon transform cancer treatments into actual cures.