Feature
Of Mice and Men: Tracking the Origins of
Metastatic Prostate Cancer
In most cases, prostate cancer is a treatable disease. Typically slow growing tumors that occur in men at a median age of 70 years are often treated effectively by interfering with androgen hormone signaling. But in 10 percent of cases, prostate cancers metastasize, become resistant to androgen deprivation therapy, and turn lethal. Kathleen Kelly, Ph.D., Chief of CCR’s Cell and Cancer Biology Branch, has a long-standing interest in understanding the transformation from normal prostate cells into primary cancer and then into metastatic disease. Led by a desire to identify the earliest origins of prostate cancer, Kelly turned to a model system that allows her to study the cells that give rise to the disease as well as trace its metastatic spread.
Cancer’s Original Sin
Kathleen Kelly, Ph.D., and members of her laboratory.(Photo: R. Baer)
Prostate tumors initially require androgen hormone signaling to survive, so androgen-deprivation therapy (ADT), using drugs that inhibit androgen- receptor signaling, has been a highly effective therapeutic option for many patients. Over the years, the drugs that can inhibit androgen receptor signaling have improved such that the time between when the prostate cancer patient is treated and when he succumbs to the disease has increased. But, when the cancer progresses, it is almost always linked to the development of androgen-independence or “castrate- resistant” prostate cancer. Metastasis is invariably associated with a castrate- resistant form of the disease.
“When I first started thinking about the source of metastases,” said Kelly, “one of the things I found very interesting about the healthy prostate is that when you take away androgen, the prostate shrinks and involutes. And when you add androgen back, it grows.” There aren’t very many dividing cells in the normal prostate, but manipulations of androgen provided a striking demonstration of the existence and importance of androgen-independent stem cells in the healthy prostate.
“One of the ideas in the field that hasn’t been proven or disproven yet, is that in prostate cancer, an immature undifferentiated cancer cell ultimately gives rise to resistance and metastases,” said Kelly. “The hypothesis is that this cancer stem cell doesn’t require—or has unique mechanisms for obtaining— androgen receptor signaling, so it survives androgen-deprivation therapy.”
To test the hypothesis, Kelly wanted to look at the cells that initiate prostate cancers and follow their progression.
Building a Better Model
Many human cancers can be effectively studied through xenografts, in which primary tumor cells are injected into mice with compromised immune systems so that they do not reject the foreign cancer cells. Prostate cancer is unusual in that it is extremely difficult to reproduce in such a model system. Furthermore, although they can be kept alive, primary prostate tumor cells do not thrive in culture conditions.
There are a handful of prostate cancer cell lines, which were mostly derived from metastases and not primary tumors, so they have multiple mutations and have been in culture for many years. Thus, they are problematic as a tool to study the properties of cancer stem cells, as they might not exist in a living organism.
“The approach I decided to take was a mouse model,” said Kelly. “I chose an aggressive model, because I thought there was a higher chance we could study a metastatic process.” Kelly, with her graduate student, Philip Martin, D.V.M., made a mouse with deletions of two tumor suppressors— PTEN and TP53—in prostate epithelial cells. Mutations of PTEN and TP53 occur at fairly high frequency in human populations and are often associated with aggressive, castrate-resistant, and metastatic disease.
In addition to the two genetic deletions, Kelly and Martin introduced a light-emitting reporter gene—luciferase—which allowed tracking of transplanted cells that carried the genetic deletions.
Kelly and her colleagues have created a mouse model of prostate cancer with tumors that display a diversity of tumor types: (A) prostatic intraepithelial neoplasia; (B) adenocarcinoma; (C) and (D) prostate tumors positively stained for two epithelial filament markers (arrow: CK8 and CK5) or only one (asterisk: CK8 but not CK5); (E) vascular invasion; (F) prostatic urothelial carcinoma; (G) sarcomatoid carcinoma; (H) basal/squamous carcinoma.(Image: K. Kelly, CCR)
“The idea we had was that we would be able to sort through the tumor cells and find the tumor-initiating cells in these mice,” said Kelly, “but first we had to fully characterize our model.” So using his training in veterinary pathology, Martin led a full longitudinal study of their mouse model. (See “The Veterinary Perspective.”)
“One of the most important things we were able to show is that, unlike other mouse models of prostate cancer, this one produced cells with metastatic potential,” said Martin. The mice rapidly developed tumors composed of multiple cell types, which were lethal in approximately seven months.
Panning for Cells
In the Kelly lab, Research Fellow Wassim Abou-Kheir, Ph.D., has studied the progenitor cells undergoing transformation in the prostate of these genetically modified mice. He used selective culturing conditions to study the self-renewing capabilities of prostate cells extracted from these mice. He found that the number of progenitor cells in these mice was strongly amplified and that the cells had a greatly increased ability to self- renew compared to cells from normal mice. “They can be cultured indefinitely and will continue as progenitors. We believe that the tumor-initiating cells are within this self-renewing population,” said Kelly.
Meanwhile, at a nearby bench, Research Fellow Paul Hynes, Ph.D., is searching for tumor-initiating cells by teasing apart identifiable cells from the primary prostate tumors in these mice. The process of fractionating the cells involves separating them out on the basis of protein markers on their cell surface. “He’s finding that there is an undifferentiated tumor-initiating cell that can give rise to both basal and luminal cells,” said Kelly, referring to the two major cell types in the prostate.
The team has also found such bipotential progenitor cells in cell lines that they have created from single tumor cells, i.e., clonal cell lines, and analyzed. These cells are also immature, can give rise to both basal and luminal cells, and metastasize when grafted into the mouse prostate.
“We are really interested in determining how tumor-initiating cells are related to the different cell lineages and then understanding what their response is to androgen and androgen deprivation,” said Kelly.
