An in vivo model for explanted human breast epithelial cells has eluded researchers for several decades, leaving them at a loss for a key tool for understanding carcinogenesis and for testing novel therapies.
Now, as explained at the San Antonio Breast Cancer Symposium, Charlotte Kuperwasser, PhD, and her former postdoctoral advisor, Robert Weinberg, PhD, have developed a model that allows them to grow human breast epithelial cells in young mice.
“It is not just about the tumor cells, but also about the ecosystem that includes non-tumor cells” in the mammary tissue, noted D. Craig Allred, MD, Professor of Pathology at Baylor College of Medicine, who as a member of the program committee introduced her plenary talk.
“Tumor cells do whatever they do—good or bad—but they do it with the involvement of stromal cells, and Dr. Kuperwasser's experiments show that the stromal cells are key factors in breast cancer,” he said. “It is unusual for us to include someone so early in their career as a plenary speaker, but I think her work is going to have a huge impact.”
Mammary development occurs postnatally in both mice and humans, making it possible, at least theoretically, for researchers to create a mouse system that mimics human development, said Dr. Kuperwasser, Assistant Professor in the Departments of Anatomy and Cellular Biology at Tufts University School of Medicine and in Radiation Oncology at the New England Medical Center.
However, there are significant differences between mammary tissue in the two organisms. The primordial breast duct in the mouse is surrounded primarily by adipose tissue with only a thin rim of stroma.
In contrast, there is massive stromal proliferation during human mammary development, and it is rare to see human mammary epithelial cells in direct contact with adipose tissue.
In the late 1950s scientists discovered that they could remove the primordial mouse mammary epithelium from the mammary fat pad before the animal reaches three weeks of age.
This leaves a clear mammary gland that can support the development of transplanted mouse mammary epithelial cells. Such transplanted mouse cells develop into an anatomically normal and functional mammary gland that only lacks connection to the nipple.
However, the mouse fat pad does not support the development of human mammary epithelium and ducts. When human mammary epithelial cells are transplanted into it they fail to establish themselves or proliferate.
Incorporating Stroma into Transplantation System
Dr. Kuperwasser said she theorized that given that stromal tissue surrounds human mammary epithelium while adipose tissue is more prominent in the mouse tissue, co-transplantation of epithelial cells and stromal cells might induce human mammary growth in the mouse fat pad.
When human mammary fibroblasts genetically labeled with green fluorescent protein were injected into the fat pad, there was little engraftment or proliferation. If, however, the cells were mixed with an equal number of similar cells that had been irradiated, which is known to activate fibroblasts and induce expression of matrix proteins, proteases and growth factors, the human cells did establish themselves in the mouse fat pad and proliferate.
The result is a “humanized” mouse fat pad, Dr. Kuperwasser explained, in which human fibroblasts are intermingled with mouse adipose cells.
Next the team prepared human mammary organoids from normal mammoplasty tissue. The organoids are composed of myoepithelial and luminal epithelial cells that are clustered together but no longer resemble structures in human breast tissue.
Injection of these mammary organoids along with primary normal human mammary fibroblasts into already humanized mouse fat pads induced ductal and lobular outgrowths and the formation of terminal ductal lobular structures.
Genomic fluorescent in situ hybridization and detection of the green fluorescent protein showed that the newly formed structures arise from human tissue.
The structures express proteins typical of human mammary epithelium. If animals carrying such transplants are sacrificed late in pregnancy, the human-derived tissue expresses ß-casein.
“At this point it seems we are able to faithfully recapitulate human mammary development,” Dr. Kuperwasser said.
Normal Stroma Suppresses Abnormal Growth
Given the strong influence the stroma had on human mammary epithelial outgrowth, the researchers hypothesized that injecting human mammary organoids alone, without the addition of normal primary fibroblasts, into humanized mouse pads might induce abnormal growth.
To test this hypothesis, mammary organoids were isolated from 10 apparently normal mammoplasty patients and injected into humanized mouse mammary pads. Of these, seven produced normal looking but stunted structures.
Three, however, resulted in hyperplasias. None of the samples resulted in abnormal growth when co-injected with normal primary fibroblasts.
“The results suggest that normal stroma suppresses abnormal growth,” Dr. Kuperwasser said.
The results also indicate, as has been suggested by postmortem data and recent research by Thea Tlsty, PhD, and colleagues at the University of California, San Francisco, that a significant number of healthy women have abnormal mammary epithelial cells. (Dr. Kuperwasser was careful to point out that neither her results nor those of others indicate what proportion, if any, of these cells would become cancerous during a woman's lifetime.)
In a second variation of the technique, human epithelial cells that overexpress either TGFß or hepatocyte growth factor (HGF) were used to humanize the mouse fat pad stroma. When human mammary epithelial organoids were subsequently injected into these genetically modified stroma, 16 of 18 injections resulted in tumors.
None of the organoid samples resulted in tumor formation when co-injected with normal fibroblasts. Thus, inclusion of normal stroma prohibits abnormal outgrowth.
Interestingly, in a single animal transplanted with organoids alone, there was a variety of proliferation patterns, from normal to DCIS to invasive tumors, which may make the model particularly valuable for studying disease progression.
Dr. Kuperwasser said that based on these data, she thinks stromal cells play a suppressive role in cancer development, which is a novel hypothesis.
“A lot of people are thinking about changes in stroma as an active event that promotes cancer, but we're interested in investigating tumor suppressive effects of normal stroma,” she said. “Maybe it is something that is lost from the stromal cells rather than something gained that causes permissiveness.”
Her team is still working out how to design an experiment to test that notion.
Already one pharmaceutical company has licensed the technology to start working on drug development, and two more are interested, she noted.
Dr. Kuperwasser is also currently testing the system's ability to support tumor xenografts, but she doesn't yet have conclusive data on whether it will work. She plans to focus her efforts on learning what in the stroma is necessary to suppress abnormal growth and what kinds of stromal changes accompany different breast cancers in humans.