flourescently labeled microscopy cross-section of bone with four zoomed-in sections
A mouse femur (left) with bone-healing NG2+ cells in pink, blood vessels in blue, and cancer cells in green. The four smaller images (right) show direct interactions between cancer cells and NG2+ cells in some areas.
Weijie Zhang, et al.

New research may help explain why long-dormant cancer cells can suddenly grow more aggressive. Cancer cells can enter the blood stream and invade bones and other tissue soon after a primary tumor forms. Within bone, these disseminated tumor cells (DTCs) hide out in the perivascular niche, a space that surrounds blood vessels, where they can remain dormant for long periods before inexplicably awakening, ready to colonize the surrounding tissue. Colonization—the last step in bone metastasis—often occurs years after removing primary tumors, and its effects are estimated to kill hundreds of thousands in the US annually. 

“If cancer is already in the bone, what triggers it to regrow?” asks Xiang H.-F. Zhang. The Baylor College of Medicine cancer researcher is trying to answer that question. 

Case studies showing bone metastasis following dental implant surgeries, as well as epidemiological studies indicating the risk of bone metastasis increases after experiencing bone fractures, have led researchers to posit that the process of bone remodeling after an injury could jumpstart cancer cell division. In a study published October 26 in Cancer Discovery, Zhang and his colleagues find that after a bone fracture, DTCs in mice hitch a ride with perivascular stem cells that the body dispatches to injury sites to begin the healing process. Once they reach the fracture site, cancer cells appear to proliferate in tandem with bone remodeling, during which damaged bone is resorbed and new bone forms in its place. 

“This is a very important study because it validates clinical data which show that increased bone resorption promotes tumor growth in bone, and it provides a mechanism by which this can occur,” Theresa A. Guise, a cancer researcher who studies mechanisms of bone metastasis at the University of Texas MD Anderson Cancer Center and who was not involved in this work,  writes in an email to The Scientist. “These results advance our understanding of bone metastases and indicate a possible reason why patients with bone metastases do poorly when they have a fracture.” 

Zhang’s earlier research showed that bone metastasis tends to occur in osteogenic niches, which are areas where bone-building cells such as osteoblasts and their cellular precursors reside. That finding led him and his colleagues to dig deeper into why bone remodeling may exacerbate the spread of cancer cells. To confirm the effect, they focused on the time when bone remodeling is most active: after a fracture. The team implanted murine-derived tumors with bioluminescent genes into mice. After about 17 days, the researchers removed the tumors and fractured some of the animals’ femurs. After allowing bone remodeling to occur for about 17 more days, bioluminescent imaging revealed that cancer cells had spread pervasively through the fractured femurs and hardly at all in the uninjured bone. 

The researchers then explored mechanisms that could explain this result by repeating the bioluminescent imaging experiment in mice that were depleted of certain bone cell types. Injured femurs from mice with reduced levels of NG2+ cells—a type of stem cell that can differentiate into bone-forming osteoblasts—showed less bone metastasis than fractured bones from mice with normal levels of these cells, suggesting that the cells play a role in conveying cancer cells to fracture sites. NG2+ cells, which like DTCs reside in the perivascular space, appear to migrate to the osteogenic niche, where, Zhang says, his experiments indicate that “the normal function of those cells is to contribute to bone remodeling.” A spatial analysis of tissue with bone metastasis showed NG2+ cells and DTCs in adjacent and often overlapping locations, providing further evidence that a connection between the two cell types may explain the link between bone remodeling and tumor growth. 

Capturing real-time migration of the two types of cells in animals is nearly impossible with today’s technology, says Zhang, but in vitro experiments showed that NG2+ and cancer cells have the ability to adhere to one another and travel in tandem. “So definitely, these two cell types can work together,” he says, adding that two forms of an adhesion molecule called cadherin serve as the glue between them. The form of adhesion molecule expressed by the stem cells is called N-cadherin, while the type made by cancer cells is called E-cadherin. 

To find out whether the processes they’d observed in mice are likely to apply to people too, Zhang and colleagues analyzed in situ protein expression in surgically removed samples of early-stage human bone metastases caused by various cancer types—including breast, prostate, colon, and lung. The researchers found NG2+ and N-cadherin in cells adjacent to cancer cells, suggesting that bone remodeling may influence metastasis regardless of the original cancer cell type, though Zhang notes that the researchers have not yet studied this in detail. 

“The techniques used here are state-of-the-art and the data are comprehensive,” Guise says, adding, “the icing on the cake is that the authors validated their experimental findings in human bone metastases, which further strengthens this study.” 

In further experiments, mice with depleted NG2+ stem cells experienced less bone colonization of DTCs even in the absence of fractures, suggesting that even homeostatic bone remodeling, which occurs continuously, may exacerbate bone metastasis. Zhang says this implies that people who experience faster bone remodeling rates, such as those with osteoporosis, may be at higher risk for metastasis. Guise agrees with this assessment, but she notes that osteoporosis in elderly people isn’t always associated with faster bone resorption. 

It may be possible to prevent bone metastasis by attacking the bond between the N-cadherin and the E-cadherin molecules expressed by cancer cells, Zhang suggests. However, the researchers found that eliminating N-cadherin slowed the rate of bone remodeling in mice. Impeding bone healing to prevent bone metastasis isn’t a viable approach, he says, so he and his colleagues will instead focus on attacking E-cadherin in future research. 

While Guise says Zhang’s proposed therapeutic approach holds promise, she notes that “therapy would have to be targeted to bone only, as loss of E-cadherin may promote malignancy in nonbone sites.” She explains that mutations in genes that normally encode E-cadherin have been linked to breast and gastric cancer in humans. 

Interrupting the connection between cadherin molecules may do more than prevent bone metastasis, Zhang says, because as tumors grow, DTCs can move out of the bone and into other organs, where they may again metastasize. “We do think, if we can break this link, we will not only reduce risk of bone metastasis, but also further metastasis to other organs.”