Two years ago, Pamela Robey broke two bones in her leg, a consequence of snow outside and an errant boot. “I slipped and my boot heel got caught in a crack,” she says. “I was moving very fast but my leg was not.”
Generally, of course, such injuries heal on their own. But in this case, the process took longer than anticipated, in part because the blood vessels in the bones were damaged. Robey was in danger of developing what is called a “non-union” break — a fracture that wouldn’t close.
So one of Robey’s friends, an orthopedic surgeon, began contemplating contingency plans. He proposed extracting some of her bone marrow, concentrating it, and injecting it into the fracture “to see if the cells would encourage the bone to heal,” Robey says.
As it turns out, that wasn’t necessary. But Robey is an unusually well informed candidate for such a procedure. She is Chief of the Craniofacial and Skeletal Diseases Branch at the National Institute of Dental and Craniofacial Research, and she, like many other researchers, studies the use of bone marrow-derived stromal cells (BMSCs, also called mesenchymal stem cells, or MSCs), for the repair of bone defects.
Bone is a dynamic tissue, constantly turning over thanks to the push and pull of bone-building osteoblasts and bone-destroying osteoclasts, explains Carl Gregory, Assistant Professor at the Institute for Regenerative Medicine at Texas A&M Health Science Center. Researchers learned early on that they could harvest osteoblast progenitors — MSCs — from bone marrow and culture them in vitro. As a result, Gregory says, bone defects like fracture nonunion, traumatic bone injuries, as well as genetic defects like osteogenesis imperfecta, were for a long time considered the “low-hanging fruit” of the stem cell world.
Several decades later, he says, “Obviously, it didn’t turn out that way.”
In part, that’s because stem cell biology is considerably more complicated than originally thought. Jennifer Westendorf, Professor of Orthopedic Research at the Mayo Clinic, who co-authored a 2009 review on bone repair using MSCs , notes for instance that researchers don’t yet know how to positively select mesenchymal stem cells. There is no one marker or set of cell surface proteins known to purify MSCs, so at the moment, they can only be enriched based on what they are not, rather than what they are (they are not white blood cells, for instance, so those cells can be removed using antibodies to lymphocyte-specific markers). Nor do researchers know if mesenchymal stem cells from the bone marrow are identical in terms of differentiation potential, epigenetics, gene expression, and so on, to those from, say, adipose tissue.
Complicating matters, says Robey, not all bone is alike. Facial bones arise from ectoderm, but the bones of the spine and the long bones in the limbs derive from mesoderm. Long bones develop by a process called endochondral osteogenesis, in which a cartilage scaffold is laid down and then replaced with bone. Bones in the skull form by intramembranous ossification, which skips the cartilage intermediate. Thus, a process that works to heal skull trauma may not work on a broken leg, and cells from the iliac crest may not be able to restore the jaw, as some dental researchers have hoped.
“I think that’s why it’s been harder than we thought it would be,” says Robey, “because the developmental biology of bone is complicated.”
Still, researchers are reporting a steady drumbeat of successes, at least in animal models.
Gregory, for instance, reported earlier this year that applying a mixture of human MSCs and MSC-derived extracellular matrix (ECM) to a 4-mm hole in the calvarial (skull) bone of mice — a critical-sized defect, i.e., one that is too large to heal on its own — produced 80−100% bone coverage over the hole in three weeks, compared to 60% in mice that received cells alone and “marginal” coverage in cells that received only ECM. 
The cells in this study had been pretreated with GW9662, an inhibitor of peroxisome proliferator-activated receptor-gamma, which serves to block MSC adipocyte differentiation and force osteoblast differentiation. That biochemical tweak helped, but key to the study, Gregory explains, is timing. His team found that there is a relatively short window of two weeks or so, following the initial inflammatory phase, during which the bone actually repairs itself (the osteogenic phase). The bone then enters a remodeling phase, and if the gap has not closed by then, it is “highly unlikely” that it ever will.
“Once the signals at the injury, to get the cells to proliferate and repair, have gone from the injury, they’re gone forever,” Gregory explains. “So no matter how many cells you put back in there, when that time’s up, the time’s up.”
His team widened that window by supplying fresh, exogenous signals in the form of MSC-derived ECM. “We just bridged the whole gap with it, and the cells stayed,” he says.
The question now, notes an Editor’s Summary of this work, is whether similar treatment can produce similar effects in people. “Although tweaking microenvironmental cues may be the key to non-union bone healing in people, the authors will need to test in larger animal models and for longer periods of time to ensure reproducible effects of human MSCs and their ECMs,” they write.
Gregory’s finding dovetails nicely with the work of Rocky Tuan, Director of the Center for Cellular and Molecular Engineering at the University of Pittsburgh School of Medicine.
Each cell composes its own ECM, Tuan explains, and in the case of stem cells, that ECM probably helps define the stem cell “niche” that allows those cells to maintain their stemness. In other words, though the bone marrow and fat are highly different, MSCs in both tissues remain as MSCs, probably because of the ECM they weave about themselves. Conversely, “maybe one reason stem cells don’t do so well or can’t do as well when placed in a new environment” — that is, after transplantation — “is because they are devoid of what makes them comfortable in the first place.”
To test that hypothesis, Tuan studied the properties of human MSCs grown either on collagen or MSC-derived ECM. The cells grown under the latter conditions divide more rapidly, are more active, and migrate more efficiently than cells grown on collagen alone.  “So it makes sense that the matrix made by the stem cells are in general very good for the stem cells,” he concludes.
Jeffrey Karp, Co-director of Regenerative Therapeutics at Brigham and Women’s Hospital, and a Principal Investigator at the Harvard Stem Cell Institute, also studies MSC-based therapeutic strategies. Karp’s primary focus is controlling MSCs after transplantation.
“Despite this incredible control we have in the lab, when you transplant cells you lose complete control over those cells,” he says. “The cells are entirely at the mercy of the biological milieu wherever they reside.”
To address that problem, Karp’s lab has devised strategies to guide MSCs to target tissues, and to control their fate once they get there. In one study, his team chemically altered the surface of MSCs in vitro to make them adhere more efficiently to endothelial cell walls at sites of inflammation. By tagging the cell surfaces with sialyl-Lewis-X, an active site on the P-selectin glycoprotein ligand-1, which neutrophils use to invade tissue, the team enhanced the cell’s ability to penetrate tissues at the site of inflammation in a mouse model. When these cells were injected into the tail veins of mice, the team observed 50% enhanced cell localization to an inflamed ear relative to unmodified cells. 
As Karp puts it, “If you know the ZIP code of the blood vessels in a particular tissue, you can program the address on the surface of the cells and then target them to that site…. And we know a lot of ZIP codes.”
In a second study, Karp’s team got MSCs to take up nanoparticles filled with fluorescent dyes or signaling molecules and then slowly release them, thereby directing those cells to differentiate down a given pathway — all without genetic modification.
For instance, by feeding the cells particles loaded with dexamethasone, a glucocorticoid that promotes bone cell differentiation, they could drive most of the cells into osteoblasts, even — thanks to endocrine and paracrine mechanisms — those that didn’t take up the hormone in the first place. 
Karp suggests it would be possible to combine these two mechanisms, for instance, to target MSCs or osteoprogenitor cells to the bone marrow to treat osteoporosis, cardiomyocytes to the heart to repair cardiac tissue damaged in heart attacks, neural stem cells to the brain in Parkinson’s patients, and endothelial progenitor cells to sites of peripheral vascular disease to promote angiogenesis.
Such strategies may prove useful as MSCs make their way into the clinic. Indeed, Westendorf’s review notes that they have already been used in a handful of small clinical trials for bone repair.  And still others are in the planning stage. Over at NIDCR, Robey has for several years been trying to initiate a clinical trial using BMSCs for the treatment of critical-sized skull injuries, such as those that many Iraq war veterans have experienced.
The plan — on the drawing board for nearly seven years — is to extract autologous BMSCs from the patients’ bone marrow, grow them in culture, and implant back into the patient in the presence of a scaffolding material. But one thing after another has set those plans back.
In the meantime, she is pursuing a similar strategy using embryonic and induced pluripotent stem cells.  At the moment, these cells cannot match the performance she sees with BMSCs, restoring only about a quarter of the amount of bone that the BMSCs can both in terms of quantity and quality, she says.
“The BMSCs, when we put them on an appropriate scaffold, make beautiful, beautiful lamellar bone, which is the strongest type of bone,” Robey says, “whereas the IPS cells made what we call ‘woven bone,’ which is not as organized, and it’s also not as strong.”
To improve on that, Robey has been refreshing her cell biology know-how, hoping to more closely mimic embryonic bone formation in a culture dish and thus coax more of the pluripotent cells into the developmental fate she wants.
“We’re just getting started on that,” she says.
 A.H. Undale et al., “Mesenchymal stem cells for bone repair and metabolic bone diseases,” Mayo Clin Proc, 84:893−902, 2009.
 S. Zeitouni et al., “Human mesenchymal stem cell-derived matrices for enhanced osteoregeneration,” Sci Transl Med, 4:132ra55, 2012.
 H. Lin et al., “Influence of decellularized matrix derived from human mesenchymal stem cells on their proliferation, migration and multi-lineage differentiation potential,” Biomaterials, 33:4480−9, 2012.
 D. Sarkar et al., “Engineered cell homing,” Blood, 118:e184−e191, 2011.
 D. Sarkar et al., “Cellular and extracellular programming of cell fate through engineered intracrine-, paracrine-, and endocrine-like mechanisms,” Biomaterials, 32:3053−61, 2011.
 S.A. Kuznetsov, N. Cherman, P.G. Robey, “In vivo bone formation by progeny of human embryonic stem cells,” Stem Cells Dev, 20:269−87, 2010.