Intravenous pamidronate is most effective when given to babies and can be given shortly after birth to severely affected babies. 65 The treatment rapidly improves pain, which is a major problem in the initial treatment of babies with severe osteogenesis imperfecta. Radiographs show considerable remodeling of the bones with thickening of the cortices. For example, crumpled femurs and flattened vertebrae resume more normal shapes and cortical thicknesses. The results to date suggest that intravenous pamidronate is an effective form of treatment for severe forms of osteogenesis imperfecta, particularly when commenced in infancy and in early childhood. The treatment is not a cure for osteogenesis imperfecta and does not alter the underlying genetic causes of the disease. Randomized trials are likely to provide more quantitative information about the reported gains. It seems however, that intravenous pamidronate favorably affects the natural history of the severe forms of the disease irrespective of the underlying collagen mutations. However, there are many remaining questions that need to be resolved including the long-term efficacy and safety of the treatment, the duration of treatment, and the use of alternative intravenous and oral bisphosphonates, and the indications for treatment of children and adults with the milder osteogenesis imperfecta Type I. Mesenchymal Stromal Cells and Somatic Gene Therapy Bone marrow contains nonhematopoietic precursor cells that can differentiate into mature mesenchymal cells. 69 The precursor cells, referred to as mesenchymal stromal or stem cells, have the capacity to differentiate in vitro and in vivo into osteogenic, chondrogenic, fibrogenic, and adipogenic lineages. 62,64,67,68 These observations underlie the investigation of unmodified and modified MSCs for cell therapy or somatic gene therapy of osteogenesis imperfecta. In an average human bone marrow graft, there are only two to five MSCs per 1 ? 106 mononuclear cells. 76 Consequently, it is not surprising that only low levels of engraftment, approximately 1% to 2%, were observed after allogeneic whole bone marrow transplantation in a small group of children with osteogenesis imperfecta. 39 An alternative approach is to expand the number of MSCs in ex vivo cultures and then to infuse them into the recipient. 28,45,68 Pereira et al 62 infused normal mouse MSCs into irradiated transgenic mice with osteogenesis imperfecta. In bone, the cells differentiated into osteocytes and produced normal collagen with partial correction of the osteogenesis bone phenotype. The cell therapy produced stromal chimerism in which some cells were normal and some carried the osteogenesis imperfecta mutation. A higher proportion of engrafted normal cells is required, however, to achieve the level of stromal chimerism necessary to functionally correct the osteogenesis imperfecta phenotype. A similar approach also has been used in a pilot study of children with osteogenesis imperfecta. 40 Transplanted allogeneic MSCs bearing a gene marker showed a higher level of engraftment than was observed after whole bone marrow transplantation in the same children. There are, nonetheless, many obstacles to be overcome before allogeneic MSC therapy can be considered to be as safe and as effective as the bisphosphonates in the treatment of severe types of osteogenesis imperfecta. The principal obstacle involves the prevention of graft rejection and graft versus host reaction. Assuming that the latter matters can be prevented, then the next major obstacle is to optimize the cell therapy so that the bone phenotype is improved significantly. A high level of engraftment likely is to be needed but the optimal level may vary in each child depending on the degree of expression of the mutant allele and the growth characteristics of the mutant osteoblasts and mutant osteoblast precursors.