Fibrous Dysplasia as a Stem Cell Disease

JOURNAL OF BONE AND MINERAL RESEARCH Volume 21, Supplement 2, 2006 Online reference number: doi: 10.1359/JBMR.06S224 © 2006 American Society for Bone and Mineral Research

Fibrous Dysplasia as a Stem Cell Disease Mara Riminucci,1,2 Isabella Saggio,2,3 Pamela Gehron Robey,4 and Paolo Bianco2,5

ABSTRACT: At a time when significant attention is devoted worldwide to stem cells as a potential tool for curing incurable diseases, fibrous dysplasia of bone (FD) provides a paradigm for stem cell diseases. Consideration of the time and mechanism of the causative mutations and of nature of the pluripotent cells that mutate in early embryonic development indicates that, as a disease of the entire organism, FD can be seen as a disease of pluripotent embryonic cells. As a disease of bone as an organ, in turn, FD can be seen as a disease of postnatal skeletal stem cells, which give rise to dysfunctional osteoblasts. Recognizing FD as a stem cell disease provides a novel conceptual angle and a way to generate appropriate models of the disease, which will continue to provide further insight into its natural history and pathogenesis. In addition, skeletal stem cells may represent a tool for innovative treatments. These can be conceived as directed to alter the in vivo behavior of mutated stem cells, to replace mutated cells through local transplantation, or to correct the genetic defect in the stem cells themselves. In vitro and in vivo models are currently being generated that will permit exploration of these avenues in depth. J Bone Miner Res 2007;21:P125–P131. Online reference number: doi: 10.1359/JBMR.06S224 Key words: fibrous dysplasia, GNAS, skeletal stem cells, embryonic stem cells INTRODUCTION

F

IBROUS DYSPLASIA (FD) of bone (polyostotic fibrous dysplasia, McCune-Albright syndrome [MAS], OMIM 174800) is a genetic, noninherited disease caused by missense mutations in the gene encoding the ␣ subunit of the stimulatory G protein, Gs, within the GNAS complex locus on chromosome 20.(1–3) The disease may affect one or several bones and may involve extraskeletal organs. Skeletal involvement may be limited to one bone (monostotic forms) or extended to multiple bones (polyostotic forms) or the entire skeleton (panostotic forms). The proximal part of the femur and craniofacial bones are the two most commonly affected sites. The most common extraskeletal lesions include pituitary and thyroid adenomas, adrenocortical hyperplasia, and ovarian cysts, which typically associate with growth hormone excess, hyperthyroidism, Cushing’s disease, and precocious puberty, respectively. Abnormal cutaneous pigmented lesions are common (café-au-lait spots with a “coast of Maine” serpiginous profile, as opposed to the lesions observed in neurofibromatosis, sometimes noted for a “coast of California” profile). Many other organs, including the liver, pancreas, and heart, may be involved. The disease presents in infancy, childhood, or adolescence, with signs and symptoms arising from either the skeletal (fracture, pain, deformity, or variable combinations thereof) or the extraskeletal disease (precocious pu-

The authors state that they have no conflicts of interest.

berty, Cushing’s disease, cholestatic liver disease). An isolated monostotic lesion may occasionally remain subclinical for a long time and only be discovered in adult life. Fracture and deformity (with the well-known “shepherd’s crook” deformity at the severe end of the spectrum) often complicate the femoral lesions, whether these are isolated or part of a polyostotic disease. Severe facial deformity complicates extensive craniofacial bone lesions. Blindness may present as a major and sudden complication of lesions involving the skull base arising either from direct compression of optic nerves by fibrous dysplastic tissue or as a consequence of bleeding within a lesion.(4) Two mutations have been shown to cause all forms of the disease. The mutations replace the arginine residue in codon 201 with either cysteine or histidine or more rarely other amino acids.(4,5) The disease-causing mutations occur postzygotically.(1) This explains why the disease is never inherited and the fact that affected individuals are somatic mosaics. In all cases, the resulting mutated protein exhibits a greatly reduced GTPase activity, such that constitutive activation of adenylyl cyclase ensues.(6) As a net result, mutated cells constitutively generate inappropriately high levels of cAMP.(7) Organ dysfunctions are directly accounted for as the consequence of mutated cell dysfunction. This is directly intuitive for some pathological changes (autonomous endocrine hyperfunction, skin pigmentation), and less immediately apparent for others (e.g., fibrous dysplastic lesions of bone, but also muscle myxomas in the related Mazabraud’s syndrome(8) and neonatal cholestatic liver disease(9)).

1 Department of Experimental Medicine, University of L’Aquila, L’Aquila, Italy; 2San Raffaele Biomedical Science Park, Rome, Italy; Department of Genetics and Molecular Biology, La Sapienza University, Rome, Italy; 4Craniofacial and Skeletal Diseases Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, DHHS, Bethesda, Maryland, USA; 5Department of Experimental Medicine and Pathology, La Sapienza University, Rome, Italy. 3

P125

P126

RIMINUCCI ET AL.

FD AS A DISEASE OF EMBRYONIC STEM CELLS The pattern of organ involvement in all cases of McCuneAlbright syndrome typically includes multiple tissues that are known to originate from the three embryonic germ layers—ectoderm (e.g., skin, craniofacial bone), endoderm (e.g., thyroid), and mesoderm (e.g., axial and appendicular bone). Because the mutated cells in the different organs must be part of a single clone originating from an ancestral cell that is mutated in development,(1,10) this cell must be able to give rise to derivatives of all three germ layers. By default, this cell must be mutated before gastrulation, which is the time when the three germ layers are formed and their differentiation potential becomes irreversibly segregated. The disease presents with a wide spectrum of severity and organ involvement.(4) Monostotic FD lesions represent one end of the spectrum, whereas those forms in which multiple bones and multiple extraskeletal organs are involved represent the other end. The diversity of the disease burden in different patients has not been directly explained by experimental data. Nonetheless, the speculation that the frequency of organ lesions reflects the time of mutation has gained wide acceptance.(11,12) In this view, extended or limited patterns of organ involvement as observed in patients would reflect an “earlier or later” time of mutation, respectively. This implies that the two FD-causing mutations may arise in different cells, at different times, through unknown, and perhaps different mechanisms. However, as discussed below, established patterns of development, mechanisms of mutation, and data taken from general clinical observation instead indicate a single site, time, and mechanism of mutation in all cases of widespread disease and in most cases of limited disease.

FD mutations arise in pluripotent cells In all land-dwelling vertebrates, gastrulation is a major watershed in development. Before gastrulation, each individual embryonic cell has the same potency—from blastomeres to cells of the inner cell mass in the blastocyst, each cell can give rise to every intraembryonic tissue.(13) At gastrulation, the three germ layers (ectoderm, endoderm, and mesoderm) separate, and irreversible lineage determination occurs.(13) Cells in each layer lose the transgermal potential that characterizes pregastrulation cells, and all cells in their progeny become lineage restricted, with the notable exception of cells in the cranial neural crest.(14) Skeletal cells, and different bones, are not derived from a single germ layer. Bones in the appendicular and axial skeleton are derived from different spatial specifications of mesoderm (vertebrae, ribs, and sternum from somites, limb, and limb girdle bones from the lateral mesoderm).(15) With the exception of parts of the occipital and otic bones, craniofacial bones in land-dwelling vertebrates including humans,(14) or their cognate cartilage in fish,(16) are derived from the cranial neural crest, an ectoderm derivative. Taken individually, craniofacial bones and the femur are the two most common sites of FD lesions.(17) Given the dual germ layer origin of these bones, concurrence of a single craniofacial and a single femoral lesion, in the same

patient, implies that the original mutated cell must mutate before gastrulation (i.e., before separation of ectoderm, from which neural crest cells giving rise to craniofacial bones originate, and mesoderm). Hence, even if only bone lesions are present and only two bone lesions are present, the mutation must arise in a single pluripotent cell existing before the separation of germ layers, exactly as in MAS patients with widespread multiorgan involvement. Identical considerations apply to all cases in which, for example, a single extraskeletal organ is involved besides bone (e.g., patients with a single lesion in a femur and café-au-lait spots or thyroid disease). Different severity, numbers, and distribution of organ lesions, therefore, do not indicate, and are not explained by, different times of mutation. It is only the embryological origin of the different organs involved, not their number that validates deductions on the timing of mutation. If the embryologic origin of the involved organs is considered, mutation must arise before gastrulation in all cases of MAS and in most cases of FD, regardless of disease extent or severity. True monostotic forms, and polyostotic forms with no extraskeletal lesions and no involvement of the craniofacial bones, are the only forms for which a postgastrulation mutation remains theoretically possible, consistent with established patterns of vertebrate development. Because most polyostotic forms do involve the craniofacial bones, extraskeletal organs, or both,(17) a postgastrulation mutation remains a viable hypothesis for monostotic forms only.(18) In all other cases of the disease, the time of mutation is the same, and mutation must have arisen in pluripotent cells before gastrulation. Disease severity and frequency of lesions in individual cases, on the other hand, can be explained by different considerations. These include variations in the size (extent of proliferation) of the clone originating from a single mutated ancestor, and variations in migration, or in the rate of survival of the mutated cells within the clone during their growth and differentiation.(4)

Both R201C and R201H mutations arise from cytosine methylation The single codon (R201, CGT) that mutates in all cases of fibrous dysplasia includes a CpG dinucleotide. CpG dinucleotides are rare in the genome except in specific regions (CpG islands(19)) and frequently undergo methylation. Two kinds of base transitions (C→T and G→A) underlie the two mutations that alone account for virtually all clinically observed mutations (R201C, R201H) of the GNAS gene in cases of FD/MAS.(4) These two types of base transitions account for ∼30% of all point mutations known to cause human disease,(20) which in itself indicates that, although considered a “rare” disease, FD is the consequence of the most common change in the genome. This change tends to eliminate CpG dinucleotides from the genome. It should not be overlooked in this context that, because the mutation is never inherited, each patient represents a new mutation, which again speaks for a relatively frequent genetic change. Only one of the two mutations is commonly appreciated as a transition of C to T, and the R201C and R201H mutations are commonly noted as transitions in the first or sec-

FIBROUS DYSPLASIA AND STEM CELLS

P127

FIG. 1. Origin of the R201C and R201H mutations from methylation of cytosine in the CpG dinucleotide and its reverse in the sense and antisense strands of codon 201. (A and B) R201C mutation (5⬘-CGT-3⬘ → 5⬘TGT-3⬘). The C→T transition (b, underlined) is generated by hydrolytic deamination of the methylated cytosine in the CpG dinucleotide (a, bold) in the sense strand. (C and D) R201H mutation (5⬘-CGT-3⬘ → 5⬘CAT-3⬘). The C→T transition (D, underlined) is generated by hydrolytic deamination of the methylated cytosine in the reverse CpG dinucleotide (C, bold) in the antisense strand. (E) Developmentally timed methylation in embryonic pluripotent cells. After massive demethylation of the whole genome, de novo methylation occurs between the morula and blastocyst stage. De novo methylation is asymmetric and selectively occurs in pluripotent cells of the inner cell mass.

ond base in the codon, respectively. However, both mutations arise from a C→T transition, which occurs in the sense and antisense strand in a single CpG dinucleotide and its reverse, respectively. The C→T transition (5⬘-CGT-3⬘ → 5⬘-TGT-3⬘) is the consequence of a methylationdeamination sequence affecting the cytosine in the CpG dinucleotide in codon 201, sense strand (Figs. 1A and 1B). The G→A transition (5⬘-CGT-3⬘ → 5⬘-CAT-3⬘) is the consequence of the same methylation-deamination sequence in the complementary, reverse CpG dinucleotide in the antisense strand (3⬘-GCA-5⬘ → 3⬘-GTA-5⬘; Figs. 1C and 1D). Base transition in the sense strand, if the original C→T antisense transition is not corrected by DNA repair, inevitably follows at DNA replication..

Timed methylation in pluripotent cells Methylation of CpG dinucleotides, the immediate cause of both FD-causing mutations, is a pivotal mechanism of epigenetic regulation of gene expression.(21) Between fertilization and gastrulation, massive demethylation and de novo methylation are essential determinants of subsequent developmental events.(22) These early embryonic demethylation and methylation events are evolutionarily linked to parental imprinting,(23) and the GNAS locus as a whole is imprinted.(24) After fertilization, a rapid demethylation of the paternal genome occurs (active demethylation), followed by a slow and progressive demethylation of the maternal genome. The latter occurs because of the absence of the DNA methyltransferase 1 (DMT1) from the nucleus until the eight-cell stage (passive demethylation).(22) Between the formation of the morula and the formation of the blastocyst, massive methylation of the genome occurs specifically in cells that will form the inner cell mass (Fig. 1E). Once the blastocyst has formed, methylation status is asymmetric—high in the inner cell mass and low in the primitive trophectoderm.(22) Whereas the rate of hydrolysis of meth-

yl-cytosine to thymine is constant, as is presumably also the efficiency of relevant repair mechanisms, the rate of proliferation is a major determinant of the chances of spontaneous C→T transitions to be repaired.(25) Given the rapid proliferation of early embryonic cells, the chances for repair of C→T transitions are low, and those for their fixation into mutation are high. Consistent with the pluripotency of the originally mutated single cell in clinically observed cases of multiorgan disease, the very nature of the mutation as linked to early embryonic methylation events provides a temporal marker of the mutational event, placing it between the morula and the blastocyst stages (Fig. 1E). In brief, massive methylation specifically occurs in pluripotent cells in the inner cell mass, before gastrulation, as a fundamental event of early development; FD-causing mutations are both caused by methylation and arise in pluripotent cells. The fact that these mutations arise as a consequence of a universally occurring methylation event is consistent with the high frequency with which these nongerm line transmitted mutations arise, de novo, in each individual patient. The ancestor cells that mutate in the embryo represent the in vivo counterpart of the pluripotent cells that are derived in culture as embryonic stem cells.(26,27) Stated in the simplest way, therefore, FD as a disease of the organism is a disease of “embryonic stem” cells, or more accurately, of their in vivo counterpart.

FD AS A DISEASE OF POSTNATAL STEM CELLS FD as a disease of bone as an organ, in addition, is a disease of skeletal stem cells. At the time when the genetic basis of the disease was unraveled (and thereafter), bone pathology textbooks described FD as a developmental disorder in which normal bone is replaced by fibrous tissue and “metaplastic” bone, formed in the absence of osteo-

P128

RIMINUCCI ET AL.

FIG. 2. Mosaicism of FD lesions. When a single cell suspension isolated from an FD lesion is plated at low density, the CFU-F proliferates and forms a colony. When individual colonies are isolated and analyzed for mutation by direct DNA sequencing, both normal and mutant clones are identified, indicating the mosaic nature of the lesion. If clones are harvested altogether and transplanted in vivo along with a suitable carrier to an immunocompromised mouse, a miniature of FD ossicle is recreated. Normal clones, as expected, give rise to a normal ossicle, whereas mutant clones are lost and do not establish an ossicle (nonossicle), indicating that survival of mutant cells is dependent on the presence of normal cells and the lethal nature of the GNAS mutation.

blasts.(28–31) More clinically oriented descriptions would invariably bring in “altered bone remodeling” as the basis of the disease. Either way, two facts remained unnoticed: (1) osteoblasts, of course, do form FD bone, but are altered in shape and therefore difficult to recognize in sections,(5,32) and (2) the “fibrous” tissue filling the marrow spaces between the abnormal bone trabeculae is made of cells phenotypically resembling bone marrow stromal cells.(33) Whereas highlighting the basic fact that there is a change in the bone marrow structure in FD and that morphologically abnormal osteoblasts are likely abnormal in function as well, these observations opened the way to the adoption of a novel conceptual angle on the disease. This can now to be seen as a disease of the osteoblastic lineage, and therefore of the skeletal (stromal) stem cells from which the lineage emanates.(34) Clonogenic progenitors (colony-forming unitfibroblastic [CFU-F]), which include stem cells in the normal bone marrow stroma,(35) have been isolated from the abnormal bone marrow stroma of FD.(33) This allowed for the identification of both normal and mutated stromal progenitors in FD (Fig. 2), showing that the individual FD lesion is a mosaic in itself, rather than just the wrong piece of the mosaic.(33) Furthermore, this also showed that mosaic strains of stromal cells (a mixture of normal and mutant cells) derived from FD lesions could generate an FD miniature “replica” when transplanted in vivo in immunocompromised mice(33) (Fig. 2). This approach provided the first animal model of the disease, a humanized murine model, and a stem cell-based model. By showing that pure strains of mutated cells would fail to generate FD tissue in the mouse, but instead, were lost from the graft site (Fig. 2), these experiments also provided an experimental confirmation of the classical pathogenetic hypothesis of Happle.(10) The hypothesis states that mutated cells are less viable than their normal counterpart and that mosaicism is in fact a survival factor for the disease genotype.(10) This may be true not only in prenatal development, but also in postnatal life. These observations also raised the question of a critical mutational load that would be necessary to generate FD

tissue versus normal bone and bone marrow, not only in the transplantation model, but also in a patient’s bone. Stated in another way, the issue of relative frequencies of normal and mutated cells, as well as of normal and mutated progenitor cells, became relevant. More refined methods for mutation analysis (such as PNA clamping, which permits detection with an estimated sensitivity of ∼1:200 cells, as opposed to 1:4 cells for standard PCR/sequencing(36,37)) were progressively developed. This made it possible to quantitate the mutational load in nonclonal monolayers of FD-derived stromal cells. Analysis of individual clones,(34) on the other hand, allows estimation of the frequency of mutated CFU-Fs in FD-derived CFU-F cultures. This led to the observation that different clinical lesions include different proportions of assayable mutated cells, as well as of mutated clonogenic progenitors. The frequency of mutated clonogenic cells in FD lesions seems to decline progressively with age, and mutated CFU-Fs may in fact ultimately become undetectable.(38) In parallel, a reversion of the abnormal FD bone structure into normal trabecular bone and bone marrow is observed with age (unpublished data).(38,39) This suggests that, within the “mosaic” found in the bones of patients with FD, normal and mutated skeletal progenitors have a differential lifespan and in vivo history. A number of clinical observations, ranging from the agedependent occurrence of fractures(5,40) to the more or less anecdotal reports of disease “burnout,”(4,39) find a biological basis in these observations. At the same time, these observations are complementary to the observation that FD lesions, albeit caused by an inborn genetic abnormality of osteogenic cells, are not congenital themselves.(4,5) The postnatal development of FD lesions may reflect, in principle, the timed expansion of mutated cells in vivo as dictated by specific physiological and growth-related circumstances. Hence, dynamic positive and negative selection events may be operating in FD bones at different ages. Gaining a more precise definition of such dynamic changes and unraveling their determinants may bring vital information on the biology of the disease, with direct therapeutic implications.

FIBROUS DYSPLASIA AND STEM CELLS The abnormal stromal cells obtained in culture from FD lesions have also been instrumental in clarifying crucial metabolic derangements associated with FD, particularly when in vitro observations were combined with in situ characterization of the FD stroma.(5) Perhaps the best example of this is found in the demonstration that FD lesions are the prime source of the excess circulating fibroblast growth factor (FGF)-23 associated with the phosphate-wasting syndrome observed in FD.(41) The entire osteogenic lineage, from stromal cells to osteocytes, participates in the production of FGF-23 in FD. However, normal cells in the osteogenic lineage also produce FGF-23, and the production of excess FGF-23 in FD reflects disease burden (indeed, osteogenic cell mass) rather than a direct downstream effect of the causative mutation. Besides making bone a pivotal endocrine regulator of renal phosphate clearance,(41) these data characterize FD as unique within the spectrum of FGF-23–linked phosphate-wasting disorders. Unlike tumor-induced osteomalacia, and unlike both autosomal dominant and X-linked forms of hypophosphatemic rickets, excess numbers of normal producer cells sustain the excess serum level of FGF-23 in FD.(41)

P129 tool not only for the generation of suitable transgenic murine models of the disease, but also for elucidating the developmental derangements brought about by the mutation in embryonic cells. For example, transplantation of lentivector-mutated ES cells may elucidate aspects of mutated cell growth, survival, migration, and differentiation during development that are directly relevant to the natural history of the disease and cannot be elucidated if addressed through clinical observation only.

USING STEM CELLS FOR FD THERAPY FD does not have a cure. Pharmacological treatment is merely palliative, and surgery, albeit indispensable in many cases, has well-known limitations.(4) Even if prudently remaining cold about the current “hype” elicited by stem cell research for treatment of disease, these considerations, fortified by the notion that FD is a stem cell disease per se, justify (and actually call for) exploring the potential role of postnatal stem cells as a treatment tool. This involves three distinct levels of approach.

Guiding resident stem cells USING STEM CELLS FOR FD RESEARCH The use of stem cells for elucidating critical questions related to FD biology remains a viable option on multiple levels. GNAS-mutated stromal cell strains (postnatal stem cells) derived in culture from FD lesions will continue to provide a relatively easy approach to the in vitro and in vivo modeling of osteoblastic lineage dysfunction, downstream of the causative GNAS mutations. However, some limitations are inherent to the use of cells from FD lesions. First, experimental work is strictly dependent on availability and adequacy of clinical samples, which may be limited by the relative rarity of the disease itself besides clinical considerations. Second, cloning mutated cells has inherent difficulties emanating from the growth restriction of most GNASmutated CFU-Fs.(42) Third, the in vivo history of the lesions and patients from which the cells are derived introduces an additional layer of experimental variability (e.g., mutational burden).(37) Creating GNAS-mutated cells can circumvent these problems. Viral vectors capable of integration into the genome of infected cells permit efficient and stable transduction of human cells ex vivo.(43–45) Relatively unexplored with respect to their potential with human skeletal stem cells, these approaches have been limited thus far to the use of viral vectors for transfer of either marker genes (e.g., green fluorescent protein) or for transfer of genes encoding secreted proteins.(46) A specific advantage of these tools, however, has been overlooked. This resides in the fact that they can be used to transfer disease genes in stem cells to obtain direct models of the disease. The mutated GNAS cDNA cloned into lentiviral vectors can effectively and stably transfer the disease genotype and phenotype to wildtype human skeletal stem cells.(47) This opens the way to a systematic analysis of the entire transcriptome and proteome changes downstream of the causative GNAS mutation. Importantly, the same approach, when applied to murine embryonic stem (ES) cells, provides an innovative

FD lesions are dynamic mosaics of normal and mutated skeletal cells and their progenitors.(33) Dynamic events underlie a sequence of expansion and regression of a mutated clone within a bone lesion over time in vivo. Therefore, novel methods of pharmacological intervention can be conceived as aiming for alteration of the balance of normal and mutated progenitor cells, stimulating the function of nonmutated stem cells residing within a lesion, or inhibiting the function of their mutated cognates. Significant amounts of work will be necessary before this approach will become feasible. However, pharmacological targeting of osteogenic precursors (the key mediators of the disease phenotype at the organ level) seems a more rational approach than targeting osteoclasts as in current approaches.

Cell therapy Cell therapy using skeletal stem cells would imply the administration of autologous nonmutated cells with the intent of replacing the mutated progenitors and progeny at the sites of lesion. The fact that FD is a somatic mosaic makes this approach specifically conceivable for FD, at variance with other crippling genetic diseases of the skeleton (e.g., osteogenesis imperfecta). Normal skeletal progenitors can be isolated from FD patients and even from FD lesions. Conceivably, one could expand those cells in culture to obtain a transplantable cell strain. However, the only currently available way of isolating nonmutated skeletal stem cells from FD tissues is by cell cloning, a rather cumbersome approach. Identification of surface markers that would permit isolation of mutated and nonmutated cells by immunoselection would greatly extend the feasibility of this approach. Genomic and proteomic analyses are expected to accelerate the pace of advances in this particular area. Route of administration is a central question for the development of cell therapy of genetic diseases of the skel-

P130

RIMINUCCI ET AL.

eton. Despite claims to the contrary,(48) systemic infusion does not seem at present as a viable option. True “engraftment” (i.e., anatomical and functional reconstitution of the osteogenic lineage) has never been convincingly proven after systemic transplantation of skeletal stem cells. However, local administration of skeletal stem cells is feasible and actually provides the basis for current approaches to bone tissue engineering.(49,50) One can conceive of the use of autologous, nonmutated skeletal progenitors for treatment of individual FD lesions or individual events in the natural history of a lesion.

Gene therapy Stem cells are the ideal vehicles for gene therapy. The significance of skeletal stem cells for gene therapy in bone has remained elusive and yet has not escaped attention altogether. As emanating from a mutation that is both dominant and gain-of-function, FD is perhaps a paradigm of the most difficult situation that gene therapy can face. Even in stem cells, replacement of a missing gene (as is the case, for example, in human SCID caused by adenosine deaminase deficiency, a condition successfully treated by gene therapy(51)) seems a conceptually simpler task compared with the specific silencing of the mutated allele that one would have to accomplish in FD. Nonetheless, the development of strategies for silencing genes in human cells (by, for example, RNA interference(52,53)) provides the necessary basic tools for initial consideration of gene therapy in FD. In skeletal stem cells, different genes, including the Gs␣ gene, can effectively and stably be silenced by lentiviral vectors encoding short hairpin precursors of RNA interfering sequences, driven by PolIII-dependent promoters.(47,54) Compared with other strategies (e.g., gene targeting in skeletal stem cells(55)), this approach seems simpler and at least experimentally feasible. Selective silencing of the mutated allele, however, may present additional difficulties. These difficulties (and other general problems associated with gene therapy per se) notwithstanding, current work in this area, are not only scientifically attractive but emanate from the need to find a cure for a crippling disorder. Cure of FD is not currently feasible by any current conventional means. Should cure of FD become feasible in the future, it will be, by default, through innovative approaches. Work that is done or planned today goes from today’s bench to tomorrow’s bedside.

ACKNOWLEDGMENTS Work covered in this article was supported by grants from Telethon Fondazione Onlus (Grant GGP04263), MIUR, and the EU (GENOSTEM) to PB and by the DIR, NIDCR, Intramural Research Program, NIH (PGR).

REFERENCES 1. Weinstein LS, Shenker A, Gejman PV, Merino MJ, Friedman E, Spiegel AM 1991 Activating mutations of the stimulatory G protein in the McCune-Albright syndrome. N Engl J Med 325:1688–1695.

2. Schwindinger WF, Francomano CA, Levine MA 1992 Identification of a mutation in the gene encoding the alpha subunit of the stimulatory G protein of adenylyl cyclase in McCuneAlbright syndrome. Proc Natl Acad Sci USA 89:5152–5156. 3. Shenker A, Weinstein LS, Sweet DE, Spiegel AM 1994 An activating Gs alpha mutation is present in fibrous dysplasia of bone in the McCune-Albright syndrome. J Clin Endocrinol Metab 79:750–755. 4. Bianco P, Robey PG, Weintroub S 2003 Fibrous dysplasia. In: Glorieux FH, Pettifor JM, Juppner H (eds.) Pediatric Bone. Academic Press, San Diego, CA, USA, pp. 509–539. 5. Ippolito E, Bray EW, Corsi A, De Maio F, Exner UG, Robey PG, Grill F, Lala R, Massobrio M, Pinggera O, Riminucci M, Snela S, Zambakidis C, Bianco P 2003 Natural history and treatment of fibrous dysplasia of bone: A multicenter clinicopathologic study promoted by the European Pediatric Orthopaedic Society. J Pediatr Orthop B 12:155–177. 6. Bourne HR, Landis CA, Masters SB 1989 Hydrolysis of GTP by the alpha-chain of Gs and other GTP binding proteins. Proteins 6:222–230. 7. Yamamoto T, Ozono K, Kasayama S, Yoh K, Hiroshima K, Takagi M, Matsumoto S, Michigami T, Yamaoka K, Kishimoto T, Okada S 1996 Increased IL-6-production by cells isolated from the fibrous bone dysplasia tissues in patients with McCune-Albright syndrome. J Clin Invest 98:30–35. 8. Okamoto S, Hisaoka M, Ushijima M, Nakahara S, Toyoshima S, Hashimoto H 2000 Activating Gs(alpha) mutation in intramuscular myxomas with and without fibrous dysplasia of bone. Virchows Arch 437:133–137. 9. Silva ES, Lumbroso S, Medina M, Gillerot Y, Sultan C, Sokal EM 2000 Demonstration of McCune-Albright mutations in the liver of children with high gammaGT progressive cholestasis. J Hepatol 32:154–158. 10. Happle R 1986 The McCune-Albright syndrome: A lethal gene surviving by mosaicism. Clin Genet 29:321–324. 11. Shenker A, Weinstein LS, Moran A, Pescovitz OH, Charest NJ, Boney CM, Van Wyk JJ, Merino MJ, Feuillan PP, Spiegel AM 1993 Severe endocrine and nonendocrine manifestations of the McCune-Albright syndrome associated with activating mutations of stimulatory G protein GS. J Pediatr 123:509–518. 12. Spiegel AM 1996 Mutations in G proteins and G proteincoupled receptors in endocrine disease. J Clin Endocrinol Metab 81:2434–2442. 13. Stern CD 2004 Gastrulation: From Cells to Embryo. Cold Spring Harbor Laboratory, Cold Spring Harbor, MA, USA. 14. Le Douarin N, Kalcheim C 1999 The Neural Crest. Cambridge University Press, Cambridge, UK. 15. Olsen BR, Reginato AM, Wang W 2000 Bone development. Annu Rev Cell Dev Biol 16:191–220. 16. Neuhauss SC, Solnica-Krezel L, Schier AF, Zwartkruis F, Stemple DL, Malicki J, Abdelilah S, Stainier DY, Driever W 1996 Mutations affecting craniofacial development in zebrafish. Development 123:357–367. 17. Dorfman HD, Czerniak B 1998 Bone Tumors. Mosby, St Louis, MO, USA. 18. Corsi A, De Maio F, Ippolito E, Cherman N, Gehron Robey P, Riminucci M, Bianco P 2006 Monostotic fibrous dysplasia of the proximal femur and liposclerosing myxofibrous tumor: Which one is which? J Bone Miner Res (in press). 19. Fryxell KJ, Moon WJ 2005 CpG mutation rates in the human genome are highly dependent on local GC content. Mol Biol Evol 22:650–658. 20. Cooper DN, Krawczak M 1990 The mutational spectrum of single base-pair substitutions causing human genetic disease: Patterns and predictions. Hum Genet 85:55–74. 21. Bartolomei MS, Tilghman SM 1997 Genomic imprinting in mammals. Annu Rev Genet 31:493–525. 22. Santos F, Dean W 2004 Epigenetic reprogramming during early development in mammals. Reproduction 127:643–651. 23. Reik W, Davies K, Dean W, Kelsey G, Constancia M 2001 Imprinted genes and the coordination of fetal and postnatal growth in mammals. Novartis Found Symp 237:19–31. 24. Liu J, Yu S, Litman D, Chen W, Weinstein LS 2000 Identifi-

FIBROUS DYSPLASIA AND STEM CELLS

25.

26. 27.

28. 29. 30.

31.

32.

33.

34. 35. 36.

37.

38.

39.

40.

41.

42.

cation of a methylation imprint mark within the mouse Gnas locus. Mol Cell Biol 20:5808–5817. Robertson KD, Jones PA 1997 Dynamic interrelationships between DNA replication, methylation, and repair. Am J Hum Genet 61:1220–1224. Smith AG 2001 Embryo-derived stem cells: Of mice and men. Annu Rev Cell Dev Biol 17:435–462. Gardner RL 2004 Pluripotential stem cells from the vertebrate embryo. In: Lanza R (ed.) Handbook of Stem Cells, vol. 1. Academic Press, New York, NY, USA, pp. 15–26. Jaffe HL 1958 Tumors and Tumorous Conditions of Bones and Joints. Lea & Febiger, Philadelphia, PA, USA. Lichtenstein L 1977 Bone Tumors. Mosby, St Louis, MO, USA. Spjut HJ, Dorfman HD, Fechner RE, Ackerman LV 1978 Tumors of bone and cartilage. Atlas of Tumor Pathology, 2nd series, vol. 5. AFIP, Washington, DC, USA. Fechner RE, Mills SE 1992 Tumors of the bones and joints. Atlas of Tumor Pathology, 3rd series, vol. 8. AFIP, Washington, DC, USA. Riminucci M, Fisher LW, Shenker A, Spiegel AM, Bianco P, Gehron Robey P 1997 Fibrous dysplasia of bone in the McCune-Albright syndrome: Abnormalities in bone formation. Am J Pathol 151:1587–1600. Bianco P, Kuznetsov S, Riminucci M, Fisher LW, Spiegel AM, Gehron Robey P 1998 Reproduction of human fibrous dysplasia of bone in immunocompromised mice by transplanted mosaics of normal and Gs-alpha mutated skeletal progenitor cells. J Clin Invest 101:1737–1744. Bianco P, Gehron Robey P 1999 Diseases of bone and the stromal cell lineage. J Bone Miner Res 14:336–341. Owen M, Friedenstein AJ 1988 Stromal stem cells: Marrowderived osteogenic precursors. Ciba Found Symp 136:42–60. Bianco P, Riminucci M, Majolagbe A, Kuznetsov SA, Collins MT, Mankani MH, Corsi A, Bone HG, Wientroub S, Spiegel AM, Fisher LW, Robey PG 2000 Mutations of the GNAS1 gene, stromal cell dysfunction, and osteomalacic changes in non-McCune-Albright fibrous dysplasia of bone. J Bone Miner Res 15:120–128. Karadag A, Riminucci M, Bianco P, Cherman N, Kuznetsov SA, Nguyen N, Collins MT, Robey PG, Fisher LW 2004 A novel technique based on a PNA hybridization probe and FRET principle for quantification of mutant genotype in fibrous dysplasia/McCune-Albright syndrome. Nucleic Acids Res 32:e63. Bianco P, Kuznetsov S, Riminucci M, Cherman N, Fisher LW, Collins MT, Wientroub S, Gehron Robey P 2001 Age dependent demise of mutated stem cells and “normalization” of fibrous dysplasia of bone. Bone 28:S125. Sissons HA, Malcolm AJ 1997 Fibrous dysplasia of bone: Case report with autopsy study 80 years after the original clinical recognition of the bone lesions. Skeletal Radiol 26:177–183. Leet AI, Chebli C, Kushner H, Chen CC, Kelly MH, Brillante BA, Robey PG, Bianco P, Wientroub S, Collins MT 2004 Fracture incidence in polyostotic fibrous dysplasia and the McCune-Albright syndrome. J Bone Miner Res 19:571–577. Riminucci M, Collins MT, Fedarko NS, Cherman N, Corsi A, White KE, Waguespack S, Gupta A, Hannon T, Econs MJ, Bianco P, Gehron Robey P 2003 FGF-23 in fibrous dysplasia of bone and its relationship to renal phosphate wasting. J Clin Invest 112:683–692. Riminucci M, Kuznetsov SA, Cherman N, Corsi A, Bianco P,

P131

43.

44. 45. 46.

47.

48.

49. 50. 51.

52. 53. 54.

55.

Gehron Robey P 2003 Osteoclastogenesis in fibrous dysplasia of bone: In situ and in vitro analysis of IL-6 expression. Bone 33:434–442. Follenzi A, Ailles LE, Bakovic S, Geuna M, Naldini L 2000 Gene transfer by lentiviral vectors is limited by nuclear translocation and rescued by HIV-1 pol sequences. Nat Genet 25:217–222. Dull T, Zufferey R, Kelly M, Mandel RJ, Nguyen M, Trono D, Naldini L 1998 A third-generation lentivirus vector with a conditional packaging system. J Virol 72:8463–8471. Chan J, O’Donoghue K, de la Fuente J, Roberts IA, Kumar S, Morgan JE, Fisk NM 2005 Human fetal mesenchymal stem cells as vehicles for gene delivery. Stem Cells 23:93–102. Van Damme A, Chuah MK, Dell’Accio F, De Bari C, Luyten F, Collen D, Van den Driessche T 2003 Bone marrow mesenchymal cells for haemophilia A gene therapy using retroviral vectors with modified long-terminal repeats. Haemophilia 9:94–103. Piersanti S, Sacchetti B, Funari A, Remoli C, Michienzi S, Riminucci M, Saggio I, Bianco P 2006 Transfer and silencing of a disease gene (R201C-GNAS) in human skeletal stem cells (SSCs). Calcif Tissue Int 78(Suppl 1):S35–S36. Horwitz EM, Prockop DJ, Fitzpatrick LA, Koo WW, Gordon PL, Neel M, Sussman M, Orchard P, Marx JC, Pyeritz RE, Brenner MK 1999 Transplantability and therapeutic effects of bone marrow-derived mesenchymal cells in children with osteogenesis imperfecta. Nat Med 5:309–313. Bianco P, Gehron Robey P 2001 Stem cells in tissue engineering. Nature 414:118–121. Gehron Robey P, Bianco P 2004 Stem cells in tissue engineering. In: Lanza RP (ed.) Handbook of Adult and Fetal Stem Cells. Academic Press, New York, NY, USA, pp. 785–792. Aiuti A, Slavin S, Aker M, Ficara F, Deola S, Mortellaro A, Morecki S, Andolfi G, Tabucchi A, Carlucci F, Marinello E, Cattaneo F, Vai S, Servida P, Miniero R, Roncarolo MG, Bordignon C 2002 Correction of ADA-SCID by stem cell gene therapy combined with nonmyeloablative conditioning. Science 296:2410–2413. Hannon GJ 2002 RNA interference. Nature 418:244–251. Paul CP, Good PD, Winer I, Engelke DR 2002 Effective expression of small interfering RNA in human cells. Nat Biotechnol 20:505–508. Piersanti S, Sacchetti B, Funari A, Di Cesare S, Bonci D, Cherubini G, Peschle C, Riminucci M, Bianco P, Saggio I 2006 Lentiviral transduction of human postnatal skeletal (stromal, mesenchymal) stem cells: In vivo transplantation and gene silencing. Calcif Tissue Int 78:372–384. Chamberlain JR, Schwarze U, Wang PR, Hirata RK, Hankenson KD, Pace JM, Underwood RA, Song KM, Sussman M, Byers PH, Russell DW 2004 Gene targeting in stem cells from individuals with osteogenesis imperfecta. Science 303:1198– 1201.

Address reprint requests to: Paolo Bianco, MD Dipartimento di Medicina Sperimentale e Patologia Universita’ La Sapienza Viale Regina Elena 324 00161 Rome, Italy E-mail: [email protected] Received in original form October 2, 2006; revised form October 2, 2006; accepted October 12, 2006.