Fortnightly Club of Redlands

Dr. David J. Baylink is married to Colleen Baylink and has three children: Tim, Paul, and Katie. He received his M.D. degree in 1957 from Loma Linda University. Dr. Baylink subsequently took several years of post graduate clinical and research training including fellowships at Harvard University and Hammersmith Hospital in London, England. In 1966, he accepted the faculty position at the University of Washington in Seattle where he developed a large research program in Mineral Metabolism. In 1981, Dr. Baylink was recruited to Loma Linda University to set up a Mineral Metabolism Laboratory. This laboratory, with over 50 employees, is one of the largest of its kind in the world. Currently, Dr. Baylink is Distinguished Professor of Medicine, Biochemistry, Pediatrics, Orthopedic Surgery, and Oral Surgery. He has published several hundred scientific papers and book chapters in the field of Mineral Metabolism. Dr. Baylink has received numerous honors and is on the editorial board of several scientific journals. He is the only graduate of Loma Linda University to be elected into the Society of Clinical Investigation and into the American Association of Physicians. Dr. Baylink has trained over 50 clinical and basic fellows from various portions of the United States and from Europe and Asia. His hobbies include tennis, dirt bike riding, snow skiing and reading fiction.

Osteoporosis is a common disease that affects millions of patients throughout the world. We anticipate that both the diagnosis and the treatment of this disease will be revolutionized by the integration of genomics and informatics. It is predicted that a genetic algorithm will be developed to identify at risk patients before they develop osteoporosis, so that preventive measures can be instituted. The sequencing of the human genome will lead to revolutionary advances in at least three areas of osteoporosis therapy: small molecule therapy, protein therapy and gene therapy. One area of focus for future therapeutics in osteoporosis will be on osteogenic agents, which should have a high likelihood of success because the skeleton has the innate capacity to regenerate itself.

OSTEOPOROSIS is a disease of too little bone, which results in fragility fractures that occur with very little trauma. It is becoming progressively more common, partly becuase fractures that occur with very little trauma. It is becoming progressively more common, partly because it is a disease that increases in frequency in patients who are over 60 years of age, a segment of the population that is progressively increasing¹.

Photomicrograph of trabecular bone from (a) a normal susbject and (b) an osteoporotic patient. The loss of bone in the osteoporotic skelaton markedly reduces the strength of the corresponding bone, which in turn leads to fragility fractures. Reproduced from Ref.3.

Osteoporosis is much more common in elderly females than in elderly males because the estrogen deficiency that occurs at the time of menopause leads to increased bone destruction, which is not compensated by a corresponding increase in bone formation (i.e. a negative bone balance), resulting in bone loss and, eventually, osteoporosis in many females:^2,3.

The cause of osteoporosis is a combination of a low peak bone density⁴ and subsequent bone loss, which in females accelerates at menopause because of estrogen deficiency. Estrogen deficiency leads to bone resorption and is a major cause of bone loss in the immediate postmenopausal years.

Fig. 2. A model showing the two causes of osteoporosis. osteoporosis is defined as a bone dentiy of 2.5 or more standard deviations below the average bone density in young individuals (the fracture threshold)4, and results from either a low peak bone density or from a normal peak bone density with a rapid bone loss. note that in both cases, the bone density drops below the fracture threshold. Patients with osteoporosis experience a combination of these two causes of low bone density. (another potential mechanism, analogous to rapid loss, would be a modest loss over a prolonged period of time, such tha the bone density drops below the fracture threshold.).

However, in elderly females there are undoubtedly other agents that contribute to the net negative bone balance, such as dietary changes and reduced physical activity.

Osteoporosis is currently diagnosed by a bone densitometer, which determines whether the bone density is sufficiently low to make the patient at risk for fractures. This approach has the advantage of leading to a diagnosis before the morbidity of a fracture ensues, thereby allowing the use of preventive therapy. The major problem with this diagnostic procedure is that it is not available to all individuals over 50 years of age, despite the fact that >40% of women who are over 50 years of age are destined to develop an osteoporotic fractures. Moreover, because of the costs, it seems unlikely that there would be universal screening with bone densitometers to diagnose osteoporosis in the foreseeable future.

In the last decade, advances made by the pharmaceutical industry have led to the development of effective agents for the treatment of osteoporosis.

(Table 1). Abbreviations: FDA, US Food and Drug Administration; PTH, parathyroid hormone. These drugs were chosen because they are representative of US osteoporotic drugs. However there are many additional osteoporotic drugs being developed and at least four types are now in clinical trials (selective estrogen receptor modulators, bisphosphonates, fluorides and PTH analogs).

Most of these agents are antiresorptive drugs, which is a highly appropriate form of therapy because bone resorption in the elderly tends to be elevated. Antiresorptive therapies tend to cause a modest increase in bone density because they decrease bone turnover, which decreases the remodeling space and increases the average bone age, leading to a greater average mineral concentration in bone matrix. However, the most effective treatment of severe osteoporosis requires two different classes of drugs: (1) antiresorptives, which will prevent further bone loss and cause modest increases in bone density (up to 10%)7; and (2) osteogenic or bone formation drugs, which replenish the lost bone tissue to a level that is compatible with normal bone strength (Fig. 3). The rationale for osteogenic therapy in patients with severe osteoporosis is that such patients have decreases in bone density that are much greater than can be corrected by antiresorptive therapy.

Based on clinical trials with parathyroid hormone (PTH) in the form of PTH (1-34), which is the active portion of this 84 amino acid hormone, and with different fluoride preparations, we now know that bone density can be markedly increased with these two bone-forming agents^8,9. Furthermore, fluoride has been shown to cause large increases in bone density, even in patients who are over 80 years of age, suggesting that, at least in many elderly patients, there are sufficient osteoblast precursor cells to treat osteoporosis. Therefore, there is an attractive opportunity to develop agents which will regenerate bone even in those patients with severely depleted skeletons. However, these two agents have some disadvantages that will probably preclude their use to correct completely the osteoporotic bone deficit: PTH therapy causes tolerance" and only modest increases in hip bone density, while use of fluoride can result in the production of bone of poor quality¹².

Future prospects for the application of genetic risk assessment in the diagnosis of osteoporosis

It is likely that, in the next one or two decades, sufficient genetic information will become available to develop genetic algorithms to assess the risk for osteoporosis. The marriage between informatics and genomics could lead to a DNA profile on all subjects, perhaps performed at birth, that will provide information on the risk for various chronic diseases, including osteoporosis. Once these risks are identified, preventive measures for the corresponding disease risks can be instituted before the development of these diseases.

The key diagnostic goal will be to develop a genetic algorithm that will provide the risk of developing osteoporosis for all subjects at birth. We discuss this goal below in the context of: (1) potentially relevant bone processes that could be influenced by heritability; (2) methodological approaches that are being used to identify osteoporosis candidate genes, particularly peak bone density genes; and (3) the potential advantages of and difficulties associated with the development of genomic risk assessment for diagnosis of osteoporosis.

There is strong evidence that peak bone density, which occurs during the third decade of life, is ~70% genetically deterrnined¹³. There is also evidence for the genetic determination of bone size, which is another important determinant of bone strengths. Box 2 shows a list of putative candidate genes that could influence peak bone density.

Vitamin D receptor	Insulin-like growth factor 1
Estrogen receptor	Parathyroid hormone
Calcium receptor	Intrleukin 1 receptor agonist
Calcitonin receptor	Type 1 collagen
Beta 3 adrenergic receptor	Osteocalcin
Interleukin 6	Alpha II HS glycoprotein
Transforming growth factor Beta	Collagenase

Recently, twin studies have shown that the onset of menopause is genetically influenced, with a heritability index of ~62% (Ref. 15). This could be a significant observation because the earlier the menopause, the longer the period of estrogen deficiency. The effects of acute estrogen deficiency on bone loss is well appreciated. Estrogen deficiency is also associated with bone loss in the elderly¹⁶.

The rate of bone loss increases markedly after the menopause but it is highly variable from one subject to another. In twin studies of postmenopausal females, only a small portion of the variance in the postmenopausal levels of both bone formation and bone resorption markers are explained by genetic factors¹⁷. These observations suggest that genetic factors influence the development of peak bone density and also the age of development of the menopause but do not appear to have a major impact on the rate of bone loss in the elderly.

The major focus of genetic diagnoses is to focus not on the patient with osteoporosis but on the patient who is at risk for osteoporosis and to tailor treatments based on specific genetic causes. Two major clinical approaches are currently used to identify genes contributing to complex diseases such as osteoporosis: (1) complete genome scans in family studies or in sib pairs using polymorphic genetic markers (linkage analysis); and (2) association studies of polymorphisms within candidate genes in osteoporosis and in case-controlled or population studies. As much more is known about genes that regulate peak bone density compared to genes that regulate the age of development of the menopause, our focus here will be on peak bone density genes.

Animal models, such as baboons and mice, are being used to identify genetic factors that contribute to peak bone density ^18,19. The advantage of the mouse model is that a large number of animals can be bred and studied under defined experimental conditions. Figure 4.

Fig. 4. Peripheral quantitative computerized tomography (pQCT) bone images from two inbred strains of mice. (a) The high density bone (from mouse strain C3H/HsJ) and (b) the low density bone (from mouse strain C57BU6J). Mineralized bone appears white; red at the peripheries repre Bents soft tissue edges and is not relevant to the differences in bone density being demonstrated. The medullary space is much larger in the C57BU6J than in the C3H/HeJ mouse. This is typical of the osteoporotic skeleton. These hero inbred strains of mice are crossed in order to perform quantitative trait loci (OTL) analyses, which will eventually identify the genes responsible for the differential bone density in these two strains of mice.

Two different inbred strains of mice, one with a relatively low bone density and one with a relatively high bone density (Fig. 4), are being studied using quantitative trait loci (QTL) analyses to identify genes that are responsible for the differential bone density. Thus far, we have found at least five statistically significant genetic loci responsible for this differential bone density between the C3H/HeJ and CS7BL16J mouse strains20. This approach has the advantage over the types of human studies done thus far in that, theoretically, it can identify genes with only a modest effect on bone density. This is particularly relevant because, in chronic diseases, the genetic contribution usually involves several genes, each of which contributes only a few percent to the phenotypes 22. The potential problem with mouse QTL studies is that there is no guarantee that a peak bone density gene in a mouse will predict the importance of this gene in determining peak bone density in humans.

Interestingly, one family study has identified a locus on chromosome 11 that confers a high bone density²³. Whether this gene plays a role in the general population is not yet known. In the future, it is likely that identification of the majority of genes involved in determining peak bone density (including those with only a modest effect) will require a combination of linkage studies in humans and experimental animals, and association studies of candidate genes in humans. Once a candidate bone density gene is identified, functional studies will be necessary to show that the candidate gene influences bone density in order to justify using this gene in genetic testing of the risk for developing osteoporosis.

The successful completion of a genetic algorithm to predict osteoporosis risk, although technically feasible, will be very time consuming and costly.

Many elderly patients with osteoporosis have both a high bone destruction rate, requiring antiresorptive drugs, and a very low bone density that would benefit from osteogenic therapy when it becomes available. Advances in biotechnology and informatics have the potential to lead to new antiresorptive and osteogenic therapies. These will fall into at least three categories: (1) small molecules; (2) proteins; and (3) gene therapy. This third category includes cell therapy, a means by which to deliver gene therapy^24,25. The ideal therapy will be safe, therapeutically effective and specific to the target (i.e. without side effects). Another level of specificity might be required because there are probably several subtypes of the disease. It seems likely that at least some of these subtypes will require drug therapy that is specific for the individual subtype²⁶.

Box 1. Potential advantages and difficulties associated with the development of genomic risk assessment to diagnose osteoporosis

a KhouTy, M.J. and Wagerer, D.K. t1995) Epidemiological evaluation of the use of genetics to improve the predictive value of disease risk factors, Am. J. Hum. Genet. s6, 83s-844

b Moskowitz, D.W. (i996) Genetic screening and diagnosis of disease, Moi. Idled. Today 2, 27s

c Poste, G. (1998) Molecular medicine and informati~:~l~ta— healthcare,Nat.Biotechnol. 16,19-21 ~ ~ :~: ~ ~: ~

d Risch, N. and Merikangas, K. (1996) The future of genetic studies of complex human diseases, Science 273, 1516-1517

In the next 20 years, it has been estimated that the Human Genome Project will provide biochemical targets for drug interaction that will be at least one order of magnitude greater than has been previously utilized in the entire history of drug development. An estimated 500-10 000 targets will be relevant to multifactorial diseases2'. The challenge in the future will be to choose the targets that are most likely to be both safe and efficacious²⁸.

One potential target that appears to be specific for bone is the mechanism by which fluoride causes large increases in bone formation and bone density^8,10. There is now strong evidence that the molecular mechanism of action of fluoride involves inhibition of a bone-specific phosphotyrosyl protein phosphatase (Fig. 5)²⁹, Thus, the identification and characterization of this enzyme, or enzyme family, could constitute an important advance for future osteogenic therapy²⁹. Other targets for low molecular weight osteogenic therapies for osteoporosis include growth factors/cytokines and their receptors, growth factor signaling molecules (kineses and phosphatases) and osteoblast-specific transcription factors [e.g. core-binding factor al (Cbfa1)]³⁰.

An example of a target for small molecule therapy for bone resorption is cathepsin K, which might be exclusive to the osteoclast ³¹. Antisense studies show that cathepsin K is essential for osteoclastic bone resorption³², and pycnodysostosis, a genetic disease that is characterized by osteosclerosis, has recently been shown to be caused by nonsense or missense mutations in the gene encoding cathepsin K (Ref. 33). Accordingly, small molecule therapy to inhibit this enzyme could be an effective antiresorptive medication. Recent cell and molecular biological studies of the osteoclast are providing a multitude of additional putative targets for small molecule therapy to inhibit bone resorption, including osteoprotegerin, a specific inhibitor of osteoclast activity³⁴.

In addition to specificity, the ideal small molecule therapeutic agent will have appropriate absorption, distribution and metabolism profiles³⁵. All the US Food and Drug Administration (FDA) approved antiresorptive drugs available today for the treatment and prevention of osteoporosis (Table 1) have reasonable profiles in terms of efficacy and side effects, but some have marginal profiles for absorption, distribution and metabolism. None of these drugs is ideal with respect to all these criteria, and further research is required to provide improved antiresorptive therapies. Furthermore, an effective stimulator of bone formation is a major therapeutic need for osteoporosis. One of the reasons for the lack of effective osteogenic agents is that most of the known agents that stimulate bone formation, including all the major growth factors, such as the fibroblast growth factors (FGFs), bone morphogenetic proteins (BMPs), plateletderived growth factors (PDGFs), the transforming growth factor beta (TGFbeta) family and the insulin-like growth factors (IGFs), as well as prostaglandin E2, are not specific for bone and all have widespread receptors throughout all organ systems-6. Therefore, it seems likely that successful small molecule osteogenic therapy will require either more specific targets or special means by which to target nonspecific drugs to the skeleton.

Currently, there is one protein drug that is approved for the treatment of osteoporosis salmon calcitonin, which can be administered by nasal spray. This protein therapy has very few side effects. PTH (134) is a protein molecule that is in Phase III clinical trials for osteoporosis. Another protein therapy currently in Phase III clinical trials for anabolic therapy, and that might prove applicable to osteoporosis, is a complex of insulin-like growth factor 1 UGFl) and insulin-like growth factor binding protein 3 (IGFBP3) (Ref. 37). This complex has a much longer serum half-life and is less hypoglycemic than IGF alone, a pharmacokinetic feature which is thought to be a therapeutic advantaged. Another therapeutic agent that is in clinical trials and that is anabolic for bone is growth hormone-releasing peptide (GHRP), which increases serum growth hormone levels and stimulates bone formation³³. The remarkable advantage of this agent over other agents that stimulate the growth hormone/IGFaxis is that it can be administered orally. All agents that increase circulating levels of growth hormone or the IGFs also increase bone resorption, so these types of therapies will require concurrent antiresorptive therapies.

There is renewed optimism for protein therapy because of the anticipated advances in the drug delivery of proteins³⁹. As well as the nasal and oral routes, advances are being made in subcutaneous delivery systems, which deliver drugs without interruption over a period of several weeks or longer and which have negative feedback regulations. The discovery of an osteogenic agent specific for osteoblasts would secure a role for such protein therapy in the treatment of osteoporosis.

Some of the candidate genes identified by the Human Genome Project are likely to encode osteogenic proteins²⁷ and therefore have potential as gene therapy agents. Osteogenic gene therapy has potential because the skeleton has the innate capacity to regenerate itself (fracture healing is a form of regeneration). In support of the hypothesis that bone will be responsive to agents capable of regenerating the skeleton is a preclinical study in which BMP4 and PTH genes, either alone or in combination, effectively increased bone formation in rats⁴⁰. Additional evidence that osteogenic growth factors stimulate bone formation and regenerate the skeleton is that the local introduction of growth factor proteins into bony defects produces large increases in bone formation and the regenerative repair of defects⁴¹.

Observations in humans and in preclinical studies suggest that gene therapy could be used to treat bone disorders. In Fig. 6, bladder carcinoma cells that have metastasized to bone are shown. These cells markedly stimulate local bone formation, presumably by a mechanism that involves the production of osteogenic protein(s) by the metastatic cancer cells. Prostate cancer metastases are also osteogenic and have been shown to produce several kinds of growth factors, including the BMPs (Ref. 42). Cells from the urogenital tract have been known for many years to stimulate bone formation. Cells transfected with the appropriate growth factor genes and injected into appropriate sites in the marrow space should therefore stimulate bone formation at these sites. Because several growth factors are involved in the normal proliferation and differentiation of osteoblasts in bone formation processes36, optimum gene therapy for bone regeneration might require several genes that encode different growth factors. Such information will be a prerequisite for optimal gene therapy for bone regeneration.

There are several potential applications of gene therapy for local skeletal abnormalities, and it should be possible to use gene therapy in the future to improve the healing of fractures, both traumatic and osteoporotic. In addition, it might be possible to use growth factor genes, rather than autologous bone grafts, to promote fusion of vertebral bodies in patients with degenerated vertebral discs. However, the major challenge of gene therapy for the skeleton will be to apply this therapy systemically in order to regenerate either fracture-prone sites, such as the hip, or the entire osteoporotic skeleton. The first goal in gene therapy for the skeleton, whether it is local therapy or systemic therapy, will be to learn how to efficiently deliver the gene or genes of interest to the appropriate skeletal cells.

En vivo gene therapy should provide an effective means of gene delivery but requires special facilities (good manufacturing practice) for ex vivo technical procedures. With respect to in viva gene delivery, we and others have found that both differentiated bone cells, as well as undifferentiated stroll cells, are somewhat resistant to gene transfer. Hopefully, this problem will be resolved in the near future as gene delivery methods improved ^{43, 44}.

Figure 5. A model of the mechanism of action of fluoride an osteogenic agent Growth factors, such as insulin-like growth factor 1 (IGF1), stimulate tyrosine phosphorylation of mitogenic signaling proteins. Ordinarily, such signals are subsequently eliminated by endogenous phosphotyrosyl protein phosphatases (PTPPs). However, in patients treated with fluoride, there is an inhibition of a bone-specific PTPP, which leads to a prolongation of the bone mitogenic signal. This PTPP, when its sequence is identified, could serve as a target for small molecule osteogenic therapy for osteoporosis. Adapted from Ref. 8.

Successful systemic gene therapy to regenerate the entire osteoporotic skeleton will also require advances in gene targeting. This is particularly true with growth factors: the systemic administration of growth factor genes could lead to undesirable side effects from actions in several organs in the body. Monoclonal antibodies have been developed that specifically identify preosteoblasts and differentiated osteoblasts, and additional monoclonal antibodies to target other osteoblast differentiation stages could be developed using phage display libraries^45,46. The incorporation of the epitope-binding region of these monoclonal antibodies into viral envelopes or liposomes could provide a way to target genes specifically to bone cells of a given differentiation stage. Not only do we need to target the osteogenic gene to the skeleton but also, in some cases, we need to target the gene to a specific site in the skeleton, such as the hip, which is one of the major sites of osteoporotic fractures. If targeting mechanisms for selective skeletal sites do not become available in the near future, we could provide gene therapy by local inoculation into the skeletal site of interest.

Figure 6. Photomicrograph of a Haematoxylin Eosin stained bone section from a patient with bladder cancer that has metastasized to bone. wpe1.jpg (8574 bytes)

The osteoblasts are large and plump and exhibit an extensive Golgi apparatus (the unstained clear area in the cytoplasm of the osteoblast), all features of cells that are very active in bone matrix synthesis. In addition, osteoblasts occupy the majority of the available bone surface. In bone sections from normal adults, the osteoblasts are smaller, flatter and occupy 15% of the trabecular bone surface. This increase in osteoblast cell number and size is probably the consequence of the metastatic bladder cells, which presumably produce osteogenic agents to stimulate osteoblast proliferation and activity. These metastatic cells suggest that if normal marrow cells were transfected with the appropriate growth factor genes, a marked stimulation of osteoblast number and activity and, thus, bone formation could occur. Osteoblastic metastases to bone are somewhat unusual; most metastatic cancer to bone is osteolytic and results in bone destruction.

Another aspect of gene therapy is to develop vectors that achieve the desired pharmacokinetics of the protein produced. An elegant resolution of this potential problem might be to include an inducible promoter into the gene constructs. Such promoters could be induced by orally active drugs that are specific for the promoter. This should provide a way to control the local bone level of the osteogenic gene product and to control the duration of treatment with the local growth factor gene product.

The prospects for molecular medicine reflected upon in this review strongly support the view that one or more of the three therapeutic approaches described here will lead to a systemic therapy to regenerate the skeleton. One reason for this optimism is evidence that suggests that many aged individuals retain the ability to respond to experimental osteogenic drugs with large increases in bone formation and in bone density. Advances in genomic risk assessment have the potential to make osteoporosis not only treatable but also preventable.

1. National Osteoporosis Foundation (1998) Osteoporosis: review of the evidence for prevention, diagnose and treahnent and cost effective analyst, Osteoporosis Int. Suppl, 4, S7-S80

2 Riggs, B.L., K'aosla, S. and Melon, L.J., III (1998) A unitary model for involm tional osteoporosis: estrogen deficiency causes both type I and type 11 often paresis in postmenopausal women and contributes to bone loss in aging men, J. Bone Miner. Res. 13, 763-773.

3 Baylink, D.l., Jennings, J.C. and Mohan, S. (]999) Calcium and bone homes!: stasis and changes with aging, in Principles of Geriatric Medicine and Gerontology (4th edn) (Hazzard, W.R. et al., ads), pp. 1041-1056, McGrawHill.

4 Teegardeo, D. et al. (1995) Peak bone mass in young women, J. Bone Miner. Res. 10, 711-715.

5 Kanis, J.A. et al. (1994) Perspective: The diagnosis of osteoporosis, J. Bone Miner. Res. 9,1137-1141.

6 Melton, L.J. et aL (1992) Perspective.. How many women have osteoporosis? J. Bone Miner. Res. 7,1005-1010.

7 Liberrnan, U. et aL (1995) Effect of oral alendronate on bone minerd density and the incidence of fractures in postmenopausal osteoporosis, New EngL J. Med. 333, 1437-1443.

8 Lau, KH.W., Akesson, K., Libanati, C.R. and Baylink, D.1. (1998) Osteogenic actions of fluoride: its therapeutic use for estabBsbed osteoporosis, in Anabolic Treatments for Osteoporosis (Whitfield, J.E and Morely, P., ads), pp. 207-254, CRC Press

9 Cosman, E and Lindsay, R. (1998) Is parathyroid hormone a therapeutic option for osteoporosis? A review of the conical evidence, Calcif Tissue Int. 62, 475-480.

10 Parley, S.M.G. et al. (1990) Fluoride therapy for osteoporosis promotes a progressive increase in spied bone density, J. Bone Miner. Res. 5, S37-S42.

11 Neer, R. et aL (1990) Treatment of postmenopausal osteoporosis with daPyPTH plus calcitriol, in Osteoporosis (3rd vol.) (Cbristiansen, C. and Overgaard, K., eds), pp. 1314-1317, Academic Press

12 Turner, C. and Dunipace, A. (1993) On fluoride and bone strength, Calcif Tissue Int. 53, 289-290.

13 Slemenda, C.W. et al. (1991) Genetic determinants of bone mass in adult women: a reevaluation of the twin model and the potential importance of gene interaction on heritability estimates, J. Bone Miner. Res. 6, 561-567.

14 Brinckmann, R. Biggemann, M. and Hilweg, D. (1989) Prediction of the compressive strength of human lumbar vertebrae, Spine 14, 606-610.

15 Snieder, H., MacGregor, A.J. and Spector T.D. (1998) Genes control the cessation of a woman's reproductive life: A twin study of hysterectomy and age at menopause, J. Clin. Endocrinol. Metab. 83, 1875-1880.

16 Cummings, S.R. et aL (1998) Endogenous hormones and the risk of hip and vertebral fractures among older women, Study of osteoporotic fractures research group, New EnBL J. Med. 11, 733-738.

17 Garnero, P. et aL (1996) Genetic influence on bone turnover in postmenopausal twins, J. Clin. Endocrw~ol. Afesab. 81, 140-146.

18 Beamer, W.G., Donabue, L.R., Rosen, C.J. and Baylink. D.J. (1996) Genetic vari. abilib in adult bone density among inbred strains of mice, Bone 18, 397-403.

19 Stewart, A. (1998) Baboons provide clues to osteoporosis, MoL Med. Today 4, 3

20 Beamer, W.G. et al. (1998) Location of genes regulating volumetrtc bone mineral density in C57BL/61 (low) and C3H/HeJ (high) inbred stralas of mice, Bone 23, 51-62.

21 Risch, N. and Melikangas, K (1996) The future of genetic dudies of complex human diseases, Science 273, 1516-1517.

22 Housman, D. and Ledley, F.D. (1998) Why pharmacogenomics? Why now? Nat. 8iotechnol. 16, 492-493.

23 Johnson, M.L. et al. (1997) Linkage ol a gene causing bigh bone mass to human chromosome 11, Atn. I. Hum Genet. 60, 233-239.

24 Prockop, D.J. (1997) Marrow stromal cells as stem cells for nonnematopoleffc tissues, Science 276, 71-74.

26 Crooke, S.T. (1998) Opthnizing tbe impact of genornics on drug discovery and development, Nat. Biotechnol. 16, 29-30.

27 Drews, J. (1998) Biotechndogy's metamorphosis inh a drug discovery incus. try, Nat. BiotechnoL 16, 22-24.

28 Poste, G. (1998) Molecular medicine and informationbased targeted health. care, Nat. Biotechnol 16, 19-21.

29 Lau, KH.W. and Baylink, D.J. (1998) Molecular mechanism of action of fluoride on bone cells, J. Bone Miner. Res. 13, 1660-1738.

30 Ducy, P. et aL (1997) OSF2/Cbfal: A transcripffonal activator of osteablast differentiation, Cell 89, 747-754.

31 Dodds,~ R.A. et al. (1998) Cathepsin K mRNA detection is restricted h osteoclasts during fetal mouse development, J. Bone Minec Res. 13, 6736-8246.

32 Inui, T. et aL (1997) Cathepsin K antisense oligodeoxyoucleotide inhibits osteo. clastic bone resorption, ]. BioL Chern. 272, 8109-8112

33 Gelb, B.D., Sbi, aR, Chapman, H.A. and Desnick, R.J. (1996) Pyenodysostosis, a ]ysosomal disease caused by Catnepsin K deficiency, Science 273, 1236-1238.

34 Simonet, W.S. et aL (1989) Osteoprotegerin: A novel secreted protein involved in tbe regulaUon of bone density, Cell 89, 309-319.

35 Haseltine, W.A. (1998) The power of genomics to transform the biotechnology industry, Nat. Biotechnol. 16, 2527

36 Mohan, S. and Baylink, D.J. (1991) Bone growth factors, Clln. Orthop. Relat. Res. 263, 30-48.

37 Adams, S., Rosen, D. and Sonuner, A. (1998) rhIGFl{GFBP3 (Son~atoKine) therapy for the treatment of osteoporosis, in Molecular Mechanisms to Regulate the Actions of lGFs (Takana, K., Hirzuka, N. and Takahasi, S.I., eds), pp. 327-330, Elsevier.

38 Svensson, J. et aL (1998) Treabnent with the oral growth hormone secrete. gogue MK 677 increases markers of bone formation and bone resorption in obese young males, ]. Bone Miner. Res. 13,1158-1166.

40 Fang, J. et al. (1996) Sthnulation of new bone formation by direct transfer of osteogenic plasmld genes, Proc. Natl, Acad. Sci. U. S. A. 93, 5753-5758

41 Mohan, S. and Baylink, D.J. (1996) Therapeutic pobotial of TGF,0, BMP and FGF in the treahnent of bone loss, in Principles of Bone Biology (Bilezikian, J.P., Raisz,L.G. and Rodan, G.A., eds), pp. 1111-1123, Academic Press.

43 Friedmann, T (1997) Overcoming the obstacles to gene therapy, Sci Am 276, 96-101.

45 Sheets, M.D. et aL (1998) Efficient construction of a large nonimmune phage antibody library: The production of high~affinib human singlechain anti. bodies to protein antigens, Proc. NatL Acad. Sci. U. S. A. 95, 6157-6162.

Bmder, S.R, Horowitz, M.C., Mosca, J.D. and Haynesworth, S.E. (1997) Monoclonal anffbodies reacffve with human osteogenic cell surface anffgens, Bone 21, 225-235.

47 Stewart, A. (1998) Orally acffvated protein therapy heads towards clinical trials, MoL Med. Today 4, 95.

48 Delmas, RD. et al. (1997) Effects of raloxifene on bone mineral density, serum cholesterol concentrations, and uterine endometrium in postmenopausal women, New EngL J. Med. 337, 1641-1647.

69 Kraenzlin, M.E. et aL (1996) The effect of intranasal salmon calcitonin on postmenopausal bone turnover as assessed by biochemical markers: Evidence of maximal effect after 8 weeks of continuous treatment, Calcif T`ssue Int. 58. 216-220.

50 Akesson, K., Lau, KH.W. and Baylink, D.J. (1997) Raffonale for acffve vitamin D analog therapy in senile osteoporosis, Calcif T`ssue Int. 60, 100-105.