Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The relative contribution of vitamin D receptor (VDR), collagen type 1, α-1 (COL1A1), tumor necrosis… Taylor, Ian Wesley 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0077.pdf [ 10.13MB ]
Metadata
JSON: 831-1.0077371.json
JSON-LD: 831-1.0077371-ld.json
RDF/XML (Pretty): 831-1.0077371-rdf.xml
RDF/JSON: 831-1.0077371-rdf.json
Turtle: 831-1.0077371-turtle.txt
N-Triples: 831-1.0077371-rdf-ntriples.txt
Original Record: 831-1.0077371-source.json
Full Text
831-1.0077371-fulltext.txt
Citation
831-1.0077371.ris

Full Text

THE RELATIVE CONTRIBUTION OF VITAMIN D RECEPTOR ( V D R ) , C O L L A G E N T Y P E 1, a - l (COL1A1), T U M O R N E C R O S I S F A C T O R R E C E P T O R 2 (TNFR2), P O L Y M O R P H I S M S , P H Y S I C A L A C T I V I T Y A N D B O N E M I N E R A L - F R E E L E A N MASS T O B O N E PARAMETERS IN CHILDREN by IAN WESLEY TAYLOR B . H . K . in Exercise Science, The University of British Columbia, 2000 A thesis submitted in partial Mfilknent of the requirements for the degree of Master - of Science  in  T H E F A C U L T Y O F G R A D U A T E STUDIES  School of Human Kinetics  We accept this thesis as conforming to the required standard  The University of British Columbia December 2002 © Ian Wesley Taylor, 2002  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or by his or her representatives.  It is  understood  that  copying or  publication of this thesis for financial gain shall not be allowed without my written permission.  j  Derjartment of  HIAWIAM  kjw£ti£S  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  T)/? £ •M/wUv  7-3 2-OQ"Z f  Abstract  Background This study sought to investigate the relationship of physical activity (PA), dietary calcium and 3 candidate gene (VDR, COL1A1 and TNFR2) genotypes on bone mass in children (n = 327, age 10.33 ± 0.65). The study also sought to investigate the effect of PA and genotype on bone mineral-free lean mass (BMFL). Finally, the relationships between bone mass and B M F L and PA, B M F L and genotype and PA and genotype interactions were investigated.  Methods Anthropometric data (height, sitting height and weight) was determined using standard techniques. Bone mass, and B M F L were assed using total body D X A scan using a Hologic Q D R 4500. Dietary calcium, PA and maturity were assessed using previously validated questionnaires. V D R Fokl and V D R Bsml genotypes were determined by standard restriction fragment length polymorphism techniques. C O L I A l genotype was determined by a novel TaqMan technique and TNFR2 genotypes and haplotypes were determined by a novel automated sequenceing protocol. Associations between PA or candidate gene genotype and either B M F L or bone mass was first controlled for inter-subject differences in body size and maturity.  Results PA was significandy associated with B M F L in boys (p = 0.038). P A score was associated with a 3-5% difference in proximal femur BMC, femoral neck B M C and femoral neck aBMD but not lumbar spine BMC in boys as well as a 4% difference in femoral neck aBMD in girls. Average dietary calcium intake was not associated with differences in bone mass in children. V D R Bsml and V D R Fokl genotype did not have a relationship to bone mass or B M F L in children. C O L I A l Ss or ss genotype is associated with 4.8% higher femoral neck B M C in boys but not B M F L in either sex. T N F R 2 A593G gg genotype was associated with a 3.8% higher B M F L in boys (p =0.038) and a 3.4% higher femoral neck B M C in boys (p = 0.045). Girls with a TNFR2 T598G tg genotype had a 3.3% higher femoral neck B M C (p = 0.029). T N F R 2 G593-G598/G593-T598 haplotype was associated with a 10% higher femoral neck B M C in girls. For boys, when the B M F L by PA interaction term was added to the model it explained significandy more (2.5-3.9%, p = 0.004) of the variance in femoral neck BMC, femoral neck aBMD and lumbar spine aBMD than the main effect for P A alone. When the B M F L by V D R Fokl genotype interaction was added to the model the main effect of the V D R Fokl genotype became significant where boys with the FF genotype had a 1.4% greater femoral neck aBMD than boys with the F f or ff genotype. For girls, significant interactions between T N F R 2 haplotype and B M F L changed the model such that girls with the G593-G598/G593-T598 haplotype had a 10-11% greater femoral neck B M C or aBMD than girls with other TNFR2 haplotypes. Girls with the Ss or ss genotype had a 4% greater femoral neck aBMD after a significant interaction between COL1A1 genotype and P A was accounted for.  Conclusions High levels of PA are associated with increased B M F L and bone mass. T N F R 2 genotypes are associated with both lean mass and bone mass in a complex fashion suggesting that the TNFR2 genotypes and interactions between B M F L and TNFR2 genotypes affect and moderate a combined lean mass/bone mass effect. COL1A1 Ss and ss genotypes are associated with high bone mass particularly in girls with high PA.  ii  TABLE OF CONTENTS  ABSTRACT ii LIST O F T A B L E S vi LIST O F FIGURES viii A B B R E V I A T I O N S U S E D I N THIS THESIS xii ACKNO WLEGMENTS xiv CHAPTER 1 1 Introduction 1 1.1 Definition of Osteoporosis 1.2 Incidence and Health Care Cost Associated with Osteoporosis in Canada 1.4 The Relationship between Peak Bone Mass and Osteoporosis CHAPTER 2 3 Literature Background for the Proposed Study 3 2.1 Bone Cytology 2.1.1 Osteoblasts and Osteocytes 2.1.2 Osteoclasts 2.1.3 Osteoblast-Osteoclast Coupling 2.2 Bone Histology 2.3 The Dynamics of Bone Growth and Maintenance 2.4 The Bone-Remodeling Transient 2.5 Bone Densitometry 2.6 Three Dimensional Bone Measurement 2.8 Normal Patterns of Bone Growth in Children 2.9 Bone Strength 2.10 Lifestyle Factors 2.10.1 Dietary Calcium and Calcium Supplementation in Children and Adolescents 2.10.2 Physical Activity 2.10.2.1 Prospective Physical Activity Intervention Studies 2.10.3 Calcium and Physical Activity Interaction 2.11 Bone Mass as a Multifactorial Polygenic Trait 2.11.1 Heritability Estimates of Absolute Values of aBMD 2.11.2 Heritability Estimates of Bone Turnover Parameters 2.11.3 Heritability Estimates of Change in aBMD 2.11.4 Bone Parameter Segregation Analysis to Determine Models of Genetic Transmission. 2.12 Identifying Genes Related to Bone Mass 2.12.1 The Candidate Gene Approach 2.12.1.1 Evidence from the Candidate Gene Approach 2.12.2 Positional Cloning Approach 2.12.2.1 Human Studies Applying the Positional Cloning Approach 2.12.2.2 Animal Studies Applying the Positional Cloning Approach 2.13 The Vitamin D Receptor (VDR) 2.13.1 Human Studies of V D R Polymorphisms iii  1 2 2  3 3 6 8 8 10 11 12 13 14 15 17 18 22 24 27 28 28 29 29 30 30 31 31 31 32 35 37 38  2.13.1.1 Allelic Frequencies in Different Racial Populations 40 2.13.1.2 Association of V D R Genotype with Fracture Risk 42 2.13.1.4 V D R Genotype and Bone Mineral during Childhood 43 2.13.1.5 V D R Start Codon Polymorphism 45 2.13.1.6 Interacdon Effects of V D R Genotypes and Environment 47 2.14 Collagen Type I, a - l (COL1A1) 49 2.14.1 Human Studies of COL1A1 Polymorphisms 50 2.14.1.1 Association of COL1A1 Polymorphism with aBMD and Fracture Risk 50 2.14.1.2 Association of COL1A1 Genotype and Bone Mineral Accrual in Children 55 2.14.1.3 Interaction between V D R and C O L I A l Genotype in aBMD and Fracture risk 55 2.15 Tumour Necrosis Factor Receptor 2 (TNFR2) 57 2.15.1 Human Studies 58 2.15.1.1 T N F R 3' U T R Polymorphisms associated with variance in aBMD 58 2.16 Summary of the Literature Review 59 CHAPTER 3 61 Study Design 61 3.1 Research Aims and Hyotheses 61 3.2 Subjects 64 3.3 Instruments and Procedures 65 3.3.1 Anthropometry 65 3.3.2 Bone Mineral Assessment 65 3.3.3 Physical Activity and Dietary Calcium Intake 66 3.3.4 Maturity Assessment 67 3.3.5 Genotype Analysis 67 3.3.5.1 Extraction and Purification of Genomic D N A 67 3.3.5.2 PCR, R E Analysis and Electrophoretic Seperation of V D R Polymorphisms 68 3.3.5.3 COL1A1 Spl Polymorphism Detection using the TaqMan™ System 71 3.3.5.4 Automated Sequencing of the TNFR2for the 3 Polymorphisms within this Region.... 74 3.3.6 Statistical Analysis 77 CHAPTER 4 79 Results 79 4.1 Descriptive Results 79 4.1.1 Anthropometric Descriptive Results 79 4.1.2 Age and Maturity Descriptive Results 79 4.1.3 Polymorphism Frequencies 80 4.1.4 Bone Parameter Descriptive Results 80 4.1.5 Bivariate Correlations of Variables 81 4.2 G L M of B M F L Mass and Bone to Determine Effects of PA, Genes and Interactions 83 4.2.1 Covariate Modeling of Bone Mineral-free Lean Mass 83 4.2.2 Effects of Physical Activity on B M F L Mass after accounting for Covariates 84 4.2.3 Effects of Candidate Polymorphisms on B M F L Mass after Accounting for Covariates. 85 4.2.4 Covariate Modeling of Bone Parameters 87 4.2.5 Effects of P A and Dietary Calcium on Bone after accounting for Covariates 89 4.2.6 Effects of Candidate Genes on Bone Parameters after Accounting for Covariates 92 4.3 Effects of P A and Genotype Moderated by B M F L on Bone Parameters 98 iv  4.3.1 Effects of Bone Mineral-free Lean Mass by Physical Activity Interactions on Bone 98 4.3.2 Effects of Candidate Gene by Bone Mineral-Free Lean Mass Interactions on Bone 100 4.4 Effects of Candidate Gene and Physical Activity Interactions on Bone Parameters 108 CHAPTER 5 110 Discussion 110 5.0.1 Genotype and Haplotype Frequencies by Ethnicity Ill 5.1 The Role of Body Size and Composition, P A and Genotype on B M F L Mass 112 5.1.1 Independent Effects of Physical Activity Levels on Bone Mineral-Free Lean Mass 113 5.1.2 Independent Effects of Genotype on Bone Mineral-Free Lean Mass 114 5.2 The Role of Body Size, Composition and Maturity on Bone Mass in Children 114 5.3 Independent Effects of Average P A and Dietary Calcium Intake on Bone Mass 116 5.3.1 Average Physical Activity Score and Bone Mass 116 5.3.1.1 Proximal Femur Bone Mass 116 5.3.1.2 Lumbar Spine Bone Mass 117 5.3.2 Average Daily Calcium Intake and Bone Mass 118 5.4 Independent Effects of V D R Fokl, BsmI, COL1A1 and T N F R 2 Genotype on Bone.... 119 5.4.1 V D R BsmI and Fokl Genotype and Bone Mass 119 5.4.2 COL1A1 Spl Genotype and Bone Mass 120 5.4.3 TNFR2 A593G, T598G and T620C Genotype and Bone Mass 122 5.5 Effects of PA on Bone are Different by Levels of Bone Mineral-Free Lean Mass 124 5.6 Effects of Genotype on Bone are Different by Levels of Bone Mineral-Free Lean Mass 126 5.6.1 Effect of V D R Fokl Genotype Moderated by B M F L Mass on Bone Mass 126 5.6.3 Effect of T N F R 2 Haplotype Moderated by B M F L Mass on Bone Mass 127 5.7 Effects of Physical Activity Moderated by Levels of Genotypes 129 5.7.1 Effects of PA Moderated by COL1A1 Genotype on Bone Mass 130 5.8 Unique Apects of the Study 131 5.9 Limitations of the Study 131 5.9.1 Degrees of Separation between Independent and Dependent Variables 131 5.9.2 Testing for Effects by Ethnicity 132 5.9.3 Adjustment for Multiple Comparisons 133 5.10 Future Directions 134 5.11 Summary and conclusions : 134 5.11.1 Independent Effect of P A or Dietary Calcium on B M F L Mass and Bone Mass 134 5.11.2 Independent Effect of Candidate Genotypes on B M F L Mass and Bone Mass 135 5.11.3 The Effect of P A and Genotype on Bone are Moderated by Levels of B M F L 136 5.11.4 The Effect of P A on Bone is Moderated by Levels of Candidate Genotype 137 REFERENCES 138 APPENDIX 1 151  v  LIST O F T A B L E S  lable Name  Vase  Table 2.1: Summary table of prospective calcium supplementation bone parameters Table 2.2: Summary table of prospective studies of childhood physical activity effect on bone parameters.  °  21 22  Table 2.3: Linkage data from Koller et al, 2000 Table 2.4 : Mouse and corresponding human chromosomal regions implied in B M C accrual.  35  Table 2.5: Summary table of V D R Bsml genotype effects by ethnicities  42  Table 2.6: Summary table of V D R Bsml genotype effects in paediatric populations  45 47  Table 2.7: Summary table of V D R Fold genotype effects.  Table 2.8: Summary table of COL1A1 Spl genotype effects  Table 3.1: P C R reaction reagents, concentrations and volumes in V D R Fold and V D R Bsml reactions.  33  55 68  Table 3.2: Oligoucleotide forward and reverse primer sequence for the P C R amplification of the V D R Fold and V D R Bsml regions of interest  68  Table 3.3: Thermal Cycler conditions for V D R Fold and V D R Bsml amplification Table 3.4: Reagents for V D R Fold and V D R Bsml restriction enzyme digestion  69 69  Table 3.5: T a q M a n ™ primers and probes for detection of the COL1A1 Spl polymorphism.  73  Table 3.6: List of reagents per 25 ul reaction using the T a q M a n ™ system to detect COL1A1 Spl polymorphisms.  73  Table 3.7: T a q M a n ™ P C R reaction conditions for COL1A1 Spl polymorphism detection.  73  Table 3.8: Primer sequences for the P C R amplification of the 3 ' U T R of T N F R 2  74  Table 3.9: Reagents, concentrations and volumes used in the P C R amplification of the polymorphic region of the 3 ' U T R of T N F R 2 .  74  Table 3.10: P C R thermal cycler conditions for the amplification of the polymorphic region in the 3 ' U T R of T N F R 2 .  75  Table 3.11: Thermal cycler conditions for the sequencing reaction of the polymorphic region of the 3 ' U T R of T N F R 2 .  76  Table 4.1: Descriptive results (mean ± SD) for height (cm), sitting height (cm), leg length (cm), weight (kg), total fat mass (g), B M F L (g) in the total group and for boys and girls.  79  Table 4.2: Descriptive results for age (mean ± SD) and Tanner Breast and Pubic Hair Stages (frequency and percentage).  79  Table 4.3: Genotype frequencies for the total group and for Caucasian and Asian subgroups  80  Table 4.4. Descriptive results (mean ± SD) of the measured bone parameters Table 4.5: Correlation matrix for Anthropometric variables in boys  81  Table 4.6: Correlation matrix for Anthropometric variables in girls vi  81 81  Table 4.7: Bivariate correlations, Pearson-R (2-tailed significance) between anthropomorphic variables and bone parameters Table 4.8: Bivariate correlations, Pearson-R (2-tailed significance) between anthropomorphic variables and bone parameters Table 4.9: Variance accounted (R ) for B M F L by the covariates Table 4.10: Results of the General Linear Model examining the contribution of P A to BMFL Table 4.11: The main effects of genotypes on B M F L in boys and girls after accounting for the covariates Table 4.12: The main effects of genotypes on B M F L in the Caucasian boys and girls after accounting for the covariates Table 4.13: The main effects of genotypes on B M F L in the Asian boys and girls after accounting for the covariates Table 4.14: Variance accounted (R ) for proximal femur B M C by the covariates Table 4.15: Variance accounted (R ) for proximal femur B M C by the covariates Table 4.16: Variance accounted (R ) for femoral neck B M C by the covariates Table 4.17: Variance accounted (R ) for femoral neck B M C by the covariates Table 4.18: Variance accounted (R ) for femoral neck a B M D by the covariates Table 4.19: Variance accounted (R ) for femoral neck a B M D by the covariates Table 4.20: Variance accounted (R ) for lumbar spine B M C by the covariates Table 4.21: Variance accounted (R ) for lumbar spine B M C by the covariates Table 4.22: The effect of average dietary calcium intake and P A on proximal femur BMC Table 4.23: The effect of average dietary calcium intake and P A on femoral neck BMC Table 4.24: The effect of average dietary calcium intake and P A on femoral neck aBMD Table 4.25: The effect of average dietary calcium intake and average P A score on lumbar spine B M C Table 4.26: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in boys Table 4.27: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian boys 2  2  2  2  2  2  2  2  2  2  2  Table 4.28: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian boys Table 4.29: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in girls Table 4.30: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian girls Table 4.31: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian girls Table 4.32: Significant differences of T N F R 2 A593G-T598G haplotypes in girls Table 4.33: The effects P A moderated by B M F L in boys and girls for proximal femur BMC  vu  85 86 86 87 88 88 88 88 89 89 89 90 90 91 92 93 94 94 96 97 97 98 99  Table 4.34: T h e effects of P A moderated by B M F L in boys and girls for femoral neck BMC Table 4.35: The effects of P A moderated by B M F L in boys and girls for femoral neck aBMD Table 4.36: The effects of P A moderated by B M F L in boys and girls for Lumbar Spine BMC Table 4.37: The effect of candidate genes polymorphisms ( V D R BsmI, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in boys Table 4.38: The effect of candidate genes polymorphisms ( V D R BsmI, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian boys Table 4.39: The effect of candidate genes polymorphisms ( V D R BsmI, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian boys Table 4.40: The effect of candidate genes polymorphisms ( T N F R 2 A593G, T N F R 2 T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D , and lumbar spine B M C in boys Table 4.41: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian boys Table 4.42: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian boys Table 4.43: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in boys Table 4.44: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian boys  99 99 99 100  101  101  102  102  102  103 103  Table 4.45: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian boys Table 4.46: The effect of candidate genes polymorphisms ( V D R BsmI, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in girls Table 4.47: The effect of candidate genes polymorphisms ( V D R BsmI, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian girls Table 4.48: The effect of candidate genes polymorphisms ( V D R BsmI, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian girls Table 4.49: The effect of candidate genes polymorphisms ( T N F R 2 A593G, T N F R 2 T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D , and lumbar spine B M C in girls  103  Table 4.50: The effect of candidate genes polymorphisms ( T N F R 2 A593G, T N F R 2  105  viii  104  104  104  105  T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D , and lumbar spine B M C in Caucasian girls Table 4.51: The effect of candidate genes polymorphisms ( T N F R 2 A593G, T N F R 2 T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D , and lumbar spine B M C in Asian girls Table 4.52: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in girls Table 4.53: T h e effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Caucasian girls Table 4.54: T h e effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C in Asian  106  106 107  107  gids  Table 4.55: Interaction of COL1A1 Spl genotype and P A in girls Table 4.56: Interaction of COL1A1 Spl genotype and P A in boys  ix  108 109  LIST O F F I G U R E S  Figure Name  Page  Figure 2.1: Assembly, secretion and elongation of hydroxyapatite crystals in osteoblasts. Figure 2.2: Mechanical stimulus transduction via osteocyte canaliculi to bone lining cells followed by recruitment of osteoblasts due to production of PGs and IGFs. Figure 2.3: Osteoclast (Oc) resorbing bone matrix (MM) via the brush border (BB). Figure 2.4: Diagram of an osteoclast highlighting the molecular aspects of bone resorbing function. Figure 2.5: A schematic of a growing long bone (left). Diagram highlighting important microstructral features of bone (Right). Figure 2.6: Diagramitic representation of Bone-Remodeling transient Figure 2.7: Areal B M D overestimates actual volumetric B M D in this idealized example.  4 5 6 7 9 12 13  Figure 2.8: Total body peak bone mineral content velocity ( T B P B M C V ) curve demonstrating differences in velocity at peak, age at peak B M C and P H V for boys and girls.  14  Figure 2.9: Geometric adaptation of bone in long bones to increase strength in bending or torsion.  16  Figure 2.10: Diagram showing (A) H S A regions (B) Standard D X A proximal femur regions (C) Schematic of the cross-sectional geometric variables calculated by the H S A program.  17  Figure 2.11: Schematic of V D R m R N A and protein product highlighting the V D R SCP and 3'UTR polymorphisms  38  Figure 2.12: COL1A1 Spl polymorphism results in increase in C o l l a l expression relative to Colla2 in osteoblasts  51  Figure 2.13: TNFR2 polymorphisms results in the 10 exon of the 43 Kb TNFR2 gene.  58  Figure 3.1: Sample electrophoresis of Bsml digest of V D R Bsml polymorphisms. Figure 3.2: Sample electrophoresis of Fold, digest of V D R Fold polymorphisms.  70 71 72 75  th  Figure 3.3: Basic chemistry of the T a q M a n ™ System. Figure 3.4: P C R amplification followed by E x o / S A P - I T degradation of excess nucleotides and unused primers prior to automated sequencing. Figure 4.1: Significant effect of P A on B M F L in boys Figure 4.2: Significant difference in B M F L for girls by TNFR2 A593G genotype Figure 4.3: Significant difference in B M F L for Asian girls by TNFR2 A593G genotype Figure 4.4: The Effect of P A on proximal femur B M C in boys after accounting for the covariates  84 85 86 90  Figure 4.5: The effect of P A on femoral neck B M C in boys after accounting for covariates  91  Figure 4.6: The effect of average P A score on femoral neck a B M D in boys and girls  92  after accounting for the covariates Figure 4.7: Effect of COL1A1 Spl genotype on femoral neck B M C in boys Figure 4.8: Effect of T N F R 2 A593G genotype on femoral neck B M C in boys Figure 4.9: Effect of V D R Fold genotype on proximal femur B M C in Caucasian boys Figure 4.10: Effect of T N F R 2 A593G genotype on femoral neck B M C in Caucasian boys Figure 4.11: Effect of T N F R 2 T598G genotype on femoral neck B M C in girls Figure 4.12: Effect of T N F R 2 T620C genotype on lumbar spine B M C in girls Figure 4.13: Significant differences in femoral neck B M C in girls by T N F R 2 A593GT598G haplotype Figure 4.14: The main effect of V D R Fold, genotype for femoral neck a B M D in boys after accounting for the interaction with B M F L Figure 4.15: M a i n effects of T N F R 2 A593G-T598G haplotype on femoral neck B M C after the haplotype by B M F L interaction is accounted for in girls Figure 4.16: Main effects of T N F R 2 A593G-T598G haplotype on femoral neck B M C after the haplotype by B M F L interaction is accounted for in Caucasian girls Figure 4.17: Main effects of COL1A1 genotype on femoral neck a B M D after the genotype by B M F L interaction is accounted for in girls Figure 4.18: Main effects of COL1A1 genotype on femoral neck B M C after the genotype by B M F L interaction is accounted for in Caucasian boys Figure 5.1: Schematic of effectors of bone mass  xi  93 94 95 95 96 97 98 100 106 107 108 109 111  ABBREVIATIONS USED IN THIS THESIS  3' UTR: 3' Untranslated region aBMD: Areal Bone Mineral Density BMC: Bone Mineral Content BMDvol: Volumetric bone mineral density BMFL: Bone mineral-free lean mass BMI: Body Mass Index BMP: Bone Morphogenic Protien BMU: Bone Multicellular Unit bp: base-pair C O L 1 A 1 : Collagen Type I, a - l CSA: Cross-sectional Area CSMI: Cross sectional moment of inertia CT: Calcitonin DXA: Dual-energy X-ray Absorptiometer DZ: Dizygotic E C M : Extra Cellular Matrix ESTs: Expressed Sequence Tags FFQ: Food Frequency Questionnaire FRA1: Fos-related Antigen-1 HBGC: Healthy Bones Genetic Cohort HSA: Hip Structral Analysis IGF: Insulin-like Growth Factor IL: Interleukin LOD: Logarigthm of the Odds M-CSF: Multiple-Colony Stimulating Factor MIM: Mendelian Inheritance in Man MRI: Magnetic Resonance Imaging MZ: Monozygotic NS: non-significant ODF: Oscteoclast differentiation factor PA: Physical activity PAQ-C: Physical activity questionaire for children PBM: Peak Bone Mass pBMC: peak bone mineral content PCR: Polymerase Chain Reaction PDGFp: Platelet-derived growth-factor p P G : Prostaglandin PHR: Parathyroid Hormone Receptor PICP: Carboxyterminal propeptide of type I procollagen PTH: Parathyroid Hormone QCT: Quantitative Computed Tomography QTL: Quantitative Trait Loci RANK: Receptor Activator of NF-k-B xii  r E R : rough E n d o p l a s m i c Reticulum R T - P C R : R e v e r s e transcriptase polymerase chain reaction S C P : Start c o d o n polymorphism S N P : Single Nucleotide P o l y m o r p h i s m T B : Total B o d y T B E : Tris-Borate-EDTA T G F p : Transforming Growth Factor (5 TNF-oc: T u m o r N e c r o s i s F a c t o r - a T N F - p : T u m o r N e c r o s i s Factor-|3 TNFR2: T u m o r N e c r o s i s Factor R e c e p t o r 2 T N F R S F 1 B : T u m o r N e c r o s i s Factor Receptor S u p e r Family M e m b e r  1B U V B : Ultraviolet-B V D R : Vitamin D R e c e p t o r V 0 2 M a x : M a x i m a l o x y g e n volume c o n s u m e d during high intensity exercise  xiii  ACKNOWLEDGMENTS  The author wishes to thank Heather M c K a y for her academic and career advice through out the length of this degree.  T h e coliegial advice of the U B C Bone G r o u p lead by D r s M c K a y and Khan has  expanded my understanding o f bone biology and the world of academia tremendously over the course of the last 2 years. In particular the author wishes thank D r . Kerry MacKelvie for her collection o f the bone data and organization of Healthy Bones II.  The author wishes to thank the thesis committee  members, D r . Jack Taunton, D r . Sylvie Langlois and D r . Ross MacGillivray for their support and instruction in completion o f this document. T h e author also wishes to thank the bone research lab at the University o f Aberdeen lead by Professor Stuart Ralston.  Their help was instrumental for  collection o f the genetic data. T h e author wishes to thank the R T A MacGillivray lab for their help in the early days of data collection and lasting friendships to help smooth over the rough days.  Finally,  the author wishes to thank Mary Lee, Rick and Kate Taylor and Nicole Champagne for sticking with me through the ups and downs o f this degree.  xiv  CHAPTER 1  Introduction Bone is a dynamic tissue. A t any given period of life a complex pattern o f genetic, physiological and environmental factors interact to contribute to the equilibrium o f bone metabolism. The intricate relationship between intrinsic and extrinsic influences on bone metabolism present challenging questions to bone researchers and that demands novel, multifactorial approaches.  It is well established that bone parameters are under considerable genetic influence [1]. However, the roles o f specific genes have yet to be completely described. Furthermore, the complexities of environmental and other physiological variables interacting with the genetics framework regulating bone metabolism continue to remain largely undefined.  Bone disease has provided many insights into the delicate balance between numerous variables related to bone metabolism. Bone mineral, the bone constituent that provides much o f bones' mechanical strength, is perhaps the most widely studied parameter of bone metabolism as it is direcdy related to osteoporosis.  Furthermore, osteoporosis is the most prevalent bone disease and hence is the most  pressing topic in current bone research.  /. / Definition of Osteoporosis Osteoporosis is a disease characterized by low bone mass and structural deterioration of bone tissue from continued disruption o f skeletal homeostasis, leading to bone fragility and an increased susceptibility to fractures o f the proximal femur, spine, and wrist from low impact trauma [2]. The clinical significance o f osteoporosis arises from the increased risk o f fracture.  1  7.2 Incidence and Health Care Cost Associated with Osteoporosis in Canada The increasing prevalence of osteoporosis is a mounting health care concern in Canada. One in four women over fifty and one in eight men over fifty are affected [2]. Currently there are an estimated 1.4 million Canadians that suffer from osteoporosis [2]. Osteoporosis represents a major cause of morbidity in elderly populations. There are 17,500 new cases of proximal femur fracture due to osteoporosis each year in Canada. Proximal femur fracture results in death for 20% of elderly patients, and permanent disability in 50% of those who survive [2]. Likewise, mortality rates in women due to osteoporosis are higher than the combined mortality rates from malignancies of the breast and ovaries [2]. Treatment of osteoporotic fracture costs the Canadian health care system an estimated $1.3 billion a year [2]. Once established, osteoporosis is difficult to reverse, but early interventions by means of therapeutic [3], dietary [4] and lifestyle changes [5] result in decreased risk of osteoporotic fracture. Clearly, the key to relieving the financial burden of reactive health care as well as decreasing morbidity due to osteoporotic fracture is preventing the onset of osteoporosis. 1.4 The Relationship betiveen Peak Bone Mass and Osteoporosis Adult bone mass can be conceived as the peak bone mass (PBM) minus subsequent bone loss [6]. Furthermore, adult bone mass is the best predictor of osteoporotic fracture [7]. The majority of research has probed the dynamics of bone loss. Therefore, fracture prevention has focused mainly on reducing or preventing bone loss by modifying lifestyle factors (i.e. diet and exercise) and pharmacotherapeutics. The converse approach, enhanced P B M , is mired in controversy as there is little data to support claims that enhanced P B M results in decreased fracture risk later in life. Further, age of P B M is contentious and there are few longitudinal studies of bone mass accrual prior to P B M [6].  2  CHAPTER 2 Literature B a c k g r o u n d for the P r o p o s e d Study  2.1 Bone Cytology There are three distinct cell types in bone: osteoblasts, osteocytes and osteoclasts. However, there are only two unique cell lines as osteocytes are derived from mature osteoblasts. Thus for a parsimonious discussion of the cell types of bone, I will focus on: (1) the osteoblast/osteocyte lineage and (2) the osteoclast lineage.  2.1.1 Osteoblasts and Osteocytes New bone is formed by osteoblasts at the bone surface synthesizing and secreting organic extracellular matrix (ECM) which subsequently becomes mineralized [8]. The ultrastructural features of osteoblasts are indicative of the high metabolic activity of these cells. Osteoblasts have a rough endoplasmic reticulum (rER) that stains extremely basophilic and a well-developed Golgi apparatus, indicative of a high degree of protein secretion [8]. Osteoblasts are mononuclear with their nucleus eccentrically opposite the bone surface. There are up to three nucleoli within the nucleus again underscoring the high rate of transcription and translation in these cells [8].  The primary role of osteoblasts is to secrete both the protein and inorganic components of bone E C M . One fifth of the total protein expressed by the osteoblast is type I collagen which is secreted into the matrix to provide tensile strength and a meshwork on which hydroxyapatite crystal may form. Osteoblasts also secrete a variety of other non-collagenous proteins that aid in: (A) linking osteocytes to the E C M (i.e. integrins and osteonectin) [9,10], (B) osteoclast signalling (e.g. osteocalcin) [11], and (C) a host of growth factors and cytokines that aid in regulating cell activation, growth and turnover [e.g. transforming growth factor P (TGFp), bone morphogenic proteins (BMPs), insulin-like growth  factors (IGFs)][12]. Furthermore, mature osteoblasts express numerous endocrine receptors such as the nuclear oestrogen receptor and 1,25-dihydroxy-vitamin D  3  receptor ( V D R ) and the plasma  membrane-bound parathyroid hormone receptor (PHR) to regulate their activity [13,14].  The compressive strength o f bones comes from the inorganic component o f bone, the layers o f hard, plate-like crystals o f hydroxyapatite [Ca (PO ) (OH) ]. These crystals are typically 20-80 nm long and 10  4  6  2  approximately 2-5 nm thick [8]. Osteoblasts secrete the matrix vesicles that have sufficiently high concentrations o f hydroxyapatite to seed crystallization. These crystals are then deposited within the collagenous meshwork to nucleate the larger crystal latticework o f hydroxyapatite in calcified bone.  Figure 2.1: Assembly, secretion and elongation of hydroxyapatite crystals in osteoblasts. Calcium and inorganic phosphate form hydroxyapatite in the matrix vesicle Following exocytosis of the matrix vesicle the hydroxyapatite crystal begins to grow and ruptures the matrix vesicle.  http://www.vet.ohio-state.edu / docs / ATCenter/Histo/Bone Osteoblasts embedded in the mineralized E C M become osteocytes that form a syncytical network connected through characteristic canaliculi permitting the passage o f extracellular fluid [15]. Newly formed osteocytes retain many cytological features o f osteoblasts including abundant and well organized granular r E R and a large Golgi region, characteristic o f active protein synthesizing cells.  4  However, osteocytes located further from the active bone-forrning surface have sparse granular r E R and less Golgi apparatus [8]. Despite the appearance of older osteocytes as inactive cells there is increasing evidence that these cells have a role in transduction of the mechanical loading signal, thereby acting as the mechano-sensors o f bone [16,17]. It has been suggested that these signals are transmitted by canalicular flow induced production o f prostaglandins (PGs) and I G F s [18]. Such communication between the osteocytes/bone lining cells provides a means o f inducing osteoclastic and osteoblastic activity in areas o f bone undergoing mechanical strain.  N e w bone  Osteoprogeniior  Mechanical signal  Praosteobtasi  Osteocyte  ^ Bone  ' lining cell  5  2.1.2 Osteoclasts  http://www.cal.nbc.upenn.edu/saortho/ chapter_01 /OlF4.jpg  The major bone resorbing cells are osteoclasts. They are large multinucleated cells, with usually 4 to 20 nuclei, and can be up to 100 pm wide in pathological states [19]. The ruffled brush border adjacent to the area of matrix resorption surrounded intra-cellularly by an anchoring actin ring are morphological indicators of osteoclasts [19]. However it is often difficult to separate osteoclasts and macrophage polykaryons based on morphology alone. Such ambiguity is due in part from the common lineage of macrophages and osteoclasts from haematopoetic mononuclear cells in bone marrow [19].  Osteoclasts resorb bone by producing proteolytic enzymes to digest the organic fraction of the bone matrix and by decreasing the local pH to dissolve the inorganic fraction (hydroxyapatite) of the bone matrix. Osteoclasts use the enzyme Carbonic Anhydrase II to produce protons which are then pumped across the ruffled border through a proton pump similar to the vacuolar ATPase in the intercalated cells of the kidney[20]. The low p H microenvironment in the area beneath the brushed border, also known as Howship's lacunae, is optimal for the digestion of bone matrix proteins by osteoclast secreted proteolytic enzymes [19]. Furthermore, to maintain electrochemical balance within 6  the osteoclast a number o f ion transporters on the basolateral side (opposite the ruffled border) are expressed (e.g. N a / H +  +  antiporter, C a  ATPase, a K  2 +  +  channel, N a / K +  +  ATPase, H C 0 ' / C r 3  exchanger) [21].  B O N E RESORPTION INDUCERS 1 y —  \ I  1 I  f  t  I t  /". ^f' P e m l n r r a H r « d CofUgrn v  —  ColLfjtn.ie  Soluble Fragment*  -  /  /  ^iyi^:^:':-;  <?>Col1ts«n f t l n g r L r d o v r r  http://healthcare.partners.org/orthopaedic/hoj2000/html/articlesll.htm  Figure 2.4: Diagram of an osteoclast highlighting the molecular aspects of bone resorbing function. Protons are produced cytosolically then p u m p e d through the ruffled border. Mature lysosomes exocytose  to release a n u m b e r o f cystein-proteinases  (cys-proteinases)  w i c h are  activitated by the l o w p H m i c r o - e n v i r o n m e n t i n the H o w s h i p ' s Lacuna. I n turn the active cysproteinase cleave and activate collagenase that was secreted i n inactive procollegenase f o r m by osteoblast during mineralization. T h e cys-proteinase as well as the collegenase digest the protein component o f the bone matrix.  The lysosomes destined for exocytosis from the ruffled border o f osteoclast are full o f inactive cysteine-protease, cathepsin K [22]. These enzymes digest type I collagen fibrils by forming ring-like complexes around the collagen fibril and then cleaving it, leaving the collagen fibrils susceptible to  7  digestion from other protease [23]. Further collagen solubulization of the bone matrix is accomplished by collagenase that is secreted by the osteoblast in procollagenase form during osteoid secretion. Cathepsin K cleaves the propeptide sequence from procollagenase which is rendered fully active by the low p H microenvironment [24]. There is increasing evidence that the collagen fragment left by collagenase digestion activates osteoclasts [25]. As such, osteoblasts next to resorbing osteoclasts have been shown to have higher expression of collagenase [24]. These findings suggest a molecular mechanism of osteoblast-osteoclast coupling by co-localizing osteoclastic activity with osteoblastic activity.  2.1.3 Osteoblast-Osteoclast Coupling Osteoclastic bone resorption and osteoblastic bone deposition are closely linked during bone remodeling (section 2.3, p. 10). Therefore, a section of bone that is being resorbed and subsequently rebuilt can be conceptualized as a "Basic Bone Multicellular Unit (BMU)" [26]. Under balanced conditions, bone is resorbed by osteoclasts only to be replaced by equivalent new bone by osteoblasts.  2.2 Bone Histology The mature human skeleton provides mechanical support, protection of vital organs, haemopoetic activity and attachment sites for muscles. There are twotypesof bone: immature, disordered woven bone and mature, well-organized lamellar bone. Woven bone comprises all bone at birth and in later years is found at sites of fracture repair. Lamellar bone replaces woven bone in the maturing skeleton by age four. Lamellar bone arranges the collagen fibrils along lines of force in cortical bone [27]. Such organization gives the maturing skeleton anisotropic properties [26]. Bone tissue can be separated into two compartments: cortical (compact) and trabecular (cancellous) bone. Although cortical and trabecular bone contain the same cellular and matrix constituents they are 8  functionally and structurally different [28]. Cortical bone is found primarily in the diaphyses of long bones comprising approximately 80% of total bone weight. As such, cortical bone represents 80 90% of the body's calcified tissue by weight. Functionally, cortical bone provides mechanical strength and protects bone. Conversely, trabecular bone provides mosdy metabolic activity (e.g. haematopoesis) and is therefore only 15 — 25% calcified. In adults trabecular bone is found predominantly in the short bones, such as the vertebral bodies or in the flat bones, such as the pelvis and in the metaphysis and epiphysis of long bones (Figure 2.5) [26],  Osteons  Growth plate Trabecular  -  TJ Epiphysis  Haversian canal  Metaphysis Resorption spaces  bone Cortical bone  Trabecular  Endosteum  (cancellous)  Diaphysis  ijone Periosteum  Lacunae  Fused growth plate Conical (compact) bone  Figure 2.5: A schematic of a growing long bone (left). D i a g r a m highlighting important  microstructral features of bone (Right). Reprinted from Khan etal, 2001.  [26]  Cortical bone is organized into haversian systems. They consist of a neural-vascular canal concentrically surrounded by lamellae of bone, also called osteons [29]. Typically haversian canals are arranged in parallel along the longitudinal axis of bone and are connected by perpendicular vessels, Volkmann's canals [29]. 9  Both cortical and trabecular bone is under constant reorganization in response to changing physiological and environmental stimuli [30]. Although the proportion of trabecular calcified bone volume is much less compared with cortical bone the surface-to-volume ratio is much greater for trabecular bone. As bone remodelling is dependent on available bone surface, remodelling rate is approximately ten-fold higher for trabecular bone compared with cortical bone[30].  2.3 The Dynamics of Hone Groivth and Maintenance There are three discrete processes that transform the strength, shape and architecture of healthy bone in children and adolescents: growth, modeling and remodeling [31]. Bone growth is the enlargement of the skeleton that terminates with the cessation of linear growth during late adolescence. Bone growth is genetically determined and mediated by autocrine and endocrine processes. Bone modeling refers to the addition of bone to both periosteal and endocorticol surfaces undergoing high mechanical strain (e.g. due to loading). The process occurs without prior resorption of bone and results in net gain of bone thereby increasing the strength of the skeleton. Bone added regionally is termed macromodelling. Macromodelling improves the geometric properties of bone by increasing bone mass on surfaces that improve bending or compressive strength. Likewise, minimodelling changes the orientation of trabelucae regardless of changes in mineralization or bone size. The alignment of archlike trabeculae in the femoral neck corresponds with the lines of force along this region of the femur (anisotropically). Such strengthening of bone occurs primarily during bone growth. Remodeling is the progression by which damaged bone is replaced with new bone. This process occurs primarily during adulthood but is also observed during growth. The cycles of remodeling determine the geometric properties of a bone area and bone mass. While resorption and deposition of bone is normally in equilibrium, disruptions of skeletal homeostasis (e.g. due to physical inactivity, 10  disease, age or drug administration) results in incomplete filling o f the resorption cavity, causing bone loss over the long term [31].  2.4 The Bone-Remodeling Transient The bone-remodeling transient is a theoretical construct that proposes that short-term environmental changes in the form of interventions may show temporary effects to bone mineral metabolism not seen over longer durations [32]. Such a phenomenon arises as the effect o f a shift in the normally equilibrated bone resorption/formation cycle. A t any period o f time a fraction o f B M U s are in resorption phase while others are in the process of depositing or mineralizing bone. A s such, the areas of remodeling contain a lower degree o f mineralization than those in the formation and mineralization phase. Quantification o f bone by non-invasive measures (e.g. dual energy X-ray absorbtiometry; section 2.5, p. 14) evaluates the bone mineral content not the amount o f bone tissue. Areas in active resorption represent undermineralized areas thereby underestimating the amount o f bone tissue. It follows that interventions that alter resorption can lead to a temporary increase in apparent mineralization. However, sustained changes in the amount o f bone tissue may not be present [32].  The bone remodeling transient is short lived as, eventually, the bone remodeling cycle reaches a new equilibrium and the proportion o f resorbing units and formation/mineralization units reach a new plateau. T h e converse effect when an intervention is ceased can also be observed [32]. Consequently to establish whether an intervention had a persistent impact on bone tissue it is necessary to monitor changes that occur during the first cycle of remodeling independently o f ensuing cycles thereby accounting for the effects o f the bone remodeling-transient. Although there is no steadfast rule for such a period it has been recommended that 6 months is an adequate time delay to account for the bone-remodeling transient in children and adolescents [31].  11  Figure  2.6:  representation Remodeling  Diagramitic of  BoneLow  transient.  activity remodeling bone  (top) vs.  high activity remodeling bone (mid). T h e apparent decrease i n a B M D is due to increased resorption spaces i n actively remodeling bone Old bone  upon  New bone  remodeling their w i l l  Osteoid Resorption space  aBMD  cessation  of  however increased  be a higher  i n the neawly  remodeled  bone. Reprinted f r o m K h a n  et al,  2001 [26]  2.5 Bone Densitometry There are several methods of assessing bone mass, however dual-energy X-ray absorptiometry (DXA) has emerged as the method of choice for clinical use as it is highly accurate, has low radiation dose per measurement and is relatively rapid [33]. Dual-energy X-ray absorptiometry is able to correct for photon attenuation when evaluating bone that is beneath soft-tissue and has a high photon flux, allowing decreased examination time, higher resolution and longer source life [33]. Scan times have been reduced to a few minutes in newer generation D X A machines by using a fan-shaped beam of xrays. Dual-energy X-ray absorptiometry measurements are 2 dimensional, measuring both bone mineral content (BMC) and cross-sectional area of the bone. Areal bone mineral density (aBMD) measures are calculated as B M C per area chosen and are reported as grams per square centimeter squared (g/cm ). Thus, aBMD is not a true measure of three dimensional density in the geometric 2  sense. The lack of three dimensional density data results in an underestimate of measured bone mineral density (g/cm ) in smaller individuals in comparison to larger counterparts when in fact actual 2  bone mineral density (g/cm ) may be equal (Figure 2.7, p. 13). Additionally, there is a positive 3  correlation between height and weight and aBMD. Such correlations are due to bone size. Ergo, corrections for bone size are necessary when evaluating inter-subject differences in aBMD [34]. A low 12  BMC, unadjusted for bone size in an ill person may reflect small bone area, that is, a small skeleton, not a reduced bone mass. It is important to know whether the D X A measurement is reflecting a small bone area or low B M C corrected for bone area [35]. Corrections for such differences in bone size are particularly important when evaluating heterogeneous groups involving men, women and children of different maturities. [26]  Block B  Block A  1 cm Total material'density (g/cm )  1 g/crri  Total material (bone mineral content) (g)  1 gram  Bone area (cm )  1 cm  3  2  Areal bone mineral density (g/cm ) 2  1 g/cm  2 cm  3  1 g/cm  3  8 grams  4 cm  2  2  2 g/em  2  2  Figure 2.7: Areal B M D overestimates actual volumetric B M D in this idealized example. Reprinted from Khan et al, 2001. [26]  2.6 Three Dimensional Bone Measurement  True measures of bone volume, geometry and trabecular vs. cortical ratio require the use of 3 dimensional Quantitative Computed Tomography (QCT). However, Q C T measurement yields a high dose of ionizing radiation. Consequently, Q C T maybe hazardous for evaluation of volumetric B M D in children and healthy adults. However, Q C T scans of the peripheral sites (i.e. radius, tibia, etc.) has similar radiation levels as those of D X A scans as they focus on less vulnerable body regions. Cylindrical models approximating bone volume have been devised to transform aBMD data to volumetric B M D (BMD ). Therefore, vo]  13  lumbar vertebrae B M D , = aBMD [4/(71 * width of the vertebrae)] vo  femoral neck B M D , = aBMD (4/7t) * (height/ area of femoral neck) vo  [36]. Such geometric modeling has shown a strong, positive correlation with Magnetic Resonance Imaging (MRI) measures of bone volume (r = 0.849, p < 0.001) and therefore is a valid approximation of true bone volume [37].  2.8 Normal Patterns of Bone Growth in Children Bone densitometry in children is particularly difficult to interpret as changes in bone geometry, bone size and B M C occur rapidly especially during BMC TB Velocity Curve Cubic Spline  puberty. The increment rates of aBMD and B M C  450-i a Boys Age of Peak Peak Value  400  l  3 5 0  1  p2ak°v2u? 32'f  Age PHV 11.77 yrs  3 8  ]n  4  4  y e a r S )  boys aged 13-17 years) [36,  -  l - It has been previously shown that lumbar  4 0  Size Adjusted 342  spine B M C continues to increase in boys and not  1  | i 200  girls post puberty [35]. Likewise, B M C  »- 15fr| O  s m  ns a g e c  A Glrlt  300-1  £.250  (gj  size Adjusted 3«4  Age PHV 13.44 yrs  at  increase are highest in girls and boys at puberty  14.05 40S  accumulation particularly in the femoral neck is  100^ 50-  greater and lasts longer in boys during puberty [36,  0-  40]. This may be due to continued periosteal bone 9 10 11 12 13 14 15 16 17 18 19  growth in boys after puberty [41]. Nevertheless,  Age in Years Figure 2.8: Total body peak bone mineral content velocity (TB PBMCV) curve demonstrating differences in velocity at peak, age at peak B M C and P H V for boys and girls. Notice that the pubertal growth spurt (PHV) is approximately 0.6 years before PBMCV. Reprinted from Bailey etal 1999. [40]  14  -. >m p  S  1115  B  M  D  ,. , ,  m  ^creases shghtly post puberty  v o .  despite a slowing of bone expansion, BMC i  n  c  r  e  a  s  e  a  n  d  h  e  J  g  h  t  ^  ^  ^  BMD  v d  may indicate an endosteal expansion in bone in late puberty and early adulthood in boys but  not girls [36, 42]. Pre- and early-pubertal periods may be the optimal time frame for accumulation of maximal B M C for girls and boys as approximately one-quarter o f the adult peak B M C (pBMC) is gained during the 2 years around puberty [6]( Figure 2.8, p. 14).  As longitudinal studies spanning the entire pubertal period are time consuming and expensive, most of the research literature surrounding skeletal growth and bone mineral accrual has been from crosssectional and short-term longitudinal studies. A limitation o f studies in growing children is the difficulty in controlling for the considerable difference in maturation for chronologically age-matched children. F o r example, a cross-sectional study involving supposed pre-pubertal girls aged ~10 years may in fact have a population o f girls with a wide range o f pubertal stages, thereby blunting the effect of any bone accrual intervention due to differences in pubertal status on bone mineral accrual effects. Longitudinal designs have the benefit o f adjusting chronological age to biological age by normalizing pre- and pubertal subjects to maturational landmarks such as peak height velocity ( P H V ) (Figure 2.8, p.14) [40, 43].  Likewise, rates o f B M C accrual can be monitored in longitudinal designs across  puberty, a variable missed in cross-sectional studies.  2.9 Bone Strength Bone's ability to resist fracture under mechanical loading is not only dependant on bone mass but also material properties, geometry and tissue quality [44]. The elastic modulus and shear modulus, measures o f intrinsic stiffness in bending and torsion respectively, are correlated to the geometry of a section o f bone. Stiffness determines the degree of deformation that bone undergoes during a given load. Stiffness of a hollow cylinder such as bone is defined as the product o f the elastic modulus and  15  the areal moment of inertia of the cross-sectional area undergoing bending. The areal moment of inertia in a hollow cylindrical model can be further defined as:  I = 7t(r -r^/4 4  0  Where r„ is the outer radius and t, is the inner radius of the cylinder. A n analogous relationship exists for torsional loading except that the elastic modulus is replaced by the shear modulus and the areal moment of inertia is replaced by the polar moment of inertia. In response to mechanical loading bone responds by adjusting the bone geometry independently from the material properties (i.e. B M C or aBMD). Clinical evaluation of D X A scans does not assess the contribution of altered geometry to fracture risk.  CSMI  Figure  2.9:  Geometric  adaptation of bone i n long bones to increase strength in bending  or  torsion.  Increased modeling on the appropriate surfaces increase cross-sectional moment of inertia (CSMI) for bending or polar moment of inertia (J). Reprinted from Fonvood, 2001. [441  Polar momani ol  ineniB increased  When considering whole-bone strength it is important to consider geometric parameters in conjunction with B M C and aBMD. As such, Beck etal, [45] developed the hip structural analysis (HSA) program to evaluate the section modulus, cross-sectional area (CSA), cortical thickness, subperiosteal width and endosteal diameter in the proximal femur using D X A data. Without high doses of ionizing radiation from QCT, measures of strength at the hip can be approximated with 16  D X A . However, it should be underlined that HSA does not evaluate the actual cortical and trabecular widths at the proximal femur sites and therefore remains prone to the errors of the cylindrical model approximations.  Narrow Neck Femoral  Nock  Intertrochanteric Trocliantnr  Shaft  C. a - Subperiosteal width b - Endosteal width c - Cortical thickness (average Ihicknoss ol (he cortex represented by the solid grey area)  Figure 2.10: Diagram showing (A) H S A regions (B) Standard D X A proximal femur regions (C) Schematic of the cross-sectional geometric variables calculated by the HSA program. Reprinted from Petit e/tf/,2002. [46]  Geometric adaptations have been studied in vivo in only a handful of exercise loading studies. The cross-sectional area of the humerus in the dominant arm of tennis players can be as much as 35% larger in males and 28% larger in females than their non-dominant arm [47]. Furthermore, exercise interventions in young girls showed that early-pubertal girls increase their CSA and section modulus at the femoral neck [46]. These studies underline the importance of looking at both bone mass and bone geometry when examining mechanical loading.  2.10 Lifestyle Factors Lifestyle factors such as diet and exercise have had implications as both treatment for, and preventative therapies against, low bone mass. Such therapies have few or no side effects and are applicable to all classes of patients. Furthermore, the preventative element of lifestyle therapies allows a proactive versus reactive management of at-risk cases. 17  2.10.1 Dietary Calcium and Calcium Supplementation in Children and Adolescents Recently, calcium supplementation has been shown to increase bone mineral accretion during adolescence [48]. Murphy et al, [48] found that milk consumption before the age of 25 years was correlated with aBMD in women aged 44 to 74 years. As such, milk consumption before the age 25 years predicted aBMD at the proximal femur even when age, menopausal status, smoking and alcohol use, physical activity and parity were controlled. Dietary calcium intake before age 25 accounted for only 1.5 to 2.0% of the total variance in aBMD for these women. Nieves et al, [49] studied the relationship between teenage and adult diets and aBMD of the forearm and total proximal femur in 139 adult women (30 - 39 years). Dietary calcium intake was assessed using food frequency questionnaires (FFQ) for food consumption during both adult and teenage (age 13 — 17 years) periods. Women who had calcium intake greater than the median for both teenage and adult periods had a significantly greater femoral neck aBMD than those who consumed less calcium than the median for both time periods.  Children whose calcium intake was below the recommended daily intake (800 mg/day) provided the most compelling evidence for a relationship between sufficient calcium intake and higher childhood aBMD. Distal radius B M C in 5 year olds from Hong Kong was positively related to calcium intake since birth. In comparison with same age children from Mainland China, Lee et al, [50] showed that children from Hong Kong who had a mean calcium intake of 542 mg/day had a radial B M C 14% higher than children from Mainland China who had a mean daily calcium intake of 244 mg/day [50, 51]. Such compelling evidence has not been seen in other observational studies of childhood calcium intake due to the proposed threshold effect. That is, calcium intakes above the threshold amount have  18  no further relationship with aBMD, however a clear relationship between calcium intake and aBMD can be observed for intake below the threshold [52].  Barr etal, [53] studied 45 girls (aged 10.5 ± 0.7 years) to evaluate the effect of calcium intake and eating attitudes on B M C over two years. They showed that levels of oral control and habitual calcium intake significandy predicted total body B M C , lumbar spine B M C and 2-year change in total body BMC. As such, oral control accounted for 0.9 — 7.6% of the variance while calcium intake accounted for 1.6 5.3%) of the variance in total body B M C and lumbar spine B M C at baseline, 2 years and for 2 year change [53].  Bonjour et al, [54] studied change in aBMD, B M C and bone size over one year in 77 calcium supplemented (850 mg/day) versus 67 non-supplemented pre-pubertal girls (aged 7.9 ± 0.1 years). They showed a greater increase in aBMD at the distal radius (43% greater), femoral neck (40% greater) and lumbar spine (8% greater) for supplemented girls. Furthermore, the greatest increase was seen in supplemented girls who had a spontaneous calcium intake below the median (880 mg/day) prior to study. Likewise, low-calcium consumers experienced a significant increase in mean B M C , bone size and height although this increase may have been a function of growth and not due to supplementation [54]. However, there were a larger proportion of girls in later pubertal stages in the intervention group. Therefore, bone gains and in particular height gains in the calcium supplementation group are more likely due to maturational differences.  A follow up study of 118 girls from the initial calcium supplementation study of Bonjour e/al, 1997 was taken 3.5 years after the intervention [55]. A significant difference in all sites measured remained in the intervention group. Due to the unbalanced maturational status in each of the experimental 19  groups, the analysis was further subdivided into early and late pubertal status. A s the group sizes between experimental groups in early pubertal stages (Tanner Breast Stage 1 -3; n = 42 placebo, n = 39 calcium supplementation) were approximately equal, this subset was further analyzed. They showed a significant difference only in femoral neck (p = 0.04) and femoral diaphyseal (p = 0.03) a B M D between calcium supplemented and control girls in early stages o f puberty, 3.5 years after the intervention [55].  Johnston et al, [56] showed that when one sibling of a pair o f monozygotic twins was supplemented with 1000 mg of calcium-citrate malate per day for 3 years during pubescence, a significant increase in a B M D (mean = 1.4%) was observed in the supplemented twin compared to the placebo treated twin at most skeletal sites. Furthermore, there is even preliminary evidence to suggest that supplementation of calcium to college-aged students increases bone mineral accretion even in late phases of skeletal growth [57].  Nowson et al, [58] compared the inter-pair difference of calcium supplementation (on a B M D in 42 female twin pairs (age 10 — 17 years). They found that after 18 months the supplementation group had a greater lumbar spine a B M D (+1.62%).  After the initial 6 months o f supplementation there was a  significant difference at the proximal femur and lumbar spine a B M D , 1.27% and 1.53% respectively, between supplemented and placebo twins. The inter-pair difference in the changes of a B M D beyond 6 months was not significant. It should be noted that this cohort had a wide range o f maturity; however the majority o f the girls were post-pubertal (mean age 14 years). T h e effect of calcium supplementation in this study may have been blunted by the effect o f maturity-related bone gain.  20  A double blind, controlled study examined calcium supplementation of 7 year olds (87 boys and 75 girls) in China with spontaneously low calcium diet (280 mg/day) [59]. The radial B M C and radial BMC/bone width increased significantly in the supplementation group over the 18 month intervention period. However, the 18 month follow-up after cessation of supplementation showed that the initial increase seen in the treatment group disappeared without high calcium supplementation [60].  Table 2.1: Summary table of prospective calcium supplementation bone parameters.  Study  Lee  et al,  Population  [59]  Johnston et  al,  [56]  N o w s o n et al, [58]  Bonjour et al, [54]  Bonjour et al, [55]  Type of  Outcome  supplementation  Variable  7-year old Chinese boys and girls  300 mg/d C a supplementation  BMC  2.5% T in distal radius B M C Difference not observed on follow up  Pubescent M Z twins post-pubescent female M Z and D Z twin pairs  1000 mg/d C a supplementation 500 mg/d C a supplementation  aBMD  1.4% t in a B M D  Femoral neck and lumbar spine a B M D  1.3-1.5% t i n femoral neck and lumbar spine a B M D after 6 months of supplementation  Pre-pubescent girls  850 mg/d C a supplementation  aBMD, B M C and bone size  t a B M D . Low C a consumers had significant increase in a B M D , B M C and bone size with supplementation  Early s targe pubertal girls  3.5 year follow-up of [54]  Femoral neck a B M D and femoral diaphysis aBMD  t femoral neck a B M D in early pubertal girls 3.5 years after supplementation  2 +  2 +  2 +  2+  21  Results  2 +  2.10.2 Physical Activity Physical activity that increases internal strain on bone has been proven in animal studies to be beneficial for increasing bone mineral accrual. The global geometry of bones is determined by genetics, however it has been well documented that aBMD, internal and external bone diameters, and architectural properties of bone respond to external forces placed on the skeleton [61]. Over time increased internal strain increases bone mass. However, as bone mass increases there is a proportional decrease in strain as the load is spread over a greater fractional dimension of bone, thereby setting bone mass to a new equilibrium [5]. Strain elicits a response in bone in reaction to three strain variables: strain magnitude, strain rate and strain distribution.  In turkey ulna models, it has been shown that complete absence of strain by 8-week immobilization results in a 20% decrease in cross sectional area (CSA) while a daily treatment of 100 cycles of 1000 to 2500 uE (microstrains; 1 U.E equals a 0.0001% deformation of the bone), a range within normal daily loading for locomotion, resulted in a maintained CSA through periosteal and endocorticol bone formation [62]. Daily strains greater than 4000 p E resulted in a 40% increase in turkey ulna CSA.  High strain rate stimulates osteogenesis. Strain rate is identified as the time over which strain develops following load (strain) application [63]. Dynamic, compressive strains at 4000 p E applied to rat ulnae resulted in 54-67% greater periosteal bone formation in those that were loaded at the highest strain rate compared to those loaded at moderate to low strain rates [64].  Strains that cause a unique loading distribution to the skeleton may elicit the greatest response in bone. Bone responds to strains that are perceived as different by adapting architecturally to reduce the error [65]. When low magnitude, normally distributed strains, typical of locomotion were applied to a rat 22  ulna there was no osteogenic response [66]. However, equally low strain magnitudes typical of daily activities applied in unusual strain distributions elicited a substantial osteogenic response in sheep ulnae [67].  Exercise may have its most significant effect on the prevention of osteoporotic fracture by increasing bone mass and bone strength in the growing skeleton [40]. The maximum amount of bone is obtained within the first 2-3 decades of life, which is ultimately a key determinant of final bone mass and osteoporotic fracture risk [68]. From early evidence it was clear that active children had greater aBMD than inactive children [69]. Bone responds only to loads greater than the physiological loading zone [70]. Furthermore, there is greater bone formation if the mechanical strains are in novel patterns [71]. Higher impact activity, such as ballet and gymnastics has a significantly greater effect on increase bone mass than low impact activity in girls [72-74]. Likewise, aBMD at the lumbar spine and femoral neck of male junior weight lifters was 24% - 31% greater than age-matched controls as well as reference values for adults [75].  Bass etal, [76] showed that among 45 pre-pubertal female gymnasts (aged 10.4 ± 0.3 years), the change in aBMD over one year in weight bearing sites (arms, legs and spine) was 30% - 85% greater than agematched, non-active controls. Evaluation of retired gymnasts (mean 8 + 1 years since retirement) showed that they had a 0.5 - 1.5 standard deviation higher aBMD than-age-matched controls. This suggests that pre-pubertal exercise that enhances aBMD may have a residual benefit of increased aBMD in adulthood.  Bailey etal, [40] conducted a six-year prospective observational trial of 53 adolescent girls and 60 adolescent boys [age at peak B M C velocity, 12.54(0.86) for girls 14.05(0.95) for boys] to determine the 23  effect of physical activity (PA) levels on the growing skeleton. Children in the lowest P A quartile (i.e. inactive children) where found to have a significandy lower peak B M C velocity (331 g/year vs. 409 g/years), B M C accrued for 2 years around peak height velocity (582 g vs. 699 g) and total body (9-17% difference), femoral neck (7-11% difference) and lumbar spine B M C (18% difference) 1 year after peak height velocity, then their moderately and highly active counterparts.  2.10.2.1 Prospective Physical Activity Intervention Studies There is growing evidence that increased high impact exercise at a young age can have significant positive effects on bone mass. Morris et al, [77] studied 71 girls aged 9- and 10-years old. Thirty-eight of the girls participated in a 10-month intervention involving three, 30-minute sessions of varied P A per week in addition to their regular physical education classes. This group was compared with thirtythree girls who only participated in their regular physical education class. Gains in the intervention group for aBMD and B M C were 5.5% greater for total body, 3.6% greater for the lumbar spine and 10.3% greater for the femoral neck when compared with controls. When weight and height adjustments are made for femoral neck aBMD, the gains in the intervention group ceased to be significantly different.  Fuchs et al, [78] showed that a mixed sample of boys and girls between the ages of 5 and 8 years old in a 7-month jumping program significandy increased lumbar spine (+3.1%) and femoral neck (+4.5%) BMC (p <0.05). Likewise, children aged 6 to 10 years old had 1.2% greater trochanter aBMD when involved in a high impact program introduced within a regular physical education class [79]. A follow up after a 7-month detraining period in initially pre-pubertal children showed that the 4% increase in femoral neck B M C was maintained. However, there was no difference between previous intervention subjects and controls at the lumbar spine after follow up [80]. 24  The relative difference in aBMD due to high impact exercise is two-times greater if started prior to menarche in girls [81]. Heinonen etal, [82] conducted a randomized control trial of a 9-month stepaerobic program in pre-menarcheal and post-menarcheal girls (age 10-15 years). They found that the trained pre-menarcheal but not post-menercheal girls had a significantly (p = 0.012) higher lumbar spine (+3.3%) and femoral neck (+4%) aBMD then controls. However, pQCT scans of the distal tibia did not show a significant difference in cortical CSA, cortical density or density weighted polar sectional area between trained children and controls in either pre- or post-menarcheal girls.  Petit etal, [46] used H S A (section 2.9, p. 19-21) to examine a randomized control trial of a school based exercise program in pre- and early pubertal girls (aged 9-12 years). Although the pre-pubertal girls showed no significant difference in HSA parameters, the early pubertal girls (Tanner Stage 2-3) showed significant gains in femoral neck (+2.6%) and intertrochanteric (+1.7%) aBMD compared to controls. Furthermore, these changes were underscored by increased CSA and reduced endosteal expansion thereby increasing the section modulus of the femoral neck (4.0%).  Bradney et al, [83] conducted an 8-month exercise trial with 20 pre-pubertal boys (aged 8.4-11.8 years). The intervention was a regimen of 3 x 30-minute weight bearing exercise sessions per week. At final measurement the increase in aBMD at the lumbar spine and T O T A L B O D Y was approximately double that for boys in the exercise group then age matched controls. Further, in the exercise group femoral midshaft cortical thickness increased by 0.97%/month due to a 0.93%/month decrease in endocortical diameter without periosteal expansion thus a B M D , increase of Vo  1.14%/month while controls did not change in these bone parameters.  25  MacKelvie et al, [84] conducted a 7-month jumping intervention in prepubertal 92 Asian and Caucasian boys (10.2 ± 0.6 years). Boys in the intervention group with a Body Mass Index (BMT) at the beginning o f the intervention at or below the 75 percentile gained significantly more total body th  and lumbar spine B M C , proximal femur and trochanteric a B M D (+2%) than controls at or below the 75 percentile for B M L F o r the prepubertal control and intervention boys in the highest quartile for th  B M I there were no significant differences in B M C or a B M D gained over the intervention period. These data suggest that jumping interventions increase bone mass accrual in prepubertal boys of average or low weight-for-height while having no effect on bone mineral accrual in high weight-forheight boys.  MacKelvie et al, [85] compared 87 girls in a 7-month jumping intervention with 90 normal activity control girls (age 8.7 —11.7 years). When the girls were divided into pre pubertal (Tanner stage 1) and Early Pubertal (Tanner Stage 2 and 3) they showed that there was no difference in 7-month change in bone mass between intervention and control girls in the prepubertal group. Conversely, girls in the intervention group that were in early pubertal stages had a 1.5 — 3.1% greater femoral neck and lumbar spine bone mass than controls. MacKelvie et al, concluded that in girls, early puberty may be an opportune period during growth for exercise interventions to have a positive effect on bone health.  It is unclear whether increased bone mass during the growing years aids in preventing low bone mass in later life. Although preliminary evidence supports the hypothesis that increased bone mass due to higher levels o f physical activity in childhood [74, 76] results in higher bone mass in adulthood, a prospective study that confirms this notion has not been carried out.  26  Table 2.2: Summary table of prospective studies of childhood physical activity effect on bone parameters. Study  Population  Type of Intervention  Outcome  Results  Variable  Fuchs et al, [78]  Children, 5-8 years  7-month jumping program  BMC  4% t lumbar spine, 3% t femoral neck B M C for intervention group vs. controls Follow up after detraining showed no difference in lumbar spine B M C  McKay et al, [79]  Children, 6-10 years  10 minutes, 3x/week jumping  aBMD  1.2% t trochanteric a B M D for intervention group vs. controls  MacKelvie et al, [84]  prepubertal boys, 10.2 + 0.6 years  10 minutes, 3x/week jumping  BMC, aBMD  2% "l in total body and lumbar spine B M C , proximal femur and trochanteric a B M D in intervention group boys at or below the 75 %tile for BMI 1.5 - 3.1% t in femoral neck and lumbar spine bone mass in early pubertal intervention group girls 2x greater total body and lumbar spine a B M D for intervention vs. controls th  MacKelvie et al, [85]  Pre and Early pubertal girls, 8.7 -11.7 years  10 minutes, 3x/week jumping  BMC, aBMD, B M D V O L  Bradney et al, [83]  prepubertal boys, age 8 . 4 - 11.4 years  3x 30 minutes weight bearing exercise/week  aBMD  Morris et al, [77]  Girls, 9 - 1 0 years  3x 30 minute varied PA / week  BMC  Petit et al, [46]  Girls, 9-12 years  10 minutes, 3x/\veek jumping  HSA (aBMD)  Heinonen et al, [82]  Girls, 10-15 years  9-month step aerobics  Femoral neck, lumbar spine aBMD  5.5% t total body, 3.6% T lumbar spine 10.5% T femoral neck B M C for intervention group vs. controls Early pubertal girls ] femoral neck (2.6%) and intertrochanteric (1.7%) a B M D . Increased CSA and section modulus in H S A . 3.3% and 4% \ in lumbar spine and femoral neck a B M D compared to controls  2.10.3 Calcium and Physical Activity Interaction Neither calcium supplementation nor physical activity act on bone metabolism in isolation. Indeed, it has been shown that bed rest and immobilization result in bone loss despite calcium supplementation  27  [86]. Conversely, Lanyon  etal,  [87] showed that despite a loading regimen that had been previously  demonstrated to increase B M C in turkey ulna, turkeys that were on a calcium deficient diet lost bone compared with their supplemented counterparts.  A study o f 8 to 16 year old boys and girls demonstrated that weight-bearing activity predicted the highest increase in bone, measured at the distal radius, in the presence o f high calcium intake for children under 11 years o f age [88]. Equally, Slemenda  etal,  [89] showed a significant effect of P A and  distal radius, proximal femur and lumbar spine B M C accrual over a 3-year period. A significant effect of calcium intake was also observed for prepubertal children. Nevertheless, there was no evidence to suggest an interaction between calcium supplementation and P A on B M C accrual.  2.11 Bone Mass as a Multifactorial Polygenic Trait The processes o f formation and resorption regulate the amount o f bone mass in the skeleton in any region, at any stage o f life. These processes are influenced by several different genes and many nongenetic factors. Identifying genes that either directly or indirectly influence bone mass is thereby confounded by the effects of other genes that affect the dynamics o f the formation/resorption process that has yet to be identified.  In addition, these genes interact and are expressed within a  heterogeneous environment.  2.11.1 Heritability Estimates ofAbsolute Values of aBMD Heritability analyses o f a B M D during the late 1980s and early 1990s using twin models showed that as much as 65% - 92% o f the variance in a B M D may be attributed to genetic factors [90-93]. Furthermore, the genetic contribution to a B M D remains strong in elderly populations [69]. Likely, the large discrepancy in heritability estimates is due to higher degrees of common environment in monozygotic than dizygotic twins thereby inflating the estimates of genetic influence [69]. Heritability 28.  estimates using sibling models has yielded substantially lower, albeit significant, estimates of genetic contribution to aBMD [94, 95].  There appears to be a common set of genetic and environmental factors that influence aBMD at various skeletal sites. Nguyen et al, [96] showed that the genetic correlation between lumbar spine and femoral neck aBMD was 0.64 whereas the environmental correlation was 0.57. This suggests that the genetic factors that influence femoral neck aBMD are more likely to influence lumbar spine aBMD (and vice versa) than environmental factors, such as diet and PA.  2.11.2 Heritability Estimates of Bone Turnover Parameters The genetic contribution to serum levels of markers of bone turnover is well established. Genetic factors account for 65% of the variance in markers of bone formation such as C-terminal propeptide of type I collagen (PICP), osteocalcin and bone specific alkaline phosphotase (BSAP) between premenopausal female twins [97, 98]. Furthermore, each standard deviation increase in BSAP (6 pg/litre) was associated with a 4% lower aBMD in the femoral neck and lumbar spine [99]. Yet, the genetic relationship between aBMD and BSAP is likely due to unshared genes. Variance in BSAP concentration accounts for only 12% of the total genetic variance in lumbar spine aBMD and 4% of the variance in femoral neck aBMD [99],  2.113 Heritability Estimates of Change in aBMD Data on the genetic contribution to rates of loss of aBMD during adulthood are scarce. Data on genetic contributions to bone mass accrual during skeletal growth are also lacking. It has been shown that the intraclass correlation for change in vertebral aBMD of monozygotic (MZ) and dizygotic (DZ) pre- and post-menopausal women followed for 5 years is higher for M Z twins than D Z sisters, however this difference was not statistically significant [100]. Likewise, male M Z and D Z twins 29  followed for 14 years showed that M Z twins had a higher correlation (r = 0.61) 2  than D Z twins (r  2  -  0.41) on rate o f bone loss at the distal radius (NS) [92].  2.11.4 Bone Parameter Segregation Analysis to Determine Models of Genetic Transmission The method of genetic transmission o f a B M D within families is determined to allow a genetic model to be established upon which molecular studies can be based. Gueguen  etal,  [101] studied 129 nuclear  families from the Nancy-Vandoeuvre region of France and created a genetic model o f a B M D . Through maximum-likelihood testing o f segregation analysis and variance component analysis, several potential models o f genetic transmission were deemed unsuitable (i.e. classical mendelian, completely environmental, sporadic). T h e best model was a polygenic transmission mode o f inheritance, assuming no-major gene effect, with a residual (environmental) effect. Furthermore, through decomposition o f the variance components, the residual component was separated into: 1) variance due to common environment and 2) variance due to residual factors o f the individual. This analysis revealed that the best model with the above assumptions of polygenic transmission involved no residual variance due to common environment. However, the residual factors o f the individual did show a relationship with age but not gender [101].  2.12 Identifying Genes Related to Bone Mass There are two possible routes by which to identify genes involved in osteoporosis: positional cloning through genome scans and candidate gene approaches. Both methods have been used successfully for genes that are involved in classical Mendelian diseases (e.g. Huntington's Disease, Cystic Fibrosis, etc.). However, these approaches are mired in methodological problems associated with a complex inheritance such as bone mass. Although both these approaches have been used to identify some of the genes in multifactorial polygenic disease (e.g. breast cancer, asthma, schizophrenia), by and large  30  these approaches in isolation are not likely to find all the genes involved in these complex genetic diseases [102].  2.12.1 The Candidate Gene Approach The candidate gene approach tests previously identified genes, and their alleles, for their association with a disease. The most troublesome aspect of this approach in studying bone mass is that each candidate gene, when investigated in isolation, is only likely to have a minor effect on the trait. Therefore, some may reach or not reach statistical significance by chance alone or due to an artifact (e.g. population admixture) in the population studied. Statistical adjustments can be made to correct for this, but as a candidate gene has an a priori probability of being associated with the disease such statistical modifications are overly-corrective.  2.12.1.1 Evidence from the Candidate Gene Approach The classic series of papers reporting an association of polymorphisms in the 3' non-coding region of the Vitamin D Receptor (VDR) gene by Morrison  etal,  used the candidate gene approach [103,104].  These polymorphisms were first believed to have a strong predictive power of B M C in the lumbar spine. Follow up investigations with other populations and meta-analysis have shown a much weaker association between these polymorphisms and lumbar spine aBMD than previously reported [105]. Similarly, other genes [e.g. Oestrogen receptor 1 (ER1), Collagen loci (COL1A1), Insulin-like growth factor-1 (IGF1), Interleukin-1 (11-1), etc.] have been shown to be associated with aBMD and/or bone turnover [106-110].  2.12.2 Positional Cloning Approach The positional cloning approach uses a different paradigm to determine the genomic position of genes involved in disease. By using highly polymorphic markers distributed throughout the genome (e.g. 31  microsatellite markers), investigators can follow the segregation of these markers with the disease within farnilies. Such a study design is known as a "linkage study". Conversely, non-familial populations can be scanned for the same markers to determine if there are particular markers that cosegregate with the phenotype. Such a study is known as an "association study". Both linkage and association studies assume that the gene of interest is near, chromosomally linked, with the markers that are segregating with the phenotype. More refined co-segregation analysis of the marker and disease can be examined to pin point the region or regions of the genome on which the genes of interest reside. Positional cloning has been reviewed in detail elsewhere [102].  Association studies are prone to spurious results due to non-genetic factors that may suggest association of a region of the genome by chance segregation of the non-genetic factor and a particular marker allele. Thus, the Type I error rate in population genome scans is a proportion of each specific loci's significance level by the number of loci tested across the genome. Hence, to have an overall genome Type I error rate of 0.05 each locus' significance level must be adjusted accordingly [1].  2.12.2.1 Human Studies Applying the Positional Cloning Approach Koller et al, 2000 studied sister pairs and 1 parent when possible in 464 Caucasian- and 131 AfricanAmerican pre-menopausal women (aged 20 - 45 years) [111]. They evaluated the association of 270 evenly distributed (average distance between markers = 12.9 cM) autosomal microsatellite markers with aBMD at the lumbar spine, greater trochanter, femoral neck and Ward's triangle. They used multi-point quantitative linkage analysis with the genomic markers and aBMD at each site, setting a Logarithm of the Odds (LOD) score of greater than 1.85 suggesting regions of interest. Results are listed in Table 2.3.  32  Table 2.3: Linkage data from Roller et al, [111]. maximum L O D score for the markers in the regions.  Chromosomal Region  lq21-23 5q33-35 6pll-12 llql2-13 14q31-32 22ql2-13  aBMD  Site  Lumbar spine Femoral neck Lumbar spine Lumbar spine Trochanter Lumbar spine  Max  Chromosomal region, a B M D site and  LOD  Score in  Max  LOD  Score in  Caucasian subpopulation  combined population  3.11 1.87 1.94 1.97 1.99 2.13  3.86 2.23 1.93 1.75 1.99 0.99  These results show that the greatest evidence for linkage in lumbar spine aBMD was on chromosomal region l q near marker D1S484. Within this region there are several candidate genes that are involved in bone metabolism (i.e. osteocalcin, interleukin-6 receptor, calcium binding proteins). There was also suggestive linkage for femoral neck aBMD at the 5q33-35 locus near marker D5S422 and for lumbar spine aBMD at 6pll-12. Candidate genes in these regions are osteonectin and platelet derived growth factor-P (PDGFP) at the 5q region and BMP 6 and 5 at the 6p region. The sharp decline in L O D score values for the 22q 12-13 region when the larger population was used suggests that the initially high L O D score is likely false positive evidence of linkage. The llql2-13 locus had been previously described as a region linked with an autosomal recessive form of familial osteoporosis [112-114]. O f particular interest, there were no genomic regions that were in linkage with more than one skeletal site suggesting that different genes are involved in aBMD phenotype in different skeletal sites.  To further investigate the 11 ql2-13 genomic region, Deng etal, [115] investigated 635 individuals from 53 pedigrees with normal variation in aBMD after being adjusted for age, sex and weight. They investigated 5 polymorphic markers within the 27 cM 1 lql2-13 region. They found that there was no evidence of significant linkage to any of the markers in the 1 lql2-13 region with variance in aBMD.  33  However, the population was different from previous studies of that region and therefore this region cannot be discounted as potentially moderating a B M D in non-Chinese populations.  A study conducted in China sought to define genomic regions involved in distal and proximal forearm a B M D [116]. Ninety-six families with 153 sib-pairs and a total o f 218 subjects (age 30.3 ± 6.3 years) were studied. A peak L O D score of 2.15 was found at chromosome 2 near marker D2S2141, D2S1400 and D2S405 for both distal and proximal forearms a B M D . A L O D score o f 1.67, suggesting linkage, at chromosome 13 near marker D13S788 and D13S800 for distal forearm a B M D was also reported. Candidate genes in these regions are calmodulin 2 (2p21.3-p21.1), a putative serine/threonine kinase (2p23-24), pro-opiomelanocortin (2p23.3), collagen type 4, a - l (13q34) and collagen type 4, a-2 (13q34). These candidate genes are between 2 and 7 M b from the markers.  A report by Zee et al, 2001 investigated linkage around the V D R chromosomal region (12ql2-14) [117]. Femoral neck, Ward's triangle, Trochanter and lumbar spine a B M D (adjusted for age, age , sex, 2  B M I , height, alcohol consumption, caffeine intake, smoking status and estrogen supplementation) was tested for linkage with 4 markers in the 12ql2-14 region in 1062 subjects (aged 37 - 89 years) from the Framingham Heart Study. O f these subjects 169 sib-pairs were analyzed separately. There was no evidence of linkage in these subjects and the 12ql2-14 region.  Devoto et al, 2001 continued on the prior evidence that genomic region lp36 was linked with femoral neck a B M D [118,119]. Through further investigation of the lp36 region they defined a significantly linked region around lp36.2 - lp36.3. Within these regions there are no known bone specific genes however, there are 16 genes o f unknown function as well as approximately 40 gene-like sequences of unknown function that may have bone specific genes. Furthermore, within this region there are 34  several genes of known osteoclastic and osteoblastic function, albeit such function is not exclusive to such cell lines. As such, members of the Tumour Necrosis Factor (TNFR) Superfamily (TNFR2, TNFRSF14) and PLOD the gene encode lysyl hydroxylase, a collagen modifying gene are within this region.  2.12.2.2 Animal Studies Applying the Positional Cloning Approach Whole genome scans can use inbred laboratory mouse strains to detangle issues of genetic and environmental heterogeneity. Klein et al, [120] used inbred mouse strains in a novel approach to scanning the genome. Two genetically distinct mouse strains with discordant mean values for body weight and aBMD were used as progenitors of 24 genetically and phenotypically novel recombinant inbred strains. By crossing the progenitor strains Klein et al allowed a random shuffle of genomic markers that were fixed in either strain prior to crossing. By analyzing phenotypic values of aBMD with the presence of markers from one strain or the other they were able to identify ten genomic regions implied in B M C accrual. Table 2.4 shows the results of this mouse genomic scan with the syntenic human regions as well as candidate genes within these regions. Analysis of the genetic diversity of these mice was continued using single nucleotide polymorphisms (SNPs) to scan the genome [121]. The SNP analysis found another region near the centromere of mouse chromosome 13 that had not been previously implicated in bone metabolism.  Table 2.4 : Mouse and corresponding human chromosomal regions implied in B M C accrual. Reprinted from Klein et al, [120]. Mouse chromosomal position  Human chromosomal Position  Candidate genes  1:74  lq32  p  2:6  10ql5-14  IL1  2:86  20qll-12  BMP2  7:44  llql3 or llq21  IGF1 receptor, P T H , C T , FRA1  11:57  17qll-12  COL1A1  14:2  3pl4  BMP4  35  15:43  llpl3 or 22ql3  16:19  3q27-28  18:48  18q21 (?)  ?  19:53  10q23-26  ?  PDGFfi Ca2+ sensitive receptor  Note that several genes previously implicated in bone mineral metabolism are listed in the candidate regions. However, V D R , which has been mapped to human chromosomal region 12ql2-14, was not found to have a significant effect on B M C accrual in the recombinant inbred strains. It is unknown whether this is an artifact of the mouse strains used and further investigation using other inbred strains is required to determine this. Furthermore, candidate genes have been noted from the chromosomal regions indicated. However, any gene or polymorphic region of sequence within the candidate regions is equally likely to be involved in bone mineral metabolism. From these data, interactions between genes and the environment are missed as the environment is homogenous for all genetically divergent strains.  Shimuzu etal, [122] studied a cross of 2 inbred mouse strains: SAMP6 (bred for senile osteoporosis) and SAMP2 (high peak aBMD). Two chromosomal regions were in strong linkage with low cortical thickness: mouse chromosome 11 between D l l M i t 9 0 and D l l M i t 5 9 had a L O D score of 10.8. Candidate genes in this region are collagen type 1, a - l (COL1A1) and granulocyte stimulating factor. Likewise, mouse chromosome 13 around region D13Mitl35 had a L O D score of 5.8. There are no putative bone metabolism genes in that region.  Nine-hundred and eighty-six female mice were studied by Beamer et al, [123] to further scan the genome for quantitative trait loci (QTL), regions that contribute subdy to the phenotype. These mice were crosses of progenitors (C57BL/6J and C3H/HeJ) divergent in femoral (50% divergence) and vertebral (9% divergence) aBMD. They found markers on 10 mouse chromosomes (1, 2, 4, 6,11,12, 36  13,14,16 and 18) carrying QTLs for femoral aBMD while markers on 7 mouse chromosomes (1, 4, 7, 9,11,14 and 18) carried QTLs for vertebral aBMD. A pair-wise evaluation using A N O V A did not reveal significant gene-by-gene interaction between QTLs common to either bone site. QTLs for femoral aBMD accounted for 35.1% of the variance, while QTLs for vertebral aBMD accounted for 23.7% of the total population variance in aBMD at either site. This study underlines the genetic complexity of multiple genes and gene interactions that are acting both in concert and individually at different sites of the skeleton.  Further investigations of the genetic, gene-by-gene and gene-by-environment influences on aBMD are warranted. Genes in isolation that have shown meager relationships with aBMD may prove, in the end, to explain much more of the variance in aBMD if interactions with the environment and other bone mineral metabolism genes are included in the analysis. Furthermore, it has been shown that the total genomic influence on aBMD is greater during the first 2 to 3 decades of life [124]. Therefore, genetic influence on aBMD and the change in aBMD over the first two decades of life may show a more significant relationship than previously reported older subject, in cross-sectional aBMD analysis.  2.13 The Vitamin D Receptor (VDR) Mendelian Inheritence in Man (MIM) ascension #601769, Gene map locus 12ql2-14 (12ql3.11)  The vitamin D receptors are intracellular polypeptides of 50 to 60 kD that specifically bind 3  l,25(OH) D and interact with target-cell genomic sequences to produce a variety of biologic effects. 2  3  Multiple sequence alignment of the deduced amino acid sequence showed that the V D R protein belongs to the superfamily of trans-acting transcriptional regulator)' factors, including the steroid and thyroid hormone receptors (NR1 subfamily) [125].  37  The V D R gene contains 11 exons and spans approximately 75 kb. T h e noncoding 5' end of the V D R gene includes exons 1A, IB, and 1C, while its translated product is encoded by 8 additional exons (2 to 9) [126]. Three unique m R N A isoforms are produced as a result o f the differential splicing of exons IB and 1C [126]. T h e D N A sequence upstream to exon 1A is G C - r i c h and does not contain an apparent T A T A box. A n intron fragment 3' of exon 1C conferred retinoic acid responsivity when fused to a reporter gene suggesting a molecular mechanism for the observed ability o f retinoic acid to induce V D R expression [126].  Figure 2.11: Schematic of V D R m R N A and protein product highlighting the V D R SCP and 3'UTR polymorphisms.  2.13.1 Human Studies of VDR Polymorphisms The synthesis o f osteocalcin is induced by calcitriol, the active hormonal form o f vitamin D , through a specific vitamin D - V D R responsive element in the osteocalcin gene promoter. In studies o f twins, variation in serum osteocalcin levels was shown to have a major genetic component [97] and to be 38  closely correlated with the genetic diversity in aBMD as measured by D X A [91]. Morrison etal, [103] set out to ascertain whether variability in circulating osteocalcin levels reflected allelic variation in the V D R gene. They analyzed the relationship between polymorphic sequence in V D R (Figure 2.11) and serum osteocalcin levels in a cohort of normal subjects. They found that the B / b RFLP in the V D R gene predicted circulating osteocalcin levels independent of age or menopausal status. Their research demonstrated that common alleles of V D R were functionally different and worked in trans to contribute to physiological variation in osteocalcin levels. Morrison et al, [104] described preliminary analyses in M Z and D Z twin pairs, and showed that the greater diversity in lumbar spine aBMD between D Z pairs could be explained by V D R polymorphisms. The common allelic variants in the V D R gene can be used to predict differences in aBMD and accounted for up to 75% of the total genetic effect on bone density in healthy persons. The genotype associated with lower aBMD was over-represented in post-menopausal women with aBMDs more than 2 SD below values for young normal women. Minigene analysis of the two most frequent haplotypes, BAt and baT, showed significant differences in luciferase reporter gene activity (1.44:1 and 1.38:1) in kidney and bone cell lines [104]. Likewise, reverse transcriptase-polymerase chain reaction analysis (RT-PCR) support the hypothesis that allelic differences in the 3'-untranslated region (3'UTR) have significandy different mRNA levels of V D R [104]. Mocharla etal, [127] detected a 30% decrease in m R N A levels in t alleles compared to T genotyped individuals. Conversely, examination of peripheral blood mononuclear cells of BB, Bb and bb genotyped individuals showed no difference in V D R m R N A levels.  Durrin etal, [128] showed that the 3'UTR Restriction Fragment Length Polymorphisms (RFLPs) are in linkage disequilibrium with two alleles of a poly(A) microsatellite also (S = A . and L = A _ ) in the 14  39  17  lg  24  3'UTR of V D R in Caucasian and African-American populations. Although careful examination of the 3' region was undertaken there was no evidence that the previously described RFLPs or poly(A) microsatellite was associated with destabflization of the mRNA product. However, three sub-regions within the 3'UTR were identified as destabilizing sequences that where associated with decreased mRNA half-life. Thus, it is plausible that some of the contradictory evidence surrounding the V D R 3'UTR region is due to these uninvestigated sequences.  2.13.1.1 Allelic Frequencies in Different Racial Populations Spotila et al, [129] investigated the V D R BsmI genotype frequency in three ethnic regional groups [Southern European (Italian, Greek), Eastern European (Ashkenazi Jewish, Polish, Hungarian), Western European (British, French)] living in the Montreal area. The frequency of the V D R BsmI allele in the 3 ethnic/regional groups did not significantly differ from the Australian population, of British and Irish descent, in previous reports [104]. Pearson % analysis of the 3 ethnic/regional and 2  the 3 possible genotypes (BB, Bb and bb) suggested a relationship between ethnic group and genotype (p = 0.048, X  2 =  9.572, df = 4). Differences in the 3 groups appeared to be greatest as an under-  representation of the bb genotype in Southern European subjects and a lower frequency of BB genotypes in Eastern European subjects. Nevertheless, there was no significant association between V D R BsmI genotype and differences in ethnic/regional difference in mean aBMD.  The frequency of V D R alleles has also been examined in southern Chinese women [130]. They showed that there was a lower frequency of the B haplotype in southern Chinese women versus Caucasian populations, 0.7% BB vs. 10% -17%, respectively. This is consistent with previously reported allele frequencies in Korean and Japanese populations [131,132]. Tamai etal, [133] compared V D R BsmI genotype frequency in 90 Japanese women (age 70 ± 10 years) with osteoporosis with 36 40  healthy controls (age 43 + 17 years). There was no significant difference between either group and B and b allele frequency (osteoporosis group: BB 5.6%, Bb 12.2%, bb 82.2%; control group BB 3.3%, Bb 17.4%, bb 79.3%). Furthermore, this evidence showed that unlike Caucasian V D R Bsml genotype data, the B allele does not correspond with lower aBMD in Asian populations.  Bell et al, [134] compared 39 young African-American men (age 20 — 40) with 44 age-, height- and weight- matched Caucasian men to investigate if V D R genotype explained mean differences in aBMD due to ethnicity. It was previously established that African-Americans had a higher mean femoral neck and lumbar spine aBMD [135]. Serum levels of osteocalcin and PICP were also tested to investigate if these markers explained differences in absolute aBMD. When both races were grouped together the VDRApal site was associated with higher serum concentrations of PICP and higher lumbar spine aBMD. V D R genotypes were not significandy different in either ethnic group hence V D R genotype did not explain the differences in aBMD associated between these African- and Caucasian American men.  Fleet etal, [136] investigated V D R genotype and ethnic differences in aBMD for pre-menopausal women (aged 20- 40 years). They found there was no significant difference in V D R Bsml genotype distribution between African- and Caucasian-American women (Caucasian Frequency: BB 12%, Bb 45.8%, bb 42.2%; African Frequency: BB 8.3%, Bb 48.6%, bb 43.1%). There was not a significant difference for race by genotype interaction on age- or BMI-adjusted aBMD for either the femoral neck or the lumbar spine despite a significandy higher femoral neck and lumbar spine aBMD for AfricanAmerican women. When the groups were collapsed mean aBMD in the femoral neck was lower in BB genotyped women than the bb (8.1%, p = 0.034) or Bb (9.3%, p = 0.015) after controlling for BMI, age, race and race by genotype interaction. Similarly, adjusted lumbar spine aBMD was 6.4% (p =. 41  0.036) lower in B B genotyped women vs. B b genotyped women. There was no significant difference between Bb and bb genotype when the groups were collapsed. This study suggests that although a B M D is associated with V D R  BsmI genotype  in a sample of pre-menopausal women o f different  races, these genotypes do not explain racial differences in a B M D between African- and CaucasianAmerican women [135].  Table 2.5: Summary table of V D R BsmI genotype effects by ethnicities.  Study  Frequency by Race  Genotype Effect  Spotila et al, [129]  Western European — Reference Eastern European - lower B B frequency Southern European - lower bb frequency Chinese population - B B = 0.7%  N o difference in a B M D by V D R  Kung et al, [130] Tamai et al, [1331 Bell et al, [134] Fleet et al, [136]  BsmI  genotype  N o genotype effect  Caucasian population — BB — 10-17% Japanese Women - B B = 4%  N o significant association of fracture risk by V D R BsmI genotype  African-American men, Caucasian-American men — no difference in allele frequency African-American women, Caucasian-American women — no difference in allele frequency  N o significant association of a B M D by V D R genotype When ethnicities were collapased the BB genotyped women had a lower femoral neck aBMD  2.13.1.2 Association of VDR Genotype with Fracture Risk In a study 171 pre- (n = 25) and postmenopausal (n = 146) women (aged 59.6 ± 0.75 years) from the northeast o f Scodand, Houston et al, [137] found that women with the B B genotype had a higher femoral neck a B M D than individuals with the bb genotype.. Morrison et al, [104] had previously reported that Australian subjects with the bb genotype had 15% higher a B M D at the femoral neck than those with the B B genotype. Houston et al, [137] also compared the V D R genotype of a group of 44 patients with severe osteoporosis who had vertebral compression fractures and compared them with healthy age- and gender-matched controls. T h e prevalence of any specific V D R genotype was the same in those who had suffered an osteoporotic fracture and those who had not. They suggested 42  that V D R polymorphisms do not cause reduced aBMD, but rather are in linkage disequilibrium with a disease-causing locus in some populations.  Ensrud etal, [138] conducted a case-cohort study within a prospective study of 9,704 women aged 65 years or older. Vitamin D receptor Taql and Apal polymorphisms in women who experienced firstincident hip (n = 181), vertebral (n = 127), and other fractures (n = 223) were compared with those of control women selected randomly from the cohort. Average length of follow-up was 5.4 to 6.5 years. Allele frequencies did not differ between fracture cases and their respective controls. Vitamin D receptor genotype was not significandy associated with the risk of proximal femur, vertebral, or other fractures. The authors concluded that determination of these V D R polymorphisms is not a clinically useful test for the prediction of fracture risk in elderly women.  2.13.1.4 VDR Genotype and Hone Mineral during Childhood Sainz et al, [139] studied the relation of V D R genotype to skeletal development and variation in size, volume, and density of bone in 100 pre-pubertal American girls of Mexican descent (age 9.2 ± 1.4 years). They found an association with femoral shaft and vertebral cortical bone density measured by QCT. Girls with V D R Apal aa and V D R Bsml bb genotypes had 2% to 3% higher femoral shaft cortical bone density (p = 0.008 and p = 0.04 respectively) and an 8% to 10% higher lumbar spine cortical bone density (p = 0.01 and p = 0.03) than girls who had VDRApal A A and V D R Bsml BB genotypes. There was no association between the cross-sectional area of the lumbar vertebrae or the cross-sectional or cortical area of the femoral shaft and the V D R genotype.  Lorentzon etal, [140] investigated V D R gene polymorphisms in 90 healthy Caucasian males and their relation with parameters of body length at birth, and parameters of body length, aBMD, and bone area, at age 16.9 ± 0.3 years and at age 19.2 + 0.7 [140]. Boys with the V D R Bsml BB genotype were 43  shorter at birth [birth height for BB 49.3 ± 2.7 cm, Bb 50.7 ± 2.3 cm, bb 51.2 ± 1.8 cm (p = 0.01)] and grew less from birth to age 16.9 ± 0.3 [adolescent height for BB 175 ± 6 cm, Bb 180 ± 5 cm, bb 180 ± 5 cm (p < 0.05)] than their Bb and bb counterparts. Both during puberty (age 16.9 ± 0.3) and after puberty (age 19.3 + 0.7), the V D R BsmI BB genotype boys were shorter (p = 0.005 - 0.008) and had lower bone area of the humerus (BB 80.9 ± 8.5 cm , Bb 87.8 ± 9.7 cm , bb 90.8 ± 11 cm ), femur (BB 2  2  2  209 ± 20 cm , Bb 233 ± 27 cm , bb 231 ± 27 cm ), and total body (BB 2626 ± 152 cm , Bb 2816 ± 178 2  2  2  2  cm , bb 2815 ± 196 cm ) than the Bb and bb boys (p = 0.01, p = 0.02, p = 0.004 respectively). The 2  2  allelic variants were not related to aBMD at any site. The authors concluded that a prediction model including parental height, birth height, birth weight, and V D R alleles could predict up to 39% of the total variation in adult height in their study population. The V D R allelic variants alone contributed to 8% of the total variation.  Tao etal, [124] explored the relationship between V D R polymorphisms and aBMD in 114 (68 female, 46 male) healthy pre-pubertal Caucasian Australian children aged 7.2 + 0.3 years. Females homozygous for the tt V D R TaqI allele had a significantly lower aBMD than V D R TaqI T T homozygotes in the left femoral shaft, femoral neck, lumbar spine, and total body. V D R TaqI Tt heterozygotes had an intermediate aBMD at these sites. The V D R TaqI tt homozygotes were also significandy shorter and lighter. These effects were not observed in males. When B M D  vol  was  estimated using cylindrical bone models, a significant relationship between the V D R TaqI polymorphism and BMD  V()I  was seen only in the lumbar spine of the girls. The authors commented  that V D R allelic variation might be responsible for some of the variation in aBMD and B M D  vol  and  postnatal growth in pre-pubertal girls. The authors did not provided an explanation for the observed sex difference. Gunnes et al, [141] studied 494 pre-pubertal Norwegian children and Baroncelli etal, 44  [142] studied 209 pre-pubertal Italian children. Neither study showed a significant relationship between aBMD and the V D R 3' polymorphisms (Bsml, Apal, Taql).  Table 2.6: Summary table of V D R  Bsml  genotype effects in paediatric populations.  Study  Population  Genotype Effect  Sainz etal, [139]  100 prepubertal Mexican-American Girls  Lorentzon et al, [1401 Gunnes et al, [141] Baroncelli et al, [142]  90 Caucasian boys  V D R Bsml bb genotype was associated with 2-3% higher femoral shaft cortical bone density and 8-10% higher lumbar spine cortical bone density V D R Bsml BB genotype is associated with shorter birth, pubertal and adult height N o association of V D R Bsml genotype and childhood bone density  494 prepubertal Norwegian Children 209 prepubertal Italian children  N o association of V D R  Bsml  genotype and childhood bone density  2.13.1.5 VDR Start Codon Polymorphism Arai etal, [143] explored a T—>C polymorphism in the translation start codon in exon 2 of the V D R gene. Two-hundred and thirty-nine Japanese women (aged 24 — 70 years) were genotyped for the transition (f representing A T G at the translation start codon, F representing A C G at the translation start codon, assayed by use of Fokl restriction endonuclease). The next in-frame A T G sequence is 12 bp downstream therefore the T—»C polymorphism results in the disruption of translation of the first four amino acids of V D R . Thirty-two (13%) subjects were ff, 75 (31%) were FF and 132 (55%) were Ff.  Investigation of the lumbar spine aBMD in 110 pre-menopausal Japanese women (age 24 — 45 years) showed that the F F genotype was associated with a 12% higher lumbar spine aBMD than ff homozygotes [143]. The f polymorphism produced a 50 kD protein product whereas the F polymorphism produced a 49.5 kD product. This represents a consistent size difference where translation of the f allele-encoded protein +10 to +12 nucleotides is initiated upstream of the F allele. 45  The activity of the F allele-encoded protein was 1.7 times higher than the f allele product when assayed with vitamin D response elements. This polymorphism was the first that linked V D R activity and aBMD differences.  Harris et al, [144] investigated the V D R start codon polymorphism (SCP) in 72 African-American and 82 Caucasian-American pre-menopausal women (aged 20 - 40 years). Although a significant difference in haplotype frequency was seen by race (ff homozygotes were 4% in African-American vs. 18% in Caucasian-American and FF homozygotes were 65% in African-American vs. 37% in Caucasian-American) the relationship between race and V D R SCP genotype with aBMD at any site measured was not significant. However, when SCP genotype was included in the analysis of covariance models that compared aBMD of the African-American and Caucasian-American women the estimated differences in femoral neck aBMD between the 2 groups was reduced by 35%. In the group as a whole, ff genotyped women had a femoral neck aBMD that was 7.4% lower than that of FF genotyped women. Femoral neck aBMD was 4.3% lower and lumbar spine aBMD was 12.1% lower in Caucasian-American women with ff genotypes compared with Caucasian-American women with the F F genotype.  Ferrari etal, [145] investigated the association of V D R SCP and aBMD in 177 Caucasian, premenopausal women (age 39.5 ± 8.6 years) and 155 Caucasian pre-pubertal girls (age 8.1 + 0.8 years). There were no significant relationships between V D R SCP genotype and aBMD or body size in either group. Likewise, Lucotte etal, [146] examined V D R SCP genotypes in 124 post-menopausal French women (aged 45 - 90 years). They did not show a relationship between V D R SCP alleles and height, weight or aBMD at the lumbar spine or femoral neck. However, when subjects greater than 75 years  46  were eliminated from the group a significandy lower aBMD at the femoral neck was observed for V D R SCP ff genotyped subjects compared to Ff and F F subjects.  Ames et al, [147] studied the V D R SCP in 72 healthy children (aged 7 - 1 2 years) [147]. They monitored dietary calcium absorption (using stable isotope techniques) bone calcium deposition rates and total body aBMD. They found that children with the V D R Fokl FF genotypes had a mean calcium absorption that was 41.5% greater than children who where ff homozygotes and 17% greater than Ff heterozygotes. Likewise, aBMD was 8.2% and 4.8% greater in F F genotyped children than ff and Ff genotyped children, respectively. Periods of high calcium accretion (i.e. early puberty or high remodeling rates) are particularly sensitive to the V D R SCP isoforms of V D R due to the varying kinetics of intestinal calcium absorption.  Table 2.7: Summary table of V D R Fold genotype effects.  Study  Population  Genotype Effect  Arai et al, [1431 Harris etal, [144]  110 pre-menopausal Japanese Women  V D R Fokl F F genotype is associated with a 12% greater lumbar spine a B M D then ff genotype V D R Fokl ff genotype is associated with a 7.4% femoral neck a B M D for the whole group.  Ferrari etal, [1451 Lucotte et al, [146] Ames etal, [147]  72 African-American, 82 CaucasianAmerican Significant difference in allele frequency by ethnicity 177 Caucasian, pre-menopausal women, 155 pre-pubertal girls 124 French post-menopausal women 72 pre-pubertal children  N o significant association with V D R Fokl genotype and aBMD Lower femoral neck a B M D in ff V D R Fokl genotype women vs those with the F f or F F genotype 8.2%, 4.8% greater a B M D for F F vs F f children  2.13.1.6 Interaction Effects of VDR. Genotypes and Environment Ferrari etal, [148] investigated the relationship between V D R SCP and 3'UTR polymorphisms, serum parathyroid hormone (PTH) levels and aBMD in a random control trial of calcium-phosphate supplementation in young, healthy men. Among V D R Bsml BB homozygotes, aBMD z-scores were 47  significantly lower only in subjects who also carried the f allele at the V D R SCP polymorphic site. Parathyroid hormone levels were significandy higher in the V D R BsmI BB genotype at baseline and remained so under either a low or a high calcium-phosphate diet. Moreover, those V D R BsmI BB subjects on the low calcium-phosphate diet, had significandy decreased tubular phosphate resorptive capacity and plasma phosphate levels.  Salamone etal, [149] investigated the interaction between V D R genotype and diet and/or exercise in 470 premenopausal Caucasian women (age 46.9 + 1.9 years). They showed that subjects with a high dietary calcium intake (> 1036 mg/day) who had a V D R BsmI BB or Bb genotype had 5.9% (p = 0.06) higher femoral neck aBMD than subjects with the V D R BsmI BB or Bb genotype and low dietary calcium intake (<1036 mg/day). The same effect was seen with the V D R BsmI bb genotype, however the magnitude of the difference in aBMD was substantially smaller (0.13%). The dietary calcium and genotype interaction was not significantly different at the lumbar spine. A similar interaction was observed for the relationship between high weekly PA (> 1789 kcal/wk) and the V D R BsmI bb genotype which together were associated with a 6.2% (p < 0.01) increase in femoral neck aBMD compared with subjects who had a V D R BsmI bb genotype and partook of low P A (<1789 kcal/wk). A modest increase (0.97%) in femoral neck aBMD was also seen in those with the V D R BsmI Bb and BB genotype who were also more physically active.  Tsuritani etal, [150] conducted a 1-year study of the interaction of the VDR BsmI polymorphism and an exercise intervention on lumbar spine, femoral neck and calcaneal aBMD in postmenopausal women. There was no significant difference between the three genotypes and aBMD after the one year exercise intervention, although a significant difference in lumbar spine aBMD was observed between V D R BsmI bb exercisers and V D R BsmI bb controls (2.35% vs. -1.25%). This suggests that 48  bb genotyped postmenopausal women may be more responsive to exercise than other V D R genotypes.  Blanchet etal, [151] examined the V D R BsmI polymorphism and P A in 575 women aged 42 to 85 years (median age 63 years old). Overall there was no significant relationship between the levels o f leisure time P A or V D R BsmI genotype with a B M D of the proximal femur or lumbar spine. However, women who exercised more than 3 times a week with a V D R BsmI bb genotype had a lower lumbar spine a B M D than women who were active and carried the V D R BsmI B B genotype. Furthermore, when the population was separated by the median age of the group a similar result was found in the older population (greater than 63 years old).  2.14 Collagen Type I, a-1 (C0L1A1) Mendelian Inheritance in M a n (MIM) Ascension #120150, Gene M a p Locus 17q21.31-q22 Collagen is a fibrous protein complex that has a triple-stranded rope-like structure. T h e majority o f collagen in the skin, tendons, and bone is the same protein complex containing 2 a - l polypeptide chains, coded by the C O L 1 A 1 gene, and 1 a-2 chain, coded by the C O L I A 2 gene. Differences in the collagens from bone, tendon and skin tissues are a function of the degree o f post-translational modification o f collagen including hydroxylation of proline and lysine residues to maintain stability of the coiled-coil structure, aldehyde formation for cross-linking with other collagen fibrils, and glycosylation.  The mechanical strength o f bone is dependent on collagen fibrils [152]. Cross-links connecting the non-helical ends of a collagen molecule (telopeptides) with the adjacent triple-helical part of nearby collagen molecules determine the tensile properties of bone [153]. X-ray diffraction and  49  stereochemical investigations have revealed that differences in cross-linking profiles result in differences in molecular packing of collagen fibrils [154,155]. Correct alignment of collagen fibrils likely affects bone mineralization as nucleation of calcium appatite crystals begins in the gap region between cross-links [155].  2.14.1 Human Studies of COL.1A1 Polymorphisms  2.14.1.1 Association of COL1A1 Sp1 Transcription Factor Consensus Sequence Polymorphism with aBM Geometry and Fracture Risk Grant etal, [107] described a novel G—»T polymorphism at the first base of a binding site for the transcription factor Spl in the first intron of COL1A1. They examined 299 postmenopausal British women (aged 61.3 ± 0.74 years), and found a significant association between the s allele (Spl consensus sequence) and lower aBMD. Women with the SS genotype (homozygotes for the non-Spl consensus sequence) had a significantly higher aBMD at the proximal femur and lumbar spine (p = 0.03) than individuals with the Ss. A similarly reduced aBMD at both sites was also observed for ss genotyped women when compared with Ss genotyped women. However due to the rarity of the ss genotype this relationship did not reach significance. Furthermore, relative risk of compression fracture was 2.97 (95% confidence interval) in individuals who carried at least one s allele.  Biochemical analysis of Spl transcription factor binding to the S and s alleles, osteoblastic C o l l A l to Col2Al protein production in S and s alleles and quantification of mRNA product of the S and s alleles was investigated by Mann etal, [156]. They showed that the Spl transcription factor binds poorly to the S allele compared to the s allele. RNase protection assays showed that COL1A1 mRNA is significantly higher in Ss osteoblasts compared with SS osteoblasts. Furthermore, the ratio of C o l l A l to CollA2 protein secretion was significandy higher in Ss than SS genotyped osteoblasts  50  (2.36:1 in Ss vs. 2:1 in Ss cells). Studies of core bone samples from the femoral head showed that Ss genotyped samples were significantly weaker but had similar inorganic constituents to SS genotyped samples. The authors suggested the increased transcription and translation of C o l l A l due to the binding of the Spl transcription factor resulted in formation of Collal homotrimers which would be structurally weaker in bone matrix than the typical 2 Collal + 1 CollA2 heterotrimers.  Sp1 Binding = increased transcription of C0L1A1 COL1A1 Gene (S Allele)  G (bp 296)  g j  COL1A1 Gene (s allele)  T (bp 296)  COL1A1 and COL1A2 expression and secretion to E C M  Normal 2 X C o l l a l + 1 Col1a2 heterotrimer Collagen fibrils  'Weak' 3 X C o l l a l homotrimer Collagen fibrils  Figure 2.12: COL1A1 S p l polymorphism results i n increase i n C o l l a l expression relative to Colla2 i n osteoblasts therefore V allele produces C o l l a l homotrimers which are structurally weaker than C o l l a l / C o l l a 2 heterotrimers (S allele).  Uitterlinden et al, [157] reported the COL1A1 Spl polymorphism in relation to aBMD and osteoporotic fracture in 1,778 postmenopausal women (aged 66 ± 7 years) in the Netherlands. The 1,194 women with the SS genotype had 2% higher aBMD at both the femoral neck (p = 0.003) and the lumbar spine (p = 0.02) as compared with the 526 women with the Ss genotype; the 58 women with the ss genotype had reductions of 4% at the femoral neck (p = 0.05) and 6% at the lumbar spine (p = 0.005) compared with the women with the Ss genotype. Women with the Ss and ss genotypes 51  were over-represented among the 111 women who had experienced a non-vertebral fracture. Therefore, the relative risk for osteoporotic fracture in this population was 1.5 times greater for those who carried the s allele. Furthermore, evidence from French pre- (aged 31 — 57 years) and postmenopausal (aged 63 ± 12.3 years) women and Spanish men with idiopathic osteoporosis (aged 50.4 ± 10.3 years) supported this association [158-160].  Keen etal, [161] studied COL1A1 Spl polymorphisms in 185 Caucasian women (aged 45 - 64 years) [161]. They observed that lumbar spine aBMD was 0.47 g/cm less and femoral neck aBMD was 0.026 2  g/cm less in Ss and ss genotypes compared to SS genotypes. Urinary pyridoline to creatine fraction 2  was assessed to determine Coll A l protein turnover in the various haplotypes. They found that the urinary pyridinoline/creatine ratio was 44.6 nmoles/nmoles in SS individuals versus 54.1 nmoles/nmoles in Ss and ss genotypes. As pyridinoline concentration is a biochemical marker of total body collagen turnover, Keen et al, suggested that COL1A1 Spl genotype may in fact be a polymorphism that affects total body collagen [161].  Alvarez et al, [162] compared the COL1A1 allele frequency in 20 pre-menopausal women (aged 25 52 years) who were diagnosed with primary osteoporosis with 24 healthy age- and sex- matched controls. The patients with primary osteoporosis had a COL1A1 genotype frequency of SS (50%), Ss (45%), ss (5%) therefore allele frequency was S (72.5%) and s (27.5%). In contrast non-affected controls had a COL1A1 haplotype frequency of SS (87.5%), Ss (12.5%), ss (0%) thus allele frequency was S (93.75%) and s (6.25%). The odds ratio between the s allele and primary osteoporosis in premenopausal women was 5.68.  52  McGuigan etal, [163] studied the association of the COL1A1 Spl polymorphism with other SNPs in the COL1A1 genomic region in 93 patients with osteoporotic compression fracture (aged 70.2 ±5.6 years) with 88 aged match controls. The absence of the Spl and presence of Rsol cut site (in intron 5) polymorphisms were in strong linkage disequilibrium (% = 77.87, p<0.001). A significant association 2  only between the COL1A1 Spl polymorphism and fracture risk was also observed (% = 11.15, df = 1, 2  p = 0.001). Qureshi etal, [164] examined the association between the COL1A1 Spl genotype and femoral neck geometry. They showed that although COL1A1 Spl genotype was not related to hip axis length or femoral neck width, the femoral neck-shaft angle was increased by 2° in carriers of the Ss/ss compared to SS carriers. Previous studies have indicated that increased femoral neck-shaft angle may increase the risk of proximal femur fracture in the event of a sideways fall. Therefore, the COL1A1 Spl binding site polymorphism may contribute to fracture risk independent of aBMD.  Proximal femur fracture rate is dissimilar in racial groups, with Africans and Asians having significandy lower rates of fracture than Caucasians [165]. As such, Beavan etal, [165] examined the prevalence of the COL1A1 Spl polymorphism in populations from China, Gambia and Northern Europe. The frequency of the s allele was 8% in the African population, 0% in the Asian population and approximately 20% in the North European population. The population difference in the prevalence of the s genotype explains between 1.7 and 6.8% of the difference in proximal femur fracture rates between Asian and European countries.  Hustmyer etal, [166] studied the COL1A1 Spl polymorphism in 40 M Z and 56 D Z female twins aged (21 - 49 years). They found no significant association between Spl genotype and aBMD at the lumbar 53  spine or femoral neck. Likewise, the incidence of the C O L I A l V allele was not greater in a group of 54 subjects with idiopathic osteoporosis than the healthy twin group. Furthermore, D Z twins discordant for COL1A1 Spl genotype did not show association with aBMD at the lumbar spine or femoral neck with either COL1A1 allele.  Meta-analysis of the effect of C O L I A l genotype and aBMD, BMI and osteoporotic fracture risk was reported by Mann etal, [156]. They reported that unadjusted lumbar spine aBMD Z-scores were 0.13 (p = 0.0003) and 0.20 (p = 0.004) standard deviations lower in Ss and ss genotypes, respectively compared with SS homozygotes. Unadjusted femoral neck aBMD Z-scores were 0.10 (p = 0.0008) and 0.27 (p = 0.0001) standard deviations lower for Ss and ss genotypes, respectively. Age and BMI adjusted aBMD scores at the femoral neck and lumbar spine reported similar values. Lumbar spine aBMD Z-scores were 0.14 (p = 0.003) and 0.19 (p = 0.02) standard deviations lower in Ss and ss haplotypes versus SS homozygotes. Femoral neck aBMD Z-scores were 0.12 (p = 0.002) and 0.23 (p = 0.003) standard deviations lower in Ss and ss haplotypes compared to SS homozygotes. They also examined yield strength of bone core samples from the femoral head in either Ss or SS haplotypes. They showed that SS genotyped bone core samples were significandy stronger (p = 0.031) compared with Ss genotyped bone core samples.  Mann etal, [156] through meta-analysis examined the fracture risk odds ratio by COL1A1 genotype in patients with osteoporosis. Individuals with Ss genotype were 52% more likely to have an osteoporotic fracture compared with SS genotyped individuals. They concluded that fracture risk was greater by COLI A l genotypes than fracture risk expected based on low aBMD or BMI alone. Therefore, C O L I A l genotype had an effect on bone strength independent of its effect on aBMD.  54  Table 2.8: Summary table of COL1A1 Spl genotype effects.  Study  Population  Genotype Effect  Grant et al, [107]  299 post-menopausal British women  Uitterlinden et al, [157]  1778 post-menopausal women  Alverez et al, [162]  20 premopausal women with primary osteoporosis vs. 24 healthy age matched controls 93 patients with compression fracture vs. 88 age matched controls 153 patients with hip fracture vs. 183 healthy controls s allele frequency is: 20% in Caucasian populations 8% in African populations 0% in Asian populations 40 M Z and 56 D Z Caucasian twins  SS genotype was associated with a significantly greater femoral neck and lumbar spine a B M D than Ss genotyped women. Relative risk of compression fracture was 2.97 is carrying the s allele. Women with the SS genotype had a 2% higher femoral neck and lumbar spine a B M D than women with the Ss genotype. Relative risk of fracture risk is 1.5% if carrying the s allele. Odds ratio of s allele in osteoporosis group was 5.68.  McGuigan et al, [163] Qureshi et al, [164] Beavan et al, [165]  Hustmyer et al, [166] Mann et al, [156]  Meta analysis of C O L 1 A 1 association studies with fracture risk and lumbar spine a B M D and femoral neck a B M D  s allele was significandy associated with fracture risk 2 degree increase in femoral neck-shaft angle in Ss/ss genotyped patients vs. SS genotyped patients S allele explains 1.7-6.8% o f the difference in hip fracture rates between Asian and Caucasian populations.  N o association between C O L 1 A 1 genotype and lumbar spine or femoral neck a B M D s allele is associated with lower lumbar spine and femoral neck a B M D . s allele is associated with 52% greater chance of hip fracture.  ype and Bone Mineral Accrual in Children 2.14.1.2 Association of COL1A1 Genoi Sainz et al, [167] studied the Spl binding-site polymorphism and measurements of the size and the volumetric bone density by Q C T of vertebral bone in 109 healthy, prepubertal girls. The 22 girls with the Ss genotype and 1 girl with the ss genotype had 6.7% and 49.4% lower trabecular bone density, respectively in the lumbar spine than girls with the SS genotype. However, there was no association between vertebral size and COL1A1 genotypes.  2.14.1.3 Interaction between VDR and COL1A1 Genotype in aBMD and Fracture risk V D R is a transcription factor that regulates the expression of COL1A1. Therefore, polymorphisms in these two genes may have an interactive effect on bone metabolism [168, 169]. Uitterlinden etal, [170] 55  studied the interaction of the 3' V D R polymorphisms with the Spl polymorphism of C O L I A l in 1004 postmenopausal Dutch women (aged 55-80). They showed that the V D R genotype baT increased risk of fracture 1.8 times and 2.6 times respectively for heterozygotes and homozygous (p = 0.009). They showed that individuals with a Ss and ss C O L I A l genotype and a baT V D R genotyped had a 2.1 times (baT heterozygotes) and 4.4 times (baT homozygotes) higher odds rado than other V D R haplotypes types. Valimaki etal, 2001 did not show a similar V D R / C O L 1 A 1 interaction on fracture risk, aBMD or markers of bone metabolism in 513 early postmenopausal (<5 years postmenopause) and 172 elderly (>85 years) Finnish subjects [171].  McGuigan et al, [172] examined 460 Irish men and women (aged 22.3 + 1.6 years) for a V D R BsmI and COL1A1 Spl genotype interaction in aBMD of the lumbar spine and proximal femur. The V D R BsmI Bb/bb genotype had a significandy lower lumbar spine (p = 0.044) but not proximal femur aBMD. There was no significant difference in aBMD by COL1A1 Spl genotype. Regression analysis of the genetic and environmental influences on aBMD was also examined. Lumbar spine aBMD in women was predicted by weight ((3 = 0.408, p < 0.0001), V D R BsmI genotype ((3 = 0.205, p = 0.005) and carbohydrate intake (P = 0.143, p = 0.046). Proximal femur aBMD in men was significandy predicted by weight (P = 0.306, p < 0.0001), dietary vitamin D intake (P = 0.238, p = 0.002) and oestrogen receptor genotype (p = 0.201, p =0.008). Lumbar spine aBMD in men was predicted by PA (P = 0.198, p = 0.006), weight (p = 0.261, p < 0.0001), oestrogen receptor genotype (P = 0.180, p = 0.011) and total energy intake (P — 0.174, p —0.014). Proximal femur aBMD in men was predicted by PA (P = 0.221, p = 0.002), weight (p = 0.187, p = 0.009) and alcohol intake (P = 0.161, p = 0.024). The regression models as a whole explained approximately 18% of the peak bone mass variance in women and 14% in men. 56  2.15 Tumour Necrosis Factor Receptor 2 (TNFR2) [fa.k.a. Tumour Necrosis Factor Receptor Superfam Member 1B (TNFRSF1B)] Mendelian Inheritance in M a n (MTM) Ascension #191191, Gene M a p Locus Ip36.3-p36.2 Tumour Necrosis Factor Receptor 2 (TNFR2) is a transmembrane receptor in cells of myeloid origin (i.e. monocyte/monocytes, osteoclasts and myocytes). In these cells T N F R 2 mediates the response to the cytokines Tumour Necrosis Factor-a ( T N F - a ) and Tumour Necrosis Factor-P (TNF-P) in cell proliferation and apoptosis [173]. T N F R 2 has been shown to be highly expressed in osteoclast precursors and arbitrates the effects of T N F - a and T N F - P on osteoclastogenesis [174,175]. The 26 kb, 10 exon T N F R 2 gene encodes a 75 k D protein that is a part of a superfamily of transmembrane receptors that includes the Receptor Activator of NF-K-B ( R A N K ) osteoclast differentiation factor [176].  One candidate loci for regulation of a B M D in humans is in genomic region lp36. Initially a genomewide screen in seven families with low a B M D and a cohort of sib-pairs with low a B M D showed strong linkage in the lp36 region [119,177]. These results were subsequently confirmed in separate families using variance components linkage analysis [118]. The region with the highest L O D score spans 28 c M between lp36.2-36.3. T N F R 2 is within this region suggesting positional and functional candidacy for the role of T N F R 2 polymorphisms in the regulation of a B M D .  57  •Exon I  2 TNFR2.Genfi(43 Kb)  3 4 S.  j  0  ?  8  <3  10  3'UTR  |  620 1 G T T I GCG ,£TTG /CiGAAAGCCTCTGCTGCCATCiG .GTGT rt w C  • "593 598  Figure 2.13: T N F R 2 polymorphisms results in the 10 gene.  th  exon of the 43 Kb T N F R 2  2.15.1 Human Studies 2.15.1.1 TNFR 3' UTR Polymorphisms associated with variance in aBMD  Spotilla etal, [178] examined a microsatellite polymorphism and 3 SNPs within 27 nucleotides of each other in the TNFR2 gene in 2 populations of individuals with low bone mineral density and a reference control group. They showed that the distribution of 3' U T R SNP alleles was significandy different between the low aBMD group and the reference population. Also, carriers of at least one TNFR2 3'UTR allele 1 defined as an Adenine at nucleotide position 593, Guanine at nucleotide position 598 and Tyrosine at nucleotide position 620 had a significandy lower lumbar spine aBMD.  Albagha etal, [179] continued to investigate the relationship between the T N F R 2 3'UTR SNPs and aBMD in a population of 1240 perimenopausal women from the U K (aged 45 - 54 years). There was a significant association between femoral neck aBMD and the presence of an adenine at nucleotide position 593, however there was no significant relationship between lumbar spine aBMD and any of the polymorphisms in the T N F R 2 3'UTR. Since the 3'UTR SNPs were in linkage disequilibrium, haplotype analysis was also investigated for association with aBMD. Following the nomenclature of Spotilla etal, [178], allele 5 (Adenine at nucleotide position, 593, Tyrosine at nucleotide position 598 58  and Cytosine at nucleotide position 620) was associated with a significandy lower femoral neck aBMD then all other haplotypes tested. Regression analysis of the T N F R 2 3'UTR haplotype showed that carrying of the A T C haplotype accounted for 1.2% of the variance in proximal femur aBMD. This is in keeping with the magnitude of effect of similar regression analyses with other aBMD candidate genes such as V D R and the oestrogen receptor gene[105,180].  To measure the plausible effect of the TNFR2 3'UTR polymorphisms on the R N A structure a computer simulation was run using RNAdraw [179,181]. The allele 5 stem-loop structure was significandy altered compared to the most common allele (allele 4: Guanosine at nucleotide position 593, Tyrosine at nucleotide position 598 and Tyrosine at nucleotide position 620). The altered stemloop structure could potentially affect R N A processing, translation or stability. However, functional studies of the T N F R 2 3'UTR polymorphisms on R N A and protein product must be undertaken to confirm this hypothesis.  2.16 Summary of the Literature Kevieiv The literature provides clear support for the positive relationship between childhood P A and augmented bone mass and bone mass accrual. The results from calcium supplementation studies are less clear. The evidence related to the relationship of V D R Bsml genotype with bone mass is varied. However, the B allele in adult populations was associated with an increased incidence of osteoporotic fracture and decreased bone mass. The body of evidence that investigated the association of the V D R Fokl polymorphism and bone suggests that the F allele is associated with high V D R in vitro activity as well as increased aBMD and decreased fracture risk. The's' allele of the COL1A1 Spl polymorphisms was generally associated with an aBMD-independent increase in fracture risk. This is likely due to the increased transcription of COL1A1 in proportion to COL1A2. The increased transcription results in 59  increased formation o f C O L 1 A 1 homotrimers that are weaker than the C o l l al / C o l l a2 heterotrimers in high abundance in normal bone E C M . T N F R 2 is a candidate gene for bone mass in a chromosomal region that has been previously implicated in familial osteoporosis (lp36.2-36.3). The polymorphisms in the 3 ' U T R o f T N F R 2 suggest that the A593-T598-C620 haplotype is associated with lower bone mass at both the proximal femur and lumbar spine in adult women.  60  CHAPTER 3 Study D e s i g n  3.1 Research Aims and Hyotheses There are 7 aims this study wishes to investigate:  i.  T o determine the main effects of P A on B M F L mass  a.  ii.  Hypothesis: High P A is related to high B M F L mass  T o determine the main effects o f candidate gene genotype on B M F L mass  a.  Hypothesis: Differences in V D R  Bsml  genotype will be significant predictors of B M F L  mass in all boys and girls and Caucasian and Asian subgroups.  b.  Hypothesis: Differences in V D R  Fokl genotype  will be significant predictors o f  B M F L mass in all boys and girls and Caucasian and Asian subgroups.  c.  Hypothesis: Differences in C O L 1 A 1 genotype will be significant predictors o f B M F L mass in all boys and girls and Caucasian and Asian subgroups.  d.  Hypothesis: Differences in T N F R 2 A 5 9 3 G genotype will be significant predictors of B M F L mass in all boys and girls and Caucasian and Asian subgroups.  e.  Hypothesis: Differences in T N F R 2 T 5 9 8 G genotype will be significant predictors of B M F L mass in all boys and girls and Caucasian and Asian subgroups.  f.  Hypothesis: Differences in T N F R 2 T 6 2 0 C genotype will be significant predictors o f B M F L mass in all boys and girls and Caucasian and Asian subgroups. 61  g.  Hypothesis: Differences in T N F R 2 haplotype will be significant predictors o f B M F L mass in all boys and girls and Caucasian and Asian subgroups.  T o determine the main effects of P A and dietary calcium on proximal femur, femoral neck and lumbar spine B M C and femoral neck a B M D  a.  Hypothesis: P A is a significant predictor o f proximal femur, femoral neck and lumbar spine bone mass  b.  Hypothesis: Dietary calcium intake is a significant predictor o f proximal femur, femoral neck and lumbar spine bone mass  T o determine the main effects o f candidate gene genotype on proximal femur, femoral neck and lumbar spine bone mass  a.  Hypothesis: Differences in V D R  BsmI genotype will be  significant predictors o f  proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  b.  Hypothesis: Differences in V D R  Fokl genotype will be  significant predictors of  proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  c.  Hypothesis: Differences in C O L 1 A 1 genotype will be significant predictors of proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  62  d.  Hypothesis: Differences in T N F R 2 A 5 9 3 G genotype will be significant predictors o f proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  e.  Hypothesis: Differences in T N F R 2 T 5 9 8 G genotype will be significant predictors o f proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  f.  Hypothesis: Differences in T N F R 2 T 6 2 0 C genotype will be significant predictors o f proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  g.  Hypothesis: Differences in T N F R 2 haplotype will be significant predictors of proximal femur, femoral neck and lumbar spine bone mass in all boys and girls and Caucasian and Asian subgroups.  T o determine the interactions of B M F L mass and P A on proximal femur, femoral neck or lumbar spine bone mass  a.  Hypothesis: T h e interaction between P A and B M F L mass will significandy predict proximal femur, femoral neck and lumbar spine bone mass.  T o determine the interactions of B M F L and candidate gene genotype on proximal femur, femoral neck or lumbar spine bone mass  63  a.  Hypothesis: The interaction between candidate gene genotype or haplotype and B M F L mass will significandy predict proximal femur, femoral neck and lumbar spine bone mass.  vii.  To determine the interactions of candidate gene genotypes and PA on proximal femur, femoral neck or lumbar spine bone mass  a.  Hypothesis: The interaction between PA and candidate gene genotype will significandy predict proximal femur, femoral neck and lumbar spine bone mass.  3.2 Subjects Three-hundred and eighty-three children (195 female, 188 male) were recruited from elementary schools in the Richmond, British Columbia school district, in October of 1999 as a part of the "Healthy Bones Study". This randomized control trial investigated the response of bone to a high impact loading intervention over two years. The present study utilizes data collected during the first year of that study (described in section 3.3). O f the 383 subjects, 327 consented to genetic analysis, a further 7 subjects were excluded from analysis for conditions that could affect bone development or measurement (Down's syndrome, polycystic kidney disease, club foot, cerebral palsy, fracture and immobilization within a year of measurement). Heretofore, the subset population involved in genetic analysis [n = 320; 162 (51%) boys and 158 (49%) girls] will be referred to as the Healthy Bones Genetic Cohort ( H B G Q . Mean age of the H B G C at time of recruitment was 10.34 + 0.65 years. Ethnic distribution of subjects was 47% (n = 150) Caucasian (European ancestry), 37% (n = 117) Asian (Chinese, Japanese, Korean or Philippino), and 16% (n = 53) other (Biracial, Indian/Pakistani, Fijian and African). Two significant ethnic groups were established; Asian and Caucasian. To determine if members of the "other" ethnic group were phenotypically and genotypically members of 64  either the Asian or Caucasian subgroup, analysis by ethnic group was tested for significant differences in (a) bone mass and (b) genotype frequencies (data not shown). Members of the Indian/Pakistani sub-group had a similar distribution of all genotype alleles and bone parameter as the Caucasian group, but were significandy different from the Asian sub-population. Therefore, members of the Indian/Pakistani sub-population were added to the Caucasian ethnic group. Final group size for the Caucasian (European or Indian/Pakistani Ancestry) sub-population was 165 and the Asian (Chinese, Japanese, Korean or Philippino Ancestry) sub-group was 115. All parents or legal guardians of subjects signed informed consent Appendix 1. This study was approved for human investigation by the University of British Columbia Clinical and Behavioural Science Ethics Board.  3.3 Instruments and Procedures 3.3.1 Anthropometry  Height was measured twice to the nearest 0.1 cm using a wall-mounted stadiometer and headboard. Stretch stature for sitting and standing height were measured using standard methods, by gently extending upward traction from the base of the mastoid processes. Leg length equalled the difference between standing and sitting heights. Weight was measured twice to the nearest 0.1 kg with a digital scale. Means were used for analysis. If the two measurements differed more than 0.4 cm (for height measurements) or 0.1 kg (for weight measurements), a third measure was taken and the average of the two closest values was used. B M F L mass (g) and fat mass (g) were measured from total body D X A scans.  3.3.2 Bone Mineral Assessment  The total body, lumbar spine and proximal femur and its femoral neck subregion were examined with a Hologic Q D R 4500W D X A (Hologic, Inc. Madison, WT). Bone mineral content (BMC, grams) at the lumbar spine, proximal femur and femoral neck and areal bone mineral density (aBMD, 65  grams/cm ) at the femoral neck are reported (bone parameters). Bone densitometry is a safe, painless, 2  extremely low radiation procedure. The total radiation dose per examination is equivalent to the background radiation one would receive making a one-way flight from Vancouver to Halifax on a commercial airline. The procedure required that the child lay on the padded examination table for positioning and measurement. Total time of the procedure is approximately 15 minutes. Measurements were made by one of two trained and qualified technicians and all scans were analyzed by a single researcher using standardized procedures as outlined in the Hologic Users Guide [182]. A pilot precision study was previously conducted with the Q D R 4500W D X A . Lumbar spine, proximal femur and femoral neck B M C and aBMD were measured three times within one hour in 17 healthy young adults [85]. The coefficients of variation for B M C and aBMD for the lumbar spine were less than 0.8%. The coefficients of variation for B M C at the proximal femur and femoral neck varied from 1.2% to 3.5%, aBMD varied from 0.5% to 1.2% at these sites. A spine and anthromorphic phantom was scanned daily to maintain quality assurance of the Q D R 4500W D X A .  333 Physical Activity and Dietary Calcium Intake Physical activity was measured by a 7-day physical activity recall questionnaire (PAQ - C) designed to discriminate between children with low, average and high general physical activity (appendix 1). The PAQ-C has been previously validated against other self-administered physical activity questionnaires, motion-sensors and interview assisted recall (r = 0.39 — 0.63) [183]. The average PA score, heretofore known as PA, was derived as the mean of 6 assessments (during the fall, winter and spring over two years between October 1999 and June 2001). The correlation coefficient over the 6 assessments was a = 0.88 for boys and 0.90 for girls.  66  Dietary calcium intake for subjects was approximated from a food frequency questionnaire (FFQ) that has been previously validated in mixed ethnicity high school children [184] (appendix 1). Dietary calcium intake is expressed in milligram per day (mg/d).  3.3.4 Maturity Assessment Pubertal stage maturity was self- or parental assessed using Tanner staging [185]. This method comprises a series of line drawings of breast and pubic hair, for girls, and pubic hair, for boys, depicting the 5 stages of pubertal development (appendix 1). The child chooses the drawing he/she feels most closely resembles his/her physical appearance. Generally, self-assessed Tanner stage is consistent with physician rating [186].  3.3.5 Genotype Analysis Genotyping of all consenting patients was done for V D R Bsml, V D R Fokl, C O L I A l Spl and TNFR2 3'UTR polymorphisms following previous protocols [103,143, 179,187].  3.3.5.1 Extraction and Purification of Genomic DNA Buccal (cheek) cell samples from the H B G C subjects were taken in June 2000 using the Epicentre MasterAmp™ Buccal Swab D N A Extraction Kit (Madison, WT) following the manufactures protocol. Each subject was sampled with two buccal swab brushes, one on either cheek. Between July and October 2000 a single investigator used the Epicentre MasterAmp™ D N A extraction solution to separate high molecular weight genomic D N A from the buccal swabs, following the manufacturers protocol. In a spectrophotometric assay ( A / A ) of sample purity ( D N A and R N A vs. protein 260  280  debris) the samples were found to be on average 54% pure nucleic acid. After extraction each sample had approximately 0.5 ml of solution. All D N A samples were diluted 1:10 with milliQ distilled water to a working concentration of approximately 0.5 pg/pi. 67  3.3.5.2 Polymerase Chain Reaction Amplification, Restriction Enzyme Analysis and Electrophoretic Sepera VDR Fokl and BsmI Polymorphisms The V D R Fokl and V D R BsmI polymorphism were analyzed by standard RFLP analysis following Morrison et al, 1994 and Arai etal, 1997 [104,143]. Both the V D R regions of interest were amplified by PCR using the following Qiagen (Valencia, CA) reagents per 25 pi reaction: Table 3.1: PCR reaction reagents, concentrations and volumes in V D R VDR BsmI reactions.  Reagent (Final Concentration) milliQ distilled water  Volume  Qiagen P C R Buffer (10X)  2.5 ul  d N T P mix (10 m M of each dNTP)  2ul  Qiagen Q-solution (5X)  5ul  Fold  and  11.875 |al  Qiagen Taq (5 U / u l )  0.125 ul  Forward Primer (10 uM)  1.25 (il  Reverse Primer (10 uM) Sample D N A (1:10 dilution) Total  1.25 ul 1 ul 25 ul  Forward and reverse oligonucleotide primers for V D R Fokl and V D R BsmI PCR amplification were ordered from Sigma-Genosys (The Woodlands, TX). Primers' nucleotide sequence are listed (Table 3.2). Table 3.2: Oligoucleotide forward and reverse primer sequence for the PCR amplification of the V D R Fold and V D R BsmI regions of interest. Designation Oligonucleotide Sequence V D R Fokl Forward Primer 5 ' A G C T G G C C C T G G C A C T G A C T C T G C T C T 3 ' V D R Fokl Reverse Primer 5'ATGGAAACACCTTGCTTCTTCTCCCTC3' V D R BsmI Forward Primer 5 ' C A A C C A A G A C T A C A A G T A C C G C G T C A G T G A 3 ' V D R BsmI Reverse Primer 5 ' A A C C A G C G G G A A G A G G T C A A G G G 3 '  PCR cycle reactions were performed on a Peltier thermal cycler, The D N A Engine DYAD™ (MJ Research™, Inc.; Waltham, MA). PCR temperature cycles were as follows: 68  Table 3.3: Thermal Cycler conditions for V D R Fold and V D R Reaction Step V D R Fold Primary Denaturation Denaturation Primer Annealing Extension Final Extension VDR  Temperature (°C.)  Time  Cycles  94°C 94°C 60°C 72°C 72°C  4 Mins. 50 Sees. 1 Min. 1.5 Mins. 10 Mins.  1 33 33 33 1  94°C 94°C 55°C 72°C 72°C  4 Mins. 50 Sees. 1 Min. 1.5 Mins. 10 Mins.  1 33 33 33 1  Bsml  amplification.  Bsml  Primary Denaturation Denaturation Primer Annealing Extension Final Extension  The P C R reactions o f for the V D R  Fokl  and V D R  Bsml  regions of interest produced 256 bp and 870  bp amplicons, respectively. Restriction fragment length polymorphism analysis of the amplicons was done using restriction enzymes from N e w England Biolabs (Mississauga, Ont.). Restriction enzyme digestion was done in 20 pi volumes overnight to ensure complete digestion. V D R incubated at 3 7 ° C while V D R  Bsml  Table 3.4: Reagents for V D R Fold and V D R  Reagent (Stock concentration) V D R Fold V D R Fokl PCR product  Volume  NEBuffer4(10X)  2ul  N E B Fokl Enzyme (5 U/ul) milliQ distilled water  0.25 ul  Total  20 ul  VDR  Bsml  restriction enzyme digestion.  10 ul  7.75 ul  Bsml Bsml  PCR product  10 ul  NEBuffer2(10X)  2ul  N E B Bsml Enzyme (5 U / u l ) milliQ distilled water  0.25 ul  Total  20 ul  digestion was  digestion was incubated at 6 5 ° C . Restriction digestion was done  with the following reagents:  VDR  Fokl  7.75 ul  69  Ten microlitres of either restriction enzyme digest product was mixed with 1 pi of Orange-G loading dye and electrophoresed in 1.5% agarose/TBE (Tris-Borate-EDTA, p H 7.5) gel stained with ethidium bromide and visualized with a U V B transilluminator. Molecular weight of bands was verified by comparison with a (j)X174/HincII digest low molecular weight marker run in parallel to the samples. V D R Fokl digestion product produced bands at 265 bp for homozygous uncut samples, bands at 196 bp and 69 bp for samples homozygous for the cut site and bands at 265 bp, 196 bp and 69 bp for heterozygotes (Figure 3.1). V D R Bsml digestion product produced a band at 870 bp for homozygous uncut samples, 640 bp and 230 bp bands for homozygous samples with the cut site and bands at 870 bp, 640 bp and 230 bp for heterozygous samples (Figure 3.2). Genotyping was repeated in approximately 20% of the sample to ensure accurate, repeatable genotyping. All genotypes had concordant results.  Figure  3.1:  Sample  electrophoresis of Bsml H O  bp  * * * 1M§  *****  mm • *  640  H o m o z y g o u s b b genotypes show a  single  band  at  640 b p ,  heterozygous B b genotypes show a band at 870 b p and 640 b p and  bp  m bb  digest  of V D R Bsml polymorphisms.  homozygous B B genotype show a band at 870 b p .  Bb BB  70  2 65 bp  Figure 3.2: Sample electrophoresis of Fold digest of V D R Fold polymorphisms. Homozygous ff genotypes show a single band at 196 bp, heterozygous F f genotypes show a band at 265 bp and 196 bp and homozygous BB genotype show a band at 265 bp.  196 bp  f f. Ff FF 3.3.5.3 C0L1A1 Sp1 Polymorphism Detection using the TaqMan™ System  Analysis of the COL1A1 Spl polymophism with the TaqMan™ System from P E Biosystems (Foster City, CA) has been described elsewhere [187]. The basic chemistry of the TaqMan™ protocol is illustrated in Figure 3.3. Briefly, 2 fluorogenic probes with oligonucleotide sequences complimentary to the sequence of interest were designed to anneal to either allele specifically under temperatures that are precisely matched to the PCR primer annealing temperatures. The fluoreogenic probes are labelled with allele specific 5'-fluorescent reporter and a 3'-quencher dye (TAMRA™). The 3'-TAMRA™ quencher inhibits priming from the 3' end of the probe and due to its close conformation to the 5'allele specific probe it reduces latent fluorescence from the intact complex oligonucleotide probe. During the PCR reaction the 5'—»3' exonuclease activity of the Taq D N A polymerase releases the 5'allele specific fluorescent dye from the oligonucleotide thereby causing an increase in allele dependent fluorescence.  71  Polymerization  Figure 3.3: Basic chemistry of the TaqMan™ System. The fluorescent emission of the probe specific fluorescent marker (R) is reduced by the proximity of the 3'-TAMRA™ quencher (Q). 5'—>3' exonucleose acdvity of the Taq DNA polymerase releases the 5' fluorescent marker from the oligonucleotide probe and causes an allele specific increase in fluorescence.  Forward , Primer  -y  -3" Reverse Primer  -y  Strand D i spl a cern ent  3'-  y-  Cleavage  r• y  Polymerization Completed IR  5'. 3'3'-4www.uld.uni-freiburg.de/core-facility/ taqman/taqindex.html  Allelic discrirnination of the COL1A1 Spl genotype using the TaqMan™ system was performed with probes and primers listed in Table 3.5. The F A M (6-carboxy-fluorescein) probe was designed to detect the S allele (no consensus sequence for Spl binding) while the T E T probe (6-carboxy-4,7,2',7'tetrachloro-fluoroscein) was designed to detect the s allele (consensus sequence for Spl binding).  72  Table 3.5: T a q M a n ™ primers and probes for detection of the COL1A1 Spl polymorphism. Nucleotide of the allele specific probe is underlined. Designation Forward Primer Reverse Primer F A M Probe (S allele) T E T Probe (s allele)  Oligonucleotide Sequence 5'GATGTCTAGGTGCTGGAGGTTAGGG3' 5'CCTTCTCTCCTCCCTCGCC3' 5'GGTCCACTTACCCCCGCCC3' 5'GGTCCACTTAACCCCGCCC3'  The PCR reagents per 25 ul reaction are listed in in Table 3.6. TaqMan™ PCR temperature conditions performed using a Peltier thermal cycler, The D N A Engine DYAD™ (MJ Research™, Inc.; Waltham, MA) are provided (Table 3.7). Table 3.6: List of reagents per 25 ul reaction using the TaqMan™ system to detect COL1A1 Spl polymorphisms. Reagent (Final Concentration) TaqMan™ Mastermix (PE Biosystems, Foster City, CA)  Volume  Forward Primer (50uM)  0.5 ul  Reverse Primer (50 |aM) F A M Probe (50 nM)  0.5 ul  T E T Probe (100 nM)  0.167 ul  milliQ distilled water  10.25 ul  Template D N A (1:10 dilution stock)  1 ul  Total  25 ul  0.083 ul  Table 3.7: T a q M a n ™ PCR polymorphism detection. Reaction Step Primary Denaturation Denaturation Annealing and Extension  12.5 ul  reaction  conditions  Temperature (°C) 95° C 95° C 60° C  Time 10 Mins. 15 Sees. 1 Min.  for  COL1A1  Spl  Cycles 1 40 40  At the end point of the PCR an automated laser detection system (ABI PRISM 7200 Sequence Detector, P E Biosystems, Foster City, CA) measured the fluorescent spectrum of each reaction vessel. The emission spectrum for the F A M dye is within the blue spectra while the T E T probe is within the green spectra. A linear combination of the fluorescent emissions is collected and calibrated versus a measure of the background fluorescence in designated no D N A template control reactions. The 73  contribution o f each component dye to the observed spectrum in the reaction vessel is plotted on a scatter diagram with the F A M and T E T as the ordinate and abscissa axes. A relative increase in either F A M or T E T fluorescence indicates homozygosity for the alleles each dye represents whereas an increase in both dye fluorescence indicates heterozygosity for the sample. Genotyping was repeated in approximately 20% o f the sample to insure accurate allele determination. Greater than 99% of the repeated genotypes were accurate upon retesting.  3.3.5.4 Automated Sequencing of the TNFR2 3 'UTRfor the Determination of the 3 Polymorphisms withi Region The determination o f the 3 SNPs in the 3 ' U T R of T N F R 2 using automated sequencing o f P C R product has been previously described [179]. Initially, a 192 bp P C R amplicon flanking the 3 ' U T R polymorphisms was amplified using the reagents and oligonucleotide primers listed in Tables 3.8 and 3.9. A Peltier thermal cycler, T h e D N A Engine D Y A D ™ (MJ R e s e a r c h ™ , Inc.; Waltham, M A ) was used for P C R amplification in the conditions (Table 3.10).  Table 3.8: Primer sequences for the PCR amplification of the 3'UTR of T N F R 2 Designation Forward Primer Reverse Primer  Oligonucleotide Sequence 5' A G G A C T C T G A G G C T C T T T C T 3 ' 5TCACAGAGAGTCAGGGACTT3'  Table 3.9: Reagents, concentrations and volumes used in the PCR amplification of the polymorphic region of the 3'UTR of TNFR2. Reagent (Starting concentration) milliQ distilled water  Volume  Qiagen P C R Buffer (1 OX)  2.5 ul  dNTP mix  2ul  Q-Solurion  5 ul  Qiagen Taq D N A Polymerase (5 U/ul)  0.125 ul  Forward Primer (10 U.M)  1 ul  Reverse Primer (10 uM) Template D N A (1:10 working stock)  lui  Total  25 ul  12.375 ul  1 ul 74  Table 3.10: P C R thermal cycler conditions polymorphic region in the 3'UTR of TNFR2. Step Primary Denaturation Denaturation Annealing Extension Final Extension  Temperature (°C) 94° C 94° C 55° C 72° C 72° C  for the amplification of the  Time 4 Mins. 50 Sees. 1 Min. 1.5 Min. 10 Mins.  Cycles 1 33 33 33 1  After P C R amplification 6 pi o f the P C R product was mixed with 1 pi o f E x o / S A P - I T ™ (Exonuclease I/Shrimp Alkaline Phosphatase, U S B Corp., Cleaveland, O H ) to remove excess d N T P s and unused primers that would interfere with the sequencing reaction chemistry (Figure 3.4). T h e 7 p i E x o / S A P IT and P C R product mixture was then incubated at 3 7 ° C for 30 minutes followed by incubation at 80° C for 30 minutes.  Figure 3.4: PCR amplification followed by Exo/SAP-IT degradation of excess nucleotides and unused primers prior to automated sequencing.  TARGET DNA  PRIMER B  PRIMER A  4dNTPs  ONA polyrnoraso. thorrnnl cyciin«i  ^nucleotides  MWfPm  Exonuclonso I Alknlino P h o s p h a t o a s a  Doublo-sU.indod PCS ftaMJuet  a f C , IS min than 8» C 13 min arc, 15 min than OtTC, 14 min >  I  DQuMo-.stF.uitlad PCR Product N u c l e o s i d e s , Pi  So.quortc.ing  http://www.apbiotech.com/product/publication/lsn/19/ pages/pl 5-full.htm+exo-sap&hl=en&ie=UTF-8  75  After incubation 7 ul o f milliQ distilled water was added to the purified P C R product thereby diluting it 1:2. T h e sequencing reaction consisted of 7 ul of the purified, diluted P C R product, 2 pi o f D Y E n a m i c ™ E T Terminator K i t with Thermo S e q u e n a s e ™ II D N A Polymerase (Amersham Bioscience, Uppsala, Sweden) and 1 pi o f T N F R 2 Reverse Primer (Table 3.8) at 2.5 u M . The conditions o f the sequencing reaction on the Peltier thermal cycler, T h e D N A Engine D Y A D ™ (MJ R e s e a r c h ™ , Inc.; Waltham, M A ) are listed (Table 3.11). Table 3.11: Thermal cycler conditions for the sequencing reaction of the polymorphic region of the 3'UTR of T N F R 2 .  Step Denaturation Annealing Extension  Temperature (°C) 95° C 55° C 60° C  Time 20 Sees. 15 Sees. 1 Min.  Cycles 25 25 25  The sequencing reaction product was then precipitated following a standard Ethanol/Sodium Acetate (NaAc) protocol. Briefly, in a fresh tube, 11.5 pi o f milliQ distilled water, 8 pi o f 3 M N a A c and 8.5 pi of the sequencing reaction was mixed by pipetting. Eighty microlitres o f 95% ethanol was added and the mixture was vortexed. This reaction mixture incubated in the dark at - 2 0 ° C for 30 minutes. After incubation the mixture was centrifuged at 15,000 g, 4° C for 20 minutes. T h e supernatant was aspirated and the pellet was washed with 70% ethanol, centrifuged at 15,000 g and aspirated 3 times. Prior to sample loading on the sequencer the sample was resuspended using the sequencing buffer (Amersham Biosciences, Uppsala, Sweden).  Automated sequencing was run on a M e g a B A C E 1000 (Amersham Biosciences, Uppsala, Sweden) following the manufactures protocol [188] with minor variations as described below. Prior to Linear Polyacrylamide (LPA) matrix injection milliQ distilled water was preinjected into the capillaries at 3 k V 76  for 40 seconds. This step was recommended by the manufacturer for small fragment sequencing to ensure clean capillaries devoid of LPA matrix from previous sequencing runs. Capillary electrophoresis and laser detection of the fluorescent terminators required approximately 60 minutes at 12 kV in this protocol.  The fluorescent emissions were analyzed with the manufacturer's base caller, Cimarron 2.1 Slim Phredify (Amersham Biosciences, Uppsala, Sweden). Proper base calling was verified and polymorphic bases were analyzed by examining the chromatograms using Lasergene v5.0, SeqMan software (Madison, WT).  Haplotype analysis of the TNFR2 A593G-T598G genotypes was conducted. Labelling of the TNFR2 A593G-T598G haplotypes followed the previous classification scheme of Fernendez-Real et al, [189] with minor modification. As such, TNFR2 A593G-T598G haplotypes were labelled thus: Allele 1 (Al) = G593-T598, Allele 2 (A2) = A593-T598, Allele 3 (A3) = A593-G598, Allele 4 (A4) = G593-G598.  3.3.6 Statistical A nalysis All analysis was performed using SPSS 11.0.1 (SPSS Inc, Chicago, ILS) or SYSTAT 10.0 (Systat Software Inc., Richmond, CA). Descriptive analysis was done using the "Descriptives" Subcommand from SPSS 11.0.1. Significant group differences in genotype frequency were determined using the Non-parametric, yf command from SPSS 11.0.1. Significant differences in A N C O V A modeling were determined using the "General Linear Model" subcommand from SPSS 11.0.1. Multivariate outliers were examined using the "Outlier" subcommand from SYSTAT 10.0. Generally, statistical significance was set to p < 0.05, however exact p-values are reported through out for a transparent examination of the significance of the results.  77  T o determine the independent effects o f P A or candidate genes on either B M F L or bone mass A N C O V A modeling was used. The following linear modeling equations were used to determine the main effects o f P A , dietary calcium or candidate genes on B M F L or bone mass:  B M F L y J = Tanner Stage  var  (Pubic Hair - boys, Breast - girls) [covariate] + Age [covariate] + Total var  Height^ [covariate] + Total F a t ^ [covariate] + P A ^ or Candidate gene  Bone Mass Length  var  Var  = Tanner Stage  var  factor  (independent effect)  (Pubic Hair - boys, Breast - girls) [covariate] + Age [covariate] + Leg var  or Sitting height^,. + Total Fat  Candidate gene  factor  var  [covariate] + B M F L y .  ir  [covariate] + P A  v a r  or C a l c i u m ^ or  (independent effect)  Interactions o f B M F L by P A , B M F L by candidate genes or P A by candidate genes on bone mass were investigated in a similar manner. T h e following equation is an example o f the interaction equation used to determine the interaction o f P A and C O L 1 A 1 genotype. Similar linear equation models were used to determine the interaction and main effects o f the aforementioned factors.  Bone Mass Length  var  Var  = Tanner Stage  (Pubic Hair - boys, Breast - girls) [covariate] + Age [covariate] + Leg var  or Sitting height^,. + Total Fat , [covariate]  C O L 1 A 1 genotype  1  var  v ar  factor  (main effect) + P A  v a r  + BMFLv  a r  [covariate] + P A  * C O L l A l genotype  Var = variance 78  factor  v a r  (main effect) +  (interaction effect)  CHAPTER 4 Results 4.1 Descriptive Results 4.1.1 Anthropometric Descriptive Results Anthropometry outcomes for the total group and for boys and girls are listed (Table 4.1). Table 4.1: Descriptive results (mean ± SD) for height (cm), sitting height (cm), leg length (cm), weight (kg), total fat mass (g), B M F L (g) in the total group and for boys and girls.  Height (cm) Sitting Height (cm) Leg Length (cm) Weight (kg) Total Fat Mass (g) B M F L (g)  Total Group 142.2 ± 7.4 75.2 ± 3.9 67.0 ± 4.1 37.0 ± 9.5 9601 ± 5449 26156 ± 4 5 9 3  Boys 142.4 ± 7.0 75.3 ± 3.6 67.0 ± 4.0 37.4 ± 10.1 9264 ± 5930 26900 ± 4640  Girls 142.0 ± 7.7 75.1 ± 4.2 66.9 ± 4.3 36.5 ± 8.8 9948 ± 4902 25394 ± 4430  4.1.2 Age and Maturity Descriptive Results Table 4.2 lists the age and maturity descriptive results of the group and the sub-groups of boys and girls. Table 4.2: Descriptive results for age (mean ± SD) and Tanner Breast and Pubic Hair Stages (frequency and percentage).  Age (years) Tanner Stage  Breast  Tanner Pubic Hair Stage  Total Group 10.33 ± 0.65 N/A  Boys 10.39 ± 0.65 N/A  1: 235 (73.7%) 2: 38 (11.9%) 3: 8 (2.5%) 4: 1 (0.3%) 5: 1 (0.3%)  1: 142 (87.7%) 2: 18 (11.1%) 3:. 2 (1.2%) 4: 0 (0%) 5: 0 (0%)  Girls 10.27 ± 0.64 1: 63 (39.9%) 2: 81 (51.3%) 3: 12 (7.6%) 4: 0 (0%) 5:2(1.3%) 1: 93 (59.2%) 2: 20 (12.7%) 3: 6 (3.8%) 4: 1 (0.6%) 5: 1 (0.6%)  79  4.1.3 Polymorphism Frequencies The frequencies of the polymorphisms in this study ( C O L 1 A 1 S p l , V D R  BsmI, VDR Fokl, T N F R 2  A 5 9 3 G , T N F R 2 T 5 9 8 G and T N F R 2 T620C) for the whole group and boys and girls are listed (Table 4.3). Table 4.3: Genotype frequencies for the total group and for Caucasian and Asian subgroups. Significant differences in allele distribution by ethnicity are also provided  COL1A1 Spl  VDR  BsmI  VDR Fold  TNFR2 AS93G  TNFR2 T598G  T N F R 2 T620C  TNFR2 Haplotype  Total Group  Caucasian  Asian  Significant Differences by Ethnicity  SS: 265 (83.3%) Ss/ss: 53 (16.7%)  SS: 120 (73.6%) Ss/ss: 43 (26.4%)  X (1) = 35.89, p < 0.001  BB: 30 (9.6%) Bb: 108 (34.4%) bb: 176 (56.1%) FF: 147 (46.4%) Ff: 117 (36.9%) ff: 53 (16.7%) A / A : 97 (30.9%) A / G : 100(31.8%) G / G : 117(37.3%) T / T : 293 (93.3%) T / G : 20 (6.4%) G / G : 1 (0.3%)  BB: 28 (17.3%) Bb: 76 (46.9%) bb: 58 (35.8%) FF: 78 (48.1%) Ff: 58 (35.8%) ff: 26 (16.0%) A / A : 43 (26.7%) A / G : 51 (31.7%) G / G : 67 (47.6%) T / T : 143 (88.8%) T / G : 17 (10.5%) G / G : 1 (0.6%)  T / T : 78 (24.8%) T / C : 127 (40.4%) C / C : 109 (34.7%) A l / A l : 113 (36.2%) A 1 / A 2 : 89 (29.8%) Al/A4:4(1.3%) A 2 / A 2 : 85 (28.4%) A 2 / A 3 : 8 (2.7%)  T / T : 59 (36.6%) T / C : 69 (42.9%) C / C : 33 (20.5%) A l / A l : 63 (40.6%) A 1 / A 2 : 45 (29.0%) Al/A4:4(2.6%) A 2 / A 2 : 35 (22.6%) A 2 / A 3 : 8 (5.2%)  SS: 115 (100%) Ss/ss: 0 (0%) Not Polymorphic BB: 0 (0%) Bb: 12(10.7%) bb: 100 (89.3%) FF: 46 (40%) Ff: 46 (40%) ff: 23 (20%) A / A : 40 (35.4%) A / G : 34 (30.1%) G / G : 39 (34.5%) T / T : 113 (100%) T / G : 0 (0%) G / G : 0 (0%) Not Polymorphic T / T : 8 (7.1%) T / C : 41 (36.3%) C / C : 64 (56.6%) A l / A l : 39 (34.5%) A 1 / A 2 : 34 (30.1%) Al/A4:0(0%) A 2 / A 2 : 40 (35.4%) A 2 / A 3 : 0 (0%)  2  X (2) = 79.21, p < 0.001 2  X (2) = 1.91, p = 0.385 2  X (2) = 2.575, p = 0.276 2  X (2) = 13.52, p < 0.001 2  X (2) = 48.95, p < 0.001 2  X (4) = 13.26, p - 0.010 2  4.1.4 Bone Parameter Descriptive Results Descriptive results of the measured bone parameters are provided for the whole group and for boys and girls (Table 4.4).  80  Table 4.4. Descriptive results (mean ± SD) of the measured bone parameters.  Proximal Femur B M C (g) Femoral Neck B M C (g) Femoral Neck aBMD (g/cm ) Lumbar Spine B M C (g) 2  Total Group 16.3 ± 3.8 2.8 ± 0.5 0.7 ± 0.1 25.2 ± 5.8  Boys 16.7 ± 3.9 2.9 ± 0.5 0.7 ± 0.1 24.7 ± 4.7  Girls 15.9 ± 3.6 2.6 ± 0.4 0.6 ± 0.1 25.8 ± 6.7  4.1.5 Bivariate Correlations of Variables The subsequent tables are the bivariate correlation matrices o f selected independent variables. T h e multivariate regression models to follow were developed based on these relationships and the biological relevance o f the correlations. Table 4.5: Correlation matrix for Anthropometric variables in boys, Peason-R (2-tailed significance). Variables listed are Age (years), height (cm), weight (kg)Total Fat (g), B M F L (g) and PA. 2  Age  Age  Height  1  0.391 (<0.001) 1  Height Total Weight Total Fat BMFL  Total Weight 0.243 (0.002) 0.706 (<0.001) 1  Total Fat 0.167 (<0.05) 0.521 (<0.001) 0.950 (<0.001) 1  BMFL  PA  0.301 (<0.001) 0.834 (<0.001) 0.919 (<0.001) 0.752 (<0.001) 1  0.156 (0.051) 0.150 (0.060) 0.080 (0.320) 0.008 (0.923) 0.154 (0.054) 1  Avg. PA Score  Table 4.6: Correlation matrix for Anthropometric variables in girls, Peason-R (2-tailed significance). Variables listed are Age (years), height (cm), weight (kg)Total Fat (g), B M F L (g) and PA. 2  Age Height Total  Age  Height  1  0.438 (<0.001) 1  Total Weight 0.251 (<0.001) 0.712 (<0.001) 1  Total Fat 0.098 (0.225) 0.467 (<0.001) 0.915  BMFL  PA  0.364 (<0.001) 0.836 (<0.001) 0.894  0.139 (0.364) -0.095 (0.241) -0.024  81  Weight Total Fat BMFL  (<0.001) 1  (<0.001) 0.639 (<0.001) 1  PA  (0.766) -0.017 (0.836) -0.025 (0.756) 1  Table 4.7: Bivariate correlations, Pearson-R (2-tailed significance) between anthropomorphic variables and bone parameters. Correlations for Age, Height, Total Weight, Total Fat, B M F L and P A and proximal femur, femoral neck and lumbar spine B M C , femoral neck a B M D in boys. 2  Proximal Femur BMC Femoral Neck BMC Femoral Neck aBMD Lumbar Spine BMC  Age  Height  Total Fat 0.505 (<0.001)  BMFL  PA  0.728 (<0.001)  Total Weight 0.676 (<0.001)  0.375 (<0.001)  0.785 (<0.001)  0.235 (0.003)  0.300 (<0.001)  0.617 (<0.001)  0.569 (<0.001)  0.413 (<0.001)  0.671 (<0.001)  0.248 (0.002)  0.240 (0.003)  0.442 (<0.001)  0.477 (<0.001)  0.361 (<0.001)  0.541 (<0.001)  0.227 (0.005)  0.343 (<0.001)  0.666 (<0.001)  0.538 (<0.001)  0.344 (<0.001)  0.697 (<0.001)  0.143 (0.073)  Table 4.8: Bivariate correlations, Pearson-R (2-tailed significance) between anthropomorphic variables and bone parameters. Correlations for Age, Height, Total Weight, Total Fat, B M F L and P A and proximal femur, femoral neck and lumbar spine B M C , femoral neck a B M D in girls. 2  Proximal Femur BMC Femoral Neck BMC Femoral Neck aBMD Lumbar Spine BMC  Age  Height  Total Fat 0.539 (<0.001)  BMFL  PA  0.730 (<0.001)  Total Weight 0.747 (<0.001)  0.308 (<0.001)  0.817 (<0.001)  0.020 (0.881)  0.237 (0.002)  0.645 (<0.001)  0.727 (<0.001)  0.568 (<0.001)  0.747 (<0.001)  0.069 (0.400)  0.078 (0.335)  0.394 (<0.001)  0.542 (<0.001)  0.441 (<0.001)  0.533 (<0.001)  0.122 (0.133)  0.367 (<0.001)  0.733 (<0.001)  0.631 (<0.001)  0.370 (<0.001)  0.785 (<0.001)  -0.029 (0.722)  82  4.2 General Unear Modeling of Bone Mineral-free Lean Mass and Bone Parameters to Accountfor Covariates and Determine Effects of Physical Activity, Calcium Intake, Polymorphisms and Interactions I have taken a stepwise approach to; (1) investigate the contribution o f P A and candidate genotypes to total body B M F L and (2) investigate the contribution o f B M F L , candidate gene genotypes and P A to the prediction o f bone mass. Finally, I will evaluate the contribution o f the following interaction terms to the bone parameters in General Linear Models: (a) B M F L by P A , (b) B M F L by candidate gene genotype and (c) candidate gene genotype by P A interactions. A s genotypes were discordant by ethnicity the contribution o f genotype to B M F L mass or bone parameters will also be examined per ethnicity. Genotype by B M F L mass and genotype by P A interaction effects on bone mass were also examined by ethnicity and for the whole study population.  Multivariate outliers were examined with Cook's distance and leverage for all data points in multivariate space. Following Lunneborg 1994's criterion, cases with leverage greater than 3(k/n), where k = the number of predictors (dependent variables and covariates) and n is the number of cases by sex, were considered outliers [190]. Cases with Cook's distances much greater than 1.00 were considered to have high influence. High influence and/or multivariate outliers were eliminated from the analysis casewise per model.  4.2.1 Covariate Modeling of Bone Mineral-free Lean Mass Bone Mineral-free Lean Mass must be controlled for size and maturity. Clearly, larger frames and more mature children will have a greater amount of muscle mass and therefore higher B M F L values. Height and total body fat were included as covariates just as were age and maturity [Tanner breast (girls) and Tanner pubic hair (boys) stage].  83  Table 4.9: Variance accounted (R ) for B M F L by the covariates: total fat mass, height, age and Tanner pubic hair or Tanner breast stage in boys and girls; significance of each covariate and effect of each predictor within the General Linear Model. 2  Boys Significance Height (cm) Total Fat (g) Age (years) Tanner Breast Tanner Pubic Hair R  p < 0.001 p < 0.001 p = 0.591  2  partial-!) 0.592 0.461 0.002  2  -  -  p = 0.013 0.841  0.039  Girls Significance p p p p  < 0.001 < 0.001 = 0.742 = 0.012  parti al-T) 0.510 0.242 0.001 0.042  -  2  -  0.788  4.2.2 Effects of Physical Activity on Bone Mineral-free Lean Mass after accountingfor Covariates In this section I investigated the unique contribution o f P A on B M F L with size, age and maturity controlled. Furthermore, the statistical significance of the variable is also listed. F o r P A effects that are statistically significant I have provided a scatter plot o f the standardized residuals of B M F L after accounting for the covariates versus P A . Dietary calcium intake was not investigated, as there was no evidence to suggest a relationship between B M F L and dietary calcium intake in the present study. Table 4.10: Results of the General Linear Model examining the contribution of PA to B M F L . The effect size (partial-ri ) of P A and the significance of that variable are also listed. 2  PA  Boys Significance  partial-r|  p = 0.039  0.028  •  1.5  2.0  o °  ° n•  2.5  £  3.0  2  Girls Significance  partial-T)  p = 0.249  0.009  ° S n  3.5  D  2  •  4.0  4.5  Avg. Physical activity score Figure 4.1: Significant effect of PA on B M F L in boys. B M F L increases with increasing PA in boys. 84  4.2.3 Effects of Candidate Gene Polymorphisms on Bone Mineral-Free Lean Mass after Accountingfor Co In the following section, I investigated the independent effect o f candidate genotype on B M F L mass. The aforementioned covariates were controlled followed by the addition o f a single genotype variable to the B M F L mass General Linear Model. Table 4.11: The main effects of genotypes on B M F L in boys and girls after accounting for the covariates. The effect size (parrial-T) ) for the candidate gene genotype and the significance of that variable are also provided. 2  Boys Significance COL1A1 VDR VDR  Bsml Fold  T N F R 2 A593G  P P P P  = 0.307 = 0.862 = 0.921 = 0.815  partial-T| 0.007 0.002 0.001 0.003  2  Girls Significance  P = 0.976 P = 0.447  P = 0.601 ag > aa gg P 0.038 P = 0.585 P = 0.677 P = 0.885 =  partial-T| <0.001 0.011 0.007 0.030  2  =  T N F R 2 T598G T N F R 2 T620C T N F R 2 Haplotype  0.590 P = 0.396 P = 0.885 P  =  0.002 0.012 0.005  0.002 0.005 0.005  25600.00  gg  ag  aa  TNFR2 A593G Figure 4.2: Significant difference in B M F L for girls by T N F R 2 A593G genotype. Girls with the T N F R 2 gg and ag genotype have a significandy higher B M F L then girls with aa genotype.  85  Table 4.12: The main effects of genotypes on B M F L in the Caucasian boys and girls after accounting for the covariates. The effect size (partial-ri ) for the candidate gene genotype and the significance of that variable are also provided. 2  COL1A1  VDR BsmI VDR Fold TNFR2 TNFR2 TNFR2 TNFR2  A593G T598G T620C Haplotype  Boys Significance  partial-r|  p p p p p p p  0.000 0.033 0.009 0.028 0.008 0.017 0.045  = = = = = = =  0.955 0.294 0.730 0.354 0.451 0.532 0.368  2  Girls Significance p p p p p p p  = = = = = = =  0.770 0.285 0.256 0.122 0.644 0.133 0.394  parti al-r) 0.001 0.033 0.035 0.056 0.003 0.089 0.056  2  Table 4.13: The main effects of genotypes on B M F L in the Asian boys and girls after accounting for the covariates. The effect size (partial-r) ) for the candidate gene genotype and the significance of that variable are also provided. 2  Boys Significance COL1A1  partial-ri  T N F R 2 A593G  Not p = p = p =  T N F R 2 T598G T N F R 2 T620C T N F R 2 Haplotype  Not polymorphic p = 0.704 0.013 p = 0.594 0.020  VDR BsmI VDR Fold  Polymorphic 0.693 0.003 0.489 0.027 0.594 0.020  2  Girls Significance  partial-T|  p = 0.897 p = 0.607 gg=ag>aa p = 0.012  0.000 0.021 0.170  p = 0.988 p = 0.112  0.001 0.170  2  26000.00  gg  ag  aa  TNFR2 A593G genotype Figure 4.3: Significant difference in B M F L for Asian girls by T N F R 2 A593G genotype. Asian girls with the T N F R 2 gg and ag genotype have a significandy higher B M F L then Asian girls with aa genotype. 86  4.2.4 Covariate Modeling of Bone Parameters As describe in the previous section to specifically examine effects o f the variables of interest on proximal femur, femoral neck or lumbar spine B M C or femoral neck a B M D , covariates must be included in the models to account for (1) variance in body size and (2) differences in maturity. T o control for diversity in body size the covariate leg length was used for proximal femur B M C and femoral neck B M C and a B M D measures while sitting height was used to control lumbar spine B M C measures. I used total fat mass and B M F L mass to control for body mass, as these covariates effectively control for mass without controlling for the mass o f bone itself. Tanner breast (girls) and pubic hair (boys) stage were used to control for differences in maturity between subjects. Chronological age differences were also controlled. Analysis was done within each sex independendy to control for gender differences. Tables 4.14 — 4.21 outline the total variance accounted for in proximal femur, femoral neck and lumbar spine B M C and femoral neck a B M D , the statistical significance o f each covariate and the effect o f each covariate on the total model (partial-T} ). 2  Table 4.14: Variance accounted (R ) for proximal femur B M C by the covariates: B M F L , total fat mass, leg length, age and Tanner pubic hair stage in boys; significance of each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Leg Length Tanner Pubic Hair Stage Age R = 0.700  p p p p p  < 0.001 = 0.013 = 0.271 = 0.778 = 0.012  partial-rf  0.341 0.040 0.008 0.001 0.041  2  87  Table 4.15: Variance accounted (R ) for proximal femur B M C by the covariates: B M F L , total fat mass, leg length, age and Tanner Breast Stage in girls; significance of each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Leg Length Tanner Breast Stage Age R = 0.718 |  p p p p p  < 0.001 = 0.385 = 0.033 = 0.948 = 0.980  partial-rf  0.344 0.005 0.030 0.001 0.000  2  Table 4.16: Variance accounted (R ) for femoral neck B M C by the covariates: B M F L , total fat mass, leg length, age and Tanner pubic hair stage in boys; significance o f each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Leg Length Tanner Pubic Hair Stage Age R = 0.544  p p p p p  < 0.001 = 0.005 = 0.524 = 0.816 = 0.137  partial-rf  0.238 0.050 0.003 0.000 0.014  2  Table 4.17: Variance accounted (R ) for femoral neck B M C by the covariates: B M F L , total fat mass, leg length, age and Tanner Breast Stage in girls; significance of each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Leg Length Tanner Breast Stage Age R = 0.632  p p p p p  < 0.001 = 0.008 = 0.331 = 0.359 = 0.914  partial-rf  0.246 0.046 0.006 0.006 0.000  2  Table 4.18: Variance accounted (R ) for femoral neck aBMD by the covariates: B M F L , total fat mass, leg length, age and Tanner pubic hair stage in boys; significance o f each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Leg Length Tanner Pubic Hair Stage Age R = 0.421  p p p p p  < 0.001 = 0.029 = 0.055 = 0.838 = 0.412  partial-rf  0.124 0.031 0.029 0.000 0.004  2  88  Table 4.19: Variance accounted (R ) for femoral neck aBMD by the covariates: B M F L , total fat mass, leg length, age and Tanner Breast Stage in girls; significance of each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Leg Length Tanner Breast Stage Age R = 0.371 |  p p p p p  < 0.001 = 0.059 = 0.647 = 0.745 = 0.204  partial-rf  0.152 0.023 0.001 0.001 0.011  2  Table 4.20: Variance accounted (R ) for lumbar spine B M C by the covariates: B M F L , total fat mass, sitting height, age and Tanner pubic hair stage in boys; significance o f each covariate and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Sitting Height Tanner Pubic Hair Stage Age R = 0.622  p p p p p  < 0.001 < 0.001 = 0.050 = 0.250 = 0.040  partial-rf  0.247 0.152 0.024 0.008 0.027  2  Table 4.21: Variance accounted (R ) for lumbar spine B M C b total fat mass, sitting height, age and Tanner Breast Stas;e in girls; sig and effect of each predictor within the General Linear Model. 2  Significance  Bone Mineral-Free Lean Mass Total Fat Mass Sitting Height Tanner Breast Stage Age R = 0.757  p p p p p  < 0.001 < 0.001 < 0.001 = 0.093 = 0.757  partial-rf  0.089 0.085 0.255 0.019 0.001  2  4.2.5 Effects of Physical Activity and Dietary Calcium on Bone Parameters after accounting for Covari Tables 4.22 - 4.25 display the variance accounted for (R change) by the P A or average dietary calcium 2  intake after accounting for the aforementioned covariates.  Furthermore, the statistical significance o f  the variable is also listed. Scatter plots o f the standardized residuals o f the bone parameter after accounting for the covariates versus P A or average dietary calcium intake are provided fro statistically significant results.  89  Table 4.22: The effect of average dietary calcium intake and PA on proximal femur BMC. The variance accounted for by (R change) of the environmental variable and the significance of that variable are also listed. 2  boys R Change 0.001 0.012 2  average calcium intake PA  1.5  2.0  2.5  Significance P = 0.823 P = 0.028  3.0  girls R Change 0.002 0.001 2  3.5  4.0  Significance P = 0.330 P = 0.373  4.5  Avg. Physical Activity Score Figure 4.4: The Effect of PA on proximal femur B M C in boys after accounting for the covariates. The trend line shows a significant increase in proximal femur B M C as P A increases.  Table 4.23: The effect of average dietary calcium intake and PA on femoral neck B M C . The variance accounted for by (R change) of the environmental variable and the significance of that variable are also listed. 2  boys R Change 0.001 0.021 2  average calcium intake PA  Significance p = 0.864 p = 0.013  girls R Change 0.002 0.009 2  90  Significance p = 0.421 p = 0.090  1.5  2.0  2.5  3.0  3.5  4.0  4.5  Avg. Physical Activity Score Figure 4.5: The effect of PA on femoral neck B M C in boys after accounting for covariates. The trend line shows a significant increase in femoral neck B M C as P A increases. Table 4.24: The effect of average dietary calcium intake and PA on femoral neck aBMD. The variance accounted for by (R change) of the environmental variable and the significance of that variable are also listed. 2  boys R Change 0.003 0.025 2  average calcium intake PA  Significance p = 0.386 p = 0.019  girls R Change 0.007 0.021 2  91  Significance p = 0.235 p = 0.034  CO  •o c ro  55  -3,. 1.5  2.0  2.5  3.0  3.5  4.0  4.5  Avg. Physical Activity Score  Figure 4.6: The effect of average PA score on femoral neck aBMD in boys and girls after accounting for the covariates. The trend line shows a significant increase in femoral neck a B M D as average P A score increases. The trend line was approximately equal for both sexes therefore on the boys graph was included.  Table 4.25: The effect of average dietary calcium intake and average PA score on lumbar spine B M C . The variance accounted for by (R change) of the environmental variable and the significance of that variable are also listed. 2  boys R Change 0.000 0.000 2  average calcium intake PA  Significance p = 0.808 p = 0.839  girls R Change 0.000 0.000 2  Significance p - 0.944 p = 0.371  4.2.6 Effects of Candidate Gene Polymorphisms on Bone Parameters after Accountingfor Covariates The six polymorphisms (VDR Bsml, V D R Fokl, COL1A1 Spl, TNFR2 A593G, TNFR2 T598G, TNFR2 T620C and TNFR2 A593G-T598G Haplotype) investigated were individually added to the regression equation for boys and girls for the whole group and Asian and Caucasian sub-groups to examine their effects on bone parameters after having accounted for size and maturity covariates. Tables 4.26 and 4.32 list the p-value of the individual polymorphisms for each bone parameter in boys and girls, respectively. A measure of effect size (r) ) and R change values for the polymorphisms with 2  92  2  significant effects was also included. Bar graphs illustrating  significant differences in  the levels of the  polymorphisms follow the tables for boys and girls.  Table 4.26: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in boys. P-values indicate if the differences by the levels of the polymorphism are statistically significant. Where effects were significant the r) of the polymorphism (a measure of the effect between genotypes of the polymorphism) were included. 2  V D R BsmI  VDR  COL1A1  Fold Proximal Femur B M C Femoral Neck B M C  p = 0.304 p = 0.126  p = 0.342 p = 0.426  Femoral Neck aBMD Lumbar Spine B M C  p = 0.601 p = 0.757  p = 0.662 p = 0.176  p = 0.109 Ss/ss > SS p = 0.007 T) = 0.048 p = 0.141 p = 0.541 2  TNFR2 A593G p = 0.704 gg > aa p = 0.045 r | = 0.034 p = 0.618  COL1A1  p - 0.647  SS  genotype  Figure 4.7: Effect of COL1A1 Spl genotype on femoral neck B M C in boys. Boys with the Ss or ss genotype have a greater femoral neck B M C than those with the SS genotype.  93  T N F R 2 T620C  p = 0.331 p = 0.084  p = 0.796 p = 0.312  p = 0.144 p = 0.277  p = 0.187 p = 0.299  2  3.10  Ss/ss  T N F R 2 T598G  3.00 co  gg  ag  aa  TNFR2 A593G Figure 4.8: Effect of T N F R 2 A593G genotype on femoral neck B M C in boys. Boys with the gg genotype have a greater femoral neck B M C than those with the ag or gg genotype. Table 4.27: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Caucasian boys. P-values indicate i f the differences by the levels o f the polymorphism are statistically significant. Where effects were significant the T | of the polymorphism (a measure of the effect between genotypes of the polymorphism) were included. 2  V D R Bsml  VDR Fold  COL1A1  Proximal Femur B M C  p = 0.234  p = 0.523  Femoral Neck B M C  p = 0.175  FF>Ff=ff p = 0.053 i f = 0.048 p = 0.298  Femoral Neck aBMD Lumbar Spine B M C  p = 0.332 p = 0.835  p = 0.431 p = 0.234  p = 0.667 p - 0.559  p = 0.265  TNFR2 A593G p = 0.175  TNFR2 T598G p = 0.145  T N F R 2 T620C  gg>ag=ag p = 0.015 i f = 0.112 p = 0.122 p = 0.455  tt>tg p = 0.032 r|2 = 0.062 P = 0.115 p = 0.272  p = 0.719  p = 0.998  p = 0.520 p = 0.738  Table 4.28: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Asian boys. P-values indicate if the differences by the levels of the polymorphism are statistically significant. Where effects were significant the T| of the polymorphism (a measure of the effect between genotypes of the polymorphism) were included. 2  V D R Bsml Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  p p p p  = 0.981 = 0.935 = 0.573 = 0.497  VDR Fold p = 0.846 p = 0.787 p = 0.882 p = 0.919  COL1A1 Not polymorphic  94  TNFR2 A593G p - 0.435 p = 0.360 p = 0.318 p = 0.393  T N F R 2 T598G  T N F R 2 T620C  Not polymorphic  p p p p  = 0.392 = 0.331 = 0.068 = 0.136  17.80  ff  Ff  FF  VDR FOK1 genotype Figure 4.9: Effect of V D R Fold, genotype on proximal femur B M C in Caucasian boys. Boys with the F F genotype have a greater proximal femur B M C than those with the F f or ff genotype.  3.10  gg  ag  aa  TNFR2 A593G genotype  Figure 4.10: Effect of T N F R 2 AS93G genotype on femoral neck B M C in Caucasian boys. Boys with the gg genotype have a greater femoral neck B M C than those with the ag or aa genotype.  95  Table 4.29: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in girls. P-values indicate if the differences by the levels of the polymorphism are statistically significant. Where effects were significant the r\ of the polymorphism (a measure of the effect between genotypes of the polymorphism) were included. 2  V D R BsmI Proximal Femur B M C Femoral Neck B M C  p = 0.619 p = 0.277  VDR Fokl p = 0.227 p = 0.455  Femoral Neck aBMD Lumbar Spine B M C  p = 0.239 p = 0.992  p = 0.720 p = 0.774  COL1A1 p = 0.114 p = 0.084  TNFR2 A593G p = 0.884 p = 0.838  p = 0.244 p = 0.846  p = 0.390 p = 0.976  T N F R 2 T598G  T N F R 2 T620C  p = 0.889 tg/gg > tt p = 0.029 Tf=0.033 p = 0.097 p = 0.531  p = 0.697 p = 0.340  2.9  tg/gg  tt  Figure 4.11: Effect of T N F R 2 T598G genotype on femoral neck B M C in girls. Girls with the tg or gg genotype have a greater Femoral Neck B M C than those with the tt genotype.  96  p = 0.296 cc = tc > tt p = 0.028 n2=0.037  26.5  UJ 24.0  Figure 4.12: Effect of T N F R 2 T620C genotype on lumbar spine B M C in girls. Girls with the cc genotype have a greater LS B M C than those with the tt genotype.  Table 4.30: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Caucasian girls. P-values indicate if the differences by the levels of the polymorphism are statistically significant. Where effects were significant the l"| of the polymorphism (a measure of the effect between genotypes of the polymorphism) were included. 2  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  VDR BsmI  VDR  p p p p  p p p p  = = = =  0.737 0.910 0.750 0.640  COL1A1  Fold = = = =  0.111 0.366 0.490 0.515  p p p p  = = = =  0.481 0.406 0.931 0.738  TNFR2 A593G p = 0.928 p = 0.423 p = 0.465 p = 0.395  T N F R 2 T598G  T N F R 2 T620C  p p p p  p p p p  = = = =  0.401 0.142 0.437 0.181  = = =  0.931 0.412 0.337 0.250  Table 4.31: Significant effects of the 6 polymorphisms on proximal femur B M C , femoral neck BMC, aBMD and lumbar spine B M C in Asian girls. P-values indicate if the differences by the levels of the polymorphism are statistically significant. Where effects were significant the r] of the polymorphism (a measure of the effect between genotypes of the polymorphism) were included. 2  V D R BsmI  VDR  COL1A1  Fold Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  p p p p  = = = =  0.691 0.661 0.940 0.789  p p p p  = = = =  0.737 0.813 0.519 0.750  Not Polymorphic  97  TNFR2 A593G p = 0.575 p = 0.571 p = 0.418 p = 0.871  T N F R 2 T598G  T N F R 2 T620C  Not Polymorphic  p p p p  = = = =  0.252 0.209 0.642 0.335  Table 4.32: Significant differences of T N F R 2 A593G-T598G haplotypes in girls, Caucasian girls and Asian girls. A l = G593-T598, A 2 = A593-T598, A3 = A593-G598, A 4 = G593-G598.  Site/Measurement Femoral Neck B M C Femoral Neck B M C Femoral Neck B M C Femoral Neck B M C Femoral Neck a B M D  Sex girls girls girls Caucasian girls girls  Haplotype Comparison A 1 / A 4 vs. A l / A l A 1 / A 4 vs. A 1 / A 2 A 1 / A 4 vs. A 2 / A 2 A 1 / A 4 vs. A 2 / A 2 A 1 / A 4 vs. A 1 / A 2  significance p = 0.042 p = 0.033 p - 0.054 p = 0.045 p = 0.051  3.00  A1/A1  A1/A2  A1/A4  A2/A2  TNFR2 Haplotype Figure 4.13: Significant differences in femoral neck B M C in girls by T N F R 2 A593G-T598G haplotype. Presence of the T N F R 2 A 4 allele is associated with significandy greater femoral neck B M C . The difference is the same for femoral neck a B M D and femoral neck B M C differences in Caucasian girls. A l = G593-T598, A 2 = A593-T598, A 3 = A593-G598, A 4 = G593-G598.  4.3 Effects of Environmental Stimuli and Genotype Moderated by Bone Mineral-Free Lean Mass on Bone Parameters 4.3.1 Effects of Bone Mineral-free Eean Mass by Physical Activity Interactions on Bone Parameters The following section examines the interactive effects of PA and B M F L on proximal femur, femoral neck, lumbar spine B M C and femoral neck aBMD moderated by B M F L . 98  Table 4.33: The effects PA moderated by B M F L in boys and girls for proximal femur BMC. The effect of the addition of the interaction term on the regression (R change), the significance and effect (partial r| ) of the interaction and avg. PA score with the interaction in the model is also provided. 2  2  boys R Change  Sig.  0.034  p = 0.218 p < 0.001 p = 0.371  2  PA BMFL PA*BMFL  Partial-r| 0.010 0.087 0.005  girls R Change  Sig.  0.000  p = 0.553 p = 0.005 p = 0.653  2  2  Parti al-T) 0.002 0.054 0.001  Table 4.34: The effects of PA moderated by B M F L in boys and girls for femoral neck BMC. The effect of the addition of the interaction term on the regression (R change), the significance and 2  effect (partial n ) of the interaction and P A with the interaction in the model is also provided. 2  boys R Change  girls Sig.  Partial-r)  PA  p = 0.022  BMFL  p < 0.001 p = 0.054  0.025  2  PA*BMFL  0.013  R Change  Sig.  Partial-r|  0.0355  p = 0.916  0.000  0.099  p = 0.126  0.016  p = 0.704  0.001  2  2  0.000  2  Table 4.35: The effects of PA moderated by B M F L in boys and girls for femoral neck aBMD. The effect of the addition of the interaction term on the regression (R change), the significance and 2  effect (partial r) ) of the interaction and P A with the interaction in the model is also provided. 2  boys R Change 2  PA BMFL PA * B M F L  0.016  girls Sig.  Partial-T|  p = 0.002  0.066  p < 0.001  0.114  p = 0.004  0.054  R Change 2  2  0.000  Sig.  Partial-!"!  p = 0.777  0.001  p = 0.200  0.011  p = 0.955  0.000  2  Table 4.36: The effects of PA moderated by B M F L in boys and girls for Lumbar Spine BMC. The effect of the addition of the interaction term on the regression (R change), the significance and 2  effect (partial T| ) of the interaction and P A with the interaction in the model is also provided. 2  boys R Change  girls Sig.  Parti al-T)  PA  p = 0.019  BMFL  p < 0.001 p = 0.014  0.039  2  PA*BMFL  0.016  Sig.  Parti al-r|  0.036  p = 0.825  0.000  0.131  p < 0.001  0.134  p = 0.538.  0.003  2  R Change 2  0.001  99  2  2  4,3.2 Hffects of Candidate Gene Polymorphisms by Bone Mineral-Free Lean Mass Interactions on Bone P The following section examines the interactive effects of candidate gene genotype and B M F L on proximal femur, femoral neck, lumbar spine B M C and femoral neck a B M D moderated by B M F L .  Table 4.37: The effect of candidate genes polymorphisms (VDR Bsml, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in boys. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-T| ) of the result is also provided. 2  Proximal Femur BMC Femoral Neck BMC Femoral Neck aBMD Lumbar Spine BMC  VDR BsmI*BMFL p = 0.824  VDR Bsml P = 0.911  VDR FokI*BMFL p = 0.652  VDR Fokl  COUA1*BMFL  COL1A1  p = 0.527 .  p = 0.974  p = 0.892  p = 0.866  P = 0.872  p = 0.102  p = 0.070  p = 0.788  p = 0.985  p = 0.836  P = 0.880  p = 0.040 i f = 0.045  p = 0.031 r\ = 0.048  p = 0.524  p = 0.433  P = 0.481  p = 0.067  p = 0.060  p = 0.628  p = 0.525  p - 0.525  2  .700  ff  Ff  FF  VDR Fokl  Figure 4.14: The main effect of V D R Fold genotype for femoral neck aBMD in boys after accounting for the interaction with B M F L . Boys with the F F genotype have a greater femoral neck a B M D than those with the F f or ff genotype. 100  Table 4.38: The effect of candidate genes polymorphisms (VDR Bsml, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur BMC, femoral neck BMC, aBMD and lumbar spine B M C in Caucasian boys. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-r) ) of the result is also provided. 2  Proximal Femur BMC Femoral Neck BMC Femoral Neck aBMD Lumbar Spine BMC  C0IJA1*BMF1.,  COLM;  p = 0.205  p = 0.996  p = 0.955  p = 0.054  p = 0.034  p = 0.645  p = 0.726  P 0.721  p = 0.080  p = 0.057  p = 0.537  p = 0.512  P 0.373  p = 0.104  p = 0.067  p = 0.325  p = 0.300  VDR  VDR  VDR  BsmI*BMFL  Bsml  FokI*BMFI-.  p = 0.285  P 0.302  p = 0.267  p = 0. 904  P 0.984  p = 0.712  p = 0.411  VDR  Fokl  Table 4.39: The effect of candidate genes polymorphisms (VDR Bsml, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur BMC, femoral neck BMC, aBMD and lumbar spine B M C in Asian boys. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-ri ) of the result is also provided. 2  Proximal Femur BMC Femoral Neck BMC Femoral Neck aBMD Lumbar Spine BMC  VDR  VDR  VDR  Bsml*BMFL  Bsml  FokI*BMFL  p = 0.500  P = 0.504  p = 0.593  p = 0.575  p = 0.941  P = 0.930  p = 0.429  p = 0.376  p = 0.292  P = 0.254  p = 0.202  p = 0.183  p = 0.832  P = 0.938  p = 0.171  p = 0.169  VDR  101  Fokl  COL  1A1  *BMFL  Not Polymorphic  COL1A1  Table 4.40: The effect of candidate genes polymorphisms (TNFR2 AS93G, T N F R 2 T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD, and lumbar spine B M C in boys.  T h e significance o f the genotype by B M F L interaction and the main  effect o f the genotype with the interaction i n the regression are also included. I n the case o f significant results the effect (partial-r| ) o f the result is also provided. 2  TNFR2  TNFR2  TNFR2  TNFR2  TNFR2  A393G*BMFL  A593G  T598G*BMFL  T598G  T620C*BMFL  T620C  p = 0.057  P = 0.054  p = 0.257  P = 0.254  p = 0.818  P = 0.816  Femoral Neck B M C  p = 0.101  P = 0.091  p = 0.556  P = 0.351  p = 0.618  P = 0.503  Femoral  p = 0.067  P = 0.059  p = 0.138  P = 0.084  p = 0.353  P = 0.253  p = 0.277  P = 0.301  p = 0.267  P = 0.321  p = 0.154  P = 0.293  Proximal Femur  TNFR2  BMC  Neck aBMD Lumbar Spine B M C  Table 4.41: The effect of T N F R 2 AS93G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Caucasian boys. T h e significance o f the genotype by B M F L interaction and the main effect o f the genotype with the interaction i n the regression are also included. I n the case o f significant results the effect (partial-ri ) o f 2  the result is also p r o v i d e d . TNFR2  TNFR2  A393G*BMFL  A593G  p = 0.086  Femoral Neck B M C Femoral  Proximal  TNFR2 T398G*BMFL  TNFR2  TNFR2  T598G  T620C*BMFL  P 0.154  p = 0.130  P 0.202  p = 0.519  P 0.523  p = 0.269  P 0.492  p = 0.054  P 0.116  p = 0.774  p = 0.042  P 0.821  P 0.079  p = 0.045  P 0.081  p = 0.965  P 0.960  p = 0.313  P 0.334  p = 0.583  P 0.708  p = 0.204  P 0.112  Femur  TNFR2 T620C  BMC  Neck aBMD Lumbar Spine B M C  Table 4.42: The effect of T N F R 2 AS93G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Asian boys.  The  significance o f the genotype by B M F L interaction and the main effect o f the genotype with the interaction i n the regression are also included. I n the case o f significant results the effect (partial-T| ) o f 2  TNFR2  TNFR2 A593G  TNFR2  TNFR2  T598G*BMFI^  T598G  p = 0.140  P 0.230  N o t polymorph] c  p = 0.155  A593G*BMFI^ Proximal  TNFR2 T620C*BMFL  TNFR2 T620C  p = 0.520  P 0.566  P 0.168  p = 0.340  P 0.428  p = 0.641  P 0.710  p = 0.484  P 0.631  p = 0.102  P 0.573  p = 0.077  P 0.076  Femur BMC Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  102  -  Table 4.43: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in boys. The significance of the genotype by BMFL interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-ri ) of the result is also provided. 2  TNFR2  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  p p p p  HapJotype*BMFL  = 0.490 = 0.146 = 0.363 = 0.596  TNFR2  p p p p  Haplotype  - 0.446 = 0.113 = 0.381 = 0.571  Table 4.44: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Caucasian boys. The significance of the genotype by BMFL interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-T) ) of the result is also provided. 2  TNFR2  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  p p p p  Haplotype*BMFL  = 0.186 = 0.196 = 0.084 = 0.188  TNFR2  p p p p  Haplotype  = 0.233 = 0.238 = 0.117 = 0.191  Table 4.45: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Asian boys. The significance of the genotype by BMFL interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-r) ) of the result is also provided. 2  TNFR2  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  p p p p  = 0.204 = 0.228 = 0.641 = 0.573  Haplotype*BlviFL  TNFR2  p p p p  = = = =  103  Haplotype  0.114 0.168 0.710 0.701  Table 4.46: The effect of candidate genes polymorphisms (VDR Bsml, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in girls. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-f| ) of the result is also provided. 2  VDR  VDR  VDR  VDR  BsmI*BMFL  Bsml  FokI*BMFL  Fokl  Proximal Femur BMC  p = 0.257  p = 0.243  p = 0.367  p = 0.523  p = 0.082  P = 0.071  Femoral Neck B M C  p = 0.897  p = 0.815  p = 0.618  p = 0.567  p = 0.084  P = 0.138  Femoral Neck aBMD Lumbar Spine B M C  p = 0.693  p = 0.587  p = 0.422  p = 0.393  p = 0.406  P = 0.504  p - 0.449  p = 0.458  p = 0.938  p = 0.958  p = 0.356  P = 0.387  COLI A1  *BMFL  COL1A1  Table 4.47: The effect of candidate genes polymorphisms (VDR Bsml, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Caucasian girls. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-!") ) of the result is also provided. 2  VDR  VDR  VDR  VDR  Bsml*BMFl^  Bsml  FokI*BMFL  Fokl  Proximal Femur BMC Femoral Neck B M C  p = 0.486  p = 0.544  p = 0.111  p = 0.128  p = 0.218  P 0.430  p = 0.775  p = 0.802  p = 0.135  p = 0.153  p = 0.053  P 0.172  Femoral Neck aBMD Lumbar Spine B M C  p = 0.916  p = 0.870  p = 0.111  p = 0.114  p = 0.315  P 0.390  p = 0.475  p = 0.434  p = 0.592  p = 0.695  p = 0.186  P 0.300  COL  1A1  *BMFL  COL1A1  Table 4.48: The effect of candidate genes polymorphisms (VDR Bsml, V D R Fold and COL1A1 Spl) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD and lumbar spine B M C in Asian girls. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-ri ) of the result is also provided. 2  Proximal Femur BMC Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  VDR  VDR  VDR  VDR  BsmI*BMFl^  Bsml  FoM*BMFL  Fokl  p = 0.892  p = 0.840  p = 0.404  p = 0.379  p = 0.389 p = 0.649  p = 0.439 p = 0.662  p = 0.678 p = 0.273  p = 0.667 p = 0.243  p = 0.420  p = 0.455  p = 0.760  p = 0.821  104  COL1A1*BMFL  Not polymorphic  COL1A1  Table 4.49: The effect of candidate genes polymorphisms (TNFR2 A593G, T N F R 2 T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD, and lumbar spine B M C in girls. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-r| ) of the result is also provided. 2  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  TNFR2  TNFR2  TNFR2  TNFR2  TNFR2  A393G*BMFL  A593G  T598G*BMFL  T598G  T620C*BMFJ^  T620C  p = 0.318  P 0.350  p = 0.325  P 0.401  p = 0.952  P = 0.940  p = 0.649  P 0.751  p = 0.644  P 0.491  p = 0.260  P 0.191  p = 0.423  P 0.515  p = 0.234  P 0.195  p = 0.209  P 0.146  p = 0.125  P 0.138  p = 0.291  P 0.412  p = 0.057  P 0.091  TNFR2  Table 4.50: The effect of candidate genes polymorphisms (TNFR2 A593G, T N F R 2 T598G and TNFR2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD, and lumbar spine B M C in Caucasian girls. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-r| ) of the result is also provided. 2  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD Lumbar Spine B M C  TNFR2  TNFR2  TNFR2  TNFR2  TNFR2  A393G*BMFL  A593G  T598G*BMFL  T598G  T620C*BMFL  T620C  p = 0.789  P = 0.731  p = 0.275  P = 0.355  p - 0.703  P = 0.651  p = 0.735  P = 0.534  p = 0.187  P = 0.303  p = 0.072  P = 0.072  p = 0.407  P = 0.478  p = 0.086  P = 0.143  p = 0.353  P = 0.357  p = 0.809  P = 0.743  p = 0.557  P = 0.684  p = 0.144  P = 0.131  105  TNFR2  Table 4.51: The effect of candidate genes polymorphisms (TNFR2 A593G, T N F R 2 T598G and T N F R 2 T620C) moderated by B M F L on proximal femur B M C , femoral neck B M C , aBMD, and lumbar spine B M C in Asian girls. The significance of the genotype by B M F L interaction and the main effect of the genotype with the interaction in the regression are also included. In the case of significant results the effect (partial-r) ) of the result is also provided. 2  TNFR2  TNFR2  TNFR2  A393G*BMFL  A593G  T598G*BMFL  Proximal Femur B M C Femoral Neck B M C Femoral Neck aBMD  p = 0.663  P 0.593  Not polymorphic  p = 0.137  Lumbar Spine B M C  TNFR2 T598G  TNFR2  TNFR2  T620C*BMFL  T620C  p = 0.962  P 0.949  P 0.104  p = 0.926  P 0.969  p = 0.103  P 0.076  p = 0.979  P 0.944  p = 0.207  P 0.306  p = 0.816  P 0.921  Table 4.52: The effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur BMC, femoral neck B M C , aBMD and lumbar spine B M C in girls. The significance of the haplotype by B M F L interaction and the main effect of the haplotype with the interaction in the regression are also included. In the case of significant results the effect (partial-r| ) of the result is also provided. 2  TNFR2  Haplotype*BMFL  TNFR2  Haplotype  Proximal Femur B M C Femoral Neck B M C  p = 0.152 p = 0.004 i f = 0.109  p = 0.006 T | = 0.100  Femoral Neck aBMD  p = 0.003 T | = 0.110 p = 0.182  p = 0.003 i f = 0.112 p = 0.194  Lumbar Spine B M C  p = 0.154 2  2  3.30  A1/A1  A1/A2  A1/A4  A2/A2  T N F R 2 Haplotype  Figure 4.15: Main effects of T N F R 2 A593G-T598G haplotype on femoral neck B M C after the haplotype by B M F L interaction is accounted for in girls. Girls with the A 4 allele have a greater femoral neck a B M D than girls with other haplotypes. The main effects for femoral neck a B M D , had a similar profile and therefore this graph will be used as a proxy for that graph.  106  Table 4.53: T h e effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C i n Caucasian girls.  The  significance of the haplotype by B M F L interaction and the main effect of the haplotype with the interaction in the regression are also included. In the case of significant results the effect (partial-ri ) of the result is also provided. 2  TNFR2  Haplotype*BMFL  TNFR2  Haplotype  Proximal Femur B M C  p = 0.348  Femoral N e c k B M C  p = 0.003 T | = 0.220  p = 0.006 i f = 0.199  Femoral N e c k a B M D  p = 0.004 i f = 0.212  p = 0.004 r f = 0.212  Lumbar Spine B M C  p = 0.304  p = 0.340  p = 0.409 2  Table 4.54: T h e effect of T N F R 2 A593G-T598G haplotype moderated by B M F L on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C i n A s i a n girls.  The  significance of the haplotype by B M F L interaction and the main effect of the haplotype with the interaction in the regression are also included. In the case of significant results the effect (partial-ri ) of the result is also provided. 2  TNFR2  Proximal Femur B M C Femoral N e c k B M C Femoral N e c k a B M D Lumbar Spine B M C  p p p p  Haplotype*BMFL  = 0.663 = 0.137 = 0.110 = 0.107  TNFR2  p p p p  Haplotype  = 0.593 = 0.104 = 0.106 = 0.106  3.40  A1/A1  A1/A2  A1/A4  A2/A2  TNFR2 Haplotype Figure 4.16: M a i n effects of T N F R 2 A593G-T598G haplotype on femoral neck B M C after the haplotype by B M F L interaction is accounted for i n Caucasian girls. Girls with the A 4 allele have a  greater femoral neck B M C than girls with other haplotypes. The main effects for femoral neck a B M D , had a similar profile and therefore this graph will be used as a proxy for that graph.  107  4.4 Effects of Candidate Gene Polymorphism and Physical Activity Interactions on Bone Parameters In the following section I examined the interactions between candidate gene polymorphisms and on proximal femur B M C , femoral neck B M C , a B M D and lumbar spine B M C . Table 4.55: Interaction of COL1A1 Spl genotype and PA in girls. The interaction was significant at the femoral neck a B M D measurement. The significance and the effect size of the interaction are also reported.  ethnicity  C0L1A1  Femoral aBMD  Neck  all  Femoral aBMD  Neck  Caucasian  Proximal BMC  Femur  Caucasian  Femoral Neck B M C  Caucasian  p = 0.014 partial-rf = 0.042 p = 0.021 partial-rf = 0.073 p = 0.043 partial-rf = 0.056 p = 0.033 partial-rf = 0.063  Site/Measurement  ^  Spl*  PA  COL1A1  PA  p = 0.012 partial-rf = 0.044 p = 0.022 partial-rf = 0.072 p = 0.036 partial-rf = 0.060 p = 0.027 partial-rf = 0.067  P = 0.731  .680  o  Ss/ss  SS  COL1A1  Figure 4.17: Main effects of COL1A1 genotype on femoral neck aBMD after the genotype by B M F L interaction is accounted for in girls. Girls with the Ss/ss genotype have a greater femoral neck a B M D than girls with other haplotypes. The main effects for femoral neck B M C , femoral neck a B M D and proximal femur B M C in Caucasian girls had a similar profile and therefore this graph will be used as a proxy for those graphs.  108  P = 0.312 P = 0.508 P = 0.854  Table 4.56: Interaction of COL1A1 Spl genotype and PA in boys. The interaction was significant at the femoral neck a B M D measurement. The significance and the effect size of the interaction are also reported.  Site/Measurement  ethnicity  C0L1A1  Proximal BMC  Caucasian  p = 0.009 partial-rf = 0.093  Femur  Spl*  PA  COL1A1  PA  p = 0.012 partial-rf = 0.087  p = 0.013 partial-rf = 0.085  16.90  15  Ss/ss  SS  COL1A1 genotype Figure 4.18: Main effects of COL1A1 genotype on femoral neck B M C after the genotype by B M F L interaction is accounted for in Caucasian boys. Boys with the Ss/ss genotype have a greater femoral neck B M C than girls with other haplotypes.  109  CHAPTER 5  Discussion I will focus on B M F L as the primary variable o f interest and discuss:  1.  The role o f body size (height), composition (total body fat) age and maturity on B M F L .  2.  The role o f P A on B M F L with body size, body composition, age and maturity variables controlled.  3.  The role o f bone mass candidate genes on B M F L with body size, body composition, age and maturity variables controlled in the whole group and by ethnicity.  I will also focus on bone mass at the proximal femur, femoral neck and lumbar spine as the primary variable of interest and discuss:  1.  T h e role o f body size (leg length or sitting height), composition (total body fat and B M F L ) and maturity on bone mass.  2.  The role o f P A and dietary calcium intake on bone with body size, body composition, age and maturity variables controlled.  3.  The role o f candidate genes on bone mass with body size, body composition and maturity variables controlled in the whole group and by ethnicity.  Finally, I will discuss the additional variance accounted for in bone mass related to B M F L by P A , B M F L by candidate gene, P A by candidate gene interactions. I will also discuss the main effects of the variables that are interacting after having accounted for these interactions. 110  The overall premise of this thesis is that genes and environmental stimuli, such as PA, affect bone independently and in a complex combinatorial fashion. This relationship is further complicated by the relationship of these effectors on B M F L which is known to be closely associated with bone mass in both paediatric [191] and young adult populations [192]. Seeman et al, [193] utilized a twin model and observed that half the covariance in lean body mass and femoral neck aBMD was attributable to the same genetic factors. It was suggested that the association between muscle strength and aBMD may be determined by genes regulating body composition and size (muscle mass + bone mass) [193]. However, the role of B M F L as a mediator of gene effects on bone has received very litde attention. Finally, not only is there an independent effect of B M F L , PA and candidate genes on bone mass, there is also a perceived, if not measurable, effect of the interactions between these factors on bone. The present study sought a primary inspection of these novel interrelationships in children.  B o n e specific p o l y g e n e s  Bone Mass  B M F L specific p o l y g e n e s  BMFL  r Endocrine/Maturity  y  ^  Height  PA  calcium  t _ Body size polygenes  Figure 5.1: Schematic of effectors of bone mass. The relationship between height, maturity, Physical Activity (PA), dietary calcium, Bone Mineral-Free Lean Mass (BMFL) and genes to bone mass.  5.0.1 Genotype and Haplotype Frequencies by Ethnicity There was a significant difference in genotype and frequency between Asian and Caucasian subpopulations. The allele distribution of V D R Bsml, TNFR2 T620C and T N F R 2 haplotype were 111  significantly different between Asian and Caucasian sub-populations (p < 0.001). In the Asian population C O L 1 A 1 and T N F R 2 A 5 9 3 G genotypes were non-polymorphic thus there was a significant difference in the distribution of those genotypes by ethnicity (p < 0.001). There were not significant differences in V D R  Fokl  and T N F R 2 A 5 9 3 G genotypes by ethnicity. Further analysis of  genotype effects on B M F L mass and bone mass investigated the whole group and Asian and Caucasian sub-populations. T h e sample size o f the ethnicity by sex sub-populations were small (5070) and therefore, potentially under-powered to discern significant differences by levels of genotype in ethnic sub-populations. If a significant difference in the group as a whole was not confirmed by significant results in a ethnic sub-populations examination of the magnitudes o f effect and trend within an ethnic sub-population was examined for confirmation of the genotype effect.  5.1 The Ro/e of Body Si%e and Composition, Physical Activity and Genotype on Bone Mineral-Free I^ea Lean mass ( B M F L ) is consistently the single greatest predictor o f bone mass at any site in children and adolescents [194]. Bone mineral-free lean mass is derived from total body D X A scan and is a measure dominated by the amount o f muscle mass o f the subject [195]. There is a paucity o f literature that has direcdy investigated the effect o f lean mass on bone mass.  There are a number o f potential pathways  through which genes and P A might act to affect B M F L mass. First, the significant effects of genes may be on global body size which, in turn, affects muscle (lean) mass. Second, genes may have an independent and direct affect on B M F L . Third, the well established relationship between P A and lean mass could be modulated by the presence or absence o f specific genotypes.  The effects o f P A or genotype on B M F L were investigated by first controlling for the effects o f total height and total fat mass. Total height and B M F L had a simple correlation (r ) between 0.83 — 0.89 (p 2  <0.001) for boys and girls, respectively. Likewise, total fat mass and B M F L had a simple correlation  112  (r ) between 0.63 — 0.75 (p <0.001) for boys and girls, respectively. It is clearly important to control 2  for stature as a taller subject will have greater tissue mass due simply to a greater body size. Subjects with greater fat mass will likely also have a greater absolute B M F L mass in response to carrying a greater load during locomotion. Controlling for age and maturation [Tanner Breast (for girls) or Pubic hair stage (for boys)] eliminates the maturity related increase in B M F L [196]. There is considerable inter-subject variability in biological maturity that is not accounted for due to the categorical nature o f Tanner staging. Recently, alternative methods o f assessing biological maturity in pre- and adolescent children using ratios o f sitting height, leg length and total height has been developed as a more accurate tool o f assessing pubertal maturity levels [43].  5.1.1 Independent Effects of Physical Activity Levels on Bone Mineral-Free Lean Mass P A was positively associated with B M F L in boys (partial-rj = 0.028, p < 0.039). Boys in the highest 2  quintile o f P A had an approximately 3% greater B M F L than boys in the lowest quintile o f P A . The positive effect o f P A on B M F L was not observed for girls. The lack o f association in girls may be a result of uncontrolled endocrine changes due to an increased number o f higher pubertal stage (Tanner 2+) girls than boys during this measurement period, differences that effected B M F L in a P A independent fashion[196].  Girls typically increase in percentage fat mass o f total weight as they enter  higher stages o f puberty while boys increase in B M F L as they enter higher stages o f puberty [197]. Alternatively, the lack o f association in girls may be due to significandy lower mean P A (2.87 vs. 3.12, p < 0.001) in girls than boys. Furthermore, there was a substantial difference in P A distribution by sex, (SD for boys: 0.54 vs. S D for girls: 0.49). Taken together, the lack o f association between B M F L and P A in girls may be due to maturity stage, range in maturity o f the group and homogeneity of P A for the girls.  113  5.1.2 Independent Effects of Genotype on Bone Mineral-Free Lean Mass In girls, there was a statistically significant effect o f T N F R 2 A 5 9 3 G genotype on B M F L . Girls with a T N F R 2 A 5 9 3 G gg or ag genotype had a 3.0% greater B M F L than girls with a T N F R 2 A 5 9 3 G aa genotype. This effect was confirmed in Asian girls (7.2% difference in B M F L mass between Asian girls with an ag vs. aa genotype, p = 0.012) but not Caucasian girls. However, a similar but nonsignificant trend was observed in Caucasian girls (4% greater B M F L mass in gg and ag vs. aa genotype, p = 0.122). T o the best o f my knowledge there have been no previous reports linking T N F R 2 genotype with muscle mass or B M F L . Likewise, there are no biochemical or cytological links in the literature between the T N F , II and T N F R cytokine network with skeletal muscle mass. There are, however, links with T N F R 2 genotype and obesity [189] suggesting that T N F R 2 genotypes are associated with other tissue types specifically those that effect total body size and composition. Furthermore, Expressed Sequence Tags (ESTs) from muscle tissue cultures have shown that T N F R 2 is expressed in high copy numbers in muscle (GenBank entry number B i l l 5 1 6 2 , BF204631, BG828205 and BE299738) [198], Taken together, the present data suggest a role o f T N F R 2 genotypes in moderating B M F L mass as well as previous reports that have shown the same genotype are associated with bone mass and adiposity.  5.2 The Role of Body Si^e, Composition and Maturity on Bone Mass in Children Previous investigations o f boys and girls of this age have demonstrated that examination of bone parameters requires prudence in accounting for covariates [84, 85]. It is well known that children of the same chronological age vary dramatically in both their maturity level at any given age (timing) but also in the rate or tempo at which they proceed through readily identifiable maturational stages [199]. That is, although all children follow a similar growth trajectory, they do it in very different ways. Thus, to examine the effect o f dietary calcium intake, P A or genotype on bone mass in children it is 114  manifestly important to control  the bone measurement for body  size and maturity (section 2.5, p. 12-13). I  have selected covariates whose relationship to the primary outcome variable have been previously established using hierarchical regression models [84, 85]. Thus, we developed a linear regression model that used the least number of covariates, and acknowledged both the biological and statistical relevance o f the predictors.  Body size and composition accounted for approximately 55-70% o f the variance in B M C , depending on the site measured. These finding are consistent with previous studies conducted by our group [84, 85]. Measures o f body size (leg length or sitting height) and composition ( B M F L and total fat) accounted for the greatest amount of the variance (~ 50%) in bone mass at the proximal femur and spine. It should also be noted that in boys there was a higher variance in fat mass than observed in girls (9264 ± 5930 g vs. 9948 ± 4902 g). It is likely that the increased variance in fat mass is the reason that fat mass was a significant predictor o f proximal femur B M C in boys and not girls. Tanner breast (girls) or pubic hair (boys) stage and age, in contrast, accounted for only ~ 5-10% o f the bone mass variance. This may be a function o f the maturational homogeneity o f this primarily prepubertal group at this measurement period. We would anticipate an increased maturational divergence as these children enter early puberty, as was seen to some extent for girls. Eighty-eight percent o f boys measured were in Tanner pubic hair stage 1 while only 59% o f girls were in Tanner breast stage 1. Girls achieve puberty, on average, 2 years in advance o f boys [200]. In addition the small effect of maturity in correcting for bone mass may also be due to the collinearity between body size and body composition covariates and maturity covariates within the model.  115  5.3 Independent Effects ofAverage Physical Activity Score and Dietary Calcium Intake on Bone Mass 5.3.1 Average Physical Activity Score and Bone Mass 5.3.1.1 Proximal Femur Bone Mass We showed a positive relationship in boys between high P A and high proximal femur, femoral neck B M C and femoral neck a B M D and high P A and high femoral neck a B M D in girls. There was a mean difference in proximal femur and femoral neck B M C or femoral neck a B M D between the lowest and highest quintiles o f P A o f 4-6% (p = 0.013 - 0.034) depending on the site measured. It should be noted that P A was only a significant predictor o f femoral neck a B M D in girls. A s was noted with the P A and B M F L association, it is likely that the lower mean value and lack o f variance in P A observed in the girls decreased the association o f P A with bone mass. It is likely that P A is associated with other measures o f proximal femur bone mass in girls. However, due to the higher variance in proximal femur B M C and femoral neck B M C associations with P A were not significant. Jones and Dwyer [201], found a similar correlation (2% inter-quartile range) in 8 year old children between femoral neck a B M D (when controlling for size or lean mass) and the highest and lowest quartile o f leg muscle strength. T h e magnitude o f the differences between extremes o f P A on proximal femur or femoral neck B M C or femoral neck a B M D are within the same ranges o f effect o f prospective, randomizedcontrol studies, with 7-10 month exercise interventions (3-5% difference in proximal femur B M C between exercise intervention and controls) [78, 79].  P A accounted for between 1 -2% o f the variance in proximal femur, femoral neck B M C or femoral neck a B M D .  This is in keeping with other studies correlating P A (hours/week) with total body  a B M D that is in the order o f 2% in boys aged 4 to 20 years [202]. T h e benefits o f high levels of P A on B M C or a B M D during childhood are becoming increasingly apparent as prospective data are establishing the causal link [77-79, 84]. Activities that involve higher than usual forces have a more  116  profound effect on bone than usual childhood activities [203]. That is, there is a segregation of increased bone mass accrual associated with degrees of loading and forces placed on bones in athletic training (i.e. squash training>aerobic dancing=speed skating training>sedentary lifestyle) in women. A similar effect has also been observed for adolescent female tennis players, where increased activity in the dominant arm of tennis players was associated with a higher bone mass than the dominant arm of age matched controls [68].  Bone responds to high loading by increased formation thereby increasing bone mass [77], size [82] and moment of inertia [204] thereby increasing bone strength. Furthermore, there is evidence that the effect of PA may have a ceiling response in bone. Rats trained for 6 weeks at moderate levels of exercise (60% of maximal oxygen uptake, V 0 Max) increased osteogenesis [205] while rats trained at 2  80% of V 0 Max or greater for a similar period decreased bone mass [206]. Thus, a threshold like 2  behaviour for increased bone mass can be imagined for bones response to PA. This may be due to a decease in mechano-sensitivity of bone cells to high levels loading exercise during extended periods [207]. Similarly, studies of turkey's undergoing varied cycle number and magnitudes of loading found that there is an optimal amount of strain and cycle number that induces the greatest amount of bone gain beyond which there is no further increase in bone [208]. Beyond high levels of training and/or PA there is no increased benefit of PA in animal models, however this has not been studied in children.  53.1.2 Lumbar Spine Bone Mass The benefits of high PA were not observed at the lumbar spine BMC. It has been suggested that bone mass at the proximal femur reflects loading patterns more so than bone mass at the lumbar spine, that is influenced primarily by endocrine factors [209]. However, high physical activity due to gymnastic 117  training in girls did result in significandy higher lumbar spine bone mass compared with girls involved in non-gymnastic type exercise [210]. Likewise, a study o f an 8-month jumping intervention off a high (61 cm) box showed a 3.1% increase in lumbar spine B M C [78]. These studies suggest that in order for P A to have an effect on lumbar spine bone mass a significant loading strain must be placed on the lumbar spine during the training. Such strains can be placed on the lumbar spine if the exercise involves higher impact landing then typical P A . Thus, the relationship between P A and lumbar spine bone mass may not have been seen in this study due to a lack of loading o f the trunk during the subjects' daily activity.  5.3.2 Average Daily Calcium Intake and Hone Mass Average daily calcium intake did not affect bone mass in either sex at any site measured in this study. In previous studies where an effect o f daily calcium intake was demonstrated, marked benefits on childhood bone formation were noted when supplementation was provided for spontaneously low dietary calcium intake subjects or when spontaneously low dietary calcium intake subjects were compared with moderate to high calcium intake subjects [50, 51, 54]. A s such, there is evidence that the effect of dietary calcium intake has a ceiling effect. That is, beyond a dietary calcium intake that is approximately the recommended daily intake (between 700 and 800 mg/day) there is litde to no benefit for increased bone mass [52]. T h e mean (SD) average daily calcium intake o f the subjects in the present study was 852 ± 365 m g / d for boys and 777 ± 366 m g / d for girls. Furthermore, 50% (inter-quartile range) o f subjects o f either sex had average daily calcium intakes between approximately 590 and 1100 mg/day. T h e lack o f effect o f average daily calcium intake in the present study likely reflects that most subjects had moderate to high average daily calcium intakes. Therefore, no positive effects o f calcium intake could be extrapolated from the distribution of average daily calcium intake in these subjects. T h e distribution o f calcium intake was not divergent enough in this population to 118  make further investigations o f this variable in combination with genotypes or B M F L . These findings are consistent with our previous work where calcium did not predict absolute [211] or change values [79, 84] for bone mass at any site in children.  5.4 Independent Effects of VDRFokl and BsmI, COL1A1 Sp1 and TNFR2 A593G, T598G, T620C Genotype on Bone Mass We selected to study specific candidate genes based on previous evidence that their polymorphisms were associated with variance in bone mass [104,107,143,179]. By and large these studies have focused on adult, female and post-menopausal subjects and related levels o f these polymorphisms to a B M D , B M C or fracture. There have only been meta-analyses for C O L 1 A 1 [156] and the V D R 5' polymorphisms  (BsmI,Apal  and  TaqI)  [105] and these meta-analyses have focused on combining  results from adult populations.  The discussion presented in the following section focuses on the significant effects o f genotype levels on bone mass. A l l significant effects o f genotypes accounted for 1.5 — 3.5% o f the variance in B M C (all sites) or a B M D (femoral neck only) which is consistent with other studies o f these genotypes and the tenets of Quantitative Trait Theory with regards to the individual effect o f polymorphisms on the quantitative trait [105,107,120,143,156,179].  5.4.1 VDR BsmI and Fokl Genotype and Bone Mass Neither the V D R  Fokl nor the  VDR  BsmI genotypes were independent  predictors of bone mass in  either sex at any site measured. There is considerable debate in the literature whether the V D R polymorphisms have an effect on bone mass in adult or paediatric populations (section 2.13, p. 36-47) [104,131,132,136,137,139]. There have been many studies either suggesting [104,137,139] or refuting [131, 132,136] an association between the V D R 3 ' U T R polymorphisms and bone mass. 119  Although there were no significant differences in bone parameters by V D R  Bsml and V D R Fokl in  boys or girls as a whole, there was a significant difference in proximal femur B M C in Caucasian boys (6.2% greater proximal femur B M C for Caucasian boys with the F F genotype than Caucasian boys with the F f or ff genotype). These findings are in agreement with Ames  etal,  [147]that found that  intestinal calcium absorption and a B M D was 8% greater in F F genotype Caucasian children vs F f and ff genotyped children (7-12 years old). Ames  etal,  [147] suggested that the V D R F o H f allele has a  lower V D R activity in intestinal calcium absorption and therefore the concomitant lower a B M D is due to decreased serum calcium during growth.  5.4.2 COL1A1 Sp1'Genotype and Bone Mass There has only been one study to date investigating the effect o f C O L 1 A 1 genotypes in children (109 healthy, prepubertal Mexican-American girls) [167]. The 22 girls with the Ss genotype and one girl with the ss genotype had 6.7% and 49.4% lower cancellous bone density in the lumbar spine than girls with the SS genotype measured by Q C T . In contrast, there was no association between the size of the vertebrae and C O L I A 1 genotypes. The current study was o f a different modality ( D X A vs. Q C T ) . T o date it is the only reported study of the association o f C O L I A l and paediatric bone mass measured by DXA.  Boys with the Ss or ss genotype had a 4.8% greater femoral neck B M C than those with an SS genotype. These relationships were not significant at the lumbar spine, however, a similar trend was observed for total proximal femur B M C (3%, NS). This significant effect was not confirmed in the Caucasian sub-population. However, a similar but non significant trend was observed (Ss/ss genotyped Caucasian had a 4% greater femoral neck and proximal femur B M C than SS genotype Caucasian boys, p = 0.265-0.523) The effect of the C O L 1 A 1 S p l polymorphism effect is different in  120  this population than the consensus effect from the meta-analysis conducted with data from a sample of adult, Caucasian women that reported higher lumbar spine a B M D associated with the SS C O L 1 A 1 genotype rather than the Ss or ss C O L 1 A 1 genotype [156]. Likewise, these findings disagree with the lumbar spine trabecular volume data from Q C T in pre-pubertal, Mexican-American girls that showed a greater lumbar spine trabecular volume in girls with the SS C O L I A l genotype rather than the Ss genotype [167].  It has been previously shown that in elderly populations the presence o f the C O L 1 A 1 's' allele predicts fracture in an aBMD-independent manner [157, 212]. Studies o f core bone samples from the femoral head showed that Ss genotyped samples were significandy weaker but had similar B M C than SS genotyped samples. M a n n  etal,  2001 showed in RNase protection assays C O L 1 A 1 m R N A was  significantly higher in Ss osteoblasts compared with SS osteoblasts [156]. T h e authors suggested the increased transcription and translation o f C o l l A l resulted in formation o f C o l l a l homotrimers which would be structurally weaker in bone matrix than the typical 2 C o l l a l + 1 C o l l a 2 heterotrimers. Therefore, the increase in femoral neck mass in the present study may be a means o f increasing femoral neck cross sectional moment of inertia to compensate for otherwise weaker bones, caused by C o l l a l homotrimer formation in the E C M , in individuals with an Ss or ss C O L 1 A 1 genotype. However, we did not measure this in the current study. Previous work from our lab has identified that these same girls increased femoral neck C S A thereby increasing cross sectional moment o f inertia in response to an exercise intervention [46]. Therefore, we know that the structure o f the femoral neck accommodates increased strain by increasing C S A in girls. This was not the case for boys (unpublished data). If indeed the s allele o f C O L 1 A 1 causes decreased structural strength o f bone  121  E C M , we could assume that the increased strain on the femoral neck would result in an increase in femoral neck CSA and therefore BMC. However, this was not assessed.  5.4.3 TNFR2 A593G, T598G and T620C Genotype and Bone Mass Analysis of the three 3'UTR TNFR2 polymorphisms (A593G, T598G, T620C) in the present study was unlike previous investigations of these polymorphisms in other populations that used the combined haplotype of the three polymorphisms. As data were collected using automated sequencing, some of the TNFR2 haplotype phases could not be discerned without collecting parental haplotypes (i.e. any haplotype with 2 or more heterozygotes at any of the 3 polymorphic sites).. The loss of n from analysis because of intractable TNFR2 haplotype prohibited such analysis. Instead, I investigated the haplotypes of the first two polymorphisms, TNFR2 A593G and T598G. The haplotype phase of the first 2 polymorphisms (TNFR2 A593G-T598G) minimized the loss of subjects and retained the pertinent data of the haplotypes. There are only two studies to date that have measured these polymorphisms in relationship with bone parameters. Both these studies focused on adult, Caucasian women [178,179].  Boys with the TNFR2 A593G gg genotype had a significandy greater femoral neck B M C (3.4%) than boys with an aa genotype. The heterozygous ag TNFR2 A593G genotype was not significandy different from either homozygous genotypes. These findings were supported in Caucasian boys where Caucasian boys with the TNFR2 A593G gg genotype had an 8.7% greater femoral neck B M C than Caucasian boys with an ag or aa genotype. Likewise, there were similar but non-significant trends in Caucasian boys for greater femoral neck aBMD (p = 0.122) and proximal femur B M C (p = 0.175) that support the findings that there is a higher proximal femur bone mass in Caucasian boys with the gg TNFR2 A593G genotype than Caucasian boys with the ag or aa genotype. These findings are in 122  general agreement with the findings of Albagha etal, [179] that showed that femoral neck aBMD is lower for adult women with the TNFR2 aa genotype versus those with the gg genotype. The effect of TNFR2 A593G genotype on femoral neck B M C in this study mirrors the effect of T N F R 2 A593G genotype on B M F L in girls where girls with the TNFR2 A593G gg genotype had an approximately 3% greater B M F L than girls with a TNFR2 A593G aa genotype. It is possible that the aforementioned effect of TNFR2 A593G genotype on B M F L is due to an effect on global body mass (muscle mass + bone mass) and not strictly a genotype effect on bone or B M F L mass.  Girls with the tg TNFR2 T598G (and one subject with the only confirmed and reported gg) in the present study had 3.3% greater femoral neck B M C than those with a tt genotype. This trend also existed for femoral neck aBMD, although this relationship was not statistically significant (p = 0.097). Likewise, a similar but non-significant trend was seen in Caucasian girls for femoral neck BMC (p = 0.142). Neither, Spotila etal, [181] nor Albagha etal, [178, 179] showed a significant relationship between T N F R 2 T598G genotype alone and bone parameters.  TNFR2 T620C genotype showed a dominant effect of the c allele in so much as girls with either the cc or tc TNFR2 genotype had a 3.7% greater lumbar spine BMC. There were non-significant but similar trends in both Caucasian and Asian girls that confirmed the same effect of higher lumbar spine BMC with cc or tc genotyped subjects than tt genotyped subjects (p = 0.250 - 0.335). Albagha et al, [179] reported the opposite effect (tt>cc) of TNFR2 T620C on femoral neck aBMD. However, they found no significant effect at the lumbar spine.  Girls with a A 1 / A 4 TNFR2 A593G-T598G haplotype had a 11.4% greater femoral neck B M C and femoral neck aBMD than girls with a A l / A l , A 1 / A 2 or A 2 / A 2 TNFR2 A593G-T598G haplotype (p  123  = 0.033 - 0.054). Similarly, Caucasian girls with the A 1 / A 4 haplotype had a significandy greater femoral neck B M C than Caucasian girls with A 2 / A 2 haplotype. T h e presence o f the A 4 allele (G593G598) of the haplotype provides an 'advantageous haplotype'. This is in keeping with the T N F R 2 A 5 9 3 G and T N F R 2 T 5 9 8 G genotype data, shown above, wherein those with a gg T N F R 2 A 5 9 3 G genotype had a greater femoral neck B M C than those with an aa T N F R 2 A 5 9 3 G genotype. Those with the tg T N F R 2 T 5 9 8 G genotype had greater femoral neck B M C than those with the tt T N F R 2 T 5 9 8 G genotype. Likewise, this data is in keeping with previous studies of T N F R 2 haplotype that showed that A593-T598-(C620) haplotype was associated with lower bone mass in adult female populations [181,182]  Computer simulation experiments of the m R N A structure of T N F R 2 3 ' U T R variants has shown that the A593-T598-C620 haplotype has a stem loop structure that is altered compared with other T N F R 2 haplotypes [179]. Although there have been no  in vitro or in vivo experiments  to substantiate the claims,  it is possible that the T N F R 2 3 ' U T R polymorphisms affect m R N A structure and therefore m R N A processing and stability. T N F R 2 is a key receptor for the intercellular cytokines T N F - a and T N F - P that are, in turn, important determinants of osteoclastogenesis. Thus, a difference in T N F R 2 expression would result in a differential response of osteoclast progenitors to these cytokines [174, 175].  5.5 Effects ofAverage Physical Activity Score on Bone are Different by Levels of Bone Mineral-Free Lean A significant interrelationship between B M F L and P A for femoral neck B M C and a B M D and lumbar spine B M C was observed in boys. A significant interaction term for P A and B M F L means that the effect of P A on bone mass is not linear across the range of measured B M F L . T h e interaction term only added an addition 1.3-1.6% of the accounted variance for femoral neck B M C and a B M D . 124  Likewise, there was no substantial change in the effect o f P A on femoral neck B M C or a B M D (5-5.7% difference in femoral neck B M C or a B M D by first vs. fifth quintiles o f P A ) with the addition o f the P A by lean mass interaction. More notably, when the B M F L by P A interaction term was added to the model, the main effect o f P A on lumbar spine B M C became significant. T h e interaction counted for only 1.6% o f the variance in lumbar spine B M C but with this additional variance accounted for there was a significant difference in lumbar spine B M C by P A in boys. P A predicted a 1.5% difference in lumbar spine B M C between the lowest and highest quintiles o f P A score. This suggests that lumbar spine B M C only responds to high levels o f P A at certain levels of B M F L . Thus, the strains placed on the lumbar spine by P A only cause a significant difference in lumbar spine B M C when the effect of P A is moderated by degrees o f lean mass.  Nordstrom & Lorentzon [213] studied a parent-son cohort to investigate the effect o f heritability in determining a B M D , lean mass and muscle strength in adolescent boys while controlling for P A . They showed that lean mass and a B M D were under shared genetic control and that the inter-relationship between a B M D and lean mass is likely due to a general (poly)genetic body size trait.  If this line of  thought is extended into the current interaction between P A and lean mass on bone, we can suppose that the polygenetic body size trait affects the positive effect o f P A on bone mass. Muscle strength and lean mass are the paramount source of mechanical strain placed on bone. Hence, PA's effect on bone mass is moderated by the amount of muscle strength and therefore mechanical strain placed on the bone during activity.  This particular argument falls in line with the Utah Paradigm o f bone physiology as offered by H . M . Frost [214]. T h e Utah Paradigm proposes that:  125  Bones have the main purpose ojproviding only enough strength to keep voluntary physical loads, whet normal, or supranormal, from causing spontaneousfractures [215]  As such, bone responds to strain and strain thresholds to maintain bone strength by modeling or remodeling bone according to mechanical strain. These strains on bone in non-traumatic situations are by and large placed on bone by muscle. The poor lever arms in locomotion require muscles to apply over 2 kg of muscle force on bones to move each kilogram of body weight [216]. Ergo, muscle forces cause the largest voluntary bone loads and strains [217] and strongly influence bone strength and mass.  5.6 Effects of Genotype on Bone are Different by Levels of Bone Mineral-Free Lean Mass Interactions between candidate gene genotypes and B M F L were observed. The following section will discuss the significant interactions between candidate genotypes and B M F L only as how they change the main effects of genotype on bone mass.  5.6.1 Effect of VDR Fokl Genotype Moderated by Bone Mineral-Free Lean Mass on Bone Mass A statistically significant interaction between V D R Fokl genotype and B M F L was observed to change the main effect of V D R Fokl genotype on femoral neck aBMD in boys. This interaction accounted for an additional 3.2% of the variance in femoral neck aBMD in boys. Once this interaction was accounted for, the main effect of V D R Fokl genotype on femoral neck aBMD was significant such that V D R Fokl F F genotype boys had a 1.4% greater femoral neck aBMD than those with an ff genotype. A similar but non-significant effect was observed for Caucasian boys for femoral neck BMC, femoral neck aBMD and lumbar spine B M C (p = 0.054 - 0.104) and for Asian boys for femoral neck aBMD and lumbar spine B M C (p = 0.171 - 0.202). There was not a statistically significant 126  difference in bone mass at any site by V D R Fokl genotype when the interaction was not included in the model.  Arai et al, [143] showed that the V D R Fokl F allele produced a more active form of the vitamin D receptor than the f allele. Likewise, they showed that in a population of Japanese women the F allele was associated with a greater lumbar spine aBMD than the f allele. Similarly, Lucotte etal, [146] showed that the F F V D R Fokl genotype was associated with higher femoral neck aBMD in a group of post-menopausal French women. Finally, the V D R Fokl F allele was associated with an higher total body aBMD and increased intestinal absorption of stable-isotope labelled calcium in 7-12 year old children [147], The results of the present study indicate that indeed there is an association between V D R Fokl genotype and femoral neck aBMD only when the variance of the V D R Fokl genotype by B M F L interaction is accounted for.  The V D R has a central role in bone metabolism as a factor important in osteoblast and osteoclast differentiation and in serum calcium homeostasis [218-220]. As such, differences in activity of the receptor may be effective in changing bone mass only in context of other biochemical and cytological as well as tissue level differences. Differences in childhood bone mass due to differences in V D R genotype-moderated differences in calcium absorption may only be significandy different when the demands of other body tissues (e.g. muscle mass) exceed the intestinal absorption activity of V D R determined, in part, by V D R Fokl genotype. This complex relationship between the polygenes of the body size trait, calcium absorption, bone mass and activity of V D R require further investigation.  5.6.3 Effect o/TNFR2 A593G-T598G Haplotype Moderated by Bone Mineral-Free Lean Mass on Bone Mass There was a significant difference in TNFR2 A593G-T598G haplotype effects by levels of B M F L in girls at femoral neck B M C and aBMD measures. The B M F L by T N F R 2 A593G-T598G haplotype 127  interaction accounted for 4.4% and 7.1% of the variance in femoral neck B M C and a B M D respectively. T h e main effects o f T N F R 2 A 5 9 3 G - T 5 9 8 G haplotype with this interaction accounted for are the same as the main effects o f T N F R 2 A 5 9 3 G - T 5 9 8 G haplotype without the interaction where girls with the A 1 / A 4 haplotypes had a 10-11% greater femoral neck B M C and a B M D than girls with a A l / A l , A 1 / A 2 or A 2 / A 2 haplotype. The same T N F R 2 haplotype by B M F L interaction was significant for femoral neck B M C and a B M D when the analysis was done in Caucasian girls alone (p = 0.003 - 0.004). This analysis can be decomposed to the individual alleles o f the T N F R 2 A 4 9 5 3 G T 5 9 8 G haplotype that are effecting bone mass. The A 4 allele (G593-G598) results in a higher than average femoral neck B M C and a B M D . Furthermore, the A 2 (A593-T598) allele results in a decreased femoral neck B M C or a B M D when compared with other T N F R 2 A 5 9 3 G - T 5 9 8 G haplotype alleles. This in keeping with the results of Albagha  etal,  [179] that showed that the A593-T598-(C620) allele of  the T N F R 2 3 ' U T R haplotypes was associated with a lower femoral neck a B M D compared with other haplotypes.  A differential effect o f bone mass to T N F R 2 3 ' U T R haplotypes by levels o f B M F L is interesting in light o f the effect o f T N F R 2 A 5 9 3 G genotype on B M F L itself (section 5.1.2, p. 113). The results of the effect o f T N F R 2 3 ' U T R genotype on B M F L and also the T N F R 2 genotype/haplotype results on bone mass moderated by B M F L imply that the T N F R 2 3 ' U T R polymorphisms may affect bone by way o f an overall body mass trait effect. It is important to note that inter-subject differences in height were controlled during these analyses. Therefore, T N F R 2 genotype/haplotype differences in bone mass or B M F L mass are not due to increased height of the subject but increased bone mass or B M F L mass for a given height.  128  The effects o f candidate gene genotype on bone mass may actually be due to an effect o f the genotype on body size which will affect both B M F L and bone mass. T h e T N F R 2 3 ' U T R polymorphisms have been previously implicated in obesity, leptin levels and Type II diabetes in young subjects [189] suggesting that the T N F R 2 3 ' U T R polymorphisms effect is observed in other tissue types and is related to total tissue mass.  5.7 Effects of Physical Activity Moderated by Levels of Genotypes There is surprisingly little evidence in the literature surrounding interactions between genotypes and exercise levels despite over 10 years of investigation into the effects o f candidate genotypes on bone [103]. In addition, research dating back over 100 years has demonstrated the benefits of mechanical loading on bone [221]. There is clearly a relationship between genotypes and physical activity effects on bone since despite controlling for most intrinsic factors (i.e. maturational stage, body size differences, etc.) the effects of prospective exercise interventions in children have considerable intersubject variation p 8 , 79]. It is reasonable to assume that the inter-subject differences in response to exercise interventions are due, in part, to differences in the genetic profile o f the subjects. The following section investigates the effects o f P A on bone measurements stratified by levels of genotype to determine if there is a differential effect o f P A by genotype.  It is notable that the interaction between genotype and P A in the following section are all at the proximal femur. A s discussed in section 5.3.1 (p.l 13 -120) positive effects o f high P A are seen only at the proximal femur. It has been previously suggested that effects of P A on bone are seen at the proximal femur and not the spine as lumbar spine bone mass is largely controlled by endocrine factors [209]. Therefore, the difference in effect o f P A by genotype is observed at a region where P A has a significant effect.  129  5.7.1 Effects of Average Physical Activity Score Moderated by C0L1A1 Genotype on Bone Mass A n interaction between P A and C O L I A l genotype was a significant predictor o f proximal femur B M C in Caucasian boys and girls and femoral neck B M C and a B M D in Caucasian girls as well as femoral neck a B M D in the combined group o f girls. This interaction accounted for 2.8% of the variance in femoral neck a B M D in the combined group o f girls. Prior to adding the P A by C O L 1 A 1 genotype interaction, C O L 1 A 1 genotype was not an independent predictor o f femoral neck a B M D in girls (section 5.4.2, p. 121-130). However, with the interaction added to the model girls with the Ss or ss C O L 1 A 1 genotype have an approximately 4% greater femoral neck a B M D than those with the SS genotype (p = 0.014). Similar interaction effects of C O L 1 A 1 genotype by P A were observed for proximal femur B M C , femoral neck B M C and femoral neck a B M D in Caucasian girls. The interactions in Caucasian girls accounted for between 4-6% of the variance in the aforementioned sites and measurements. Likewise, in Caucasian girls the main effect o f C O L 1 A 1 genotype was not a significant predictor o f proximal femur B M C or femoral neck B M C or a B M D prior to addition of the C O L 1 A 1 by P A interaction. W h e n the interaction term was added a similar diffence in proximal femur B M C and femoral neck B M C or a B M D was observed in Caucasian girls as the combined group of girls (~5% greater proximal femur B M C , femoral neck B M C or a B M D in Caucasian girls with the Ss/ss genotype than Caucasian girls with the SS C O L I A l genotype, p = 0.022 - 0.036). In Caucasian boys there was a significant interaction between C O L 1 A 1 genotype and P A for proximal femur B M C (p = 0.009). Similarly, the main effect for C O L 1 A 1 genotype for proximal femur B M C was such that Caucasian boys with a Ss or ss genotype had a 1% greater proximal femur B M C than Caucasian boys with a SS C O L 1 A 1 genotype.  As suggested earlier C O L 1 A 1 genotype affects bone strength in a B M D independent fashion [157, 212] and bone core samples from C O L 1 A 1 Ss subjects have a lower yield strength than those from SS 130  subjects [156]. Structural weakness of bone samples from Ss subjects is likely due to the formation o f coll al homotrimers in the bone E C M decreasing bending strength independently o f B M C in the E C M . The interaction term represents a differential response o f bone mass to high P A by C O L I A l genotype. It has been previously shown that exercise interventions increase C S A and section modulus in the femoral shaft and femoral neck [46]. Thus, in this study, there may be an increase in proximal femur bone mass due to P A inversely proportional to the bone structural weakness conferred by the C O L 1 A 1 s allele. Therefore, girls and boys with the C O L 1 A 1 Ss or ss genotype may have a high proximal femur bone mass to offset the structural weakness conferred by the V allele. When high strain is placed on the proximal femur due to high P A bone responds by proportionally increasing proximal femur bone mass as a means of increasing bone strength.  5.8 Unique Apects of the Study This study is the first to simultaneously investigate 6 genotypes (2 V D R , 1 C O L I A l and 3 T N F R 2 ) and physical activity on bone mass in children. This study is also the first to investigate the effect o f T N F R 2 genotypes and haplotypes in a paedeatric population. Similarly, this is the first account to investigate the interrelationships o f P A and candidate genotypes in children. Although, the relationship between lean mass and bone mass is well established this study is the first to investigate how the effects o f genotype and P A on bone mass are moderated by lean mass. Likewise, this is the first report that investigated the effect of previously recognized bone candidate gene genotypes with regard to variance in B M F L mass.  5.9 Limitations of the Study 5.9.1 Degrees of Separation between Independent and Dependent Variables The greatest short-coming o f this thesis is the lack of direct connection between the variable being tested (independent variable - P A or genotype) and the variable being measured (dependant variable 131  bone mass or lean mass). This is a common problem when dealing with genetic association studies as there is no direct logical means o f connecting the trait studied with the genetic loci in question, nor is any implicit statement o f such a connection made. Instead the assumptions o f an association study are such that the genetic locus under investigation, not necessarily the gene itself or its biological function, is associated with the outcome variable. This is the State o f the A r t in the sense that researchers cannot use direct experimental manipulation o f genetic loci in human populations to make causal statements of genetic loci on the affected trait. Although use o f experimental animal models and in  vitro experiments  resolve these issues, effects on a quantitative trait such as bone mass is currently  difficult to manipulate in these manners. However, previous  in vitro and animal model  evidence has  suggested that V D R , C O L I A l and T N F R 2 gene products and P A are indeed involved in bone metabolism and the genotypes studied and mechanical loading are direcdy related to bone tissue and cell function in  in vitro models.  The separation o f dependent and independent variables in association studies o f bone mass is further complicated by the effects o f covariates that are also affected by the independent variable. T h e convolution o f the relationship between the dependent variable and independent variable makes it even more difficult to connect them logically. Clearly there are a multitude o f changes and effectors that are not measured and are between the actual minute effect of genotype or P A and the measured outcome, B M C or a B M D .  5.9.2 Testingfor Effects by Ethnicity Dividing groups by ethnicity resulted in relatively small groups (»'s in the order o f 50 to 70 for Asian by sex or Caucasian by sex). T h e statistical power to detect differences in bone mass by levels o f genotype with the covariates included in these models required a sample size greater than 120 for 80% 132  power (calculations not shown). Although many o f the genotype effects remained statistically significant when investigated by ethnicity some genotypes were not statitistically significant likely due to small sample size not because o f a lack o f effect of the gene in the population studied (i.e. C O L 1 A 1 in Caucasian populations).  5.9.3 Adjustmentfor Multiple Comparisons The results o f this study often required multiple comparisons o f the same dependant variable (i.e. B M F L mass or bone parameters) with independent tests o f numerous independent variables (i.e. environmental stimuli and genotypes). Rigorous statistical analysis would demand that the significance level (a-level) be adjusted for multiple comparisons proportionally to the number o f comparisons (i.e. a Bonferroni Correction). However, Bonferroni corrections for multiple comparisons over-correct in cases such as quantitative traits like bone mass [222, 223]. A s this study was not powered to investigate the combined effects of the genotypes and P A variables, the independent effects were investigated in place. This is mosdy a matter of practicality as the effects o f exercise and genotype are small, typically between 1-8%. A s such, a study design that could investigate all these variables simultaneously (i.e. an A N C O V A model with all 6 polymorphisms plus exercise variables) would require an n values in the order o f the thousands, even tens o f thousands. T h e cost for such a large study would be prohibitive for such a study design except, perhaps, in context o f a multi-centre, national agency-type program.  The literature surrounding both exercise and genotype data has by and large been focused on single effects o f a genotype or P A , therefore the Type I error rate (a-level) has been set at 0.05. Based on the a-level set in the literature that investigated indepedent effects o f genotype or P A on bone mass, the a-level in this study has remained at 0.05 without a Bonferroni correction for multiple 133  comparisons. In combining the independent effects from each model it is tempting to be skeptical of the results that are not corrected. However, there are several arguments for the data to be kept in its unadjusted form. First, there is an  a priori assumption  that the polymorphisms and environmental  stimuli affect bone in combinatorial fashion. T h e evidence from segregation analysis [101] (section 2.11.4, p. 36) suggests that both genetic and environmental variables affecting bone have minor independent effects that combine in a complex fashion for a total effect that is greater than the sum of its parts. N o t correctting for multiple comparisons allows the significant findings to be presented but with the proviso that they be interpreted conservatively. In most part, this thesis is hypothesis generating rather than hypothesis solving. That being said, with a larger sample the trends we are seeing in this data would likely be confirmed with statistically rigorous adjustment for multiple comparisons.  5.10 Future Directions Many of the unresolved issues of this thesis are related to differences in bone strength not represented by D X A measures o f B M C or a B M D . These structural differences may be dissimilar by levels of genotype or P A . Further analysis o f association of bone strength with genotype, P A and interactions would certainly be useful in determining the true nature o f the associations between the effects o f the genotypes or P A on bone strength.  5.11 Summary and conclusions 5.11.1 Independent Effect of PA or Dietary Calcium Intake on Bone MineralFree Lean Mass and Bone M The results o f the current study both refute and support the research hypotheses stated previously in this document. K e y findings are summarized below.  134  Summary  1.  Variance in P A Score is related to variance in B M F L in boys.  2.  High P A score is related to high bone mass at the proximal femur in boys.  3.  Differences in average daily calcium intake were not associated with differences in bone mass.  Conclusions  General physical activity is associated with lean and bone mass suggesting that increasing a child's general physical activity will increase both tissues masses.  5.11.2 Independent"Effectof Candidate Gene Genotypes on Bone Mineral-Free Eean Mass and Bone Ma Summary 1.  VDR  BsmI  and V D R  Fokl genotype did not  have a relationship to bone mass or B M F L in  children.  2.  C O L I A l Ss or ss genotype was associated with higher femoral neck B M C in boys but not B M F L in either sex.  3.  T N F R 2 A 5 9 3 G gg genotype was associated with higher B M F L in boys and higher femoral neck B M C in girls.  4.  T N F R 2 T 5 9 8 G tg genotype was associated with higher femoral neck B M C in girls.  5.  T N F R 2 A 5 9 3 G - T 5 9 8 G A 4 allele (G593-G598) was associated with a higher femoral neck B M C in girls than other T N F R 2 A 5 9 3 G - T 5 9 8 G haplotype alleles.  135  Conclusions The V C O L 1 A 1 allele association with higher bone mass in children may provide a biological means to offset structural bone weakness that occurs due to the formation o f C o l l a l homotrimers in bone E C M . However, we did not investigate this directly. T N F R 2 genotype and haplotype are associated with both a higher B M F L and bone mass suggesting that the T N F R 2 genotype may influence body size in combination and not high bone mass or lean mass individually.  5.11.3 The Effect of PA and Candidate Gene Genotype on Bone Mass are Moderated by Levels of BMF Summary 1.  There was a significant interaction between P A and B M F L which changed the effect of P A on bone in boys.  2.  A significant interaction between B M F L and V D R  Fokl genotype was  With this interaction added to the model girls with the V D R femoral neck a B M D than girls with other V D R  3.  Fokl  observed for girls.  F F genotype had a higher  Fokl genotypes.  There was a significant interaction between T N F R 2 haplotype and B M F L . When this interaction was added girls with the T N F R 2 A 5 9 3 G - T 5 9 8 G haplotype G593-G598 allele had a greater femoral neck B M C and a B M D than girls with other haplotype alleles and girls with the A593-T598 allele had a lower femoral neck B M C and a B M D than girls with other haplotype alleles.  Conclusions Increases in bone mass due to high P A are different by levels of B M F L mass. Therefore, the effect of P A on bone mass is mediated by B M F L mass. T h e V D R  136  Fokl genotype has  a differential effect on  femoral neck a B M D by B M F L . When this interaction was included the V D R F F genotype was associated with higher femoral neck a B M D . This suggests that V D R  Fokl genotype is  associated with  high femoral neck a B M D only in the presence o f other body size related variables. T N F R 2 genotype/haplotypes affect both lean and bone mass therefore interactions between lean mass and T N F R 2 genotypes/haplotypes effect bone mass by potentiating a body mass trait effect on bone.  5.11.4 The Effect of PA on Bone Mass is Moderated by Levels of Candidate Gene Genotype Summary 1.  There was a significant interaction between P A and C O L 1 A 1 genotype on proximal femur bone mass in boys and girls. With the interaction accounted for, boys and girls with the Ss or ss C O L 1 A 1 genotype had a higher proximal femur bone mass than those with a SS C O L 1 A 1 genotype.  Conclusions In response to high P A , subjects with the Ss or ss C O L 1 A 1 genotype may increase femoral neck strength proportionally to the structural weakness conferred by the V C O L 1 A 1 allele.  137  REFERENCES  1.  Nguyen T V , Blangero J, and Eisman J A , Genetic epidemiological approaches to the search for osteoporosis genes. J Bone Miner Res,  2000.15(3): p. 392-401.  2.  Hanley D A and Josse R G , Prevention and management of osteoporosis: consensus statementsfrom the Scientific Advisory Board ojthe Osteoporosis Society ojCanada. 1. Introduction. Cmaj, 1996.155(7): p.  3.  Josse R G , Prevention and management oj osteoporosis: consensus statements jrom the Scientific Advisory Board oj the Osteoporosis Society oj Canada. 3. Effects oj ovarian hormone therapy on skeletal and  921-3.  extraskeletal tissue in women. Cmaj, 1996.155: p. 929-34.  4.  Murray T M , Prevention and management oj osteoporosis: consensus statementsjrom the Scientific Adviso Board oj the Osteoporosis Society oj Canada. 4. Calcium nutrition and osteoporosis. Cmaj, 1996. 155: p. 935-9.  5.  Prior J C , Barr SI, Chow R, and Faulkner R A , Prevention and management oj osteoporosis: consensus statementsjrom the Scientific Advisory Board oj the Osteoporosis Society oj Canada. 5. Physical activity as therapy jor osteoporosis. Cmaj, 1996.155(7): p. 940-4.  6.  Bailey D A ,  7.  during puberty: a longitudinal analysis. J Bone Miner Res, 2000.15(11): p. 2245-50. Melton LJ, 3rd, K a n S H , Wahner H W , and Riggs B L , Lifetimefracture risk: an approach to hip fracture risk assessment based on bone mineral density and age. J Clin Epidemiol, 1988. 41(10): p.  Martin A D ,  M c K a y H A , Whiting S, and Mirwald R, Calcium accretion in girls and boys  985-  94.  8.  Puzas J E , Osteoblast Cell Biology - Lineage and Functions, in Primer on the Metabolic Bone Diseases and  9.  Bennett J H , Moffatt S, and Horton M , Cell adhesion molecules in human osteoblasts: structure and  Disorders of Mineral Metabolism, MJ Favus, Editor. 1996,  Lippincott-Raven: Philadelphia, p.  11-6.  junction. Histol Histopathol, 2001.16(2): p. 603-11. 10.  Aarden E M , Wassenaar A M , Alblas MJ, and Nijweide PJ, Immunocytochemicaldemonstration oj  11.  Chenu C , Colucci S, Grano M , Zigrino P, Barattolo R, Zambonin G , Baldini N , Vergnaud P,  12.  Delmas P D , and Zallone A Z , Osteocalcin induces chemotaxis, secretion oj matrix proteins, and calciummediatedintracellular signaling in human osteoclast-like cells. J Cell Biol, 1994.127(4): p. 1149-58. McCarthy T L , Ji C , and Centrella M , Links among growth jactors, hormones, and nuclearfactors with  extracellular matrix proteins in isolated osteocytes. Histochem Cell Biol, 1996.106(5): p. 495-501.  essential roles in bone jormation. Crit Rev 13.  Oral Biol Med,  Haussler M R , Haussler C A , Jurutka PW, Thompson P D ,  2000.11(4): p. 409-22.  Hsieh J C , Remus LS, Selznick S H ,  and Whitfield G K , The vitamin D hormone and its nuclear receptor: molecular actions and disease states. Endocrinol, 1997.154 S u p p l : p. S57-73.  14. 15.  Goltzman D , Interactions ojPTH and PTHrP with the PTH/PTHrP receptor and ivith downstream signalingpathways: exceptions thatprovide the rules. J Bone Miner Res, 1999.14(2): p. 173-7. Duncan R L and Turner C H , Mechanotransduction and thejunctional response oj bone to mechanical strain. Calcif Tissue Int,  16.  1995.  57(5): p. 344-58.  Mullender M G and Huiskes R, Osteocytes and bone lining cells: which are the best candidatesjor mechano sensors in cancellous bone? Bone, 1997.  17.  20(6): p. 527-32.  Burger E H and Klein-Nulend J, Mechanotransduction in bone-role oj the lacuno-canalicular network. FasebJ, 1999.13 S u p p l : p. SI01-12.  138  18.  Signal transduction pathways involved influidflow-induced'PGE2production by cultured osteocytes. A m J Physiol, 1999. 276(1  Ajubi N E , Klein-Nulend J , Alblas M J , Burger E H , and Nijweide PJ,  Pt 1): p  E171-8.  21.  Mundy G R , Bone-Resorbing Cells, in Primer on the Metabolic Bone Diseases and Disorders ofMine Metabolism, M J Favus, Editor. 1996, Lippincott-Raven: Philadelphia. Blair H C , Teitelbaum SL, Ghiselli R, and Gluck S, Osteoclastic bone resorption by a polarised vacuolar proton pump. Science, 1989. 245(4920): p. 855-7. Komarova S V , D i x o n SJ, and Sims S M , Osteoclast ion channels: potential targetsfor antiresorptive dr  22.  Delaisse J M , Engsig M T , Everts V , del Carmen Ovejero M , Ferreras M , L u n d L , V u T H ,  19. 20.  Curr Pharm Des, 2001. 7(8): p. 637-54. Werb Z , Winding B , Lochter A , Karsdal M A , Troen T , Kirkegaard T , Lenhard T , Heegaard A M , N e f f L , Baron R, and Foged N T ,  Proteinases in bone resorption: obvious and less obvious roles  Clin C h i m Acta, 2000. 291(2): p. 223-34.  23. 24. 25. 26. 27. 28.  Collagenolytic activity of cathepsin K is specifically modulated by car resident chondroitin sulfates. Biochemistry, 2000. 39(3): p. 529-36. Zhao W , Byrne M H , Boyce B F , and Krane S M , Bone resorption induced by parathyroid hormone is strikingly diminished in collagenase-resistant mutant mice.] Clin Invest, 1999.103(4): p. 517-24. Holliday L S , Welgus H G , Fliszar CJ, Veith G M , Jeffrey JJ, and Gluck SL, Initiation of osteoclast bone resorption by interstitial collagenase. J Biol Chem, 1997. 272(35): p. 22053-8. K h a n K M , M c K a y H A , Kannus P, Bailey D A , Wark J D , and Bennell K L , Physical Activity and Bone Health. 2001, Champaign: Human Kinetics. Baron R E , Anatomy and Ultrastructure of Bone, in Primer on the Metabolic Bone Diseases and Dis ofMineral Metabolism, M J Favus, Editor. 1996, Lipincott-Raven: Philadelphia, p. 3-10. Baron R, Chakraborty M , Chatterjee D , H o m e W , Lomri A , and Raveslooot J - H , Biology of the Osteoclast, in Physiology and Pharmacology of Bone, T J Martin, Editor. 1993, Springer-Verlag: N e w L i Z , H o u W S , and Bromme D ,  York. p. 111-47. 29.  Tortora G J and Grabowsld SR, Principles  of Anatomy and Principles.  9th ed. 1999, N e w York:  WilleyJohn&Son. 30. 31.  Embryology and Cellular Biology of Bone, in Metabolic bone disease & Clinically Related Disorders, S M Krane, Editor. 1990, W B Saunders Publisher: Philadelphia. Barr SI and M c K a y H A , Nutrition, exercise, and bone status inyouth. Int J Sport Nutr, 1998. 8(2): p. Raisz L G and Roden G A ,  124-42.  32. 33.  The bone-remodeling transient: implicationsfor the interpretation of clinical studies of change. J Bone Miner Res, 1994. 9(10): p. 1515-23. Holbrook T L , Barrett-Connor E , Klauber M , and Sartoris D , A population-based comparison of quantitative dual-energy X-ray absorptiometry with dual-photon absorptiometry of the spine and Heaney RP,  Tissue Int, 1991. 49(5): p. 305-7. 34.  Carter D R and O r r T E ,  Skeletal development and bonefunctional adaptation. J Bone  Miner Res, 1992.  7 S u p p l 2: p. S389-95. 35. 36.  Whole body bone mineral content in healthy children and adolescents. Arch Dis Child, 1997. 76(1): p. 9-15. Kroger H , Kotaniemi A , Kroger L , and Alhava E , Development of bone mass and bone density of the spine andfemoralneck-aprospective study of 65 children and adolescents. Bone Miner, 1993. 23(3): p  Molgaard C , Thomsen B L , Prentice A , Cole T J , and Michaelsen K F ,  171-82.  37.  Comparison of different modelsfor interpreting bone mineral density measurements using DXA andMRI technology. Bone, 1995.17(2): p. 157-9.  Kroger H , Vainio P, Nieminen J , and Kotaniemi A , 139  38. 39.  Katzman D K , Bachrach L K , Carter D R , and Marcus R, Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls, J Clin Endocrinol Metab, 1991. 73(6): p. 1332-9. Theintz G , Buchs B , Rizzoli R, Slosman D , Clavien H , Sizonenko P C , and Bonjour JP,  Longitudinal monitoring of bone mass accumulation in healthy adolescents: evidencefor a marke 16years of age at the levels of lumbar spine andfemoral neck in female subjects. J Clin Endocrinol 1992. 75(4): p. 1060-5.  40.  41. 42.  A six-year longitudinal study of the relationship ofphysical activity to bone mineral accrual ingrowing children: the university o Saskatchewan bone mineral accrual study. J Bone Miner Res, 1999. 14(10): p. 1672-9. Mosekilde L , Sex differences in age-related changes in vertebral body si%e, density and biomech competence in normal individuals. Bone, 1990.11(2): p. 67-73. N e u C M , M a n z F , Rauch F , Merkel A , and Schoenau E , Bone densities and bone si%e at the distal radius in healthy children and adolescents: a study usingperipheral quantitative computed tomog Bailey D A , M c K a y H A , Mirwald R L , Crocker PR, and Faulkner R A ,  2001.28(2): p. 227-32. 43. 44.  An Assessment of maturity from anthropometric measurements. M e d Sci Sports Exerc, 2002. 34(4): p. 689-94. Forwood M R , Mechanical effects on the skeleton: are there clinical implications'? Osteoporos  Mirwald R, Baxter-Jones A D , Bailey D A , and Beunen G P ,  12(1): p. 77-83. 45. 46.  47. 48. 49.  Beck T J , Ruff C B , Warden K E , Scott W W , Jr., and Rao G U ,  bone mineral data. A structural approach.  Int, 20  Predictingfemoral neck strength from  Invest Radiol, 1990. 25(1): p. 6-18.  A randomised school-basedjumping interventionconfers site and maturity-specific benefits on bone structuralpr A hip structure analysis (HSA) study. J Bone Miner Res, 2002.17(3): p. 363-72. Jones H H , Priest J D , Hayes W C , Tichenor C C , and Nagel D A , Humeral hypertrophy in response to exercise. J Bone Joint Surg A m , 1977. 59(2): p. 204-8. Murphy S, K h a w K T , May H , and Compston J E , Milk consumption and bone mineral density in middle aged and elderly women. Bmj, 1994. 308(6934): p. 939-41. Nieves J W , G o l d e n A L , Siris E , Kelsey J L , and Lindsay R, Teenage and current calcium intake are related to bone mineral density of the hip andforearm in women aged 30-39'years. A m J Epidem Petit M A , M c K a y H A , Mackelvie K J , Heinonen A , K h a n K M , and Beck T J ,  141(4): p. 342-51.  50. 51. 52.  Bone mineral content of two populations of Chinese children with different calcium intakes. Bone Miner, 1993. 23(3): p. 195-206. Lee W T , Leung SS, Lui SS, and Lau J , Relationship between long-term calcium intake and bone miner content of children agedfrom birth to 5years. B r J Nutr, 1993. 70(1): p. 235-48. Matkovic V and Heaney R P , Calcium balance during human growth: evidencefor threshold behavior. Lee W T , Leung SS, N g M Y , Wang SF, X u Y C , Zeng W P , and Lau J ,  J Clin Nutr, 1992. 55(5): p. 992-6.  53.  Eating attitudes and habitual calcium intake in peripubertalgirls are associated with initial bone mineral content and its change over 2years. J Barr SI, Petit M A , Vigna Y M , and Prior J C ,  Bo  Res, 2001.16(5): p. 940-7. 54.  Bonjour JP, Carrie A L , Ferrari S, Clavien H , Slosman D , Theintz G , and Rizzoli R, Calciumenrichedfoods and bone mass growth in prepubertalgirls: a randomised, double-blind, placebo-co Clin Invest, 1997. 99(6): p. 1287-94.  55.  Bonjour JP, Chevalley T , A m m a n n P, Slosman D , and Rizzoli R, Gain in bone mineral mass in prepubertalgirls 3.5years after discontinuation of calcium supplementation: afollow-up study. 358(9289): p. 1208-12. 140  Lan  56.  57. 58.  Johnston C C , Jr., Miller J Z , Slemenda C W , Reister T K , H u i S, Christian J C , and Peacock M , Calcium supplementation and increases in bone mineral density in children. N Engl J Med, 1992. 3 82-7. Recker RR, Davies K M , Hinders S M , Heaney RP, Stegman M R , and Kimmel D B ,  young adult women. Jama,  N o w s o n C A , Green R M , Hopper J L , Sherwin A J , Y o u n g D , Kaymakci B , Guest CS, Smid M , Larkins R G , and Wark J D ,  adolescence.  59.  60 61.  Bone gain in  1992. 268(17): p. 2403-8.  A. co-twin study of the effect of calcium supplementation on bone density  Osteoporos Int, 1997. 7(3): p. 219-25.  Lee W T , Leung SS, Wang S H , X u Y C , Zeng W P , Lau J, Oppenheimer SJ, and Cheng J C ,  Double-blind, controlled calcium supplementation and bone mineral accretion in children accusto calcium diet. A m J Clin Nutr, 1994. 60(5): p. 744-50. Lee W T , Leung SS, Leung D M , and Cheng ]C,A follow-up study on the effects of calcium-supplement withdrawal andpuberty on bone acquisition of children. A m J Clin Nutr, 1996. 64(1): p. 71 -7. Rubin C T and Lanyon L E , Regulation of bone formation by applied dynamic loads. J Bone Joint Surg A m , 1984. 66(3): p. 397-402.  62. 63.  Rubin C T and Lanyon L E , 1985. 37(4): p. 411-7. Marcus R,  Regulation of bone mass by mechanical strain magnitude.  Mechanisms of exercise effects on bone., in Principles ofBone Biology,  Calcif Tissue Int  L R Bilezikian and  Rodan, G , Editors. 1996, Academic Press: N e w York. p. 1135-46.  64. 65. 66.  Strain rate as a controlling influence on adaptive modelling in response t dynamic loading of the ulna in growing male rats. Bone, 1998. 23(4): p. 313-8. Lanyon L E , Usingfunctional loading to influence bone mass and architeture. Objectives, mechani relationships with estrogen of the mechanically adaptive process in bone. Bone, 1996.18: p. 37S-43S Torrance A G , Suswillo R F L , and Lanyon L E , Non-invasive loading of the rat ulna in vivo induces strain-related modeling response uncomplicated by trauma orperiostealpressure. Calcif Tissue Int, Mosley J R and Lanyon L E ,  54(241-47). 67. 68.  Lanyon L E , Goodship A E , Pye CJ, and MacFie J H , Biomech, 1982.15(3): p. 141-54.  Mechanically adaptive bone remodelling. J  Haapasalo H , Kannus P, Sievanen H , Pasanen M , Uusi-Rasi K , Heinonen A , Oja P, and Vuori  I, Development of mass, density, and estimated mechanical characteristics of bones in Caucasian Bone Miner Res, 1996.11(11): p. 1751-60.  69.  70. 71.  Genetic determinants of bone mass in adult women: a reevaluation of the twin model and the potential importance of ge on heritability estimates. J Bone Miner Res, 1991. 6(6): p. 561-7. Frost H M , Bone "mass" and the "mechanostat": a proposal. Anat Rec, 1987. 219(1): p. 1 -9. Lanyon L E , XJsingfunctional loading to influence bone mass and architecture: objectives, mechan relationship with estrogen of the mechanically adaptive process in bone. Bone, 1996.18(1 Suppl): p. Slemenda C W , Christian J C , Williams CJ, Norton J A , and Johnston C C , Jr.,  43S. 72.  Cassell C , Benedict M , and Specker B,  swimmers.  Bone mineral density in elite 7- to 9yr-oldfemale gymnasts and  M e d Sci Sports Exerc, 1996. 28(10): p. 1243-6.  Gymnastic training and bone  73.  Dyson K , Blimkie CJ, Davison K S , Webber C E , and Adachi J D ,  74.  K h a n K M , Bennell K L , Hopper J L , Flicker L , Nowson C A , Sherwin A J , Crichton K J ,  density inpre-adolescentfemales.  M e d Sci Sports Exerc, 1997. 29(4): p. 443-50.  Self-reported ballet classes undertaken at age 10-12years and hip bone mineral density in later life. Osteoporos Int, 1998. 8(2): p. 165-73. Harcourt PR, and Wark J D ,  141  75.  Conroy B P , Kraemer W J , Maresh C M , Fleck SJ, Stone M H , Fry A C , Miller P D , and Dalsky  G P , Bone mineral density in elitejunior Olympic iveightlifters. M e d Sci Sports Exerc, 1993. 25(10): p. 1103-9. 76.  Bass S, Pearce G , Bradney M , Hendrich E , Delmas P D , Harding A , and Seeman E , Exercise  before puberty may confer residual benefits in bone density in adulthood: studies in active prepubertal and retired female gymnasts. J Bone Miner Res, 1998.13(3): p. 500-7. 77.  Morris F L , Naughton G A , Gibbs J L , Carlson JS, and W a r k J D , Prospective ten-month exercise  intervention inpremenarchealgirls:positive effects on bone and lean mass. J Bone Miner Res, 1997. 12(9): p. 1453-62.  78. 79.  Fuchs R K , Bauer JJ, and Snow CM, Jumping improves hip and lumbar spine bone mass in prepubescent children: a randomised controlled trial. J Bone Miner Res, 2001. 16(1): p. 148-56. M c K a y H A , Petit M A , Schutz R W , Prior J C , Barr SI, and K h a n K M , Augmented trochanteric bone  mineral density after modifiedphysical education classes: a randomised school-based exercise intervention study in prepubescent and early pubescent children. J Pediatr, 2000. 136(2): p. 156-62. 80. 81.  Fuchs R K and Snow C M , Gains in hip bone massfrom high-impact training are maintained: a randomised controlled trial in children. J Pediatr, 2002. 141(3): p. 357-62. Kannus P, Haapasalo H , Sankelo M , Sievanen H , Pasanen M , Heinonen A , Oja P, and Vuori I,  Effect of starting age ofphysical activity on bone mass in the dominant arm of tennis and squash players. A n n Intern M e d , 1995.123(1): p. 27-31. 82.  Heinonen A , Sievanen H , Kannus P, Oja P, Pasanen M , and V u o r i I, High-impact exercise and  83.  Bradney M , Pearce G , Naughton G , Sullivan C , Bass S, Beck T , Carlson J , and Seeman E ,  bones of growing girls: a 9-month controlled trial. Osteoporos Int, 2000.11(12): p. 1010-7. Moderate exercise during growth in prepubertal boys: changes in bone mass, sise, volumetric density, and bone strength: a controlledprospective study. J Bone Miner Res, 1998.13(12): p. 1814-21. 84.  MacKelvie K J , M c K a y H A , Petit M A , Moran O , and K h a n K M , Bone mineral response to a 7-  month randomised controlled, school-basedjumping intervention in 121 prepubertal boys: associations with ethnicity and body mass index. J Bone Miner Res, 2002.17(5): p. 834-44. 85.  Mackelvie K J , M c K a y H A , Khan K M , and Crocker PR, A school-based exercise intervention  augments bone mineral accrual in early pubertalgirls. J Pediatr, 2001.139(4): p. 501-8. 86.  Bloomfield SA, Changes in musculoskeletal structure andfunction with prolonged bed rest. M e d Sci Sports Exerc, 1997. 29(2): p. 197-206.  87.  Lanyon L E , Rubin C T , and Baust G , Modulation of bone loss during calcium insufficiency by controlled dynamic loading. Calcif Tissue Int, 1986. 38(4): p. 209-16.  88.  Gunnes M and Lehmann E H , Physical activity and dietary constituents as predictors offorearm cortical and trabecular bone gain in healthy children and adolescents: a prospective study. Acta Paediatr, 1996. 85(1): p. 19-25.  89.  Slemenda C W , Reister T K , H u i SL, Miller J Z , Christian J C , and Johnston C C , Jr., Influences on  skeletal mineralisation in children and adolescents: evidencefor varying effects of sexual maturation andphysical activity. J Pediatr, 1994.125(2): p. 201 -7. 90.  Dequeker J, Nijs J, Verstraeten A , Geusens P, and Gevers G , Genetic determinants of bone mineral  91.  Pocock N A , Eisman J A , Hopper J L , Yeates M G , Sambrook P N , and Eberl S, Genetic  content at the spine and radius: a twin study. Bone, 1987. 8(4): p. 207-9. determinants of bone mass in adults. A Win study. J Clin Invest, 1987. 80(3): p. 706-10. 92.  Christian J C , Y u P L , Slemenda C W , and Johnston C C , Jr., Heritability of bone mass: a longitudinal ~ " ' in aging male twins. A m J H u m Genet, 1989. 44(3): p. 429-33. J  J  142  93.  Y o u n g D , Hopper J L , N o w s o n C A , Green R M , Sherwin A J , Kaymakci B , Smid M , Guest CS,  Larkins R G , and Wark J D , Determinants of bone mass in 10- to 26-year-oldfemales: a twin study. J Bone Miner Res, 1995.10(4): p. 558-67.  94.  Tylavsky F A , Bortz A D , Hancock R L , and Anderson JJ, Familial resemblance of radial bone mass between premenopausal mothers and their college-age daughters. Calcif Tissue Int, 1989. 45(5): p. 265-7  95.  Krall E A and Dawson-Hughes B, Heritable and life-style determinants of bone mineral density. J Bone  96.  Nguyen T V , Howard G M , Kelly PJ, and Eisman J A , Bone mass, lean mass, andfat mass: same genes or same environments? Km] Epidemiol, 1998.147(1): p. 3-16.  Miner Res, 1993. 8(1): p. 1-9.  97.  Kelly PJ, Hopper J L , Macaskill G T , Pocock N A , Sambrook P N , and Eisman J A , Geneticfactors in bone turnover. J Clin Endocrinol Metab, 1991. 72(4): p. 808-13.  98.  Tokita A , Kelly PJ, Nguyen T V , Q i J C , Morrison N A , Risteli L , Risteli J , Sambrook P N , and  Eisman J A , Genetic influences on type I collagen synthesis and degradation: further evidencefor geneti regulation of bone turnover. J Clin Endocrinol Metab, 1994. 78(6): p. 1461-6. 99.  Harris M , Nguyen T V , Howard G M , Kelly PJ, and Eisman J A , Genetic and environmental  correlations between boneformation and bone mineral density: a twin study. Bone, 1998. 22(2): p. 141 100.  Kelly PJ, Nguyen T , Hopper J , Pocock N , Sambrook P, and Eisman J , Changes in axial bone  density with age: a twin study. J Bone Miner Res, 1993. 8(1): p. 11-7. 101.  Gueguen R, Jouanny P, Guillemin F , Kuntz C , Pourel J, and Siest G , Segregation analysis and  variance components analysis of bone mineral density in healthyfamilies. J Bone Miner Res, 1995. 10(12  p. 2017-22. 102.  Strachan T and Read A P , Human Molecular Genetics 2. 1999, N e w York: Willey-Liss.  103.  Morrison N A , Yeoman R, Kelly PJ, and Eisman J A , Contribution of trans-actingfactor alleles to  normalphysiological variability: vitamin D receptor gene polymorphism and circulating osteocalcin. Pro Acad Sci U S A , 1992. 89(15): p. 6665-9. 104.  Morrison N A , Q i J C , Tokita A , Kelly PJ, Crofts L , Nguyen T V , Sambrook P N , and Eisman  J A , Prediction of bone densityfrom vitamin D receptor alleles. Nature, 1994. 367(6460): p. 284-7. 105.  Cooper G S and Umbach D M , Are vitamin D receptorpolymorphisms associated with bone mineral density? A meta-analysis. J Bone Miner Res, 1996.11(12): p. 1841-9.  106.  Q i J C , Morrison N A , Nguyen T V , White C P , Kelly PJ, Sambrook P N , Eisman J A , and :,  Oestrogen receptor genotypes and bone mineral density in women and men. The Journal of Bone & Mi Research, 1994.10(S170). 107.  Grant SF, Reid D M , Blake G , Herd R, Fogelman I, and Ralston S H , Reduced bone density and  osteoporosis associated with a polymorphic Sp1 binding site in the collagen type I alpha 1 gene. Nat 1996.14(2): p. 203-5. 108.  Kobayashi S, Inoue S, Hosoi T , Ouchi Y , Shiraki M , and O r i m o H , Association of bone mineral  density with polymorphism of the estrogen receptorgene. J Bone Miner Res, 1996.11(3): p. 306-11. 109.  Keen R W , Woodford-Richens K L , Lanchbury JS, and Spector T D , Allelic variation at the  interleukin-1 receptor antagonist gene is associated with early postmenopausal bone loss at the spine 1998. 23(4): p. 367-71. 110.  Rosen CJ, Kurland E S , Vereault D , Adler R A , Rackoff PJ, Craig W Y , Witte S, Rogers J, and  Bilezikian JP, Association between serum insulin growthfactor-1 (IGF-I) and a simple sequence repeat IGF-Igene: implicationsforgenetic studies of bone mineral density. J Clin Endocrinol Metab, 1998. 83(7): p. 2286-90. 111.  Koller D L , E c o n s M J , M o r i n P A , Christian J C , H u i SL, Parry P, Curran M E , Rodriguez L A , Conneally P M , Joslyn G , Peacock M , Johnston C C , and Foroud T , Genome screenfor QTLs 143  contributing to normal variation in bone mineral density and osteoporosis. J Clin Endocrinol Metab, 2000. 85(9): p. 3116-20.  112.  113.  Gong Y, Vikkula M , Boon L, Liu j , Beighton P, Ramesar R, Peltonen L, Somer H , Hirose T, Dallapiccola B, De Paepe A, Swoboda W, Zabel B, Superti-Furga A, Steinmann B, Brunner H G , Jans A , Boles R G , Adkins W, van den Boogaard MJ, Olsen BR, and Warman ML, Osteoporosis-pseudoglioma syndrome, a disorder affecting skeletal strength and vision, is assigned to chr region 11 q12-13. A m J Hum Genet, 1996. 59(1): p. 146-51. Johnson M L , Gong G , Kimberling W, Recker SM, Kimmel D B , and Recker RB, Linkage of a gene causing high bone mass to human chromosome 11 (11q12-13). A m J Hum Genet, 1997. 60(6): p. 1326-32.  114.  115.  116.  Heaney C, Shalev H , Elbedour K, Carmi R, StaackJB, Sheffield V C , and Beier DR, Human autosomal recessive osteopetrosis maps to 11q13, a position predicted by comparative mapping of the m osteosclerosis (oc) mutation. Hum Mol Genet, 1998. 7(9): p. 1407-10. Deng HW, X u F H , Conway T, Deng X T , Li JL, Davies K M , Deng H , Johnson M , and Recker RR, Is population bone mineral density variation linked to the marker D11S987 on chromosome 11q12-13 J Clin Endocrinol Metab, 2001. 86(8): p. 3735-41. Niu T, Chen C, Cordell H , Yang J, Wang B, Wang Z, Fang Z, Schork NJ, Rosen CJ, and Xu X , A genome-wide scan for loci linked to forearm bone mineral density. Hum Genet, 1999. 104(3): p. 22633.  117.  11.8.  Zee RY, Myers R H , Hannan MT, Wilson PW, Ordovas J M , Schaefer EJ, Lindpaintner K, and Kiel DP, Absence of linkagefor bone mineral density to chromosome 12q12-14 in the region of the vitam receptor gene. Calcif Tissue Int, 2000. 67(6): p. 434-9. Devoto M , Specchia C, Li H H , Caminis J, Tenenhouse A, Rodriguez H , and Spotila L D , Variance component linkage analysis indicates aQTL forfemoral neck bone mineral density on chromosom 1p36. Hum M o l Genet, 2001.10(21): p. 2447-52.  119.  120. 121. 122.  123.  124.  Devoto M , Shimoya K, Caminis J, Ott J, Tenenhouse A , Whyte MP, Sereda L, Hall S, Considine E , Williams CJ, Tromp G , Kuivaniemi H , Ala-Kokko L, Prockop DJ, and Spotila L D , First-stage autosomalgenome screen in extendedpedigrees suggests genes predisposing to low bone density on chromosomes 1p, 2p and 4q. Eur J Hum Genet, 1998. 6(2): p. 151-7. Klein RF, Mitchell SR, Phillips TJ, Belknap J K , and Orwoll US, Quantitative trait loci affectingpeak bone mineral density in mice. J Bone Miner Res, 1998.13(11): p. 1648-56. Grupe A, Germer S, Usuka J, Aud D, Belknap JK, Klein RF, Ahluwalia M K , Higuchi R, and Peltz G , In silico mapping of complex disease-related traits in mice. Science, 2001. 292(5523): p. 1915-8. Shimizu M , Higuchi K , Bennett B, Xia C, Tsuboyama T, Kasai S, Chiba T, Fujisawa H , Kogishi K, Kitado H , Kimoto M , Takeda N , Matsushita M , Okumura H , Serikawa T, Nakamura T, Johnson T E , and Hosokawa M , Identification ofpeak bone massQTL in a spontaneously osteoporotic mouse strain. Mamm Genome, 1999.10(2): p. 81-7. Beamer W G , Shultz K L , Donahue LR, Churchill G A , Sen S, Wergedal JR, Baylink DJ, and Rosen CJ, Quantitative trait lociforfemoral and lumbar vertebral bone mineral density in C57BL/ 6] and C3H/HeJ inbred strains of mice. J Bone Miner Res, 2001.16(7): p. 1195-206. Tao C, Y u T, Garnett S, BriodyJ, Knight J, Woodhead H , and Cowell CT, Vitamin D receptor alleles predict growth and bone density in girls. Arch Dis Child, 1998. 79(6): p. 488-93; discussion 934.  125.  Baker AR, McDonnell D P , Hughes M , Crisp T M , Mangelsdorf DJ, Haussler MR, Pike JW, Shine J, and O'Malley BW, Cloning and expression of full-length cDNA encoding human vitamin D receptor. Proc Nad Acad Sci U S A , 1988. 85(10): p. 3294-8. 144  126.  Miyamoto K , Kesterson R A , Yamamoto H , Taketani Y , Nishiwaki E , Tatsumi S, Inoue Y , Morita K , Takeda E , and Pike J W ,  gene and its promoter. 127.  Structural organisation of the human vitamin D receptor chromos  M o l Endocrinol, 1997.11(8): p. 1165-79.  Mocharla H , Butch A W , Pappas A A , Flick J T , Weinstein RS, D e Togni P, Jilka R L , Roberson  SC, Quantification of vitamin D receptor mRNA by competitive polymerase chain reaction in PBMC: lack of correspondence with common allelic variants, J P K , Parfitt A M , and Manolagas  Bone  1997.12(5): p. 726-33. 128. 129.  Durrin L K , Haile R W , Ingles SA, and Coetzee G A , Vitamin D receptor 3'-untranslated region polymorphisms: lack of effect on mRNA stability. Biochim Biophys Acta, 1999.1453(3): p. 311-20. Spotila L D , Caminis J, Johnston R, Shimoya K S , O'Connor M P , Prockop D J , Tenenhouse A ,  Vitamin D receptor genotype is not associated ivith bone mineral density in ethnic/ regionalgroups. Calcif Tissue Int, 1996. 59(4): p. 235-7. K u n g A W , Yeung SS, and Lau K S , Vitamin D receptor gene polymorphisms andpeak bone mass in southern Chinese women. Bone, 1998. 22(4): p. 389-93. L i m S K , Park Y S , Park J M , Song Y D , Lee E J , K i m K R , Lee H C , and H u h K B , Lack of association betiveen vitamin D receptorgenotypes and osteoporosis in Koreans. J Clin Endocrino and Tenenhouse H S ,  130. 131.  1995. 80(12): p. 3677-81. 132.  Tokita A , Matsumoto H , Morrison N A , Tawa T , Miura Y , Fukamauchi K , Mitsuhashi N , Irimoto M , Yamamori S, Miura M , Watanabe T , Kuwabara Y , Yabuta K , and Eisman J A ,  Vitamin D receptor alleles, bone mineral density and turnover in premenopausal Japanese ivom  Miner Res, 1996.11(7): p. 1003-9. 133.  134.  Tamai M , Yokouchi M , Komiya S, Mochizuki K , Hidaka S, Narita S, Inoue A , and Itoh K ,  Correlation between vitamin D receptor genotypes and bone mineral density in Japanese patien osteoporosis. Calcif Tissue Int, 1997. 60(3): p. 229-32. Bell N H , Morrison N A , Nguyen T V , Eisman J, and Hollis B W , Apal polymorphisms of the vitamin D receptorpredict bone density of the lumbar spine and not racial difference in bone density in  Lab Clin Med, 2001. 137(2): p. 133-40. 135.  Grisso J A , Kelsey J L , Strom B L , O'Brien L A , Maislin G , LaPann K , Samelson L , and Hoffman S,  Risk factorsfor hip fracture in black women. The Northeast Hip Fracture Study Group  E n g l J M e d , 1994. 330(22): p. 1555-9. 136.  Fleet J C , Harris SS, W o o d RJ, and Dawson-Hughes B, The Bsml vitamin D receptor restriction fragment length polymorphism (BB) predicts low bone density in premenopausal black and whit Bone Miner Res, 1995.10(6): p. 985-90.  Vitamin D receptorpolymorphism, bone mineral density, and osteoporotic vertebral fracture: studies in a UKpopulation. Bone, 1996.18(3): p. 249-5  137. •  Houston L A , Grant SF, Reid D M , and Ralston S H ,  138.  Ensrud K E , Stone K , Cauley J A , White C , Z m u d a J M , Nguyen T V , Eisman J A , and  Vitamin D receptor gene polymorphisms and the risk of fractures in older women. Study of Osteoporotic Fractures Research Group. J Bone Miner Res, 1999.14(10): p. 1637-45. Sainz J , V a n Tornout J M , Loro M L , Sayre J, Roe T F , and Gilsanz V , Vitamin D-receptorgene polymorphisms and bone density in prepubertal American girls ofMexican descent. N E n g l J Med, Cummings SR,  139.  337(2): p. 77-82. 140.  Vitamin D receptor gene polymorphism is associated with birth height, growth to adolescence, and adult stature in healthy Caucasian men: a cross-se longitudinal study. J Clin Endocrinol Metab, 2000. 85(4): p. 1666-70.  Lorentzon M , Lorentzon R, and Nordstrom P,  145  141.  Gunnes M , BergJP, Halse J, and Lehmann E H , Lack of relationship between vitamin D receptor genotype andforearm bone gain in healthy children, adolescents, and young adults. J Clin Endoc  142.  Vitamin-D receptor genotype does notpredict bone mineral density, bone turnover, and growth in prepube  1997. 82(3): p. 851-5. Baroncelli G l , Federico G , Bertelloni S, Ceccarelli C , Cupelli D , and Saggese G ,  H o r m Res, 1999. 51(3): p. 150-6. 143.  Arai H , Miyamoto K , Taketani Y , Yamamoto H , Iemori Y , Morita K , Tonai T , Nishisho T , M o r i S, and Takeda E , A vitamin D receptorgene polymorphism in the translation initiation codon: protein activity and relation to bone mineral density in Japanese women. J Bone Miner Res, 1997.1 915-21.  144.  Harris SS, Eccleshall T R , Gross C , Dawson-Hughes B , and Feldman D , The vitamin D receptor start codon polymorphism (Fokl) and bone mineral density in premenopausal American black an women. J Bone Miner Res, 1997.12(7): p. 1043-8. Ferrari S, Rizzoli R, Manen D , Slosman D , and Bonjour JP, Vitamin D receptorgene start codon polymorphisms (Fokl) and bone mineral density: interaction with age, dietary calcium, and 3'-e polymorphisms. J Bone Miner Res, 1998.13(6): p. 925-30. Lucotte G , Mercier G , and Burckel A , Fok I polymorphism at the human vitamin D receptor gene loc in Europeans and Africans. Clin Genet, 1999. 55(4): p. 281-2. Ames S K , Ellis K J , G u n n S K , Copeland K C , and Abrams SA, Vitamin D receptor gene Fokl polymorphism predicts calcium absorption and bone mineral density in children. J Bone Miner Res  145.  146. 147.  14(5): p. 740-6. 148.  Bone mineral mass and calcium and phosphate metabolism inyoung men: relationships with vitamin D receptor allelic polymorphisms Ferrari S, Manen D , Bonjour JP, Slosman D , and Rizzoli R,  Endocrinol Metab, 1999. 84(6): p. 2043-8. 149.  Salamone L M , Glynn N W , Black D M , Ferrell R E , Palermo L , Epstein RS, Kuller L H , and Cauley J A , Determinants ofpremenopausal bone mineral density: the interplay of genetic Bone Miner Res, 1996.11(10): p. 1557-65.  150.  Tsuritani I, Brooke-Wavell K S , Mastana SS, Jones PR, Hardman A E , and Yamada Y ,  and lifes  Does vitamin D receptorpolymorphism influence the response of bone to brisk walking in postmenopau  H o r m Res, 1998. 50(6): p. 315-9. 151.  152. 153. 154.  Blanchet C , Giguere Y , Prud'homme D , Dumont M , Rousseau F, and D o d i n S, Association of Med Sci Sports Exerc, 2002. 34 24-31.  physical activity and bone: influence of vitamin D receptor genotype. Einhorn T A ,  Bone strength: the bottom line.  Calcif Tissue Int, 1992. 51(5): p. 333-9.  Knott L , Whitehead C C , Fleming R H , and Bailey A J , Biochemical changes Biochem J, 1995. 310 ( Pt 3): p. 1045-51.  osteoporotic avian bone.  in the collagenous matrix of  Mechanic G L , Katz E P , Henmi M , Noyes C , and Yamauchi M , Locus of a histidine-based, stable trifunctional, helix to helix collagen cross-link: stereospecific collagen structure oftypeI skin fibrils Biochemistry, 1987. 26(12): p. 3500-9.  155. 156.  157.  Yamaguchi T , Tamaki K , Kaufman H E , Katz J, and Shaw E L , Histologic study of a pair of human corneas after anterior radial keratotomy. A m J Ophthalmol, 1985.100(2): p. 281-92. Mann V , H o b s o n E E , L i B , Stewart T L , Grant SF, Robins SP, Aspden R M , and Ralston S H ,  A COL1A1 Sp1 binding site polymorphism predisposes to osteoporoticfracture by affecting bone quality. J Clin Invest, 2001.107(7): p. 899-907. Uitterlinden A G , Burger H , Huang Q , Yue F, McGuigan F E , Grant SF, Hofman A , van Leeuwen JP, Pols H A , and Ralston S H ,  Relation of alleles of the collagen type Ialphal gene to bone 146  density and the risk of osteoporoticfractures in postmenopausal women.  N E n g l J M e d , 1998. 338(1  1016-21. 158.  Garnero P, Borel O , Grant SF, Ralston S H , and Delmas P D , Collagen Ialphal Sp1 polymorphism, bone mass, and bone turnover in healthy French premenopausal women: the OFELY study. J  B  Res, 1998.13(5): p. 813-7.  159. 160.  Roux C , Dougados M , Abel L , Mercier G , and Lucotte G , Association of a polymorphism in the collagen I alphal gene ivith osteoporosis in French women. Arthritis Rheum, 1998. 41(1): p. 187-8. Peris P, Alvarez L , Oriola J, Guanabens N , Monegal A , de Osaba M J , Jo J, Pons F , Ballesta A M , and M u n o z - G o m e z J,  CollagentypeIalphal gene polymorphism in idiopathic osteoporosis in me  Rheumatology (Oxford), 2000. 39(11): p. 1222-5. 161.  162.  Keen R W , Woodford-Richens K L , Grant SF, Ralston S H , Lanchbury JS, and Spector T D ,  Association ofpolymorphism at thetypeI collagen (COL1A1) locus with reduced bone mineral den increasedfracture risk, and increased collagen turnover. Arthrids Rheum, 1999. 42(2): p. 285-90. Alvarez L , Oriola J , Jo J , Ferro T , Pons F , Peris P, Guanabens N , Duran M , Monegal A ,  CollagentypeI alphal gene Sp1 polymorphism in premenopausal women with primary osteoporosis: improved detection ofSpl bin polymorphism in the collagentype1 gene. Clin Chem, 1999. 45(6 Pt 1): p. 904-6. McGuigan F E , Reid D M , and Ralston S H , Susceptibility to osteoporoticfracture is determined by alle variation at the Sp1 site, rather than otherpolymorphic sites at the COL1A1 locus. Osteoporos I Martinez de Osaba M J , Rivera-Fillat F, and Ballesta A M ,  163.  11(4): p. 338-43. 164.  Qureshi A M , McGuigan F E , Seymour D G , Hutchison J D , Reid D M , and Ralston S H ,  Association between COLIA1 Sp1 alleles andfemoral neck geometry.  Calcif Tissue Int, 2001. 69(2):  67-72.  165. 166.  Polymorphism of the collagen type Ialphal gene and ethnic differences in hip-fracture rates. N E n g l J Med, 1998. 339(5): p. 351-2. Hustmyer F G , L i u G , Johnston C C , Christian J, and Peacock M , Polymorphism at an Sp1 binding site of COL1A1 and bone mineral density in premenopausalfemale twins and elderly fracture pati Beavan S, Prentice A , Dibba B, Y a n L , Cooper C , and Ralston S H ,  Osteoporos Int, 1999. 9(4): p. 346-50.  167. 168.  Sainz J, V a n Tornout J M , Sayre J, Kaufman F, and Gilsanz V , Association of collagentype1 alphal gene polymorphism with bone density in early childhood. J Clin Endocrinol Metab, 1999. 84(3): p. 853 Slack J L , Liska D J , and Bornstein P, Regulation of expression of thetypeI collagen genes. A m J Med Genet, 1993. 45(2): p. 140-51.  169.  Pavlin D , Bedalov A , Kronenberg M S , Kream B E , Rowe D W , Smith C L , Pike J W , and  Analysis of regulatory regions in the COL1A1 gene responsiblefor 1,25-dihydroxyvita mediated transcriptional repression in osteoblastic cells. J Cell Biochem, 1994. 56(4): p. 490-501. Lichder A C ,  170.  Uitterlinden A G , Weel A E , Burger H , Fang Y , van Duijn C M , H o f m a n A , van Leeuwen JP,  Interaction between the vitamin D receptor gene and collagentypeIalphal gene in s for fracture. J Bone Miner Res, 2001.16(2): p. 379-85. and Pols H A ,  171.  Valimaki S, Tahtela R, Kainulainen K , Laitinen K , Loyttyniemi E , Sulkava R, Valimaki M , and  Relation of collagentypeI alpha 1 (COLIA 1) and vitamin D receptor genotypes to bon turnover, andfractures in early postmenopausal womenand to hip fractures in elderly people. 20 Kontula K , 48-56.  172.  McGuigan F E , Murray L , Gallagher A , Davey-Smith G , Neville C E , Van't H o f R, Boreham C , and Ralston S H ,  Genetic and environmental determinants ofpeak bone mass inyoung men and w  Bone Miner Res, 2002.17(7): p. 1273-9. 147  173.  Smith C A , Davis T , Anderson D , Solam L , Beckmann M P , Jerzy R, Dower S K , Cosman D , and G o o d w i n R G , A receptorjor tumor necrosisjactor defines an unusualfamily oj cellular and vira proteins. Science, 1990. 248(4958): p. 1019-23.  174.  A b u - A m e r Y , Erdmann J , Alexopoulou L , Kollias G , Ross F P , and Teitelbaum SL,  Tumor necrosisjactor receptors types 1 and 2 differentially regulate osteoclastogenesis. J Biol  Chem, 20  p. 27307-10. 175.  Kobayashi K , Takahashi N , Jimi E , Udagawa N , Takami M , Kotake S, Nakagawa N , Kinosaki M , Yamaguchi K , Shima N , Yasuda H , Morinaga T , Higashio K , Martin T J , and Suda T , Tumor  necrosisjactor alpha stimulates osteoclast differentiation by a mechanism independent oj the OD RANK interaction. J E x p M e d , 2000.191(2): p. 275-86. 176.  H s u H , Lacey D L , Dunstan C R , Solovyev I, Colombero A , Timms E , T a n H L , Elliott G , Kelley M J , Sarosi I, Wang L , X i a X Z , Elliott R, Chiu L , Black T , Scully S, Capparelli C , Morony S, Shimamoto G , Bass M B , and Boyle WJ, Tumor necrosisjactor receptorjamily  member RANK mediates osteoclast differentiation and activation induced by osteoprotegerin ligand.  Pro  Sci U S A , 1999. 96(7): p. 3540-5. 177.  178.  Spotila L D , Caminis J , Devoto M , Shimoya K , Sereda L , Ott J, Whyte M P , Tenenhouse A , and Prockop D J , Osteopenia in 37 members of sevenfamilies: analysis based on a model oj dominant M o l M e d , 1996. 2(3): p. 313-24. Spotila L D , Rodriguez H , K o c h M , Adams K , Caminis J, Tenenhouse H S , and Tenenhouse A ,  Association oj a polymorphism in the TNFR2 gene with low bone mineral density. J Bone  Miner Re  2000.15(7): p. 1376-83. 179.  180.  Albagha O M , Tasker P N , McGuigan F E , Reid D M , and Ralston S H , Linkage  disequilibrium between polymorphisms in the human TNFRSFIB gene and their association with bone mass in perimenopausal women. H u m M o l Genet, 2002.11(19): p. 2289-95. Rubin L A , Hawker G A , Peltekova V D , Fielding LJ, Ridout R, and Cole D E , Determinants oj peak bone mass: clinical and genetic analyses in ayoungjemale Canadian cohort. J Bone Miner Re 14(4): p. 633-43.  181. 182. 183.  RNAdraw: an integratedprogramjor RNA secondary stmcture calculation and analysis under 32-bit Microsoft Windows. Comput A p p l Biosci, 1996.12(3): p. 247-9. Hologic, ModelQDR-4500 User's Guide. 1996, Waltham, M A : Hologic, Inc. Crocker P R , Bailey D A , Faulkner R A , Kowalski K C , and M c G r a t h R, Measuring general levels oj physical activity: preliminary evidencejor the Physical Activity Questionnaire jor Older Children. Me Matzura O and Wennborg A ,  Sports Exerc, 1997. 29(10): p. 1344-9.  184.  Barr SI, Associations oj social and demographic Diet Assoc, 1994. 94(3): p. 260-6, 9; quiz 7-8.  Foetus into man.  185.  Tanner J M ,  186.  Duke P M , Litt IF, and Gross R T ,  variables with calcium intakes oj high school stud  1978, Cambridge: Harvard Press.  Adolescents' self-assessment of sexual maturation.  Pediatrics, 1980.  66(6): p. 918-20. 187. 188. 189.  McGuigan F E and Ralston S H , Single nucleotide polymorphism TaqMan. Psychiatr Genet, 2002.12(3): p. 133-6.  Amersham P B , DYEnamic ET Terminator Cycle Sweden: Amersham Pharmacia Biotech.  detection: allelic discrimination using  Sequencing Kit User's Manual.  2001, Uppsala,  Fernandez-Real J M , VendrellJ, Ricart W , Broch M , Gutierrez C , Casamitjana R, Oriola J, and Richart C , Polymorphism of the tumor necrosisfactor-alpha receptor 2 gene is associated with obe levels, and insulin resistance inyoung subjects and diet-treated type 2 diabetic patients. Diabete 23(6): p. 831-7. 148  190.  Lunneborg C E ,  Modelling Experimental and Observational Data.  Press. 191. 192.  Nelson D A and Barondess D A ,  Three Ethnic Groups.  1994, Belmont, C A : Duxbury  Whole Body Bone, Fat and Lean Mass in Children: Comparison of  A m J Physical Anthropology, 1997.103: p. 157-62.  Ogle G D , Allen JR, Humphries IRJ, L u P W , Briody J N , Morley K , Howman-Giles R, and Cowell C T ,  Body-composition assessment by dual-energy x-ray absorptiometry in subjects aged  Clin Nutr, 1995. 61: p. 746-53. 193.  Do geneticfactors explain associations between muscle strength, lean mass and bone density'? A twin study.  Seeman E , Hopper J , Y o u n g N , Formica C , Goss P, and Tsalamandris C ,  Am  1996. 270: p. E320-E7.  194.  Association of lean tissue andfat mass with bone mineral content in children and adolescents. Obes  Pietrobelli A , Faith M S , Wang J , Brambilla P, Chiumello G , and Heymsfield SB,  Res, 2002.  56-60.  195.  Total-body skeletal muscle mass: estimation by a new dual-energy X-ray absorptiometry method. A m J Clin Nutr, 2002. 76(2 K m J , Wang Z , Heymsfield SB, Baumgartner R N , and Gallagher D , 378-83.  196. 197. 198. 199. 200. 201. 202.  Timing and magnitude ofpeak height velocity andpeak tissue velocitiesfor early, average, and late maturing boys and girls. A m J H u m a n Biol, 2001.13(1) Bray G A , D e L a n y J P , Harsha D W , Volaufova J , and Champagne C M , Body composition of African American and white children: a 2-year follow-up of the BAROC study. Obes Res, 2001. 9(10): p. 605 N C B I , UniGene Cluster Hs.256278 Homo sapiens. 2002. Tanner J M , Growth at Adolescence. 1962, Oxford: Blackwell Scientific Publications. M c K a y H A , Bailey D A , Mirwald R L , Davison K S , and Faulkner R A , Peak bone mineral accrual and age at menarche in adolescent girls: a 6-year longitudinal study.] Pediatr, 1998.133(5): p. 682Jones G and Dwyer T , Bone mass in prepubertal children: gender differences and the role ofphysi and sunlight exposure. J Clin Endocrinol Metab, 1998. 83(12): p. 4274-9. Boot A M , de Ridder M A , Pols H A , Krenning E P , and de Muinck Keizer-Schrama S M , Bone mineral density in children and adolescents: relation to puberty, calcium intake, andphysical ac Iuliano-Burns S, Mirwald R L , and Bailey D A ,  Endocrinol Metab, 1997. 82(1): p. 57-62.  Bone mineral density in female athletes representing sports with different loading characteristics of the skelet  203.  Heinonen A , Oja P, Kannus P, Sievanen H , Haapasalo H , Manttari A , and Vuori I,  204.  Cullen D M , Smith R T , and Akhter M P ,  205.  Bourrin S, Palle S, Pupier R, Vico L , and Alexandre C ,  17(3): p. 197-203.  206. 207. 208. 209. 210.  mechanical loading. J A p p l  Time coursefor boneformation with long-term external  Physiol, 2000. 88(6): p. 1943-8.  in three %pnes of the rat tibia. J Bone  Effect ofphysical training on bone adaptation  Miner Res, 1995.10(11): p. 1745-52.  Adverse effects of strenuous exercise: a densitometric and histomorphometric study in the rat. J A p p l Physiol, 1994. 76(5): p. 1999-2005. Robling A G , Hinant F M , Burr D B , and Turner C H , Shorter, more frequent mechanical loading sessions enhance bone mass. M e d Sci Sports Exerc, 2002. 34(2): p. 196-202. Q i n Y X , Rubin C T , and M c L e o d K J , Nonlinear dependence of loading intensity and cycle number in maintenance of bone mass and morphology. J Orthop Res, 1998.16(4): p. 482-9. Parfitt A M , Hormonal influences on bone remodeling and bone loss: application to the managem hyperparathyroidism. A n n Intern Med, 1996.125(5): p. 413-5. Helge E W and Kanstrup IL, Bone density in female elite gymnasts: impact of muscle strength and hormones. M e d Sci Sports Exerc, 2002. 34(1): p. 174-80. Bourrin S, Genty C , Palle S, Gharib C , and Alexandre C ,  149  Lifestyle determinants of bone mineral: a prepubertal Asian- and Caucasian-Canadian boys and girls. Calcif Tissue Int, 2000.  211.  M c K a y H A , Petit M A , K h a n K M , and Schutz R W ,  212.  Langdahl B L , Ralston S H , Grant SF, and Eriksen E F ,  co^mPansonb^een  An Sp 1 binding site polymorphism in the COLIA1 gene predicts osteoporoticfractures in both men and women. J Bone Miner Res, 1998.13(9 1384-9.  213.  Nordstrom P and Lorentzon R, Influence of heredity and environment on bone density in adolescen parent-offspring study. Osteoporos Int, 1999.10(4): p. 271-7. Frost H M , Introduction to a New Skeletal Physiology. V o l . I  214. II. 1995, Pueblo, C O : T h e Pajaro Group, Inc. 215.  Frost H M , p. 911-7.  Muscle, bone, and the Utah paradigm: a 1999 overview.  M e d Sci Sports Exerc, 2000. 32(5  Skeletal Tissue Mechanics. 1989, Berlin: Springer-Verlag. Muscle strength, bone mass, and age-related bone loss. J Bone Miner Res, 1997.12(10): p.  216.  Martin R B , Burr D B , and Sharkey N A ,  217.  Burr D B , 1547-51.  218.  Rao L G , L i u LJ, Rawlins M R , M c B r o o m RJ, Murray T M , Reddy G S , Uskokovic M R , Rao D S , and Sutherland M K , The biological activities of 1 alpha,25-dibydroxyvitamin D3 and its synthetic a  1 alpha,25-dihydroxy-16-ene-vitamin D3 in normal human osteoblastic cells and human osteosa cells are modulated by 17-beta estradiol and dependent on stage of differentiation. Biol Pharm Bu 24(3): p. 242-8. 219.  220.  221. 222. 223.  Takeda S, Yoshizawa T , Nagai Y , Yamato H , Fukumoto S, Sekine K , Kato S, Matsumoto T , and Fujita T , Stimulation of osteoclastformation by 1,25-dihydroxyvitamin D requires its binding to D receptor (VDR) in osteoblastic cells: studies using VDR knockout mice. Endocrinology, 1999.140 p. 1005-8.  Amling M , Priemel M , Holzmann T , Chapin K , Rueger J M , Baron R, and Demay M B , Rescue of the skeletalphenotype of vitamin D receptor-ablated mice in the setting of normal mineral ion hom formal histomophometric and biomechanical analyses. Endocrinology, 1999.140(11): p. 4982-7. Wolff J, The law of bone remodeling. 1892, Berlin: Springer-Verlag. Lander E S and Botstein D , Mapping mendelianfactors underlying quantitative traits using RFLP link maps. Genetics, 1989.121(1): p. 185-99. Lander E and Kruglyak L , Genetic dissection of complex traits: guidelinesfor interpreting and reporti linkage results. N a t Genet, 1995.11(3): p. 241-7.  150  A P P E N D I X 1  151  T  H  UN  R  O  I V E R S IT V  DR  F  I T 1 S H  C O L U M  B I A  2)0, Wm Mcrtiotisl G y « i 60S! University is«uk«-ifd Vi»in*o><vt-r, B.C. Canada VST I'll r»*:{0r>f) 8 / S t S C i  fth (6(>ii S22 5S3E  lealthy Bonos Study Consent Form for Families C t o t r a t O MI inerti *  d '"(!.' I |  t <  , ,<  h i > \ < Mi . !' r '!. i >' > '  ,  .  < ,i j , < 1 1  <  ;  (  , 'i il i >,  j  (t  j  11 '.i vg two sesftng sessions (approximately two 11 j « d one a! the end of the school year ai she A ;!«cte she foJiowng procedures:  || r | i f .  III  ii , 1  i i i nl , i , il i i ii i  I) u  (  i I i  f  i i *, t \  ! M«J sludy shi!; person will II i sory gueitlonnri'O will te  Characterise of the bone trt im hect .Tres be assessed toy quantitative ultrasound This is a very simple ptoceC-M thai only requires <ftai your am temto $ ) i tot 2 minutes wHft his or he-' Iteef resting on the measurement device. There ate neither risks or rfisconilort sssoctoled wish this prr.<«Jur«. Your child's w k * j body, hip and spine bone slalus wil! be evateteij by a bom deratarnek-r This procedure is painless and routinely used in modem medical pfHd.ce il requites only that she child ties s i * <wt the padded meaMxemani laWe tor abswl IS minutes. AJiltough Ihe bone measurement is X-ray based, the total patters! elective dose per session w«; be less Shan JO milKrem wiiieh is sinvlar So the background radiation one would receive maRin<} a one-way (light Horn Vsncorivf-f So Halifax, io put this in po-spociive. the annual background ra&itiorviri Vancouver due lo natural seances is around 550 miirero per year, lae cower permissible level let the general poputsfion is 500 mijlirem per yeas. These values eao be used to cec-paf" the reteSve dsk ol toss than 10 naSr-wi exposure: Srocn she bone density procedure bon<; density o^asurements wiii tie conducted by a 'rained operator. Less than IS minuies is requirert lor aii the beas (JieasisfK'neni proceduies,  i  I  t  me only, wo m l do a coef analysis ot She Worrechanfeal tees invoked to i cc al She UBC Biomechanics Labor alary, and simpty requires fhas i s mma i$ measured Ths will rake approximately tO minutes So complete.  in  I  '  '  , I  f  i  I  11  rti '  ;  < ' M  '  I  I  i' i  I iI  rt  •' 1  < '  ' '<  "  (  i  I n  1  I' )  in  ' ii> IN" • n i l  Price? 1 of 3  152  '  K i  l / i  ! I  n  l  '|. I , 1 ,  I  I  S s i  |I  I!  i  i  '  n x  l (  ,  i  ' » , » ! ' ! , ,  1  '  1)1.>( .  ' I,  a l J I  (.1  III  I  *  |, | | ,  j Js {  (  ,  1  to mi stud* reports or research  ins researcnteam(w. Mcrsay, Kerry Maewivie, and Mota  chdi  Please Pe assured iha! you may ask questions 31 any time. We yon and your  w3 be  glad lo discuss year  cnft/s  results with  when tftey nave become available- and we welcome your comments and suggestions. Should  you have any concerns arV.trf mis study re v.-isft Minor rnten^tstNi please contsci Dr. Heather McKay (S223120) or Kerry MaeKeMe ff?Z-03gt)at She Unveisriy of Bntis*> Caiturrfcia. Si you terns any concerns about you* child's rigtt's or irenfncni as a research subject, please contact Or. fi. P. Sprafey at the oflfec ol Research  Page 2 of 3  153  T H E UNIVERSITY  OF  B R I T ! S 11 C O L U M B ! A  School of H u m a n Kts 210, War Mcroorb! C.ym 60S! University Beuirvaia Vattiutuvei, B . C . Canada v'ij'f J X i  HEALTHY BONES STUDY CONSENT FORM Parent;'? C o n s e n t S t a t e m e n t ;  (please prim m en a m e ofo n e or boih parents) understand the purpose  ant) procedures oi l i t i s  s t u d y as described io p a r f i c i p a i c in [ p l e a s e  and  1 voluntaiPy agree  in a l l o w  m y child,  _ . . „ _ T h e c e n i r a l c o m p o n e n t s o l t h e H e a t h y B o n e s S t u d y (height, weigh!, q u e s t i o n n a i r e s , bone- m i n e r a l density, ultrasound, a n d biomechnnical Ijumping lorce) measurements), .  J d on o ! a g r e e i o h a v e m y child participate in Hie central c o m p o n e n t s o f'he Healthy B o n e s Study.  P l e a s e c h e c k ( • ! t h e s e c o n d a r y c o m p o n e n t b e l o w il y o u a g r e e lo ( h e c e n t r a ) c o m p o n e n t s o l the H e a l t h y B o n e s S t u d y a n d w i s h y o u r c h i l d t o p a r t i c i p a t e i n this p a r t of t h e s t u d y . B o m M e l a i e d g e n e analysts from a c h e e s smear. I d o n o t a g r e e l o h a v e m y c h i l d to p a r t i c i p a t e i n t h e b o n e - i e l a i e d y e n s a n a l y s t s . I u n d e r s t a n d t h a i a i a n y t i m e d u r i n g t h e s t u d y w e will b e Iree » w i t h d r a w w i t h o u t j e o p a r d i z i n g a n y m e d i c a l m a n a g e m e n t , e n p l o y m e m o< e d u c a t i o n a l o p p o r t u n i t i e s . I u n d e r s t a n d 9 r e c o n t e n t s o l a n t r u e ' ; p a g e s o f i t * , f o r m , the p r o p o s e d p r o c e d u r e s a n d possible risks. I h a v e i a d the opportunity to as'* questions a n d h a v e r e c e n t ! s a t i s l a c i o r y a n s w e r s l o a l l i n q u i r i e s r e g a r d i n g this s t u d y .  Signature ol ParenVSusrdran  Dale  Signature ot Witness  Dale  Signature onflviisiigaiof  ,  Dale  Child's .Sja'orrfn;: I u n d e r s t a n d t h e p u r p o s e a n d p r o c e d u r e s o l l i t i s s t u d y a s d e s e r t e d a n d I v o i u r a a n y a g r e e to p a r t i c i p a t e . I u n d e r s t a n d i h a i a i a n y t i m e d u r i n g t h e s i u d y , 1 wJit b e f i e e So v m h d t a v / w s i h o u l jeopBKfting a n y m e d i c a l m a n a g e m e n t , e n p t o y r r i e M e re d u c a t i o n a l opportunities. I u n d e r s t a n d t h e c o n t e n t s o l t h e c o n s e n t totrn, t h e p r o p o s e d p r o c e d u r e s a n d p o s s i b l e r i s k s . I n a v e h a d t h e o p p o r t u n i t y t o as'?, q u e s t i o n s a n d h a v e r e c u r v e d s a t i s f a c t o r y a n s w e r s i o a l l inquiries r e g a t d i n g this s t u d y . .  Signature oi Child  ~ Pans? 3 o l 3  154  mmim mmwmmim. mimmmrnmwMmmemn, •wmmmmm'i&imv&Ammm  ®mm¥m®mmm%mmAm UBC Bone And Mineral Measurement: Lab  UBC 'Biomechanics Laboratory  t0^m-^%m^m^.n^m^  155  ;fi:fi:l!E£i>_  K  _  nmimi\mmmm?mrym  Wlli!l^#iI^J3«^tiS^»_ (ll?*f M P SI M - F J ^ g ^ )  .. wmmmmmmmmnfflmmmmt.  F3J®  aw  157  158  HEALTH? B O N E S STUDY: PERSONAL; DATA FORM, FALL NAME:  _  AGE:  .' „  „„  TODAY'S DATE:  GENDER:  _  SCHOOL:  1999 „  6lRTHDATE:__  „ „  GROUP NUMBER: 11213 (circle one)  _ GRADE  Have you participated in the Healthy Bones Study before? yes / no (circle one) What language do your pareni(s) speak aHiome?  L _  What country were you bom in?  „  _ _  _  _  What country was your mother bom in?  _,.  What country was your fattier born in?.  :  How ionythave you jtved-to Canada.?_  _  „  _,  •.  ;  If you Have no! lived-in Canada all-your life, w h e r e ; d i d : f i v e before coming to Canada? ., _. : ., _ _ _ ;  ANTHROPOMETRY AND FITNESS TESTING Measurer: . „ HEIGHT:  ;.  _  WEIGHT  ,  _  _  SITTING HEIGHT CALF GIRTH  ,  VERTICAL JUMP - standing,  jump!  „ ' jump2_  ,  jump 3 „  LONG JUMP ULTRASOUND: 8UA  _____  SOS  _  _  QUI.  159  ;  „.  Is y o u r c h i l d c u r r e n t l y t a k i n g a n y m e o T c a t t o r e r ? : „ , If y e s , v r h a i m e d t c a t i o n ( s ) i s y o u r c h i l d t a k i n g ?  yes  What are these fnedreaiion(s) for?  B o n e History H a s y o u r c h i l d e v e r b e e n - h o s o t e f e e d , c o n f i n e d t o b e d o r h a d a i i m b i m r o o b i f e e d (I.e., a r m i n a c a s t ) " _ yes no If Jf»s, list.•condition, a p p r o x i m a t e d a t e a n d t i m e i n v o l v e d (Example: wrist b a c t u r e summer, I960 !0 w e e k s ) Reason.  Date •  Time involved  Is t h e r e a h i s t o r y o l w r i s t , h i p , o r s p i n e f r a c t u r e s i n y o u r f a m i l y ? If. yes, I n d i c a t e w h o w a s . a f f e c t e d . mother father • . maternal (jraodmother paternal giyndmother • . _ „ . ntatemat grandfather ' paternal grandfather Is ( h e r e a h i s t o r y o f o s t e o p o r o s i s i n y o u r ' f a m i l y ? tl y e s ; i n d i c a t e w h o w a s a f f e c t e d mother maternal grandmother raatemal grandfather  „ „ '  :  yes  y  no  father p a t e r n a l grandmother • paternal grandfather  te t h e r e a h i s t o r y o f a n y o t h e r b o n e d i s e a s e i n y o u r M y ? yes , no If y e s , p l e a s e i n d o l e I h e f a m i l y member® a f f e c t e d  1 2.  ' _ ' '"  W h a t i s t h e n a m e o i t h e c o n r j i t i o n j s ) a f f e c t i n g this tnnVily m e m b e r ?  1  • • '  •  Physical Aciiiviiy  H o w w o u l d y o u r a l e t h e p h y s i c a l a c t i v i t y l e v e l o l y o u r child? (physical a c t i v i t y i s defined a s v i g o r o - . r s a c t i v i t y t h a t m a k e s t h e m s w e a t a n d / o r b r e a t h e hard) _ •„„ I n a c t i v e _ Someiiriwsactive  ! _ \ Moderately active : O f t e n .active •'_ V e r y a c t i v e  THANK YOU FOR YOUR PARTICIPATION  161  Healthy Bones Study Food Frequency Questionnaire: Spring 2001 Name/Grade: Date: „ We would like to know about some Of the foods you eal. For each food listed please fill in how often you usually eat a portion of the-size stated. II you eat the food: « every day or more than once a day, fill tn how many times you have it per day • less than once a day bu! more than one a week, fill in the times per week • less than once a week, bu! more than once a .•month, (ill in Ihe times per month • less often than once a-month, or never eai il, put an 'X' under'cfa not .eat'. Example: Janice has a glass of orange juice every.rnornirig, along with (wo slices of toast. She usually has two sandwiches ai lunch, and eats trench fries about 3 times per week. She almost never eats-caulifiower. •  ; '-Vs.-Per day's '" ^ f^eweek-. • .'Per montli 1 ' ..' • 3 ; „_J3_ _ 1  Orange Juice, 1 cup French fries, regular serving Cauliflower, % cup.(125 ml)' Bread or toast, 1 slice  NUMBER OF TIMES I EAT THE FOOD Per day Per week • Per month ;  Btead or toast, 1 slice or 1 roll  Dorr S eai  v  X  :  _ Don't eai  Muffin, 1 large Pizza, 1 medtuncslee  _  ....  .  .  Cheeseburger or veggiehurger with cheese  _  „  _  Cheese: 1 slice processed OR 1 piece  hard cheese {plain or in sandwich) Broccoli, ,!4c.up.(125mt)  '  _  _ _  Gar-Ian (Chinese broccoli}, - % cup  ,  ;.. „;  :,„ _  ;  Bok-choi {Chines&cabbagej,:^cup ice cream (large:scoop) ' ' .'••'•  ... _  :  :  .  Frozen yogurt (targe scoop) • Fast food milkshake  .  Collage cheese, >i cup  _  Yogurt, small {17A mi) carton or equivalent •  .  . ..  Canned salmon or sardines with bones, K small can • '  •  Soli drink, 1 can or large glass  _  l  162  Per day .  Per week  Pet month  Don't  Tofu, 2 oz (60 gni) Milk oil cereal:  '  ;  __  _  ;  Orange juice, 1-cup Milk (any type-induding chocolate), 1 cup.  _.  Macaroni & cheese. 1 cup (?50 ml)  I usually drink (choose- one. only) milk OR chocolate milk OR „soy milk OR rice milk Are yen allergic to anytfeods?' _  „  _NO  _Y'£S:,(whal foods?.  Do.yow use any vitamin and/or mineral supplements? (This question is not-about medications) Multivitamin MullivUamin/ruineral iron Vitamin C Calcium Other  Daily , _ _  >3x/v/eeK  1-3x/\veek ,  ..„  _. ..,  _  <1Aveek  _  „  What rs me branclfriame of She supplement?  THANK  YOU!  16J  .  Healthy Bones Activity Questionnaire: Spring 2001 N a n , e  Sex:  ^-  M„  -• -  -  .  _  Age:,, Grade:.  W e w o u l d l i k e lo k n o w a b o u t t h e p h y s i c a l a c t i v i t y y o u h a v e d o n e in t h e l a s t 7 d a y s . T h i s i n c l u d e s s p o r t s o r d a r . c e I h a t - m a k e y o u s w e a t o r m a k e y o u r l e g s feel t i r e d , o r g a m e s t h a t m a k e y o u hurl a n d p u f f , l i k e t a g , s k i p p i n g , r u n n i n g , and climbing. Remember: A  T h e s e a r e n o right o r w r o n g a n s w e r * - t h i s i s n o t a l e s t ,  B  P t e a s e a n s w e r all q u e s t i o n s a s h o n e s t l y a n d g c c m a t e t y a s y o u c a n - W i s v e r y a ' m p o r t n m  1. P H Y S I C A L A C T I V I T Y ! N Y O U R S P A R E T I M E ( t h i s d o e s n o t i n c l u d e - P . E c l a s s e s ) . H a v e y o u d o n e a n y o l t h e f o l l o w i n g a c t i v i t i e s i n t h e p a s t 7 d a y s ? If y e s , - h o w m a n y t i m e s a n d f o r h o w long? ( R e i n e m b e i , r e c e s s i s 1 5 m i n u t e s l o n g , a n d l u n c h i s u s u a l l y Vi a n h o u r (30 minutes)) •Tick o n l y one c i r c l e per row*.  No-  •  1-2  Skipping  0  •  0"  four.Square  0  Creative P l a y g r o u n d  0  0  0  0  Tag  0  Bicycling  0 • 0  Jogging or r u n n i n g  0  G 0 0 0  0 0 0 0  0  Walking for exercise  Aerobics  0  •0  0  0  Swimming  Q  0  0  0  0  Baseball sofibaii  0  0  .0  0  0  Dance  0  0  0  0  0  Football  0  0  0  0  0  Badminton Skateboarding  , •  'Soccer*'"  '  >  -  ".-  "  Street-Hockey" Volleyball  •  •©/.••  • 0  0  .0 .0.'  0  0  0  0-' •G .0 . J  '  .0 • :  ,  0  c  • 5-8 - .7.orrnore-irmes '0 .- - ' 0  "••0 [ - 6.':  0 0 0  0  0 0 .. 0 0..• 0 -0-:.  0 0 .. , 0 0 0  --.o :  ' 0 . 0 0 0 0  0  0. 0  .'•O'V 0 0  ©  0  0 0  F l o o r Hockey  0  Basketball  0  tee skating.  0  0  0  C r o s s - c o u n t r y skiing  0 0  0 0  0  0  0  0  0 0  0  Ice hockey/rtngeile •  0  0  Other:  0  t>  0  0  0  0  0  0  0  0  164  time per  session  2. In the m i days, during your PHYSICAL EDUCATION (PE) CLASSES, how often were you very active {playing hard, running, jumping and throwing)? Check only one. 0  I don't do PE  0  Hardly ever  0  ''•Sometimes  0  Quiteofteu  0  Always  •  3. In the l a s t 7 days, w h a t did you do most ol t h e lime a t R E C E S S ? C h e c f c only- o n e . 0  Sat rlown {isfcir*}. reading, dornr? school work) • ,  0  'Stood arooorj or walker! around.  0  Ran or { t o y e d a J t t e bit.  0  Ran around a n d played quite a hit.  0  Ran and played hard mew of lh<? lima.  4. I n the last 7 days, what did you normally do AT L U N C H { b e s i d e s eating lunch}? Check only one. 0  Sat down (talking, reading, doing school work)  0. Stood around or walked around. 0  Ran o r p l a y e d a Imle bit.  0' Ran around and played quite a bit. .. 0  Ran andplayed hardmos? of the time,  5 . I n the last 7 ' d a y s . o n how m a n y days R I G H T A F T E R ' S C H O O L , d i d ' y o u do s p o c t s . - d a n c e , or' day games i n which  you were very a c t i v e ? Check only one. • 0 Norsev 0  t time last week.  0  2 or 3 times,  0  4 tiroes last week  0  5 times last wesk  6. I n t h e las! 7 days, on how many E V E N I N G S did you d o sports, dance, or play games i n which you wore very active? Check only one' 0  . 0  Morse.  Y l i m e last-week.  0  ^.Slimes.  0  4 - 5. times, last w e e k .  0  6 • i limes, l a s t weft-.  " 165  7. H o w m a n y limes d i d y o u d o sports, dance, o r p i s y games i n w h i c h you were v e r y active L A S T W E E K E N D ? C h e c k o n l y one. 0  None.  0  liime.  0  ?-3times.  0  4 -5 times.  0  6 o r m o r e times:  8. W h i c h OUE c l the following t i v e s t a t e m e n t s , describes you best for the l a s t ? d a y s ? R e a d a l l 5 b e f o r e d e c i d i n g o n the o n e a n s w e r that d e s c r i b e s you. 0  A l l o r m o s t of m y t r e e t i m e w a s s p e n t d o i n g t h i n g s t h a t i n v o l v e d l i t t l e p h y s i c a l e f f o r t ( e . g . w a t c h i n g T V ,  h o m e w o r k , p l a y i n g compile? g a m e s . N i n t e n d o ) , '  0 Isometimes  (1-2  tiroes.jasi w e e k )  •  d i d p h y s i c a l t h i r i g s in m y free t i m e  (e,g,  played  sports  went  running, s w i m m i n g , hike riding, did aerobics), 0  I o f t e n , ( 3 - 4 t i m e s l a s t - w e e k ) d i d p h y s i c a l t h i n g s i n my free t i m e .  0  I q u i t e o f t e n (5-6 t i m e s l a s t w e e k ) d i d p h y s i c a l t h i n g s i n m y f r e e l i m e . .  0  I v e r y o f t e n (7 o r m o r e t i m e s l a s t w e e k ) d i d p h y s i c a l t h i n g s i n m y t r e e t i m e .  9. H o w m a n y h o u r s p e r d a y - d i d y o u w a t c h t e l e v i s i o n or p t a y N i n t e n d o tost w e e k ? { e a c h s h e w i s u s u a l l y a haft h o u r o r 30 m i n u t e s ) . C h e c k o n l y o n e , 0  | w a t c h e d l e s s t h a n 1 h o u r o r h a v e no T V ,  0  I w a t c h e d m o r e t h a n 1 . h o u r hei less I r a n 2  0  I w a t c h e d m o r e than 2 h o u r s b u t l e s s t h a n 3.  .0  I w a t c h e d m o r e t h a n 3hours b u t l e s s than <S,  0 10.  I watched more than 4 h o u r s ,  W e n ; y o u $*:>. tasl week, or d i d a n y t h i n g • p r e v e n t y o u f r o m d o i n g y o u r n o r m a l p h y s i c a l a c t i v i t i e s  0 0  7  Yes No  if yes. w h a t prevented y o u ?  „  _  _  166  _.  1!., Mark H o w often you #} physical activity (like playing sports., games, c l o i n q dance o r any mm physical activity) fo? e a c h day last week ( t h i s includes P . E , l u n c h , recess, alter school, evenings, spare fane, etc). None  Little Bit  -Wednesday•Thursday Friday-  wisu  0  0.  Tuesday  Medium  Laura o«  0  Monday  0  .0  Q v' ; \ , C L -  0  0 0 0  0 © 0 0 0  Often  Very  0 0 0 0 0  0 0 G 0 0 0  0  0  •0  0 0 0  Oiten  -12. Do you 'participate, in organized s p o r t or activities (music lessons, Chinese s c h o o l , tutoring, air! auides. toy scouts) outside of school?  0 0  Yes N o  ! I yes, what spoii(s) or/activities -do you-.do?  Slow many n i g h t s d u r i n g the week do you do these activities? ( i f you •. c h e c k the circle beside "2" a n d write swimming lessons o n the line, « « • • 0  '.  activity..  0  2  activity.  0  3  activity:  0  4  activity:  0  5  activity:  0  6  aciivily:  0  7  activity:  .  _  _  . v u m n j i - s s r i : nights o l live week, «r. •*, <tc <u .» <H activity o n a l i n e ) .  „  . _„ _  ;._ _....„... _  THANK YOU!  4  167  „  _  891  a.  T3  C CT  Q. Q  6'  3  CO  —j :x  CD  "0  c  o  cn Q.  (D <  m  CU.  3  fc  (8 •  CI)  3"  Put  iO  a.  fi)  9  O  s/> CD  X  w"  —«  3  "C> O 6)  ct) .  3"  O C  O m 'JK" Q, ©  CJ  o.  3" fD a .  m <g  O eft  a  a a"  ro a>  Q~t ••<  •><  <t>  c  cj>  » •OC  cu  I. 3* 3  m cc  Q.  •<  <; O  o  3  fi}  3 15.  O'  o  CT CO  '51  CB t/> 'O  cr  'ir  3  cu  3-  ct)  3  0>  O"  c  «  CD  3  W  O"  O < ft)'  s  0,  3'  .Sa»  o o  3"  3: —>  0 3  -  su  o  "O  .ci 3  1  3  5 •a  ma  CB O  'O  ST, (Q CO. if> • CO  0)  33  oi  m  . a.  3  Ci  cr  3n  d XI  o  3'  ..... O  o S\ c to. o. o 2, o X> c ,^ cr cB o*  >  03 CO CD eft  O  ft) 3  «  3c  w  3  o  S  w  <¥' ct)  £  .ca  O O  co  3  c  CO 3"  jM  t»  CD  a.  o  O  o o  •3"  3  CO w  o  3 <»  cu 3  to  CB to  JS'  o 0)  03 3'  <a  in  O c -* • •cr  o.  3-  Ci  3  ca  ',/>  3  o  ®  Q.  3  03  Ct)  Q. O  5' 05  3  o o  o H  0>  3  3" 3  CO  o  (tt Ui 3-  03 < CB  Q.  Ct)  o 5>  ta  t o  Ct)  a  3T  zr  CS fi]  a  to o 3 u>  n_  a: co co 3  cr su  "a "O  CB 3  691  C  CT  5"  *  O  O  3"  ~/~ CD  TJ  <o O •<  O CO'  m  Z  •O  so SI; o> H. as  O  o  "O — a» a n n o w H = ° 2* 2. 3" 3 « »' o o" £ 5 n Sf. ST 3 < o » - . c <o ^ . a l 3 1 3-c5' O ^ 3  O • s>  3  p-  ft)  =: Pa. w cry o  w  •  C  as -5" a" £*• Xf  5  § w,  3  GO ?  o c  J i S to 7 2 O K 5 ' 3  £ c  3 - 3  s  *  o w  • a) m  m  ~*  ST.  ai  o  P  su  05 TO  3  3 ™  0)  «"a. o » 3 3  fti  rt  S» ° o  •S H °  u  3"  ~.  fi> 5} < ~o  ai  3 ato• to 3  8 a)  6" o  XJ  3"  <9 01  "o  R-  3-  §  3  f r> A»  w  ce  a."  a; e> "P- m d  q  :  1:5 - o T a> 3^ S* 5 ' a. If j.. <u 2: to to s > < ~* ra  3 Q  a  -*  SU  _  X"  P  3  3  CO  3  0 ^  fl»  5  z >  s.  m  H e a l t h y B o n e s S t u d y S e l f A s s e s s m e n t of Maturity S t a t u s : G i r l s  A s you k e e p g r o w i n g o v e r the next f e w y e a r s , y o u wil! s e e c h a n g e s in y o u r b o d y . T h e s e c h a n g e s h a p p e n at different a g e s for different c h i l d r e n . Y o u m a y a l r e a d y be s e e i n g  some c h a n g e s ,  a n d s o m e of  your-friends may have a l r e a d y  gone  through s o m e c h a n g e s . S o m e t i m e s si is important to k n o w h o w a p e r s o n is g r o w i n g without h a v i n g . a d o c t o r e x a m i n e t h e m . It cart b e h a r d for a p e r s o n to d e s c r i b e t h e m s e l v e s in w o r d s , s o c h i l d r e n g o through. p u b i c hair growth  on  doctors-have d r a w i n g s  of s t a g e s that all  T h e r e a r e 5 d r a w i n g s of b r e a s t growth, a n d 5 d r a w i n g s of the n e x t p a g e . . A l l y o u n e e d  look like y o u n o w . Put  one  check mark  to do  on the line at  is pick  the  drawings  that  the d r a w i n g that is c l o s e s t  to y o u r s t a g e of d e v e l o p m e n t for b r e a s t growth, a n d o n e c h e c k m a r k at the d r a w i n g that is c l o s e s t to y o u r s t a g e of p u b i c hair  growth.  Put the s h e e t in the  e n v e l o p e a n d s e a l it s o that your a n s w e r will b e kept private,  S p r i n g 2001  Name  Please put a check mark an the drawing that looks most like (1) your stage of breast development, and (2) your stage of pubic hair development. Seal your response in the envelope provided. Thank you!  Have you had your 1 period? Y e s f i  No  If yes, do you remember w h e n ? Month  ,  THANK Y O U .  171  Year  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0077371/manifest

Comment

Related Items