A Proteomic Approach to Identify Biomarkers of Growth Hormone and Aging (312 Pp.)

Home , Allometry, Thymopoietin

A Proteomic Approach to Identify Biomarkers of

Growth Hormone and Aging

A dissertation presented to

the faculty of

the College of Arts and Sciences of Ohio University

In partial fulfillment

of the requirements for the degree

Doctor of Philosophy

Juan Ding

August 2009

This dissertation titled A Proteomic Approach to Identify Biomarkers of

Growth Hormone and Aging

JUAN DING

has been approved for

the Department of Biological Sciences

and the College of Arts and Sciences by

John J. Kopchick

Goll-Ohio Professor of Molecular Biology

Benjamin M. Ogles

Dean, College of Arts and Sciences 3

ABSTRACT

DING, JUAN, Ph.D., August 2009, Biological Sciences

A Proteomic Approach to Identify Biomarkers of Growth Hormone and Aging (312 pp.)

Director of Dissertation: John. J. Kopchick

The growth hormone (GH)/ insulin-like growth factor-1 (IGF-1) axis is an

important regulator of longitudinal growth, metabolism and aging. The physiological role

of GH is manifested in mouse models with altered GH signaling. For example, bovine (b)

GH transgenic mice are giant, lean, insulin resistant and short-lived. In contrast, GH

receptor (R) gene knockout (GHR-/-) mice are dwarf, obese, insulin sensitive and long-

lived. In order to uncover the biomarkers of GH and aging, two-dimensional gel

electrophoresis (2-DE) has been used to analyze the mouse plasma proteome. The bGH,

GHR-/- mice and their respective controls at different ages were analyzed. Female GHR-

/- and control mice were also analyzed for gender differences.

Several biomarkers of aging and GH actions have been discovered. Biomarkers

for aging include proteins whose plasma levels increased (isoform(s) of transthyretin

(TTR), haptoglobin (Hp), immunoglobulin kappa chain) and decreased (isoform(s) of

albumin, serum amyloid A-1 (SAA-1) and peroxiredoxin-2). Biomarkers for chronic GH

action include upregulation of apolipoprotein (apo) E, Hp, mannose-binding protein-C,

specific isoform(s) of alpha-2 macroglobin (a2m), retinol-binding protein-4 (RBP-4), and

downregulation of apoA4, TTR, specific isoforms of apoA1, a2m and RBP-4.

Importantly, isoforms of Hp and clusterin increased markedly in bGH mice during aging

compared to wild type (WT) mice, possibly indicating an accelerated aging phenotype of 4 these mice. Gender-specific biomarkers include plasma proteins that increased such as isoform(s) of apoE and RBP-4 or decreased such as isoform(s) of clusterin, albumin, Hp and hemoblogin-β in females. Interestingly, some proteins showed no change during male aging but significant increase (isoform(s) of apoA1, apoA4, TTR and albumin) or decrease (isoform(s) of RBP-4 and albumin) during female aging. In addition, several proteins (Hp, apoE and RBP-4) showed an interaction between genotype (GHR-/- and

WT) and gender, suggesting interplay between GH and sex.

Finally, short-term effects of GH/IGF-1 were studied in mice injected with GH and/or IGF-1. GH injection resulted in upregulation of isoform(s) of apoE, Hp, RBP-4,

SAA-1, albumin and a2m; and downregulated isoform(s) of apoA4 and clusterin. For

IGF-1, isoform(s) of Hp, albumin and TTR were upregulated. Further, biomarkers differentiating between GH and IGF-1 actions have been revealed, including isoform(s) of apoE and TTR. The biomarkers of GH/IGF-1 injection will provide potential candidates for detecting GH and IGF-1 doping.

Approved: ______

John J. Kopchick

Goll-Ohio Professor of Molecular Biology

DEDICATION

This work is dedicated to my mother, Shi Hongxia, who passed away in the summer

2005 when I was in the USA pursuing my Ph.D. degree, a Pacific Ocean away. I wish I

could have been there with you in those last few days. Losing you has been the hardest thing in my life. Sorry, mother, for not being there for you, while you had been there for

me for 23 years. May you rest in peace!

ACKNOWLEDGEMENTS

I would like to give my special thanks to my advisor, Dr. John Kopchick. It has been a great pleasure and honor to work with him, a renowned scientist as well as a great mentor, or even a father, who has taught me dedication, profession and perfection in research, writing and everything else in life.

I would also extend my thanks to my committee members, for guidance and support in the past four years. Dr. Darlene Berryman has helped me a lot, both in dissertation and everyday lab life; and I admire her as a scientist and a woman. Dr. Xiaozhuo Chen, the faculty advisor for Chinese students, has kindly helped me in transition to graduate student life in USA. He has also provided insight in understanding my research in relation with obesity and diabetes, both subjects are his specialty. Last, but not least, Dr.

Calvin James, who always put me at ease with his kind character, who always raised the

‘tough’ questions in molecular biology, reminding me to look at the big picture outside the limitation of my proteomic work.

I would also like to thank my colleagues. Elahu Gosney and Amanda Palmer provided the mice necessary for this work. Dr. Edward List taught me careful design of animal experiments. I thank my fellow student Lucila Sackmann-Sala for much help and support. I thank Dr. Shigeru Okada and Dr. Bruce Kelder for kind guidance. And finally, thanks to all my lab mates who helped make the lab life at EBI much richer than just ‘lab life’.

In addition to research, I have also enjoyed teaching undergraduate students.

Thanks to Drs Joan Cunningham, Lorie LaPierre and John Zook, I have learned the art of 7 teaching and passing on not only knowledge, but also enthusiasm to others. Dr Robert

Colvin, the MCB chair, has given me guidance over the years; many thanks to him as well.

I would also like to take this opportunity to acknowledge fundings from Department of Biological Sciences, Molecular and Cellular Biology program, Diabetes Research

Initiative, World Anti-Doping Agency and State of Ohio’s Eminent Scholars Program for support of my work.

Lastly, I want to thank my parents, my brother and my husband Yueran Yan. They have always been there for me through the laughter and the tears. Thanks to them, I have had more laughter than tears!! 8

TABLE OF CONTENTS

Page

Abstract ...... 3 Dedication ...... 5 AcknowledgEments ...... 6 List of Tables ...... 11 List of Figures ...... 12 List of Abbreviations ...... 15 Chapter I. Background ...... 17 Review of GH and Mouse Models with Altered GH Signaling ...... 17 GH/IGF-1 axis ...... 17 Characteristics of GH transgenic mice ...... 23 Characteristics of GHR-/- mice ...... 30 Overview of Plasma Proteome and Biomarkers ...... 34 Plasma proteome ...... 34 Proteomic techniques to study plasma proteome ...... 37 Plasma proteins as biomarkers ...... 43 Mouse plasma proteome ...... 44 Introduction to 2-DE ...... 46 Sample preparation ...... 46 Contaminant removal ...... 47 Protein quantification, solubilization/denaturation/reduction/alkylation ...... 48 Electrophoresis ...... 49 Staining of protein spots ...... 50 Plasma as a sample for 2-DE ...... 50 Review of GH Doping and Detection ...... 53 Recombinant human GH doping in sports ...... 53 Current approaches to detect GH doping ...... 55 Novel biomarkers for GH are needed ...... 58 Chapter II. Research Objectives ...... 61 Chapter III. Biomarkers of Aging: Proteomic Analysis of Plasma from Wild Type Male Mice of Different Ages ...... 65 Abstract ...... 65 Introduction ...... 66 Materials and Methods ...... 68 Experimental animals ...... 68 Body composition measurement ...... 68 Plasma collection ...... 68 Fasting glucose and insulin measurements ...... 68 2-DE ...... 69 Quantification of proteins ...... 70 Statistical analysis ...... 70 9

Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography (LC)-MS/MS ...... 70 Results ...... 74 Body composition, insulin and glucose levels during aging ...... 74 Identification of plasma proteins...... 75 Plasma proteins that increased during aging ...... 77 Plasma proteins that decreased during aging ...... 83 Discussion ...... 88 TTR ...... 90 Hp ...... 93 Ig kappa ...... 95 Prx-2 ...... 96 Albumin ...... 97 SAA-1 ...... 100 Conclusion ...... 102 Chapter IV. Biomarkers of Chronic GH Action: Proteomic Analysis of Plasma from Male bGH Transgenic and Control Mice ...... 104 Abstract ...... 104 Introduction ...... 105 Materials and Methods ...... 107 Experimental animals ...... 107 Body composition measurement ...... 107 Plasma IGF-1 measurement ...... 107 Fasting glucose and insulin measurements ...... 108 2-DE and quantification of proteins ...... 108 Statistical analysis ...... 108 Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography (LC)-MS/MS ...... 109 Results ...... 110 Body weight and body composition ...... 110 IGF-1, fasting plasma insulin and glucose levels ...... 112 Detection of plasma proteins in bGH mice versus WT controls by 2-DE ...... 114 Discussion ...... 130 Body composition, IGF-1 and insulin sensitivity ...... 130 Plasma proteins that were significantly different in bGH mice versus controls ..... 134 Conclusion ...... 150 Chapter V. Biomarkers of GH and Gender: Proteomic Analysis of Plasma from Growth Hormone Receptor Gene Disrutped Mice...... 153 Abstract ...... 153 Introduction ...... 154 Materials and Methods ...... 156 Experimental animals ...... 156 Body composition measurement ...... 156 Plasma IGF-1 measurement ...... 156 10

Fasting glucose and insulin measurements ...... 157 2-DE and quantification of proteins ...... 157 Statistical analysis ...... 157 Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography (LC)-MS/MS ...... 157 Results ...... 158 Body weight and body composition ...... 158 IGF-1, fasting glucose and insulin levels ...... 160 Plasma proteins that exhibited significant gender differences ...... 161 Plasma proteins that exhibited significant genotypic differences ...... 165 Plasma proteins that exhibited significant interaction between gender and genotype ...... 173 Discussion ...... 175 Body composition, glucose and insulin ...... 175 Plasma protein characteristics ...... 176 Conclusion ...... 186 Chapter VI. Biomarkers of GH And IGF-1 Doping: Proteomic Analysis of Plasma from Wild Type Mice Injected with GH and IGF-1 ...... 189 Abstract ...... 189 Introduction ...... 191 Materials and Methods ...... 194 Experimental animals and treatment protocol ...... 194 Body composition measurement ...... 195 Fasting blood glucose ...... 196 2-DE and quantification of proteins ...... 196 Statistical analysis ...... 196 Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography (LC)-MS/MS ...... 196 Results ...... 197 Body composition ...... 197 Fasting blood glucose ...... 198 Plasma proteins altered by GH/IGF-1 ...... 199 Plasma proteins altered by GH ...... 200 Plasma proteins altered by IGF-1 ...... 208 Plasma proteins differentially altered by GH and IGF-1 ...... 211 Discussion ...... 213 Conclusion ...... 219 Chapter VII. Summary and Discussion of Plasma Biomarkers of GH and Aging ...... 222 Chapter VIII. Future Directions ...... 251 References ...... 253 Appendix A: Identification of Plasma Proteins ...... 289 Appendix B: Original Gel Images ...... 308

LIST OF TABLES

Table 1 Comparison of insulin sensitivity in bGH versus WT control mice ...... 24 Table 2. Lipid profiles of GH versus WT control mice...... 25 Table 3. Comparison of insulin sensitivity parameters in GHR-/- versus WT control mice ...... 31 Table 4. Lipid profiles of GHR-/- versus WT control mice...... 32 Table 5. Organization of four papers in this dissertation...... 63 Table 6. Biomarkers of aging in WT mice...... 102 Table 7. Quantification of proteins up- or downregulated in bGH mice ...... 116 Table 8. Proteins significantly different in males and females (p<0.01) ...... 162 Table 9. Proteins significantly different in GHR-/- compared to WT mice (p<0.01) .... 166 Table 10. Biomarkers indicative of lack of GH action and gender difference...... 186 Table 11. GH and IGF-1 treatment scheme...... 195 Table 12. Biomarkers of GH, IGF-1 and their difference...... 220 Table 13. Comparison of chronic and acute actions of GH on plasma proteins...... 221 Table 14. Summary of protein changes during aging, and via GH action or due to gender differences...... 225 Table 15. MS and MS/MS scores of proteins ...... 289 Table 16. LC/MS/MS identification of selected proteins...... 300 Table 17. Mw and pI of proteins...... 302

LIST OF FIGURES

Figure 1. GH/IGF-1 axis...... 18 Figure 2. GH signaling pathways...... 20 Figure 3. Image of an adult bGH, wild type (WT) and GHR-/- mouse...... 22 Figure 4. Twenty-two proteins make up 99% of plasma protein mass...... 36 Figure 5. Plasma protein abundance spans 10 orders of magnitude...... 36 Figure 6. SWISS-2DPAGE map of human plasma proteins ...... 38 Figure 7. A schematic flow chart of protein isolation by 2-DE, then identified by MS and MS/MS...... 40 Figure 8. Mouse plasma proteome...... 45 Figure 9. 2D gel images of plasma proteins...... 51 Figure 10. Relative numbers of proteins identified in LMW proteome...... 52 Figure 11. History of GH in clinical practice and abuse by professional athletes...... 54 Figure 12. Individual datapoint plot of P-III-P against IGF-I levels in male subjects after 3 weeks of treatment...... 57 Figure 13. Body composition of WT mice at different ages...... 74 Figure 14. Four-hour fasting plasma glucose and insulin levels...... 75 Figure 15. 2D gel reference map of identified proteins...... 76 Figure 16. Plasma proteins that significantly increased in WT mice during aging...... 78 Figure 17. 3-D view of the intensity of Ig kappa (spot 1) in WT mice at different ages.. 79 Figure 18. 3-D view of intensity of Ig kappa (spot 2) of WT mice at different ages...... 80 Figure 19. 3-D view of intensity of Ig kappa (spot 3) of WT mice at different ages...... 81 Figure 20. 3-D view of intensity of TTR isoform 3 of WT mice at different ages...... 82 Figure 21. 3-D view of intensity of Hp isoforms 2 and 3 of WT mice at different ages.. 83 Figure 22. Plasma proteins that significantly decreased in WT mice during aging ...... 84 Figure 23. 3-D view of intensity of Prx-2 in WT mice at different ages...... 84 Figure 24. 3-D view of intensity of albumin isoforms 6-9 in WT mice at different ages...... 85 Figure 25. 3-D view of intensity of albumin isoform 18 of WT mice at different ages.. . 86 Figure 26. 3-D view of intensity of SAA-1 in WT mice at different ages...... 87 Figure 27. Total protein levels of TTR and RBP-4 in WT mice during aging...... 91 Figure 28. Albumin and TTR levels during human aging...... 92 Figure 29. Schematic representation of albumin fragments...... 98 Figure 30. Body weight and composition of bGH and WT mice during aging...... 111 Figure 31. Body composition of bGH and WT mice during aging...... 112 Figure 32. IGF-1, insulin and glucose levels...... 113 Figure 33. Plasma proteins up- or downregulated in bGH compared to WT mice...... 115 Figure 34. Comparison of apoE isoforms in WT and bGH mice of different ages...... 117 Figure 35. ApoE increased in bGH compared to WT mice at all ages...... 118 Figure 36. One isoform of MBP-C increased in bGH compared to WT mice...... 119 Figure 37. Hp in bGH and WT mice...... 120 Figure 38. 3-D view of four isoforms of Hp (spots 10-13) and one isoform of RBP-4 (spot 9)...... 121 13

Figure 39. Comparison of TTR and Hp isoforms in WT and bGH mice at different ages...... 122 Figure 40. TTR decreased in bGH compared to WT mice at all ages...... 123 Figure 41. ApoA1 decreased in bGH compared to WT mice...... 124 Figure 42. RBP-4 showed isoform-specific change in bGH compared to WT mice.. ... 125 Figure 43. Comparison of a2m isoforms in WT and bGH mice at different ages...... 126 Figure 44. A2m showed isform-specific changes in bGH compared to WT mice...... 127 Figure 45. Clusterin showed significant interaction between genotype and aging...... 128 Figure 46. 3-D view of clusterin isoforms...... 129 Figure 47. Body weight from 9 to 24 months and body composition at 12 months...... 158 Figure 48. IGF-1 and fasting glucose and insulin levels...... 161 Figure 49. 2D gel map of identified proteins of interest...... 162 Figure 50. Plasma proteins that exhibited significant differential change during aging between males and females...... 164 Figure 51. Cropped gel image from a representative mouse in each group showing increased intensity of apoA4 and decreased intensity of apoE in GHR-/- versus WT mice...... 166 Figure 52. 3-D view of intensity of apoA4 (spots 1-3) and apoE (spots 4, 6 and 7)...... 167 Figure 53. 3-D view of intensity of apoA1 (spots 8-11)...... 168 Figure 54. 3-D view of intensity of RBP-4 (spot 16), TTR (spot 17) and Alb (spots 19 and 20)...... 169 Figure 55. 3-D view of intensity of RBP-4 (spot 14)...... 170 Figure 56. 3-D view of intensity of RBP-4 (spots 15) and Alb (spot 18)...... 171 Figure 57. 3-D view of intensity of RBP-4 (spot 12) and Alb (spot 21)...... 172 Figure 58. 3-D view of intensity of RBP-4 (spot 13)...... 173 Figure 59. Plasma proteins that exhibited significant interaction between gender and genotype...... 174 Figure 60. 3-D view of intensity of Hp (spots 23 and 24)...... 174 Figure 61. RBP-4 showed isoform-specific interaction between GH and gender...... 182 Figure 62. GH and IGF-1 treatment scheme...... 195 Figure 63. Body weight and body composition change after the injections...... 197 Figure 64. 12-hour fasting glucose levels. No significant difference was found among five groups...... 199 Figure 65. 2-D gel map of proteins that changed significantly during GH or IGF-1 treatment or after the respective washout period...... 200 Figure 66. Proteins that responded to GH and returned to control levels after the GH washout period...... 201 Figure 67. 3-D view of apoE (isoforms 1-4) and a2m...... 202 Figure 68. 3-D view of clusterin isoforms...... 202 Figure 69. Proteins that decreased after the GH washout period...... 203 Figure 70. Proteins that were increased after washout of GH injection...... 204 Figure 71. 3-D view of intensity of Hp isoforms 1-3 and RBP-4...... 205 Figure 72. 3-D view of intensity of SAA-1...... 205 14

Figure 73. Multiple albumin fragments were increased after washout of GH injection...... 206 Figure 74. 3-D view of intensity of Alb (spots 1 and 2) and TTR (spot 1)...... 206 Figure 75. 3-D view of intensity of Alb spots 3 and 4...... 207 Figure 76. 3-D view of intensity of Alb (spots 5 and 6)...... 208 Figure 77. Proteins that were increased by IGF-1...... 209 Figure 78. 3-D view of intensity of TTR (spot 3). 5...... 210 Figure 79. Alb fragments that were increased by IGF-1...... 211 Figure 80. Proteins that were changed by GH and IGF-1 differentially. 5...... 212 Figure 81. Alb fragments that were changed by GH and IGF-1 differentially...... 212 Figure 82. Summary of biomarker indications of GH action and aging...... 239 Figure 83. Inflammation and lipotoxicity contribute to insulin resistance...... 244 Figure 84. Identification of a2m (isoform 4) by MS/MS...... 306 Figure 85. Representative gel images of WT and bGH mice at different ages...... 308 Figure 86. Snap shot of all 2D gels of WT mice at different ages...... 309 Figure 87. Snap shot of all 2D gels of bGH mice at different ages...... 309 Figure 88. Representative gel images of WT and GHR-/- mice of both genders at different ages...... 310 Figure 89. Snap shot of all 2D gels of male WT and GHR-/- mice at different ages. ... 311 Figure 90. Snap shot of all 2D gels of female WT and GHR-/- mice at different ages. 311 Figure 91. Representative gel images of control, GH, IGF-1 injections and washout groups...... 312 Figure 92. Snap shot of all 2D gels of mice in the injection study...... 312

LIST OF ABBREVIATIONS

2-DE: two-dimensional electrophoresis A2m: α-2 macroglobulin Alb: albumin ALS: acid-labile subunit apoA1: apolipoprotein A1 apoA2: apolipoprotein A2 apoA4: apolipoprotein A4 apoE: apolipoprotein E bGH: bovine growth hormone Clu: clusterin DAG: diacylglyceride DTT: dithiothreitol ELISA: enzyme-linked immunosorbent assay FFA: free fatty acid GH: growth hormone GHR-/-: growth hormone receptor null or knockout GHR: growth hormone receptor GHRH: GH releasing hormone HDL: high-density lipoprotein ICTP: type I collagen telopeptide IEF: isoelectric focusing Ig kappa C: Ig kappa constant region Ig kappa V-III: Ig kappa variable region III Ig kappa: Immunoglobulin kappa chain IGF-1: insulin-like growth factor-1 IGFBP: IGF-1 binding protein IL: interleukin IPG: immobilized pH gradient IR: insulin receptor IRS: insulin receptor substrate IRS: insulin receptor substrate JAK: Janus kinase JNK: c-Jun N-terminal kinase LC: liquid chromatography LDL: low density lipoprotein MALDI: matrix assisted laser desorption ionization MAPK: mitogen-activated protein kinase MBP-C: mannose-binding protein-C MS/MS: tandem mass spectrometry MS: mass spectrometry MT: metallothionein Mw: molecular weight 16

PAGE: polyacrylamide gel electrophoresis PEPCK: Phosphoenolpyruvate carboxykinase PGC-1: PPAR-γ coactivator-1 pI: isoelectric focusing point PI3K: phosphoinositide-3 kinase P-III-P: procollagen type III P-I-P: C-terminal propeptide of type I procollagen PKC: protein kinase C PKC: protein kinase C PPAR-γ: peroxisome proliferator-activated receptor-γ Prx-2: peroxiredoxin-2 PTM: post-translational modification RBP-4: retinol-binding protein-4 rhGH: recombinant human growth hormone SAA-1: serum amyloid A-1 SDS: sodium dedecyle sulfate SELDI: surface-enhanced laser desorption ionization STAT: signal transducer and activator of transcription TFA: trifluoroacetic acid TG: triglyceride TNF: tumor necrosis factor TOF: time of flight TTR: Transthyretin UCP: uncoupling protein VLDL: very low-density lipoprotein WT: wild type

CHAPTER I. BACKGROUND

Review of GH and Mouse Models with Altered GH Signaling

GH/IGF-1 axis

Growth hormone (GH), a 191 amino acid (22-kDa) peptide hormone secreted by

somatotrophs of the anterior pituitary gland, is regulated by growth hormone releasing

hormone (GHRH) and somatostatin (Figure 1), both produced in the hypothalamus, and

ghrelin which is produced in the stomach as well as several other tissues (Kopchick &

Andry, 2000; G. S. Tannenbaum, 1991; A. O. Wong et al., 2006). GH secretion is

induced by GHRH, which is stimulated by stress, sleep and exercise (Garlaschi et al.,

1975; Reinhard W. Holl et al., 1991; Kopchick, 2003; Krulich et al., 1974; Manson &

Wilmore, 1986; Osterman et al., 1974; Rice et al., 1975; Sutton & Lazarus, 1976). One protein induced by GH in tissues such as liver, muscle, fat and bone is insulin-like growth factor-1 (IGF-1), which mediates many of GH’s growth-promoting activities via endocrine, paracrine and/or autocrine routes. Thus, GH possesses both direct and indirect effects via IGF-1. These two physiological effects by GH have been termed the ‘dual effector system’ of GH action (Green et al., 1985).

Both GH and IGF-1 can exert control of GH synthesis and secretion. Elevated blood IGF-1 levels will inhibit GHRH and GH release. This regulatory arrangement has been termed the long-loop feedback system. GH also negatively inhibits its own secretion and this regulatory system is called the short-loop feedback scheme (Figure 1).

Figure 1. GH/IGF-1 axis. Reproduced from (Kopchick & Andry, 2000).

GH plays important physiological roles in longitudinal growth and lactation as well as in the regulation of lipids, carbohydrates, and protein metabolism (Kopchick & Andry,

2000). GH also has a potent anabolic effect on bone and muscle and is lipolytic and anti lypogenic on adipose tissue (Cheek & Hill, 1974; Fryburg & Barrett, 1993; Katznelson,

2008; N. Moller & Jorgensen, 2009; Velloso, 2008; Widdowson et al.; Woodhouse et al.,

2006). 19

GH exerts its potent action by signaling through GH receptors (GHRs) located on cell plasma membranes. GHR belongs to class I cytokine receptor superfamily and is a single pass transmembrane protein. Upon binding of GH to pre-dimerized GHR via two binding sites (W. Y. Chen et al., 1991; Cunningham et al., 1991), the two receptors rotate relative to each other (Brown et al., 2005), which leads to activation of intracellular

Janus-kinase 2 (JAK2) (Frank et al., 1994), a kinase that phosphorylates itself as well as the GHR on multiple tyrosine (Y) residues Y332, Y487, Y534, Y566, and Y627 (L. H.

Hansen et al., 1996; X. Wang et al., 1996). JAK2 further activates Signal Transducers and Activators of Transcription (STATs), which, upon tyrosyl phosphorylation, dimerize and enter the nucleus to regulate the transcription of GH-specific genes, among which is

IGF-1 (Kopchick & Andry, 2000; Lanning & Carter-Su, 2006; Waters et al., 2006). Other major signaling pathways downstream of GHR include SHC-MAPK, IRS-PI3K and PKC

(Figure 2) (Argetsinger & Carter-Su, 1996; Herrington & Carter-Su, 2001; Kopchick et al., 1999; Lanning & Carter-Su, 2006; Lichanska & Waters, 2008).

Figure 2. GH signaling pathways. Reproduced from (Argetsinger & Carter-Su, 1996).

IGF-1, primarily induced by GH, is thought to mediate some of GH’s growth- promoting actions (Figure 1). About 75% of circulating IGF-1 comes from liver (Yakar et al., 1999), a primary GH target tissue. Other tissues such as muscle, fat and bone also express IGF-1, which acts via paracrine/autocrine routes. The paracrine/autocrine IGF-1 action is sufficient in mediating growth, because a liver-specific IGF-1 knockout (LID) mouse has normal growth and development activities, despite a 75% reduction in circulating IGF-1 (Yakar et al., 1999). In addition, GH and IGF-1 have their own specific effects. For example, GH can stimulate growth further in rats that have been stimulated with IGF-1 maximally (Fielder et al., 1996), and injection of GH and IGF-1 produce synergistic effects on bone growth (Ohlsson et al., 1998). GH stimulates the differentiation of precursor bone cells directly and induces IGF-1, which promotes the clonal expansion of the differentiated cells subsequently (Isaksson et al., 1982). Thus, the 21 growth promoting effect of GH is mediated in part by IGF-1, but GH has IGF-1 independent actions as well, futher demonstrating the dual effects of GH action.

The GH/IGF-1 axis is involved in aging, as attenuated IGF-1 signaling over long term leads to increased lifespan in a wide range of organisms including yeast, worms, fruit flies, rodents and primates (Bonafè & Olivieri, 2009; Brown-Borg, 2009;

Holzenberger et al., 2003; Tatar et al., 2003). Several mouse models with diminished

GH/IGF-1 axis have extended lifespan (Longo & Finch, 2003). Ames and Snell dwarf mice have disrupted development of somatotrophs, lactotrophs and thyrotrophs in the pituitary due to mutations of Prop-1 and Pit-1, respectively. Although possessing a deficiency in GH as well as prolactin and thyroid stimulating hormone, these mice have increased lifespan, primarily due to decreased GH/IGF-1 action (A. Bartke, 2001; Wolf et al., 1993). The GHR-/- mouse also possesses an increased lifespan; in fact, it is the longest-lived laboratory mouse (Prize, 2009). The phenotypic characteristic of this mouse will be discussed below. Caloric restriction (CR) has been shown to increase lifespan in many species. Although the mechanism is still unclear, it is thought to be overlapping with an attenuated GH-IGF-1 axis (Bonkowski et al., 2009; Bonkowski et al., 2006; R.

A. Miller et al., 2002).

In humans, elevated levels of GH in prepubertal children result in acromegalic gigantism (Ayuk & Sheppard, 2006). Excess GH secretion in adults leads to acromegaly, featured with enlarged bones (of the hands, feet, and head) and swelling soft tissues due to GH’s fluid retention actions. Complications of acromegaly include hypertension, diabetes, sleep apnea, kidney and heart problems (Chanson & Salenave, 2008). On the 22 other hand, GH deficiency (D) in children, if left untreated, results in dwarfism (Ayuk &

Sheppard, 2006). GHD also occurs in adults, with symtoms of increased fat mass, reduced muscle size and strength and insulin resistance (Ho, 2007). GH insensitivity due to GHR mutations results in Laron syndrome (LS) (Laron, 1984) characterized by short statue, central obesity, as well as reduced muscle mass, force and exercise capacity (Ben-

Dov et al., 2003; Brat et al., 1997; Laron et al., 2006).

Mouse models with altered GH/IGF-1 axis have been developed to study the action of GH/IGF-1. Bovine (b) GH transgenic mice, reminiscent of human acromegaly, have elevated IGF-1 levels in blood, are gigantic, lean, insulin resistant and die prematurely due to kidney lesions, liver tumors (Quaife et al., 1989) and perhaps heart dysfunction

(Bollano et al., 2000; Izzard et al., 2009). On the other hand, GHR-/- mice mimic LS humans (Kopchick & Laron, 1999; Laron, 1984), and have very low blood levels of IGF-

1, are dwarf, obese, insulin sensitive and long-lived (Berryman et al., 2004; Coschigano et al., 2003). Figure 3 shows the image of a giant bGH, a wild type (WT) and a dwarf

GHR-/- mouse.

Figure 3. Image of an adult bGH, wild type (WT) and GHR-/- mouse. 23

Characteristics of GH transgenic mice

GH transgenic mice have high plasma insulin levels, with normal or low glucose

(Table 1), suggesting these animals are insulin-resistant. Interestingly, they show normal

or better glucose tolerance (A. Bartke et al., 2004; Olsson et al., 2005), at least when

young, perhaps due to a robust pancreas (Parsons et al., 1995) that secrets increased

insulin in response to a glucose challenge (Olsson et al., 2005). Peripheral tissues such as

liver, muscle and fat all have shown insulin resistance (del Rincon et al., 2007; Fernando

P. Dominici et al., 1999a; F. P. Dominici et al., 1999b), confirming the insulin-resistant

state in these giant mice. Interestingly, a porcine GHR transgenic mouse model also

shows insulin resistance in adipose tissue (pooled from three dopots: inguinal,

parametrial and retroperitoneal) of 9-12 week-old mice (X. L. Chen et al., 2001),

demonstrating an anti-insulin effect of increased GH signaling.

GH transgenic mice are lean and have altered lipid profiles compared to control

mice (Table 2). They have decreased plasma free fatty acids (FFA), triglycerides (TG)

and increased cholesterol including high-density lipoprotein (HDL) and low-density

lipoprotein (LDL) cholesterol. Liver cholesterol content also increases, but TG decreases.

In addition, TG decreases in kidney, remains the same in muscle and increases in white

adipose tissue (WAT) (del Rincon et al., 2007; Fernando P. Dominici et al., 1999a; F. P.

Dominici et al., 1999b). Liver expression of genes involved in β-oxidation and

lipogenesis are also reduced (Olsson et al., 2003). Plasma adiponectin, an insulin

sensitizer, decreases, whereas resistin, TNF-α and IL-6 increase in GH mice (Zhihui

Wang et al., 2007), which may contribute to insulin resistance of these mice. 24

Table 1 Comparison of insulin sensitivity in bGH versus WT control mice Genetic Age Gender Glucose a Insulin a Glucose Tissue background and tolerance insulin transgene used sensitivity b C57BL/6J Not Not ↔ ↑ PEPCK-bGH specified specified (Valera et al., 1993) C57BL/6J 3-5 mo Not ↔ ↑ Muscle ↓ PEPCK-bGH specified Liver ↓ (Fernando P. Dominici et al., 1999a; F. P. Dominici et al., 1999b) C57BL/6J 8 mo Both ↓ (non- ↑ (non- MT-bGH (Frick fasted) fasted) et al., 2001a) C57BL/6J 6 mo Male ↔ ↑ MT-bGH (Olsson et al., 2003) C57BL/6J 3 mo Male ↓ (ns) ↑ PEPCK-hGH mice (A. Bartke et al., 2004) C57BL/6J 5-6 mo Male ↔ ↑ ↔ MT-bGH (Olsson et al., 2005) C57BL/6J 5 mo Male ↓ ↑ (ns) MT- bGH (Berryman et al., 2006) C57BL/6J 7 mo Male ↔ ↑ PEPCK-bGH (Zhihui Wang et al., 2007) C57BL/6J 3 mo Male WAT ↓ MT- bGH (del Rincon et al., 2007) a fasting blood glucose and insulin unless otherwise indicated b tissue insulin sensitivity as measured by selected key insulin signaling molecules Abbreviations: WAT: white adipose tissue, mo: month; ns: non-significant; ↑ increased as compared to WT mice, ↓ decreased, ↔ no change.

Table 2. Lipid profiles of GH versus WT control mice Genetic Age Gender Plasma Plasma Plasma Liver Tissue TG background and FFA a TG a cholesterol a cholester transgene used ol C57BL/6J 8 mo Both ↓ ↓ ↑ ↑ Liver ↔ MT-bGH (Frick apoB ↔ Secretion ↓ et al., 2001a) LDL ↑ HDL ↑/ ↔ C57BL/6J 6 mo Male ↓ fed MT-bGH (Olsson ↔ fasted et al., 2003) C57BL/6J 3 mo Male ↑ ↑ free↑ Liver ↓ PEPCK-hGH esterified↓ mice (A. Bartke et al., 2004) C57BL/6J 5-6 mo Male ↓ ns ↑ Liver MT-bGH (Olsson apoB ↔ Secretion ↔ et al., 2005) Clearance ↓ C57BL/6J 7 mo Male ↔ ↓ ↑ ↓ PEPCK-bGH Muscle ↔ (Zhihui Wang et WAT ↑ al., 2007) C57BL/6J 1 year Both Liver ↓ MT-bGH (Palmer Kidney ↓ et al., 2009) a. Fasting levels unless otherwise indicated ↑ increased as compared to control mice, ↓ decreased, ↔ no change.

The bGH mice have an increased lean mass and reduced fat mass compared to WT

mice (Berryman et al., 2004; Palmer et al., 2009). They are also more resistant to diet-

induced obesity (Berryman et al., 2006; Olsson et al., 2005). Interestingly, bGH mice

have more absolute fat mass than WT mice before 6 months and when normalized to

body weight, they still have more fat mass% than WT before 4 months of age (Palmer et

al., 2009), suggesting an initial role of GH in adipocyte differentiation and proliferation.

Moreover, GH has recently been shown to have depot-specific effects on adipose tissue.

Preadipocytes isolated from subcutaneous region of GHR-/- mice are able to normally

differentiate and proliferate, whereas adipocytes from parametrial fat pads are unable to

do so (Flint et al., 2006), suggesting that GH is required for adipose tissue development 26 in the parametrial but not subcutaneous depot. However, adult bGH mice have proportionally less subcutaneous fat mass than other depots compared to WT mice

(Berryman et al., 2006), suggesting a more pronounced lipolytic action of GH in subcutaneous fat. Consistent with this finding is that GHR-/- mice have increased fat mass, with a large proportion in the subcutaneous region (Berryman et al., 2004;

Berryman et al., 2006), indicating that a lack of GH action primarily affects subcutaneous fat. Thus, subcutaneous fat seems to be the most GH sensitive among all fat pads.

Whether this is due to increased GHR remains to be determined and is currently under study (Miles, et al, in preparation).

Regarding energy intake and expenditure, 6-month old male bGH mice have similar normalized food and calorie intake (per body weight), increased oxygen consumption, increased liver uncoupling protein-2 (UCP-2) expression and 0.5 oC higher body

temperature versus controls (Berryman et al., 2006; Olsson et al., 2005). However,

unchanged body temperature, reduced resting metabolic rate and reduced cold-induced

thermogenesis in GH versus control mice are also reported (Moura et al., 1998). The

discrepancy is likely due to different experimental conditions, for example, the length of

the measurement time and free moving versus constriction of the mice.

GH mice live on average for ~ 14 months, significantly shorter than WT littermates,

which live for ~ 2 years (Berryman et al., 2004; McGrane et al., 1988; Steger et al.,

1993). The pathophysiology of this animal model has been studied. Acromegalic patients often have kidney problems (Flyvbjerg, 1990; O'Shea & Layish, 1992). Likewise, GH transgenic mice have enlarged kidney (normalized with body weight) and develop 27 progressive glomerulosclerosis (T. Doi et al., 1990; T. Doi et al., 1988; Quaife et al.,

1989), which features over-production of kidney mesangial extracellular matrix (T. Doi et al., 1991), increased inducible nitric oxide synthase expression in mesangial cells (S.

Q. Doi et al., 2000), and glomerular podocyte damage (Reddy et al., 2007; Wanke et al.,

2001). Together, these changes result in glomerulosclerosis and ultimately death of the animal.

Another abnormal organ in these mice is the liver. GH mice have disproportionally enlarged liver with large hepatocytes and develop liver tumors at old ages (Orian et al.,

1989; Quaife et al., 1989; Shea et al., 1987), featuring increased rates of hepatocellular proliferation (Snibson et al., 2001; Snibson et al., 1999), sustained liver inflammation and hepatocellular injury with reduced natural killer T cells (Hardy et al., 1997). In line with this, GH mice have enlarged spleen and show altered immunity, including deficiency in T helper 2 cytokine production (Gonzalo et al., 1996) and self-antibody production leading to arthritic disorders (Ogueta et al., 2000). Reduced capacity to fight oxidative stress may also be a cause of premature aging. It was found that female hGH mice had early declines of antioxidant enzyme superoxide dismutase (SOD) in the kidney and reduced levels of glutathione peroxidase (GPx) in the liver, increasing the risk of

ROS damage (Steven J. Hauck & Bartke, 2001).

One complication of acromegaly is hypertension and increased cardiovascular disease risks (Colao et al., 2001). The cardiovascular physiology of GH mice has been studied. The bGH mice have enlarged hearts even after normalization to body size

(Bohlooly et al., 2001a; Dilley & Schwartz, 1989; Izzard et al., 2009). Female bGH mice are reported to have normal blood pressure, a normal vascular wall mass and cross- 28 sectional areas of vessels in proportion to body size, normal organization and branching patterns of vessels, but increased wall-to-lumen ratio of mesenteric vasculature (Dilley &

Schwartz, 1989). The latter feature is similar to that seen in hypertensive vascular remodeling. Normal blood pressure but impaired cardiac function including suppressed systolic function and impaired energy reserve are reported in 6-8 month old female bGH mice (Bollano et al., 2000). However, cardiac contractile function is enhanced in isolated ventricular myocytes from GH mice; therefore , the impaired cardiac function is unlikely due to dysfunctional cardiocytes (Colligan et al., 2002). However, cardiocyte apoptosis is affected in GH mice. Chronic exposure to GH affects cardiac apoptosis in an age- dependent manner such that young (3 mo) male bGH mice have reduced apoptosis whereas older (9 mo) ones have increased apoptosis versus control mice (Bogazzi et al.,

2008). Increased cardiac apoptosis may contribute to the impaired cardiac function of GH mice at older ages. GH mice also have endothelial dysfunction in carotid artery caused by increased oxidative stress, which is more pronounced in aged (6 mo) versus young (2.5 mo) female mice (I. J. Andersson et al., 2006a). Metabolically, heart is also affected by chronic GH actions. Compared to age-matched controls, young male bGH hearts have increased but older bGH mice have decreased fatty acid oxidation (Bogazzi et al., 2009).

Others report increased blood pressure but with normal vasculature functions in male and female bGH mice 5-8 mo of age (Bohlooly et al., 2001a; Izzard et al., 2009).

Consistent with this, GHR-/- mice have reduced blood pressure (E. Egecioglu et al.,

2007; Izzard et al., 2009), plasma rennin and aortic eNOS expression (E. Egecioglu et al.,

2007). The renin-angiotensin-aldosterone-system plays a central role in hypertension. GH mice of both genders have increased aldosterone levels independent of rennin, potentially contributing to hypertension observed in some studies involving this mouse model 29

(Bielohuby et al., 2009). The controversy regarding blood pressure and cardiac functions of GH mice is not clear, but different ages used in different studies may be a factor.

Since GH affects connective tissue, male bGH mice have thicker skin, but this effect seems to be dependent on sex hormones, since male bGH mice have normalized skin when castrated (Wanke et al., 1999). Male hGH mice have thicker skin than controls with increased dermis, whereas female hGH mice have thinner skin than controls due to a reduced hypodermis (Serrat et al., 2007). These studies demonstrate another sexually dimorphic effect of GH and its interaction with sex hormones related to skin thickness.

Several other abnormalities of bGH mice have been noted. BGH mice have normal lymphatic vessels in normal skin, but increased lymphatic vessels ingrowth into the

granulation tissue of full-thickness skin wounds (Banziger-Tobler et al., 2008). They

display more spontaneous locomotor activity (Bohlooly et al., 2001b; Soderpalm et al.,

1999). They have altered metabolism of neurotransmitters in the brain (A. Bartke et al.,

2002). BGH mice display reduced sympathetic responsiveness as well as plasma and

tissue noradrenaline concentrations in 2-6 mo old females (I. J. Andersson et al., 2004).

In summary, chronic GH excess promotes a giant and lean body phenotype. The

organisms are hyperinsulinemic with normal or low glucose levels, and a relatively short

lifespan. The accelerated aging seen in GH transgenic mice involves abnormal pathology

of multiple organs, with altered metabolism and increased oxidative stress, leading to

liver pathology and tumors, kidney damage, and cardiac dysfunction. One note is that

many of these adverse phenotypes are observed when animals get older rather than

young, suggesting it is important to study them in an extended age span rather than only

at young ages. The gender difference also calls for more research in both genders rather

than merely males. 30

Characteristics of GHR-/- mice

GHR-/- mice have been shown repeatedly to be insulin sensitive (Table 3). GHR-/- mice have lower glucose levels, although they tend to normalize at older ages. They possess lower insulin levels throughout lifespan compared to WT mice. Glucose tolerance, however, is impaired, perhaps due to a hypo-pancreas phenotype (Liu et al.,

2004; Robertson et al., 2006) that results in reduced insulin secretion relative to WT mice upon glucose stimulus (Guo et al., 2005). Regarding tissue insulin signaling, the GHR-/- liver has similar insulin sensitivity compared to control mice; however, the GHR-/- heart and muscle are more sensitive to insulin (Table 3).

GHR-/- mice are obese and accumulate excess subcutaneous fat (Berryman et al.,

2004; Berryman et al., 2006; Liu et al., 2004). However, they have decreased plasma TG and cholesterol as well as liver cholesterol (Table 4). Plasma FFA, liver TG as well as muscle cholesterol and TG are all similar compared to controls. GHR-/- mice have lower body temperature (S. J. Hauck et al., 2001), increased energy intake (normalized to body weight) when young (2-4 mo) but similar compared to controls when older (9 mo)

(Berryman et al., 2006; Coschigano et al., 2003; Emil Egecioglu et al., 2006). GHR-/-

mice have increased oxygen consumption (Berryman et al., 2006; Westbrook et al., 2009)

with a lower respiratory quotient (RQ) values indicative of fat oxidation (Westbrook et

al., 2009).

Table 3. Comparison of insulin sensitivity parameters in GHR-/- versus WT control mice Genetic Age Gender Glucose a Insulin a Glucose Tissue background tolerance b insulin sensitivity c C57BL/6J 3-5 mo Not ↓ ↓ Liver ↔ (F. P. Dominici et specified al., 2000) C57BL/6J (S. J. 5 mo Both ↓ ↓ Hauck et al., 2001) 129/Ola-BALB/c 1.5, 2, 3, Male 1.5-9 mo ↓ and C57BL/6J 5, 7, 9, 11 ↓, (Coschigano et al., mo 11 mo ↔ 2003) 129/Ola-BALB/c- 3, 10, 60, Both ↓ ↓ Insulin C57BL/6J 180 days tolerance ↑ (Liu et al., 2004) C57BL/6J (A. 3 mo Male ↓ ↓ Bartke et al., 2004) C57BL/6J (Guo et 2-4 mo Both ↓ ↓ ↓ al., 2005) C57BL/6J 21 mo Male ↓ (ns) ↓ Liver ↓ (Al-Regaiey et al., 2005) C57BL/6J 21 mo Male Liver ↑ (Masternak et al., Muscle ↔ 2005b) 129/Ola-BALB/c 3.5 mo Male ↓ (E. Egecioglu et al., 2005) C57BL/6J 5 mo Male ↓ ↓ (Berryman et al., 2006) C57BL/6J 3 mo Male Heart ↑ 3mo (Masternak et al., 21 mo ↔ 21 mo 2006) 129/Ola-Balb/c- 6 mo Not Liver ↔ C57BL/6J specified Muscle ↔ (Robertson et al., but delayed 2006) signaling C57BL/6J (Al- 21 mo Male Muscle ↑ Regaiey et al., 2007) C57BL/6J (Giani et 14 mo No Heart ↑ al., 2008) specified C57BL/6J 12 mo male ↔ ↓ Liver ↔ (Bonkowski et al., Muscle ↑ 2009) a. fasting blood glucose and insulin b. sometimes insulin tolerance as indicated c. tissue insulin sensitivity as measured by selected key insulin signaling molecules ↑ increased as compared to control mice, ↓ decreased, ↔ no change. 32

Table 4. Lipid profiles of GHR-/- versus WT control mice Genetic Age Gender Plasma Plasma Plasma Liver Tissue background FFA a TG a cholesterol a cholesterol TG C57BL/6J (A. 3 mo Male ↔ ↓ free↓ Liver ↔ Bartke et al., esterified↑ 2004) 129/Ola-BALB/c- 3-4 mo Both ↓ ns ↓ ns C57BL/6J HDL↓ ns (Liu et al., 2004) 129/Ola-BALB/c 3.5 mo Male ↓ ↓ (E. Egecioglu et apoB↓ HDL↓ al., 2005) LDL↓ C57BL/6J 21 mo Male ↔ ↔ ↓ Muscle ↔ Muscle (Masternak et ↔ al., 2005a) a. Fasting levels ↑ increased as compared to control mice, ↓ decreased, ↔ no change.

GHR-/- mice are long-lived, and the mechanism is thought to be similar to that of

caloric restriction (CR), since CR and GHR disruption have a partial overlap in terms of

gene expression (R. A. Miller et al., 2002). Also, CR does not further improve the

lifespan of GHR-/- mice (Bonkowski et al., 2009).

In contrast to GH transgenic mice which have progressive glomerulosclerosis,

GHR-/- mice are protected from streptozotocin-induced kidney damage (Bellush et al.,

2000). However, GHR-/- mice do not have increased antioxidant enzymes or free radical

scavenge in liver and kidney compared to controls (S. J. Hauck et al., 2002), suggesting

at least in these two tissues, oxidative stress is not the mechanism to explain the

protection from aging.

GHR-/- mice have altered cardiac parameters but maintain a normal function. They

have reduced heart weight and volume (E. Egecioglu et al., 2007; Izzard et al., 2009), but

manage to produce normal cardiac output to meet their oxygen requirement (E. Egecioglu

et al., 2007). They have slightly reduced heart beat rate, reduced blood pressure, but 33 normal contractile response to norepinephrine and acetylcholine (E. Egecioglu et al.,

2007; Izzard et al., 2009).

As with LS individuals (Shevah & Laron, 2007), GHR-/- mice have decreased cancer incidence, including breast, lung, pituitary, lymph and prostate (Ikeno et al., 2009;

Kineman et al., 2001; Z. Wang et al., 2005; X. Zhang et al., 2007).

In summary, GHR-/- mice are obese, have altered lipid profiles, yet have increased insulin sensitivity and longevity. Together with GH transgenic mice, they make exellent animal models to study GH actions. Many issues remain unresolved such as the mechanism of longevity with altered GH signaling. Genomic and proteomic tools provide approaches that are now being used in attempt to add to the knowledge base of molecular mechanisms responsible for GH’s physiological effects. My proteomic studies hopefully will contribute to this field of study.

Overview of Plasma Proteome and Biomarkers

Plasma proteome

Blood is composed of blood cells and plasma. In general, blood cells include red

blood cells or erythrocytes, white blood cells or leukocytes and platelets or thrombocytes.

Serum is the cell-free, clear liquid of blood devoid of clotting factors, whereas plasma

contains the coagulation factors, and is usually obtained by treating blood with anti- coagulants such as heparin. Other than clotting factors, plasma and serum contain the same proteins. For simplicity, ‘plasma’ is used throughout this dissertation unless when serum, not plasma, is involved.

Plasma proteins can be classified into several groups (N. Leigh Anderson &

Anderson, 2002). Classical or typical plasma proteins are the resident proteins that carry out specific functions in plasma. These are secreted primarily by liver and intestine and are usually larger than 45 kDa and are, thus, not filtered by the kidney. Examples include albumin, apolipoproteins, transferrins and transthyretin. Plasma proteins smaller than 45 kDa often form multimers or bind to other proteins/lipids/polysaccharides in order to escape kidney filtration. Immunoglobulins are a second group of plasma proteins.

Hormones and cytokines are a third group of plasma proteins with varying molecular masses and half-lives. These are usually present in relatively small concentrations (ng/ml or less). Unusually high levels of hormones may cause deleterious effects. For example, excess GH causes acromegaly. The last group includes tissue leakage products, proteins released into blood stream due to cell death or damage, as well as aberrant secretions released from tumors or diseased tissues. Examples include cardiac troponins, creatine 35 kinase, and myoglobin that are used in the diagnosis of myocardial infarction (N. Leigh

Anderson & Anderson, 2002).

The term proteome refers to the entire set of proteins expressed in a certain cell, tissue, or organism at a specific time under a defined condition (N. L. Anderson &

Anderson, 1998; James, 1997). The plasma proteome is the entire protein content present in plasma. The plasma proteome contains subsets of other tissue proteomes, and is, therefore, the most complex proteome (N. Leigh Anderson & Anderson, 2002). As expected, many of the identified plasma proteins are cellular proteins (presumably released due to cell rupture); however unexpectedly, many examples of nuclear, cytoplasmic, and kinesin complex proteins have been identified (N. Leigh Anderson et al., 2004). In the mouse proteome, it is found that a large proportion of the identified proteins are membrane proteins, the majority of which are presumably shed extracellular domains of membrane proteins and receptors (Hood et al., 2005).

One major challenge in studying plasma proteome is the dynamic range of proteins.

The majority of serum proteome is composed of albumin, transferrins, immunoglobulins and apolipoproteins. In this regard, 22 proteins comprise 99% of the plasma proteome

(Figure 4). The most and least abundant proteins differ by 10 orders of magnitude

(Figure 5) (N. Leigh Anderson & Anderson, 2002). For example, albumin has a concentration of 35-50 g/L, whereas the concentration of GH in normal human plasma is

0-3 ng/L. The abundant proteins mask the less abundant ones making it difficult to study the latter.

Figure 4. Twenty-two proteins make up 99% of plasma protein mass. Reproduced from (N. Leigh Anderson & Anderson, 2002).

Figure 5. Plasma protein abundance spans 10 orders of magnitude. X axis shows representative protein names. Reproduced from (N. Leigh Anderson & Anderson, 2002).

Proteomic techniques to study plasma proteome

Traditional methods to study plasma proteins include chemical fractionation,

biochemical assay of enzyme activity, targeted detection using polyclonal or monoclonal

antibodies, as well as commercialized kits such as enzyme-linked immunosorbent assays

(ELISAs). Recently, proteomic techniques, especially 2-DE, have been developed to

study the proteome of various tissues including blood (Banks et al., 2000; Naaby-Hansen

et al., 2001; Pandey & Mann, 2000; Rabilloud, 2000).

Since its introduction in 1975 (Klose, 1975; O'Farrell, 1975), 2-DE has been used to

separate many proteins including plasma proteins. In the current human plasma proteome

SWISS-2DPAGE database (Sanchez et al., 1995) , there are 1966 proteins (spots on the

gel) detected, 626 of which have been identified. Seventy entries are present in SWISS-

PROT database (database, 2009) (the small # of plasma proteins in SWISS database may

be due to the low rate of conversion from literature reports to database, as for an entry to

appear in the database, stringent confirmation is needed). An annotated (protein spots

with idenfication) 2-DE gel map for human plasma proteins is shown in Figure 6.

Figure 6. SWISS-2DPAGE map of human plasma proteins (database, 2009). X axis: isoelectric focusing point; y axis: molecular weight in kDa.

2-DE is only the isolating step. Following 2-DE and staining of the separated proteins, they are identified by mass spectrometry (MS), the gold standard of protein identification in modern proteomics research (Packer & Harrison, 1998). Figure 7 shows a typical flow chart of this process. Proteins are digested by an enzyme such as trypsin, chymotrypsin or pepsin, with trypsin the most commonly used, and the resultant fragments are resolved as a mass spectrum based on time of flight (TOF) of individual peptides through a MS analyzer. In MALDI-TOF, individual peptides are mixed with a solvent and a matrix then placed on a probe tip. Under a vacuum condition, the solvent 39 evaporizes, leaving the peptides dispersed evenly into the matrix, which then is charged by a pulsed laser beam. The charged peptides then go through an electromagnetic field and reach the detector after a TOF that reflects mass (m)/charge (z) specific to the peptides. Since all peptides have the same charge of 1+, the TOF corresponds to the mass of the peptides. Therefore, the spectrum conveys information on peptide mass distributions. The spectrum can then be compared to proteins in the database that have been digested with trypsin in-silico, and depending on the degree of similarity, the target protein can be matched to a known protein in the database. In MS/MS, each digested peptide is further ionized into sub-fragments, which undergo another round of TOF resulting in several spectra for each peptide. Because the peptides are ionized into random fragments, an MS/MS spectrum provides information that can be used to deduce the amino acid sequence of a given peptide. Together, MS and MS/MS provide more confident identification of a given protein. One common search engine used for peptide mass fingerprint (data derived from MS) and MS/MS ion search is MASCOT at

Matrixscience (http://www.matrixscience.com/search_form_select.html).

Figure 7. A schematic flow chart of protein isolation by 2-DE, then identified by MS and MS/MS. 2-DE: two-dimensional electrophoresis, MS: mass spectrometry, PMF: peptide finger print.

Proteins are often subjected to post-translational modifications (PTMs), such as phosphorylation, glycosylation, acetylation, oxidation and methylation. There are 593

PTMs listed in Unimod database (http://www.unimod.org/modifications_list.php?). Some of the PTMs are in vivo or in vitro processes (i.g., phosphorylation) whereas others are derived from experimental conditions (i.g., carbamidomethyl of Cys by alkylation with iodoacetamide to prevent disulfide bond reformation in 2-DE). PTMs can be identified by

MS and MS/MS by selecting fixed or variable PTMs. For example, if a protein is known to have ‘carbamidomethyl’ group on the Cys residue (as seen in proteins treated with iodoacetamide in 2-DE), then ‘carbamidomethyl’ can be selected for ‘fixed modification’ so that MASCOT will add that group on Cys residue of proteins in the database in-silico 41 to compare with the actual MS and MS/MS data. On the other hand, if a protein is suspected to be phosphorylated, then ‘phosphorylation’ can be selected for ‘variable modification’; by doing this, the search engine will modify proteins in the database with possible phosphorylations in-silico and compare to the actual MS and MS/MS data for best matches.

The advantage of 2-DE coupled with MS and MS/MS is that 2-DE itself provides information on Mw and pI of intact proteins, which can serve as an internal confirmation for protein identification. Also, a given protein is unequivocally identified to a particular spot. Further, protein spots on the gel with suspected PTMs may be correlated with MS fingerprint and MS/MS spectra. Thus, 2-DE based proteomic approach has a relative high identification rate, about 50-70% (Conrotto & Souchelnytskyi, 2008). The disadvantage of 2-DE is low-throughput and that experienced experts are required for reproducible results (Conrotto & Souchelnytskyi, 2008).

Liquid chromatography (LC) is another proteomic technique. It is commonly used for high-throughput analysis of samples. In LC, a sample mixture is digested by a protease such as trypsin, and the mixed peptides are loaded onto a column and separated using hydrophobic phase chromatography. Peptides eluted from the column are then ionized and subsequently identified by MS/MS. There are quite a few derivatives of this technique, including reverse phase LC and ion exchange LC. The advantage of this technique is that automation has been developed such that the eluted peptides are directly subjected to MS ionization; thus, high-throughput, high reproducibility and high sensitivity are characteristics of this technique. The major disadvantage of LC is that 42 peptides, instead of intact proteins, are eluted and analyzed, therefore peptide information has to be merged to resolve the identity of the original protein; thus, relatively low protein identification rates are associated with this technique. Approximately 5-10% of the spectra can be identified, although this is to an extent compensated for by the large number of spectra generated (Conrotto & Souchelnytskyi, 2008).

Recently, a new technique based on protein binding called protein arrays or protein chips has been developed (MacBeath & Schreiber, 2000). Proteins that are known to bind specific proteins are placed on a chip, and the sample is then exposed to the chip, allowing specific binding to occur. Following washing, fluorescence or radio isotope intensities are determined. Limitations with this technique include difficulty in protein quantification and the ability to generate binding antibodies (Conrotto & Souchelnytskyi,

2008). Also, significant knowledge about protein-protein interaction and binding affinity is required for the protein bound to the chip and the protein that will recognize it.

With the advancement in technology, the list of identified plasma proteins is growing. In 2002, about 300 proteins were known (N. Leigh Anderson & Anderson,

2002). Since then, using ultra-high-efficiency strong cation exchange LC/reversed-phase

LC/MS/MS, 800 proteins have been identified in human plasma (Shen et al., 2004).

Compiling from all previous reports, a non-redundant list of 1175 distinct gene products is generated for human plasma proteome (N. Leigh Anderson et al., 2004). Subsequently,

81 more proteins have been identified by multidimensional (multiple rounds of separation based on protein charges and hydrophobicity) LC separation (Sheng et al., 2006), adding to the growing list and results in a total of 1256 identified proteins in human plasma. 43

Plasma proteins as biomarkers

Because blood is easy to obtain and the procedure is less invasive compared to tissue biopsies, it is used routinely for disease diagnostics. Plasma proteins have long been used as biomarkers for diseases. For example, cardiac troponins T (Gerhardt & Ljungdahl,

1998), creatine phosphokinase (Michelson, 1980) and lactate dehydrogenase (Metais &

Sacrez, 1963) in plasma are highly sensitive markers of heart injury. Similarly, liver transaminases (Molander et al., 1955) as well as interleukin-18 levels (Vecchiet et al.,

2005) are closely associated with liver injury. Additionally, plasma biomarkers are used

for clinical diagnosis for various diseases including cancer (Bates, 1991), inflammation

(Ridker et al., 2000), diabetes (Dworacka & Winiarska, 2005), multiple sclerosis

(Angelucci et al., 2005), tuberculosis (Djoba Siawaya et al., 2009), HIV infection (Mine

et al., 2009; Rose et al., 2009) and infections by various pathogens.

Plasma proteins can serve as biomarkers not only for diseases, but also for any

specified condition, state or process. For example, this project was started with the hope

of identifying biomarkers for aging. Also, plasma proteins that change as a function of

hormonal action have been determined. For example, serum IGF-1 levels increase in

response to GH. Thus, IGF-1 is a plasma indicator of GH activity (Kemp et al., 1981; D.

C. Moore et al., 1982; Reiter & Lovinger, 1981).

Similar to research in determining the plasma proteome, proteomic techniques

including 2-DE have been used to discover novel biomarkers of diseases (X. Huang et

al., 2006; Villanueva et al., 2006; R. Zhang et al., 2004). For example, squamous cell

carcinoma antigen 1 was found to be a biomarker of tongue cancer (X. Huang et al., 44

2006). Aiming to uncover specific proteins that show significantly different expression patterns from the entire protein set, proteomic techniques provide a promising approach for biomarker discovery. In order to aid the discovery and characterization of biomarkers, a Human Proteome Detection and Quantitation Project has been founded, attempting to establish splice variants, PTMs, protein-protein interactions, and tissue localization of a given protein via proteomic means (N. Leigh Anderson et al., 2009).

Mouse plasma proteome

The mouse and human share a high degree of genome similarity or homology

(Carver & Stubbs, 1997) and have very similar chromosomal organization and content

("Initial sequencing and comparative analysis of the mouse genome", 2002). About 80%

of mouse proteins have a strict 1:1 orthologue in the human genome ("Initial sequencing

and comparative analysis of the mouse genome", 2002; Pennacchio, 2003). It is much

easier to conduct biological research using the mouse, since genetic and environmental variations can be controlled. Therefore, the mouse is a good model organism for the study of human diseases.

The mouse plasma proteome is less well studied than human. As in human plasma, the majority of the mouse plasma proteome is composed of a few abundant protein species. In a recent study, 12,300 unique peptides that originated from 4567 unique proteins, about 16% of all known mouse genes (which is ~28,000), have been identified through multidimensional fractionation and MS/MS approach in mouse plasma (Figure

8) (Hood et al., 2005). Thus much more research remains to be performed in the mouse plasma proteomic field. 45

Figure 8. Mouse plasma proteome. Adapted from (Hood et al., 2005).

As has been mentioned earlier, the SWISS-2DPAGE database includes 2D annotation of human plasma proteins, but to date, no reference gel exists in the database for the mouse plasma proteome. However, many reports have been published using 2-DE on mouse plasma or serum proteins to uncover biomarkers of diseases including lung adenocarcinoma, hepatocellular carcinoma, tongue cancer, injuries from burn or radiation, as well as biomarkers of therapeutics against snake venom and liver cirrhosis

(Bijon & Jürgen, 2007, 2009; Duan et al., 2004; Gazzana & Borlak, 2008; Guipaud et al.,

2007; Hao et al., 2002; Hsueh-Fen et al., 2004; X. Huang et al., 2006; W. Li et al., 2007;

Ma et al., 2003; Ren et al., 2007; Roberto et al., 2008; Skehel et al., 2000; Yuichiro et al., 2006). Among these papers, a few have published reference 2D PAGE maps (Bijon &

Jürgen, 2007, 2009; Duan et al., 2004; Gazzana & Borlak, 2008; Guipaud et al., 2007;

Ren et al., 2007; Roberto et al., 2008; Skehel et al., 2000). Therefore, 2-DE has become an increasingly popular technique for biomarker discovery in mouse models. 46

Introduction to 2-DE

The proteomic technique 2-DE has been used to separate proteins for several decades. It was first introduced in 1975 (Klose, 1975; O'Farrell, 1975), and a few years later, in 1977, 2-DE was used to isolate human plasma proteins (L. Anderson &

Anderson, 1977). In the first dimension, proteins with different isoelectric points (pIs) are focused at different positions in an immobilized pH gradient (IPG) strip, in which a series of immobilized reagents with a broad pI range are chemically linked to an acrylamide matrix strip. In the second dimension, proteins from the IPG strips are separated by sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDS-PAGE) based on size.

With varying gel concentrations, it is possible to separate proteins in different molecular weight ranges. Following SDS-PAGE, the gels are stained to visualize the proteins, and the gel image is captured using software that determines the density of each protein spot.

From these images, proteins in the gel are profiled and quantified. Proteins of interest are then removed from the gel and identified by MS and/or MS/MS.

Sample preparation

Proteins must be solubilized, denatured, and reduced before being absorbed onto

IPG strips. Cells and/or tissues are usually lysed to release the protein content. Body

fluids such as plasma, serum, urine, or cerebral spinal fluid do not need lysing.

For cultured cells, gentle lysis methods are sufficient to release cellular proteins (T.

S. Berkelman, T, 1998). Osmotic lysis works well with blood cells and cultured cells,

which are suspended in a hypotonic solution and burst because of osmotic pressure

(Ramsby et al., 1994). Sonication is also a method employed to lyse cells. One cautionary 47 note is that prolonged sonication can overheat the sample. Therefore, samples are usually sonicated with short bursts and cooled on ice between bursts. Another method is freeze- thaw lysis (Toda et al., 1994). Cells are rapidly frozen using liquid nitrogen then thawed in one or more cycles. For tissues that are difficult to disrupt such as liver, kidney, and heart, a mortar and pestle can be used to grind the tissues frozen in liquid nitrogen (Gorg et al., 1988). Various mechanical homogenization devices are also available to efficiently lyse many types of tissues (Goldberg, 2008; Posch et al., 1995).

Cells and tissues are usually lysed in the presence of detergents to solubilize and denature proteins. Ionic detergents are avoided since they interfere with the subsequent

IEF. If ionic detergents such as SDS are used, the sample must be diluted by excess non- ionic detergents or removed by acetone precipitation. Centrifugation is usually carried out afterwards to collect the protein supernatant, or differential centrifugation can be performed to recover proteins in various cellular organelles (Castle, 2003).

Contaminant removal

Often tissue or blood samples need to be further processed to remove contaminants

such as salt, DNA, RNA, and polysaccharides that may interfere with the separation or

visualization of proteins on a 2D gel. Excess salt and other small ionic molecules make the sample more conductive and should be kept at a minimum prior to IEF. Dialysis or precipitation can be used to desalt the sample. DNA and RNA can increase sample viscosity, and may result in background protein smearing or ‘clogged’ gel pores.

DNase/RNase is used to hydrolyze the nucleic acid polymers in the sample. Large

polysaccharides also can block gel pores and cause horizontal streaking. They are 48 removed by precipitation with ammonium sulfate or phenol/ammonium acetate. Lipids can reduce the solubility and change the pI and/or molecular weight (Mw) of proteins.

They can be removed by excess detergents or acetone precipitation (T. Berkelman, 2008).

This is especially important when analyzing adipose tissue samples.

Protein quantification, solubilization/denaturation/reduction/alkylation

After the preliminary sample treatment, the protein content of the sample must be

quantified to determine the volume and mass to be used for 2-DE. Standard protein

assays, such as the Bradford assay, are used to determine protein concentration in a sample. In Bradford assay, a protein specimen, along with a known gradient concentration of a standard protein such as bovine serum albumin, reacts with a

Rosaniline dye, and the protein concentration in the specimen is determined by colorimetry (Bradford, 1976). The proper amount of protein needed for 2-DE (for example, 500ug) is calculated and diluted in a buffer containing urea, thiourea, non-ionic and/or zwitterionic detergents, a reducing agent, carrier ampholytes and protease inhibitors (Kopchick et al., 2002). Urea is a neutral chaotrope that denatures proteins

(Rabilloud & Chevallet, 2000). Heating aids the solubilization but should be avoided because with heat urea will break down into ammonium cyanate, which can cause protein charge modifications (Cejka et al., 1968). Thiourea is often used in addition to urea to improve solubilization and denaturation. Non-ionic detergents such as NP-40 and Triton

X-100, and/or zwitterionic detergents such as CHAPS, are included to further solubilize proteins and prevent hydrophobic aggregation. Carrier ampholytes are added to create a pH gradient during IEF. A reducing agent such as dithiothreitol (DTT), beta- 49 mercaptoethanol, or tributylphosphine (TBP) is added to reduce disulfide bonds

(Kopchick et al., 2002). Following reduction, the sulfhydryl groups on the cysteine residues should be alkylated by an agent, such as iodoacetamide, to prevent sulfide bond reformation (Herbert et al., 2001). During sample preparation, protease inhibitor cocktails are usually added at the time of solubilization/denaturation/reduction, and are often added as early as cell/tissue lysis (T. S. Berkelman, T, 1998).

Electrophoresis

After sample preparation, proteins are first separated by IEF. Two methods exist for creating a pH gradient. One uses carrier ampholytes to generate pH gradients in polyacrylamide tube gels. A more widely-used method is the use of IPG strips. The strips are generated by covalently linking acidic and basic buffering groups to a polyacrylamide gel strip. This eliminates the ‘drifting’ which can occur in the carrier ampholyte-based method. IPG strips are commercially available in various pH ranges with linear or non- linear pH gradients that amplify a specific pI range (T. S. Berkelman, T, 1998).

After IEF, proteins in the strip are subjected to the second dimension SDS-PAGE

(Laemmli, 1970). In this procedure, the proteins are coated with SDS in proportion to their mass (Shapiro et al., 1967), which makes them negatively charged. Because of the

‘pores’ in the polyacrylamide, larger proteins move more slowly and smaller proteins faster; thus, the separation is based on the mass of the proteins. Also, one can selectively separate proteins in a particular molecular mass range using specific concentration of acrylamide in the resolving gel. Larger proteins would be separated better in lower percentages of acrylamide and vice versa for smaller proteins. 50

Staining of protein spots

Following SDS–PAGE, proteins are stained using a variety of dyes including

coomassie blue or a silver stain. Coomassie blue staining is simple and quantitative but

relatively insensitive to small amount of proteins. Silver staining is sensitive but the

linearity is not satisfactory. In addition, silver stain interferes with subsequent MS

analysis (Lauber et al., 2001). Fluorescent stains such as sypro red, ruby, and orange

have high sensitivity, an excellent linear range and does not affect subsequent MS

analysis.

Plasma as a sample for 2-DE

Blood is the most common tissue collected for clinical diagnostics. Centrifugation

at 7,000 x g for 10min effectively removes the blood cells and the supernatant plasma is readily soluble. The sample can be treated for 1st dimension IEF directly, since normally very low concentrations of DNA, RNA, large polysaccharides or lipids are present in blood. See Figure 9B for a typical 2D gel image of mouse plasma proteins.

As mentioned in the previous section, a few abundant proteins dominate in plasma, obscuring the presence of other less abundant proteins. Only 1% of plasma proteome, albeit in low abundance, contains a large amount of unique proteins and therefore represents a biomarker reservoir. In a typical 2D gel of plasma proteins, albumin will appear as a huge blot masking many other proteins (Figure 9A). Therefore, it is necessary to deplete albumin and other high-abundance proteins in order to resolve the remaining proteins. Human albumin can be depleted with immunoaffinity resin (Steel et al., 2003), 51 resulting in greatly improved resolution of the 2D image (Echan et al., 2005). However, albumin depletion often causes loss of other proteins that bind to it (Granger et al., 2005).

Figure 9. 2D gel images of plasma proteins. A, 2D gel image of plasma proteins with albumin smear. Adapted from (Ding et al., 2009). B, 2D gel image of mouse plasma proteins. Enclosed in the box are smaller proteins that are not affected by albumin smear.

Recently a new technique using a hexapeptide ligand library has been developed

(Guerrier et al., 2008). Plasma proteins are bound by their respective hexapeptide ligands; the more abundant a protein is, the more likely it will be bound, resulting in enriched population of less abundant proteins. This technique minimizes the loss of proteins as can occur in immunodepletion, but requires a large sample volume (~1 ml plasma). This is not a problem in human samples, but for animal studies such as mice or rats, it is difficult to obtain such a large volume per animal.

A method used to circumvent the albumin smear as well as the pitfalls of albumin- depletion is simply to selectively study only proteins with a relative small molecular weight. Albumin’s molecular mass is ~70 kDa, the majority of which consists of a protein smear above 50 kDa on a 2-D gel. By increasing the 2nd dimension gel 52 concentration to 15%, proteins larger than 45 kDa are clustered as horizontal streaks in the upper region of the gel (Figure 9B). When focusing only on proteins smaller than 45 kDa, the albumin smear is not encountered (Figure 9B).

This method has limitation on the size range of proteins measured, but provides a faster, simpler way to study plasma protein profile. There are already a handful of proteins in the low molecular weight (LMW) range that have been identified, and studies have shown that LMW proteins below 30 kDa contain comprehensive categories of proteins including structural proteins, nuclear proteins, transcription factors, channels, receptors, enzymes, secreted proteins, coagulation and complement factors, transport and binding proteins, proteases, protease inhibitors, homones and cytokines (Figure 10)

(Tirumalai et al., 2003) and have great potential for biomarker discovery (Petricoin et al.,

2002).

Figure 10. Relative numbers of proteins identified in LMW proteome. Hypothetical proteins are proteins that have not been experimentally identified but whose sequences are predicted from known open reading frames. Reproduced from (Tirumalai et al., 2003). 53

Review of GH Doping and Detection

Recombinant human GH doping in sports

Because GH promotes bone, cartilage and muscle growth as well as fat degradation,

athletes often abuse recombinant human GH (rhGH) to enhance performance. RhGH is

on the list of prohibited substances by the World Anti Doping Agency (WADA)

(WADA, 2009). Since GH was first extracted from human pituitary, it has been used in clinical practice to treat children with GH deficiency (D) (Figure 11). RhGH has been

approved by the FDA to treat children with the following conditions: Prader-Willi

Syndrome, children born small for gestational age, Turner’s Syndrome, Noonan’s

Syndrome, chronic renal failure, idiopathic short stature and cystic fibrosis (Krysiak et al., 2007); in adults, rhGH has been approved to treat adult GHD (Cummings & Merriam,

1999), wasting syndrome of AIDS (Mulligan et al., 1993; Schambelan et al., 1996) and short bowl syndrome (Krysiak et al., 2007). However, with the beneficial effects of GH come the abuse and misuse of GH among athletes. Although officially on the prohibited list as early as 1989, it continues being abused and more and more cases and scandals have been exposed (Figure 11). It is estimated that the dose of rhGH abuse is 10-25

IU/day (Saugy et al., 2006), which is much higher than therapeutic doses for adult GHD, which is 0.2-0.3mg/day (~0.6-0.9 IU/day) (Ho, 2007). Moreover, rhGH is thought to be used in combination with other doping agents such as anabolic steroids or erythropoietin

(Saugy et al., 2006).

Figure 11. History of GH in clinical practice and abuse by professional athletes. Reproduced from (Holt et al., 2009b).

It is difficult to detect GH doping. First, rhGH is identical to the endogenous hGH.

Thus, it is difficult to directly detect ‘doped’ GH levels in the blood. Second, endogenous hGH is secreted in a pulsatile manner and the levels are affected by many environmental factors such as exercise, sleep, stress and nutritional status (Garlaschi et al., 1975;

Reinhard W. Holl et al., 1991; Kopchick, 2003; Krulich et al., 1974; Manson & Wilmore,

1986; Osterman et al., 1974; Rice et al., 1975; Sutton & Lazarus, 1976). Third, GH has a very short plasma half life of about 15min (R. W. Holl et al., 1993). Exogenous GH is cleared quickly from circulation (Owens et al., 1973). Thus, ‘the window of opportunity’ for detection of rhGH is only 24-36h (Wu et al., 1999). Urine samples, often used to test drug doping among athletes, is not feasible because urine GH levels are very low (Albini 55 et al., 1988) and poorly correlated with serum GH levels (Flanagan et al., 1997). To date, blood is the most common biological fluid used for GH doping detection.

Current approaches to detect GH doping

To date, there are two approaches to detect rhGH doping in blood. One is based on

the different isoforms of hGH. Endogenous GH, produced in the somatotrophs of the

pituitary gland, has several forms including the most abundant 22 kDa isoform, a 20 kDa

isoform (Baumann) generated by alternative precursor RNA splicing (Leung et al., 2002),

and other minor-isoforms with sizes of 17.08 kDa and 17.84 kDa. These latter isoforms

are detected by proteomic analysis in the human pituitary gland but not plasma (Zhan et

al., 2005), although the degraded GH isoforms might be due to a sample extraction

artifact (Baumann). Also, hGH is subjected to PTMs including acetylation, deamidation,

phosphorylation and possibly glycosylation (Baumann; Zhan et al., 2005). It is well

known that IGF-1 is induced by GH, resulting in a feedback inhibition of endogenous

hGH secretion. Since rhGH is composed of only the 22 kDa isoform, a change in the ratio

of 22 kDa isoform over the total GH makes it possible to detect whether exogenous rhGH

has been used (Bidlingmaier et al., 2000; Wu et al., 1999).

The limitation of this approach is that the window of opportunity for detection is

short (Nelson & Ho, 2008) and it is impractical to draw blood from athletes daily. The

GH isoform method was adopted by WADA for the 2004 Athens and 2006 Turin

Olympic Games. However, no positive results were found possibly due to the above

reasoning. 56

A second approach to detect GH doping is to determine GH-dependent biomarkers with longer half lives than GH. Two groups of proteins are currently used for this purpose. One is IGF-1/ IGF binding proteins (IGFBPs) and the other includes proteins involved in bone and collagen turnover. A study entitled ‘The GH-2000 project’ was initiated to search for GH specific biomarkers in human serum (Dall et al., 2000; Holt et al., 2009b; Longobardi et al., 2000; Sonksen). In a randomized, double blind, placebo- controlled study involving healthy volunteers of both sexes, daily rhGH was administrated (0.1 IU/kg /day and 0.2 IU/kg/day) for 4 weeks followed by a 8-week wash-out period and resulted in increases in IGF-1, acid- labile subunit (ALS) and

IGFBP-3 levels (Dall et al., 2000). Markers of bone and collagen turnover were also increased. These included osteocalcin, a C-terminal propeptide of type I procollagen (P-I-

P), a bone resorption marker type I collagen telopeptide (ICTP), and a soft tissue marker procollagen type III (P-III-P). Of these markers, only P-III-P and osteocalcin remained significantly elevated after the 8-week wash-out period (Longobardi et al., 2000). IGF-1 and P-III-P are considered ideal candidate markers because of little variation from day-to- day, or by exercise and gender (Sonksen). In ’the GH-2000 project’, females were found to be less responsive than males to GH treatment as has been reported by others (Nelson et al., 2008). This is presumably due to the antagonistic actions of estrogen on GH action

(Meinhardt & Ho, 2007). Although females are less responsive, the GH-2000 project has established IGF-1 and P-III-P (Figure 12) as two biomarkers of GH that may be used to detect GH doping.

Figure 12. Individual datapoint plot of P-III-P against IGF-I levels in male subjects after 3 weeks of treatment. Reproduced from (Sonksen, 2009).

Injury is quite common among athletes and it can cause changes in serum bone and collagen markers. Concerns about injury and ethnicity differences on GH doping markers resulted in a new phase of study: the GH-2004 project (Holt et al., 2009a; Holt et al.,

2009b). Elite and amateur athletes have been followed longitudinally for sport-related injury. IGF-1 does not change after injury. A raise in P-III-P serum level is found; however, this rise is minimal compared to GH’s effect (Erotokritou-Mulligan et al.,

2008a). Ethnicity reveals some differences in IGF-1 and P-III-P, but still within 99% limit of confidence (Erotokritou-Mulligan et al., 2008b; Erotokritou-Mulligan et al.,

2009). Similarly, another study found that ethnicity is a moderate effecter on GH- responsive markers (Nelson et al., 2006).

Another independent double blind, placebo controlled rhGH injection study, the

Kreischa study, confirmed IGF-1 and P-III-P as GH-responsive markers (Kniess et al.,

2003). These two studies have developed respective algorithm to calculate a score to 58

predict the likelihood of GH doping. The GH-2000 fomula [GH-2000 test score= -

2.269+0.7207 log (P-III-P)+0.5210 log(IGF-1)] is able to detect 90% of those receiving

rhGH in the Kreischa study with a similar sensitivity. The Kreischa formula [Kreischa

test score= -13.465+0.0272×IGF-I+0.0398×IGFBP-3+1.367P-III-P-0.00271×(IGF-IP-III-

P)] can correctly identify 41% of individuals receiving GH in the GH-2000 study

(Erotokritou-Mulligan et al., 2007).

Novel biomarkers for GH are needed

GH is affected by many factors such as exercise, gender, age, body mass index

(BMI), etc. The effect of maximum exercise on endogenous hGH levels has been examined in elite athletes (Ehrnborg et al., 2003). HGH shows a peak concentration by the end of the exercise period and decreases to baseline levels within 30-60 min.

Similarly, IGF-I, IGFBP-3, ALS and IGFBP-2 increase following exercise, with IGF-I,

IGFBP-3 and ALS returning to baseline by 3-4 days. The bone markers ICTP and P-III-P

increase after exercise but return to normal within 120 min whereas osteocalcin and P-I-P

are not affected by exercise (Ehrnborg et al., 2003). This study demonstrates that although exercise increases the levels of several biomarkers, the effect is acute and diminishes in a relatively short period of time (J. D. Wallace et al., 1999; J. D. Wallace et

al., 2000).

Demographic factors and sport activity or category have been examined for their

effects on GH responsive biomarkers (Nelson et al., 2006). Age is the major contributor

to variability in these marker values that significantly decrease with age. Gender

significantly confounds the variability of IGFBP-3 and ALS, whereas BMI and sport 59 category are only modest effecters. Taken together, these results point out that reference ranges of biomarkers must be established for different factors including age and gender.

In addition, the candidate biomarkers IGF-1 and P-III-P have relatively large variations among individuals or even within the same person. For example, serum samples obtained over a 2- to 3-week period from 1103 elite athletes have shown that the within-subject coefficient of variation (CV, a value to measure the variance relative to the mean) is nearly 21% for the IGF axis markers and 15% for the collagen markers (Nguyen et al., 2008). In addition, in some cases, IGF-1 levels cannot properly reflect that of GH.

For example, some acromegalic patients have high levels of IGF-1 but GH levels are normal; thus, the correlation of GH and IGF-1 is poor. Also, after treatment to normalize

GH, IGF-1 does not always normalize (Barkan, 2007; Clemmons, 2006; Popovic, 2005;

Stonesifer et al., 1981). Thus, other plasma biomarkers for GH action would be valuable.

It is generally agreed that a single biomarker is not a robust indicator of GH doping; rather, a combination of multiple biomarkers will generate more robust and reliable results.

Theoretically, the whole plasma proteome could be profiled before and after rhGH

treatment and those that are differentially regulated could be identified. Thus, proteomics

represents a novel and highly promising approach to discover additional biomarkers for

rhGH doping. Surface plasmon resonance (SPR) immune assays (Gutiérrez-Gallego et

al., 2009) and surface-enhanced laser desorption ionization (SELDI)–TOF MS (Chung &

Baxter, 2009) are two emerging proteomic techniques to uncover new GH specific

markers. SPR is the phenomenon which occurs when light is reflected off thin metal 60 films. With this technique, a thin metal film is immersed in a medium. A change in the light reflective angle (θ) and the light intensity can be observed when a difference in the medium occurs such as when GH binds to its antibody. SPR immune assays thus can differentiate between different GH isoforms which produce different θ and/or light intensities when binding to an antibody (Gutiérrez-Gallego et al., 2009).

SELDI-TOF MS is similar to MALDI-TOF MS introduced earlier in the section.

Like in MALDI, digested peptides are mixed with the matrix, but in SELDI, the mixture is applied to a surface that binds to the peptides with different affinities, instead of subjected to TOF directly as in MALDI. Thus, SELDI has one dimension of peptide separation before TOF. Recently, hemoglobin α-chain has been discovered to be a biomarker of GH action using this SELDI-TOF MS technique (Chung et al., 2006), although the robustness of this marker is brought into question because subsequent analysis find this marker to be at very low levels or even undetectable, possibly due to a different materiasl of the tubes used for sample collection (Chung & Baxter, 2009). To date, a validated novel biomarker of GH is yet to be discovered. One project of this Ph.D. dissertation involves using a classic proteomic tool, 2-DE, to discover biomarkers for

GH.

CHAPTER II. RESEARCH OBJECTIVES

My research interest is to discover novel biomarkers of GH as it relates to aging.

These markers will not only serve as indicators of GH and aging, but also provide insight into the mechanism(s) of aging as well as the in vivo interaction between GH and aging.

Since GH is involved in the aging process, it would be interesting to study GH action in the context of aging. Mouse models with altered GH signaling, specifically, bGH and

GHR-/- mice were used in my research. My specific research objectives are to discover the biomarkers of aging and GH action.

Longitudinal studies involving bGH transgenic mice and littermate controls were carried out to identify biomarkers of aging in the context of constant and elevated levels of GH action. Longitudinal studies of GHR-/- mice and littermate controls (male and female) were carried out to identify biomarkers of aging in the context of the lack of GH action. Blood from these mice were collected periodically and plasma samples gathered for 2-DE analysis. Plasma proteins that exhibited significant differences during aging, between different genotypes as well as gender differences were identified. To complement the longitudinal studies, a cross-sectional study involving WT male mice was designed to confirm the biomarkers of aging found in the longitudinal study.

My studies of GH transgenic and GHR-/- mice involved mice in which the transgene was constitutively present or absent. I have therefore exended my studies of the plasma proteome to mice injected with either GH or IGF-1.The goal was to discover biomarkers of short-term action of GH and IGF-1. This study will hopefully determine

GH and IGF-1 responsive proteins that could then be evaluated in human GH doping 62 studies. Ultimately, it may result in the development of new targets for GH doping detection.

I will organize my results in four papers in the following four chapters (chapters III to VI). The first one (chapter III) is centered on aging biomarkers in WT male mice, both in a longitudinal and a cross-sectional setting. The first paper will also include the 2D map and identification of all plasma proteins pertinent to this dissertation. The first paper also describes in detail the key techniques and methods such as 2-DE and MS which are referred to by subsequent papers. The second paper (chapter IV) involves plasma proteins found in male bGH compared to WT mice during different ages. The third one (chatper

V) focuses on GHR-/- and WT mice of both genders at different ages. The final manuscript (chatper VI) is the short-term GH/IGF-1 injection of WT male mice. Each of these papers is independent; however, each subsequent paper also relates to the previous ones. The four papers are summarized in Table 5.

Table 5. Organization of four papers in this dissertation Paper 1 Paper 2 Paper 3 Paper 4 Chapter III Chapter IV Chapter V Chapter VI Mouse WT (C57BL/6J) WT and bGH WT and GHR-/- WT (C57BL/6J) (C57BL/6J) (C57BL/6J) Gender Male Male Male and female Male

Ages Longitudinal: 2, 4, Longitudinal: 2, Longitudinal: 8, Cross-sectional: 3 8, 12, 16, 19 4, 8, 12, 16, 19 16, 24 months months months months Cross-sectional: 6, 12, 24 months Objective Biomarkers of Biomarkers of Biomarkers of lack Biomarkers of aging chronic GH of GH action; short-term GH action; aging biomarkers of action (GH pattern of bGH gender; interaction doping); versus WT mice of gender and biomarkers of aging; interaction IGF-1 doping; of gender and differential effects genotype of GH and IGF-1 Parameters Plasma proteomics measured Body weight and body composition Fasting glucose and insulin (except for paper 4) IGF-1 (except for paper 4)

My hypothesis is as the following: 1) The levels of a subset of plasma proteins will change during aging. These proteins will be biomarkers of aging. 2) The levels of a subset of plasma proteins will be up or downregulated in bGH compared to WT mice.

These proteins will be biomarkers of (chronic) GH action. 3) The levels of a subset of plasma proteins will be up or downregulated in GHR-/- compared to WT mice. These proteins will be biomarkers of lack of GH action. 4) A subset of plasma proteins will show a different trend of change during aging in bGH and/or GHR-/- mice compared to

WT mice; these will be biomarkers of the interaction between aging and GH. 5) A subset of plasma proteins will show a difference between males and females. 6) A subset of 64 plasma proteins will be different in GH or IGF-1 injected mice versus controls. These proteins will represent biomarkers of short-term GH or IGF-1 action. 7) A subset of plasma proteins may be different bewteen GH and IGF-1 treatments. These proteins will be biomarkers to differentiate between GH and IGF-1 actions.

CHAPTER III. BIOMARKERS OF AGING: PROTEOMIC ANALYSIS OF PLASMA

FROM WILD TYPE MALE MICE OF DIFFERENT AGES

Abstract

To discover biomarkers for aging, we analyzed plasma proteins of normal male mice using two-dimensional electrophoresis coupled with mass spectrometry. This proteomic approach allowed us to identify 86 plasma proteins, 13 of which exhibited significant changes during aging. Interestingly, many of these proteins exist as protein isoforms with distinct masses and/or charges probably resulting from yet unknown post- translational modifications. The levels of several plasma proteins changed through aging.

For example, several proteins increased significantly including one isoform of transthyretin, two isoforms of haptoglobin and three isoforms of immunoglobulin kappa chain. Conversely, several proteins decreased significantly during aging, including peroxiredoxin-2, serum amyloid A-1 and five isoforms of albumin. Thus, proteomic analysis of plasma proteins has identified several plasma markers of mouse aging.

Introduction

Aging is an inevitable process that eventually leads to death. Through medical intervention, it is now being debated whether there are viable therapeutic targets that would delay the onset of or slow the aging process. If possible, therapeutic agents may ultimately extend human life span, and perhaps more importantly, improve the quality of life or the healthspan of the elderly (Benedetti et al., 2008; Boudrault et al., 2008; Nass et al., 2009; Topinkova, 2008).

Biomarkers of aging are biological characteristics that indicate or represent certain age periods. For example, there are markers for puberty- Tanner stages of puberty

(Marshall, 1986), adulthood, middle age and old age. Traditional biomarkers for old age include wrinkles, gray hair, reduced muscle strength, and increased susceptibility to infections. In a general term, old age is associated with frailty, a condition featured with disability and comorbidities (Nass et al., 2009; Topinkova, 2008). These macroscopic traits, however, are composed of complex physiological processes and are therefore difficult to study. Other aging biomarkers currently in use include evaluation of cardiorespiratory function (blood pressure, forced expiratory capacity, etc.), hemotology

(red and white blood cells, hemotocrit [percentage volume that red blood cells occupy in blood]), and serum biochemical markers including the levels of albumin, globulin and blood urea nitrogen (Nakamura & Miyao, 2007, 2008). It has been found that systolic blood pressure and blood urea nitrogen increase with chronological age, whereas forced expiratory volume in 1 second, hemotocrit and albumin decrease (Nakamura & Miyao,

2007). Another example of a biomarker is dehydroepiandrosterone sulfate, an adrenal 67 steroid sulphate, which has been found to decline during normal aging in humans

(Orentreich et al., 1992) and rhesus monkeys (Lane et al., 1997).

Serum/plasma proteins are relatively easy to access and are widely used for disease diagnosis. Serum/plasma biomarkers of aging will not only facilitate the elucidation of the physiology processes of aging, but also may provide insight in therapeutic targets and intervention strategies to delay and/or slow aging.

Proteomic techniques including two-dimensional electrophoresis (2-DE) have been used to study aging (Ballesteros et al., 2001; Gromov et al., 2003; Miura et al., 2007;

Sato et al., 2006; Schöneich, 2003; Toda, 2000; Yuri Miura, 2005). This technique has led to the identification of many candidate proteins that are altered by aging. For example, Prx-2, an antioxidant enzyme, is able to be induced by irradiation in cultured astrocytes from young but less so from old rats; also phosphorylation of α-tubulin is increased with aging (Miura et al., 2007; Yuri Miura, 2005). Senescent bacteria cells produce aberrant isoforms of proteins that are oxidized (Ballesteros et al., 2001). Old human skin expresses altered levels of certain proteins, including manganese-superoxide dismutase, tryptophanyl-tRNA synthetase, the p85ß subunit of phosphatidylinositol 3- kinase, and proteasomal proteins PA28-α (Gromov et al., 2003). N-glycosylated proteins are found to accumulate in aged rat cerebral cortex (Sato et al., 2006). Since proteomics reveals the total detectable proteins in a given sample, this approach will make it possible to identify proteins involved in the aging process. Here we use 2-DE as a tool to uncover biomarkers of aging in mouse plasma. 68

Materials and Methods

Experimental animals

For the longitudinal study, male C57 BL/6J mice (n=8) were followed through 2, 4,

8, 12, 16 and 19 months of age. These ages represent post-puberty (2 months), young

adult (4 months), adult (8 months), middle-age (12 and 16 months) and old age (19

months) (Richard A. Miller & Nadon, 2000). For the cross-sectional study, three different

groups of male C57 BL/6J mice of 6 months (n=6), 12 months (n=8) and 24 months

(n=8) were used. These represent young adult, middle age and old age in mice,

respectively (Richard A. Miller & Nadon, 2000). Mice were housed 2-3 per cage at room

temperature (22 oC) in a 12-hour light, 12-hour dark cycle. All mice were fed ad lib with

a typical chow diet. Animal protocols were approved by Ohio University’s Institutional

Animal Care and Use Committee.

Body composition measurement

Body composition was measured using an NMR analyzer the Bruker Minispec (The

Woodlands, TX) as described previously (List et al., 2009; Palmer et al., 2009). The data

are presented as absolute mass (gram or g) of lean mass, fat mass and fluid mass. Each

animal was measured twice and the mean value was reported.

Plasma collection

Blood was collected into heparinized capillary tubes following tail tip clipping after

heat lamp exposure. Whole blood was centrifuged at 7000 x g for 10 minutes at 4 ºC and

the resulting plasma was stored at -80 ºC.

Fasting glucose and insulin measurements

Mice used in the longitudinal study were fasted for 4 hours and bled at 3 PM. Blood

glucose was measured by ONE TOUCH glucometer from Lifescan (Milpitis, CA). 69

Plasma insulin levels were determined using an ultrasensitive rat/mouse insulin ELISA kit (ALPCO, Windham, NH).

2-DE

2-DE was performed as previously described (Ding et al., 2009; Kopchick et al.,

2002; List et al., 2007; Qiu et al., 2005; Sackmann-Sala et al., 2009). 2-DE was carried

out within a week after plasma collection to avoid possible sample degradation.

Plasma proteins were treated for 2 hours at room temperature with sample buffer containing 8M urea, 1.8M thiourea, 4% zwitterionic detergent (CHAPS), and 5mM reducing agent tributylphosphine, in a protease inhibitor cocktail containing AEBSF,

aprotinin, leupeptin, bestatin, pepstatin A and E-64 (Sigma-Aldrich, Inc., St. Louis, MO).

Then 15mM iodoacetamide was added for alkylation for 30min. 500ug protein was loaded onto a 17cm immobile pH gradient gel (IPG) strip with a broad linear pI range of

3–10 (Bio-Rad, Hercules, CA). After actively rehydrated (50V) for 12 hours at 20 ˚C using a Protean IEF cell (Bio-Rad), the strips were subjected to first dimensional isoelectric focusing (IEF) at 4000V for 60000V-Hr. When IEF was completed, strips were incubated in buffer containing 2% (w/v) sodium dodecyl sulfate (SDS), 0.5M

Tris/HCl (pH 6.8), 20% (v/v) glycerol for 25min. The middle section of the strip with pI

5-8 bearing the majority of plasma proteins was removed and subjected to second

dimension SDS polyacrylamide gel electrophoresis (PAGE). Only the middle section of

the strip was used instead of the full length because the middle section contained most of

the proteins already and the full length strip required a large gel. Polyacrylamide (15%)

gels (8 x 7cm) were used for the 2nd dimension electrophoresis at a current of 25mA/gel until a total of 250 V-Hr was reached. After electrophoresis, the gels were fixed in a

solution containing 40% ethanol and 2% acetic acid overnight followed by washing three 70 times in 2% acetic acid. The gels were then stained with SYPRO Orange (1:5000)

(Molecular Probes, Eugene, OR).

Quantification of proteins

Gel images were captured using a laser-scanner Pharos FX plus (Bio-Rad) with an

excitation wavelength of 488nm and an emission wavelength of 604nm. Proteins were

matched across all images using PDQuest (Bio-Rad) software and manually checked and

corrected when necessary. For quantification, the intensity of each protein spot was

determined according to the fluorescence signal strength, and then normalized by the total

density of each image using PDQuest software. The data were exported to an excel

sheet, log-transformed and then subjected to statistical analysis.

Statistical analysis

All statistical analyses were performed using SPSS version 14.0 software (Chicago,

IL). For the longitudinal experiment, data were subjected to ‘repeated measures’ for

significant effect of aging (p<0.05 for insulin or glucose and p<0.01 as a more stringent

cutoff for protein intensity). For the cross-sectional experiment, data were subjected to

one way ANOVA (p<0.01 for protein intensity). In the analysis of body composition, one way ANOVA was applied followed by LSD as a post hoc test (p<0.05). All data were

presented as the mean ± SEM.

Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography

(LC)-MS/MS

Proteins of interest were excised manually from the SDS-PAGE gels, lyophilized

and shipped to Protea Biosciences, Inc. (Morgantown, WV) for MS and MS/MS analyses 71 using matrix assisted-laser desorption ionization (MALDI)-time of flight (TOF) and

MALDI-TOF-TOF. Selected proteins were also subjected to LC-MS/MS.

Protein in-gel digestion (Performed at Protea Biosciences, Inc.)

Gel plugs were treated with acetonitrile and 50mM ammonium bicarbonate, then reduced and alkylated with 250mM dithiothreitol (DTT) (60min / 55°C) and 650mM iodoacetamide (60min / room temperature / in the dark). Digestion was performed with

500ng trypsin in 50mM ammonium bicarbonate buffer overnight. Peptides were extracted using 5% formic acid in 50% acetonitrile (dehydration), followed by rehydration with

50mM ammonium bicarbonate. For each extraction step, the solution was aspirated, collected, and collated. Three extraction cycles (dehydration and rehydration) were performed per sample. The recovered peptides were lyophilized, reconstituted in 10mM acetic acid, and re-lyophilized to yield a purified, protein digest.

MALDI target spotting (Performed at Protea Biosciences, Inc.)

After in-gel digestion by trypsin, proteins were analyzed by ABI 4800 MALDI TOF-

TOF analyzer. A C18 ProteaTip was washed and then equilibrated a using 0.1%

trifluoroacetic acid (TFA) / 50% acetonitrile solution and a 0.1% TFA / 2% acetonitrile

solution. The remaining reconstituted protein digest solution in an auto sampler vial

(~65% of sample) was loaded onto the C18 ProteaTip by aspirating and expelling the sample 5-10 times within the sample vial. The bound sample was washed twice with the

0.1%TFA / 2% acetonitrile solution by aspirating and expelling 20µL of the wash solution 5-10 times. The sample was spotted directly onto a MALDI target that was pre- 72 spotted with 0.6µL MALDI matrix α-cyano-4-hydroxycinnamic acid (CHCA) using 1µL

of an elution solution (0.1%TFA / 90% acetonitrile).

The MALDI MS parameters used for analyses were: MS acquisition in reflector

mode; positive ion mode; mass range = mass/charge (m/z) = 850 – 4000; 400 laser shots

per spectrum; minimum signal/noise (S/N) = 10 for MS acquisition; 15 strongest

precursors chosen for MS/MS; minimum S/N = 30 for MS/MS precursors; MALDI spot

interrogated until at least 4 peaks in the MSMS spectra achieved a S/N = 70.

Protein identification from combined MS and MS/MS data (Performed at Protea

Biosciences, Inc.)

The Applied Biosystems GPS Explorer ™v3.6 program was used for MS/MS data

processing with Mascot (Mascot, 2009) as the search engine. Other parameters included:

sample type: gel samples; digestion enzyme: trypsin; species: mouse; database: NCBInr;

type of search: combined MS and MS/MS; mass values: monoisotopic; protein mass:

unrestricted; peptide mass tolerance: ± 1 Da; maximum missed cleavages: 2; variable

modifications: oxidation (M), carbamidomethyl (C); exclusion mass list: 1151.8, 1358.9,

1795.1, 2211.4, 2225.4, 2283.

LC-MALDI spotting

Some protein spots that did not yield good identification by MS/MS were selected

for LC- MALDI analysis by ABI Tempo LC MALDI using Tempo LC MALDI v.2.00.09

as a data acquisition and processing program. Lyophilized digested samples were

reconstituted in 7µL of 0.1% TFA in deionized water and 5µL of sample was injected

into a separation column (Merck Chromolith CapRod monolith column – 150 X 0.1mm) 73 with a separation gradient using 0.1% acetic acid 2% acetonitrile and 0.1 % acetic acid in

90% acetonitrile for 30 min.

Protein identification from LC-MALDI MS/MS data

The ABI Protein ProteinPilot Software 2.0 Program was used for MS/MS data processing with ABI’s MS/MS search engine software Paragon. Other parameters included sample type: identification; digestion enzyme: trypsin; special factors: gel-based

I.D.; species: mouse; I.D Focus: biological modifications; database: NCBInr.

Manual confirmation of protein identification from MS and MS/MS data

(performed at Ohio University)

MS and MS/MS data were manually submitted to MASCOT (Mascot, 2009) for confirmation of the results reported by Protea Biosciences. For MS data, the searching criteria were as follows: NCBInr as the database; mouse as the species; trypsin digestion; maximum one missed cleavage; fixed carbamidomethylation of Cys; variable modifications of oxidation-M (methionine), pyro-Glu; monoisotopic; and 50 ppm of peptide mass or parent tolerance. For MS/MS ion search, in addition to the above conditions, a peptide charge of +1 and a fragment mass tolerance of 0.5 Da were used.

Results

Body composition, insulin and glucose levels during aging

There were significant age differences in body weight, lean, fat and fluid masses

(one-way ANOVA, p<0.01, cross-sectional setting). Body weight and lean mass became significantly higher in mice of 12 months compared to 6 months and did not change thereafter (Figure 13A,C). Fat mass increased from 6 months to 12, 19 and 24 months, but decreased from 19 to 24 months (Figure 13D). As a result, body weight increased as

mice aged but slightly decreased after 19 months. Fluid mass became higher at 9 and 12

months and increased further at 19 and 24 months (Figure 13E). The changes in body

composition can be seen in Figure 13B, in which lean and fluid masses increased,

whereas fat mass increased till 19 months and then decreased at 24 months.

Figure 13. Body composition of WT mice at different ages. Data were from different animals during aging and subjected to one-way ANOVA followed by LSD as a post-hoc 75 test. Male mice of 6 mo (n=12), 9 mo (n=8), 12 mo (n=8), 19 mo (n=8) and 24 mo (n=9) were used. A, body weight; B, stacking bars of mean body composition; C, lean mass; D, fat mass; E, fluid mass. Different letters denote significant difference (p<0.05). mo: month.

Fasting blood insulin levels significantly increased during aging (repeated measures, p<0.05; longitudinal setting) with the significant increase occurring between 9 and 12 months (paired sample t test, p<0.05, Figure 14). After 12 months, the fasting insulin levels tended to remain relatively high, although not significantly different from 12 months. Fasting glucose levels did not change during aging (Figure 14).

Figure 14. Four-hour fasting plasma glucose and insulin levels. Repeated measures revealed no significant difference in glucose, but significant change (p<0.05) in insulin levels during aging. Paired sample t test was used to compare insulin levels at two different ages. Different letter denote significant difference (p<0.05).

Identification of plasma proteins

Overall, 86 protein spots representing the majority of the plasma proteins observed in this gel system, were identified by MS, MS/MS and LC/MS (Figure 15). Detailed MS

and MS/MS or LC-MS/MS scores are listed in Table 15 and Table 16; Mw and pI of 76 these spots were listed in Table 17 of appendix A). Among these identified proteins, 13 showed significant changes in plasma levels during aging (repeated measures, p<0.01).

Figure 15. 2D gel reference map of identified proteins. Proteins in the red boxes are identified and if more than one spot is indicated, isoforms are numbered 1, 2, etc. In 77 order to simplify the labeling of protein and protein isoforms, two duplicate gel images are presented as A and B. Abbreviations: Mw: molecular weight; pI: isoelectric point; apoA4: apolipoprotein A-4; a2m: α-2 macroglobin; apoE: apolipoprotein E; TTR: transthyretin; MBP-C: mannose binding protein-C; Ig kappa: immunoglobulin kappa chain; Prx-2: peroxiredoxin-2; RBP-4: retinol binding protein-4; SAA-1: serum amyloid A-1, Hbb: hemoglobin β-chain.

Plasma proteins that increased during aging

The levels of six plasma proteins were significantly increased during aging in the

longitudinal study (Figure 16 and Figure 15). The increase of these proteins during aging was confirmed in the cross-sectional study (data not shown). These proteins include three isoforms of immunoglobulin kappa chain (Ig kappa), of which spots 1 and 2 corresponded to Ig kappa variable III region (Ig kappa V-III) and spot 3 Ig kappa constant region (Ig kappa C). In addition, isoform 3 of transthyretin (TTR) and isoforms

2 and 3 of haptoglobin (Hp) also increased. Notice that the three Ig kappa chain isoforms did not change from 2 to 8 months (young to adult), but started to increase at 12 or 16 months (middle, old age). On the other hand, TTR isoform 3 and Hp isoforms 2 and 3 were increased as early as 4 or 8 months. 78

Figure 16. Plasma proteins that significantly increased in WT mice during aging (repeated measures, p<0.01). Protein intensities are log-transformed and presented in arbitrary unit (y axis). Protein isoform # corresponds to Figure 15.

Figure 17 shows the 3-D view of intensity of Ig kappa (spot 1) during aging. Notice that the spot was barely detectable from 2-8 months, but became apparent at 16 and 19 months.

Figure 17. 3-D view of the intensity of Ig kappa (spot 1) in WT mice at different ages. The spot intensity was converted into this topographic view of a peak using PDQuest. At the upper right corner is the original gel image of the spot, which is circled and indicated by an arrow.

Similarly, Ig kappa spots 2 and 3 became detectable only after 12 months (Figure

18 and Figure 19).

Figure 18. 3-D view of intensity of Ig kappa (spot 2) of WT mice at different ages. At the lower right corner is the original gel image of the spot, which is circled and indicated by an arrow.

Figure 19. 3-D view of intensity of Ig kappa (spot 3) of WT mice at different ages. At the upper right corner is the original gel image of the spot, which is circled and indicated by an arrow.

TTR isoform 3 was detectable as early as 2 months of age with a very low intensity.

It continued to increase to 19 months when the spot became visually obvious (Figure 20).

Figure 20. 3-D view of intensity of TTR isoform 3 of WT mice at different ages. At the upper right corner is the original gel image of the spot, which is circled and indicated by an arrow.

Hp isoforms 2 and 3 were non-detectable at 2 and 4 months, increased during aging and had a relatively high intensity at 16 and 19 months (Figure 21). 83

Figure 21. 3-D view of intensity of Hp isoforms 2 and 3 of WT mice at different ages. At the upper right corner is the original gel image of the two spots, which are circled and indicated by arrows.

Plasma proteins that decreased during aging

Seven proteins were significantly decreased during aging in the longitudinal study

(p<0.01, Figure 22). The decrease of these proteins was confirmed in cross-sectional study (data not shown). These proteins include peroxiredoxin-2 (Prx-2), five isoforms of albumin (isoforms 6-9 and 18), and serum amyloid protein-A1 (SAA-1). Notice that all these proteins had little change from 2 to 8 months (young to adult), but started to decline rapidly from 12 months (middle age) to 16 and 19 months of age.

Figure 22. Plasma proteins that significantly decreased in WT mice during aging (repeated measures, p<0.01). Protein intensities are log-transformed and presented in arbitrary unit (y axis). Protein isoform # corresponds to Figure 15.

Prx-2 started at relatively high levels but decreased after 8 months (Figure 23).

Figure 23. 3-D view of intensity of Prx-2 in WT mice at different ages. At the upper right corner is the original gel image of the spot, which is circled and indicated by an arrow.

Albumin (isoforms 6-9) all decreased from 8 months onward (Figure 24). 85

Figure 24. 3-D view of intensity of albumin isoforms 6-9 in WT mice at different ages. At the lower right corner is the original gel image of the four spots, which are circled and indicated by arrows.

Albumin isoform 18 decreased after 12 months (Figure 25).

Figure 25. 3-D view of intensity of albumin isoform 18 of WT mice at different ages. At the upper right corner is the original gel image of the spot, which is circled and indicated by an arrow.

SAA-1 decreased slightly but in a statistically significant manner (p<0.01) as mice aged. A topographic view of the spot intensity during aging is shown in Figure 26.

Figure 26. 3-D view of intensity of SAA-1 in WT mice at different ages. At the upper right corner is the original gel image of the spot, which is circled and indicated by an arrow.

Discussion

As mice aged, lean mass increased from young to middle age and did not change much from middle to old age; whereas fat mass increased at middle age, but started decreasing after 19 months (old age). As a result, body weights increased during aging with a slight weight loss after 19 months of age (Figure 13). A subset of these data (from

6 to 12 months) are in agreement with previous results in which mice were followed up to one year of age (Palmer et al., 2009). In humans, fat gain is associated with aging, but wasting at old age is also common. Unlike mice, wasting in humans happens with muscle as well fat loss (Fanciulli et al., 2009). This is likely due to species difference, but it is also possible that lean mass loss may be observed if mice are followed to ages older than

24 months. One note is that the sum of lean, fat and fluid mass is always smaller than body weight measured by a scale (comparing Figure 13 A and B). The reason for this discrepancy is that body compostion was measured by an NMR analyzer which utilizes the nuclear magnetic resonance of hydrogen, resulting in certain tissues not being measured (i.g., bone mineral).

Another characteristic of human aging is reduced insulin sensitivity, which was found in mice as well. Mice in this study had normal fasting glucose levels throughout life, but increased insulin levels at middle and old age (Figure 14), suggesting decreased insulin sensitivity at these later ages. Similar trend of insulin levels going up during aging in WT mice was also reported previously (Coschigano et al., 2003). The onset of increased insulin coincided with fat mass gain, suggesting a possible role of fat in insulin resistance that is commonly known (Guy-Grand & Bour, 1972; P. L. Huang, 2009; 89

Mikhail, 2009; Muhlhausler & Smith, 2009). However, insulin did not decrease at old age when fat loss occurs, suggesting other mechanisms are responsible for insulin- resistance at old ages.

In order to uncover other biomarkers of aging, we used 2-DE as a proteomic approach to resolve mouse plasma proteins in the natural course of mouse aging. One problem in isolating plasma proteins is that a few species of abundant proteins occupy most of the protein mass (Figure 4 in chapter I) (N. Leigh Anderson & Anderson, 2002).

Albumin, for example, comprises of 60% of all serum protein mass (Rothemund et al.,

2003). The high-abundance proteins result in high background smearing on 2D gels

(Thadikkaran et al., 2005). Relatively cheap, reproducible commercial kits are available for removal of human albumin from serum or plasma; however, no such kits are available for the treatment of mouse blood. To circumvent this problem, we used high concentration of polyacrylamide (15%) to resolve proteins with Mw below 45 kDa

(Figure 9B), which is smaller than albumin (~70 kDa).

We have quantified and identified 86 spots in mouse plasma during normal aging.

As can be seen from the gel profiles, these proteins represent the majority (in mass abundance) of all the proteins resolved (Figure 15). The protein distribution pattern on the 2D gel is comparable to a previous study that reported 57 identified mouse plasma proteins (Duan et al., 2004). We have discovered 13 spots corresponding to 6 distinct gene products that showed significant changes (repeated measures, p<0.01) as mice aged.

These are described below. 90

TTR

Mouse TTR is in theory 13.6 kDa (without the signal peptide) with a pI of 5.77.

Nine isoforms of TTR have been identified (Figure 15). The observed Mw of TTR

identified on 2D gel is ~14 kDa for isoforms 4-9, ~21 kDa for isoform 3, and ~40kDa for

isoforms 1 and 2 (Figure 15 and Table 17 in Appendix A). TTR is known to exist in

human plasma in different forms due to PTMs such as phosphorylation, cysteinylation,

dihydroxylation and glutathionylation (Hutchinson et al., 1996; Terazaki et al., 1998;

Zheng Wang et al., 2004). In SWISS-2DPAGE database, human TTR is resolved as

multiple spots on a 2D gel, including ~14 kDa and ~38 kDa forms, with the larger form

thought to be multimers of TTR (database, 2009). Similar TTR isoforms have also been

identified in mouse serum by (Duan et al., 2004). Also Okada et al from our group have

reported similar forms of TTR (manuscript submitted).

We have found that total TTR levels (sum intensity of all identified TTR isoforms;

this is not necessarily the total TTR in plasma as there may be TTR spots not identified

on the gels. However, this is a measurement of ‘total’ TTR in this study) fluctuated but increased significantly (p<0.01) as mice reached old age (Figure 27). In particular, isoform 3 increased throughout aging (Figure 20).

Figure 27. Total protein levels of TTR and RBP-4 in WT mice during aging. ‘Total protein’ is defined as the sum of all isoforms. Total TTR and RBP-4 changed in a similar pattern during aging.

The majority of TTR in plasma is secreted from the liver and forms a homotetramer that binds and transports thyroxine (T4) and retinol binding protein-4 (RBP-4). TTR binding to T4 is thought to increase T4’s half life, facilitating its controlled release and preventing the toxic effects of high-concentration free thyroid hormone (Schussler,

2000). TTR also binds to RBP-4 to increase its half life by reducing kidney filtration.

There are four RBPs with a primary role of binding and transporting retinol in both humans and mice. RBPs 1-3 are present intracellular whereas RBP-4 is found in the blood.

Interestingly, total RBP-4 levels at different ages were in parallel with those of total

TTR (Figure 27), both fluctuating with an increasing trend toward old ages. Since TTR binds to RBP-4 and prevents its kidney filtration, this suggests that the sum of all isoforms of TTR and RBP-4 reflects to a large extend the total protein levels in plasma.

RBP-4 is associated with obesity and insulin resistance (Haider et al., 2007; Kowalska et 92 al., 2008; Kohzo Takebayashi et al., 2007a). Plasma TTR and RBP-4 both increase in type II diabetic subjects (Raila et al., 2007). In this regard, increased total TTR may reflect increased insulin resistance as mice aged. Interestingly, in both male and female humans, TTR levels also increase as a function of age (Figure 28C,D) (Robert et al.,

1999a; Robert et al., 1999b). This trend of TTR increase resembles that of mouse aging in Figure 27.

Figure 28. Albumin and TTR levels during human aging. The closed circles (●) represent the observed median level. The lines represent the predicted median values. A and C, males; B and D, females. Adapted from (Robert et al., 1999b).

A condition called senile cardiac TTR amyloidosis has been found to be the primary cause of death in supercentenarians, people over 110 years old (Leslie, 2008). These 93 supercentenarians are usually free of cancer, cardiovascular diseases and neurodegenerative diseases. However, ‘clogging’ of blood vessels at autopsy has been found to be comprised of TTR amyloid in supercentenarians. Also, TTR can form plaques that deposit systemically (all over the body) (Ando, 2006). Finally, certain TTR mutants or variants have been found to be more susceptible to amyloid formation (Ando,

2006).

As mentioned earlier, TTR is modified by different PTMs. One specific PTM, a thiol conjugation, is found to be dependent on the age of a person and its increase is indicative of symptomatic amyloid disease (Suhr et al., 1999). Further, certain mutant

TTR is more likely to be modified by thiol conjugation than the wild type (Suhr et al.,

1999). Thus, PTMs of TTR may be related to amyloid formation and/or toxicity.

Therefore, it will be interesting to identify the PTM responsible for isoform 3.

Hp is composed of α and β subunits that are cleaved from a prepropeptide and

connected by disulfide bonds. Hp is ~31 kDa (without an 18-amino acid signal peptide).

However, in 2-DE, the two subunits are disassociated by reduction and alkylation. The

theoretical Mw of the α-subunit is ~9.5 kDa and the β-subunit is ~27 kDa. The spots

observed in Figure 15 matches the prediction of the α-subunit of Hp, not the β, by

MS/MS as well as LC-MS/MS (see Table 15 and Table 16 in Appendix A for details). It

is known that β-subunit is glycosylated, but no report has confirmed glycosylation for the

α-subunit. In the SWISS-2DPAGE database for human plasma, the Hp β-subunit resolves as multiple spots at ~45-48 kDa, which is beyond the limits of detection of our gel 94 system. In a previously resolved mouse serum 2D system, Hp α-subunit is also reported at similar Mw (15 kDa) and pI (5+), whereas Hp β-subunit resolved at 42 kDa with pI more acidic than 5 (Duan et al., 2004). Hp β-subunit has not been identified from our gels perhaps because of this. However, since Hp is synthesized as one peptide which is then cleaved into two chains, presumably the level of α-subunit is representative of the full- length protein. Among the four Hp isoforms identified in our gels, isoforms 2 and 3 increased during aging.

Hp is mainly secreted by liver, but also by adipose and lung tissues. In case of hemolysis in blood vessels (as opposed to normal degradation of red blood cells in the spleen), Hp binds to hemoglobin to protect against heme-induced oxidative stress. Hp is also an acute phase protein (proteins that are induced in the liver by inflammation) highly induced in the liver by inflammation and injury (Dobryszycka, 1997). In elderly people, a higher concentration of Hp is correlated with infection or inflammation (Katz et al.,

1990). It is found that in healthy aging population, serum concentration of Hp is negatively correlated with cognitive performance (Teunissen et al., 2003). In one proteomic study engaging 2-DE, Hp is found preferably in old aorta tissues but rarely in young, potentially indicating senescent vessels (Song et al., 1985).

In diabetes, acute phase proteins, including Hp, (Fleck, 1989) increase in the serum

(McMillan, 1989). Hp has been known to be a major susceptibility gene in diabetic vascular complications in humans (Asleh & Levy, 2005). Different alleles in humans are thought to have different antioxidant efficiencies, resulting in different susceptibilities to diabetic nephropathy, retinopathy, and cardiovascular disease (Bessa et al., 2007; 95

Nakhoul et al., 2007). As increased inflammation appears to be a common theme for both diabetes and aging (Dowling & Simmons, 2009), it is possible that Hp plays a role in both processes. Thus, increased Hp isoforms during normal aging observed in this study could indicate a state of increased inflammation at older ages. Hp was not detected by 2-

DE at 2 months, consistent with a previous study which does not detect Hp in normal male (6 weeks old) mouse plasma until inflammation has been induced by burn injury or infection (Duan et al., 2004).

Ig kappa

Ig kappa chain is an immunoglobulin light chain. Three spots have been identified

as Ig kappa. Two of them were matched to Ig kappa variable region III (spots 1 and 2)

and one to Ig kappa constant region (spot 3) (see Figure 15B). Spots 1 and 2 have sequences matched to multiple Ig kappa V-III, which is a variable region including multiple variants, for example, Ig kappa V-III region PC 7043 and region MOPC 63. The theoretical Mw of these Ig kappa V-III regions is about 12 kDa. Ig kappa C is 11.8 kDa with pI 5.23 in theory. All three spots are about 30 kDa on the gel, perhaps due to

multimerization of Ig kappa chains. All three increased during aging, it possibly indicates

that mice have undergone increased immune insults as they aged. Unfortunately, the

nature of the Ig kappa chain is not known, so it is difficult to say if this is merely a

reflection of increased concentration of immunoglobulins in old animals or if these are

Igs targeting a specific antigen, or if the process reflects an infection or autoimmune

disease. 96

Prx-2

Prx-2 has a theoretical Mw of 22 kDa and pI of 5.2. On the gel, it appears to be at

the location of Mw 25 kDa and pI 6.1. The known PTMs of this protein include

acetylation and phosphorylation (Swiss-Prot). We have found that Prx-2 decreased in

normal aging.

Prx-2 is an anti-oxidant enzyme involved in reducing peroxide levels in cells. It is

present in cytoplasm rather than being secreted. However, it is present in plasma at a

relatively large concentration and has also been identified by 2-DE in mouse serum

samples with similar Mw and pI (Guipaud et al., 2007); the same study shows that serum

Prx-2 levels decreases after mice have been exposed to skin irradiation. Thus Prx-2 seems

to be reduced by irradiation stress.

In C. elegans, Prx protects against oxidative stress and heat stress, as well as

promotes longevity as evidenced by shortened lifespan of the loss-of-function mutant

(OlÃhovÃ et al., 2008). A proteomic study using 2-DE identified Prx-2 as being induced

in primary astrocyte cultures from young rats (1 month) but not adult or aged rats (9 and

24 months) after exposure to X-ray irradiation (Miura et al., 2007). Prx-2 inhibits mouse

embryonic fibroblast senescence by inhibiting Ras-ERK-NFκB pathway (Han et al.,

2006). Together, the evidence suggests that Prx-2 protects against oxidative stress and

inhibits aging. Interestingly, its concentration decreased as mice aged in the current

study, perhaps indicating reduced capacity to battle reactive oxygen species (ROS) at

older ages. 97

Albumin

Full-length albumin has an Mw of ~ 70 kDa. As mentioned earlier, the high-

concentration SDS-PAGE used in our 2-DE resolved proteins smaller than 45 kDa;

therefore, full-length albumin would not be detected on these 2D gels. In fact, the albumin spots identified were all fragments. We have identified a total of 23 albumin fragments (Figure 15B). Isoforms 6-9 were ~24 kDa and matched to an N terminus fragment of albumin by MS/MS; isoform 18 was ~12 kDa and also matched to an N terminal fragment of albumin by MS/MS (Table 15 in Appendix A). In addition, other spots of different sizes were also identified as albumin, including isoforms 1-4 (32 kDa),

isoform 5 (28 kDa), isoform 10 (24 kDa), isoforms 11-17 (16 kDa), isoforms 19 and 20

(10 kDa) and isoforms 21-23 (8 kDa). Of these, isoforms 5, 14-17 and 20 were C-

terminal fragments; isoforms 1-4, 6-10, 11-12, 18, 19 and 21-23 were N-terminal

fragments (Table 15). Each fragment is represented in albumin in Figure 29.

Figure 29. Schematic representation of albumin fragments. Mouse full length albumin amino acid sequence is displayed. As shown, each color-coded box represents matched sequence by MS/MS of a particular albumin fragment. Numbers in the boxes represent the isoform number of albumin consistent with Figure 15.

Small Mw albumin fragments have also been found in mouse serum by 2-DE (Duan et al., 2004) and in human serum as reported in the SWISS-2DPAGE database (database,

2009). In NCBInr protein sequences database (NCBInr), there are several entries reporting mouse serum albumin fragments, including a 20.4 kDa N-terminal fragment

(the Mw is calculated from primary sequence, ACCESSION: CAC81903, GI:26986064) 99 and a 7.8 kDa fragment from the internal sequence (CAA31458.1 GI:899334). In addition, there are numerous entries in NCBInr of albumin fragments in a wide range of species (NCBInr). Thus, it appears that it is quite common for animals to have plasma albumin existing as multiple forms with different sizes.

We have found that five albumin fragments decrease as mice aged. It is not clear how fragmentation of albumin in plasma is regulated. One study found a proteolytic activity degrading albumin in urine and identified α−1 antitrypsin as an anti-proteolytic protein to inhibit albumin fragmentation (Magistroni et al., 2009).

In humans, a lower serum albumin concentration is found in the elderly (Figure

28A,B) (Gom et al., 2007; Nakamura & Miyao, 2007; Robert et al., 1999a; Robert et al.,

1999b). In rats, serum albumin concentration increases from 3 months to 7 months of

age, then decreases in 12 and 20 months (Barber et al., 1995). Interestingly, the five

mouse albumin fragments observed in this studiy showed little change from 2 to 8

months, but decreased from 8 months to 19 months (Figure 22, Figure 24 and Figure

25). However, the reduced levels of these albumin fragments were not necessarily due to

the reduction of total albumin concentration as a function of mouse age because some of

the other albumin fragments actually increased during aging (not reported because p

values were not below cut-off value of 0.01). In addition, only albumin fragments

smaller than 45 kDa were resolved on the gels. It is very possible that while total

albumin in blood decreased at old age, certain fragments increased whereas others

decreased or remained the same. Regardless, the five albumin fragments identified here 100

were consistent in both longitudinal and cross-sectional samples and significantly

(p<0.01) decreased with age.

The function of albumin mainly involves transportation of small molecules such as

water, Ca2+, Na+, K+, fatty acids, hormones, bilirubin and drugs and regulation of

colloidal osmotic pressure of blood. In terms of age-related diseases, higher serum

albumin levels are found to be associated with reduced cardiovascular mortality and

coronary heart disease and stroke incidence (Gillum, 2000). Albumin is also considered

an antioxidant protein that scavenges ROS (Halliwell & Gutteridge, 1990), a negative

acute phase protein that is decreased by inflammation (Fleck, 1989), and a common

nutritional marker with decreased levels indicating malnutrition.

SAA-1

SAA-1 is a 12 kDa protein secreted by liver. The observed Mw of this protein on the gel is 11.5 kDa. It showed a significant decrease as mice aged. SAA-1 is found to be associated with high-density lipoprotein (HDL). It is known that lipid profiles change during aging (Bittner, 2003; Karlamangla et al., 2004; Klag et al., 1993; Kronmal et al.,

1993). For example, non HDL cholesterol increases in the elderly (Bittner, 2003) and

HDL decreases during aging (Wilson et al., 1994). Total cholesterol is positively correlated with increased mortality at age 40, but this correlation becomes increasingly less obvious at ages 50, 60 and 70, and at age 80, total cholesterol is negatively correlated with mortality (Kronmal et al., 1993). HDL is considerd the ‘good’ cholesterol that is associated with decreased cardiovascular disease risks and better cognitive function

(Atzmon et al., 2002), and shows increased levels in centenarians (Nir et al., 2001). Low- 101 density lipoprotein (LDL), on the other hand, is considered ‘bad’ cholesterol and is associated with increased risks for cardiovascular disease (Bittner, 2003). Since SAA-1 is an apolipoprotein for HDL, the decrease of SAA-1 may therefore reflect or contribute to aging by an altering the lipid profile.

Other than lipid transportation, SAA-1 is also an acute phase protein induced by cytokines in response to infection or inflammation (Fleck, 1989). Interestingly, in humans, SAA-1 is also the precursor of reactive amyloid fibrils in type AA amyloidosis, a condition caused by deposition of insoluble fibrillar amyloid proteins (a degraded N-

terminus fragment of SAA) in the extracellular space in many organs and tissues (Merlini

& Bellotti, 2003). The chronic high expression of SAA in prolonged inflammation is

considered to be essential for amyloidogenesis (Yamada, 2006). Several alleles were

discovered that had different susceptibilities to amyloid formation, e.g., SAA1.1 was more susceptible than SAA1.5 (van der Hilst et al., 2008). To our knowledge, a role of

SAA-1 in aging has not been reported.

102

Conclusion

Longitudinal analysis of mouse plasma proteome by 2-DE revealed two important results. 1) Many proteins exist as multiple isoforms. 2) In normal male mice, certain plasma proteins undergo significant changes during aging. These proteins and/or their isoforms, therefore, are biomarkers of mouse aging (Table 6).

Table 6. Biomarkers of aging in WT mice Increased during aging ↑ Decreased during aging ↓

Hp isoforms 2 and 3 Prx-2 Two Ig kappa V-III isoforms (spots 1 and 2); Albumin isoforms 6-9 and 18 one Ig kappa C isoform (spot 3) TTR isoform 3 SAA-1 Spot # corresponds to those labeled in Figure 15.

The plasma proteins described in this report are involved in many aspects of aging or age-related diseases. Prx-2, Hp and albumin are antioxidant proteins that protect against oxidative stress, which is an important aspect of aging. SAA-1 and TTR have primary roles in transportation/metabolism of lipids or RBP-4 and T4, but are also involved in amyloid formation in blood vessels or tissues, a phenomenon associated with old age. In addition, albumin is involved in cardiovascular health and TTR in insulin resistance, both contributing heavily to morbidity and mortality at old age. Hp, SAA-1, albumin and TTR are acute phase proteins that indicate inflammation, which is likely associated with aging. 103

Of particular interest is that the different isoforms of the same protein did not always show the same profile relative to each other or with the total protein levels during aging (i.g., albumin and TTR). That is, some isoforms may increase, while others decrease. It is possible that different isoforms actually have different functions, contributing differentially to aging. Alternatively, isoform composition can affect the function of a given protein, and changes of the composition may be physiologically important and therefore reflective of aging. It is intriguing to study the PTMs responsible for the apparent isoforms. These plasma proteins can be modified in the secretory pathway inside the cell (e.g. hepatocyte) or after they are released into circulation. The regulation of generation of these PTMs is the underlying phenomena that future studies will address. Unraveling the PTMs and their regulation by aging may shed light on some of the molecular mechanisms of aging. Extension of these results to humans will be of obvious importance.

Are these proteins merely the markers of the aging process or do they actually play a causative role in aging? The functions of many of these proteins are not completely known; even less or nothing is known about the specific isoforms detected on the 2D gels. Future studies are needed to investigate the roles of these proteins/isoforms, the nature of their PTMs, and their functions during aging. With the elucidation of the mechanisms responsible for the phenotypic changes, we may be able to use these biomarkers as potential therapeutic targets that may ultimately combat aging, thereby increasing the lifespan and perhaps more importantly healthspan of an individual. 104

CHAPTER IV. BIOMARKERS OF CHRONIC GH ACTION: PROTEOMIC

ANALYSIS OF PLASMA FROM MALE BGH TRANSGENIC AND CONTROL

MICE

Abstract

Bovine growth hormone (bGH) transgenic mice are giant, lean, insulin-resistant and possess life spans shorter than wild type (WT) mice. To discover biomarkers indicative of

GH action and perhaps indicators of an accelerated aging phenotype, we analyzed mouse plasma proteins of different ages by two-dimensional electrophoresis (2-DE) and identified proteins of interest by mass spectrometry (MS) and MS/MS. We found many proteins whose plasma levels changed between bGH and control mice. Several of these proteins existed as multiple isoforms presumably due to post-translational modifications

(PTMs). Specifically, apolipoprotein (apo) E, haptoglobin (Hp) and mannose-binding protein-C increased significantly (p<0.01) in bGH mice versus controls; whereas transthyretin, apoA1 and retinol-binding protein-4 (RBP-4) significantly decreased. In addition, RBP-4 and alpha-2 macroglobulin exhibited isoform-specific changes in bGH mice, suggesting GH regulations of these proteins at PTM level. Importantly, two isoforms of Hp and clusterin exhibited a significant interaction between genotype and aging, increasing markedly during bGH aging with no change or less dramatic changes during WT aging. Thus, these two proteins are biomarkers reflecting an age-dependent difference in bGH mice.

105

Introduction

Growth hormone (GH) has been shown to be involved in the regulation of longitudinal growth, lactation and metabolism (Kopchick & Andry, 2000). GH acts on various tissues including bone, muscle, liver and fat. GH induces the expression of a potent growth factor called insulin-like growth factor-1 (IGF-1), which acts on target tissues through the IGF-1 receptor (R) via endocrine/paracrine or autocrine actions. The

GH/IGF-1 axis has been implicated in aging, as an attenuated GH/IGF-1 axis has been found to be beneficial for prolonged lifespan in a wide range of organisms (Andrzej

Bartke, 2005; Bonafè & Olivieri, 2009; Brown-Borg, 2009; Holzenberger et al., 2003;

Tatar et al., 2003).

The physiological effect of GH can be seen in transgenic mouse models that over- express GH. Bovine (b) GH transgenic mice have markedly elevated serum levels of

IGF-1, are gigantic, lean, insulin-resistant (Fernando P. Dominici et al., 1999a; Valera et al., 1993) and die prematurely due to liver, kidney and heart problems (A. Bartke, 2003;

Berryman et al., 2004; Knapp et al., 1994; Quaife et al., 1989). In addition, GH transgenic mice have decreased reproductive life span, increased astrogliosis (abnormal increase of astrocyte number due to death of nearby neurons), and premature changes in cognitive function, hypothalamic neurotransmitter turnover, and plasma corticosterone levels (A. Bartke, 2003).

Serum/plasma samples are routinely used for disease diagnosis. There are an estimated 10,000 proteins in human plasma and 1175 distinct proteins/peptides have been identified (N. Leigh Anderson et al., 2004). In mice, 12,300 unique peptides that 106 originate from 4567 unique proteins have been identified through a multidimensional fractionation and tandem mass spectrometry (MS) approach (Hood et al., 2005). Plasma, therefore, has a huge repertoire of proteins, some of which may be potential markers of

GH action.

In this study, plasma proteins of bGH mice were compared to wild type (WT) littermates during aging using a two-dimensional electrophoresis (2-DE) technique coupled with MS and MS/MS for protein identification. In this study, 29 novel plasma proteins/isoforms involved in the action of GH have been identified. These results may be beneficial for better understanding the pathphysiology of acromegaly, as well as in providing a background for the discovery of biomarkers of GH action in humans especially when recombinant human (rh) GH is misused or abused. RhGH doping is currently an important issue on the agenda of the World Anti-Doping Agency (WADA)

(WADA, 2009).

107

Materials and Methods

Experimental animals

The bGH transgenic mice were generated by injecting a linearized plasmid

containing a metallothionein transcriptional regulatory region (promoter) driven bGH

cDNA into the pronucleus of C57BL/6J embryos as described previously (Berryman et

al., 2004; McGrane et al., 1988; Wagner et al., 1981; Yang et al., 1993). WT littermates

were used as control animals. Male bGH mice (n=9) and WT (n=8) in the background of

C57 BL/6J were followed during aging at 2, 4, 8, 12, 16 and 19 months of age. These

ages represent post-puberty (2 months), young adult (4 months), middle age (8 and 12

months) and ‘old age’ (16 and 19 months) in mice (Richard A. Miller & Nadon, 2000).

Mice were housed 2-3 per cage at room temperature (22 oC) in a 12-hour light, 12-

hour dark cycle, fed ad lib with a standard chow diet. Animal protocols were approved by

Ohio University’s Institutional Animal Care and Use Committee.

Body composition measurement

Body composition was measured at 9, 13 and 19 months using mq 7.5 NMR analyzer, the Bruker Minispec (The Woodlands, TX) as described previously (List et al.,

2009; Palmer et al., 2009). The data are presented as absolute mass (gram, or g) as well as percentage of body weight (%) of lean mass, fat mass and fluid mass. Each animal at each age was measured twice and the mean value used.

Plasma IGF-1 measurement

Blood was collected with heparinized capillary tubes by tail tip clipping following

heat lamp exposure. Whole blood was centrifuged at 7000 x g for 10 minutes at 4 ºC and 108 the resulting plasma was stored at -80 ºC. Plasma IGF-1 levels were determined using the

DSL-10-29200 mouse/rat IGF-1 ELISA kit purchased from Diagnostic Systems

Laboratories, Inc. (Webster, TX).

Fasting glucose and insulin measurements

Mice were fasted for 4 hours and bled at 3 PM. Blood glucose was measured using a

ONE TOUCH glucometer from Lifescan (Milpitis, CA). Plasma was collected as

described above and insulin levels measured using an ultrasensitive rat/mouse insulin

ELISA kit (ALPCO, Windham, NH).

2-DE and quantification of proteins

See chapter III.

Statistical analysis

All statistical analyses were performed using SPSS 14.0 software (Chicago, IL).

Data for IGF-1 and body composition were subjected to t test with significance value

p<0.05. Data for body weight, fasting glucose and insulin were subjected to repeated

measures with significance value p<0.05. For protein intensity, the log-transformed data

were subjected to repeated measures for five different age points (2, 4, 8, 12 and 16

months), with genotype as a fixed factor (p<0.01). The data at 19 months was excluded

from repeated measures because 5 bGH mice died at 19 months. For genotypic difference

at each age, an independent t test was used (including at 19 months). Data of bGH mice

(n=4) at 19 months was also graphed to show the trend of any change. All data were

presented as mean ± SEM. 109

Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography

(LC)-MS/MS

See chapter III.

110

Results

Body weight and body composition

Whereas WT mice showed steady body weight gain from weaning (3/4 months) to

19 months, bGH mice reached a weight plateau at 9 months and started losing weight

thereafter (Figure 30A). However, throughout their lifespans, with the exception of ¾

months and 19 months, bGH mice had significantly greater weights than WT at each age

point (p<0.00001).

Body composition was examined at 9, 13 and 19 months. WT and bGH mice

showed marked differences in lean, fat and fluid mass at each age (Figure 30B, C, D;

Figure 31A). At each age, bGH mice had significantly higher lean mass than WT

(p<0.0001) and significantly lower fat mass than WT animals (p<0.01). At 9 and 13

months, bGH mice had significantly increased fluid mass than WT (p<0.001 and p<0.01,

respectively), but no significant difference at 19 months. Noteably, whereas WT mice

had significant age-related fat mass gain (paired sample t test, p<0.05), bGH mice

showed significant fat mass loss during aging (paired sample t test, p<0.01). Also, WT

mice showed increased fluid mass during aging (paired sample t test, p<0.05) whereas

bGH mice showed no significant change.

As shown above, in terms of absolute mass, bGH mice had an increase in lean and

fluid mass (except at 19 months) and reduced fat mass compared to WT. When data in

grams were converted to percentage of body weight, bGH mice still had increased lean%

and decreased fat % compared to WT mice but fluid % became not significantly

different between genotypes (Figure 31B). 111

Figure 30. Body weight and composition of bGH and WT mice during aging. A, body weight change during aging. One-way, repeated measures revealed a significant change of body weight during aging (p<10-38), a significant genotype difference (p<10-10) and a significant interaction between aging and genotype (p<10-9). B, lean mass. C, fat mass. Paired sample t test revealed that WT mice had fat mass gain from 9 to 13 months (p<0.05) whereas bGH had decreased fat mass (p<0.01). D, fluid mass. Paired sample t test revealed that WT mice had fluid mass gain from 9 to 13 months (p<0.05) whereas bGH had no change. * p<0.01, ** p<0.001, *** p<0.0001 comparing of bGH to WT of the same age by t test.

112

Figure 31. Body composition of bGH and WT mice during aging. A, stacking bar graph of body composition in bGH and WT mice. B, normalized body composition as percentage of body weight. * p<0.01, ** p<0.001, comparing of bGH to WT of the same age.

IGF-1, fasting plasma insulin and glucose levels

At 6 months, bGH had significantly higher IGF-1 levels than WT mice

(p<0.0001, Figure 32A). According to one-way repeated measures, fasting insulin levels were changed by aging significantly (p<0.01), but not by genotype. There was a significant interaction between aging and genotype (p<0.01, Figure 32B). bGH mice had an initial increased levels of insulin (from 3 to 5 months) and then they decreased at older ages. At 3 months, bGH mice had higher insulin than WT (p<0.05), however, from 12 months onward, bGH mice had significantly lower insulin levels than WT (p<0.05 for 12 months, p<0.01 for 16 months, and p<0.05 for 19 months). For fasting glucose levels, a 113 significant aging effect (p<0.01) and a genotype effect (p<0.001) as well as an interaction between aging and genotype (p<0.001) were observed (Figure 32C). The bGH mice had comparable levels of glucose with WT mice at young ages; however, from 12 to 19 months, bGH mice had significantly lower glucose levels than WT mice (p<0.01, Figure

32C).

Figure 32. IGF-1, insulin and glucose levels. A, IGF-1 levels at 6 months of age. B, 4- hour fasting insulin levels during aging. One-way repeated measures showed a significant aging effect (p<0.01) and a significant interaction between aging and genotype (p<0.01), with no genotype difference. C. 4-hour fasting glucose levels during aging. One-way repeated measures showed a significant aging effect (p<0.01) and a genotype effect (p<0.001) as well as an interaction between aging and genotype (p<0.001). #p<0.05, *p<0.01, **p<0.001, ***p<0.0001 comparing of bGH to WT of the same age by t test. 114

Detection of plasma proteins in bGH mice versus WT controls by 2-DE

Plasma proteins of WT and bGH mice at six different ages were analyzed (see

Figure 85, Figure 86 and Figure 87 in Appendix B for original gel images). A number of

proteins changed significantly in bGH mice using one-way repeated measures. Thirteen

proteins increased while fourteen decreased significantly in bGH mice compared to WT

controls, and 2 proteins (haptoglobin (Hp) and clusterin) showed a significant interaction

between genotype and aging (p<0.01, Figure 33).

Proteins upregulated in bGH versus WT are numbered in red (# 1-13) in Figure 33,

including all five isoforms of apolipoprotein E (apoE, # 2, 4-7), two isoforms of α-2

macroglobulin (a2m) (# 1 and 3), mannose binding protein-C (MBP-C, # 8), one isoform

of retinol-binding protein-4 (RBP-4, # 9), and all four isoforms of Hp (# 10-13). Proteins

downregulated in bGH mice are numbered in purple (# 14-28) in Figure 33, including

eight isoforms of transthyretin (TTR, # 14, 20, 22-27), two isoforms of a2m (# 15 and

16), three isoforms of apolipoprotein A1 (apoA1, # 17-19) and one isoform of RBP-4 (#

21). Two isoforms of clusterin (spots 28 and 29) and all four isoforms of Hp (spots 10-

13) showed significantly different aging patterns between bGH and WT mice, which are

labeled with circles in Figure 33. The quantification of fold change in these proteins is

listed in Table 7.

115

Figure 33. Plasma proteins up or downregulated in bGH compared to WT mice (p<0.01). Proteins are marked by green crosses for their locations on the gel and numbered 1-13 in red (upregulated proteins), and 14-27 in purple (downregulated). Spots in circles (spots 10-13, 28 and 29) are proteins that had a significant interaction between genotype and aging (p<0.01).

116

Table 7. Quantification of proteins up or downregulated in bGH mice Spot # Protein ID Isoform # Ratio of bGH/WT 1 A2m 1 ↑3.8 2 apoE 1 ↑3.0 3 A2m 2 ↑1.9 4 apoE 2 ↑6.2 5 apoE 3 ↑7.4 6 apoE 4 ↑6.7 7 apoE 5 ↑3.9 8 MBP-C 1 ↑8.2 9 RBP-4 4 ↑3.2 10 Hp 1 ↑10.4 11 Hp 2 ↑11.0 12 Hp 3 ↑6.5 13 Hp 4 ↑22.6 Total apoE Sum ↑6.1 Total Hp Sum ↑8.3 14 TTR 1 ↓0.26 15 A2m 5 ↓0.74 16 A2m 6 ↓0.43 17 apoA1 4 ↓0.33 18 apoA1 5 ↓0.54 19 apoA1 6 ↓0.35 20 TTR 3 ↓0.09 21 RBP-4 2 ↓0.59 22 TTR 4 ↓0.43 23 TTR 5 ↓0.38 24 TTR 6 ↓0.45 25 TTR 7 ↓0.36 26 TTR 8 ↓0.29 27 TTR 9 ↓0.19 Total TTR Sum ↓0.40 Total RBP-4 Sum ↓0.75 Spot # corresponds to Figure 33; isoform # corresponds to Figure 15. The ratio bGH/WT is calculated as the mean value of bGH/mean value of WT (pooled values of all ages). ‘↑’ means up-regulation and ‘↓’ means down-regulation in bGH versus WT mice. Total protein is defined as the sum of all identified isoforms of a given protein.

All five isoforms of apoE increased dramatically in bGH mice at all ages compared to age-matched WT controls (Figure 34 and Figure 35). These isoforms did not show significant changes during aging at a significance level p<0.01. However, total apoE, the sum of all isoforms, decreased significantly during aging in bGH (p<0.001) and WT mice 117

(p<0.05), with total apoE levels significantly higher (more than 6 fold) in bGH mice

(p<0.001).

Figure 34. Comparison of apoE isoforms in WT and bGH mice of different ages. A, cropped image of one representative gel from WT and bGH at each age. The red arrows point to individual apoE isoforms of mice at 4 months. B, 3-D view of apoE isoforms in one WT and one bGH mouse at 8 months of age. Spot intensity was converted into topographic view of a peak using PDQuest. The spot # corresponds to Figure 33. 118

Figure 35. ApoE increased in bGH compared to WT mice at all ages. Individual isoforms as well as total apoE are quantified at different ages in bGH and WT mice. Spot # corresponds to Figure 33. Total apoE is calculated as the sum intensity of all apoE isoforms. The Y axis represents log-transformed intensity.

Spot 8, an MBP-C isoform, increased in bGH mice significantly (p<0.01, Figure

36). It also showed a significant change relative to age in both genotypes (p<0.01), increasing from young to middle age and decreasing in old mice. 119

Figure 36. One isoform of MBP-C increased in bGH compared to WT mice. The spot number corresponds to that in Figure 33. A, MBP-C levels (log-transformed intensity) during aging in bGH and WT mice. B, 3-D view of the intensity of MBP-C from one WT and one bGH mouse at 12 months. At the upper right corner is the original gel image of the spot, which is circled and is indicated by an arrow.

All four Hp isoforms increased in bGH mice (Figure 37, Figure 38 and Figure

39A). All isoforms showed a significant increase in bGH mice during aging (p<0.01), but only spots 11 and 12 increased as a function of age in WT mice (p<0.01). All Hp isoforms increased during aging to a greater extent in bGH mice, thus showing a significant interaction between genotype and aging (p<0.01). Hp was not detectable in

WT or bGH mice at 2 and 4 months, but became detectable after 8 months, with increasing levels found during aging. This occurred more so in bGH than WT mice. Total

Hp, the sum of all isoform intensities, increased during aging (p<0.01 for WT and p<0.001 for bGH), with similar levels at 2 and 4 months but higher levels in bGH mice at later ages (p<0.01 for interaction) and overall higher levels in bGH mice (p<0.01).

120

Figure 37. Hp in bGH and WT mice. Individual isoform as well as total Hp are quantified at different ages in bGH and WT mice. Spot # corresponds to Figure 33. Total Hp is calculated as the sum intensity of all Hp isoforms. The Y axis represents log-transformed intensity.

121

Figure 38. 3-D view of four isoforms of Hp (spots 10-13) and one isoform of RBP-4 (spot 9). The spot # corresponds to Figure 33. 122

Eight TTR isoforms decreased in bGH mice at all ages compared to age-matched controls (Figure 39, Figure 40 and Figure 43B). Spots 22 and 25 (Figure 40) also showed a significant increase during aging in both genotypes (p<0.01). The other isoforms did not change significantly with age. Total TTR, i.e., the sum of all isoform intensities, increased with age in both WT and bGH mice (p<0.01 for WT and p<0.001 for bGH), with a greater increase of the levels in bGH mice, although they were always lower in bGH than WT mice (p<0.001).

Figure 39. Comparison of TTR and Hp isoforms in WT and bGH mice at different ages. A, cropped image of one representative gel from WT and bGH at each age. Arrows indicate individual isoforms of TTR and Hp as labeled. The spot # corresponds to Figure 33. For each gel image, spots to the left of the dashed vertical line are Hp and those to the 123 right are TTR. B, 3-D view of TTR isoforms (spots 22-26) in one WT and one bGH mouse at 8 months of age.

Figure 40. TTR decreased in bGH compared to WT mice at all ages. Eight individual isoform as well as total TTR are quantified at different ages in bGH and WT mice. Spot # corresponds to Figure 33. Total TTR is calculated as the sum intensity of all TTR isoforms. The Y axis represents log-transformed intensity. 124

Three isoforms of apoA1 decreased in bGH compared to WT mice; no significant changes were seen during aging (Figure 41). The other three isoforms of apoA1 (shown in Figure 15 in chapter III) did not show a difference between genotypes or during aging.

Total apoA1, the sum of all isoform intensities, also decreased (p=0.011). Thus, apoA1 was not changed by age and showed only a genotype difference.

Figure 41. ApoA1 decreased in bGH compared to WT mice. A, three isoforms of apoA1 decreased significantly (p<0.01) and total apoA1 also decreased (p=0.011). Total apoA1 is calculated as the sum intensity of all identified apoA1 (see Figure 15 for all apoA1 isoforms). The Y axis represents log-transformed intensity. B, 3-D view of apoA1 125 isoforms in one WT and one bGH mouse at 8 months of age. The spot # corresponds to Figure 33.

One isoform of RBP-4 (spot 9) decreased, whereas another one (spot 21) increased in bGH compared to WT mice (Figure 42 and Figure 38). However, total RBP-4 decreased in bGH mice versus WT mice. Both isoforms also showed significant fluctuation (decreased at middle age but increased again at old age) during aging in both genotypes (p<0.01). Total RBP-4, sum of all RBP-4 isoforms, fluctuated similarly in both genotypes during aging (p<0.001 for both), with the levels in bGH always lower than WT

(p<0.001).

Figure 42. RBP-4 showed isoform-specific change in bGH compared to WT mice. A, quantification of spot intensities. Total RBP-4 is calculated as the sum intensity of all identified RBP-4 (see Figure 15 for all RBP-4 isoforms). The Y axis represents log- 126 transformed intensity. B, 3-D view of intensity of RBP-4 from one WT and one bGH mouse at 12 months. Arrows indicate the 3-D peak as well as the original gel image corresponding to the spot. The spot # corresponds to Figure 33.

A2m was identified as seven isoforms with similar Mw but pIs ranging from 5.6 to

7 (Figure 15 in chapter III). Of these, spots 1 and 3 increased, whereas spots 15 and 16 decreased in bGH mice (Figure 43 and Figure 44). None of the isoforms showed significant changes during aging. Total a2m did not change during aging or between genotypes.

Figure 43. Comparison of a2m isoforms in WT and bGH mice at different ages. A, cropped image of one representative gel from WT and bGH at each age. Arrows indicate individual isoforms. B, 3-D view of intensity of a2m isoforms from one WT and one bGH mouse at 8 months. The spot # corresponds to Figure 33. 127

Figure 44. A2m showed isform-specific changes in bGH compared to WT mice. Spot # corresponds to Figure 33. The Y axis represents log-transformed intensity.

Two isoforms of clusterin increased markedly during aging in bGH but not WT mice (Figure 45 and Figure 46). Similar trend was found in other four isoforms but did not reach significance level of 0.01, therefore not shown. Total clusterin, the sum of all isoform intensities, did not change in WT, but increased significantly in bGH mice

(p<0.001), with lower levels than WT at young ages but higher levels than WT at later ages. Both spots 28 and 29 as well as total clusterin showed marked increase in bGH mice during aging, with no significant change during WT aging, resulting in lower levels of clusterin when young, but higher levels in older bGH versus WT mice.

128

Figure 45. Clusterin showed significant interaction between genotype and aging. A, cropped image of one representative gel from WT and bGH at each age. Of the six isoforms shown in circles, spots 28 and 29 showed significant interaction between genotype and aging (p<0.01). Arrows indicate these two spots in all gels. Other 4 isoforms showed similar changes without reaching a significant threshold. Spot # corresponds to Figure 33. B, quantification of clusterin in bGH and WT mice during aging. The Y axis represents log-transformed intensity.

129

Figure 46. 3-D view of clusterin isoforms.

130

Discussion

Body composition, IGF-1 and insulin sensitivity

BGH mice had increased IGF-1, body weight, lean and fluid mass, and reduced fat

mass, all consistent with features of acromegaly in humans (Chanson & Salenave, 2008).

These data suggest that bGH mice are a good model to study acromegaly.

BGH mice had similar body weight as WT at weaning (3/4 month), suggesting that the growth-promoting effect of GH is not manifested before 3 weeks of age. Close examination of the mice at different ages revealed that from 3/4 to 19 months, bGH mice reached a weight plateau at 8 months and started losing weight after 9 months, whereas

WT mice did not show weight loss till 19 months. Weight loss, therefore, could be a rough estimation of getting ‘old’. By this notion, WT mice started getting ‘old’ at 19 months at which age they started losing weight (chapter III). The early-onset weight loss of bGH mice reflects their shortened lifespan. In fact, 5 out of 9 bGH mice died at shortly before the age of 18 months. This early death rate in GH transgenic mice has been reported previously (Berryman et al., 2004; McGrane et al., 1988; Steger et al., 1993).

As expected, bGH mice had significantly higher weight than WT mice of the same age throughout their lifespan, manifesting the potent effect of GH in growth promotion and anabolism. WT had steady levels of lean mass from 9 months (middle age) to 19 months (old), whereas bGH mice showed no change of weight and even a slight decrease at 19 months. At each age, bGH had significantly higher lean mass than WT, showing

GH’s anabolic action on lean mass. WT mice also had increased fat mass during aging

(up to 19 months); in contrast, bGH had decreased fat mass as a function of age. In 131 chapter III of this dissertation, WT mice started losing fat and weight at 19 months; it seemed that bGH mice started to decline from 9 months, perhaps reflecting the accelerated aging in this mouse model. The fluid mass increased in WT throughout 19 months, but did not change in bGH mice. However, at each age, bGH mice had significantly higher fluid mass, perhaps due to the fluid retention effect of GH (Carroll &

Van den Berghe, 2001; Clemmons et al., 1987; O'Sullivan et al., 1994). GH and IGF-1 promote fluid retention by stimulating renin-angiotensin-aldosterone system, which is central to fluid homeostasis (J. Moller et al., 1999). One cautionary note is that when normalized to body weight, fluid% remained the same in bGH and WT mice. The larger error bar at 19 months was probably due to the reduced sample size (n=4). Curiously, there may be a problem with urination in the very old bGH mice, since at dissection a huge, full bladder was noted (unpublished observation). It is interesting that in WT mice, fluid mass continued to increase from 19 to 24 months (chapter III), but remained unchanged in bGH mice from 9 to 13 months, possibly suggesting that the normal age- related fluid retention in mice is affected by high levels of GH. Together, body composition data suggest that for WT mice, 19 months is an ‘indicator’ for old age; however this ‘indicator’ is much earlier in bGH mice where 9 months can be considered old. Indeed, the mean lifespan is 409 ± 45 days (about 14 months) for bGH mice versus

602 ± 34 days (about 21 months) for WT in this mouse line (mentioned in (Berryman et al., 2004) as unpublished observations, and similar in other hGH or bGH lines (McGrane et al., 1988; Steger et al., 1993). The body composition data before 13 months in this 132 study was consistent with a previous study, in which bGH and WT mice were followed to one year of age (Palmer et al., 2009).

As previously shown (chapter III), fasting insulin levels increased in WT mice during aging; however, insulin levels increased only from 3 to 5 months in bGH mice.

Although this study only looked at body composition after 9 months, a previous study reported that from 1.5 to 5 months of age, bGH mice had increased fat mass as a function of age similar to WT mice and strikingly more absolute fat mass than age-matched WT animals (Palmer et al., 2009). Therefore, the increase in insulin with age may reflect fat increase in bGH mice during this time period. From 1.5 to 3 months of age, bGH mice also have a greater percentage of fat mass than WT, but the normalized fat mass become less than WT after 3 months (Palmer et al., 2009). This may be a part of the reason why bGH mice had higher insulin levels than WT at 3 months but was not higher at 5 and 9 months when they started showing less normalized fat than WT mice. This implies a role of fat in insulin levels as has been shown numerous times (Guy-Grand & Bour, 1972; P.

L. Huang, 2009; Mikhail, 2009; Muhlhausler & Smith, 2009).

BGH mice are known to be hyperinsulinemic, however, most of the literature reported this finding in bGH mice of younger ages (Table 1 in background chapter) and few studies looked beyond 8 months. From this study, it appeared that hyperinsulinemia occurs only in young mice (perhaps due to a greater fat mass of bGH than WT when young), normalizes at middle ages (bGH started having less fat than WT mice at 6 months), and develops into hypoinsulinemia at older ages (much less fat). The reason for this phenomenon is not known. In addition to the fat mass theory above, another 133 possibility is that beta cells in young bGH mice are stimulated by high levels of GH and produce high amounts of insulin, as has been shown previously (A. Bartke et al., 2004;

Fernando P. Dominici et al., 1999a; F. P. Dominici et al., 1999b; Olsson et al., 2003;

Olsson et al., 2005; Zhihui Wang et al., 2007); however, chronic exposure to high GH leads to accelerated aging and this ‘aging’ phenotype may be manifested in multiple organs, including pancreas, which may fail to secret insulin as bGH mice get old. In this regard, GH/IGF-1 is known to stimulate beta cell proliferation and insulin secretion

(Nielsen et al., 2001; Okuda et al., 2001; Parsons et al., 1995; Rhodes, 2000). However, no data to date have demonstrated an aging effect on pancreatic beta cells in bGH mice.

Although it was reported that bGH mice as old as 8 months still had higher insulin levels than controls (Frick et al., 2001a), there was also a report that showed non-significantly different insulin levels (only showing an increased trend) compared to controls at 5 months of age (Berryman et al., 2006).

Normal glycemia was maintained in WT mice during aging, but decreased in bGH mice such that they had normal glucose up to 9 months but significantly lower levels in

12, 16 and 19 months. Similarly, lower than control glucose levels, although sometimes not significant, were observed in 3-month old hGH mice (A. Bartke et al., 2004), 11- month old (Rocha et al., 2007) and 8-month old bGH mice (Frick et al., 2001b). It seems that hypoglycemia develops in bGH mice at older ages. It is interesting that after 9 months, both fasting insulin and glucose dropped as a function of age and the levels were lower than those of WT mice. One possible explanation is that the prematurely ‘old’ bGH liver failed to secret normal amounts of glucose, since fasting blood glucose is mainly 134 derived from hepatic output (Nordlie et al., 1999; Sheehan, 2004). In support of this, a transcriptional coactivator regulating hepatic glucose production, PGC-1, which normally is induced by fasting in mouse liver to upregulate gluconeogenesis-related enzymes

(Yoon et al., 2001), was lower in bGH than WT mice (Al-Regaiey et al., 2005).

Conversely, PGC-1 was elevated in GHR-/- versus WT mice (Rocha et al., 2007). In addition, a key enzyme in gluconeogenesis, PEPCK, was lower in bGH mice in response to fasting (Valera et al., 1993). The levels of insulin, glucose, and body weight decline after 9 months of age, which again suggests that in bGH mice, 9 months may be a

‘definition’ of ‘old’. Additional support for this comes from data regarding cardiovascular health. In 8 or 9 month old bGH mice, cardiac function significantly decreases versus age-matched controls in terms of energy reserve, cardiocyte apoptosis and fatty acid metabolism (Bogazzi et al., 2009; Bogazzi et al., 2008; Bollano et al.,

2000).

As mentioned in the introduction of this chapter and the background chapter I of this dissertation, bGH mice age prematurely and develop multiple organ pathology. The data on body composition, glucose and insulin data confirmed that bGH mice possess an accelerated aging process. Thus, it is possible that these age-related differences are reflected in changes in plasma proteins that could ultimately provide biomarkers of aging.

Plasma proteins that were significantly different in bGH mice versus controls

ApoE is increased in bGH mice

Five apoE isoforms have been identified (Figure 15 in chapter III). ApoE has a

theoretical Mw of 34 kDa and the observed Mw on the gel for four of the isoforms (spots 135

4-7 in Figure 33) was ~34 kDa. An additional ApoE isoform (spot 2 in Figure 33) was

~38 kDa. The PTM for the size difference of this isoform is not clear. This additional isoform of 38kDa has not been reported previously. Although not reported in mice, in human plasma, apoE is glycosylated and glycated, which could in theory result in protein spots of slightly different Mw and also different pIs.

All five apoE isoforms increased significantly in bGH versus WT mice, with total apoE increasing 6 fold, indicating a positive effect of GH action. Consistent with this, hypophysectomized (Hx, removal of pituitary gland) female rats had decreased levels of

apoE, but upon GH treatment, serum apoE levels nearly normalized (Frick et al., 2002;

Jan Oscarsson et al., 1991b; Sjoberg et al., 1994a; Sjoberg et al., 1994b). This apoE increasing effect was not mediated by IGF-1 (Sjoberg et al., 1994b). GH has been shown to increase apoE by increasing the translation and secretion rate of apoE from hepatocytes (Sjoberg et al., 1994a). Continuous, but not pulsatile, treatment of GH increased apoE in Hx rats (J. Oscarsson et al., 1991a; Jan Oscarsson et al., 1991b). This suggested a sex-dimorphic effect of GH in regulating apoE expression, since males have a pulsatile secretion of GH during a 24 hour period whereas females have a more constant secretion pattern in mice, rats and humans (Leppaluoto et al., 1975; Norstedt &

Palmiter, 1984; Saunders et al., 1976; Gloria Shaffer Tannenbaum & Martin, 1976).

These reports are in agreement with the data presented in this study, since bGH mice have relatively high, constant levels of serum GH and should therefore upregulate production of ApoE. The novel data in this report are that different plasma isoforms of

ApoE are all up regulated. 136

Several human studies have also confirmed GH’s effect on increasing apoE.

Acromegalic patients had increased apoE levels in blood (Wildbrett et al., 1997). GH treatment in adult GH deficient patients for one year increased serum apoE levels

(Johannsson et al., 1995). GH treatment of obese men showed an increase in apoE at 2 weeks although not 8 weeks (Svensson et al., 1999). However, in one study, GH treatment of 30 abdominally obese men for 9 months showed no change in serum apoE levels (Svensson et al., 2000). One possible reason is the dose difference of GH: bGH mice and acromegalic humans have much higher GH concentrations than that used for treatment of human subjects. Another likely reason is that the human subjects in these studies are obese, versus the lean mice used in this study.

ApoE is secreted primarily by the liver and functions to transport lipids via chylomicron, very-low-density lipoprotein (VLDL), LDL and high-density lipoproteins

(HDL). VLDL is made by liver and transports lipids to other tissues to be used.

Relatively high levels of VLDL are linked to increased risk of atherosclerosis (Veniant et al., 2008). LDL particles are formed from VLDL particles by the loss of triglyceride and gain of cholesterol ester. VLDL and LDL are considered ‘bad’ cholesterol because of a positive association with increased cardiovascular risks. HDL transports lipids from peripheral tissues back to the liver and is considered the ‘good’ cholesterol that is protective to cardiovascular health. ApoE-/- mice, deficient in apoE that transports and mediates the uptake of VLDL and LDL, have hypercholesterolemia and develop spontaneous atherosclerosis (Nakashima et al., 1994). Interestingly, GH exacerbated the atherosclerotic lesions in these mice even further, whether fed with chow or high fat diet 137

(Irene J. Andersson et al., 2006b). bGH mice have increased plasma total cholesterol, including both HDL and LDL cholesterol (Frick et al., 2001a; Olsson et al., 2005).

Perhaps increased apoE is made to transport the increased cholesterol.

It is interesting that apoE levels decreased in mice during aging. If GH stimulates apoE, then this is not surprising, since serum GH levels decrease with age. ApoE also is considered an antioxidant that helps to fight oxidative stress (Davignon, 2005; Miyata &

Smith, 1996; Tarnus et al., 2009). Liver DNA damage increases with age in apoE-/- mice, but not in control mice (Folkmann et al., 2007). ApoE has also been shown to be coorelated with a downregulation of inflammation in the central nervous system in response to immune insult (Lynch et al., 2001). Since increased oxidative stress and reduced ability to fight it is one marker of aging (Dowling & Simmons, 2009), it appears that less apoE at old age reflects the reduced anti-oxidative capacity in the elderly.

Contrary to the result in mice, apoE in human serum appears to increase with age

(Schiele et al., 2000). This may be due to species difference.

Hp is increased in bGH mice and shows a significant interaction between genotype and aging

As discussed in chapter III, four Hp isoforms have been identified (see chapter III for Mw, pI and possible PTMs for Hp). All four Hp isoforms increased, with total Hp increasing 8 fold in bGH versus WT mice. Hp is an acute phase protein primarily secreted by liver. Acute phase proteins are induced in the liver by inflammation, infection or other immune insults. Hp has been found to increase in response to GH treatment in

Turner’s syndrome (TS) patients (Gravholt et al., 2004). GH also increased Hp 138 production in a human hepatoma cell line (Derfalvi et al., 2000). In the mouse, Hp increased following whole-body exposure to gamma-rays (Rithidech et al., 2009). In pigs and lambs treated with acute insults (endotoxin and yeast), serum Hp as well as GH levels increased (L. G. Moore et al., 1995; Spurlock et al., 1998).

Since Hp is an indicator of inflammation, and that Hp increased during aging in both bGH and WT mice, it may indicate that mice had increased inflammatory state as they age. bGH mice had higher Hp levels than WT, and interestingly, two of the Hp isoforms showed an enhanced increase during aging in bGH relative to WT mice. This suggests that bGH mice may have an increased inflammation process during aging compared to WT mice.

Clusterin showed significant interaction between genotype and aging

We have identified 6 clusterin isoforms (Figure 15 in chapter III). The clusterin precursor is cleaved into α- and β-subunits, which are held together by five disulfide bonds. The heterodimer is extensively glycosylated with sulfated N-linked carbohydrates and secreted (Pajak & Orzechowski, 2006). Incidentally, the 6 isoforms observed on our

2D gels were clusterin β-chain, as indicated by MS/MS matching (Table 15 in appendix

A). Clusterin β-chain has a theoretical Mw of 23.9 kDa and pI 6.12. The observed values on the gel were 34-36 kDa and pI 5.28, 5.37, 5.47, 5.59, 5.69 and 5.77. These are likely the result of PTMs such as glycosylation and phosphorylation. Clusterin is known to be glycosylated with sulphated carbohydrates, which could explain the increased Mw and a shift of pI to become more acidic. Phosphorylation may also contribute to this pI shift.

According to an on-line tool to calculate Mw and pI with possible phosphorylation 139 modification at Swiss-Prot, ‘ScanSite pI/Mw’ (http://scansite.mit.edu/calc_mw_pi.html),

clusterin β-chain would have pIs of 5.73, 5.57, 5.43 and 5.3 if it were to be

phosphorylated on 1, 2, 3, or 4 residues (this tool has been applied to all proteins that

have been identified in this dissertation; only certain clusterin and a2m isoforms (shown

later) can be explained by phosphorylation). Interestingly, several of these numbers are

very similar to the observed pIs. For example, with 1 phospho (P) group, the theoretical pI would be 5.73 and spot 20 (Figure 15 in chapter III) has a pI of~ 5.77; with 2 P groups, theoretical pI 5.57 and spot 18 has a pI of ~ 5.59; with 3 P groups, the theoretical pI 5.43 and spot 17 is at ~ 5.47; and with 4 P groups, the theoretical pI would be 5.3 and spot 15 was at ~5.28. Therefore, it is very likely that the difference in pI of clusterin is due to differential levels of phosphorylation. It is curious that the clusterin α-chain, with a theoretical Mw 25 kDa and pI 5.09, was not observed on our gels. One possibility is that

α-chain is sulphated and if also phosphorylated, then the pI will shift to a pI less than 5, which is beyond the pI range of the gel analysis performed in this work (pI 5-8). In a previous report (Duan et al., 2004), where a broader pI range was used for 2-DE, clusterin was identified as two distinct protein ‘spot’ trains, one with pIs of ~5-6, and the other with pI of ~4-5 with a slightly larger Mw. Although not specified in that paper, it is likely that the two trains corresponded to β- and α- chains, respectively (notably, MS/MS sequence coverage was 22.5% for the more acidic train and 31% for the other, implying that only a fragment of the precursor was matched). Similarly in a human plasma 2D gel database, clusterin is also resolved as two trains; one at a pI of ~4.7-5.1 and the other at a pI of~5.2-5.8. Since clusterin is synthesized as a precursor containing both chains, the 140 quantity of β-chain should in theory represent that of the full-length clusterin. Therefore, the finding regarding clusterin β-chain should reflect that of total clusterin levels.

Clusterin did not change in WT mice with age, but increased significantly in bGH mice. Initially lower in bGH mice, clusterin increased to higher levels than WT at later ages. This was true for two isoforms as well as total level, with a similar trend in other isoforms, but not reaching p<0.01 of statistical significance. Total clusterin decreased significantly (p<0.01) in bGH versus WT mice, although not markedly (87% of WT); and this decrease was only because of the lower levels at 2 and 4 months of age. At older ages, clusterin was higher in bGH than WT mice.

Clusterin is expressed in several tissues, mainly in liver, brain and stomach. It is a molecule involved in a wide range of biological processes, and surprisingly has several names, such as sulfated glycoprotein 2 and apolipoprotein-J. Clusterin binds to cells, membranes and hydrophobic proteins for scavenging abnormally aggregated or denatured proteins and is also involved in apoptosis. It is an apolipoprotein that associates with

HDL. Interestingly, clusterin is found to accumulate in atherosclerotic plaques (Jordan-

Starck et al., 1994; Witte et al., 1993). To our knowledge, no report to date mentioned a relationship between GH and clusterin. Thus, changes in clusterin isoforms may be a marker of GH action, specifically, spots 28 and 29 that significantly increased in bGH mice at older ages.

More data concerning clusterin and aging is presented below. In humans at autopsy, it was found that the pituitary content of clusterin increased with age, suggesting a role of clusterin in the aging of the pituitary (Ishikawa et al., 2006). Also, secreted clusterin was 141 higher in the kidney samples of patients with renal cell carcinoma than healthy subjects

(Kurahashi et al., 2005; Miyake et al., 2002). In mice, clusterin is expressed in the kidney during embryonic development but diminishes subsequently (French et al., 1993;

Harding et al., 1991). However, clusterin is induced in kidney under certain injury/disease conditions such as ischemic renal injury, acute kidney injury by ablation and glomerular hypertension (Correa-Rotter et al., 1992; M. E. Rosenberg & Paller,

1991; Singhal et al., 1998). In fact, clusterin is used as a marker for severity of kidney tubular damage in rats (Hidaka et al., 2002). This is very interesting because bGH mice are known to develop progressive glomerulosclerosis as they age (Quaife et al., 1989) and glomerulosclerosis is detected as early as 7 months (T. Doi et al., 1990). Note that clusterin (spots 28, 29 and total) in bGH mice started to surpass controls at 8 months

(Figure 45 and Figure 46).

Additional data on clusterin and kidney damage has been reported. Young clusterin

-/- mice have normal kidneys; however, 21-month old knockout mice develop glomerulopathy (Mark E. Rosenberg et al., 2002). In humans, lower levels of plasma and urine clusterin are associated with renal diseases (Ghiggeri et al., 2002). These studies suggest a protective role of clusterin against renal disease. Therefore, it is possible that bGH mice secret higher levels of clusterin as a protective mechanism for the progressive kidney disease.

It is possible that the higher level of clusterin noted in this study were secreted from the damaged kidneys of bGH mice at 8 months and older. If this possibility could be confirmed, then clusterin could be a marker of excess-GH-induced kidney damage, either 142 directly by GH or indirectly from kidney damage via a yet unknown mechanism.

However, bGH mice had lower than normal clusterin in plasma at 2 and 4 months, suggesting that at young ages, GH suppressed clusterin expression. Clusterin thus could be a very important factor when studying age-related diseases in bGH mice, because it is suppressed in young bGH mice, but is increased dramatically in bGH mice as a function of age and eventually surpassing levels in WT mice.

One isoform of MBP-C is increased in bGH mice

Three isoforms have been identified as MBP-C (Figure 15 in chapter III). MBP-C has a theoretical Mw 24 kDa and pI ~5. PTMs of MBP-C include hydroxylation on proline residues and N-linked glycosylation. The three isoforms identified on 2D gel were ~32 kDa, possibly could be explained by glycosylation and proline oxidation. The isoform reported in this paper increased in bGH mice. The other two isoforms were also higher in bGH mice. However, due to large variability between individual mice, they were not reported here. MBP-C, also called mannan-binding lectin (MBL), is primarily synthesized by the liver and is also an acute phase protein. It binds mannose and N- acetylglucosamine in a calcium-dependent manner. It is an innate immune defense protein that is involved in defense against pathogens by activating the classical complement pathway independently of antibodies (Turner, 2003). In the serum, MBP-C is composed of 18 subunits (6 units of homotrimers) (Kurata et al., 1994). Different human alleles of MBP-C have been identified that affect immune functions, e.g., genetic variations in MBP-C are associated with susceptibility to hepatitis B virus infection 143

(HBV infection) (Thio et al., 2005). Also, low levels of MBP-C were associated with susceptibility to frequent and chronic infections (Sumiya et al., 1991; Turner, 2003).

GH has been linked to modulation of innate immunity (T. K. Hansen, 2003; Kelley et al., 2007). GH treatment of Turner’s syndrome patients increases MBP-C level in plasma (Gravholt et al., 2004). GHD patients have lower and acromegalic patients have higher baseline MBP-C levels than healthy subjects. Also, when treated with GH, healthy subjects have 2-fold increase and GHD patients 4-fold increase in MBP-C, whereas acromegalic patients treated with octreotide or pegvisomant had reduced MBP-C by 33%

(T. K. Hansen et al., 2001). This MBP-C increasing effect of GH was independent of

IGF-1 as IGF-1 treatment did not change MBP-C levels, both in the serum and in cultured hepatocytes (C. M. Sørensen, 2006; T. K. Hansen et al., 2001).

In addition to innate immunity, GH also stimulates acquired immunity (Blazar et al., 1995; Dialynas et al., 1999; Postel-Vinay et al., 1997). GH is thought to have a positive action on T cell development, cytokine production, B cell development, antibody production and priming neutrophils and monocytes for superoxide anion secretion

(Hattori, 2009). However, controversy exists and bGH mice do not appear to have an immune system that better combats infection. On the contrary, they produce less specific antibodies when challenged with tetanus toxoid while GH deficient Ames dwarf mice produce normal levels of antibody (Hall et al., 2002). Thus the complete role of GH in immune function is still not clear. However, the increase of MBP-C observed in bGH mice may be a good indicator of GH action in innate immunity.

144

TTR decreased in bGH mice

As discussed in chapter III, 9 TTR isoforms have been identified in 2D gels. Eight of these isoforms decreased in bGH compared to WT mice at all ages examined.

Interestingly, it has been reported that serum TTR is decreased by GH treatment in Hx male rats of 2 months of age (Vranckx et al., 1994).

TTR is primarily secreted by the liver. It forms a homotetramer that transports thyroxine (T4) and RBP-4. TTR is also synthesized by retinal pigment epithelium (RPE), choroid plexus (Ong et al., 1994; Schreiber, 2002) and pancreas (Itoh et al., 1992;

Jacobsson et al., 1990). TTR synthesized and secreted by choroid plexus is the major protein in the cerebrospinal fluid (CSF) and is involved in the transport of T4 across blood brain barrier (Schussler, 2000). TTR also regulates β-cell function. In cell cultures,

TTR promoted insulin secretion upon glucose stimulation and protected pancreatic β- cells from apoptosis (Refai et al., 2005).

TTR and RBP-4 were found to be decreased in type I diabetic patients and type I diabetic rats (Itoh et al., 1992; Refai et al., 2005; Tuitoek et al., 1996). TTR is also an acute phase protein that is suppressed during inflammation, infection or malnutrition

(Fleck, 1989; Ingenbleek & Young, 1994; Lasztity et al., 2002; Seres, 2005). The lower levels of TTR are associated with higher rate of morbidity/mortality and recovery from acute and chronic disease (Fuhrman et al., 2004). It is possible that decreased TTR in bGH mice indicated an increased inflammation state in these giant mice. In fact, TTR is a negative acute phase protein that is suppressed by inflammation (Fleck, 1989). Therefore the lower levels of TTR in bGH mice indicate an inflammatory state in these mice. 145

Interestingly, total TTR levels in the plasma correlated with those of RBP-4 in WT mice, and to a less extend, in bGH mice (Figure 40 and Figure 42A), which suggests retention of RBP-4 in the plasma from kidney clearance by TTR. Also, TTR increased during aging in both WT and bGH mice (Figure 40). In particular, as shown in chapter

III, one isoform of TTR was a positive biomarker of aging. Together, TTR is suppressed by GH but increased by aging.

ApoA1 decreased in bGH mice

Six apoA1 isoforms have been identified (Figure 15 in chapter III). The theoretical

Mw and pI of apoA1 are 28 kDa and 5.4, respectively. Observed values of these isoforms on our gels were 28 kDa and pI ranging from 5.73-6.4. Of these, spots 17-19 decreased in bGH mice.

ApoA1, synthesized by liver and intestine, is the major protein component in HDL, which is thought to have antiatherogenic and anti-aging effects (Nofer et al., 2005).

Plasma concentration of apoA1 is closely associated with HDL cholesterol, and upregulation of endogenous apoA1 level, either by increasing production or by decreasing clearance, is considered to be a promising approach to improve cholesterol generated cardiovascular disease (Santos-Gallego et al., 2008).

Acromegalic patients had increased serum apoA1 levels after treatment with somatostatin analogues (Sideris P. Delaroudis, 2008). However, several studies have shown decreased apoA1 in GHD patients or Hx rats and increased apoA1 when treated with GH (Cenci et al., 2008; Cuneo et al., 1993; Jan Oscarsson et al., 1991b; Svensson et al., 2000). Thus, too much and too little GH both reduce apoA1. bGH mice mimic 146 acromegaly in humans, and the total apoA1 in bGH mice indeed was decreased

(p=0.011). Importantly, three isoforms decreased with p<0.01, with others not changing significantly. This suggested that over-expression of GH decreased apoA1 plasma levels, possibly affecting the PTMs, and therefore the composition of apoA1 isoforms in plasma.

It is not clear how this is regulated, but may be relevant to cardiovascular disease propensity when GH is over-expressed.

RBP-4 levels decrease as well as isoform-specific changes in bGH mice

RBP is in theory 21.6 kDa with a pI of 6. We observed 7 isoforms, with isoforms 1-

3 being 22 kDa and isoforms 4-7 being 19 kDa, and a wide range of pI from 5.28 to 7.61

(Figure 15). It appeared that the 22 kDa isoforms were the full length, and the 19 kDa isoforms may be a fragment. The different pIs of the 19 kDa RBP-4 isoforms are likely the result of phosphorylation (unpublished observation by Okada).

RBP-4 is mainly secreted by liver and adipose tissue. RBP-4 binds to retinol and delivers it from liver to peripheral tissues. In blood, RBP-4 is bound to TTR homotetramer, which increases its half life by reducing kidney filtration. RBP-4 has been associated with obesity and insulin resistance (Haider et al., 2007; Kowalska et al., 2008;

Kohzo Takebayashi et al., 2007a) and has been found to increase in type II diabetic subjects (Raila et al., 2007).

Since RBP-4 is associated with obesity, it is not surprising that total RBP-4 in bGH mice was lower than WT, as bGH mice had much less fat than WT mice. Similar results have been reported. For example, serum RBP-4 was decreased in GHD children after 6 weeks of GH treatment (Kemp & Canfield, 1983). In Hx rats, vitamin A (a retinol that is 147 bound by RBP-4) was cleared from blood much slower than controls, and a single GH injection decreased vitamin A in plasma and increased its level in the urine significantly

(Ahluwalia et al., 1980). These results suggest that GH promotes vitamin A clearance, possibly by clearance of RBP-4 and thus decreasing its levels in plasma. Since GH also decreases TTR, GH may reduce RBP-4 through reducing TTR. Interestingly, like TTR,

RBP-4 is also considered a negative acute phase protein (Fleck, 1989), whose decreased levels indicate inflammation. Thus, by decreasing both TTR and RBP-4, excess GH may promote inflammation in bGH mice.

In addition to total RBP-4 reduction, RBP-4 also showed isoform-specific changes.

For example, despite overall decrease and in particular spot 21, spot 9 actually increased.

Notice that these two isoforms were of different sizes. Spot 9 was 19kDa and spot 21 was

22kDa (Figure 33). This indicated that GH favored the accumulation of 19 kDa rather than 22 kDa RBP-4. The nature of the PTM responsible for the size difference may therefore be regulated by GH. For example, GH may upregulate a proteolytic enzyme that cleaves the full-length RBP-4 to give rise to the 19kDa fragment.

A2m showed isoform-specific changes in the plasma of bGH mice

Seven isoforms have been identified as a2m (Figure 15 in chapter III and appendix

A). In mice and rats, a2m is composed of two subunits- 165 kDa and 35 kDa, presumably through intracellular proteolytic processing of a precursor protein by an yet to be defined proteinase (van Leuven et al., 1992). In this study, the 35 kDa subunit was detected

(Figure 33), and MS/MS also matched peptides in this subunit (Table 15 and Figure 84 in appendix A). The 165kDa subunit was beyond the range of our gels and therefore was 148 not detected; however, since the two subunits are held together by disulfide bond, and are disrupted during the 2-DE process, this 35kDa subunit should represent the presence of the intact protein. The 35kDa subunit of a2m has a theoretical pI of 7.09 and the seven observed isoforms had pI ranging from 5.65 to 7.08, which means isoform 7 (refer to

Figure 15) is most likely the un-modified form and isoforms 1-6 have been modified by

PTM(s). A2m is phosphorylated (Villen et al., 2007) and glycosylated on multiple Asn residues in both subunits (Ghesquiere et al., 2006). These PTMs are likely the reason for a2m to exist as multiple isoforms with similar sizes but different pIs on the 2D gel. Using the above-mentioned tool to calculate Mw and pI with possible phosphorylation modification (http://scansite.mit.edu/calc_mw_pi.html), it was revealed that isoform 2

(observed pI 5.85) was consistent with addition of three phosphate groups (predicted pI

5.81) and isoform 3 (observed pI 5.08) was consistent with addition of two phosphate

groups (predicted pI 5.08). Other isoforms have observed pIs not explained by

phosphorylation alone, and may be subjected to glycosylation and/or a combination of

phosphorylation and glycosylation.

Interestingly, although total a2m did not change, four of these isoforms showed differential regulation by GH. Spot 1 and 3 increased whereas spots 15 and 16 decreased in bGH mice, suggesting that GH regulated the PTMs of plasma a2m.

A2m is a plasma proteinase inhibitor that is mainly secreted by liver (Lorent et al.,

1994). It is also expressed in ovary, heart, stomach, kidney, lung and uterus (He et al.,

2005). It binds to all four classes of proteases (including aspartic, métallo, cysteine, and serine proteases) to form a complex, which is cleared via receptor-mediated endocytosis 149

(Van Leuven et al., 1994). Both native (free) and conformationally transformed (when binding to a protease) a2m can bind a variety of cytokines, growth factors and hormones, and is thought to play roles in hormonal control, immune modulation, signal transduction, growth and differentiation (Tayade et al., 2005). Interestingly, in plasma, the conformationally transformed a2m binds GH, although with weaker affinity than GH binding protein (BP) (Adham et al., 1969; Kratzsch et al., 1995, 1996). A2m also binds to IGFBP-1 and modifies IGF-I/IGFBP-1 actions resulting in enhanced IGF effects

(Westwood et al., 2001). A2m is an acute phase protein induced by inflammation (de

Boer et al., 1993; Fleck, 1989). A knockout mouse model deficient in a2m were normal under standard conditions but had an increased mortality rate and more severe symptoms when subjected to acute pancreatitis (Umans et al., 1999).

GH increased a2m production in a human hepatoma cell line (Derfalvi et al., 2000).

Acromegalic patients had increased a2m levels (Pozzilli et al., 1978). Patients with somatotropic adenoma had elevated levels of a2m in blood, but normalized one year after

Hx, together with reduced serum GH levels (Zargarova & Konnova, 1983). A2m is induced by prolactin in rat ovarian granulosa cells (Dajee et al., 1996) and decidual cells

(Barkai et al., 2000) through the Jak2-Stat5 pathway. Although it has not been reported, it is possible that GH induces a2m in the liver through a similar mechanism. My results indicate a novel regulation of GH on a2m through PTM, rather than at the total level change.

150

Conclusion

2-DE of the plasma of bGH and WT mice at different ages revealed that several proteins increased (apoE, Hp and MBP-C) or decreased (TTR, RBP-4 and apoA1).

• All five apoE isoforms identified increased in bGH mice compared to WT

mice at all ages, resulting in a total increase of 6 fold.

• All four Hp isoforms increased in bGH mice compared to WT mice, resulting

in a total increase of more than 8 fold.

• Out of three isoforms of MBP-C, one increased more than 8 fold in bGH

compared to WT mice.

The above isoforms and/or proteins are positive biomarkers of GH.

• Out of nine isoforms of TTR, eight decreased at all ages, resulting in a 60%

reduction of total TTR in bGH compared to WT mice.

• Out of six apoA1 isoforms, three decreased, ranging from 46-66% reduction

compared to WT levels. Moreover, total apoA1 decreased in bGH mice

(p=0.011). Thus, apoA1, in particular these three isoforms, are negative

biomarkers of GH.

• One isoform of RBP-4 (spot 9) increased 3 fold in bGH mice; however,

another isoform (spot 21) decreased 40%. The total RBP-4 is reduced 25% in

bGH mice compared to WT mice. Therefore, total RBP-4 as well as spot 21

are negative biomarkers of GH, but spot 9 is a positive marker.

• Similarly, out of the seven a2m isoforms, two increased (spots 1 and 3, four

and two fold, respectively) and two decreased (spots 15 and 16, 26% and 151

57% reduction, respectively). Thus, spots 1 and 3 of a2m are positive

markers and spots 15 and 16 are negative markers of GH.

Of particular interest is that Hp and clusterin exhibited significant interaction between GH and aging. All four Hp isoforms increased more vibrantly in bGH mice during aging than in WT mice. Two isoforms of clusterin as well as total clusterin increased during aging to a higher extent in bGH mice versus controls. Therefore, Hp and clusterin are possibly related to the accelerated aging in bGH mice.

Many of the proteins are secreted by liver as acute phase proteins (proteins produced by liver to promote inflammation, activate the complement cascade, and stimulate chemotaxis of phagocytes in response of acute infection, injury and inflammation) either induced (Hp, MBP-C, a2m) or suppressed (TTR, RBP-4) by acute inflammation, infection or other immune insults. One common emerging theme is that in the bGH mice,

‘positive’ (protein that is induced by inflammation) acute phase proteins (Hp, MBP-C) increase, but ‘negative’ ones (protein that is suppressed by inflammation, such as TTR and RBP-4) decrease, suggesting an increased inflammatory state in bGH mice. In fact, many acute phase proteins are induced by cytokines through Jak-Stat signaling pathways.

GH signals through a Jak2-Stat5 pathway (Frank et al., 1994; Kopchick & Andry, 2000;

Lanning & Carter-Su, 2006; Waters et al., 2006). Acute phase proteins are a part of innate immune system and are important players in combating acute infection and injury.

However, the chronic upregulation of these proteins in the absence of an ‘acute infection/injury’ in bGH mice reflect a inflammation-prone phenotype. This could explain partially the shortened lifespan of bGH mice, since increased inflammation is 152 associated with aging (Brüünsgaard & Pedersen, 2003; Ferrucci et al., 2005; Roubenoff et al., 1998).

Another theme appears to be that GH increases VLDL and LDL-associated lipoprotein (apoE), but decreased HDL-associated lipoprotein (apoA1 and clusterin). It is known that HDL is the ‘good’ cholesterol and beneficial to cardiovascular system and

VLDL and LDL are ‘bad’ and increase cardiovascular disease risks. This could also contribute to reduced lifespan of bGH mice. However, bGH mice are reported to have increased HDL and LDL, and decreased VLDL at 6 months (Frick et al., 2001a). Since apoE is also an apolipoprotein for HDL, it is possible that the increased HDL cholesterol in bGH mice is carried by apoE, compensating for reduced apoA1 and clusterin which are carriers of HDL. This suggests that bGH mice not only have altered lipid profiles, but also altered apolipoproteins associated with HDL and LDL. The significance of this needs further study.

In conclusion, several plasma proteins have been identified that are up- or downregulated in bGH mice. These proteins and their specific isoforms are biomarkers of chronic GH action. Some proteins show isoform-specific changes, thus it will be important to identify PTMs corresponding to these isoforms. These biomarkers of GH indicate an increased inflammatory state and an altered lipid profile in bGH mice.

153

CHAPTER V. BIOMARKERS OF GH AND GENDER: PROTEOMIC ANALYSIS OF

PLASMA FROM GROWTH HORMONE RECEPTOR GENE DISRUTPED MICE

Abstract

Growth hormone receptor gene knockout (GHR KO or GHR-/-) mice are dwarf, insulin sensitive and long-lived despite having an obese phenotype. We hypothesize that the plasma protein profiles of these mice in comparison with wild type (WT) control mice would possess novel biomarkers for the lack of GH action and may also be correlated with longevity. In this paper, we have analyzed plasma proteins of GHR-/- and WT mice of both genders using two-dimensional electrophoresis (2-DE) and identified several plasma markers for the lack of GH action. GHR-/- mice had increased apolipoprotein A-4

(apoA4), decreased apoE and isoform-specific changes in retinol-binding protein-4

(RBP-4) and apoA1. Interestingly, certain proteins showed gender difference. Female mice had increased isoform(s) of apoE and RBP-4; decreased apoA1, clusterin, Hp and hemoglobin beta chain; and isoform-specific changes in albumin. A significant interaction between gender and aging was found in the levels of isoform(s) of apoA4, apoA1, transthyretin, albumin and RBP-4. Specific isoform(s) of apoE, RBP-4 and Hp showed a significant interaction between gender and genotype. Together, these data demonstrated marked difference between GHR-/- and WT mice, as well as a gender specific difference in plasma proteins. 154

Introduction

Growth hormone receptor/binding protein gene knockout (GHR KO or GHR-/-) mice are dwarf and have markedly reduced serum IGF-1 and elevated GH levels. They are obese, yet insulin sensitive and long-lived (Berryman et al., 2004; Coschigano et al.,

2000; Coschigano et al., 2003; Liu et al., 2004). Many studies on GHR-/- mice show phenotypic differences including their increased adipose tissue mass, particularly in the subcutaneous depot (Berryman et al., 2004; Berryman et al., 2006; Liu et al., 2004).

These mice also have reduced muscle fiber size (Schuenke et al., 2008; Sotiropoulos et al., 2006) and reduced bone length and bone mineral density (Sjogren et al., 2000). GHR-

/- mice have a disproportionally enlarged brain compared to WT animals (Asa et al.,

2000; Ransome et al., 2004) and have less age-associated memory loss that is seen in wild type (WT) mice (Kinney et al., 2001). Additionally, GHR-/- mice have delayed sex maturation and are sub-fertile (A. Bartke, 2000; Chandrashekar et al., 2001a;

Chandrashekar et al., 2001b; Keene et al., 2002; Zhou et al., 1997). Despite a smaller heart phenotype, GHR-/- mice have normal cardiovascular function (E. Egecioglu et al.,

2007; Izzard et al., 2009). GHR-/- mice are resistant to streptozotocin -induced kidney damage (Bellush et al., 2000), in contrast to bGH mice, which progressively develop glomerulosclerosis (T. Doi et al., 1990; T. Doi et al., 1988; Quaife et al., 1989). Finally,

GHR-/- mice have decreased cancer incidence, including breast, lung, pituitary, lymph and prostate (Ikeno et al., 2009; Kineman et al., 2001; Z. Wang et al., 2005; X. Zhang et al., 2007). The longevity seen in this mouse model is thought to be at least in part 155 overlapping with life extension in caloric restricted animals (Bonkowski et al., 2009; R.

A. Miller et al., 2002).

Despite much research on the GHR-/- mice, little has been done at the proteomic level with even less associated with the life span of these mice. Proteomics explores the entire set of proteins in a given tissue under a certain condition. Novel proteins that are differentially regulated in GHR-/- mice versus WT will provide insight into the long- lived phenotype of this mouse model. Specifically, we hope to identify plasma proteins that serve as potential markers of the lack of GH action and increased longevity. In this longitudinal study, plasma proteins were analyzed in both male and female GHR-/- mice of adult (8 months), middle age (16 months) and old (24 months) relative to WT controls using a two-dimensional gel electrophoresis (2-DE). Proteins differently expressed in the plasma were identified by mass spectrometry (MS) and tandem MS (MS/MS).

156

Materials and Methods

Experimental animals

The GHR-/- mouse line was described previously (Zhou et al., 1997). The mice were

originally generated in the background of 129Ola/BalbC. They were subsequently

backcrossed with C57BL/6J mice for at least eight generations. GHR-/- mice of both

genders (n=6 for both male and female) and their wild type (WT, or GHR+/+) littermate

controls (n=5) were analyzed at the following ages: 8, 16 and 24 months. These ages

represent young-middle (8 months), middle (16 months) and old age (24 months) in mice

(Richard A. Miller & Nadon, 2000). Mice were housed 2-3 per cage at room temperature

(22 oC) in a 12-hour light, 12-hour dark cycle. Mice were fed ad lib with chow diet.

Animal protocols were approved by Ohio University’s Institutional Animal Care and Use

Committee.

Body composition measurement

Body composition of 12-month old mice was measured using mq 7.5 NMR analyzer

the Bruker Minispec (The Woodlands, TX) as described previously (List et al., 2009;

Palmer et al., 2009). Each animal was measured twice and the mean value used.

Plasma IGF-1 measurement

Blood was collected with heparinized capillary tubes by tail tip clipping following

heat lamp exposure. Whole blood was centrifuged at 7000 x g for 10min at 4 ºC and the

resulting plasma was stored at -80 ºC. Plasma IGF-1 levels were determined using the

DSL-10-29200 mouse/rat IGF-1 ELISA kit by Diagnostic Systems Laboratories, Inc.

(Webster, TX). 157

Fasting glucose and insulin measurements

Mice were fasted for 4 hours and bled at 3 PM. Blood glucose was measured by

ONE TOUCH glucometer from Lifescan (Milpitis, CA). Plasma was collected as

described above and insulin levels were measured by an ultrasensitive rat/mouse insulin

ELISA kit (ALPCO, Windham, NH).

2-DE and quantification of proteins

See chapter III.

Statistical analysis

All statistical analyses were performed using SPSS 14.0 software (Chicago, IL).

Data for IGF-1 was subjected to independent t test with p<0.05 as significant; data for

body composition, fasting glucose and insulin were subjected to 2-way ANOVA test with gender and genotype as two factors (p<0.05). Data for body weight were subjected to repeated measures with significance value p<0.05. For protein intensity, the log- transformed data were subjected to repeated measures of three ages to examine an aging effect, with genotype and gender as two fixed factors to examine the effects of GH and sex (p<0.01). All data are presented as mean ± SEM.

Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography

(LC)-MS/MS

See chapter III.

158

Results

Body weight and body composition

As mice aged from middle age (9 months) to old age (24 months), they showed

significant weight change (two-way repeated measures, p<0.001, Figure 47 A). All

animals gained weight from 9 to 16 months, and then showed a slight weight loss. GHR-

/- mice had significantly reduced body weight than WT mice (p<0.001). There was no

significant gender difference using repeated measures. However, if each age was viewed

separately with 2-way ANOVA, female WT had reduced weights relative to male WT

mice at 9 (p<0.01) and 12 months of age (p<0.05). Interestingly, GHR-/- had a smaller

gender difference than WT mice. Also, female mice (both WT and GHR-/-) were less

susceptible to weight loss at 24 months than males.

Figure 47. Body weight from 9 to 24 months and body composition at 12 months. A, body weight of mice of different ages. Two-way repeated measures showed a significant 159 effect of aging (p<0.001), genotype (p<0.001) and interaction between aging and genotype (p<0.05). B, body composition at 12 months in absolute mass. C, stacking bar graph of mean body composition in each group at 12 months. D, normalized lean mass (percentage of body weight) at 12 months. E, normalized fat mass (percentage of body weight) at 12 months. F, normalized fluid mass (percentage of body weight) at 12 months. Different letters denote significant difference (p<0.05).

Body composition was measured at 12 months of age (middle age). The absolute mass of all body composition components was reduced in GHR-/- compared to WT mice

(p<0.001 for body weight, lean and fluid mass, p=0.002 for fat mass, Figure 47 B). There was no gender difference in body composition in GHR-/- mice. In contrast, female WT had a reduced body weight (p=0.033) and lean mass (p=0.001) relative to male WT mice.

Male and female WT mice had similar fat and fluid masses (Figure 47B). From Figure

47 C, it is clear that the weight difference was entirely due to reduced lean mass in female

WT mice, whereas fat and fluid mass were similar in both genders. Similarly, normalized lean mass (percentage of body weight) was significantly lower in GHR-/- compared to

WT (p<0.01) without a gender effect (Figure 47 D); whereas normalized fat mass was significantly higher in GHR-/- mice (p<0.01) without a gender effect (Figure 47 E).

Normalized fluid mass, on the other hand, exhibited a significant interaction between gender and genotype (p<0.01), being higher in WT female than WT male, but the same in

GHR-/- females and males (Figure 47 F). Together, GHR-/- mice had disproportionally increased fat and decreased lean mass with no change in fluid mass compared to WT mice. There was no gender difference in GHR-/- mice, whereas in WT mice, females had reduced lean mass resulting in reduced total body weight. 160

IGF-1, fasting glucose and insulin levels

IGF-1 levels were determined for male mice at 6 months of age (Figure 48 B). IGF-

1 was significantly lower in GHR-/- than WT mice (p<0.0001) as has been reported

previously (Coschigano et al., 2000; Coschigano et al., 2003). Two-way ANOVA was

used to test for the effects of gender and genotype on 4h-fasting glucose and insulin

levels at two ages: 9 and 24 months. At 9 months, GHR-/- mice had significantly lower

fasting glucose levels than WT controls (p<0.05), and females had significantly higher

levels than males (p<0.05). However, no significant difference was found among the four

groups at 24 months (Figure 48A). Fasting insulin levels, on the other hand, exhibited interesting features. At both ages, insulin was significantly lower in GHR-/- mice than

WT controls (p<0.001). At 9 months, insulin was significantly lower in females than males in both genotypes (p<0.01), but not significantly different between genders at 24 months (Figure 48A). These valuses are similar to those reported previously (Al-Regaiey et al., 2005; Coschigano et al., 2003).

Together, these data suggest that GHR-/- mice (both male and female) have lower insulin levels than their WT counterparts at adult and old age, but have lower glucose levels than WT only at 9 months of age. Regarding gender difference, females have lower insulin levels than males at an old age, but not at a younger age. There is no gender difference in glucose levels.

161

Figure 48. IGF-1 and fasting glucose and insulin levels. A, 4-hour fasting glucose and insulin levels at 9 and 24 months. Different letters denote significant difference (p<0.05). B, IGF-1 levels of male mice at 6 months. *p<0.0001.

Plasma proteins that exhibited significant gender differences

Plasma proteins that showed changes in genotype, gender and aging revealed by

two-way repeated measures were labeled 1-29 on a typical 2-D gel as shown in Figure

49. Proteins with significant gender difference are listed in Table 8 and those different by

genotype listed in Table 9.

162

Figure 49. 2D gel map of identified proteins of interest. Proteins are marked by ‘x’ and numbered 1-31. Abbreviations: pI: isoelectric focusing point; Mw: molecular weight.

Table 8. Proteins significantly different in males and females (p<0.01) Ratio Spot # Protein ID Isoform # (F/M) 4 apoE 1 ↑3.2 16 RBP-4 7 ↑2.4 Total RBP-4 Sum ↑1.2 5 Clu 3 ↓0.27 14 RBP-4 3 ↓0.31 22 Hp 2 ↓0.08 25 Alb 19 ↓0.25 26 Hbb 1 ↓0.11 27 Hbb 3 ↓0.05 Total apoA1 Sum ↓0.85 Spot # corresponds to Figure 49; isoform # corresponds to Figure 15 in chapter III. The ratio F/M was calculated as the mean value of female/mean value of male (all samples regardless of genotype). ‘↑’ means up-regulation and ‘↓’ means down-regulation.

Spots 4 (apolipoportein-E or apoE isoform) and 16 (retinol-binding protein-4 or

RBP-4 isoform) were significantly higher in female compared to male mice regardless of 163 genotype (p<0.01, Figure 49 and Table 8). Since no genotypic difference was found for these proteins, the reported quantity ratio of female (F)/male (M) was calculated from pooled values of both genotypes. The same applies to Table 9, where no significant gender difference was found for a given protein, the data from both genders were pooled together. Spot 4 increased 3 fold and spot 16 increased 2 fold in females versus males.

Total levels of RBP-4 (sum of all identified RBP-4 isoforms) also increased significantly

(p<0.01) in females, although only at 1.2 fold.

Certain isoform(s) of clusterin (Clu), RBP-4, albumin (Alb), haptoglobin (Hp) and hemoglobin β-chain (Hbb) decreased significantly in females (p<0.01, Table 8). Total apoA1 also decreased significantly (p<0.01), although no individual spot of apoA1 showed significant decrease or increase. This may be due to an accumulative effect when addting up all the individual isoforms.

In addition, 14 proteins showed significant interaction between gender and aging

(p<0.01, Figure 50), including two isoforms of apoA4, 4 isoforms of apoA1, one isoform of transthyretin (TTR) and RBP-4, as well as 6 isoforms of Alb.

Figure 50. Plasma proteins that exhibited significant differential change during aging between males and females (p<0.01). Protein # corresponds to Figure 49. A, proteins that increased during female aging but did not change in male aging. B, proteins that decreased during female aging but did not change in male aging.

165

Spots 1 and 2, two apoA4 isoforms, increased during aging in females with no aging difference in males (Figure 50, Figure 51 and Figure 52). Spots 8-11, four apoA1 isoforms, increased in female aging but did not show significant change in male aging

(Figure 50 and Figure 53). Spot 17, a TTR isoform, increased in female aging to a larger degree than in male aging (Figure 50 and Figure 54). Spot 14, a RBP-4 isoform, decreased in female aging but increased at 16 months then decreased at 24 months in male aging (Figure 50 and Figure 55). Two albumin spots, 28 and 29, increased in female aging without change in male aging (Figure 50). However, four other albumin spots decreased in female aging with no change in male aging (Figure 50; also see Figure

54, Figure 56 and Figure 57).

Plasma proteins that exhibited significant genotypic differences

Four proteins showed a significant increase and three showed a significant decrease

in GHR-/- compared to WT controls (p<0.001, Table 9). Two isoforms (spots 2 and 3) as

well as total of apoA4 were upregulated in GHR-/- mice (Figure 51 and Figure 52). Two

isoforms of RBP-4 (spots 13 and 14) were also upregulated in GHR-/- mice (Figure 55

and Figure 58). Another isoform of RBP-4 (spot 15), however, decreased in GHR-/- mice

(Figure 56). One isoform of apoA1 (spot 8, Figure 53), as well as three isoforms (spots 4,

6 and 7, Figure 51 and Figure 52) and total levels of apoE decreased in GHR-/- mice.

166

Table 9. Proteins significantly different in GHR-/- compared to WT mice (p<0.01) Spot # Protein ID Isoform # Ratio (KO/WT) 2 apoA4 3 ↑1.7 3 apoA4 4 ↑1.9 12 RBP-4 1 ↑1.8 13 RBP-4 2 ↑3.0 14 RBP-4 3 ↑4.2 Total apoA4 Sum ↑1.6 4 apoE 1 ↓0.12 6 apoE 3 ↓0.40 7 apoE 4 ↓0.60 8 apoA1 2 ↓0.83 15 RBP-4 5 ↓0.45 Total apoE Sum ↓0.61 Spot # corresponds to Figure 49; isoform # corresponds to Figure 15 in chapter III. The ratio of KO/WT was calculated as the mean value of KO/mean value of WT (all samples regardless of gender). ‘↑’ means up-regulation and ‘↓’ means down-regulation. KO: GHR-/- mice.

Figure 51. Cropped gel image from a representative mouse in each group showing increased intensity of apoA4 and decreased intensity of apoE in GHR-/- versus WT mice. Spots 1 and 2 increased as females aged but did not change during male aging. ApoA4 spots are in the oval circle and apoE spots in the squares. Protein # corresponds to Figure 49. 167

Figure 52. 3-D view of intensity of apoA4 (spots 1-3) and apoE (spots 4, 6 and 7). Spot intensity was converted into topographic view of a peak by PDQuest. Protein # corresponds to Figure 49. ApoA4 increased, whereas apoE decreased in GHR-/- versus WT in both genders. For apoA4, spots 1 and 2 showed no change during aging in males, but increased during aging in females. For apoE, spot 4 increased in females, and spot 7 was the lowest in female GHR-/- compared the other 3 groups.

168

Figure 53. 3-D view of intensity of apoA1 (spots 8-11). Protein # corresponds to Figure 49. Cropped images of original gel are shown in the center at the bottom of each graph. Spot 8 decreased in GHR-/- mice. Spots 8-11 showed increase during aging in females, but unchanged in males.

169

Figure 54. 3-D view of intensity of RBP-4 (spot 16), TTR (spot 17) and Alb (spots 19 and 20). Protein # corresponds to Figure 49. Spot 16 increased in female compared to male mice. Spot 17 increased in female aging, spots 19 and 20 decreased in female aging with no change or fluctuation in male aging.

170

Figure 55. 3-D view of intensity of RBP-4 (spot 14). The number of the spot corresponds to that in Figure 49. This spot decreased in female aging but fluctuated in male aging.

171

Figure 56. 3-D view of intensity of RBP-4 (spots 15) and Alb (spot 18). Protein # corresponds to Figure 49.

172

Figure 57. 3-D view of intensity of RBP-4 (spot 12) and Alb (spot 21). Protein # corresponds to Figure 49.

173

Figure 58. 3-D view of intensity of RBP-4 (spot 13). Protein # corresponds to Figure 49. This spot increased in GHR-/- compared to WT mice.

Plasma proteins that exhibited significant interaction between gender and genotype

Four proteins showed a significant interaction between gender and genotype

(p<0.01), i.e., whether levels in GHR-/- were higher or lower than WT depended on gender (Figure 59). Spot 7, an apoE isoform, had the same level between GHR-/- and

WT in males, but was lower in GHR-/- females relative to controls (Figure 59 and 174

Figure 52). Spot 12, an RBP-4 isoform, showed no difference between GHR-/- and WT in males, but was higher in GHR-/- females versus controls (Figure 59 and Figure 57).

Two Hp isoforms, spots 23 and 24, were lower in male GHR-/- but were higher in GHR-

/- females versus controls (Figure 59 and Figure 60).

Figure 59. Plasma proteins that exhibited significant interaction between gender and genotype (p<0.01). Protein # corresponds to Figure 49.

Figure 60. 3-D view of intensity of Hp (spots 23 and 24). Protein # corresponds to Figure 49. These images are from mice at 24 months of age. 175

Discussion

Body composition, glucose and insulin

All animals gained weight from 9 to 16 months, but males had a slight weight loss at 24 months. This was consistent with the data presented regarding WT male mice

during aging in chapter III. Interestingly, in females, body weight plateaued at 24 months,

suggesting a gender-specific effect of aging on body weight. Since females in general live

longer than males (Waldron & Johnston, 1976), the protection from weight loss at an old

age may imply that females are more ‘fit’ than males at advanced ages.

GHR-/- mice had about 50% reduction in body weight compared to WT mice.

There was no gender difference in GHR-/- mice, but female WT mice weighed less than

WT males at 9 and 12 months, although the difference was diminished at 16 and 24 months. This suggested that an intact GH signaling system had a sexually dimorphic effect on body weight. For example, GH secretion is pulsatile in males but less so in females (Leppaluoto et al., 1975; Saunders et al., 1976; van den Berg et al., 1996), and estrogen inhibits GH action in females (Meinhardt & Ho, 2007), contributing to reduced lean mass in females. Interestingly, female WT mice weights caught up with WT males at old ages, perhaps due to reduced GH and reduced estrogen (in females) at older ages.

GHR-/- mice had decreased lean mass %, increased fat mass %, and no change in fluid mass % compared to WT controls, as has been reported previously (Berryman et al.,

2004). Again, there was no gender difference found in GHR-/- mice. In contrast, female

WT mice had a reduced lean mass relative to WT males, suggesting that an intact GH signaling system is required to maintain a gender-dependent body composition. 176

Fasting glucose levels were lower in GHR-/- than WT mice at an adult age (9 months), but were similar at an old age (24 months). Fasting insulin levels were lower in

GHR-/- mice at both ages. These data were consistent with previous studies in which

GHR-/- mice were slightly hypoglycemic but normalized with age (Coschigano et al.,

2003). GHR-/- mice also had increased insulin sensitivity throughout their lifespan (Al-

Regaiey et al., 2005; Coschigano et al., 2003). Interestingly, females had lower insulin than males, similar to a previous report (Guo et al., 2005).

In summary, body weight and composition not only revealed a well-known genotype difference (more fat and less lean in GHR-/- compared to WT mice), but also a gender difference found only in WT (females had less lean mass) and not GHR-/- mice.

Blood glucose and insulin, on the other hand, showed a similar gender difference in both genotypes (lower insulin in females), in addition to the well-known genotype difference

(lower insulin and glucose in GHR-/- mice).

Plasma protein characteristics

ApoA4 increased in GHR-/- versus WT mice and increased during aging in females but not in males

ApoA4 is in theory 43 kDa with a pI of 5.27. The five identified isoforms in this study were ~43 kDa with pI ranging from 5.64-5.9 (Figure 15 in chapter III, also see appendix A). ApoA4 is known to be phosphorylated, but this is not likely the only reason for the observed pI difference, as phosphorylation shifts spots to more acidic rather than basic pI. 177

Three out of four isoforms showed interesting changes. Two of them increased in

GHR-/- mice (spots 2 and 3), as did total apoA4. In addition, spots 1 and 2 increased during aging in female compared to male mice of both genotypes.

ApoA4 is secreted by liver and intestine and is a major component of high-density lipoprotein (HDL). Most plasma apoA4 is associated with lipids (84%), about 12% associated with apoA1 and only 4% in free form (Ezeh et al., 2003). It is involved in chylomicrons and very low-density lipoprotein (VLDL) secretion and catabolism. In addition to cholesterol transportation and metabolism, apoA4 is also an anti-oxidant protein that removes superoxides and lipid hydroperoxide (Ostos et al., 2001; W.-M. R.

Wong et al., 2007). It has anti-atherosclerosis effect and reduces inflammation (Recalde et al., 2004). Over expression of apoA4 in apoE-/- mice protect these mice from diet- induced atherosclerosis (Duverger et al., 1996). ApoA4 is a marker for coronary heart disease, with lower levels indicating higher risk (Warner et al., 2001). Patients with coronary artery disease have lower levels of apoA4 (Kronenberg et al., 2000).

It is possible that increased apoA4 levels in GHR-/- mice conferred protective effects against oxidative stress and heart diseases on these long-lived mice. It is also known that females generally live longer than males, and that males have higher rates of mortality from cardiovascular diseases (Waldron & Johnston, 1976). Therefore, the increase during aging of apoA4 in females, but not in male mice, may further imply a beneficial effect of circulating apoA4 in females.

178

ApoE decreased in GHR-/- versus WT mice; one isoform increased in female versus male mice; and one isoform showed interaction between genotype and gender

Of the five apoE isoforms identified (Figure 15 in chapter III), three of them showed changes in this study. ApoE has a theoretical Mw of ~34 kDa and the observed

Mw on the gel was ~34 kDa. An additional apoE spot was also detected (spot 4 in Figure

49) which was ~38 kDa. The PTM for the size difference of this isoform is not clear.

Although not reported in mice, in human plasma, apoE is glycosylated and glycated, which could in theory result in protein spots of slightly different Mws and pIs on a 2D gel.

GHR-/- mice had decreased apoE isoforms (spots 6 and 7) as well as total apoE.

ApoE is associated with HDL and VLDL. Reduced apoE in GHR-/- mice may therefore reflect the fact that they have reduced plasma cholesterol levels (both HDL and LDL) compared to WT mice (A. Bartke et al., 2004; E. Egecioglu et al., 2005; Masternak et al.,

2005a). This is consistent with the previous result where bGH mice had increased apoE in all isoforms (chapter IV). Hypophysectomized (Hx) rats had decreased apoE which normalized upon GH replacement (Frick et al., 2002; Jan Oscarsson et al., 1991b;

Sjoberg et al., 1994a; Sjoberg et al., 1994b). In this regard, apoE shows a positive response to GH.

One isoform (spot 4, the 38kDa isoform) of apoE was increased in females, although total apoE levels did not differ. It is possible that this specific isoform, or the

PTM responsible for the Mw increase, is regulated by gender differences. Female rats have been found to have higher levels of apoE than males (Patsch et al., 1980). In 179 humans, apoE is involved in age-related diseases including Alzheimer’s disease, dementia, cognitive decline, stroke and cardiovascular diseases in a gender-specific manner (Frikke-Schmidt et al., 2004; Galasko et al., 2007; Ghebremedhin et al., 2001;

Gromadzka et al., 2007; Mortensen & Hogh, 2001; Rosvall et al., 2008). For example, apoE allele ε4 in men is associated with increased mortality whereas ε2 in women is associated with decreased mortality (Rosvall et al., 2008). ApoE ε4 is also an independent predictor of death within 1 year after stroke incidence in men but not in women (Gromadzka et al., 2007). However, ε4 allele is associated with age-related cognitive decline in women but not in men (Mortensen & Hogh, 2001). An association between the ε4 allele and Alzheimer’s disease-related senile plaques was found in men but not women over 80 years of age (Ghebremedhin et al., 2001). Thus, apoE affects males and females differently.

Different alleles of apoE could result in different isoforms on a 2-D gel. However, it is not known if the C57Bl/6J mice used in our study have different alleles of apoE.

Theoretically, these in-bred mice should be genetically identical; therefore they should have the same allele of apoE. Thus the isoforms observed on 2D gels are likely due to

PTM instead of different alleles.

Interestingly, one isoform (spot 7) of apoE showed significant interaction between gender and genotype such that there was no gender difference in WT, but was reduced in

GHR-/- females, suggesting an intact GH signaling is required for the gender difference of this apoE isoform.

180

ApoA1 decreased in female mice, increased during aging in females but not in males

Six isoforms of apoA1 were identified (Figure 15 in chapter III). Theoretical Mw and pI of apoA1 are ~28 kDa and ~5.4, respectively. Observed values of the five isoforms on our gels were 28 kDa and pI ranging from 5.73-6.4.

Spot 10, an apoA1 isoform, decreased, although total apoA1 did not change in

GHR-/- mice versus controls. Total apoA1 decreased in female mice versus male mice of both genotypes. Four isoforms (spots 8-11) increased during aging in females but did not change in males. ApoA1 associates with HDL, and is therefore considered a beneficial marker for cardiovascular disease risk. Even though total apoA1 decreased in females, it showed an increased trend during aging, possibly indicating an increased protective effect.

GHR-/- mice have reduced HDL (E. Egecioglu et al., 2005; Liu et al., 2004).

However, GHR-/- mice have normal total apoA1 in this study, although one isoform decreased compared to WT mice. Laron Syndrome patients have normal levels of HDL

(Shechter et al., 2007). Acromegalic patients have reduced HDL (Delaroudis et al.,

2008). In agreement, bGH mice had decreased apoA1 relative to WT controls (chapter

IV). It seems that excess GH leads to unfavorable HDL and apoA1 levels, whereas lack of GH action results in normal apoA1 and normal or reduced HDL levels.

RBP-4 showed an isoform-specific change in gender, genotype differences, as well as an interaction between gender and genotype

Seven isoforms of RBP-4 were observed (Figure 15 in chapter III) and discussed in chapter IV. Three isoforms of RBP-4, spots 12, 13 and 14, increased while one isoform, 181 spot 15, decreased in GHR-/- mice. Interestingly, spots 12, 13 and 14 were the larger isoforms (22kDa) whereas spots 15 and 16 were the shorter isoforms (19kDa). Thus,

22kDa forms increased in GHR-/- and the 19kDa forms decreased in GHR-/-. This is consistent with previous chapter IV in which bGH mice had increased 19kDa and decreased 22kDa RBP-4 isoforms. Thus, the PTM event that regulates this process may be subjected to GH regulation. Regarding gender differences, spot 16 (19kDa) increased while spot 14 (22kDa) decreased in females. Spot 14 also showed a decrease by age in females, but not in males. On close examination, spots 12-14, the 22kDa isoforms, increased in GHR-/- versus WT mice and was more pronounced in females than males

(Figure 61). Similarly, for spots 15 and 16, the 19kDa isoforms, decreased in GHR-/- versus WT and was greater in females than males. Since males have a pulsatile GH secretion pattern, whereas females have a more constant pattern, this suggests that the female GH pattern is more potent in increasing the 19kDa and decreasing the 22kDa

RBP-4 (or the male pattern reverses it). When GH is absent such as in GHR-/- mice, the difference between male and female is diminished in these RBP-4 isoforms (Figure 61).

Based on these data, it is possible that the relative proportion of 22 and 19kDa RBP-4 isoforms is regulated by GH in terms of both the amount and secretion pattern in circulation. In this regard, bGH mice have constitutively elevated GH levels, therefore, pronounced effect on this regulation (increasing the 19kDa and decreasing the 22kDa isoforms, see chapter IV for data). Thus, it is important to differentiate between RBP-4 isoforms as a predictor of GH effects and the effects on aging.

182

Figure 61. RBP-4 showed isoform-specific interaction between GH and gender. Spot 12: p=0.021 for genotype, p=0.04 for gender, p=0.127 for interaction; spot 13: p=0.001 for genotype, p=0.013 for gender, p=0.026 for interaction; spot 14: p<0.001 for genotype, p=0.004 for gender, p=0.099 for interaction; spot 15: p=0.107 for genotype, p=0.078 for gender, p=0.764 for interaction; spot 16: p=0.736 for genotype, p<0.001 for gender, p=0.122 for interaction. Data are pooled from all three ages and subjected to 2-way ANOVA.

RBP-4 is secreted by liver and fat, and more so in visceral than subcutaneous adipose depot (Kloting et al., 2007). It is associated with non-alcoholic fatty liver (Seo et al., 2008), insulin resistance (Reinehr et al., 2008) and type 2 diabetes (Chang et al.,

2008; K. Takebayashi et al., 2007b). Total RBP-4 levels tend to be lower in women than men (Ichihara et al., 2004). Despite the obesity seen in GHR-/- mice, they did not have increased total RBP-4 compared to WT mice. This may be because of the contribution of liver-derived RBP-4, or that the excess fat of GHR-/- mice is subcutaneous rather than visceral. 183

One isoform of TTR showed interaction between gender and aging

Nine TTR spots were identified (Figure 15 in chapter III). TTR is in theory 13.6 kDa (without the signal peptide) with a pI of 5.77. Possible PTMs of TTR are known to exist in human plasma including phosphorylation, cysteinylation, dihydroxylation and glutathionylation (Hutchinson et al., 1996; Terazaki et al., 1998; Zheng Wang et al.,

2004). The details of each isoform were discussed in chapter III. This paper will focus only on the isoform that changed in this study (spot 17, ~21 kDa).

This TTR isoform (spot #17) showed an increase with age in females more so than males. Interestingly, this was also the TTR isoform that increased during aging in male

WT mice (See chapter III). This seems to be paradoxical to the notion that females live longer. One possibility is that this TTR is actually protective and induced at older ages to counteract some of the damage imposed by aging.

Hp decreased in females versus males and showed an interaction between genotype and gender

Of the four Hp isoforms identified (Figure 15 in chapter III; also see discussions in chapters III and IV), one (spot 22) decreased in females, two (spots 23 and 24) showed interaction between genotype and gender such that they decreased markedly in females in

WT, but did not change as much in GHR-/- mice.

Hp is an acute phase protein indicating inflammation. Hp binds to heme to reduce oxidative stress. Thus, reduced levels of Hp may indicate reduced need of Hp to bind heme. Incidentally, females did show reduced hemoglobin beta levels in plasma (see below). Therefore it is possible that female mice had reduced levels of oxidative stress 184 related to heme release. Also, in WT mice, this gender difference was obvious, however, in GHR-/- mice, male and female mice both had low levels of Hp (compared to WT male) and did not differ significantly. In was shown in chapter IV that bGH mice had increased Hp relative to WT controls. Together, these data implied that low levels of Hp indicated reduced inflammation in GHR-/- (both male and female) versus WT controls; and in WT mice, lower Hp levels indicate reduced inflammation in females. This could be one of the reasons that GHR-/- mice are long-lived, bGH mice short-lived and that females are usually longer-lived than males.

Hbb decreased in females

There were 4 spots identified as Hbb (Figure 15 in chapter III; also see appendix A for MS/MS matching). Hbb is in theory 15.7 kDa with a pI of 7.26. The identified four isoforms were ~12 kDa with pI ranging from 7.24 to 7.96. Therefore isoform 1 (Figure

15) is likely the non-modified Hbb. Hbb is known to be phosphorylated, although this is not likely the PTM for other three isoforms which have more basic pIs. Also, the PTM responsible for the reduced Mw is not clear. Two of them, spots 26 and 27 (isoforms 1 and 3 in Figure 15), decreased in females (Table 8).

Hbb is a subunit of hemoglobin, the major protein in red blood cells. The presence of Hbb in the plasma may be due to blood cell lysis. It is not clear why female mice have lower Hbb. However, this may indicate that females have less oxidative stress due to heme release.

185

Alb decreased in females versus males and showed an interaction between gender and aging

As discussed in chapter III, the Alb isoforms seen on our 2D gels are most likely

Alb fragments, and the PTMs for these fragments resulting in various Mw and pI are not clear. One isoform (spot 25, 12kDa) decreased in females. Two isoforms (spots 28 and

29; 9kDa) increased by age in females but not males, whereas four isoforms (spots 18-21;

16kDa) decreased by age in females but not males. It is difficult to explain the complex behavior of these albumin fragments. One possibility is that one or more theoretical proteases cleave albumin in plasma and act differentially between male and female mice.

It is possible that this activity changes during aging between genders as well. Further studies are needed to confirm the albumin fragmentation pattern in mice during aging and its physiological significance.

186

Conclusion

This study has found that GHR-/- mice have lower glucose (in adult but not old mice), lower insulin, reduced body weights and lean mass, but increased fat mass. WT mice showed a gender difference in lean mass (reduced in females); however, this gender difference was diminished in GHR-/- mice, suggesting that GH is important in maintaining gender-specific body composition.

Biomarkers of lack of GH action as well as gender are summarized in Table 10.

Table 10. Biomarkers indicative of lack of GH action and gender difference Biomarkers indicating lack of GH Gender differences Increase in GHR-/- Decrease in GHR-/- Increase in females Decrease in females apoA4 (spots 2 and 3; apoE (spots 6 and 7; apoE (spot 4) Clusterin (spot 5) total) total) RBP-4 (spots 12-14, RBP-4 (spot 15, RBP-4 (19kDa RBP-4 (spot 14, 21kDa isoforms) 19kDa isoform) isoform spot 16 and 21kDa isoform) total) Alb (spot 22) apoA1 (total) Hbb (spots 26 and 27) Hp (spot 25) Spot # corresponds to Figure 65.

One important finding is that GHR-/- mice have elevated apoA4 levels compared to

WT mice, indicating an enhanced cardiovascular protective effect of apoA4 in this long-

lived mouse. Interestingly, although female mice (both genotypes) do not have higher

levels of apoA4, they do show an increased apoA4 toward old age versus no change

during aging in male mice; further, females also have increased apoA1 during aging in

contrast to no change in male aging. This is possibly one of the reasons females have

reduced cardiovascular mortality and live longer than males. 187

In contrast to bGH mice, which have increased cholesterol and apoE, GHR-/- mice have decreased cholesterol and apoE. Thus apoE appears to be a sensitive marker for GH action involving lipid transportation and metabolism.

Expectedly, gender differences exist in plasma proteins. Total RBP-4 increased in females by 20%, and total apoA1 decreased by 15%. However, although a 19kDa isoforms of RBP-4 increased, another 22kDa isoform decreased in female versus male mice, similar to isoform changes in bGH compared to WT mice. This suggests a role of

GH secretion pattern on PTM reglation of RBP-4. Total levels of apoE, clusterin and Hp do not change between genders, but specific isoforms show up or down-regulation in female versus male mice (Table 10). Further, females show a different aging profile on certain proteins from males. For example, apoA1 and apoA4 increase during female aging but not male aging, possibly indicating better cardiovascular protection as mice age in females. Two albumin fragments increase, whereas five albumin fragments decrease during female aging compared to little age-related changes in males. All these gender differences are independent of genotype, i.e., GHR-/- and WT mice show similar gender- specific differences.

Interestingly, some of the gender differences observed were genotype-dependent

(Hp spots 23 and 24, apoE spot 7 and RBP-4 spot 12), implying an interaction of GH with gender. GH itself is secreted in a sexually dimorphic manner (MacLeod et al.,

1991). GH also affects liver sexually dimorphically (Norstedt & Palmiter, 1984; D. J.

Waxman, 2000; David J. Waxman & O'Connor, 2006; Wiwi & Waxman, 2004). Many of the plasma proteins are secreted by liver, and liver is known to express genes in a 188 sexually dimorphic manner. For example, drug metabolizing genes, including cytochromes P450 (CYPs), sulfotransferases, glutathione-S-transferases and UDP- glucuronosyltransferases (David J. Waxman & Holloway, 2009). More than 1000 genes have been identified in mouse and rat livers with sexually dimorphic expressions, substantially affecting liver metabolic function and pathophysiology (David J. Waxman

& Holloway, 2009). The sexual dimorphism of liver gene expression is regulated by the

GH secretion pattern, with pulsatile in males vs. more continuous in females. Liver

STAT5b is phosphorylated by GH in male but not female rats (D. J. Waxman, 2000).

Pulsatile GH causes phosphorylation of STAT5b but continuous GH actually downregulates STAT5b signaling in rat liver (D. J. Waxman, 2000). Hepatic nuclear factor (HNF)-4α also responds to GH in a sexually dimorphic manner, thereby possibly mediating the sex-dependent effects of GH on the liver (Wiwi & Waxman, 2004). These studies demonstrate that sexually dimorphic pattern of GH is responsible for the sexual dimorphism seen in liver gene expressions. It will be important to establish if the plasma proteins (i.g., Hbb) observed in this study are regulated by GH, more specifically, through STAT5b and/or HNF-4α. Also, in the case of isoform-specific changes, how

PTMs of the plasma proteins (i.g., RBP-4) are regulated by GH signaling. Thus future studies are needed to reveal the functional inference of these proteins or specific isoforms related to GH and gender.

189

CHAPTER VI. BIOMARKERS OF GH AND IGF-1 DOPING: PROTEOMIC

ANALYSIS OF PLASMA FROM WILD TYPE MICE INJECTED WITH GH AND

IGF-1

Abstract

Recombinant human growth hormone (rhGH) has been and continues to be abused by athletes and is an important issue on the agenda of World Anti-Doping Agency

(WADA). Due to the short half life, pulsatile secretion pattern, and sequence identity to endogenous hGH, rhGH is very difficult to detect. Efforts have been made to discover and validate biomarkers of hGH that have relative long half lives. A combination, rather than a single biomarker, may provide an accurate and sensitive assay to detect hGH doping.

Moreover, insulin-like growth factor-1 (IGF-1) is also abused by athletes and there is currently no method to detect IGF-1 doping. In order to complement the studies ongoing in the rhGH doping area, we set out to establish GH and IGF-1 doping biomarkers in mice. Thus, mice were injected subcutaneously with recombinant bovine

(b) GH or rhIGF-1 for two weeks followed by one additional week of a drug-free washout period. It is important to be able to detect GH or IGF-1 doping after a period of drug free washout because athletes normally will stop using drugs when the competition is up-coming. Plasma proteins were collected after the drug treatment or washout period and resolved by two-dimensional electrophoresis (2-DE). Proteins that showed significant difference with GH or IGF-1 treatment were identified by mass spectrometry (MS) and

MS/MS. We have found several markers of GH action in the plasma of GH treated mice: 190 namely three isoforms of haptoglobin (Hp), one isoform of retinol-binding protein-4

(RBP-4), six isoforms of albumin, one isoform of α-2 macroglobulin (a2m) and one isoform of serum amyloid A-1 (SAA-1), all of which increased in the group receiving

GH treatment with a one week washout period. Also, one isoform of apolipoprotein A-4

(apoA4) was found to be significantly decreased following GH treatment and in the one week washout period. Additionally, ApoE and clusterin showed significant changes after

GH treatment but returned to control levels after 7 days of washout; thus, their response to GH was not long lasting and did not warrant further study. In terms of biomarkers for

IGF-1, two isoforms of Hp and TTR, as well as six isoforms of albumin, increased after

IGF-1 treatment and in the washout period. Important to the study of attempting to separate GH versus IGF-1 action, we found several plasma proteins to be differentially regulated by GH and IGF-1, including apoE, which was increased by GH but not affected by IGF-1, and one isoform of TTR and four isoforms of albumin which were increased by IGF-1 but not affected by GH. If these novel biomarkers for GH and IGF-1 are verified in humans, then perhaps novel human markers of GH vs IGF-1 action will be established and commercial kits to detect GH or IGF-1 abuse will be generated.

191

Introduction

Growth hormone (GH) is secreted by somatotrophs of the anterior pituitary gland and has a central role in stimulating longitudinal growth and regulating metabolism of protein, carbohydrates, lipids and mineral (Baumann; Kopchick & Andry, 2000). GH promotes bone, cartilage and muscle growth as well as fat degradation; it also induces insulin resistance and hyperglycemia (LeRoith & Yakar, 2007). Recombinant human GH

(rhGH) is abused by athletes to increase muscle, degrade fat and to enhance performance.

However, the positive role of short-term GH administration in muscle mass, strength and endurance as well as athletic performance have been brought into question as positive performance outcomes of GH administration have not been reported (Berggren et al.,

2005; Ehrnborg et al., 2005). Still, rhGH abuse is rampant among athletes and is on the prohibited substance list by World Anti-Doping Agency (WADA) (WADA, 2009).

Injected rhGH is difficult to detect in blood for three reasons. First, rhGH is indistinguishable from endogenous GH. Second, GH has a relatively short half-life of ~

15 minutes (R. W. Holl et al., 1993). Exogenous GH is cleared quickly from circulation and injected rhGH is also quickly cleared from the blood (Hashimoto et al., 2000; Owens et al., 1973; Jennifer D. Wallace et al., 2001). Third, GH secretion from pituitary somatotrophs is pulsatile, especially in males (Goldenberg & Barkan, 2007), thus making it hard to determine the exact time for GH measurement. Therefore, it is difficult to directly detect GH levels in the blood.

GH induces the production of insulin-like growth factor-1 (IGF-1) in the liver as well as many GH responsive organs. IGF-1 mediates many of the actions of GH, e.g., 192 muscle hypertrophy, bone growth and fluid retention. IGF-1 has been used traditionally as a marker for GH, being positively correlated with the level of GH. However, in some cases, plasma IGF-1 levels cannot properly reflect blood GH levels. For example, some acromegalic patients have high levels of IGF-1, but GH levels are normal. Thus, the correlation of GH and IGF-1 is poor. Also, after treatment to normalize GH, IGF-1 does not always normalize (Barkan, 2007; Clemmons, 2006; Popovic, 2005; Stonesifer et al.,

1981). Thus, other plasma biomarkers for GH action would be valuable.

In addition, because of the anabolic effect of IGF-1, it is being abused in the athletic world as well (Guha et al., 2009; R I G Holt, 2008). RhIGF-1 is FDA-approved for the treatment of children with short stature caused by primary IGF deficiency. Currently, there is no technology to detect IGF-1 abuse. Therefore, it is important to find biomarkers for IGF-1.

Because GH induces IGF-1, it is difficult to dissect out the specific effects of GH or

IGF-1. This has been an enigma in the GH/IGF-1 field, i.e. separating the molecular action of GH and IGF-1.

The mouse is a good model to study human biology because it shares high homology with humans at both the genomic and proteomic levels. In this study, three- month-old normal male mice were injected daily with recombinant bovine growth hormone (bGH) or recombinant human (h) IGF-1 for 14 days. Two additional groups of mice were given a 7-day washout period following these injections. Then, the immediate effect as well as a washout effect was observed by analyzing the mouse plasma proteome via two-dimensional electrophoresis (2-DE). Plasma protein levels that changed 193 significantly by GH or IGF-1 in their washout period were identified by mass spectrometry (MS) and MS/MS. Several plasma biomarkers were identified. Hopefully, they will have the potential to indicate GH and IGF-1 doping in humans as well.

194

Materials and Methods

Experimental animals and treatment protocol

Three-month old male C57 BL/6J mice were used for these studies. These represent

relatively young adult animals. Mice were housed 2-3 per cage at room temperature

(22 oC) in a 12-hour light, 12-hour dark cycle. Mice were fed ad lib with a standard

chow diet.

Bovine (b) GH and rhIGF-1 were administered to the mice via subcutaneous

injection daily for a total of 21 days. Purified recombinant bGH was a gift from

Monsanto (St Louis, MO) and was dissolved in standard phosphate buffered saline

(PBS). RhIGF-1 was a gift from Tercica Inc. (Brisbane, CA) and dissolved in a solution

called ‘vehicle’ containing 9 mg/mL benzyl alcohol, 5.84 mg/mL sodium chloride, 2

mg/mL polysorbate 20, and 0.05M acetate at a pH of approximately 5.4. Mice were

divided randomly into 5 groups (n=7 for each group). bGH was injected once per day at a

dose of 5ug/gram of body weight as previously described (List et al., 2009). IGF-1 was

administered twice per day at a dose of 5ug/gram body weight for a total of 10ug/gram

body weight per day to prevent hypoglycemic shock for the animals. To control for the

stress levels as well as the different solultions used for bGH and rhIGF-1 preparation,

each mouse was injected three times daily (one for PBS or bGH, two for vehicle or IGF-

1) as summarized in Table 11 and Figure 62. All animals were sacrificed at the end of the

study. Animal protocols were approved by Ohio University’s Institutional Animal Care

and Use Committee.

195

Table 11. GH and IGF-1 treatment scheme Group Week 1 Week 2 Week 3 1. Control PBS+vehicle(x2) PBS+vehicle(x2) PBS+vehicle(x2) 2. GH PBS+vehicle(x2) GH+vehicle(x2) GH+vehicle(x2) 3. GH washout GH+vehicle(x2) GH+vehicle(x2) PBS+vehicle(x2) 4. IGF-1 PBS+vehicle(x2) PBS+IGF-1(x2) PBS+IGF-1(x2) 5. IGF-1 washout PBS+IGF-1(x2) PBS+IGF-1(x2) PBS+vehicle(x2)

Figure 62. GH and IGF-1 treatment scheme.

Body composition measurement

Body composition was measured using mq 7.5 NMR analyzer the Bruker Minispec

(The Woodlands, Tx) as described previously (List et al., 2009; Palmer et al., 2009)

before and after the 21-day injection. Each animal was measured twice and the mean

value was determined. The data are given as change in grams (g) for lean mass, fat mass

and fluid mass. 196

Fasting blood glucose

Mice were fasted for 12 hours overnight and blood was collected with heparinized

capillary tubes by tail tip clipping. Whole blood was centrifuged at 7000 x g for 10

minutes at 4 ºC and the resulting plasma was stored at -80 ºC. Fasting glucose was

measured with the first drop of blood from tail tip at 9AM after 12h over night fasting

using ONE TOUCH glucometer from Lifescan (Milpitis, CA).

2-DE and quantification of proteins

See chapter III.

Statistical analysis

All statistical analyses were performed using SPSS 14.0 software (Chicago, IL).

Data for body composition and fasting glucose were subjected to one-way ANOVA test

using LSD test for Post Hoc comparison (p<0.05 considered significant). Log-

transformed data for proteomics were subjected to one-way ANOVA with p<0.01 considered significant, then Post Hoc comparison using LSD method was applied to detect between group difference (p<0.05). All data were presented as mean ± SEM.

Protein identification by mass spectrometry (MS), MS/MS, and liquid chromatography

(LC)-MS/MS

See chapter III. 197

Results

Body composition

After 3 weeks of treatment, the 5 groups of mice had significantly different weight

changes (p=0.009). All GH/IGF-1 treated mice had increased weight gains compared to

controls (Figure 63). The GH treated group gained the most weight, followed by IGF-1

group. Interestingly, whereas the IGF-1 washout group had slightly decreased weight

gains relative to the IGF-1 group, the GH washout group had marked decreased weights compared to the GH group. However, the weight gains were still higher than controls, which suffered from weight loss, presumably due to stress from daily injections.

Figure 63. Body weight and body composition change after the injections. Different letters denote significant differences (p<0.05). GHw: GH washout; IGFw: IGF-1 washout. 198

Lean mass gain was significantly different among the 5 groups (p=0.007). Other than controls and the GH washout group, all treated mice had a positive lean mass gain

(Figure 63). The GH group had the largest lean mass gain, although the effect was diminished after 1 week of washout; whereas both IGF-1 and IGF-1 washout groups showed positive lean mass gain. Fat gain was not significantly different among 5 groups

(p=0.13), although a trend in increased fat loss was observed in GH group (Figure 63).

Fluid gain was significantly different among 5 groups (p=0.016). Both GH and IGF-1 increased fluid gain compared with controls (Figure 63), whereas this fluid retention effect was diminished in the washout groups.

Taken together, the weight loss of control mice was attributed to lean and fluid mass loss, not fat loss. The weight gain of GH and IGF-1 groups was due to lean mass and fluid gain; and the weight gain of IGF-1 washout group was primarily due to lean mass gain. In summary, GH treatment increased lean and fluid mass, decreased fat mass; however, the effect was temporary. In contrast, IGF-1 increased lean and fluid mass, did not decrease fat mass, and maintained lean mass after a 7-day washout period.

Fasting blood glucose

Fasting blood glucose was not significantly different among the 5 groups (p= 0.195,

Figure 64).

199

Figure 64. 12-hour fasting glucose levels. No significant difference was found among five groups.

Plasma proteins altered by GH/IGF-1

A total of 165 protein spots were detected by 2-DE. Among them, a subset of spots

showed significantly different intensities among the 5 groups of mice. Some had

markedly altered levels by GH and/or in the GH washout period, which could serve as

potential biomarkers for GH doping. Some of them were altered significantly by IGF-1

and/or the IGF-1washout period, and therefore are potential biomarkers for IGF-1

doping. Very importantly, some proteins were differentially altered by GH and IGF-1,

illustrating specific effects of GH and IGF-1 on plasma proteins. Figure 65 is an annotated 2D gel image with the selected spots of interest. Notice that multiple spots were identified as the same protein probably due to post-translational modifications

(PTMs) of the identified protein. The nature of these PTMs has not been identified.

200

Figure 65. 2-D gel map of proteins that changed significantly during GH or IGF-1 treatment or after the respective washout period. Mw: molecular weight; pI: isoelectric focusing point; apoA4: apolipoprotein A-4; a2m: alpha-2 macroglobulin (35kDa subunit); Clu: clusterin; apoE: apolipoprotein E; RBP-4: retinol-binding protein-4; TTR: transthyretin; Hp: haptoglobin (alpha-chain); SAA-1: serum amyloid protein A-1.

Plasma proteins altered by GH

The plasma levels of several proteins were altered by GH treatment and returned to

control levels after the washout period. These included apoE, which increased, and

clusterin, which decreased (Figure 66).

201

Figure 66. Proteins that responded to GH and returned to control levels after the GH washout period. Significant difference was found among three groups via one-way ANOVA (p<0.01). Different letters denote significant differences (p<0.05). A, apoE isoform (2) and total apoE (defined as sum of all four isoforms) increased in GH group but remained unchanged in GH washout group. B, four isoforms as well as total clusterin decreased in GH but remained unchanged in GH washout group. Protein names and isoform # corresponds to Figure 65. GHw: GH washout.

ApoE, in particular isoform 2, increased after GH injection; however, the levels were the same as controls after the 1 week of washout period (Figure 67). ApoE was not altered by IGF-1 injections. 202

Figure 67. 3-D view of apoE (isoforms 1-4) and a2m. Spot intensity was converted into a topographic view of peaks using PDQuest.

Clusterin (isoforms 3-6) decreased immediately after GH treatment, but rose to control levels after the washout period (Figure 68).

Figure 68. 3-D view of clusterin isoforms.

Several plasma protein levels were changed following GH treatment but were found altered only after the washout period rather than immediately after GH treatment. These 203 included one isoform of TTR and apoA4 that decreased (Figure 69); Hp (Figure 71),

RBP-4 (Figure 71), a2m (Figure 67) and SAA-1 (Figure 72) that increased (Figure 70), as well as multiple isoforms of albumin that increased (Figure 73).

TTR isoform 2 decreased after 1 week of the GH washout period. ApoA4 isoform 5 decreased after 1 week of GH washout. However, other isoforms of TTR or apoA4 were not changed.

Figure 69. Proteins that decreased after the GH washout period. Significant difference was found among three groups via one-way ANOVA (p<0.01). Different letters denote significant differences (p<0.05). Protein names and isoform # corresponds to Figure 65. 204

A, quantification of TTR spot 2 and apoA4 spot 5. Y axis represents the log-transformed intensity of spots. GHw: GH washout. B, 3-D view of spot intensity.

Figure 70. Proteins that were increased significantly after washout of GH injection (one- way ANOVA, p<0.01). Different letters denote significant differences (p<0.05). Protein name and isoform # corresponds to Figure 65. GHw: GH washout.

All three isoforms of Hp increased after the GH washout period (Figure 71). One isoform of RBP-4 also increased after GH washout (Figure 71). IGF-1 treatment increased Hp, but did not change RBP-4 (Figure 71). 205

Figure 71. 3-D view of intensity of Hp isoforms 1-3 and RBP-4.

SAA-1 was slightly decreased after GH treatment, but increased after the GH washout period (Figure 72).

Figure 72. 3-D view of intensity of SAA-1. Note that SAA-1 increased in the GH washout group.

Alb isoforms including spots 1-6 all increased after the GH washout period (Figure

73, Figure 75 and Figure 76). 206

Figure 73. Multiple albumin fragments were increased after washout of GH injection. Significant difference was found among three groups via one-way ANOVA (p<0.01). Different letters denote significant differences (p<0.05). Isoform # corresponds to Figure 65. GHw: GH washout.

Figure 74. 3-D view of intensity of Alb (spots 1 and 2) and TTR (spot 1). Isoform # corresponds to Figure 65.

207

Figure 75. 3-D view of intensity of Alb spots 3 and 4. Isoform # corresponds to Figure 65.

208

Figure 76. 3-D view of intensity of Alb (spots 5 and 6). Isoform # corresponds to Figure 65.

Plasma proteins altered by IGF-1

Several proteins increased by IGF-1 injection and remained at elevated levels after

the washout period, including Hp, TTR (Figure 77), and multiple forms of Alb (Figure

79).

209

Figure 77. Proteins that were increased by IGF-1. Significant difference was found among three groups via one-way ANOVA (p<0.01). Different letters denote significant differences (p<0.05). Isoform # of Hp and TTR corresponds to Figure 65. IGFw: IGF-1 washout.

Hp isoforms 1 and 3 increased in the IGF-1 treated group and increased further after the washout period (Figure 71).

TTR isoforms 1 and 3 both increased after IGF-1 injection and isoform 1 increased further after the washout period (Figure 74). Isoform 3 remained elevated after the washout period (Figure 78).

210

Figure 78. 3-D view of intensity of TTR (spot 3). Isoform number corresponds to that in Figure 65.

Six Alb isoforms (1, 2, 4, 5, 7 and 9) all increased after IGF-1 injection and remained at the elevated levels after the washout period (Figure 74, Figure 75, Figure 76 and Figure 79).

211

Figure 79. Alb fragments that were increased by IGF-1. Significant difference was found among three groups via one-way ANOVA (p<0.01). Different letters denote significant differences (p<0.05). Isoform # corresponds to Figure 65. IGFw: IGF-1 washout.

Plasma proteins differentially altered by GH and IGF-1

Several proteins responded to GH and IGF-1 differentially, including apoE, TTR

(Figure 80) as well as individual albumin fragments (Figure 81).

ApoE (isoforms 2 and 3 and total levels) increased by GH but not IGF-1 treatment

(Figure 67). TTR (isoform 3) was not changed by GH but increased by IGF-1 treatment

(Figure 78). Alb 1, 2 and 5 did not change after GH treatment but did increase after IGF-

1 injections (Figure 74, Figure 75 and Figure 76). 212

Figure 80. Proteins that were changed by GH and IGF-1 differentially. Significant difference was found among three groups via one-way ANOVA (p<0.01). Different letters denote significant differences (p<0.05). Isoform # corresponds to Figure 65.

Figure 81. Alb fragments that were changed by GH and IGF-1 differentially. Significant difference was found among three groups via one-way ANOVA (p<0.01). Isoform # corresponds to Figure 65. 213

Discussion

Two weeks of GH or IGF-1 injections did not alter fasting glucose levels significantly. However, body weights and compositions changed. Mice gained lean mass with GH treatment but did not maintain this mass. In contrast, IGF-1 treatment was able to retain the lean mass gain even after the washout period. This is perhaps because the anabolic effect of GH is mediated through IGF-1, and IGF-1 has a much longer serum half-life than GH. Both GH and IGF-1 promoted body weight and lean mass gains, consistent with the well known anabolic effect of the GH/IGF-1 axis (Cheek & Hill,

1974; N. Moller & Jorgensen, 2009; Velloso, 2008).

Increased trend (not significant) in fluid gain in the GH and IGF-1 group indicates the fluid retention effect of GH (Carroll & Van den Berghe, 2001; Clemmons et al., 1987;

O'Sullivan et al., 1994). GH and IGF-1 promote fluid retention by stimulating renin- angiotensin-aldosterone system which is central to fluid homeostasis (J. Moller et al.,

1999).

Fat mass was slightly reduced, although not significantly, by GH. The fat decreasing effect was not seen after 7 days of washout. Apparently, the fat degrading effect was only induced by GH, not IGF-1, as IGF-1 group did not show a decrease in fat.

This effect has been reported previously (Clark et al., 1996; Dubuis et al., 1996; Fielder et al., 1996).

Thus, the weight gain that we detected was due to increased lean and fluid mass and was not due to changes in fat. The GH group had a larger weight gain because of the 214 larger lean mass gain. However, during the GH washout period, a quick rebound of weight was seen when compared to mice in the GF-1 washout group.

In summary, the body composition data confirmed lean mass anabolic and fluid retention actions of GH/IGF-1, which normalized after a one-week washout period. An exception to this statement was the lean mass and body weight gain for IGF-1 washout group. These data were consistent with a previous study with GH injection in mice (List et al., 2009).

The levels of many plasma proteins were altered during GH or IGF-1 treatment.

ApoE (isoform 2 as well as total levels) increased by GH injection, consistent with markedly increased apoE (all isoforms) that was found in bGH mice (chapter IV). ApoE has also been show to increase in individuals with acromegaly (Wildbrett et al., 1997).

However, apoE levels returned to control levels after a one week washout period, indicating that apoE is an acute marker but not a long-lasting marker of GH action.

ApoE is increased in rats with GH administration and this effect is not mediated by IGF-1

(Sjoberg et al., 1994b). GH has been shown to increase apoE by increasing the translation and secretion rate of apoE from hepatocytes (Sjoberg et al., 1994a).

ApoE is secreted primarily by the liver and functions to transport lipids via very- low-density lipoprotein (VLDL), LDL and high-density lipoproteins (HDL). The change of apoE by acute GH treatment therefore indicates an acute effect on lipoprotein metabolism. ApoE also is considered an antioxidant that helps to fight oxidative stress

(Davignon, 2005; Miyata & Smith, 1996; Tarnus et al., 2009), reduces DNA damage 215

(Folkmann et al., 2007) and inflammation (Lynch et al., 2001). Thus, short-term GH treatment may improve resistence to oxidative stress by increasing apoE.

Similarly, clusterin (isoforms 3-6 as well as total levels) decreased by GH injection, also consistent with the lower levels found in bGH mice at young ages (chapter IV). Note that bGH mice expressed high levels of clusterin when they were older, which could indicate the accelerated aging phenotype of these mice (chapter IV). However, the physiology of this age-dependent increase in clusterin is complicated. Since bGH mice of

2 and 4 months, as well as GH-injected mice (at 3 months in this study) had lower levels, clusterin could be considered a negative marker of GH action. Unfortunately, the effect was short-lived as the GH washout group had normal levels of clusterin. Clusterin is also associated with HDL, therefore the reduced clusterin by GH may indicate a lipid profile change by GH.

One isoform of TTR (isoform 2) did not change immediately after GH injection, but decreased after a week of washout, suggesting TTR lowering effect of GH was not acute, but relatively long lasting. Consistently, bGH mice had lower levels of TTR than WT at all ages. Therefore, TTR, in particular, isoform 2, is a long-lasting (at least one week) biomarker for GH. As shown in chapter IV, chronic GH excess leads to reduced levels of eight TTR isoforms including the one that decreased in this study. Perhaps this isoform is the most sensitive to GH action.

One isoform of apoA4 (isoform 5) showed similar behavior, that is, decreasing after the GH washout period but not immediately after injections. This is in contrast with bGH mice, which did not show difference in any of the apoA4 isoforms compared to WT 216 mice. However, GHR-/- mice have increased apoA4 (2 isoforms as well as total).

Therefore, apoA4 could be considered a negative marker of GH action.

Hp increased (isoforms 1-3 and total levels) in the GH washout group although not immediately after GH injections. bGH mice had increased Hp (all isoforms and total) compared to WT mice; therefore, Hp is a long-lasting biomarker of GH. As with bGH mice, this suggests that short-term treatment of GH also induces an inflammatory state in mice.

RBP-4 (one isoform of 19kDa) increased in the GH washout group although not in the GH group. This was consistent with data in chapter IV, in which bGH mice had increased RBP-4 19kDa isoforms and decreased 22kDa isoforms and decreased total

RBP-4. Further, GHR-/- mice have increased 22kDa and decreased 19kDa isoforms.

Thus, chronic GH excess and short-term GH treatment both upregulate the 19kDa isoforms of RBP-4, which can be a biomarker of GH.

A2m (one isoform) increased in GH washout group although not in GH group. This was the same isoform that was found to be increased in bGH mice. Thus, this particular isoform of a2m is a biomarker of GH. A2m is an acute phase protein, although the total a2m did not change, this isoform may play a certain role in inflammation that does not upregulate total a2m.

SAA-1 increased in GH washout group. SAA-1 was found to be decreased during normal aging in male mice (chapter III). Therefore SAA-1 is a biomarker of both aging and GH. If SAA-1 is responsive to GH, then decline over aging of SAA-1 may be due to decline to GH during aging. 217

Multiple Alb fragments (isoforms 1-6) all increased in GH washout group although not in GH group. GH increases Alb mRNA transcription in mice (M. Wang et al., 2008).

These specific albumin isoforms are therefore long-lasting biomarkers of GH.

Together, Hp, SAA-1, specific isoform of RBP-4, a2m and Alb all could be used as markers of GH action after a relatively long washout of one week.

Regarding biomarkers of IGF-1, Hp, TTR and several Alb fragments increased.

Two Hp isoforms (1 and 3), two of TTR isoforms (1 and 3), as well as six Alb isoforms

(1, 2, 4, 5, 7 and 9) increased in IGF-1 group and also remained at elevated levels or even increased further in the IGF-1 washout group. The increase in Hp and Alb was in agreement with the GH groups, although in some cases different isoforms were involved.

However, TTR showed an opposite change compared to GH treatment, where down- regulation was observed.

Finally, comparing the difference between GH and IGF-1 groups, apoE (isoforms 2 and 3 and total) was increased in GH but did not change in IGF-1 group; TTR isoform 3 increased in IGF-1 group but not GH group; and Alb isoforms 1, 2 and 5 increased in

IGF-1 group with no change in GH group. These are biomarkers differentiating between

GH and IGF-1 actions.

In general, plasma protein levels that changed by IGF-1 treatment were seen early and also lasted through the one week washout period; in contrast, in GH treated mice, protein changes usually appreared after the washout period and not acutely after the 2- week injection period, suggesting that these proteins might be regulated indirectly by GH, perhaps by IGF-1. Two exceptions are apoE and clusterin, which are altered by GH and 218 not IGF-1 treatment, and the change diminished during the GH washout period, suggesting that these two proteins are directly regulated by GH. In the case of TTR, apparently the negative effect of GH (as seen in bGH mice) is not immediate, and not likely mediated by IGF-1, since IGF-1 treatment increased TTR.

As mentioned earlier, IGF-1 and P-III-P are currently established as two biomarkers for GH doping. However, these molecules are not detected on our gels. Secreted form of

IGF-1 is 7.7 kDa with pI 8.31 in theory. Since IGF-1 is cleaved into a and b chains like insulin, the two subunits have Mw of 3.1 and 2.5 kDa, respectively. These small sizes are beyond the limit of the gels (>8 kDa), therefore not likely detected. P-III-P is present in plasma at a very low level (~10 ng/ml (Monarca et al., 1985)), therefore not likely to be detected by 2-DE (sample loading of 2-DE is 15 ul plasma per run, resulting in 0.15ng P-

III-P per gel; a spot needs to contain at least 1 ng to be detected on a 2D gel). 219

Conclusion

Two weeks of GH and IGF-1 injections revealed that several plasma proteins were changed by GH or IGF-1, and certain proteins were differentially changed by GH and

IGF-1 (summarized in Table 12). ApoE and Clu, although altered after GH injections, did not show any difference after the washout. In contrast, the levels of other proteins such as

Hp, RBP-4, SAA-1, a2m, Alb and apoA4 were different after the washout period, which represent the relative long-lasting biomarkers of GH injection.

For IGF-1, biomarkers include isoforms 1 and 3 of Hp, isoforms 1 and 3 of TTR, as well as isoforms 1, 2, 4, 5, 7 and 9 of Alb. All these proteins increased after injections and remained at higher levels after the washout period. Therefore, these proteins are biomarkers for IGF-1 doping.

Lastly, several proteins exhibit significant difference after GH and IGF-1 injections, including isoforms 2 and 3 of apoE, isoform 3 of TTR and isoforms 1, 2 and 5 of Alb.

These are useful biomarkers to differentiate between GH and IGF-1. No one has to date presented plasma markers that differentiate GH from IGF-1 action. In fact, very few physiological markers exist. GH inhibits insulin action and IGF-1 has insulin-like activities such as glucose uptake, which GH does not. GH is lipolytic and IGF-1 is not.

Clearly differences exist between GH and IGF-1 actions, and these are the first plasma markers that differentiate GH from IGF-1 action. For example, GH upreguates apoE and

SAA-1 whereas downregulates clusterin and apoA4. All these apolipoproteins are associated with HDL (apoE also with VLDL), which suggests that GH alters cholesterol composition. None of these is altered by IGF-1, suggesting that at least for 2 weeks of 220 administration, IGF-1 does not alter lipid profile as much. This agrees with body compositon data where GH reduces body fat mass whereas IGF-1 does not change fat mass.

Table 12. Biomarkers of GH, IGF-1 and their difference Biomarkers for GH Biomarkers for IGF-1 Differential Changed Changed after Changed after Changed after biomarkers after washout injections washout by GH and injections IGF-1 apoE ↑ Hp ↑ (isoforms Hp ↑ (isoforms Hp ↑ (isoforms apoE (isoform 2 and 1-3 and total) 1 and 3) 1 and 3) (isoforms 2, 3 total) and total), ↑ by GH, ↔ by IGF-1 Clu ↓ RBP-4 ↑ (one TTR ↑ TTR ↑ TTR (isoform (isoforms 3-6 isoform of (isoforms 1 and (isoforms 1 and 3), ↔ by GH and total) 19kDa) 3) 3) and ↑ by IGF- 1 Alb ↑ (isoforms Alb ↑ (isoforms Alb ↑ (isoforms Alb (isoforms 1-6) 1, 2, 4, 5, 7 and 1, 2, 4, 5, 7 and 1, 2 and 5), ↔ 9) 9) by GH and ↑ by IGF-1 A2m ↑ (one isoform) SAA-1 ↑ apoA4 ↓ (isoform 5 and total) Isoform # corresponds to Figure 65.

The above mentioned biomarkers provide a set of biomarkers for GH and IGF-1

doping in mice. Of course, extension of these results to humans is of paramount

importance. A combination of these in conjunction of the already known markers of GH,

such as IGF-1 and P-III-P, could provide more accurate and sensitive tests to detect GH

doping. 221

Many biomakers are consistent between acute GH injection and bGH mice (Table

13). For example, apoE and Hp are increased by GH; whereas TTR and clusterin are decreased. In many cases, more pronounced effect is seen in bGH mice compared to the injection, perhaps due to an effect of GH dose and exposure time. For example, all five isoforms of apoE are increase in bGH mice whereas only one isoform is increased in GH injected mice. However, several proteins show difference by chronic and acute actions of

GH, including albumin, MBP-C, apoA4 and SAA-1, demonstrating a differential effect of chronic versus acute actions of GH.

Table 13. Comparison of chronic and acute actions of GH on plasma proteins Plasma proteins Chronic GH action (bGH mice) Acute GH action (injection)

apoE All five isoforms ↑ Isoform 3 and total ↑ TTR 8 isoforms and total ↓ ↔ for total One soform ↓ after washout Hp All four isoforms ↑ Isoforms 2-4 and total ↑ Clusterin Isforms 1,2 and total↓ (2, 4 mo) Isoforms 3-6 and total ↓ RBP-4 Total ↓ Total ↔ 19kDa isoforms ↑ One 19kDa isoform ↑ after 22kDa isoforms ↓ washout A2m Total ↔ Total ↔ Isoforms 2 and 3 ↑ Isoform 2↑ after washout Isoforms 6 and 7 ↓ apoA1 Total and 3 isoforms ↓ Tend to ↑ after injection and ↓ after washout Albumin ↔ Several isoforms ↑ MBP-C ↑ ↔ apoA4 ↔ Total ↔ One isoform ↓ SAA-1 ↔ ↑ 222

CHAPTER VII. SUMMARY AND DISCUSSION OF PLASMA BIOMARKERS OF

GH AND AGING

GH is a hormone with well-characterized physiological effects including promoting bone growth, protein anabolism, fat degradation, glucose output from the liver via gluconeogenesis and fluid retention (Carroll & Van den Berghe, 2001; Cheek & Hill,

1974; Clemmons et al., 1987; Fryburg & Barrett, 1993; Katznelson, 2008; N. Moller &

Jorgensen, 2009; O'Sullivan et al., 1994; Velloso, 2008; Widdowson et al.; Woodhouse et al., 2006). Elevated levels of GH is the causative factor in the pathology found in individuals with acromegaly, while lower levels of GH and GH insensitivity are associated with dwarfism displayed in GH deficient and Laron Syndrome individuals, respectively. The FDA has approved recombinant (r) hGH to treat children with the following conditions: Prader-Willi Syndrome, children born Small for Gestational Age,

Turner’s Syndrome, Noonan’s Syndrome, chronic renal failure, idiopathic short stature and cystic fibrosis (Krysiak et al., 2007). In adults, rhGH has been approved to treat adult

GHD (Cummings & Merriam, 1999), wasting syndrome of AIDS (Mulligan et al., 1993;

Schambelan et al., 1996) and short bowl syndrome (Krysiak et al., 2007). Unfortunately,

unethical doctors can prescribe rhGH for uses other than those for which it has been

approved. This has led to misuse and abuse of rhGH. It has been and is abused by

athletes who think that it will enhance muscle mass and physical performance. However,

data suggest that GH administration has no positive role in enhancement of athletic

performance (Berggren et al., 2005; Ehrnborg et al., 2005). The current biomarkers for

detecting GH action for approved indications, or in a doping scenario, are note adequate. 223

Therefore, new biomarkers for GH action will help unravel novel target proteins regulated by GH and provide candidate markers for GH doping detection.

The GH/IGF-1 axis is involved in aging, as attenuated IGF-1 signaling over the long term leads to increased lifespan (Bonafè & Olivieri, 2009; Brown-Borg, 2009;

Holzenberger et al., 2003; Tatar et al., 2003). For example, GH transgenic mice are insulin resistant and short-lived whereas GHR-/- mice are insulin sensitive and long-lived compared to WT mice. Thus, bGH mice and GHR-/- mice are two good mouse models used to study biomarkers of aging. Longitudinal analyses of plasma proteomics of these animals and their littermate controls have revealed biomarkers of chronic GH action or chronic absence of GH action as well as aging.

In this dissertation, 86 proteins have been analyzed and identified in mouse plasma using 2-DE (Figure 15). The SWISS-2DPAGE database includes 2D annotation of human plasma proteins, but not those of mice. However, several reports have been published concerning 2D PAGE reference maps for mouse plasma or serum proteins

(Bijon & Jürgen, 2007, 2009; Duan et al., 2004; Gazzana & Borlak, 2008; Guipaud et al.,

2007; Ren et al., 2007; Roberto et al., 2008; Skehel et al., 2000). Most of the results in this dissertation are consistent or comparable with the literature. However, a few unique findings that have not been published elsewhere are presented in this dissertation and are summarized below.

1) TTR has been identified as an ~14kDa protein as one or a few (depending on the study) isoforms (Bijon & Jürgen, 2009; Duan et al., 2004; Gazzana & Borlak, 2008; 224

Skehel et al., 2000). This study has identified two TTR isoforms at ~38kDa (isoforms 1 and 2 in Figure 15A) and as many as six isoforms in the 14kDa region (isoforms 4-9).

2) ApoE has been reported as at 34kDa protein (Bijon & Jürgen, 2009; Duan et al.,

2004; Gazzana & Borlak, 2008; Skehel et al., 2000). In this dissertation, an additional isoform at 38kDa was found.

3) RBP-4 has been reported as one or two isoforms (depending on the study) at

22kDa (Bijon & Jürgen, 2009; Duan et al., 2004; Gazzana & Borlak, 2008). In this study has identified seven isoforms of RBP-4, three of ~22kDa and four of ~19kDa (isoforms

4-7) have been identified.

4) Alb fragments in the lower Mw region (<40kDa) have been reported as 1-6 isoforms (depending on the study) (Bijon & Jürgen, 2009; Duan et al., 2004; Gazzana &

Borlak, 2008; Roberto et al., 2008). In this dissertation 23 alb isoforms (Figure 15, isoforms 1-23) have been identified, many of which have never been reported. Thus, certain isoforms of the above proteins are reported for the first time in this dissertation.

Results presented in this dissertation also have shown that several plasma proteins change significantly during aging, between bGH and WT littermates, between GHR-/- and WT littermates, between male and female mice, as well as in WT groups treated with

GH or IGF-1. Discussed below are these potential biomarkers for GH action using mice injected with GH or IGF-1 as well as markers of normal aging. These proteins are shown in the 2D images in Figure 15 of chapter III. The changes in the isoforms of these proteins are summarized in Table 14.

225

Table 14. Summary of protein changes during aging, and via GH action or due to gender differences Change Change in bGH vs. Change in Change in GH Change in Change in female vs. male during WT GHR-/- vs. injection IGF-1 WT aging WT injection apoE 1-5↑, total↑ 1,3,4 and 3 and total↑ but ↔ 1↑; 4↓ in female GHR-/- total↓ ↔washout (similar between WT genders) MBP-C 1↑ 1↓(p<0.05) A2m 1,2↑; 5,6↓ 2↑ ↔ Hp 2,3↑ 1-4 and total↑; 1-3↓(p<0.05) 2-4 and total↑ 2,4↑ 2↓; 3,4↓ in female WT 1-4 and total↑ in washout (similar between GHR-/- bGH aging genders) Clu 1,2 and total↑ in 3-6 and total↓ but ↔ 3↓ bGH aging ↔washout RBP-4 2↓, 4↑, total↓ 1-3↑, 5↓, 6↑ washout ↔ 7↑, 3↓; 3↓ in female aging; total↑ 1↓ in female WT (similar between GHR-/- genders) TTR 3↑ 1, 3-9 and total↓ 6↓ washout 3,4↑ Ig kappa 1-3↑ 1↑(p<0.05) 3↓(p<0.05) apoA1 4-6↓ 2↓ 2-5↑ in female aging apoA4 3,4 and total↑ 5↓ washout ↔ 2,3↑ in female aging Albumin 6-9,18↓ 14-17,18,20↑ 14, 15, 19↓; 21,23↑ in female washout; 17,18,21,23↑ aging; 12,14-16↓ in female ↔ right after aging injection SAA-1 ↓ ↑ washout ↔ Prx-2 ↓ Hbb 1,3↓ Numbers represent isoform# of a given protein as shown in Figure 15 of chapter III. ‘Total’ is the sum of intensity of all isoforms. ↑ increase, ↓ decrease, ↔ unchanged. All changes are at a significance level of p<0.01 unless otherwise indicated (p<0.05 means 0.01

ApoE is resolved as five isoforms (Figure 15). The theoretical Mw of apoE is ~34 kDa with a pI 5.46. Isoform 1 has an Mw of 38kDa, larger than the other isoforms, which are 34kDa. The identified apoE isoforms have pIs ranging from 5.76 to 6.28. The levels of all of these isoforms are increased in bGH versus WT mice from ages 2 to 19 months

(Figure 34, Figure 35). Isoforms 1, 3, 4 and total apoE decreased in GHR-/- versus WT mice (Figure 52, Table 9). Isoform 3 and total apoE increased after 2 weeks of GH injection in WT mice (Figure 66). In particular, isoform 3, the most abundant apoE spot, consistently shows a significant (p<0.01) increase in bGH mice and in GH-injected mice, and decreases in GHR-/- mice. Together, these data indicate that apoE is a biomarker of

GH and is positively regulated (upregulated) by GH.

The upregulation of apoE by GH has been reported in humans and rats as well

(Frick et al., 2002; Johannsson et al., 1995; Jan Oscarsson et al., 1991b; Sjoberg et al.,

1994a; Sjoberg et al., 1994b; Wildbrett et al., 1997) and this effect is considered to be independent of IGF-1 (Sjoberg et al., 1994b). Consistently, in our study, IGF-1 injection did not change the levels of any apoE isoforms. Unfortunately, upon GH removal, apoE returns to normal level after 7 days as shown in GH injection washout group. Therefore, in terms of window of opportunity for GH detection, apoE would not be an ideal candidate for the detection of GH doping.

ApoE also is regulated by gender. Isoform 1, the longer form, is upregulated in female mice (Table 8). Isoform 4 is decreased only in female WT mice compared to male

WT and GHR-/- mice of both genders (Figure 59), suggesting an interaction of GH and gender on the level of this apoE isoform. Interestingly, the gender-specific effect on apoE 227 has been observed in rats such that the female pattern of GH secretion, but not male pattern, increases apoE levels (J. Oscarsson et al., 1991a; Jan Oscarsson et al., 1991b).

However, our study points out that apoE may be affected by gender and/or GH in an isoform-specific way. For example, while isoform 1 is increased in females (regardless of genotype), isoform 4 is decreased in female WT mice only. Thus, without GH action (as seen in GHR-/- mice), apoE isoform 1 still is increased in females. Additionally, only the female pattern of GH secretion decreases apoE isoform 4, but not male pattern or a total lack of GH action. This suggests that in addition to gender-specific GH pattern, other gender differences also regulate apoE, in particular the PTM of apoE.

ApoE is an apolipoprotein that transports VLDL, LDL and HDL. It is also an antioxidant protein in plasma (Davignon, 2005; Miyata & Smith, 1996; Tarnus et al.,

2009). Upregulation of apoE in bGH mice may be related to increased HDL and LDL cholesterol levels in these mice (Frick et al., 2001a). Moreover, decreased apoE may indicate decreased cholesterol levels in GHR-/- mice (A. Bartke et al., 2004; E. Egecioglu et al., 2005; Masternak et al., 2005a). It will be intriguing to unravel other physiological relevance of an increase in apoE in response to GH.

Three MBP-C isoforms have been identified (Figure 15). MBP-C has a theoretical

Mw of ~24 kDa and pI ~5. PTMs of MBP-C include hydroxylation on proline residues and N-linked glycosylation. The three isoforms identified on 2D gels were ~32 kDa, possibly due to glycosylation and proline oxidation. Isoform 1 increases in bGH mice

(Figure 36) and decreases in GHR-/- mice (p<0.05, Table 14) and therefore a biomarker of chronic GH action. MBP-C has been reported to be increased by GH treatment in 228 humans (Gravholt et al., 2004; T. K. Hansen et al., 2001) by increased hepatic synthesis of MBP-C (Sørensen et al., 2006). This effect is also IGF-1 independent (C. M. Sørensen,

2006; T. K. Hansen et al., 2001). However, in our study MBP-C is not altered by GH or

IGF-1 injection for 2 weeks. Therefore, MBP-C is more likely a biomarker of long-term

GH action and possibly longer time periods of GH treatment will result in an increase in its plasma level.

MBP-C is an innate immune defense protein against pathogens and functions to activate the classical complement pathway (Turner, 2003). Low levels of MBP-C are associated with susceptibility to frequent and chronic infections (Sumiya et al., 1991;

Turner, 2003). GH has been linked to a positive modulation of innate immunity (T. K.

Hansen, 2003; Kelley et al., 2007), and the increase of MBP-C in bGH mice and the decrease in GHR-/- mice may be an indicator of this action of GH.

Seven a2m isoforms with Mw ~35kDa and pI ranging from 5.65 to 7.08 have been identified (Figure 15). As discussed in chapter IV, these a2m isoforms correspond to the

35kDa subunit, which has a theoretical pI of 7.09. A2m shows isoform-specific alteration by GH. Isoforms 1 and 2 are increased while isoforms 5 and 6 are decreased in bGH versus WT mice (Table 7, Figure 44). None of the a2m isoforms are significantly altered in GHR-/- mice. Moreover, isoform 2 also increases in GH injected mice (Figure 70).

Interestingly, none of the a2m isoform levels are altered by IGF-1 injection, suggesting that a2m regulation by GH is independent of IGF-1. Thus, a2m is a biomarker of GH; in particular, isoform 2 can indicate both chronic and acute GH actions. This represents a significant finding that distinguishes IGF-1 from GH actions. 229

Four isoforms of Hp have been identified (Figure 15). Two of the isoforms, 2 and

3, increase as WT mice age (Figure 16, Figure 21). All four isoforms are upregulated in bGH versus WT mice (Figure 37). All four isoforms also increase during aging in bGH mice at a much faster rate than in WT mice. This result positively correlates with the accelerated aging phenotype of bGH mice. Interestingly, three of the isoforms, 1-3, decrease in GHR-/- versus WT mice (p<0.05, Table 14). These have not been reported in chapter V because of not reaching a significance cutoff value of p<0.01. In both GH and

IGF-1 injected mice, isoforms 2-4 increase, and remain elevated after the washout period

(Figure 70, Figure 77). These data suggest that Hp is a biomarker of GH and IGF-1.

Hp is also sexually dimorphic. Isoform 2 is reduced in the plasma of female compared to male mice (Table 8). Isoforms 3 and 4 are reduced in female WT mice only, with similar levels in male WT and both genders of GHR-/- mice (Figure 59), suggesting an interaction of GH and gender on the levels of isoforms 3 and 4.

Related to Hp is Hbb, which is theoretically possesses a Mw of 15.7 kDa and a pI of

7.26. Four isoforms of Hbb have been identified in this dissertation with observed Mw

~12kDa and pIs ranging from 7.24 to 7.92 (Figure 15). Isoforms 1 and 3 decrease markedly in female compared to male mice (Table 8). Hbb is a subunit of hemoglobin, which can be released into plasma due to red blood cell lysis. Incidentally, hemoglobin α- chain has been found to increase in response to GH treatment (Chung et al., 2006), but it is a different subunit (α-subunit). The presence of hemoglobin, in particular heme, imposes an oxidative stress in plasma as heme is a molecule with considerable oxidative potential (Carrell et al., 1977). Hp as an acute phase protein can be induced by such 230 stress. It binds to hemoglobin and quenches its oxidative damage. The lower levels of

Hbb in female plasma may therefore indicate a reduced need for Hp, explaining lower levels of both Hbb and Hp in females. Hp is also an indicator of inflammation. Its induction and upregulation help to combat stress, infection and injury; however, it also reflects increased inflammation, which tends to rise as a function of age (Teunissen et al.,

2003). Therefore, the higher levels of Hp in bGH mice and in mice injected with GH and

IGF-1, as well as the lower levels of Hp in GHR-/- mice and the lower levels of Hbb in female mice are consistent with the healthspan and lifespan of these animals.

Clu is identified as six isoforms in this dissertation (Figure 15). These are β- subunits (see chapter IV) of Clu with Mw 34-36 kDa and pI ranging from 5.28 to 5.77.

Clusterin β-chain has a theoretical Mw of 23.9 kDa and pI 6.12. Isoforms 1, 2 and total

Clu show accelerated increased plasma levels as bGH mice age but do not change as WT mice age (Figure 45, Figure 46). The other isoforms (3-6) show a similar pattern although not significant at p<0.01. Plasma Clu levels are lower in bGH versus WT mice at 2 and 4 months, but quickly becomes higher than WT after 8 months, thus showing a significant interaction between GH and aging. Together with Hp, Clu also indicates the accelerated aging phenotype of bGH mice. Clu is not changed by IGF-1 injection; however, isoforms 3-6 and total Clu are decreased by GH injection (Figure 69), consistent with lower Clu levels in bGH mice at young ages (mice used for injections are

3 months of age). Thus, Clu is suppressed by GH in young mice independent of IGF-1.

However, after 7 days of washout, Clu returns to normal, indicating that Clu is not a 231 long-lasting biomarker of GH. Clu is also sexually dimorphic. Isoform 3 decreases in female compared to male mice (Table 10).

Seven RBP-4 isoforms have been identified in this dissertation, of which isoforms

1-3 are ~22kDa and 4-7 are ~19kDa (Figure 15). RBP-4 is 21.6 kDa with pI 6 in theory.

RBP-4 shows isoform-specific changes by GH action. In bGH mice, isoform 2 decreases and 4 increases (Figure 42). In GHR-/- mice, isoforms 1-3 increase and 5 decreases

(Table 9). The results indicate that GH increases the 19kDa isoforms and decreases the

22kDa isoforms. Consistent with this, after the GH washout, GH injected mice have increased isoform 6, which is 19kDa (Figure 70). The regulation of RBP-4 may rely on a

GH secretory pattern. Constant GH secretion as seen in females or bGH mice favors the

19kDa over 22kDa isoforms. Isoform 7 increases while isoform 3 decreases in female compared to male mice (Table 8). Isoform 1 decreases during female aging but does not change in males (Figure 50). Female WT mice have reduced isoform 1, compared to similar levels in male WT (pulsatile GH), male and in female GHR-/- mice (no GH action) (Figure 59). Therefore, high and constant GH action upregulates the 19kDa and downregulates the 22kDa RBP-4 isoforms. RBP-4 also is associated with obesity (Haider et al., 2007; Kowalska et al., 2008; Kohzo Takebayashi et al., 2007a). Consistent with these findings is the result that total RBP-4 is decreased in bGH mice (very lean) and increased in GHR-/- mice (obese).

Nine isoforms of TTR have been identified in this dissertation (Figure 15). TTR is in theory 13.6 kDa (without the secretory signal peptide) with a pI of 5.77. Among the identified TTR isoforms, 1 and 2 have Mw of ~38kDa and are likely multimers of TTR; 232 isoform 3 is ~20kDa; isoforms 4-9 are probably single TTR molecules of ~14kDa.

Isoform 3 increases as WT mice age (Figure 16, Figure 20). In bGH mice, except for isoform 2, all other 8 isoforms decrease compared to WT mice at all ages (Figure 40), indicating a suppression action of GH. However, this effect seems to require long-term

GH action, since only isoform 6 decreases in the GH-injection/washout group (Figure

70). This suppressive effect of GH is independent of, or even antagonistic of IGF-1, since

IGF-1 injection results in increased TTR isoforms 3 and 4.

Three spots have been identified as Ig kappa chain components (Figure 15). Spots 1 and 2 are Ig kappa variable region III, and spot 3 is Ig kappa constant region. All three increase as WT mice age (Figure 16) and therefore, are biomarkers of aging.

Interestingly, spot 1 increases in bGH mice (p<0.05) and spot 3 decreases in GHR-/- mice

(p<0.05), although not reaching a significance value of p<0.01 and therefore are not reported in chapters IV and V. Perhaps the upregulation of Ig kappa in bGH and downregulation of it in GHR-/- mice reflect their reduced and increased lifespans.

Six isoforms have been identified as apoA1 in this dissertation (Figure 15). The theoretical Mw and pI of apoA1 are ~28 kDa and ~5.4, respectively. Observed values of these isoforms were ~28 kDa and pI ranging from 5.73-6.4. Isoforms 4-6 decrease in bGH mice (Figure 41), and isoform 2 decreases in GHR-/- mice (Table 9). In this scenario, excess GH and lack of GH action both downregulate apoA1. Interestingly, while no aging effect is seen in male mice, female mice have increased apoA1 (isoforms

2-5) during aging (Figure 50), suggesting an age-related cardiovascular protective role of apoA1 specifically in females. 233

Five isoforms have been identified as apoA4 in this dissertation (Figure 15).

ApoA4 is ~43 kDa with a pI of 5.27 in theory. The five identified isoforms are ~43 kDa with pI ranging from 5.64-5.9. Isoforms 3, 4 and total apoA4 increase in GHR-/- mice

(Figure 51, Figure 52). Therefore, apoA4 is a biomarker for the lack of GH action.

Isoform 2 decreases after the GH washout period (Figure 69). This isoform could serve as a biomarker for GH doping. Similar to apoA1, apoA4 also shows no change during aging in males; however, isoforms 2 and 3 of apoA4 increase during female aging. Both apoA1 and apoA4 are associated with HDL. Their levels increase as female mice age and therefore may indicate a more favorable lipid profile of females at old age.

Although full-length Alb is not resolved on our 2D gels, as many as 23 Alb fragments are detected, with Mw ranging from ~9kDa to ~33kDa (Figure 15). Isoforms

14-17, 18 and 20 increase after the GH washout period (Figure 73). Isoforms 14, 15, 17,

18, 21 and 23 increase by IGF-1 injection (Figure 79). The pattern shows that IGF-1 injection increases these isoforms relatively quickly and the increased levels are maintained after washout. However, GH treatment results in a more gradual increase in these isoforms. Thus the incease in these isoforms detected after the GH washout suggests that regulation of the levels of these Alb isoforms occur somewhat further downstream of acute GH signaling. Indeed, Alb fragments are likely the result of yet unknown proteolytic activity, which may be a target of GH and IGF-1. Although the mechanism is not clear, the fact that these Alb fragments increase after 7 days of GH washout makes them an attractive biomarker candidate for GH doping. A subset of these isoforms (#14, 15, 17 and 18) overlap with IGF-1 markers, further confirming the effect 234 of GH/IGF-1 axis on the Alb isoforms. In summary, Alb isoforms 14-17, 18 and 20 are upregulated by GH; isoforms 14, 15, 17, 18, 21 and 23 are upregulated by IGF-1; and isoforms 14, 15, 17 and 18 are upregulated by both GH and IGF-1

In addition, the Alb isoforms are also subjected to aging and gender effects.

Isoforms 6-9 and 18 decrease as WT mcie age (Figure 22). In WT females, isoform 19 decreases at all ages. Also, isoforms 21 and 23 increase as female mice age; isoforms 12 and 14-16 decrease as females age (Figure 50), while no change is found as males age for these isoforms. Together with results in injection studies, isoform 18 is decreased during aging, upregulated by GH and IGF-1 injections; isoform 14 and 15 are upregulated by

GH and IGF-1 injections, and decreased during female aging. Thus, Alb fragmentation is regulated by GH, aging and gender.

One protein spot has been identified as SAA-1 in this dissertation (Figure 15), which is 12 kDa in theory. The observed Mw of this protein is 11.5 kDa. It decreases as

WT mice age (Figure 16); therefore, it may be considered a biomarker of aging. GH injection also increases SAA-1 after the GH washout period but not immediately after

GH treatment (Figure 70), suggesting that SAA-1 is a biomarker for GH doping. Further, this action is independent of IGF-1 which does not alter its level either after injection or

IGF-1 washout. SAA-1 is an apolipoprotein that associates with HDL, especially during acute inflammation (Artl et al., 2000; Kisilevsky & Subrahmanyan, 1992). SAA- associated HDL (versus apoA1-associated HDL) has a larger binding affinity to macrophages, thereby mediating the uptake of HDL in macrophages. This notion has been hypothesized to lead to the transformation of macrophages into cholesterol-rich 235 foam cells that could deliver phospholipids and cholesterol for tissue repair during inflammation and injury (Artl et al., 2000). The foam cells are also found at atherosclerotic plaques; thus sustained inflammation may transform HDL

(antiatherogenic lipoprotein particle) into a proatherogenic particle (Artl et al., 2000).

Therefore, the data presented in this dissertation showing that GH injection results in increased SAA-1 may indicate an alteration of HDL during acute inflammation induced by GH

One protein spot has been identified as Prx-2 in this dissertation (Figure 15). Prx-2 has a theoretical Mw of 22 kDa and pI of 5.2. On the gel, it appears to be at the location of Mw 25 kDa and pI 6.1. It decreases as WT mice age and is, therefore, a biomarker of aging (Figure 22). Decreased Prx-2 (an anti-oxidant enzyme) indicates a reduced capacity to fight oxidative stress as mice age.

Biomarkers for Aging, GH and IGF-1

In conclusion, biomarkers of aging, GH and gender have been identified in mice.

Plasma biomarkers for aging include: Hp (isoforms 2 and 3), TTR (isoform 3), and Ig kappa chain (spots 1-3) that increase; and Prx-2, SAA-1 and Alb fragments (isoforms 6-

9, 18) that decrease during aging in WT male mice. These are the first plasma biomarkers discovered in a longitudinal study after screening mouse plasma proteome using 2-DE.

Importantly, a human cross-sectional study indicates that plasma TTR increases and Alb decreases as a function of age (Robert et al., 1999b), confirming these two proteins as biomarkers of aging. One or more of these proteins may be used as therapeutic target(s) to delay aging. 236

Biomarkers for chronic GH action as seen in bGH mice include apoE (all 5 isoforms), MBP-C (isoform 1), Hp (all 4 isoforms), a2m (isoforms 1 and 2) and RBP-4

(isoform 4) that increase; TTR (all isoforms except isoform 2), a2m (isoforms 5 and 6),

RBP-4 (isoform 2) and apoA1 (isoforms 4-6) that decrease. These plasma proteins indicate an increased inflammatory state and a less favorable lipoprotein profile in these giant, short-lived mice (as will be discussed below), which not only provides a better understanding of the physiology of bGH mice, but also an possible explanation of their shortened lifespan.

Biomarkers for short-term GH action that last for up to 7 days after GH injection include a2m (isoform 2), Hp (isoforms 2-3), RBP-4 (isoform 6) and Alb fragments

(isoforms 14-17, 18 and 20) that increase; TTR (isoform 6) and apoA4 (isoform 5) that decrease. This is the first time that novel biomarkers of short-term or acute GH treatment have been reported. Further, several of these marker proteins are the same for chronic GH action seen in bGH mice. Notably, these makers last at least until 7 days after the GH injection, at which time the body composition changes are diminished (Figure 63). This suggests that these are better markers of GH action than increased lean and decreased fat mass, the two traditional physiological activity read-outs of GH action. These plasma proteins are promising biomarkers for a novel GH doping detection. Note that some proteins do show a difference between acute and chronic GH conditions; thus similarities and disparities exist for chronic and acute GH effects on plasma proteins (Table 13 in chapter VI). 237

Biomarkers for short-term IGF-1 treatment that last for up to 7 days after injection include Hp (isoforms 2 and 4), TTR (isoforms 3 and 4) and Alb fragments (isoforms 14,

15, 17, 18, 21 and 23) that increase. Like GH, IGF-1 is also abused among athletes; currently there is no method to detect IGF-1 doping (Guha et al., 2009; R I G Holt,

2008). These plasma proteins therefore provide novel candidates for developing a method for IGF-1 doping detection. Note that Hp and several Alb isoforms are changed similarly in GH treatment, suggesting these are biomarkers of GH/IGF-1 actions, likely affected by

IGF-1 instead of GH directly.

Importantly, biomarkers that differentiate between GH and IGF-1 actions have been reported for the first time in this dissertation, including apoE (increased by GH but not altered by IGF-1) and TTR (decreased by GH and increased by IGF-1) (Figure 80). For decades, no plasma markers are known to be differentially regulated by GH and IGF-1, although clearly the two have different functions. For example, GH inhibits insulin action and IGF-1 has insulin-like activities such as promoting glucose uptake. GH is lipolytic and anti-lipogenic whereas IGF-1 does not have these properties. These results, therefore, fills in a large gap in this area of research and hopefully will inspire further studies on the mechanisms and biological significance of the differences observed in these plasma proteins.

Biomarkers for the absence of GH action as seen in GHR-/- mice include apoA4

(isoforms 2, 4 and total) and RBP-4 (isoforms 1-3 and total) that increase; apoE (isoforms

1, 3 and 4 and total), RBP-4 (isoform 5) and apoA1 (isoform 2) that decrease. These 238 proteins indicate a reduced inflammatory state and favorable lipoprotein profile in this long-lived mouse.

Biomarkers for gender include apoE (isoform 1) and RBP-4 (isoform 7) that increase in females; and Hp (isoform 2), Clu (isoform 3), RBP-4 (isoform 3) and Alb fragments (isoform 19) that decrease in females compared to males. Interestingly, several proteins including apoA1, apoA4 and Alb fragments show a significant increase as females age but not in males (Table 14). Very few studies look at gender difference

(Palmer et al., 2009; Serrat et al., 2007), let alone a longitudinal proteomic profile of plasma biomarkers. This result ‘opens up a new door’ for identifying gender-specific differences in proteins and how they may affect aging.

Judging from all the mouse models above, the most likely biomarkers for GH action are apoE (in particular isoform 3), Hp (in particular isoforms 2 and 3), MBP-C (isoform

1), a2m (isoform 2) and SAA-1 that increase, as well as TTR (in particular isoform 6) that decreases with GH. The best biomarkers for aging are Hp (isoforms 2 and 3) that increases and Prx-2 and SAA-1 that decrease with age.

A Model of GH’s Impact on Aging

In conclusion, I would like to present a model of aging (Figure 82) that is characterized by increased inflammation, increased cardiovascular disease risks and decreased anti-oxidative stress ability. These age-related processes are summarized in a

‘funnel’ as depicted in Figure 82. Below, I will discuss these three physiological issues in relation to the proteins discovered in this dissertation in a general way, and then more comprehensively. 239

Figure 82. Model of GH’s impact on aging . Generally, inflammation is increased as normal mice age, in bGH mice and is decreased in GHR-/- mice. For example, many of the biomarkers discovered in this dissertation are positive acute phase proteins (induced and increased by inflammation), including Hp, MBP-C, SAA-1 and a2m, as well as negative acute phase proteins (acute phase protein that is decreased by inflammation) such as TTR and RBP-4 (Ananian et al.,

2005; Fleck, 1989; Robert et al., 1999b). Hp increases during aging in WT mice. Hp and

MBP-C are increased in bGH mice and decreased in GHR-/- mice. Increased Hp and

SAA-1, as well as one isoform of a2m and the decrease in one isoform of TTR by GH injection are in line with this action of GH. On the other hand, negative acute phase proteins such as TTR and RBP-4 are decreased in bGH mice and increased in GHR-/- 240 mice. In terms of cardiovascular disease risks, many markers are increased during aging and in bGH mice but decreased in GHR-/- mice as manifested by several cardio- protective plasma proteins including albumin, apoA1 and apoA4. For example, albumin decreases during aging, apoA1 decreases in bGH mice and apoA4 increases in GHR-/- mice, suggesting increased cardiovascular disease risks as mice age and in bGH mice, but decreased risks in GHR-/- mice. Finally, antioxidant protein is decreased during aging.

For example, Prx-2 is decreased as normal mice age, suggesting decreased ability to resist oxidative stress. Thus, increased inflammation, increased cardiovascular disease risks and decreased anti-oxidation together contribute to aging, which is promoted under chronic GH excess as seen in short-lived bGH mice, and inhibited or delayed by diminished GH/IGF-1 axis as seen in long-lived GHR-/- mice. Thus, the three physiological states shown in Figure 82 and protein ‘markers’ associated with them act together to influence mouse aging. Below, I will further expand on these issues as they relate to GH action and aging.

Inflammation

Inflammation is a natural defense mechanism against infection or injury. Pathogens and other immune insults activate nuclear factor (NF) κB, a transcription factor that activates a host of genes responsible for immune defense (Hayden & Ghosh, 2008). GH is known to affect the immune system. GH transgenic mice have enlarged spleen and thymus (Dialynas et al., 1999), show altered immunity, including deficiency in T helper-2 cytokine production (Gonzalo et al., 1996) and autoimmune arthritic disorders (Ogueta et al., 2000). However, the relationship between GH and inflammation is not clear. This 241 dissertation has identified several acute phase proteins that are upregulated in bGH mice.

Acute phase proteins are induced in the liver by inflammation and are an important component of innate immune defense. GH has been shown to upregulate certain acute phase proteins such as MBP-C, TTR and RBP-4 (Gravholt et al., 2004; T. K. Hansen et al., 2001; Kemp & Canfield, 1983; Vranckx et al., 1994). It is not clear how GH induces acute phase proteins at the molecular level, although it is likely by JAK2/STAT5 pathway

(Barkai et al., 2000; Dajee et al., 1996), which is also the pathway via which many cytokines are induced. GH has been shown to activate NFκB via the PI3K-Akt pathway

(Jeay et al., 2002); thereby providing a potential cross-talk mechanism of GH and inflammation.

Again, the physiological role of GH in the inflammation response is not clear.

Acute inflammation is featured with classic signs of redness, heat, swelling and pain.

When injury or infection is cleared, these signs and symptoms of inflammation stop

(Serhan & Savill, 2005). If the cause of the inflammation sustains, chronic inflammation occurs, which is featured with increased inflammatory cytokines and reactive oxygen species and eventually tissue fibrosis. Studies suggest that GH is induced under acute inflammation. For example, serum GH levels are elevated in children suffered from septic shock (de Groof et al., 2002; Önenli-Mungan et al., 2004). During sepsis caused by lipopolysacchride (LPS), GH levels are increased (Coleman & Sartin, 1996; Lang et al., 1997). It is possible that during infection, the upregulation of GH serves as a beneficial signaling for the induction of acute phase proteins, which, together with other pro-inflammatory cytokines, are necessary to fight infection. However, GH seems to be a 242 double-edged sword, as high levels of GH are detrimental to critically ill patients (Y.

Chen et al., 2007; Takala et al., 1999). Moreover, animals under sepsis become GH resistant with suppressed JAK2/STAT5 signaling and STAT5 DNA binding (Y. Chen et al., 2007). Thus, naturally-induced GH during infection appears beneficial; however, chronic high doses of GH treatment may have an adverse effect. Clearly in bGH mice, the increase of inflammation inferred by the plasma proteins is not due to infection or injury, but possibly chronic high levels of GH, which could induce acute phase proteins and perhaps other inflammatory cytokines such as TNF-α (unpublished observation by

Valente and Berryman).

Increased inflammation is associated with aging (Brüünsgaard & Pedersen, 2003;

Ferrucci et al., 2005; Roubenoff et al., 1998). Inflammatory markers such as (IL)-1β, tumor necrosis factor (TNF)-α and IL-6 increase during aging (Kireev et al., 2009). From the results in this dissertation, it is likely that the chronic inflammation eventually leads to accelerated aging of bGH mice. There is evidence that the chronic excess of GH action causes increased oxidative stress in the liver and kidney (S. Q. Doi et al., 2000; Steven J.

Hauck & Bartke, 2001). On the other hand, GHR-/- mice have reduced levels of Hp,

MBP-C, indicating a reduced inflammatory state in these dwarf mice, therefore, increased lifespan (Figure 82).

Inflammation is also associated with obesity and insulin resistance (Taubes, 2009).

In the case of obesity-induced insulin resistance, storage of excess fat causes enlargement of adipocytes, which attract macrophages that excrete inflammatory cytokines such as IL-

1β, TNF-α and IL-6 that block insulin signaling in muscle cells via c-Jun N-terminal 243 kinase (JNK) that inhibits IRS-1 pathway (Figure 83). As mentioned earlier, few reports exist in literature regarding the mechanism of inflammation induced by GH. BGH mice actually have smaller adipocytes (presumably due to lipolytic action of GH); however, macrophage infiltration of adipose tissue and increased TNF-α in bGH mice have been observed (unpublished results by Valente and Berryman). Therefore, the increased inflammatory markers in these animals are not due to enlarged fat tissue; rather, excess

GH may cause inflammation directly by increasing TNF-α and acute phase proteins that have been discovered in this dissertation. On the other hand, life-long suppression of

GH/IGF-1 by antisense GH transgene expression in rats, a model similar to GHR-/- mice, resulted in decreased inflammation as evidenced by reduced renal infiltration of macrophages and age-related renal damage (Zha et al., 2006). In agreement with this,

GHR-/- mice are protected from kidney diseases whereas bGH mice develop glumerulosclerosis prematurely (Bellush et al., 2000; T. Doi et al., 1990; T. Doi et al.,

1988; Quaife et al., 1989). Interestingly, GHR-/- mice have depot-specific increases in adipocyte size compared to WT mice, with enlarged retroperitoneal and subcutaneous fat cells and unchanged epididimal adipocytes (Kelder et al., 2007). Therefore, mechanism other than fat cell size must exist to explain the altered insulin sensitivity in bGH and

GHR-/- mice.

244

Figure 83. Inflammation and lipotoxicity contribute to insulin resistance. DAG: diacylglyceride; IL: interleukin; TNF: tumor necrosis factor; JNK: c-Jun N-terminal kinase; IRS: insulin receptor substrate; PKC: protein kinase C. Reproduced from (Taubes, 2009).

On the other hand, how excess fat is stored has also been recognized as important in terms of insulin resistance. It seems that if fat can be stored in newly formed adipocytes instead of ‘stuffing’ the existing over-sized fat cells with lipid or accumulating lipid ectopically in tissues such as liver and muscle, then insulin sensitivity can be maintained

(Taubes, 2009). GHR-/- mice have reduced cell number in parametrial depot, but increased cell number in subcutaneous depot compared to controls (Flint et al., 2006), indicating that these obese mice are capable of generating more adipocytes in subcutaneous region to store the excess fat. This may be one of the reasons that explain 245 their insulin sensitivity despite being fat. Interestingly, GH transgenic pigs show no mature subcutaneous adipocytes (Pinkert et al., 1994), In this aspect, it is not known if bGH mice have altered number of fat cells compared to controls. GH is known to promote adipocyte differentiation and proliferation in parametrial but not subcutaneous fat cells (Flint et al., 2006); thus GH may act on adipocytes in a depot-specific manner regarding adipocyte differentiation and proliferation. More studies are needed to uncover the action of GH in fat cell development.

Inflammation may in part explain the insulin resistance of bGH mice. GH itself is well-known to have anti-insulin actions (Cotes et al., 1949) and insulin resistance is commonly observed in acromegaly (Arya et al., 1997). GH causes insulin resistance by upregulating p85 (del Rincon et al., 2007), the regulatory subunit of PI3K, a key insulin signaling molecule. Although direct evidence is needed to link GH-induced inflammation with insulin resistance, it is likely to be another mechanism of insulin resistance by GH.

Another contributor to insulin resistance is lipotoxicity (Figure 83). In this scenario, excess fatty acids lead to accumulation of intracellular diacylglycerols (DAGs), which activate serine kinases that inhibit insulin signaling. GH transgenic mice have reduced plasma and tissues levels of triglyceride and free fatty acids, but increased cholesterol (A.

Bartke et al., 2004; Frick et al., 2001a; Olsson et al., 2003; Olsson et al., 2005; Palmer et al., 2009; Zhihui Wang et al., 2007). In contrast, GHR-/- mice have reduced or unchanged plasma and tissue levels of triglyceride, free fatty acids and cholesterol (A.

Bartke et al., 2004; E. Egecioglu et al., 2005; Liu et al., 2004; Masternak et al., 2005a). 246

Unfortunately, no data have reported DAG levels in tissues. Therefore, a role of lipid metabolism in insulin sensitivity in bGH and GHR-/- mice warrants further study.

Together, the biomarkers found in this dissertation suggests that bGH mice have increased inflammation and insulin resistance, whereas GHR-/- mice have decreased inflammation and increased insulin sensitivity, contributing to the shortened and enhanced lifespans of these mice, respectively.

Cardiovascular Disease Risk Factors

A second major physiological issue related to aging is cardiovascular disease risk factors including lipoprotein profiles. In this dissertation, it has been observed that several apolipoproteins are altered by GH. ApoE, SAA-1 are increased by GH, whereas apoA1, apoA4 and clusterin are decreased by GH. All these proteins associate with HDL

(apoE also with LDL and VLDL), suggesting GH may change the protein composition of

HDL. GH is known to be lipolytic and alter lipid metabolism; thus, it is not be surprising to see apolipoprotein changes in bGH and GHR-/- mice. However, the functional differences of these HDL-associated proteins warrant further study. ApoA1 is negatively associated with cardiovascular disease risks. Therefore, the decreased apoA1 may indicate an increased cardiovascular disease risk in bGH mice. On the other hand, GHR-

/- mice have increased apoA4, also negatively associated with heart disease, suggesting a protective effect in these long-lived mice (Figure 82). One note of caution is that due to species difference, lipoprotein metabolism and its indication in cardiovascular disease risks may be different between humans and rodents. For example, mouse and rats have a very different lipoprotein cholesterol distribution from humans, with ~80% HDL and 247

~15% LDL in mouse and rat, whereas ~30% HDL and 60% LDL in humans (Fernandez

& Volek, 2006). Mouse and human show important differences in apoE gene promoter region, including only a 40% DNA sequence homology and different DNA-binding protein factors (Bryan et al., 2007). Rodent models of cardiovascular diseases, heart failure and hypertrophy exhibit differences from humans in terms of myocardium action potential, cytosolic calcium exchange and heart rate (Hasenfuss, 1998). Therefore, the associations with heart disease risks for humans mentioned above are not necessarily applicable in mice.

Oxidative Stress

The third physiological factor related to aging is oxidative stress, which is associated with aging and considered to be the cause of aging according to some theories

(Dowling & Simmons, 2009; Finkel & Holbrook, 2000; Honda & Honda, 2002). Prx-2 is an anti-oxidant enzyme that reduces reactive oxygen species. C. elegans with mutation in

Prx-2 exhibits reduced lifespan (OlÃhovÃ et al., 2008). In this dissertation, Prx-2 plasma levels are found to be decreased as a function of age in WT mice, indicating a reduced capacity to fight oxidative stress at older ages (Figure 82). This result supports the oxidative stress/aging mechanism. Although the level of this protein does not show significant changes in bGH and GHR-/- mice, some evidence does show that chronic GH excess in mice causes reduced antioxidant enzymes against oxidative stress. For example, hGH mice have early declines of antioxidant enzyme superoxide dismutase (SOD) in the kidney and reduced levels of glutathione peroxidase (GPx) in the liver (Steven J. Hauck

& Bartke, 2001). On the other hand, it is not clear whether GHR-/- mice possess more 248 antioxidant enzymes or are better at resisting stress. GHR-/- mice do not have increased antioxidant enzymes or free radical scavenge in liver and kidney compared to controls (S.

J. Hauck et al., 2002). Male GHR-/- mice are less able to resist paraquat-induced stress than WT and female GHR-/- mice show no difference than WT (S. J. Hauck et al., 2002).

In a more recent study, anti-oxidant enzymes are found to increase slightly (non- significant) in various tissues including liver, heart, kidney, skeletal muscle and brain of

GHR-/- mice (Brown-Borg et al., 2009). Thus, more studies are needed to confirm the oxidative stress status in GHR-/- mice.

Comparing Plasma Proteins in bGH and GHR-/- Mice

A unique feature of my research was the ability to use two mouse models that display two extremes of GH signaling; giant bGH and dwarf GHR-/- mice. As presented throughout this dissertation, opposite trends of change in certain plasma proteins were observed. For example, GHR-/- mice have increased RBP-4, decreased Hp and MBP-C

(p<0.05), which are the opposite in bGH mice. However, not all proteins show an opposite direction of change. For example, GHR-/- mice have normal levels of TTR, which is decreased in bGH mice. This is not entirely surprising, because if GH is involved but is not required for a process, its excess certainly produces a difference, but its absence may be rescued by other compensatory mechanisms. For example, bGH mice have elevated levels of IGF-1 protein in muscle; however, GHR-/- mice have normal levels of IGF-1 in muscle, even though circulating IGF-1 is greatly reduced (Kraemer et al., 2008). GH stimulates the immune system, however, GHR-/- mice show normal immune functions (Zhou et al., 1997), as do GH deficient Ames dwarf mice (Hall et al., 249

2002). Another example is that bGH and GHR-/- mice both have elevated levels of corticosterone compared to controls (E. Egecioglu et al., 2005; Hall et al., 2002; S. J.

Hauck et al., 2001). Thus, it is not necessarily true that all biomarkers have to show opposite changes in bGH and GHR-/- mice.

Another example is that apoA4 showed normal levels in bGH mice but elevated in

GHR-/- mice. As discussed in chapter V, apoA4 is an apolipoprotein transporting HDL

(the ‘good’ cholesterol), and an anti-oxidant protein that removes superoxides and lipid hydroperoxide (Ostos et al., 2001; W.-M. R. Wong et al., 2007). It also has anti- atherosclerosis effect, reduces inflammation (Recalde et al., 2004), and is associated with reduced risk of cardiovascular diseases (Kronenberg et al., 2000; Warner et al., 2001).

Therefore, elevated apoA4 may be an indicator of the prolonged longevity seen in GHR-

/- mice (Figure 82).

Conclusion

In conclusion, this dissertation is the first proteomic study with longitudinal monitoring of plasma proteins in the entire lifespan of bGH and GHR-/- mice along with their littermate controls or normal animals. Screening of mouse plasma allows for the identification of novel biomarkers for acute, chronic and absence of GH actions.

Additionally, analysis of the plasma of control (WT) mice provided the first longitudinal proteomic analysis of normal aging. The results will not only be helpful for developing a more sensitive and accurate method for detecting rhGH doping, but also have revealed several important physiological activities of GH in vivo that have not been systematically studied or explicitly proposed previously. The biomarkers for aging described in this 250 dissertation indicate increased inflammation and decreased anti-oxidative stress as a function of age, consistent with current theories on aging ( Brüünsgaard & Pedersen,

2003; Dowling & Simmons, 2009; Ferrucci et al., 2005; Finkel & Holbrook, 2000;

Honda & Honda, 2002; Roubenoff et al., 1998). Although not stressed in this write-up, these biomarkers could be potential therapeutic targets against which medical intervention may be designed.

I would like to propose that plasma biomarkers of chronic GH excess and absence are indicative of increased inflammation and unfavorable lipoprotein profiles in as seen in bGH mice, and reduced inflammation, and favorable lipoprotein profiles observed in

GHR-/- mice. These two physiological states ultimately contribute to the decreased and increased lifespans observed in these two mouse models, respectively (Figure 82).

Although anti-oxidative stress biomarkers decrease as a function of normal aging, a clear relationship to GH action or lack thereof was not seen. Further studies are needed to clearly show a relationship between GH and inflammation, which may ultimately unravel the mechanism of aging and the ‘fountain of youth’.

251

CHAPTER VIII. FUTURE DIRECTIONS

The identification of plasma biomarkers for GH and aging as discovered in this dissertation is only the first step in a long process of determining their function. Next, the source of the proteins in plasma needs to be resolved. Although most of the proteins in this dissertation are secreted by liver, some are expressed by a variety of tissues. Since isoforms of many proteins were discovered, twovery important questions arise: (1) What

PTM is responsible for the multiple isoforms observed on 2D gels, and (2) In what tissue does the PTM occur? As mentioned above, often one or a few specific isoforms are biomarkers of GH or aging, but not necessarily the total level of the protein. In addition, sometimes different isoforms of the same protein show opposite changes, i.g., isoforms 1 and 2 of a2m are increased in bGH mice but isoforms 5 and 6 are decreased. Unraveling the production and physiological function of the PTMs of the distinct isoforms is indeed a challenge.

Another issue would be to establish the function and mechanism of regulation of the biomarkers. For example, SAA-1 decreases with age; therefore the question becomes: what causes this decline? Similarly, since apoE is increased by GH, the outstanding question is how is it regulated and what physiological role does increased apoE play in response to GH? Understanding the mechanism responsible for the increased or decreased plasma levels of the proteins reported here is important in that it may help provide an explanation as to the mechanism by which GH acts. In a similar manner, the aging markers reported here may provide a molecule link between the aging phenotype and its underlying cause(s). 252

One way to study the functions of these proteins is to generate transgenic animals with overexpression or deletion of the corresponding genes encoding the proteins of interest. For example, Prx-2 decreases with aging. Thus, if a transgenic mouse overexpressing Prx-2 were demonstrated to be longer-lived, then one could propose a protective role of Prx-2 against aging. Similarly, Hp increases during aging and is upregulated in bGH mice. Thus, if an Hp transgenic mouse were shown to have accelerated aging such as seen in bGH mice, then this could confirm the causative role of

Hp in promoting aging.

Knowledge on biomarkers and the mechanism of their regulation must be applied to humans. Biomarkers of rhGH action can be used to develop a novel method to detect GH doping. Likewise, plasma biomarkers of IGF-1 action would make it possible to detect

IGF-1 doping. In the same way, biomarkers for aging may provide viable targets to detect precocious aging that could ultimately lead to an improvement in healthspan. Once the biomarkers are verified in humans, they may result in the generation of diagnostic kits for

GH action, IGF-1 action or aging. An unanticipated result may be the discovery of therapeutic targets for abnormal GH or IGF-1 action of disorders associated with aging.

Finally, the interplay between GH action and aging is, to say the least, interesting.

The lay press constantly advertises for the potential benefits of GH or GH related

‘concoctions’ in delaying the aging process. Yet, scientific evidence stresses that the lack of GH action promotes longevity. Thus, the relationship between GH action and aging is certainly controversial and stimulating. Perhaps the plasma markers that I discovered in my studies will help resolve this controversy. 253

REFERENCES

Adham, N. F., Chakmakjian, Z. H., Mehl, J. W., & Bethune, J. E. (1969). Human growth hormone and alpha-2-macroglobulin: A study of binding. Arch Biochem Biophys, 132(1), 175-183. Ahluwalia, G. S., Kaul, L., & Ahluwalia, B. S. (1980). Evidence of facilitory effect of growth hormone on tissue vitamin a uptake in rats. J Nutr, 110(6), 1185-1193. Al-Regaiey, K. A., Masternak, M. M., Bonkowski, M., Sun, L., & Bartke, A. (2005). Long-lived growth hormone receptor knockout mice: Interaction of reduced insulin-like growth factor i/insulin signaling and caloric restriction. Endocrinology, 146(2), 851-860. Al-Regaiey, K. A., Masternak, M. M., Bonkowski, M. S., Panici, J. A., Kopchick, J. J., & Bartke, A. (2007). Effects of caloric restriction and growth hormone resistance on insulin-related intermediates in the skeletal muscle. J Gerontol A Biol Sci Med Sci, 62(1), 18-26. Albini, C. H., Quattrin, T., Vandlen, R. L., & MacGillivray, M. H. (1988). Quantitation of urinary growth hormone in children with normal and abnormal growth. Pediatric research, 23(1), 89-92. Anderson, L., & Anderson, N. G. (1977). High resolution two-dimensional electrophoresis of human plasma proteins. Proceedings of the National Academy of Sciences, 74(12), 5421-5425. Anderson, N. L., & Anderson, N. G. (1998). Proteome and proteomics: New technologies, new concepts, and new words. Electrophoresis, 19(11), 1853-1861. Anderson, N. L., & Anderson, N. G. (2002). The human plasma proteome: History, character, and diagnostic prospects. Molecular Cellular Proteomics, 1(11), 845. Anderson, N. L., Anderson, N. G., Pearson, T. W., Borchers, C. H., Paulovich, A. G., Patterson, S. D., et al. (2009). A human proteome detection and quantitation project. Mol Cell Proteomics, 8(5), 883-886. Anderson, N. L., Polanski, M., Pieper, R., Gatlin, T., Tirumalai, R. S., Conrads, T. P., et al. (2004). The human plasma proteome: A nonredundant list developed by combination of four separate sources. Molecular Cellular Proteomics, 3(4), 311. Andersson, I. J., Barlind, A., Nystrom, H. C., Olsson, B., Skott, O., Mobini, R., et al. (2004). Reduced sympathetic responsiveness as well as plasma and tissue noradrenaline concentration in growth hormone transgenic mice. Acta Physiol Scand, 182(4), 369-378. Andersson, I. J., Johansson, M. E., Wickman, A., Bohlooly, Y. M., Klintland, N., Caidahl, K., et al. (2006a). Endothelial dysfunction in growth hormone transgenic mice. Clin Sci (Lond), 110(2), 217-225. Andersson, I. J., Ljungberg, A., Svensson, L., Gan, L.-M., Oscarsson, J., & Bergström, G. (2006b). Increased atherosclerotic lesion area in apoE deficient mice overexpressing bovine growth hormone. Atherosclerosis, 188(2), 331. Ando, Y. (2006). Transthyretin-its function and pathogenesis. Rinsho byori.The Japanese journal of clinical pathology, 54(5), 497. 254

Angelucci, F., Mirabella, M., Frisullo, G., Caggiula, M., Tonali, P. A., & Batocchi, A. P. (2005). Serum levels of anti-myelin antibodies in relapsing-remitting multiple sclerosis patients during different phases of disease activity and immunomodulatory therapy. Dis Markers, 21(2), 49-55. Argetsinger, L. S., & Carter-Su, C. (1996). Mechanism of signaling by growth hormone receptor. Physiol. Rev., 76(4), 1089-1107. Artl, A., Marsche, G., Lestavel, S., Sattler, W., & Malle, E. (2000). Role of serum amyloid a during metabolism of acute-phase HDL by macrophages. Arterioscler Thromb Vasc Biol, 20(3), 763-772. Arya, K. R., Pathare, A. V., Chadda, M., & Menon, P. S. (1997). Diabetes in acromegaly- -a study of 34 cases. J Indian Med Assoc, 95(10), 546-547. Asa, S. L., Coschigano, K. T., Bellush, L., Kopchick, J. J., & Ezzat, S. (2000). Evidence for growth hormone (GH) autoregulation in pituitary somatotrophs in GH antagonist-transgenic mice and GH receptor-deficient mice. Am J Pathol, 156(3), 1009-1015. Asleh, R., & Levy, A. P. (2005). In vivo and in vitro studies establishing haptoglobin as a major susceptibility gene for diabetic vascular disease. Vasc Health Risk Manag, 1(1), 19-28. Atzmon, G., Gabriely, I., Greiner, W., Davidson, D., Schechter, C., & Barzilai, N. (2002). Plasma HDL levels highly correlate with cognitive function in exceptional longevity. J Gerontol A Biol Sci Med Sci, 57(11), M712-715. Ayuk, J., & Sheppard, M. C. (2006). Growth hormone and its disorders. Postgrad Med J, 82(963), 24-30. Ballesteros, M., Fredriksson, A., Henriksson, J., & Nystrom, T. (2001). Bacterial senescence: Protein oxidation in non-proliferating cells is dictated by the accuracy of the ribosomes. Embo J, 20(18), 5280-5289. Banks, R. E., Dunn, M. J., Hochstrasser, D. F., Sanchez, J. C., Blackstock, W., Pappin, D. J., et al. (2000). Proteomics: New perspectives, new biomedical opportunities. Lancet, 356(9243), 1749. Banziger-Tobler, N. E., Halin, C., Kajiya, K., & Detmar, M. (2008). Growth hormone promotes lymphangiogenesis. Am J Pathol, 173(2), 586-597. Barber, B. J., Babbitt, R. A., Parameswaran, S., & Dutta, S. (1995). Age-related changes in rat interstitial matrix hydration and serum proteins. J Gerontol A Biol Sci Med Sci, 50A(5), B282-287. Barkai, U., Prigent-Tessier, A., Tessier, C., Gibori, G. B., & Gibori, G. (2000). Involvement of socs-1, the suppressor of cytokine signaling, in the prevention of prolactin-responsive gene expression in decidual cells. Mol Endocrinol, 14(4), 554-563. Barkan, A. L. (2007). Defining normalcy of the somatotropic axis: An attainable goal? Pituitary, 10(2), 135. Bartke, A. (2000). Effects of growth hormone on male reproductive functions. J Androl, 21(2), 181-188. Bartke, A. (2001). Extending the lifespan of long-lived mice. Nature, 414, 412. 255

Bartke, A. (2003). Can growth hormone (GH) accelerate aging? Evidence from GH- transgenic mice. Neuroendocrinology, 78(4), 210-216. Bartke, A. (2005). Minireview: Role of the growth hormone/insulin-like growth factor system in mammalian aging. Endocrinology, 146(9), 3718-3723. Bartke, A., Chandrashekar, V., Bailey, B., Zaczek, D., & Turyn, D. (2002). Consequences of growth hormone (GH) overexpression and GH resistance. Neuropeptides, 36(2-3), 201-208. Bartke, A., Peluso, M. R., Moretz, N., Wright, C., Bonkowski, M., Winters, T. A., et al. (2004). Effects of soy-derived diets on plasma and liver lipids, glucose tolerance, and longevity in normal, long-lived and short-lived mice. Horm Metab Res, 36(08), 550. Bates, S. E. (1991). Clinical applications of serum tumor markers. Ann Intern Med, 115(8), 623-638. Baumann, G. P. (2009). Growth hormone isoforms. Growth Hormone & IGF Research, In Press, Corrected Proof. Bellush, L. L., Doublier, S., Holland, A. N., Striker, L. J., Striker, G. E., & Kopchick, J. J. (2000). Protection against diabetes-induced nephropathy in growth hormone receptor/binding protein gene-disrupted mice. Endocrinology, 141(1), 163-168. Ben-Dov, I., Gaides, M., Scheinowitz, M., Wagner, R., & Laron, Z. (2003). Reduced exercise capacity in untreated adults with primary growth hormone resistance (laron syndrome). Clin Endocrinol (Oxf), 59(6), 763-767. Benedetti, M. G., Foster, A. L., Vantipalli, M. C., White, M. P., Sampayo, J. N., Gill, M. S., et al. (2008). Compounds that confer thermal stress resistance and extended lifespan. Experimental gerontology, 43(10), 882. Berggren, A., Ehrnborg, C., Rosen, T., Ellegard, L., Bengtsson, B. A., & Caidahl, K. (2005). Short-term administration of supraphysiological recombinant human growth hormone (GH) does not increase maximum endurance exercise capacity in healthy, active young men and women with normal GH-insulin-like growth factor i axes. The Journal of clinical endocrinology and metabolism, 90(6), 3268. Berkelman, T. (2008). Removal of interfering substances in samples prepared for two- dimensional (2-d) electrophoresis. Methods in molecular biology (Clifton, N.J.), 424, 51. Berkelman, T. S., T. (1998). 2-d electrophoresis using immobilized ph gradients, principles and methods. Uppsala: Pharmacia Biotech. Bernhard, O. K., Kapp, E. A., & Simpson, R. J. (2007). Enhanced analysis of the mouse plasma proteome using cysteine-containing tryptic glycopeptides. J Proteome Res, 6(3), 987-995. Berryman, D. E., List, E. O., Coschigano, K. T., Behar, K., Kim, J. K., & Kopchick, J. J. (2004). Comparing adiposity profiles in three mouse models with altered GH signaling. Growth Hormone & IGF Research, 14(4), 309. Berryman, D. E., List, E. O., Kohn, D. T., Coschigano, K. T., Seeley, R. J., & Kopchick, J. J. (2006). Effect of growth hormone on susceptibility to diet-induced obesity. Endocrinology, 147(6), 2801. 256

Bessa, S. S., Hamdy, S. M., & Ali, E. M. M. (2007). Haptoglobin gene polymorphism in type 2 diabetic patients with and without nephropathy: An egyptian study. European Journal of Internal Medicine, 18(6), 489. Bidlingmaier, M., Wu, Z., & Strasburger, C. J. (2000). Test method: Gh. Bailliere's best practice & research.Clinical endocrinology & metabolism, 14(1), 99. Bielohuby, M., Roemmler, J., Manolopoulou, J., Johnsen, I., Sawitzky, M., Schopohl, J., et al. (2009). Chronic growth hormone excess is associated with increased aldosterone: A study in patients with acromegaly and in growth hormone transgenic mice. Proc Soc Exp Biol Med, 0901-RM-0934. Bijon, C., & Jürgen, B. (2007). Serum proteomics of lung adenocarcinomas induced by targeted overexpression of c-raf in alveolar epithelium identifies candidate biomarkers. PROTEOMICS, 7(21), 3980-3991. Bijon, C., & Jürgen, B. (2009). A 2-de maldi-tof study to identify disease regulated serum proteins in lung cancer of c-myc transgenic mice. PROTEOMICS, 9(4), 1044- 1056. Bittner, V. (2003). Non-high-density lipoprotein cholesterol and cardiovascular disease. Curr Opin Lipidol, 14(4), 367-371. Blazar, B. R., Brennan, C. A., Broxmeyer, H. E., Shultz, L. D., & Vallera, D. A. (1995). Transgenic mice expressing either bovine growth hormone (bgh) or human GH releasing hormone (hgrh) have increased splenic progenitor cell colony formation and DNA synthesis in vitro and in vivo. Exp Hematol, 23(13), 1397-1406. Bogazzi, F., Raggi, F., Ultimieri, F., Russo, D., D'Alessio, A., Manariti, A., et al. (2009). Regulation of cardiac fatty acids metabolism in transgenic mice overexpressing bovine GH. J Endocrinol, 201(3), 419-427. Bogazzi, F., Russo, D., Raggi, F., Ultimieri, F., Urbani, C., Gasperi, M., et al. (2008). Transgenic mice overexpressing growth hormone (GH) have reduced or increased cardiac apoptosis through activation of multiple GH-dependent or -independent cell death pathways. Endocrinology, 149(11), 5758-5769. Bohlooly, Y. M., Carlson, L., Olsson, B., Gustafsson, H., Andersson, I. J., Tornell, J., et al. (2001a). Vascular function and blood pressure in GH transgenic mice. Endocrinology, 142(8), 3317-3323. Bohlooly, Y. M., Olsson, B., Gritli-Linde, A., Brusehed, O., Isaksson, O. G., Ohlsson, C., et al. (2001b). Enhanced spontaneous locomotor activity in bovine GH transgenic mice involves peripheral mechanisms. Endocrinology, 142(10), 4560-4567. Bollano, E., Omerovic, E., Bohlooly-y, M., Kujacic, V., Madhu, B., Tornell, J., et al. (2000). Impairment of cardiac function and bioenergetics in adult transgenic mice overexpressing the bovine growth hormone gene. Endocrinology, 141(6), 2229- 2235. Bonafè, M., & Olivieri, F. (2009). Genetic polymorphism in long-lived people: Cues for the presence of an insulin/IGF-pathway-dependent network affecting human longevity. Molecular and Cellular Endocrinology, 299(1), 118. Bonkowski, M. S., Dominici, F. P., Arum, O., Rocha, J. S., Al Regaiey, K. A., Westbrook, R., et al. (2009). Disruption of growth hormone receptor prevents 257

calorie restriction from improving insulin action and longevity. PLoS One, 4(2), e4567. Bonkowski, M. S., Rocha, J. S., Masternak, M. M., Al Regaiey, K. A., & Bartke, A. (2006). Targeted disruption of growth hormone receptor interferes with the beneficial actions of calorie restriction. Proc Natl Acad Sci U S A, 103(20), 7901- 7905. Boudrault, C., Bazinet, R. P., & Ma, D. W. (2008). Experimental models and mechanisms underlying the protective effects of n-3 polyunsaturated fatty acids in alzheimer's disease. The Journal of nutritional biochemistry. Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem, 72, 248-254. Brat, O., Ziv, I., Klinger, B., Avraham, M., & Laron, Z. (1997). Muscle force and endurance in untreated and human growth hormone or insulin-like growth factor- i-treated patients with growth hormone deficiency or laron syndrome. Horm Res, 47(2), 45-48. Brown-Borg, H. M. (2009). Hormonal control of aging in rodents: The somatotropic axis. Molecular and Cellular Endocrinology, 299(1), 64. Brown-Borg, H. M., Rakoczy, S. G., Sharma, S., & Bartke, A. (2009). Long-living growth hormone receptor knockout mice: Potential mechanisms of altered stress resistance. Experimental Gerontology, 44(1-2), 10. Brown, R. J., Adams, J. J., Pelekanos, R. A., Wan, Y., McKinstry, W. J., Palethorpe, K., et al. (2005). Model for growth hormone receptor activation based on subunit rotation within a receptor dimer. Nat Struct Mol Biol, 12(9), 814. Brüünsgaard, H., & Pedersen, B. K. (2003). Age-related inflammatory cytokines and disease. Immunology and Allergy Clinics of North America, 23(1), 15. Bryan, M., Yuan-Wen, G., George, M. A., & Debomoy, K. L. (2007). Important differences between human and mouse apoE gene promoters: Limitation of mouse apoE model in studying alzheimer's disease. Journal of Neurochemistry, 103(3), 1237-1257. C. M. Sørensen, T. K. H. R. S. J. C. J. S. T. (2006). Hormonal regulation of mannan- binding lectin synthesis in hepatocytes. Clinical and Experimental Immunology, 145(1), 173-182. Carrell, R. W., Winterbourn, C. C., & French, J. K. (1977). Haemoglobin--a frustrated oxidase? Implications for red cell metabolism. Hemoglobin, 1(8), 815-827 Carroll, P. V., & Van den Berghe, G. (2001). Safety aspects of pharmacological GH therapy in adults. Growth Hormone & IGF Research, 11(3), 166. Carver, E. A., & Stubbs, L. (1997). Zooming in on the humanâ€“mouse comparative map: Genome conservation re-examined on a high-resolutionâ€‰scale. Genome Research, 7(12), 1123-1137. Castle, J. D. (2003). Purification of organelles from mammalian cells. Current protocols in immunology / edited by John E.Coligan.[et al.], Chapter 8, Unit 8.1B. 258

Cejka, J., Vodrazka, Z., & Salak, J. (1968). Carbamylation of globin in electrophoresis and chromatography in the presence of urea. Biochim Biophys Acta, 154(3), 589- 591. Cenci, M. C. P., Conceição, F. L., Soares, D. V., Spina, L. D. C., Brasil, R. R. d. L. O., Lobo, P. M., et al. (2008). Impact of 5 years of growth hormone replacement therapy on cardiovascular risk factors in growth hormone-deficient adults. Metabolism, 57(1), 121. Chandrashekar, V., Bartke, A., Awoniyi, C. A., Tsai-Morris, C. H., Dufau, M. L., Russell, L. D., et al. (2001a). Testicular endocrine function in GH receptor gene disrupted mice. Endocrinology, 142(8), 3443-3450. Chandrashekar, V., Coschigano, K. T., Barta, A., & Kopchick, J. J. (2001b). Neuroendocrine and reproductive functions in transgenic mice with altered growth hormone secretion and in growth hormone receptor gene disrupted mice. In D. M. G. Castro (Ed.), Transgenic models in endocrinology (Vol. 13). Boston: Kluwer Academic Publishers. Chang, Y. H., Lin, K. D., Wang, C. L., Hsieh, M. C., Hsiao, P. J., & Shin, S. J. (2008). Elevated serum retinol-binding protein 4 concentrations are associated with renal dysfunction and uric acid in type 2 diabetic patients. Diabetes Metab Res Rev, 24(8), 629-634. Chanson, P., & Salenave, S. (2008). Acromegaly. Orphanet Journal of Rare Diseases, 3(1), 17. Cheek, D. B., & Hill, D. E. (1974). Effect of growth hormone on cell and somatic growth. In E. Knobil & W. H. Sawyer (Eds.), Handbook of physiology (Vol. 4, pp. 159-185). Bethesda, MD: American Physiology Society. Chen, W. Y., White, M. E., Wagner, T. E., & Kopchick, J. J. (1991). Functional antagonism between endogenous mouse growth hormone (GH) and a GH analog results in dwarf transgenic mice. Endocrinology, 129(3), 1402. Chen, X. L., Lee, K., Hartzell, D. L., Dean, R. G., Hausman, G. J., McGraw, R. A., et al. (2001). Adipocyte insensitivity to insulin in growth hormone-transgenic mice. Biochemical and Biophysical Research Communications, 283(4), 933. Chen, Y., Sun, D., Krishnamurthy, V. M. R., & Rabkin, R. (2007). Endotoxin attenuates growth hormone-induced hepatic insulin-like growth factor i expression by inhibiting jak2/stat5 signal transduction and stat5b DNA binding. Am J Physiol Endocrinol Metab, 292(6), E1856-1862. Chung, L., & Baxter, R. C. (2009). Detection of growth hormone responsive proteins using seldi-tof mass spectrometry. Growth Hormone & IGF Research, In Press, Corrected Proof. Chung, L., Clifford, D., Buckley, M., & Baxter, R. C. (2006). Novel biomarkers of human growth hormone action from serum proteomic profiling using protein chip mass spectrometry. The Journal of clinical endocrinology and metabolism, 91(2), 671. Clark, R. G., Mortensen, D. L., Carlsson, L. M., Carlsson, B., Carmignac, D., & Robinson, I. C. (1996). The obese growth hormone (GH)-deficient dwarf rat: 259

Body fat responses to patterned delivery of GH and insulin-like growth factor-i. Endocrinology, 137(5), 1904-1912. Clemmons, D. R. (2006). Clinical utility of measurements of insulin-like growth factor 1. Nature clinical practice.Endocrinology & metabolism, 2(8), 436. Clemmons, D. R., Snyder, D. K., Williams, R., & Underwood, L. E. (1987). Growth hormone administration conserves lean body mass during dietary restriction in obese subjects. J Clin Endocrinol Metab, 64(5), 878-883. Colao, A., Marzullo, P., Di Somma, C., & Lombardi, G. (2001). Growth hormone and the heart. Clin Endocrinol (Oxf), 54(2), 137-154. Coleman, E. S., & Sartin, J. L. (1996). Endotoxin stimulates in vitro pituitary growth hormone release in eicosanoid-dependent manner. Am J Vet Res, 57(11), 1662- 1667. Colligan, P. B., Brown-Borg, H. M., Duan, J., Ren, B. H., & Ren, J. (2002). Cardiac contractile function is enhanced in isolated ventricular myocytes from growth hormone transgenic mice. J Endocrinol, 173(2), 257-264. Conrotto, P., & Souchelnytskyi, S. (2008). Proteomic approaches in biological and medical sciences: Principles and applications. Exp Oncol, 30(3), 171-180. Correa-Rotter, R., Hostetter, T. H., Manivel, J. C., Eddy, A. A., & Rosenberg, M. E. (1992). Intrarenal distribution of clusterin following reduction of renal mass. Kidney Int, 41(4), 938-950. Coschigano, K. T., Clemmons, D., Bellushi, L. L., & Kopchick, J. J. (2000). Assessment of growth parameters and life span of ghr/bp gene-disrupted mice. Endocrinology, 141, 2608. Coschigano, K. T., Holland, A. N., Riders, M. E., List, E. O., Flyvbjerg, A., & Kopchick, J. J. (2003). Deletion, but not antagonism, of the mouse growth hormone receptor results in severely decreased body weights, insulin, and insulin-like growth factor i levels and increased life span 10.1210/en.2003-0374. Endocrinology, 144(9), 3799-3810. Cotes, P. M., Reid, E., & Young, F. G. (1949). Diabetogenic action of pure anterior pituitary growth hormone. Nature, 164(4162), 209-211. Cummings, D. E., & Merriam, G. R. (1999). Age-related changes in growth hormone secretion: Should the somatopause be treated? Semin Reprod Endocrinol, 17(4), 311-325. Cuneo, R. C., Salomon, F., Watts, G. F., Hesp, R., & Sonksen, P. H. (1993). Growth hormone treatment improves serum lipids and lipoproteins in adults with growth hormone deficiency. Metabolism, 42(12), 1519-1523. Cunningham, B. C., Ultsch, M., De Vos, A. M., Mulkerrin, M. G., Clauser, K. R., & Wells, J. A. (1991). Dimerization of the extracellular domain of the human growth hormone receptor by a single hormone molecule. Science, 254(5033), 821-825. Dajee, M., Kazansky, A. V., Raught, B., Hocke, G. M., Fey, G. H., & Richards, J. S. (1996). Prolactin induction of the alpha 2-macroglobulin gene in rat ovarian granulosa cells: Stat 5 activation and binding to the interleukin-6 response element. Mol Endocrinol, 10(2), 171-184. 260

Dall, R., Longobardi, S., Ehrnborg, C., Keay, N., Rosen, T., Jorgensen, J. O., et al. (2000). The effect of four weeks of supraphysiological growth hormone administration on the insulin-like growth factor axis in women and men. Gh-2000 study group. The Journal of clinical endocrinology and metabolism, 85(11), 4193. database, S. D. (2009). Expasy - swiss-2dpage top page. Davignon, J. (2005). Apolipoprotein e and atherosclerosis: Beyond lipid effect. Arterioscler Thromb Vasc Biol, 25(2), 267-269. de Boer, J. P., Creasey, A. A., Chang, A., Abbink, J. J., Roem, D., Eerenberg, A. J., et al. (1993). Alpha-2-macroglobulin functions as an inhibitor of fibrinolytic, clotting, and neutrophilic proteinases in sepsis: Studies using a baboon model. Infect. Immun., 61(12), 5035-5043. de Groof, F., Joosten, K. F. M., Janssen, J. A. M. J. L., de Kleijn, E. D., Hazelzet, J. A., Hop, W. C. J., et al. (2002). Acute stress response in children with meningococcal sepsis: Important differences in the growth hormone/insulin-like growth factor i axis between nonsurvivors and survivors. J Clin Endocrinol Metab, 87(7), 3118- 3124. del Rincon, J.-P., Iida, K., Gaylinn, B. D., McCurdy, C. E., Leitner, J. W., Barbour, L. A., et al. (2007). Growth hormone regulation of p85î± expression and phosphoinositide 3-kinase activity in adipose tissue. Diabetes, 56(6), 1638-1646. Delaroudis, S. P., Efstathiadou, Z. A., Koukoulis, G. N., Kita, M. D., Farmakiotis, D., Dara, O. G., et al. (2008). Amelioration of cardiovascular risk factors with partial biochemical control of acromegaly. Clin Endocrinol (Oxf), 69(2), 279-284. Derfalvi, B., Igaz, P., Fulop, K. A., Szalai, C., & Falus, A. (2000). Interleukin-6-induced production of type ii acute phase proteins and expression of junb gene are downregulated by human recombinant growth hormone in vitro. Cell Biol Int, 24(2), 109-114. Dialynas, E., Brown-Borg, H., & Bartke, A. (1999). Immune function in transgenic mice overexpressing growth hormone (GH) releasing hormone, GH or GH antagonist. Proc Soc Exp Biol Med, 221(3), 178-183. Dilley, R. J., & Schwartz, S. M. (1989). Vascular remodeling in the growth hormone transgenic mouse. Circ Res, 65(5), 1233-1240. Ding, J., List, E. O., Okada, S., & Kopchick, J. J. (2009). Perspective: Proteomic approach to detect biomarkers of human growth hormone. Growth Hormone & IGF Research, In Press, Corrected Proof. Djoba Siawaya, J. F., Chegou, N. N., Heuvel, M. M., Diacon, A. H., Beyers, N., Helden, P. V., et al. (2009). Differential cytokine/chemokines and kl-6 profiles in patients with different forms of tuberculosis. Cytokine. Dobryszycka, W. (1997). Biological functions of haptoglobin--new pieces to an old puzzle. Eur J Clin Chem Clin Biochem, 35(9), 647-654. Doi, S. Q., Jacot, T. A., Sellitti, D. F., Hirszel, P., Hirata, M. H., Striker, G. E., et al. (2000). Growth hormone increases inducible nitric oxide synthase expression in mesangial cells. J Am Soc Nephrol, 11(8), 1419-1425. Doi, T., Striker, L. J., Gibson, C. C., Agodoa, L. Y., Brinster, R. L., & Striker, G. E. (1990). Glomerular lesions in mice transgenic for growth hormone and insulinlike 261

growth factor-i. I. Relationship between increased glomerular size and mesangial sclerosis. Am J Pathol, 137(3), 541-552. Doi, T., Striker, L. J., Kimata, K., Peten, E. P., Yamada, Y., & Striker, G. E. (1991). Glomerulosclerosis in mice transgenic for growth hormone. Increased mesangial extracellular matrix is correlated with kidney mrna levels. J Exp Med, 173(5), 1287-1290. Doi, T., Striker, L. J., Quaife, C., Conti, F. G., Palmiter, R., Behringer, R., et al. (1988). Progressive glomerulosclerosis develops in transgenic mice chronically expressing growth hormone and growth hormone releasing factor but not in those expressing insulinlike growth factor-1. Am J Pathol, 131(3), 398-403. Dominici, F. P., Arostegui Diaz, G., Bartke, A., Kopchick, J. J., & Turyn, D. (2000). Compensatory alterations of insulin signal transduction in liver of growth hormone receptor knockout mice. J Endocrinol, 166(3), 579-590. Dominici, F. P., Cifone, D., Bartke, A., & Turyn, D. (1999a). Alterations in the early steps of the insulin-signaling system in skeletal muscle of GH-transgenic mice. Am J Physiol Endocrinol Metab, 277(3), E447-454. Dominici, F. P., Cifone, D., Bartke, A., & Turyn, D. (1999b). Loss of sensitivity to insulin at early events of the insulin signaling pathway in the liver of growth hormone-transgenic mice. J Endocrinol, 161(3), 383-392. Dowling, D. K., & Simmons, L. W. (2009). Reactive oxygen species as universal constraints in life-history evolution. Proc Biol Sci, 276(1663), 1737-1745. Duan, X., Yarmush, D. M., Berthiaume, F., Jayaraman, A., & Yarmush, M. L. (2004). A mouse serum two-dimensional gel map: Application to profiling burn injury and infection. Electrophoresis, 25(17), 3055-3065. Dubuis, J. M., Deal, C., Tsagaroulis, P., Clark, R. G., & Van Vliet, G. (1996). Effects of 14-day infusions of growth hormone and/or insulin-like growth factor i on the obesity of growing zucker rats. Endocrinology, 137(7), 2799-2806. Duverger, N., Tremp, G., Caillaud, J. M., Emmanuel, F., Castro, G., Fruchart, J. C., et al. (1996). Protection against atherogenesis in mice mediated by human apolipoprotein a-iv. Science, 273(5277), 966-968. Dworacka, M., & Winiarska, H. (2005). The application of plasma 1,5-anhydro-d- glucitol for monitoring type 2 diabetic patients. Dis Markers, 21(3), 127-132. Echan, L. A., Tang, H. Y., Ali-Khan, N., Lee, K., & Speicher, D. W. (2005). Depletion of multiple high-abundance proteins improves protein profiling capacities of human serum and plasma. Proteomics, 5(13), 3292. Egecioglu, E., Andersson, I. J., Bollano, E., Palsdottir, V., Gabrielsson, B. G., Kopchick, J. J., et al. (2007). Growth hormone receptor deficiency in mice results in reduced systolic blood pressure and plasma renin, increased aortic enos expression, and altered cardiovascular structure and function. Am J Physiol Endocrinol Metab, 292(5), E1418-1425. Egecioglu, E., Bjursell, M., Ljungberg, A., Dickson, S. L., Kopchick, J. J., Bergstrom, G., et al. (2006). Growth hormone receptor deficiency results in blunted ghrelin feeding response, obesity, and hypolipidemia in mice. AJP - Endocrinology and Metabolism, 290(2), E317. 262

Egecioglu, E., Bjursell, M., Ljungberg, A., Dickson, S. L., Kopchick, J. J., Bergstrom, G., et al. (2005). Growth hormone receptor deficiency results in blunted ghrelin feeding response, obesity and hypolipidemia in mice. Am J Physiol Endocrinol Metab. Ehrnborg, C., Ellegard, L., Bosaeus, I., Bengtsson, B. A., & Rosen, T. (2005). Supraphysiological growth hormone: Less fat, more extracellular fluid but uncertain effects on muscles in healthy, active young adults. Clinical endocrinology, 62(4), 449. Ehrnborg, C., Lange, K. H., Dall, R., Christiansen, J. S., Lundberg, P. A., Baxter, R. C., et al. (2003). The growth hormone/insulin-like growth factor-i axis hormones and bone markers in elite athletes in response to a maximum exercise test. The Journal of clinical endocrinology and metabolism, 88(1), 394. Erotokritou-Mulligan, I., Bassett, E. E., Bartlett, C., Cowan, D., McHugh, C., Seah, R., et al. (2008a). The effect of sports injury on insulin like growth factor-i and procollagen iii peptide: Implications for detection of growth hormone abuse in athletes. The Journal of clinical endocrinology and metabolism. Erotokritou-Mulligan, I., Bassett, E. E., Cowan, D., Bartlett, C., McHugh, C., Sonksen, P. H., et al. (2008b). The influence of ethnicity on insulin like growth factor-i and procollagen iii peptide in elite athletes and its effect on the ability to detect GH abuse. Clinical endocrinology. Erotokritou-Mulligan, I., Bassett, E. E., Cowan, D. A., Bartlett, C., McHugh, C., Sonksen, P. H., et al. (2009). Influence of ethnicity on IGF-I and procollagen iii peptide (p-iii-p) in elite athletes and its effect on the ability to detect GH abuse. Clin Endocrinol (Oxf), 70(1), 161-168. Erotokritou-Mulligan, I., Bassett, E. E., Kniess, A., Sonksen, P. H., & Holt, R. I. (2007). Validation of the growth hormone (GH)-dependent marker method of detecting GH abuse in sport through the use of independent data sets. Growth hormone & IGF research: official journal of the Growth Hormone Research Society and the International IGF Research Society. Ezeh, B., Haiman, M., Alber, H. F., Kunz, B., Paulweber, B., Lingenhel, A., et al. (2003). Plasma distribution of apoA-iv in patients with coronary artery disease and healthy controls. J. Lipid Res., 44(8), 1523-1529. Fanciulli, G., Delitala, A., & Delitala, G. (2009). Growth hormone, menopause and ageing: No definite evidence for 'rejuvenation' with growth hormone. Hum Reprod Update, 15(3), 341-358. Fernandez, M. L., & Volek, J. S. (2006). Guinea pigs: A suitable animal model to study lipoprotein metabolism, atherosclerosis and inflammation. Nutr Metab (Lond), 3, 17. Ferrucci, L., Corsi, A., Lauretani, F., Bandinelli, S., Bartali, B., Taub, D. D., et al. (2005). The origins of age-related proinflammatory state. Blood, 105(6), 2294-2299. Fielder, P. J., Mortensen, D. L., Mallet, P., Carlsson, B., Baxter, R. C., & Clark, R. G. (1996). Differential long-term effects of insulin-like growth factor-I (IGF-I) growth hormone (GH), and IGF-I plus GH on body growth and IGF binding proteins in hypophysectomized rats. Endocrinology, 137(5), 1913-1920. 263

Finkel, T., & Holbrook, N. J. (2000). Oxidants, oxidative stress and the biology of ageing. Nature, 408, 239. Flanagan, D. E., Taylor, M. C., Parfitt, V., Mardell, R., Wood, P. J., & Leatherdale, B. A. (1997). Urinary growth hormone following exercise to assess growth hormone production in adults. Clinical endocrinology, 46(4), 425. Fleck, A. (1989). Clinical and nutritional aspects of changes in acute-phase proteins during inflammation. Proceedings of the Nutrition Society, 48(03), 347-354. Flint, D. J., Binart, N., Boumard, S., Kopchick, J. J., & Kelly, P. (2006). Developmental aspects of adipose tissue in GH receptor and prolactin receptor gene disrupted mice: Site-specific effects upon proliferation, differentiation and hormone sensitivity. J Endocrinol, 191(1), 101-111. Flyvbjerg, A. (1990). Growth factors and diabetic complications. Diabet Med, 7(5), 387- 399. Folkmann, J. K., Loft, S., & Moller, P. (2007). Oxidatively damaged DNA in aging dyslipidemic apoE-/- and wild-type mice. Mutagenesis, 22(2), 105-110. Frank, S. J., Gilliland, G., Kraft, A. S., & Arnold, C. S. (1994). Interaction of the growth hormone receptor cytoplasmic domain with the jak2 tyrosine kinase. Endocrinology, 135(5), 2228-2239. French, L. E., Chonn, A., Ducrest, D., Baumann, B., Belin, D., Wohlwend, A., et al. (1993). Murine clusterin: Molecular cloning and mrna localization of a gene associated with epithelial differentiation processes during embryogenesis. J Cell Biol, 122(5), 1119-1130. Frick, F., Bohlooly-Y, M., Linden, D., Olsson, B., Tornell, J., Eden, S., et al. (2001a). Long-term growth hormone excess induces marked alterations in lipoprotein metabolism in mice. AJP - Endocrinology and Metabolism, 281(6), E1230. Frick, F., Bohlooly-Y, M., Linden, D., Olsson, B., Tornell, J., Eden, S., et al. (2001b). Long-term growth hormone excess induces marked alterations in lipoprotein metabolism in mice. Am J Physiol Endocrinol Metab, 281(6), E1230-1239. Frick, F., Linden, D., Ameen, C., Eden, S., Mode, A., & Oscarsson, J. (2002). Interaction between growth hormone and insulin in the regulation of lipoprotein metabolism in the rat. Am J Physiol Endocrinol Metab, 283(5), E1023-1031. Frikke-Schmidt, R., Sing, C. F., Nordestgaard, B. G., & Tybjaerg-Hansen, A. (2004). Gender- and age-specific contributions of additional DNA sequence variation in the 5' regulatory region of the apoE gene to prediction of measures of lipid metabolism. Hum Genet, 115(4), 331-345. Fryburg, D. A., & Barrett, E. J. (1993). Growth hormone acutely stimulates skeletal muscle but not whole-body protein synthesis in humans. Metabolism, 42(9), 1223. Fuhrman, M. P., Charney, P., & Mueller, C. M. (2004). Hepatic proteins and nutrition assessment. Journal of the American Dietetic Association, 104(8), 1258. Galasko, D., Salmon, D., Gamst, A., Olichney, J., Thal, L. J., Silbert, L., et al. (2007). Prevalence of dementia in chamorros on guam: Relationship to age, gender, education, and apoE. Neurology, 68(21), 1772-1781. Garlaschi, C., di Natale, B., del Guercio, M. J., Caccamo, A., Gargantini, L., & Chiumello, G. (1975). Effect of physical exercise on secretion of growth 264

hormone, glucagon, and cortisol in obese and diabetic children. Diabetes, 24(8), 758-761. Gazzana, G., & Borlak, J. r. (2008). Mapping of the serum proteome of hepatocellular carcinoma induced by targeted overexpression of epidermal growth factor to liver cells of transgenic mice. Journal of Proteome Research, 7(3), 928. Gerhardt, W., & Ljungdahl, L. (1998). Troponin t: A sensitive and specific diagnostic and prognostic marker of myocardial damage. Clinica chimica acta; international journal of clinical chemistry, 272(1), 47. Ghebremedhin, E., Schultz, C., Thal, D. R., Rub, U., Ohm, T. G., Braak, E., et al. (2001). Gender and age modify the association between apoE and ad-related neuropathology. Neurology, 56(12), 1696-1701. Ghesquiere, B., Van Damme, J., Martens, L., Vandekerckhove, J., & Gevaert, K. (2006). Proteome-wide characterization of n-glycosylation events by diagonal chromatography. J Proteome Res, 5(9), 2438-2447. Ghiggeri, G. M., Bruschi, M., Candiano, G., Rastaldi, M. P., Scolari, F., Passerini, P., et al. (2002). Depletion of clusterin in renal diseases causing nephrotic syndrome. Kidney Int, 62(6), 2184-2194. Giani, J. F., Bonkowski, M. S., Munoz, M. C., Masternak, M. M., Turyn, D., Bartke, A., et al. (2008). Insulin signaling cascade in the hearts of long-lived growth hormone receptor knockout mice: Effects of calorie restriction. J Gerontol A Biol Sci Med Sci, 63(8), 788-797. Gillum, R. F. (2000). Assessment of serum albumin concentration as a risk factor for stroke and coronary disease in african americans and whites. J Natl Med Assoc, 92(1), 3-9. Goldberg, S. (2008). Mechanical/physical methods of cell disruption and tissue homogenization. Methods in molecular biology (Clifton, N.J.), 424, 3. Goldenberg, N., & Barkan, A. (2007). Factors regulating growth hormone secretion in humans. Endocrinology and metabolism clinics of North America, 36(1), 37. Gom, I., Fukushima, H., Shiraki, M., Miwa, Y., Ando, T., Takai, K., et al. (2007). Relationship between serum albumin level and aging in community-dwelling self- supported elderly population. J Nutr Sci Vitaminol (Tokyo), 53(1), 37-42. Honda, Y., & Honda, S. (2002). Oxidative stress and life span determination in the nematode caenorhabditis elegans. Ann.NY Acad.Sci., 959, 466. Gonzalo, J. A., Mazuchelli, R., Mellado, M., Frade, J. M., Carrera, A. C., von Kobbe, C., et al. (1996). Enterotoxin septic shock protection and deficient t helper 2 cytokine production in growth hormone transgenic mice. J Immunol, 157(8), 3298-3304. Gorg, A., Postel, W., & Gunther, S. (1988). The current state of two-dimensional electrophoresis with immobilized ph gradients. Electrophoresis, 9(9), 531-546. Granger, J., Siddiqui, J., Copeland, S., & Remick, D. (2005). Albumin depletion of human plasma also removes low abundance proteins including the cytokines. Proteomics, 5(18), 4713. Gravholt, C. H., Leth-Larsen, R., Lauridsen, A. L., Thiel, S., Hansen, T. K., Holmskov, U., et al. (2004). The effects of GH and hormone replacement therapy on serum 265

concentrations of mannan-binding lectin, surfactant protein d and vitamin d binding protein in turner syndrome. Eur J Endocrinol, 150(3), 355-362. Green, H., Morikawa, M., & Nixon, T. (1985). A dual effector theory of growth-hormone action. Differentiation, 29(3), 195-198. Gromadzka, G., Baranska-Gieruszczak, M., Sarzynska-Dlugosz, I., Ciesielska, A., & Czlonkowska, A. (2007). The apoE polymorphism and 1-year outcome in ischemic stroke: Genotype-gender interaction. Acta Neurol Scand, 116(6), 392- 398. Gromov, P., Skovgaard, G. L., Palsdottir, H., Gromova, I., Ostergaard, M., & Celis, J. E. (2003). Protein profiling of the human epidermis from the elderly reveals upregulation of a signature of interferon-{gamma}-induced polypeptides that includes manganese-superoxide dismutase and the p85{beta} subunit of phosphatidylinositol 3-kinase. Mol Cell Proteomics, 2(2), 70-84. Guerrier, L., Righetti, P. G., & Boschetti, E. (2008). Reduction of dynamic protein concentration range of biological extracts for the discovery of low-abundance proteins by means of hexapeptide ligand library. Nature protocols, 3(5), 883. Guha, N., Sönksen, P. H., & Holt, R. I. G. (2009). Igf-i abuse in sport: Current knowledge and future prospects for detection. Growth Hormone & IGF Research, In Press, Corrected Proof. Guipaud, O., Holler, V., Buard, V., Tarlet, G., Royer, N., Vinh, J., et al. (2007). Time- course analysis of mouse serum proteome changes following exposure of the skin to ionizing radiation. Proteomics, 7(21), 3992-4002. Guo, Y., Lu, Y., Houle, D., Robertson, K., Tang, Z., Kopchick, J. J., et al. (2005). Pancreatic islet-specific expression of an insulin-like growth factor-i transgene compensates islet cell growth in growth hormone receptor gene-deficient mice. Endocrinology, 146(6), 2602-2609. Gutiérrez-Gallego, R., Bosch, J., Such-Sanmartín, G., & Segura, J. (2009). Surface plasmon resonance immuno assays - a perspective. Growth Hormone & IGF Research, In Press, Corrected Proof. Guy-Grand, B., & Bour, H. (1972). [the possible role of adipose tissue in insulin resistance and in the physiopathology of obese diabetics]. Journ Annu Diabetol Hotel Dieu, 13(0), 81-92. Haider, D. G., Schindler, K., Prager, G., Bohdjalian, A., Luger, A., Wolzt, M., et al. (2007). Serum retinol-binding protein 4 is reduced after weight loss in morbidly obese subjects. Journal of Clinical Endocrinology Metabolism, 92(3), 1168. Hall, M. A., Bartke, A., & Martinko, J. M. (2002). Humoral immune response in mice over-expressing or deficient in growth hormone. Experimental Biology and Medicine, 227(7), 535-544. Halliwell, B., & Gutteridge, J. M. C. (1990). The antioxidants of human extracellular fluids. Archives of Biochemistry and Biophysics, 280(1), 1. Han, Y.-H., Kwon, J.-H., Yu, D.-Y., & Moon, E.-Y. (2006). Inhibitory effect of peroxiredoxin ii (prx ii) on rasâ€“erkâ€“nfîºb pathway in mouse embryonic fibroblast (mef) senescence. Free Radical Research, 40(11), 1182 - 1189. 266

Hansen, L. H., Wang, X., Kopchick, J. J., Bouchelouche, P., Nielsen, J. H., Galsgaard, E. D., et al. (1996). Identification of tyrosine residues in the intracellular domain of the growth hormone receptor required for transcriptional signaling and stat5 activation. J Biol Chem, 271(21), 12669-12673. Hansen, T. K. (2003). Growth hormone and mannan-binding lectin: Emerging evidence for hormonal regulation of humoral innate immunity. Minerva Endocrinol, 28(1), 75-84. Hansen, T. K., Thiel, S., Dall, R., Rosenfalck, A. M., Trainer, P., Flyvbjerg, A., et al. (2001). Gh strongly affects serum concentrations of mannan-binding lectin: Evidence for a new IGF-I independent immunomodulatory effect of GH. J Clin Endocrinol Metab, 86(11), 5383-5388. Hasenfuss, G. (1998). Animal models of human cardiovascular disease, heart failure and hypertrophy. Cardiovasc Res, 39(1), 60-76. Hao, Z. M., Yang, X., Ying, W. T., Wang, J., Qian, X. H., & Huang, C. F. (2002). [analysis of the proteomic changes of the serum of the smad3 targeted deficient mice]. Sheng Wu Gong Cheng Xue Bao, 18(4), 452-456. Harding, M. A., Chadwick, L. J., Gattone, V. H., 2nd, & Calvet, J. P. (1991). The sgp-2 gene is developmentally regulated in the mouse kidney and abnormally expressed in collecting duct cysts in polycystic kidney disease. Dev Biol, 146(2), 483-490. Hardy, C. L., Bhathal, P. S., Snibson, K. J., & Adams, T. E. (1997). Comparison of intrahepatic lymphocytes from normal and growth hormone transgenic mice with chronic hepatitis and liver cancer. Immunology, 90(3), 412-420. Hashimoto, Y., Kamioka, T., Hosaka, M., Mabuchi, K., Mizuchi, A., Shimazaki, Y., et al. (2000). Exogenous 20k growth hormone (GH) suppresses endogenous 22k GH secretion in normal men. J Clin Endocrinol Metab, 85(2), 601-606. Hattori, N. (2009). Expression, regulation and biological actions of growth hormone (GH) and ghrelin in the immune system. Growth Hormone & IGF Research, 19(3), 187. Hauck, S. J., Aaron, J. M., Wright, C., Kopchick, J. J., & Bartke, A. (2002). Antioxidant enzymes, free-radical damage, and response to paraquat in liver and kidney of long-living growth hormone receptor/binding protein gene-disrupted mice. Horm Metab Res, 34(9), 481-486. Hauck, S. J., & Bartke, A. (2001). Free radical defenses in the liver and kidney of human growth hormone transgenic mice: Possible mechanisms of early mortality. J Gerontol A Biol Sci Med Sci, 56(4), B153-162. Hauck, S. J., Hunter, W. S., Danilovich, N., Kopchick, J. J., & Bartke, A. (2001). Reduced levels of thyroid hormones, insulin, and glucose, and lower body core temperature in the growth hormone receptor/binding protein knockout mouse. Exp Biol Med (Maywood), 226(6), 552-558. Hayden, M. S., & Ghosh, S. (2008). Shared principles in nf-[kappa]b signaling. Cell, 132(3), 344. He, H., McCartney, D. J., Wei, Q., Esadeg, S., Zhang, J., Foster, R. A., et al. (2005). Characterization of a murine alpha 2 macroglobulin gene expressed in reproductive and cardiovascular tissue. Biol Reprod, 72(2), 266-275. 267

Herbert, B., Galvani, M., Hamdan, M., Olivieri, E., MacCarthy, J., Pedersen, S., et al. (2001). Reduction and alkylation of proteins in preparation of two-dimensional map analysis: Why, when, and how? Electrophoresis, 22(10), 2046-2057. Herrington, J., & Carter-Su, C. (2001). Signaling pathways activated by the growth hormone receptor. Trends in Endocrinology and Metabolism, 12(6), 252. Hidaka, S., Kranzlin, B., Gretz, N., & Witzgall, R. (2002). Urinary clusterin levels in the rat correlate with the severity of tubular damage and may help to differentiate between glomerular and tubular injuries. Cell Tissue Res, 310(3), 289-296. Ho, K. K. Y. (2007). Consensus guidelines for the diagnosis and treatment of adults with GH deficiency ii: A statement of the GH research society in association with the european society for pediatric endocrinology, lawson wilkins society, european society of endocrinology, japan endocrine society, and endocrine society of australia. Eur J Endocrinol, 157(6), 695-700. Holl, R. W., Hartman, M. L., Veldhuis, J. D., Taylor, W. M., & Thorner, M. O. (1991). Thirty-second sampling of plasma growth hormone in man: Correlation with sleep stages. J Clin Endocrinol Metab, 72(4), 854-861. Holl, R. W., Schwarz, U., Schauwecker, P., Benz, R., Veldhuis, J. D., & Heinze, E. (1993). Diurnal variation in the elimination rate of human growth hormone (GH): The half-life of serum GH is prolonged in the evening, and affected by the source of the hormone, as well as by body size and serum estradiol. The Journal of clinical endocrinology and metabolism, 77(1), 216. Holt, R. I. G., Bassett, E. E., Erotokritou-Mulligan, I., McHugh, C., Cowan, D., Bartlett, C., et al. (2009a). Moving one step closer to catching the GH cheats: The GH- 2004 experience. Growth Hormone & IGF Research, In Press, Corrected Proof. Holt, R. I. G., Erotokritou-Mulligan, I., & Sönksen, P. H. (2009b). The history of doping and growth hormone abuse in sport. Growth Hormone & IGF Research, In Press, Corrected Proof. Holzenberger, M., Dupont, J., Ducos, B., Leneuve, P., Geloen, A., Even, P. C., et al. (2003). Igf-1 receptor regulates lifespan and resistance to oxidative stress in mice. Nature, 421(6919), 182. Hood, B. L., Zhou, M., Chan, K. C., Lucas, D. A., Kim, G. J., Issaq, H. J., et al. (2005). Investigation of the mouse serum proteome. J.Proteome Res., 4(5), 1561. Hsueh-Fen, J., Jenn-Han, C., Wei-Tse, H., Su-Ching, H., Shui-Tein, C., John Yi-Chung, L., et al. (2004). Identification of tumor-associated plasma biomarkers using proteomic techniques: From mouse to human. PROTEOMICS, 4(9), 2766-2775. Huang, P. L. (2009). A comprehensive definition for metabolic syndrome. Dis Model Mech, 2(5-6), 231-237. Huang, X., Wei, Y., Li, L., Wen, Y., Yang, J., Liu, B., et al. (2006). Serum proteomics study of the squamous cell carcinoma antigen 1 in tongue cancer. Oral Oncology, 42(1), 25. Hutchinson, W. L., Alves, I. L., Saraiva, M. J., & Pepys, M. B. (1996). Post translational modification of serum ttr. Neuromuscular Disorders, 6(Supplement 1), S16. Ichihara, K., Itoh, Y., Min, W. K., Yap, S. F., Lam, C. W., Kong, X. T., et al. (2004). Diagnostic and epidemiological implications of regional differences in serum 268

concentrations of proteins observed in six asian cities. Clin Chem Lab Med, 42(7), 800-809. Ikeno, Y., Hubbard, G. B., Lee, S., Cortez, L. A., Lew, C. M., Webb, C. R., et al. (2009). Reduced incidence and delayed occurrence of fatal neoplastic diseases in growth hormone receptor/binding protein knockout mice. J Gerontol A Biol Sci Med Sci, 64(5), 522-529. Ingenbleek, Y., & Young, V. (1994). Transthyretin (prealbumin) in health and disease: Nutritional implications. Annual Review of Nutrition, 14, 495. Initial sequencing and comparative analysis of the mouse genome. (2002). Nature, 420(6915), 520. Isaksson, O. G., Jansson, J. O., & Gause, I. A. (1982). Growth hormone stimulates longitudinal bone growth directly. Science, 216(4551), 1237-1239. Ishikawa, T., Zhu, B. L., Li, D. R., Zhao, D., Michiue, T., & Maeda, H. (2006). Age- dependent increase of clusterin in the human pituitary gland. Leg Med (Tokyo), 8(1), 28-33. Itoh, N., Hanafusa, T., Miyagawa, J., Tamura, S., Inada, M., Kawata, S., et al. (1992). Transthyretin (prealbumin) in the pancreas and sera of newly diagnosed type i (insulin-dependent) diabetic patients. The Journal of clinical endocrinology and metabolism, 74(6), 1372. Izzard, A. S., Emerson, M., Prehar, S., Neyses, L., Trainer, P., List, E. O., et al. (2009). The cardiovascular phenotype of a mouse model of acromegaly. Growth Hormone & IGF Research, In Press, Corrected Proof. Jacobsson, B., Carlstrom, A., Platz, A., & Collins, V. P. (1990). Transthyretin messenger ribonucleic acid expression in the pancreas and in endocrine tumors of the pancreas and gut. The Journal of clinical endocrinology and metabolism, 71(4), 875. James, P. (1997). Of genomes and proteomes. Biochemical and Biophysical Research Communications, 231(1), 1. Jeay, S., Sonenshein, G. E., Postel-Vinay, M.-C., Kelly, P. A., & Baixeras, E. (2002). Growth hormone can act as a cytokine controlling survival and proliferation of immune cells: New insights into signaling pathways. Molecular and Cellular Endocrinology, 188(1-2), 1. Johannsson, G., Oscarsson, J., Rosen, T., Wiklund, O., Olsson, G., Wilhelmsen, L., et al. (1995). Effects of 1 year of growth hormone therapy on serum lipoprotein levels in growth hormone–deficient adults: Influence of gender and apo(a) and apoE phenotypes. Arterioscler Thromb Vasc Biol, 15(12), 2142-2150. Jordan-Starck, T. C., Lund, S. D., Witte, D. P., Aronow, B. J., Ley, C. A., Stuart, W. D., et al. (1994). Mouse apolipoprotein j: Characterization of a gene implicated in atherosclerosis. Journal of Lipid Research, 35(2), 194. Karlamangla, A. S., Singer, B. H., Reuben, D. B., & Seeman, T. E. (2004). Increases in serum non-high-density lipoprotein cholesterol may be beneficial in some high- functioning older adults: Macarthur studies of successful aging. Journal of the American Geriatrics Society, 52(4), 487. 269

Katz, P. R., Karuza, J., Gutman, S. I., Bartholomew, W., & Richman, G. (1990). A comparison between erythrocyte sedimentation rate (esr) and selected acute-phase proteins in the elderly. Am J Clin Pathol, 94(5), 637-640. Katznelson, L. (2008). Alterations in body composition in acromegaly. Pituitary. Keene, D. E., Suescun, M. O., Bostwick, M. G., Chandrashekar, V., Bartke, A., & Kopchick, J. J. (2002). Puberty is delayed in male growth hormone receptor gene- disrupted mice. J Androl, 23(5), 661-668. Kelder, B., Berryman, D. E., Clark, R., Li, A., List, E. O., & Kopchick, J. J. (2007). Cide- a gene expression is decreased in white adipose tissue of growth hormone receptor/binding protein gene disrupted mice and with high-fat feeding of normal mice. Growth Hormone & IGF Research, 17(4), 346. Kelley, K. W., Weigent, D. A., & Kooijman, R. (2007). Protein hormones and immunity. Brain Behav Immun, 21(4), 384-392. Kemp, S. F., & Canfield, M. E. (1983). Acute effects of growth hormone administration: Vitamin a and visceral protein concentrations. Acta Endocrinol (Copenh), 104(4), 390-396. Kemp, S. F., Rosenfeld, R. G., Liu, F., Gaspich, S., & Hintz, R. L. (1981). Acute somatomedin response to growth hormone: Radioreceptor assay versus radioimmunoassay. J Clin Endocrinol Metab, 52(4), 616-621. Kineman, R. D., Teixeira, L. T., Amargo, G. V., Coschigano, K. T., Kopchick, J. J., & Frohman, L. A. (2001). The effect of ghrh on somatotrope hyperplasia and tumor formation in the presence and absence of GH signaling. Endocrinology, 142(9), 3764-3773. Kinney, B. A., Coschigano, K. T., Kopchick, J. J., Steger, R. W., & Bartke, A. (2001). Evidence that age-induced decline in memory retention is delayed in growth hormone resistant GH-R-ko (laron) mice. Physiol Behav, 72(5), 653-660. Kireev, R., Tresguerres, A., Garcia, C., Borras, C., Ariznavarreta, C., Vara, E., et al. (2009). Hormonal regulation of pro-inflammatory and lipid peroxidation processes in liver of old ovariectomized female rats. Biogerontology. Kisilevsky, R., & Subrahmanyan, L. (1992). Serum amyloid a changes high density lipoprotein's cellular affinity. A clue to serum amyloid a's principal function. Lab Invest, 66(6), 778-785. Klag, M. J., Ford, D. E., Mead, L. A., He, J., Whelton, P. K., Liang, K. Y., et al. (1993). Serum cholesterol in young men and subsequent cardiovascular disease. The New England journal of medicine, 328(5), 313. Klose, J. (1975). Protein mapping by combined isoelectric focusing and electrophoresis of mouse tissues. A novel approach to testing for induced point mutations in mammals. Humangenetik, 26(3), 231-243. Kloting, N., Graham, T. E., Berndt, J., Kralisch, S., Kovacs, P., Wason, C. J., et al. (2007). Serum retinol-binding protein is more highly expressed in visceral than in subcutaneous adipose tissue and is a marker of intra-abdominal fat mass. Cell Metab, 6(1), 79-87. 270

Knapp, J. R., Chen, W. Y., Turner, N. D., Byers, F. M., & Kopchick, J. J. (1994). Growth patterns and body composition of transgenic mice expressing mutated bovine somatotropin genes. J. Anim Sci., 72(11), 2812-2819. Kniess, A., Ziegler, E., Kratzsch, J., Thieme, D., & Muller, R. K. (2003). Potential parameters for the detection of hgh doping. Analytical and bioanalytical chemistry, 376(5), 696. Kopchick, J. J. (2003). Discovery and development of a new class of drugs: Gh antagonists. J Endocrinol Invest, 26(10 Suppl), 16. Kopchick, J. J., & Andry, J. M. (2000). Growth hormone (GH), GH receptor, and signal transduction. Mol Genet Metab, 71(1-2), 293-314. Kopchick, J. J., Bellush, L. L., & Coschigano, K. T. (1999). Transgenic models of growth hormone action. Annual Review of Nutrition, 19(1), 437. Kopchick, J. J., & Laron, Z. (1999). Is the laron mouse an accurate model of laron syndrome? Mol Genet Metab, 68(2), 232-236. Kopchick, J. J., List, E. O., Kohn, D. T., Keidan, G. M., Qiu, L., & Okada, S. (2002). Perspective: Proteomics--see "spots" run. Endocrinology, 143(6), 1990. Kowalska, I., Straczkowski, M., Adamska, A., Nikolajuk, A., Karczewska-Kupczewska, M., Otziomek, E., et al. (2008). Serum retinol binding protein 4 is related to insulin resistance and nonoxidative glucose metabolism in lean and obese women with normal glucose tolerance. Journal of Clinical Endocrinology Metabolism, 93(7), 2786. Kraemer, W. J., Vingren, J. L., Schuenke, M. D., Kopchick, J. J., Volek, J. S., Fragala, M. S., et al. (2008). Effect of circulating growth hormone on muscle IGF-I protein concentration in female mice with growth hormone receptor gene disruption. Growth Horm IGF Res. Kratzsch, J., Selisko, T., & Birkenmeier, G. (1995). Identification of transformed alpha 2- macroglobulin as a growth hormone- binding protein in human blood. J Clin Endocrinol Metab, 80(2), 585-590. Kratzsch, J., Selisko, T., & Birkenmeier, G. (1996). Transformed alpha 2-macroglobulin as a low-affinity growth hormone-binding protein. Acta Paediatr Suppl, 417, 108- 110. Kronenberg, F., Stuhlinger, M., Trenkwalder, E., Geethanjali, F. S., Pachinger, O., von Eckardstein, A., et al. (2000). Low apolipoprotein a-iv plasma concentrations in men with coronary artery disease. J Am Coll Cardiol, 36(3), 751-757. Kronmal, R. A., Cain, K. C., Ye, Z., & Omenn, G. S. (1993). Total serum cholesterol levels and mortality risk as a function of age. A report based on the framingham data. Archives of Internal Medicine, 153(9), 1065. Krulich, L., Hefco, E., Illner, P., & Read, C. B. (1974). The effects of acute stress on the secretion of LH, FSH, prolactin and GH in the normal male rat, with comments on their statistical evaluation. Neuroendocrinology, 16(5-6), 293-311. Krysiak, R., Gdula-Dymek, A., Bednarska-Czerwinska, A., & Okopien, B. (2007). Growth hormone therapy in children and adults. Pharmacol Rep, 59(5), 500-516. 271

Kurahashi, T., Muramaki, M., Yamanaka, K., Hara, I., & Miyake, H. (2005). Expression of the secreted form of clusterin protein in renal cell carcinoma as a predictor of disease extension. BJU Int, 96(6), 895-899. Kurata, H., Sannoh, T., Kozutsumi, Y., Yokota, Y., & Kawasaki, T. (1994). Structure and function of mannan-binding proteins isolated from human liver and serum. J Biochem, 115(6), 1148-1154. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage t4. Nature, 227(5259), 680-685. Lane, M. A., Ingram, D. K., Ball, S. S., & Roth, G. S. (1997). Dehydroepiandrosterone sulfate: A biomarker of primate aging slowed by calorie restriction. J Clin Endocrinol Metab, 82(7), 2093-2096. Lang, C. H., Pollard, V., Fan, J., Traber, L. D., Traber, D. L., Frost, R. A., et al. (1997). Acute alterations in growth hormone-insulin-like growth factor axis in humans injected with endotoxin. Am J Physiol, 273(1 Pt 2), R371-378. Lanning, N. J., & Carter-Su, C. (2006). Recent advances in growth hormone signaling. Rev Endocr Metab Disord, 7(4), 225-235. Laron, Z. (1984). Laron-type dwarfism (hereditary somatomedin deficiency): A review. Ergeb Inn Med Kinderheilkd, 51, 117-150. Laron, Z., Ginsberg, S., Lilos, P., Arbiv, M., & Vaisman, N. (2006). Body composition in untreated adult patients with laron syndrome (primary GH insensitivity). Clin Endocrinol (Oxf), 65(1), 114-117. Lasztity, N., Biro, L., Nemeth, E., Pap, A., & Antal, M. (2002). Protein status in pancreatitis--transthyretin is a sensitive biomarker of malnutrition in acute and chronic pancreatitis. Clinical chemistry and laboratory medicine: CCLM / FESCC, 40(12), 1320. Lauber, W. M., Carroll, J. A., Dufield, D. R., Kiesel, J. R., Radabaugh, M. R., & Malone, J. P. (2001). Mass spectrometry compatibility of two-dimensional gel protein stains. Electrophoresis, 22(5), 906. Leppaluoto, J., Laisi, U., Lybeck, H., Partanen, J., Ranta, T., & Virkkunen. (1975). Pulsatile secretion of ACTH, GH, LH and TSH in man. Acta Physiol Scand, 95(4), 470-476. LeRoith, D., & Yakar, S. (2007). Mechanisms of disease: Metabolic effects of growth hormone and insulin-like growth factor 1. Nature clinical practice.Endocrinology & metabolism, 3(3), 302. Leslie, M. (2008). Aging: Searching for the secrets of the super old. Science, 321(5897), 1764. Leung, K.-C., Howe, C., Gui, L. Y. Y., Trout, G., Veldhuis, J. D., & Ho, K. K. Y. (2002). Physiological and pharmacological regulation of 20-kda growth hormone. American Journal Of Physiology. Endocrinology And Metabolism, 283(4), E836. Li, H., Xing, X., Ding, G., Li, Q., Wang, C., Xie, L., et al. (2009). Sysptm - a systematic resource for proteomic research of post-translational modifications. Mol Cell Proteomics. 272

Li, W., Zhang, Y. Q., Wang, G. J., Wang, J., & Zhang, X. M. (2007). [serum proteome in mice after low dose radiation]. Zhongguo Shi Yan Xue Ye Xue Za Zhi, 15(1), 191- 194. Lichanska, A. M., & Waters, M. J. (2008). How growth hormone controls growth, obesity and sexual dimorphism. Trends Genet, 24(1), 41-47. List, E. O., Berryman, D. E., Palmer, A. J., Qiu, L., Sankaran, S., Kohn, D. T., et al. (2007). Analysis of mouse skin reveals proteins that are altered in a diet-induced diabetic state: A new method for detection of type 2 diabetes. Proteomics, 7(7), 1140. List, E. O., Palmer, A. J., Berryman, D. E., Bower, B., Kelder, B., & Kopchick, J. J. (2009). Growth hormone improves body composition, fasting blood glucose, glucose tolerance and liver triacylglycerol in a mouse model of diet-induced obesity and type 2 diabetes. Diabetologia. Liu, J. L., Coschigano, K. T., Robertson, K., Lipsett, M., Guo, Y., Kopchick, J. J., et al. (2004). Disruption of growth hormone receptor gene causes diminished pancreatic islet size and increased insulin sensitivity in mice. Am J Physiol Endocrinol Metab, 287(3), E405-413. Longo, V. D., & Finch, C. E. (2003). Evolutionary medicine: From dwarf model systems to healthy centenarians? 10.1126/science.1077991. Science, 299(5611), 1342. Longobardi, S., Keay, N., Ehrnborg, C., Cittadini, A., Rosen, T., Dall, R., et al. (2000). Growth hormone (GH) effects on bone and collagen turnover in healthy adults and its potential as a marker of GH abuse in sports: A double blind, placebo- controlled study. The GH-2000 study group. The Journal of clinical endocrinology and metabolism, 85(4), 1505. Lorent, K., Overbergh, L., & Van Leuven, F. (1994). Expression of mouse alpha 2m and its receptor in vivo. Annals of the New York Academy of Sciences, 737, 498. Lynch, J. R., Morgan, D., Mance, J., Matthew, W. D., & Laskowitz, D. T. (2001). Apolipoprotein e modulates glial activation and the endogenous central nervous system inflammatory response. Journal of Neuroimmunology, 114(1-2), 107. Ma, Z. C., Gao, Y., Tang, H. L., & Wang, S. Q. (2003). [the effects of si-wu-tang on serum protein of blood deficient mice induced by radiation]. Zhongguo Zhong Yao Za Zhi, 28(11), 1050-1053. MacBeath, G., & Schreiber, S. L. (2000). Printing proteins as microarrays for high- throughput function determination. Science (New York, N.Y.), 289(5485), 1760. MacLeod, J. N., Pampori, N. A., & Shapiro, B. H. (1991). Sex differences in the ultradian pattern of plasma growth hormone concentrations in mice. J Endocrinol, 131(3), 395-399. Magistroni, R., Ligabue, G., Lupo, V., Furci, L., Leonelli, M., Manganelli, L., et al. (2009). Proteomic analysis of urine from proteinuric patients shows a proteolitic activity directed against albumin. Nephrol. Dial. Transplant., 24(5), 1672-1681. Manson, J. M., & Wilmore, D. W. (1986). Positive nitrogen balance with human growth hormone and hypocaloric intravenous feeding. Surgery, 100(2), 188-197. Marshall, W. A. T., JM. (1986). Human growth: A comprehensive treatise (2nd ed.). New York: Plenum Press. 273

Mascot. (2009). from http://www.matrixscience.com/ Masternak, M. M., Al-Regaiey, K. A., Del Rosario Lim, M. M., Bonkowski, M. S., Panici, J. A., Przybylski, G. K., et al. (2005a). Caloric restriction results in decreased expression of peroxisome proliferator-activated receptor superfamily in muscle of normal and long-lived growth hormone receptor/binding protein knockout mice. J Gerontol A Biol Sci Med Sci, 60(10), 1238-1245. Masternak, M. M., Al-Regaiey, K. A., Del Rosario Lim, M. M., Jimenez-Ortega, V., Panici, J. A., Bonkowski, M. S., et al. (2005b). Effects of caloric restriction on insulin pathway gene expression in the skeletal muscle and liver of normal and long-lived ghr-ko mice. Exp Gerontol, 40(8-9), 679-684. Masternak, M. M., Al-Regaiey, K. A., Del Rosario Lim, M. M., Jimenez-Ortega, V., Panici, J. A., Bonkowski, M. S., et al. (2006). Caloric restriction and growth hormone receptor knockout: Effects on expression of genes involved in insulin action in the heart. Exp Gerontol, 41(4), 417-429. McGrane, M. M., de Vente, J., Yun, J., Bloom, J., Park, E., Wynshaw-Boris, A., et al. (1988). Tissue-specific expression and dietary regulation of a chimeric phosphoenolpyruvate carboxykinase/bovine growth hormone gene in transgenic mice. The Journal of biological chemistry, 263(23), 11443. McMillan, D. E. (1989). Increased levels of acute-phase serum proteins in diabetes. Metabolism, 38(11), 1042-1046. Meinhardt, U. J., & Ho, K. K. (2007). Regulation of growth hormone action by gonadal steroids. Endocrinology and metabolism clinics of North America, 36(1), 57. Merlini, G., & Bellotti, V. (2003). Molecular mechanisms of amyloidosis. N Engl J Med, 349(6), 583-596. Metais, P., & Sacrez, A. (1963). Serum lactate dehydrogenase in cardiac disorders. Strasbourg medical, 14, 518. Michelson, W. B. (1980). Cpk-mb isoenzyme determinations: Diagnostic and prognostic value in evaluation of blunt chest trauma. Annals of Emergency Medicine, 9(11), 562. Mikhail, N. (2009). The metabolic syndrome: Insulin resistance. Curr Hypertens Rep, 11(2), 156-158. Miller, R. A., Chang, Y., Galecki, A. T., Al-Regaiey, K., Kopchick, J. J., & Bartke, A. (2002). Gene expression patterns in calorically restricted mice: Partial overlap with long-lived mutant mice. Mol Endocrinol, 16(11), 2657-2666. Miller, R. A., & Nadon, N. L. (2000). Principles of animal use for gerontological research. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 55(3), B117. Mine, M., Bedi, K., Maruta, T., Madziva, D., Tau, M., Zana, T., et al. (2009). Quantitation of human immunodeficiency virus type 1 viral load in plasma using reverse transcriptase activity assay at a district hospital laboratory in botswana: A decentralization pilot study. J Virol Methods, 159(1), 93-97. Miura, Y., Kano, M., Yamada, M., Nishine, T., Urano, S., Suzuki, S., et al. (2007). Proteomic study on x-irradiation-responsive proteins and ageing: Search for responsible proteins for radiation adaptive response. J Biochem, 142(2), 145-155. 274

Miyake, H., Hara, S., Arakawa, S., Kamidono, S., & Hara, I. (2002). Over expression of clusterin is an independent prognostic factor for nonpapillary renal cell carcinoma. J Urol, 167(2 Pt 1), 703-706. Miyata, M., & Smith, J. D. (1996). Apolipoprotein e allele-specific antioxidant activity and effects on cytotoxicity by oxidative insults and beta-amyloid peptides. Nat Genet, 14(1), 55-61. Molander, D. W., Wroblewski, F., & Ladue, J. S. (1955). Serum glutamic oxalacetic transaminase as an index of hepatocellular integrity. The Journal of laboratory and clinical medicine, 46(6), 831. Moller, J., Nielsen, S., & Hansen, T. K. (1999). Growth hormone and fluid retention. Horm Res, 51 Suppl 3, 116-120. Moller, N., & Jorgensen, J. O. L. (2009). Effects of growth hormone on glucose, lipid, and protein metabolism in human subjects. Endocr Rev, 30(2), 152-177. Monarca, A., Petrini, C., Perolini, S., Pozzi, F., Adelasco, L., Natangelo, R., et al. (1985). Procollagen-type iii peptide serum concentrations in alcoholic and non-alcoholic liver disease. Ric Clin Lab, 15(2), 167-171. Moore, D. C., Ruvalcaba, R. H., Smith, E. K., & Kelley, V. C. (1982). Plasma somatomedin-c as a screening test for growth hormone deficiency in children and adolescents. Horm Res, 16(1), 49-55. Moore, L. G., Pfeffer, A., Chie, W. N., Miller, H. A., Rogers, K. M., & O'Keeffe, L. E. (1995). Induction of an acute phase response in lambs causes an increase in plasma levels of GH and IGF-I. J Endocrinol, 144(2), 243-250. Mortensen, E. L., & Hogh, P. (2001). A gender difference in the association between apoE genotype and age-related cognitive decline. Neurology, 57(1), 89-95. Moura, A. S., Spiers, D. E., & Lamberson, W. R. (1998). Thermogenic activity of growth hormone transgenic mice. Growth Dev Aging, 62(4), 149-159. Muhlhausler, B., & Smith, S. R. (2009). Early-life origins of metabolic dysfunction: Role of the adipocyte. Trends Endocrinol Metab, 20(2), 51-57. Mulligan, K., Grunfeld, C., Hellerstein, M. K., Neese, R. A., & Schambelan, M. (1993). Anabolic effects of recombinant human growth hormone in patients with wasting associated with human immunodeficiency virus infection. J Clin Endocrinol Metab, 77(4), 956-962. Naaby-Hansen, S., Waterfield, M. D., & Cramer, R. (2001). Proteomics--post-genomic cartography to understand gene function. Trends in pharmacological sciences, 22(7), 376. Nakamura, E., & Miyao, K. (2007). A method for identifying biomarkers of aging and constructing an index of biological age in humans. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 62(10), 1096. Nakamura, E., & Miyao, K. (2008). Sex differences in human biological aging. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 63(9), 936. Nakashima, Y., Plump, A. S., Raines, E. W., Breslow, J. L., & Ross, R. (1994). Apoe- deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb, 14(1), 133-140. 275

Nakhoul, F. M., Miller-Lotan, R., Awaad, H., Asleh, R., & Levy, A. P. (2007). Hypothesis--haptoglobin genotype and diabetic nephropathy. Nat Clin Pract Nephrol, 3(6), 339-344. Nass, R., Johannsson, G., Christiansen, J. S., Kopchick, J. J., & Thorner, M. O. (2009). The aging population - is there a role for endocrine interventions? Growth Hormone & IGF Research, 19(2), 89. NCBInr. Ncbi non-redundant protein database (http://www.Ncbi.Nlm.Nih.Gov). Nelson, A. E., & Ho, K. K. (2008). A robust test for growth hormone doping--present status and future prospects. Asian Journal of Andrology, 10(3), 416. Nelson, A. E., Howe, C. J., Nguyen, T. V., Leung, K. C., Trout, G. J., Seibel, M. J., et al. (2006). Influence of demographic factors and sport type on growth hormone- responsive markers in elite athletes. The Journal of clinical endocrinology and metabolism, 91(11), 4424. Nelson, A. E., Meinhardt, U., Hansen, J. L., Walker, I. H., Stone, G., Howe, C. J., et al. (2008). Pharmacodynamics of growth hormone abuse biomarkers and the influence of gender and testosterone: A randomized double-blind placebo- controlled study in young recreational athletes. The Journal of clinical endocrinology and metabolism, 93(6), 2213. Nguyen, T. V., Nelson, A. E., Howe, C. J., Seibel, M. J., Baxter, R. C., Handelsman, D. J., et al. (2008). Within-subject variability and analytic imprecision of insulinlike growth factor axis and collagen markers: Implications for clinical diagnosis and doping tests. Clinical chemistry, 54(8), 1268. Nielsen, J. H., Galsgaard, E. D., Moldrup, A., Friedrichsen, B. N., Billestrup, N., Hansen, J. A., et al. (2001). Regulation of beta-cell mass by hormones and growth factors. Diabetes, 50 Suppl 1, S25-29. Nir, B., Ilan, G., Magda, G., Nancy, I., & John, D. S. (2001). Offspring of centenarians have a favorable lipid profile. Journal of the American Geriatrics Society, 49(1), 76-79. Nofer, J.-R., Walter, M., & Assmann, G. (2005). Current understanding of the role of high-density lipoproteins in atherosclerosis and senescence. Expert Review of Cardiovascular Therapy, 3(6), 1071. Nordlie, R. C., Foster, J. D., & Lange, A. J. (1999). Regulation of glucose production by the liver. Annu Rev Nutr, 19, 379-406. Norstedt, G., & Palmiter, R. (1984). Secretory rhythm of growth hormone regulates sexual differentiation of mouse liver. Cell, 36(4), 805-812. O'Farrell, P. H. (1975). High resolution two-dimensional electrophoresis of proteins. J Biol Chem, 250(10), 4007-4021. O'Shea, M. H., & Layish, D. T. (1992). Growth hormone and the kidney: A case presentation and review of the literature. J Am Soc Nephrol, 3(2), 157-161. O'Sullivan, A. J., Kelly, J. J., Hoffman, D. M., Freund, J., & Ho, K. K. (1994). Body composition and energy expenditure in acromegaly. J Clin Endocrinol Metab, 78(2), 381-386. 276

Ogueta, S., Olazabal, I., Santos, I., Delgado-Baeza, E., & Garcia-Ruiz, J. P. (2000). Transgenic mice expressing bovine GH develop arthritic disorder and self- antibodies. J Endocrinol, 165(2), 321-328. Ohlsson, C., Bengtsson, B. A., Isaksson, O. G., Andreassen, T. T., & Slootweg, M. C. (1998). Growth hormone and bone. Endocr Rev, 19(1), 55-79. Okuda, Y., Peña, J., Chou, J., & Field, J. B. (2001). Acute effects of growth hormone on metabolism of pancreatic hormones, glucose and ketone bodies. Diabetes Research and Clinical Practice, 53(1), 1. OlÃ¡hovÃ¡, M., Taylor, S. R., Khazaipoul, S., Wang, J., Morgan, B. A., Matsumoto, K., et al. (2008). A redox-sensitive peroxiredoxin that is important for longevity has tissue- and stress-specific roles in stress resistance. Proceedings of the National Academy of Sciences, 105(50), 19839-19844. Olsson, B., Bohlooly-Y, M., Brusehed, O., Isaksson, O. G. P., Ahren, B., Olofsson, S.-O., et al. (2003). Bovine growth hormone-transgenic mice have major alterations in hepatic expression of metabolic genes. AJP - Endocrinology and Metabolism, 285(3), E504. Olsson, B., Bohlooly-Y, M., Fitzgerald, S. M., Frick, F., Ljungberg, A., Ahren, B., et al. (2005). Bovine growth hormone transgenic mice are resistant to diet-induced obesity but develop hyperphagia, dyslipidemia, and diabetes on a high-fat diet. Endocrinology, 146(2), 920-930. Önenli-Mungan, N., Yildizdas, D., Yapicioglu, H., Topaloglu, A. K., Yüksel, B., & Özer, G. (2004). Growth hormone and insulin-like growth factor 1 levels and their relation to survival in children with bacterial sepsis and septic shock. Journal of Paediatrics and Child Health, 40(4), 221-226. Ong, D. E., Davis, J. T., O'Day, W. T., & Bok, D. (1994). Synthesis and secretion of retinol-binding protein and transthyretin by cultured retinal pigment epithelium. Biochemistry, 33(7), 1835. Orentreich, N., Brind, J. L., Vogelman, J. H., Andres, R., & Baldwin, H. (1992). Long- term longitudinal measurements of plasma dehydroepiandrosterone sulfate in normal men. J Clin Endocrinol Metab, 75(4), 1002-1004. Orian, J. M., Lee, C. S., Weiss, L. M., & Brandon, M. R. (1989). The expression of a metallothionein-ovine growth hormone fusion gene in transgenic mice does not impair fertility but results in pathological lesions in the liver. Endocrinology, 124(1), 455-463. Oscarsson, J., Carlsson, L. M. S., Bick, T., Lidell, A., Olofsson, S. O., & Eden, S. (1991a). Evidence for the role of the secretory pattern of growth hormone in the regulation of serum concentrations of cholesterol and apolipoprotein e in rats. J Endocrinol, 128(3), 433-438. Oscarsson, J., Olofsson, S.-O., Vikman, K., & Edén, S. (1991b). Growth hormone regulation of serum lipoproteins in the rat: Different growth hormone regulatory principles for apolipoprotein (apo) b and the sexually dimorphic apo e concentrations. Metabolism, 40(11), 1191. Osterman, P. O., Wallin, G., & Wide, L. (1974). Nocturnal secretory patterns of FSH, GH, LH and TSH. Ups J Med Sci, 79(1), 55-62. 277

Ostos, M. A., Conconi, M., Vergnes, L., Baroukh, N., Ribalta, J., Girona, J., et al. (2001). Antioxidative and antiatherosclerotic effects of human apolipoprotein a-iv in apolipoprotein e-deficient mice. Arterioscler Thromb Vasc Biol, 21(6), 1023- 1028. Owens, D., Srivastava, M. C., Tompkins, C. V., Nabarro, J. D., & Sonksen, P. H. (1973). Studies on the metabolic clearance rate, apparent distribution space and plasma half-disappearance time of unlabelled human growth hormone in normal subjects and in patients with liver disease, renal disease, thyroid disease and diabetes mellitus. European journal of clinical investigation, 3(4), 284. Packer, N. H., & Harrison, M. J. (1998). Glycobiology and proteomics: Is mass spectrometry the holy grail? Electrophoresis, 19(11), 1872-1882. Pajak, B., & Orzechowski, A. (2006). Clusterin: The missing link in the calcium- dependent resistance of cancer cells to apoptogenic stimuli. Postepy higieny i medycyny doswiadczalnej (Online), 60, 45. Palmer, A. J., Chung, M. Y., List, E. O., Walker, J., Okada, S., Kopchick, J. J., et al. (2009). Age-related changes in body composition of bovine growth hormone transgenic mice. Endocrinology, 150(3), 1353-1360. Pandey, A., & Mann, M. (2000). Proteomics to study genes and genomes. Nature, 405(6788), 837. Parsons, J. A., Bartke, A., & Sorenson, R. L. (1995). Number and size of islets of langerhans in pregnant, human growth hormone-expressing transgenic, and pituitary dwarf mice: Effect of lactogenic hormones. Endocrinology, 136(5), 2013-2021. Patsch, W., Kim, K., Wiest, W., & Schonfeld, G. (1980). Effects of sex hormones on rat lipoproteins. Endocrinology, 107(4), 1085-1094. Pennacchio, L. (2003). Insights from human/mouse genome comparisons. Mammalian Genome, 14(7), 429. Petricoin, E. E., Paweletz, C. P., & Liotta, L. A. (2002). Clinical applications of proteomics: Proteomic pattern diagnostics. J Mammary Gland Biol Neoplasia, 7(4), 433-440. Pinkert, C. A., Galbreath, E. J., Yang, C. W., & Striker, L. J. (1994). Liver, renal and subcutaneous histopathology in pepck-bgh transgenic pigs. Transgenic Res, 3(6), 401-405. Popovic, V. (2005). Are there alternative tests for diagnosis of acromegaly? Journal of endocrinological investigation, 28(11 Suppl), 73. Posch, A., van den Berg, B. M., Burg, H. C., & Gorg, A. (1995). Genetic variability of carrot seed proteins analyzed by one- and two-dimensional electrophoresis with immobilized ph gradients. Electrophoresis, 16(7), 1312-1316. Postel-Vinay, M.-C., de Mello Coelho, V., Gagnerault, M.-C., & Dardenne, M. (1997). Growth hormone stimulates the proliferation of activated mouse t lymphocytes. Endocrinology, 138(5), 1816-1820. Pozzilli, P., Tamburrano, G., Mercalli, M. E., & Javicoli, M. (1978). Raised alpha-2- macroglobulin levels in acromegaly. Horm Metab Res, 10(4), 359. 278

Prize, M. M. (2009). The world's longest living laboratory mouse. from http://www.methuselahfoundation.org/ Qiu, L., List, E. O., & Kopchick, J. J. (2005). Differentially expressed proteins in the pancreas of diet-induced diabetic mice. Molecular & cellular proteomics: MCP, 4(9), 1311. Quaife, C. J., Mathews, L. S., Pinkert, C. A., Hammer, R. E., Brinster, R. L., & Palmiter, R. D. (1989). Histopathology associated with elevated levels of growth hormone and insulin-like growth factor i in transgenic mice. Endocrinology, 124(1), 40-48. R I G Holt, P. H. S. (2008). Growth hormone, IGF-I and insulin and their abuse in sport. British Journal of Pharmacology, 154(3), 542-556. Rabilloud, T. (2000). Detecting proteins separated by 2-d gel electrophoresis. Analytical Chemistry, 72(1), 48A. Rabilloud, T., & Chevallet, M. (2000). Solubilization of proteins in 2d electrophoresis. In T. Rabilloud (Ed.), Proteome research: Two-dimensional gel electrophoresis and identification methods (pp. 9-30.). Berlin: Springer. Raftery, M., Campbell, R., Glaros, E. N., Rye, K. A., Halliday, G. M., Jessup, W., et al. (2005). Phosphorylation of apolipoprotein-e at an atypical protein kinase ck2 psd/e site in vitro. Biochemistry, 44(19), 7346-7353. Raila, J., Henze, A., Spranger, J., Mohlig, M., Pfeiffer, A. F., & Schweigert, F. J. (2007). Microalbuminuria is a major determinant of elevated plasma retinol-binding protein 4 in type 2 diabetic patients. Kidney international. Ramsby, M. L., Makowski, G. S., & Khairallah, E. A. (1994). Differential detergent fractionation of isolated hepatocytes: Biochemical, immunochemical and two- dimensional gel electrophoresis characterization of cytoskeletal and noncytoskeletal compartments. Electrophoresis, 15(2), 265-277. Ransome, M. I., Goldshmit, Y., Bartlett, P. F., Waters, M. J., & Turnley, A. M. (2004). Comparative analysis of cns populations in knockout mice with altered growth hormone responsiveness. Eur J Neurosci, 19(8), 2069-2079. Recalde, D., Ostos, M. A., Badell, E., Garcia-Otin, A.-L., Pidoux, J., Castro, G., et al. (2004). Human apolipoprotein a-iv reduces secretion of proinflammatory cytokines and atherosclerotic effects of a chronic infection mimicked by lipopolysaccharide. Arterioscler Thromb Vasc Biol, 24(4), 756-761. Reddy, G. R., Pushpanathan, M. J., Ransom, R. F., Holzman, L. B., Brosius, F. C., 3rd, Diakonova, M., et al. (2007). Identification of the glomerular podocyte as a target for growth hormone action. Endocrinology, 148(5), 2045-2055. Refai, E., Dekki, N., Yang, S. N., Imreh, G., Cabrera, O., Yu, L., et al. (2005). Transthyretin constitutes a functional component in pancreatic beta-cell stimulus- secretion coupling. Proceedings of the National Academy of Sciences of the United States of America, 102(47), 17020. Reinehr, T., Stoffel-Wagner, B., & Roth, C. L. (2008). Retinol-binding protein 4 and its relation to insulin resistance in obese children before and after weight loss. J Clin Endocrinol Metab, 93(6), 2287-2293. 279

Reiter, E. O., & Lovinger, R. D. (1981). The use of a commercially available somatomedin-c radioimmunoassay in patients with disorders of growth. J Pediatr, 99(5), 720-724. Ren, Y., Wang, J., Xia, J., Jiang, C., Zhao, K., Li, R., et al. (2007). The alterations of mouse plasma proteins during septic development. Journal of Proteome Research, 6(7), 2812. Rhodes, C. J. (2000). Igf-i and GH post-receptor signaling mechanisms for pancreatic beta-cell replication. J Mol Endocrinol, 24(3), 303-311. Rice, R. W., Kroning, J., & Critchlow, V. (1975). Effects of stress on plasma corticosterone and growth hormone levels in rats with median eminence-pituitary islands. Neuroendocrinology, 19(4), 339-351. Ridker, P. M., Hennekens, C. H., Buring, J. E., & Rifai, N. (2000). C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med, 342(12), 836-843. Rithidech, K. N., Honikel, L., Rieger, R., Xie, W., Fischer, T., & Simon, S. R. (2009). Protein-expression profiles in mouse blood-plasma following acute whole-body exposure to 137cs γ rays. International Journal of Radiation Biology, 85(5), 432 - 447. Robert, F. R., Glenn, E. P., Louis, M. N., & Olga, N. (1999a). Reference distributions for the negative acute-phase proteins, albumin, transferrin, and transthyretin: A comparison of a large cohort to the world's literature. Journal of Clinical Laboratory Analysis, 13(6), 280-286. Robert, F. R., Glenn, E. P., Louis, M. N., Olga, N., Thomas, B. L., & Wendy, Y. C. (1999b). Reference distributions for the negative acute-phase serum proteins, albumin, transferrin and transthyretin: A practical, simple and clinically relevant approach in a large cohort. Journal of Clinical Laboratory Analysis, 13(6), 273- 279. Roberto, G., Ifeanyi, G. O., Erica, B., Chiara, M., John, C. A., Riccardo, C., et al. (2008). Proteomic analysis of the pathophysiological process involved in the antisnake venom effect of mucuna pruriens extract. PROTEOMICS, 8(2), 402-412. Robertson, K., Kopchick, J. J., & Liu, J. L. (2006). Growth hormone receptor gene deficiency causes delayed insulin responsiveness in skeletal muscles without affecting compensatory islet cell overgrowth in obese mice. Am J Physiol Endocrinol Metab, 291(3), E491-498. Rocha, J. S., Bonkowski, M. S., de França, L. R., & Bartke, A. (2007). Effects of mild calorie restriction on reproduction, plasma parameters and hepatic gene expression in mice with altered GH/IGF-I axis. Mechanisms of Ageing and Development, 128(4), 317. Rose, T., Lambotte, O., Pallier, C., Delfraissy, J. F., & Colle, J. H. (2009). Identification and biochemical characterization of human plasma soluble IL-7r: Lower concentrations in hiv-1-infected patients. J Immunol, 182(12), 7389-7397. Rosenberg, M. E., Girton, R., Finkel, D., Chmielewski, D., Barrie Iii, A., Witte, D. P., et al. (2002). Apolipoprotein j/clusterin prevents a progressive glomerulopathy of aging. Mol. Cell. Biol., 22(6), 1893-1902. 280

Rosenberg, M. E., & Paller, M. S. (1991). Differential gene expression in the recovery from ischemic renal injury. Kidney Int, 39(6), 1156-1161. Rosvall, L., Rizzuto, D., Wang, H. X., Winblad, B., Graff, C., & Fratiglioni, L. (2008). Apoe-related mortality: Effect of dementia, cardiovascular disease and gender. Neurobiol Aging. Rothemund, D. L., Locke, V. L., Liew, A., Thomas, T. M., Wasinger, V., & Rylatt, D. B. (2003). Depletion of the highly abundant protein albumin from human plasma using the gradiflow. Proteomics, 3(3), 279. Roubenoff, R., Harris, T. B., Abad, L. W., Wilson, P. W., Dallal, G. E., & Dinarello, C. A. (1998). Monocyte cytokine production in an elderly population: Effect of age and inflammation. J Gerontol A Biol Sci Med Sci, 53(1), M20-26. Sackmann-Sala, L., Ding, J., Frohman, L. A., & Kopchick, J. J. (2009). Activation of the GH/IGF-1 axis by cjc-1295, a long-acting ghrh analog, results in serum protein profile changes in normal adult subjects. Growth Horm IGF Res. Sanchez, J. C., Appel, R. D., Golaz, O., Pasquali, C., Ravier, F., Bairoch, A., et al. (1995). Inside swiss-2dpage database. Electrophoresis, 16(7), 1131-1151. Santos-Gallego, C. G., Ibanez, B., & Badimon, J. J. (2008). HDL-cholesterol: Is it really good? Differences between apoA-I and HDL. Biochemical Pharmacology, 76(4), 443. Sato, Y., Suzuki, Y., Ito, E., Shimazaki, S., Ishida, M., Yamamoto, T., et al. (2006). Identification and characterization of an increased glycoprotein in aging: Age- associated translocation of cathepsin d. Mechanisms of Ageing and Development, 127(10), 771. Saugy, M., Robinson, N., Saudan, C., Baume, N., Avois, L., & Mangin, P. (2006). Human growth hormone doping in sport. British journal of sports medicine, 40 Suppl 1, i35. Saunders, A., Terry, L. C., Audet, J., Brazeau, P., & Martin, J. B. (1976). Dynamic studies of growth hormone and prolactin secretion in the female rat. Neuroendocrinology, 21(3), 193-203. Schambelan, M., Mulligan, K., Grunfeld, C., Daar, E. S., LaMarca, A., Kotler, D. P., et al. (1996). Recombinant human growth hormone in patients with hiv-associated wasting: A randomized, placebo-controlled trial. Ann Intern Med, 125(11), 873- 882. Schiele, F., De Bacquer, D., Vincent-Viry, M., Beisiegel, U., Ehnholm, C., Evans, A., et al. (2000). Apolipoprotein e serum concentration and polymorphism in six european countries: The apoeurope project. Atherosclerosis, 152(2), 475. Schöneich, C. (2003). Proteomics in gerontological research. Experimental Gerontology, 38(5), 473. Schreiber, G. (2002). The evolution of transthyretin synthesis in the choroid plexus. Clinical chemistry and laboratory medicine: CCLM / FESCC, 40(12), 1200. Schuenke, M. D., Kopchick, J. J., Hikida, R. S., Kraemer, W. J., & Staron, R. S. (2008). Effects of growth hormone overexpression vs. Growth hormone receptor gene disruption on mouse hindlimb muscle fiber type composition. Growth Horm IGF Res, 18(6), 479-486. 281

Schussler, G. C. (2000). The thyroxine-binding proteins. Thyroid: official journal of the American Thyroid Association, 10(2), 141. Seo, J. A., Kim, N. H., Park, S. Y., Kim, H. Y., Ryu, O. H., Lee, K. W., et al. (2008). Serum retinol-binding protein 4 levels are elevated in non-alcoholic fatty liver disease. Clin Endocrinol (Oxf), 68(4), 555-560. Seres, D. S. (2005). Surrogate nutrition markers, malnutrition, and adequacy of nutrition support. Nutr Clin Pract, 20(3), 308. Serhan, C. N., & Savill, J. (2005). Resolution of inflammation: The beginning programs the end. Nat Immunol, 6(12), 1191-1197. Serrat, M. A., Vinyard, C. J., & King, D. (2007). Alterations in the mechanical properties and composition of skin in human growth hormone transgenic mice. Connect Tissue Res, 48(1), 19-26. Shapiro, A. L., Vinuela, E., & Maizel, J. V., Jr. (1967). Molecular weight estimation of polypeptide chains by electrophoresis in sds-polyacrylamide gels. Biochemical and biophysical research communications, 28(5), 815. Shea, B. T., Hammer, R. E., & Brinster, R. L. (1987). Growth allometry of the organs in giant transgenic mice. Endocrinology, 121(6), 1924-1930. Shechter, M., Ginsberg, S., Scheinowitz, M., Feinberg, M. S., & Laron, Z. (2007). Obese adults with primary growth hormone resistance (laron syndrome) have normal endothelial function. Growth Horm IGF Res, 17(2), 165-170. Sheehan, J. P. (2004). Fasting hyperglycemia: Etiology, diagnosis, and treatment. Diabetes Technol Ther, 6(4), 525-533. Shen, Y., Jacobs, J. M., Camp, D. G., Fang, R., Moore, R. J., Smith, R. D., et al. (2004). Ultra-high-efficiency strong cation exchange lc/rplc/ms/ms for high dynamic range characterization of the human plasma proteome. Analytical Chemistry, 76(4), 1134-1144. Sheng, S., Chen, D., & Van Eyk, J. E. (2006). Multidimensional liquid chromatography separation of intact proteins by chromatographic focusing and reversed phase of the human serum proteome: Optimization and protein database. Mol Cell Proteomics, 5(1), 26-34. Shevah, O., & Laron, Z. (2007). Patients with congenital deficiency of IGF-I seem protected from the development of malignancies: A preliminary report. Growth Horm IGF Res, 17(1), 54-57. Sideris P. Delaroudis, Z. A. E. G. N. K. M. D. K. D. F. O. G. D. D. G. G. A. M. P. M. A. S. A. I. (2008). Amelioration of cardiovascular risk factors with partial biochemical control of acromegaly. Clinical Endocrinology, 69(2), 279-284. Singhal, P. C., Gibbons, N., Franki, N., Reddy, K., Sharma, P., Mattana, J., et al. (1998). Simulated glomerular hypertension promotes mesangial cell apoptosis and expression of cathepsin-b and sgp-2. J Investig Med, 46(2), 42-50. Sjoberg, A., Oscarsson, J., Eden, S., & Olofsson, S. O. (1994a). Continuous but not intermittent administration of growth hormone to hypophysectomized rats increases apolipoprotein-e secretion from cultured hepatocytes. Endocrinology, 134(2), 790-798. 282

Sjoberg, A., Oscarsson, J., Olofsson, S. O., & Eden, S. (1994b). Insulin-like growth factor-i and growth hormone have different effects on serum lipoproteins and secretion of lipoproteins from cultured rat hepatocytes. Endocrinology, 135(4), 1415-1421. Sjogren, K., Bohlooly, Y. M., Olsson, B., Coschigano, K., Tornell, J., Mohan, S., et al. (2000). Disproportional skeletal growth and markedly decreased bone mineral content in growth hormone receptor -/- mice. Biochem Biophys Res Commun, 267(2), 603-608. Skehel, J. M., Klaus, S., Nuala, M., Annette, G., Benson, G. M., Paul, C., et al. (2000). Phenotyping apolipoprotein e*3-leiden transgenic mice by two-dimensional polyacrylamide gel electrophoresis and mass spectrometric identification. Electrophoresis, 21(12), 2540-2545. Snibson, K. J., Bhathal, P. S., & Adams, T. E. (2001). Overexpressed growth hormone (GH) synergistically promotes carcinogen-initiated liver tumour growth by promoting cellular proliferation in emerging hepatocellular neoplasms in female and male GH-transgenic mice. Liver, 21(2), 149-158. Snibson, K. J., Bhathal, P. S., Hardy, C. L., Brandon, M. R., & Adams, T. E. (1999). High, persistent hepatocellular proliferation and apoptosis precede hepatocarcinogenesis in growth hormone transgenic mice. Liver, 19(3), 242-252. Soderpalm, B., Ericson, M., Bohlooly-y, M., Engel, J. A., & Tornell, J. (1999). Bovine growth hormone transgenic mice display alterations in locomotor activity and brain monoamine neurochemistry. Endocrinology, 140(12), 5619-5625. Song, J., Stastny, J., Fosslien, E., & Robertson, A. L., Jr. (1985). Effect of aging on human aortic protein composition. Ii. Two-dimensional polyacrylamide gel electrophoretic analysis. Exp Mol Pathol, 43(3), 297-304. Sonksen, P. (2009). The international olympic committee (ioc) and GH-2000. Growth Hormone & IGF Research, In Press, Corrected Proof. Sørensen, C. M., Hansen, T. K., Steffensen, R., Jensenius, J. C., & Thiel, S. (2006). Hormonal regulation of mannan-binding lectin synthesis in hepatocytes. Clinical and Experimental Immunology, 145(1), 173-182. Sotiropoulos, A., Ohanna, M., Kedzia, C., Menon, R. K., Kopchick, J. J., Kelly, P. A., et al. (2006). Growth hormone promotes skeletal muscle cell fusion independent of insulin-like growth factor 1 up-regulation. Proc Natl Acad Sci U S A, 103(19), 7315-7320. Spurlock, M. E., Ranalletta, M. A., Cornelius, S. G., Frank, G. R., Willis, G. M., Ji, S., et al. (1998). Leptin expression in porcine adipose tissue is not increased by endotoxin but is reduced by growth hormone. J Interferon Cytokine Res, 18(12), 1051-1058. Steel, L. F., Trotter, M. G., Nakajima, P. B., Mattu, T. S., Gonye, G., & Block, T. (2003). Efficient and specific removal of albumin from human serum samples. Molecular Cellular Proteomics, 2(4), 262. Steger, R. W., Bartke, A., & Cecim, M. (1993). Premature ageing in transgenic mice expressing different growth hormone genes. J Reprod Fertil Suppl, 46, 61-75. 283

Stonesifer, L. D., Jordan, R. M., & Kohler, P. O. (1981). Somatomedin c in treated acromegaly: Poor correlation with growth hormone and clinical response. J Clin Endocrinol Metab, 53(5), 931-934. Suhr, O. B., Svendsen, I. H., Ohlsson, P. I., Lendoire, J., Trigo, P., Tashima, K., et al. (1999). Impact of age and amyloidosis on thiol conjugation of transthyretin in hereditary transthyretin amyloidosis. Amyloid: the international journal of experimental and clinical investigation: the official journal of the International Society of Amyloidosis, 6(3), 187. Sumiya, M., Tabona, P., Arai, T., Summerfield, J. A., Super, M., Levinsky, R. J., et al. (1991). Molecular basis of opsonic defect in immunodeficient children. The Lancet, 337(8757), 1569. Sutton, J., & Lazarus, L. (1976). Growth hormone in exercise: Comparison of physiological and pharmacological stimuli. J Appl Physiol, 41(4), 523-527. Svensson, J., Bengtsson, B. Å., Taskinen, M. R., Wiklund, O., & Johannsson, G. (2000). A nine-month, placebo-controlled study of the effects of growth hormone treatment on lipoproteins and LDL size in abdominally obese men. Growth Hormone & IGF Research, 10(3), 118. Svensson, J., Jansson, J. O., Ottosson, M., Johannsson, G., Taskinen, M. R., Wiklund, O., et al. (1999). Treatment of obese subjects with the oral growth hormone secretagogue mk-677 affects serum concentrations of several lipoproteins, but not lipoprotein(a). J Clin Endocrinol Metab, 84(6), 2028-2033. Swiss-Prot. Swiss-prot/trembl (http://www.Expasy.Org). Takala, J., Ruokonen, E., Webster, N. R., Nielsen, M. S., Zandstra, D. F., Vundelinckx, G., et al. (1999). Increased mortality associated with growth hormone treatment in critically ill adults. N Engl J Med, 341(11), 785-792. Takebayashi, K., Suetsugu, M., Wakabayashi, S., Aso, Y., & Inukai, T. (2007a). Retinol binding protein-4 levels and clinical features of type 2 diabetes patients. Journal of Clinical Endocrinology Metabolism, 92(7), 2712. Takebayashi, K., Suetsugu, M., Wakabayashi, S., Aso, Y., & Inukai, T. (2007b). Retinol binding protein-4 levels and clinical features of type 2 diabetes patients. J Clin Endocrinol Metab, 92(7), 2712-2719. Tannenbaum, G. S. (1991). Neuroendocrine control of growth hormone secretion. Acta Paediatr Scand Suppl, 372, 5-16. Tannenbaum, G. S., & Martin, J. B. (1976). Evidence for an endogenous ultradian rhythm governing growth hormone secretion in the rat. Endocrinology, 98(3), 562-570. Tarnus, E., Wassef, H., Carmel, J.-F., Rondeau, P., Roche, M., Davignon, J., et al. (2009). Apolipoprotein e limits oxidative stress-induced cell dysfunctions in human adipocytes. FEBS Letters, 583(12), 2042. Tatar, M., Bartke, A., & Antebi, A. (2003). The endocrine regulation of aging by insulin- like signals 10.1126/science.1081447. Science, 299(5611), 1346. Taubes, G. (2009). Prosperity's plague. Science, 325(5938), 256-260. 284

Tayade, C., Esadeg, S., Fang, Y., & Croy, B. A. (2005). Functions of alpha 2 macroglobulins in pregnancy. Molecular and Cellular Endocrinology, 245(1-2), 60. Terazaki, H., Ando, Y., Suhr, O., Ohlsson, P.-I., Obayashi, K., Yamashita, T., et al. (1998). Post-translational modification of transthyretin in plasma. Biochemical and Biophysical Research Communications, 249(1), 26. Teunissen, C. E., van Boxtel, M. P. J., Bosma, H., Bosmans, E., Delanghe, J., De Bruijn, C., et al. (2003). Inflammation markers in relation to cognition in a healthy aging population. Journal of Neuroimmunology, 134(1-2), 142. Thadikkaran, L., Siegenthaler, M. A., Crettaz, D., Queloz, P. A., Schneider, P., & Tissot, J. D. (2005). Recent advances in blood-related proteomics. Proteomics, 5(12), 3019. Thio, C. L., Mosbruger, T., Astemborski, J., Greer, S., Kirk, G. D., O'Brien, S. J., et al. (2005). Mannose binding lectin genotypes influence recovery from hepatitis b virus infection. J Virol, 79(14), 9192-9196. Tirumalai, R. S., Chan, K. C., Prieto, D. A., Issaq, H. J., Conrads, T. P., & Veenstra, T. D. (2003). Characterization of the low molecular weight human serum proteome. Molecular Cellular Proteomics, 2(10), 1096. Toda, T. (2000). Current status and perspectives of proteomics in aging research. Experimental Gerontology, 35(6-7), 803. Toda, T., Ishijima, Y., Matsushita, H., Yoshida, M., & Kimura, N. (1994). Detection of thymopoietin-responsive proteins in nude mouse spleen cells by two-dimensional polyacrylamide gel electrophoresis and image processing. Electrophoresis, 15(7), 984-987. Topinkova, E. (2008). Aging, disability and frailty. Annals of Nutrition & Metabolism, 52 Suppl 1, 6. Tuitoek, P. J., Ritter, S. J., Smith, J. E., & Basu, T. K. (1996). Streptozotocin-induced diabetes lowers retinol-binding protein and transthyretin concentrations in rats. The British journal of nutrition, 76(6), 891. Turner, M. W. (2003). The role of mannose-binding lectin in health and disease. Molecular Immunology, 40(7), 423. Umans, L., Serneels, L., Overbergh, L., Stas, L., & Van Leuven, F. (1999). Alpha2- macroglobulin- and murinoglobulin-1- deficient mice. A mouse model for acute pancreatitis. Am J Pathol, 155(3), 983-993. Valera, A., Rodriguez-Gil, J. E., Yun, J. S., McGrane, M. M., Hanson, R. W., & Bosch, F. (1993). Glucose metabolism in transgenic mice containing a chimeric p- enolpyruvate carboxykinase/bovine growth hormone gene. FASEB J., 7(9), 791- 800. van den Berg, G., Veldhuis, J. D., Frolich, M., & Roelfsema, F. (1996). An amplitude- specific divergence in the pulsatile mode of growth hormone (GH) secretion underlies the gender difference in mean GH concentrations in men and premenopausal women. J Clin Endocrinol Metab, 81(7), 2460-2467. van der Hilst, J. C. H., Yamada, T., Op den Camp, H. J. M., van der Meer, J. W. M., Drenth, J. P. H., & Simon, A. (2008). Increased susceptibility of serum amyloid a 285

1.1 to degradation by mmp-1: Potential explanation for higher risk of type aa amyloidosis. Rheumatology, 47(11), 1651-1654. van Leuven, F., Torrekens, S., Overbergh, L., Lorent, K., de Strooper, B., & van den Berghe, H. (1992). The primary sequence and the subunit structure of mouse alpha-2-macroglobulin, deduced from protein sequencing of the isolated subunits and from molecular cloning of the cdna. European journal of biochemistry / FEBS, 210(1), 319. Van Leuven, F., Umans, L., Lorent, K., Hilliker, C., Serneels, L., Overbergh, L., et al. (1994). Molecular analysis of the human and mouse alpha 2m family. Annals of the New York Academy of Sciences, 737, 163. Vecchiet, J., Falasca, K., Cacciatore, P., Zingariello, P., Dalessandro, M., Marinopiccoli, M., et al. (2005). Association between plasma interleukin-18 levels and liver injury in chronic hepatitis c virus infection and non-alcoholic fatty liver disease. Annals of Clinical and Laboratory Science, 35(4), 415. Velloso, C. P. (2008). Regulation of muscle mass by growth hormone and IGF-I. Br J Pharmacol, 154(3), 557-568. Veniant, M. M., Beigneux, A. P., Bensadoun, A., Fong, L. G., & Young, S. G. (2008). Lipoprotein size and susceptibility to atherosclerosis--insights from genetically modified mouse models. Curr Drug Targets, 9(3), 174-189. Villanueva, J., Shaffer, D. R., Philip, J., Chaparro, C. A., Erdjument-Bromage, H., Olshen, A. B., et al. (2006). Differential exoprotease activities confer tumor- specific serum peptidome patterns. The Journal of clinical investigation, 116(1), 271. Villen, J., Beausoleil, S. A., Gerber, S. A., & Gygi, S. P. (2007). Large-scale phosphorylation analysis of mouse liver. Proc Natl Acad Sci U S A, 104(5), 1488- 1493. Vranckx, R., Rouaze-Romet, M., Savu, L., Mechighel, P., Maya, M., & Nunez, E. A. (1994). Regulation of rat thyroxine-binding globulin and transthyretin: Studies in thyroidectomized and hypophysectomized rats given tri-iodothyronine or/and growth hormone. The Journal of endocrinology, 142(1), 77. WADA. (2009). The world anti-doping code, the 2009 prohibited list, international standard. Wagner, T. E., Hoppe, P. C., Jollick, J. D., Scholl, D. R., Hodinka, R. L., & Gault, J. B. (1981). Microinjection of a rabbit beta-globin gene into zygotes and its subsequent expression in adult mice and their offspring. Proc Natl Acad Sci U S A, 78(10), 6376. Waldron, I., & Johnston, S. (1976). Why do women live longer than men? J Human Stress, 2(2), 19-30. Wallace, J. D., Cuneo, R. C., Baxter, R., Orskov, H., Keay, N., Pentecost, C., et al. (1999). Responses of the growth hormone (GH) and insulin-like growth factor axis to exercise, GH administration, and GH withdrawal in trained adult males: A potential test for GH abuse in sport. The Journal of clinical endocrinology and metabolism, 84(10), 3591. 286

Wallace, J. D., Cuneo, R. C., Bidlingmaier, M., Lundberg, P. A., Carlsson, L., Boguszewski, C. L., et al. (2001). The response of molecular isoforms of growth hormone to acute exercise in trained adult males. J Clin Endocrinol Metab, 86(1), 200-206. Wallace, J. D., Cuneo, R. C., Lundberg, P. A., Rosen, T., Jorgensen, J. O., Longobardi, S., et al. (2000). Responses of markers of bone and collagen turnover to exercise, growth hormone (GH) administration, and GH withdrawal in trained adult males. The Journal of clinical endocrinology and metabolism, 85(1), 124. Wang, M., Chen, M., Zheng, G., Dillard, B., Tallarico, M., Ortiz, Z., et al. (2008). Transcriptional activation by growth hormone of hnf-6-regulated hepatic genes, a potential mechanism for improved liver repair during biliary injury in mice. Am J Physiol Gastrointest Liver Physiol, 295(2), G357-366. Wang, X., Darus, C. J., Xu, B. C., & Kopchick, J. J. (1996). Identification of growth hormone receptor (ghr) tyrosine residues required for ghr phosphorylation and jak2 and stat5 activation. Molecular endocrinology (Baltimore, Md.), 10(10), 1249. Wang, Z., Masternak, M. M., Al-Regaiey, K. A., & Bartke, A. (2007). Adipocytokines and the regulation of lipid metabolism in growth hormone transgenic and calorie- restricted mice. Endocrinology, 148(6), 2845-2853. Wang, Z., Prins, G. S., Coschigano, K. T., Kopchick, J. J., Green, J. E., Ray, V. H., et al. (2005). Disruption of growth hormone signaling retards early stages of prostate carcinogenesis in the c3(1)/t antigen mouse. Endocrinology, 146(12), 5188-5196. Wang, Z., Yip, C., Ying, Y., Wang, J., Meng, X.-Y., Lomas, L., et al. (2004). Mass spectrometric analysis of protein markers for ovarian cancer. Clinical Chemistry, 50(10), 1939. Wanke, R., Milz, S., Rieger, N., Ogiolda, L., Renner-Muller, I., Brem, G., et al. (1999). Overgrowth of skin in growth hormone transgenic mice depends on the presence of male gonads. J Invest Dermatol, 113(6), 967-971. Wanke, R., Wolf, E., Brem, G., & Hermanns, W. (2001). [role of podocyte damage in the pathogenesis of glomerulosclerosis and tubulointerstitial lesions: Findings in the growth hormone transgenic mouse model of progressive nephropathy]. Verh Dtsch Ges Pathol, 85, 250-256. Warner, M. M., Guo, J., & Zhao, Y. (2001). The relationship between plasma apolipoprotein a-iv levels and coronary heart disease. Chin Med J (Engl), 114(3), 275-279. Waters, M. J., Hoang, H. N., Fairlie, D. P., Pelekanos, R. A., & Brown, R. J. (2006). New insights into growth hormone action. J Mol Endocrinol, 36(1), 1-7. Waxman, D. J. (2000). Growth hormone pulse-activated stat5 signalling: A unique regulatory mechanism governing sexual dimorphism of liver gene expression. Novartis Found Symp, 227, 61-74; discussion 75-81. Waxman, D. J., & Holloway, M. G. (2009). Centennial perspective: Sex differences in the expression of hepatic drug metabolizing enzymes. Mol Pharmacol, mol.109.056705. 287

Waxman, D. J., & O'Connor, C. (2006). Growth hormone regulation of sex-dependent liver gene expression. Mol Endocrinol, 20(11), 2613-2629. Westbrook, R., Bonkowski, M. S., Strader, A. D., & Bartke, A. (2009). Alterations in oxygen consumption, respiratory quotient, and heat production in long-lived ghrko and ames dwarf mice, and short-lived bgh transgenic mice. J Gerontol A Biol Sci Med Sci, 64(4), 443-451. Westwood, M., Aplin, J. D., Collinge, I. A., Gill, A., White, A., & Gibson, J. M. (2001). Alpha 2-macroglobulin: A new component in the insulin-like growth factor/insulin-like growth factor binding protein-1 axis. J Biol Chem, 276(45), 41668-41674. Widdowson, W. M., Healy, M.-L., Sönksen, P. H., & Gibney, J. (2009). The physiology of growth hormone and sport. Growth Hormone & IGF Research, In Press, Corrected Proof. Wildbrett, J., Hanefeld, M., Fucker, K., Pinzer, T., Bergmann, S., Siegert, G., et al. (1997). Anomalies of lipoprotein pattern and fibrinolysis in acromegalic patients: Relation to growth hormone levels and insulin-like growth factor i. Exp Clin Endocrinol Diabetes, 105(6), 331-335. Wilson, P. W., Anderson, K. M., Harris, T., Kannel, W. B., & Castelli, W. P. (1994). Determinants of change in total cholesterol and HDL-c with age: The framingham study. J Gerontol, 49(6), M252-257. Witte, D. P., Aronow, B. J., Stauderman, M. L., Stuart, W. D., Clay, M. A., Gruppo, R. A., et al. (1993). Platelet activation releases megakaryocyte-synthesized apolipoprotein j, a highly abundant protein in atheromatous lesions. The American journal of pathology, 143(3), 763. Wiwi, C. A., & Waxman, D. J. (2004). Role of hepatocyte nuclear factors in growth hormone-regulated, sexually dimorphic expression of liver cytochromes p450. Growth Factors, 22(2), 79-88. Wolf, E., Kahnt, E., Ehrlein, J., Hermanns, W., Brem, G., & Wanke, R. (1993). Effects of long-term elevated serum levels of growth hormone on life expectancy of mice: Lessons from transgenic animal models. Mech Ageing Dev, 68(1-3), 71-87. Wong, A. O., Zhou, H., Jiang, Y., & Ko, W. K. (2006). Feedback regulation of growth hormone synthesis and secretion in fish and the emerging concept of intrapituitary feedback loop. Comp Biochem Physiol A Mol Integr Physiol, 144(3), 284-305. Wong, W.-M. R., Gerry, A. B., Putt, W., Roberts, J. L., Weinberg, R. B., Humphries, S. E., et al. (2007). Common variants of apolipoprotein a-iv differ in their ability to inhibit low density lipoprotein oxidation. Atherosclerosis, 192(2), 266. Woodhouse, L. J., Mukherjee, A., Shalet, S. M., & Ezzat, S. (2006). The influence of growth hormone status on physical impairments, functional limitations, and health-related quality of life in adults. Endocr Rev, 27(3), 287-317. Wu, Z., Bidlingmaier, M., Dall, R., & Strasburger, C. J. (1999). Detection of doping with human growth hormone. Lancet, 353(9156), 895. Yakar, S., Liu, J. L., Stannard, B., Butler, A., Accili, D., Sauer, B., et al. (1999). Normal growth and development in the absence of hepatic insulin-like growth factor i 10.1073/pnas.96.13.7324. PNAS, 96(13), 7324. 288

Yamada, T. (2006). [serum amyloid a (saa)--pathogenicity and implication of appearance in plasma]. Rinsho Byori, 54(5), 509-512. Yang, C. W., Striker, L. J., Pesce, C., Chen, W. Y., Peten, E. P., Elliot, S., et al. (1993). Glomerulosclerosis and body growth are mediated by different portions of bovine growth hormone. Studies in transgenic mice. Lab Invest, 68(1), 62-70. Yoon, J. C., Puigserver, P., Chen, G., Donovan, J., Wu, Z., Rhee, J., et al. (2001). Control of hepatic gluconeogenesis through the transcriptional coactivator pgc-1. Nature, 413(6852), 131. Yuichiro, Y., Shuji, T., Tsuyoshi, I., Koji, A., Yohei, U., Yoshio, M., et al. (2006). Proteomic analysis of serum marker proteins in recipient mice with liver cirrhosis after bone marrow cell transplantation. PROTEOMICS, 6(8), 2564-2570. Yuri Miura, M. K. K. A. S. U. S. S. T. T. (2005). Age-dependent variations of cell response to oxidative stress: Proteomic approach to protein expression and phosphorylation. ELECTROPHORESIS, 26(14), 2786-2796. Zanivan, S., Gnad, F., WickstroÌˆm, S. A., Geiger, T., Macek, B., Cox, J. r., et al. (2008). Solid tumor proteome and phosphoproteome analysis by high resolution mass spectrometry. Journal of Proteome Research, 7(12), 5314-5326. Zargarova, O. P., & Konnova, L. A. (1983). [proteinase inhibitor activity of blood plasma and cerebrospinal fluid in patients with pituitary adenomas before and after proton therapy]. Vopr Med Khim, 29(6), 96-98. Zha, Y., Le, V. T., Higami, Y., Shimokawa, I., Taguchi, T., & Razzaque, M. S. (2006). Life-long suppression of growth hormone-insulin-like growth factor i activity in genetically altered rats could prevent age-related renal damage. Endocrinology, 147(12), 5690-5698. Zhan, X., Giorgianni, F., & Desiderio, D. M. (2005). Proteomics analysis of growth hormone isoforms in the human pituitary. Proteomics, 5(5), 1228. Zhang, R., Barker, L., Pinchev, D., Marshall, J., Rasamoelisolo, M., Smith, C., et al. (2004). Mining biomarkers in human sera using proteomic tools. Proteomics, 4(1), 244. Zhang, X., Mehta, R. G., Lantvit, D. D., Coschigano, K. T., Kopchick, J. J., Green, J. E., et al. (2007). Inhibition of estrogen-independent mammary carcinogenesis by disruption of growth hormone signaling. Carcinogenesis, 28(1), 143-150. Zhou, Y., Xu, B. C., Maheshwari, H. G., He, L., Reed, M., Lozykowski, M., et al. (1997). A mammalian model for laron syndrome produced by targeted disruption of the mouse growth hormone receptor/binding protein gene (the laronâ€‰mouse). Proceedings of the National Academy of Sciences of the United States of America, 94(24), 13215-13220.

289

APPENDIX A: IDENTIFICATION OF PLASMA PROTEINS

Table 15. MS and MS/MS scores of proteins Protein name Isoform Accession Protein Protein Total Total ion Full length Start-end position of URL link to # # a score b score ion score C.I. /start-end (of peptides matched by MS/MS match C.I.% c score d % e protein) f MS/MS (ion score, ion sequence data score C.I.%) g h apoA4 1 P06728 174 100 154 100 395/21-395 80-97 (51, 99.607%) 317-326 (53, 99.751%) apoA4 2 P06728 543 100 476 100 395/21-395 52-65 (69, 99.994%) 80-97 (108, 100%) 155-163 (45, 98.361%) 295-305 (60, 99.954%) 317-326 (67, 99.989%) 327-345 (106, 100%) apoA4 3 P06728 301 100 395/21-395 210-221(62, 100%) http://www.matr 295-304(68, 100%) ixscience.com/c 295-305(62, 100%) gi/protein_view. 317 - 326 (85, 100%) pl?file=../data/2 0090620/FtipSz emT.dat&hit=A POA4_MOUSE &px=1&ave_thr esh=34&_sigthr eshold=0.05&_s erver_mudpit_s witch=0.001 apoA4 5 P06728 147 100 395/21-395 52-65 (70, 100%) http://www.matr 80-97 (70, 100%) ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSz emO.dat&hit=A POA4_MOUSE &px=1&ave_thr 290

esh=33&_sigthr eshold=0.05&_s erver_mudpit_s witch=0.001 A2m 1 Q61838 45 73.165 45 97.141 1495/1240-1495 1358-1373 (45, 97.141%) (35kDa subunit) A2m 2 Q61838 219 100 219 100 1495/1240-1495 1358-1373 (41, 96.875%) (35kDa subunit) 1418-1426 (44, 98.532%) 1427-1440 (57, 99.925%) A2m 3 Q61838 293 100 293 100 1495/1240-1495 1307-1315 (45, 98.654%) (35kDa subunit) 1323-1333 (60, 99.963%) 1358-1373 (48, 99.411%) 1418-1426 (40, 95.654%) 1427-1440 (63, 99.979%) A2m 4 Q61838 277 100 277 100 1495/1240-1495 1323-1333 (48, 99.381%) (35kDa subunit) 1358-1373 (73, 99.998%) 1418-1426 (42, 97.978%) 1427-1440 (59, 99.96%)

A2m 5 Q61838 184 100 180 100 1495/1240-1495 1323-1333 (46, 99.047%) (35kDa subunit) 1418-1426 (53, 99.825%) 1427-1440 (47, 99.235%) A2m 6 Q61838 128 100 128 100 1495/1240-1495 1427-1440 (40, 95.834%) (35kDa subunit) A2m 7 Q61838 233 100 230 100 1495/1240-1495 1427-1440 (49, 96.958%) (35kDa subunit) 1441-1459 (76, 99.994%) Clusterin 1 Q06890 176 100 159 100 448/22-448 182-193 (52, 99.601%) (β-chain) 227-447 (α- 198-213 (53, 99.648%) Clusterin 2 Q06890 233 100 218 100 chain) 167-181 (55, 99.781%) (β-chain) 22-226 (β-chain) 182-193 (58, 99.904%) 198-213 (73, 99.997%) Clusterin 3 Q06890 232 100 221 100 182-193 (77, 99.998%) (β-chain) 198-213 (72, 99.995%) Clusterin 4 Q06890 250 100 241 100 182-193 (59, 99.92%) (β-chain) 198-213 (73, 99.997%) Clusterin 5 Q06890 51 99.9 198-213 (14) http://www.matr (β-chain) 214-221 (37, 97%) ixscience.com/c 291

gi/protein_view. pl?file=../data/2 0090620/FtipSz eee.dat&hit=CL US_MOUSE&p x=1&ave_thresh =34&_sigthresh old=0.05&_serv er_mudpit_switc h=0.001 Clusterin 6 Q06890 129 100 116 100 198-213 (51, 99.675%) (β-chain) apoE 1 P08226 290 100 278 100 311/19-311 87-100 (87, 100%) 262-270 (62, 99.97%) 271-284 (62, 99.97%) apoE 2 P08226 206 100 198 100 311/19-311 87-100 (56, 99.915%) 271-284 (55, 99.892%) apoE 3 P08226 453 100 406 100 311/19-311 87-100 (108, 100%) 130-144 (75, 99.999%) 262-270 (66, 99.991%) 271-284 (80, 100%) apoE 4 P08226 256 100 311/19-311 87-100 (127, 100) http://www.matr 114-122 (38, 98%) ixscience.com/c 130-144 (85, 100) gi/protein_view. 130-144 (92, 100) pl?file=../data/2 0090620/FtipSz eeT.dat&hit=AP OE_MOUSE&p x=1&ave_thresh =34&_sigthresh old=0.05&_serv er_mudpit_switc h=0.001 apoE 5 P08226 79 99.988 72 99.997 311/19-311 43-48 (11, 0) 87-100 (21, 0) 262-270 (16, 0) 292

271-284 (23, 0) MBP-C 1 P41317 124 100 112 100 244/19-244 97-112 (43, 84.305%) 149-162 (46, 92.76%) 176-186 (23, 0) MBP-C 2 P41317 244/19-244 149-162 (14) MBP-C 3 P41317 248 100 230 100 244/19-244 97-112 (77, 99.994%) 149-162 (69, 99.959%) 176-186 (55, 98.799%) Ig kappa V-III 1 P01654 57 84.617 51 98.683 111/1-111 1-18 (51, 98.683%)

Ig kappa V-III 2 P01654 56 79.721 48 98.437 111/1-111 1-18 (41, 93.245%) [1-18 (105, 100%) 50-65 (71, 99.994)] apoA1 1 Q00623 488 100 264/25-264 34 – 46 (87, 100%) http://www.matr 93 – 100 (40, 98.5%) ixscience.com/c 120 – 129 (48, 99.8%) gi/protein_view. 131 – 139 (62, 100%) pl?file=../data/2 142 – 154 (102, 100) 0090620/FtipSz 164 – 172 (61, 100) emt.dat&hit=AP OA1_MOUSE& px=1&ave_thres h=34&_sigthres hold=0.05&_ser ver_mudpit_swit ch=0.001 apoA1 2 Q00623 264/25-264 apoA1 3 Q00623 264/25-264 apoA1 5 Q00623 213 100 264/25-264 142 - 154 (80, 100) http://www.matr 164 - 172 (63, 100) ixscience.com/c 206 - 216 (54, 100) gi/protein_view. pl?file=../data/2 0090620/FtipSz eOt.dat&hit=AP OA1_MOUSE& px=1&ave_thres 293

h=34&_sigthres hold=0.05&_ser ver_mudpit_swit ch=0.001 apoA1 6 Q00623 38 0 157 100 264/25-264 93-106 (98, 100%) http://www.matr 142-154 (43, 99.4%) ixscience.com/c 164-174 (16, 0) gi/protein_view. pl?file=../data/2 0090708/Ftiumb uaS.dat&hit=AP OA1_MOUSE& px=1&ave_thres h=33&_sigthres hold=0.05&_ser ver_mudpit_swit ch=0.001 Prx-2 1 Q61171 266 100 215 100 198/2-198 11-29 (60, 99.677%) 92-109 (67, 99.933%) 140-150 (49, 95.38%) Albumin 1 P07724 105 100 96 100 608/25-608 243-257 (53, 98.365%) 43-57 (38, 53.801%) Albumin 2 P07724 233 100 220 100 608/25-608 35-44 (48, 95.472%) 45-57 (46, 91.588%) 243-257 (93, 100%) Albumin 3 P07724 73 99.639 73 99.987 608/25-608 243-257 (39, 67.025%) Albumin 4 P07724 61 94.285 51 94.74 608/25-608 243-257 (46, 93.437%) albumin 5 P07724 243 100 608/25-608 439 – 452 (93, 100%) http://www.matr 585 – 602 (96, 100%) ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSz eTO.dat&hit=A LBU_MOUSE& px=1&ave_thres h=34&_sigthres hold=0.05&_ser 294

ver_mudpit_swit ch=0.001 albumin 6 P07724 80 99.99 70 99.996 608/25-608 45-57 (63, 99.981%) albumin 7 P07724 65 99.71 52 99.76 608/25-608 45-57 (58, 99.941%) albumin 8 P07724 55 97.44 49 99.51 608/25-608 45-57 (55, 99.878%) albumin 9 P07724 58 98.66 49 99.45 608/25-608 45-57 (55, 99.863%) albumin 10 P07724 608/25-608 25-34 (9, 0) http://www.matr ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSz eTL.dat&hit=A LBU_MOUSE& px=1&ave_thres h=34&_sigthres hold=0.05&_ser ver_mudpit_swit ch=0.001 albumin 11 P07724 69 99.88 59 99.94 608/25-608 45-57 (46, 98.902%) albumin 12 P07724 66 99.78 60 99.96 608/25-608 45-57 (55, 99.858%) albumin 13 P07724 48 84.56 40 95.66 608/25-608 45-57 (27, 13.3%) albumin 14 P07724 67 99.84 58 99.96 608/25-608 470-483 (20, 0) 509-524 (16, 0) 585-602 (21, 0) albumin 15 P07724 423 100 416 100 608/25-608 470-483 (115, 100%) 528-543 (57, 99.858%) 570-584 (51, 99.383%) 585-602 (133, 100%) 585-602 (69, 99.99%) 295

albumin 16 P07724 83 99.996 73 99.998 608/25-608 [470-483 (67, 99.994%)] 337 100 328 100 [470-483 (102, 100%) 509-524 (55, 99.78%) 585-602 (147, 100%)] albumin 17 P07724 98 100 94 100 608/25-608 [470-483 (78, 100%)] 174 100 171 100 [470-483 (89, 100%) 585-602 (59, 99.947)] Albumin 18 P07724 110 100 97 100 608/25-608 35-44 (60, 99.936%) PEX-14 Q9R0A0 Albumin 19 P07724 170 100 114 100 608/25-608 35-44 (43, 96.091) PEX-14 45-57 (65, 99.975) albumin 20 P07724 106 100 99 100 608/25-608 586-602 (70, 99.992%) albumin 21 P07724 153 100 126 100 608/25 45-57 (44, 96.917%) albumin 22 P07724 65 99.692 54 99.788 608/25 45-57 (37, 87.613%) albumin 23 P07724 96 100 83 100 608/25 45-57 (58, 99.892%)

RBP-4 1 Q00724 49 100 201/19-201 158-167 (24) http://www.matr 172-181 (26) ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSzs nR.dat&hit=RE T4_MOUSE&p x=1&ave_thresh =33&_sigthresh old=0.05&_serv er_mudpit_switc h=0.001 RBP-4 2 Q00724 91 100 201/19-201 158-167 (53) http://www.matr 172-181 (38) ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSzs tt.dat&hit=RET 296

4_MOUSE&px =1&ave_thresh= 34&_sigthreshol d=0.05&_server _mudpit_switch =0.001 RBP-4 3 Q00724 108 100 92 100 201/19-201 84-140 (55, 99.025%) 118-139 (22, 0) HP 2 Q61646 110 100 90 100 347/19-347 58-72 (34, 68.345%) (α-chain) 19-101 (α) 103-347 (β) HP 3 Q61646 59 91.546 50 99.193 347/19-347 58-72 (31, 28.094%) (α-chain) 19-101 (α) 103-347 (β) HP 4 Q61646 31 0 8 0 347/19-347 58-72 (8, 0) (α-chain) 19-101 (α) 103-347 (β) TTR 1 P07309 179 100 154 100 147/27-147 101-123 (101, 100%)

TTR 2 P07309 239 100 215 100 147/27-147 56-68 (50, 99.382%) 101-123 (129, 100%) TTR 4 P07309 120 100 147/27-147 101-123 (106, 100%) http://www.matr 56-68 (14) ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSz eSm.dat&hit=T THY_MOUSE &px=1&ave_thr esh=32&_sigthr eshold=0.05&_s erver_mudpit_s witch=0.001 TTR 5 P07309 680 100 639 100 147/27-147 55-68 (48, 98.736%) 56-68 (76, 99.998%) 69-90 (160, 100%) 84-96 (74, 99.996%) 297

101-123 (161, 100%) 124-147 (84, 100%) TTR 6 P07309 384 100 355 100 147/27-147 56-68 (57, 99.888%) 69-90 (77, 99.999%) 101-123 (171, 100%) TTR 7 P07309 38 0 336 100 147/27-147 56-68 (80, 100%) http://www.matr 69-90 (168, 100%) ixscience.com/c 124-147 (187, 100%) gi/protein_view. pl?file=../data/2 0090620/FtipSz eES.dat&hit=TT HY_MOUSE&p x=1&ave_thresh =32&_sigthresh old=0.05&_serv er_mudpit_switc h=0.001 TTR 8 P07309 144 100 147/27-147 56-68 (54, 100%) http://www.matr 101-123 (91, 100%) ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSz eEt.dat&hit=TT HY_MOUSE&p x=1&ave_thresh =33&_sigthresh old=0.05&_serv er_mudpit_switc h=0.001 TTR 9 P07309 177 100 95 100 147/27-147 101-123 (95, 100%) SAA-1 1 P05366 59 98.907 54 99.71 122/20-122 81-89 (54, 99.71%)

Hbb (1 and 2) 1 P02088 256 100 179 100 147/2-147 32-41 (59, 99.889%) P02089 68-83 (75, 99.997%) Hbb (1 and 2) 2 P02088 327 100 301 100 147/2-147 32-41 (57, 99.859%) P02089 118 100 113 100 68-83 (48, 98.842%) 298

42-60 (100, 100%)

Hbb (1 and 2) 3 P02088 686 100 147/2-147 32 - 41 (75, 100) http://www.matr P02089 42 - 60 (201, 100) ixscience.com/c 42 - 60 (139, 100) gi/protein_view. 67 - 83 (117, 100) pl?file=../data/2 68 - 83(123, 100) 0090620/FtipSz 106 - 121 (113, 100) emS.dat&hit=H 106 - 121 (94, 100) BB1_MOUSE& 122 - 133 (59, 100) px=1&ave_thres h=33&_sigthres hold=0.05&_ser ver_mudpit_swit ch=0.001 Hbb (1 and 2) 4 P02088 659 100 147/2-147 32 - 41 (66, 100) http://www.matr P02089 42 - 60 (172, 100) ixscience.com/c 67 - 83 (102, 100) gi/protein_view. 68 - 83 (113, 100) pl?file=../data/2 84 - 96 (75, 100) 0090620/FtipSz 106 - 121 (85, 100) emE.dat&hit=H 106 - 121 (96, 100) BB1_MOUSE& 122 - 133 (37, 97.3%) px=1&ave_thres h=33&_sigthres hold=0.05&_ser ver_mudpit_swit ch=0.001 apoA2 1 P09813 43 59.383 36 90.192 102/24-102 68-77 (36, 90.192%) apoA2 2 P09813 69 100 102/24-102 68-77 (69, 100%) http://www.matr ixscience.com/c gi/protein_view. pl?file=../data/2 0090620/FtipSz emL.dat&hit=A POA2_MOUSE &px=1&ave_thr esh=33&_sigthr 299

eshold=0.05&_s erver_mudpit_s witch=0.001 a. Accession number in UniProtKB/Swiss-Prot database. b. The MOWSE score is calculated by the Mascot search engine for each protein matched from the MS peak list. This score is based on the probability that peptide mass matches are non-random events. If the protein score is equal to or greater than the Mascot® Significance Level calculated for the search, the protein match is considered to be statistically non-random at the 95% confidence interval. The analysis used combined MS and MS/MS. c. The Confidence Interval (C.I. %) for the protein score d. A score calculated by weighting ion scores for all individual peptides matched to a given protein. e. The C.I. % for the total ion score. f. Total number of amino acids (in the full length protein)/position of start to end sequence (of the mature/secreted form of the protein). For example, 300/23-300 means the protein is 300 amino acids long, and the mature or secreted protein is from 23rd to 300th amino acids. g. Start to end position of protein sequences matched by MS/MS with a C.I.%>95% except for a few. For example, 130-142 (59, 99.999%) means sequence 130th to 142nd amino acids is matched by MS/MS with a C.I.% of 99.999%. h. Some proteins did not have good MS scores; however, their MS/MS scores were good enough to consider a confident identification. Web links to MS/MS data of these proteins are provided. 300

Table 16. LC/MS/MS identification of selected proteins

Protein Isoform Accession Unused Total Coverage Confidence Sequence e Modification f name # # a score b score c % d Ig kappa C 3 P01654 4.05 4.05 38.7 99.99 QNGVLNSWTDQDSK Gln->pyro-Glu@N- term WKIDGSER DSTYSMSSTLTLTKD Dethiomethyl(M)@6 EYER DSTYSMSSTLTLTKD Oxidation(M)@6 EYER DSTYSMSSTLTLTKD EYER QNGVLNSWTDQDSK Deamidated(N)@2 WKIDGSER Oxidation(W)@1 WKIDGSER FormaldehydeAdduct( W)@1 TTR 3 P07309 4 4 19.0 99.99 TAESGELHGLTT DEK TSEGSWEPFASG K

Hp 1 Q61646 3 3 7.8 99.9 LRAEGDGVYTLNDE (α-chain) K LPECEAVCGKPK Carbamidomethyl(C) @4; Carbamidomethyl(C) @8

a. Accession number in UniProtKB/Swiss-Prot database. b. Unused score is a measurement of all the peptide evidence for a protein that is not better explained by a higher ranking protein. It is the true indicator of protein confidence. A score of 2= 99 % confidence, a score of 4=99.99% confidence, etc. 301 c. Total score is measurement of all the peptide evidence for a protein and is analogous to protein scores reported by other protein identification software. d. Sequence coverage % of the matched peptides to the full-length protein e. Sequences matched by LC-MS/MS f. Amino acid modification within the matched peptide in ‘sequence’ column.

302

Table 17. Mw and pI of proteins Protein name Isoform # Theoretical Mw Observed Mw Theoretical pI c Observed pI d Known PTMs e (kDa) a (kDa) b

apoA4 1 43.0 43 5.27 5.64 S333-p apoA4 2 43.0 43 5.27 5.7 (Zanivan et al., apoA4 3 43.0 43 5.27 5.76 2008) apoA4 4 43.0 43 5.27 5.84 apoA4 5 43.0 43 5.27 5.9 A2m 1 28.2 38.3 7.09 5.65 Glycosylation (35kDa subunit) 2 28.2 38.7 7.09 5.85 (1385, 1443) 3 28.2 38.5 7.09 6.08 Phosphorylation

4 28.2 38 7.09 6.3 5 28.2 37.7 7.09 6.54 6 28.2 37.5 7.09 6.80 7 28.2 37.6 7.09 7.08 Clusterin 1 23.9 36.3 6.12 5.28 N-glycosylation (β-chain) 2 23.9 36.1 6.12 5.37 at 102,144 3 23.9 35.8 6.12 5.47 4 23.9 34.4 6.12 5.59 5 23.9 34.3 6.12 5.69 6 23.9 34.1 6.12 5.77 apoE 1 34.0 37.8 5.46 5.76 Phosphorylation apoE 2 34.0 34.7 5.46 5.89 Q308-p apoE 3 34.0 34 5.46 6 (Raftery et al., 2005) apoE 4 34.0 34.6 5.46 6.13 apoE 5 34.0 35.1 5.46 6.28 MBP-C 1 24.0 31.7 4.96 5.65 Glycosylation 303

MBP-C 2 24.0 31.7 4.96 5.75 Hydroxylation (proline) MBP-C 3 24.0 31.6 4.96 5.9 Ig kappa V-III 1 * 30 * 5.75 2 * 30.6 * 6.03 Ig kappa C 3 11.8 30.8 5.23 6.38 apoA1 1 27.9 28.1 5.42 5.73 Phosphorylation apoA1 2 27.9 28 5.42 5.8 at S56,S60,T75 apoA1 3 27.9 28.1 5.42 5.93 (H. Li et al., 2009) apoA1 4 27.9 28.8 5.42 6.07 apoA1 5 27.9 29.3 5.42 6.22 apoA1 6 27.9 29.5 5.42 6.4 Prx-2 1 21.6 25.4 5.2 6.11 Acetylation at A2 phosphorylation RBP-4 1 21.4 22.5 5.98 6.74 RBP-4 2 21.4 22.3 5.98 7.13 RBP-4 3 21.4 22 5.98 7.61 RBP-4 4 21.4 19.5 5.98 5.28 RBP-4 5 21.4 19.1 5.98 5.49 RBP-4 6 21.4 19.2 5.98 5.75 RBP-4 7 21.4 19.4 5.98 6.08 Albumin 1 65.9 31.7 5.53 6.14 Phosphorylation Albumin 2 65.9 32.7 5.53 6.29 Cleavage on Albumin 3 65.9 32.6 5.53 6.42 pair of basic residuals Albumin 4 65.9 32.6 5.53 6.58 Albumin 5 65.9 29.7 5.53 6.78 Albumin 6 65.9 22.6 5.53 5.75 Albumin 7 65.9 22.6 5.53 5.86 304

Albumin 8 65.9 23.2 5.53 6.02 Albumin 9 65.9 23.2 5.53 6.16 Albumin 10 65.9 23.1 5.53 6.32 Albumin 11 65.9 16.6 5.53 5.57 Albumin 12 65.9 16.8 5.53 5.72 Albumin 13 65.9 16.9 5.53 5.87 Albumin 14 65.9 16.7 5.53 6.1 Albumin 15 65.9 16.2 5.53 6.4 Albumin 16 65.9 16.3 5.53 6.74 Albumin 17 65.9 16 5.53 7.12 Albumin 18 65.9 11.1 5.53 5.62 Albumin 19 65.9 11.3 5.53 5.78 Albumin 20 65.9 11.6 5.53 6.2 Albumin 21 65.9 9.1 5.53 6 Albumin 22 65.9 9 5.53 6.34 Albumin 23 65.9 9.2 5.53 6.6 HP 1 9.5 14.3 4.77 5.2 Glycosylation (α-chain) on beta chain 2 9.5 14 4.77 5.4 3 9.5 14 4.77 5.6 4 9.5 14 4.77 5.85 TTR 1 13.6 41.2 5.77 6.48 N- TTR 2 13.6 40.1 5.77 6.85 Glycosylation at TTR 3 13.6 20.6 5.77 6.47 118 (Bernhard et al., 2007) TTR 4 13.6 14 5.77 6.08 TTR 5 13.6 14 5.77 6.5 TTR 6 13.6 14.3 5.77 6.66 TTR 7 13.6 13.8 5.77 6.86 TTR 8 13.6 13.8 5.77 7.24 305

TTR 9 13.6 14.5 5.77 7.79 SAA-1 1 11.8 11.5 6.07 6.8 Hbb 1 15.7 11.4 7.26 7.24 phosphorylation Hbb 2 15.7 11.8 7.26 7.47 Hbb 3 15.7 11.5 7.26 7.64 Hbb 4 15.7 12.6 7.26 7.92 apoA2 1 8.9 8.5 5.25 5.26 Cleavage on apoA2 2 8.9 9 5.25 5.68 pair of basic residues

a. Unless otherwise indicated, for secreted proteins, signal peptide was removed for calculation of theoretical Mw. In case of a protein with multiple subunits, Mw of the subunit instead of the full-length protein was calculated. Mw and pI were calculated using ‘ScanSite pI/Mw’ tool from ExPaSy website (http://scansite.mit.edu/calc_mw_pi.html). Signal peptide was determined by Swiss-Prot/TrEMBL protein database (http://ca.expasy.org). b. Observed Mw was an estimation from 2-D gel c. Theoretical pI assumes no modification of the protein. d. Observed pI was estimated from 2-D gel. e. Unless otherwise indicated, known PTMs of proteins were reported in Swiss-Prot/TrEMBL protein database (http://ca.expasy.org). * Because of the heterogeneous nature of Ig kappa variable region, no theoretical Mw or pI are given.

306

Figure 84. Identification of a2m (isoform 4) by MS/MS. A, total ion score of MS/MS. X axis: probability based mowse score (an algorithm defined by MASCOT), y axis: number of hits (matched protein in the database). As shown, one hit (a2m) in the white 307 region has extensive homology to the target protein spot. B, sequences matched by MS/MS shown in red letters. Blue box represents sequence of 35 kDa subunit of a2m; yellow box represents the signal peptide; the rest of the sequence corresponds to 165 kDa subunit of a2m. C, individual ion scores. Ten peptides are shown. The mass errors in Da and ppm are graphed below, with each red dot corresponding to one peptide.

308

APPENDIX B: ORIGINAL GEL IMAGES

Figure 85. Representative gel images of WT and bGH mice at different ages. 309

Figure 86. Snap shot of all 2D gels of WT mice at different ages.

Figure 87. Snap shot of all 2D gels of bGH mice at different ages.

310

Figure 88. Representative gel images of WT and GHR-/- mice of both genders at different ages.

311

Figure 89. Snap shot of all 2D gels of male WT and GHR-/- mice at different ages.

Figure 90. Snap shot of all 2D gels of female WT and GHR-/- mice at different ages. 312

Figure 91. Representative gel images of control, GH, IGF-1 injections and washout groups. Male mice of 3 months were used.

Figure 92. Snap shot of all 2D gels of mice in the injection study. Mice were divided into groups receiving saline, GH, IGF-1, GH then washout and IGF-1 then washout. Male mice of 3 months were used (n=7).