Determination of Biologically Relevant Vitamin D Metabolites in a Mouse Model of Non-Melanoma Skin Cancer

THESIS

Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in the Graduate School of The Ohio State University

Matthew D. Teegarden

Graduate Program in Food Science and Technology

The Ohio State University

2014

Master's Examination Committee:

Dr. Steven J. Schwartz, Adviser

Dr. Yael Vodovotz

Dr. Steven K. Clinton

Copyrighted by

Matthew Daniel Teegarden

2014

Abstract

Vitamin D is most notably associated with bone health, but a growing body of evidence suggests that it may also have a role in certain chronic diseases. An individual can produce sufficient levels of vitamin D via unprotected sun exposure, but this exposure is also a risk factor for non-melanoma skin cancer (NMSC). Despite this relationship, cell and animal studies suggest that vitamin D may have an important inhibitory role in the development of NMSC. The effects of consuming dietary vitamin

D on this disease have not been investigated. Efforts to measure vitamin D status, 25- hydroxyvitamin D3 concentration, in biological microsamples by HPLC-MS/MS have been hampered by sensitivity thresholds and an interfering metabolite, C-3epi-25- hydroxyvitamn D3. Researchers commonly utilize chemical derivatization with 4- phenyl-3H-1,2,4-triazole-3,5(4H)-dione (PTAD) for signal enhancement, but the presence of C-3epi-25-hydroxyvitamn D3 remains a problem. The objectives of this research were to develop methodology for the analysis of PTAD-derivatized 25- hydroxyvitamin D3 and C-3epi-25-hydroxyvitamn D3 in skin and serum, and to apply this methodology to a study of dietary vitamin D in a mouse model of NMSC.

Chromatographic separation of PTAD-derivatized 25-hydroxyvitamin D3 and C-

3epi-25-hydroxyvitamn D3 was assessed using C18 columns of varying length and

ii particle size. Multiple skin and serum extraction methods were evaluated, and those determined to be best fit for the purpose of this analysis were selected for use in subsequent sample processing. For skin, a 5:1 hexane/dichloromethane extraction was utilized, and a simple solid phase extraction method was selected for serum processing.

Optimized chromatographic resolution was achieved with a Luna C18 (Phenomenex) 250 mm column (3 µm packing). Metabolite detection was optimized on a Qtrap 5500 mass spectrometer. This methodology will allow for more accurate assessment of vitamin D status in samples which require PTAD-derivatization for sensitive analysis.

Using this optimized method, 25-hydroxyvitamin D3 and C-3epi-25- hydroxyvitamn D3 levels were evaluated in the skin and serum of Skh-1 hairless mice.

Equal numbers of female (ntotal =75) and male (ntotal =75) mice were placed on diets with escalating doses of vitamin D3 (25, 150, 1000 IU) for 29 weeks. Within each dietary level, n=15 mice were exposed to UVB light, one minimal erythemal dose, three times weekly for the last 25 weeks of the study. Skin and serum of three mice from each gender/diet/UV group were assayed for 25-hydroxyvitamin D3 and C-3epi-25- hydroxyvitamn D3 (ntotal=36). The skin and serum levels of these metabolites rose in a dose-dependent manner. UV-exposed mice had lower serum levels of both metabolites.

Male mice had greater levels of C-3epi-25-hydroxyvitamn D3 at the highest dose of vitamin D3. A strong correlation between skin and serum C-3epi-25-hydroxyvitamn D3 levels was observed, suggesting skin may be a source of circulating C-3epi-25- hydroxyvitamn D3. Furthermore, a role of C-3epi-25-hydroxyvitamn D3 in NMSC was postulated when levels of this metabolite were put into context with the cancer outcomes

iii of this study. It is clear that dietary administration of vitamin D3 affects the development of NMSC, and this study will inform future work on the mechanisms by which these effects are modulated.

Dedication

Dedicated to my grandfather, Joe Bloom

Acknowledgments

I would like to sincerely thank my adviser Dr. Steven J. Schwartz for his guidance throughout my master’s degree and for providing me with opportunities that have helped me grow as a scientist. Dr. Yael Vodovotz, has also helped shape my experience in food science, and I am grateful for her not-so-subtle nudge for me to pursue graduate studies in the field, as well as her continued support. Furthermore, I would like to thank Dr. Steven

K. Clinton who has provided me with valuable advice over the past few years.

This project would not have been possible without the support I received from several individuals. I am indebted to the efforts of Dr. Ken Riedl who guided me through the process of understanding advanced analytical techniques, and my fellow labmates who continue to serve as a valuable support network. In particular, I would like to thank

Jessica Cooperstone and Dr. Jennifer Ahn-Jarvis for their support in this project. I have had the immense privilege to work with Drs. Tania Oberyszyn and Kathy Tober as collaborators on this project. Their enthusiasm was contagious and uplifting! I would also like to thank The Ohio State University Food Innovation Center and The Ohio State

University Comprehensive Cancer Center Molecular Carcinogenesis and

Chemoprevention Research Program for providing the funds that made this research

vi possible. Finally, I would like to thank my family and friends for their continued support.

vii

Vita

2012...... B.S. in Food Science with Honors Research Distinction in Food Science and Nutrition

2012 to 2013 ...... University Fellow

2013 to present ...... Graduate Research Associate, Department of Food Science and Technology

Fields of Study

Major Field: Food Science and Technology

viii

Table of Contents

Abstract ...... ii Dedication ...... v Acknowledgments...... vi Vita ...... viii Table of Contents ...... ix List of Figures ...... xii List of Tables ...... xiii Chapter 1: Literature Review ...... 1 1.1 What is Vitamin D? ...... 1 1.1.1 Dietary and Endogenous Vitamin D ...... 3 1.1.2 Vitamin D Metabolism and Status ...... 7 1.2 Biological Functions and Implications of Vitamin D ...... 8 1.2.1 Basic and Known Actions ...... 8 1.2.2 Vitamin D and Non-Melanoma Skin Cancer ...... 10 1.3 Vitamin D Analysis ...... 16 1.3.1 HPLC-MS/MS Analysis ...... 17 1.3.2 Advanced Techniques for Vitamin D Analysis in Biological Samples ...... 19 1.3.3 Vitamin D Analysis in Biological Tissues ...... 24 1.4 Specific Aims ...... 26 1.4.1 Specific Aim 1: To develop methodology suitable for the accurate evaluation of vitamin D status in murine skin...... 26 1.4.2 Specific Aim 2: Apply the method developed in the first aim to samples from a study in which the effects of dietary vitamin D on NMSC were evaluated...... 26 Chapter 2: Materials and Methods ...... 28

2.1 Animals and Experimental Design ...... 28 2.2 Chemicals ...... 31 2.3 Method Development ...... 31 2.4 Sample Preparation ...... 32 2.5 HPLC-MS/MS Analysis ...... 34 2.6 Statistical Analysis ...... 35 Chapter 3: Results and Discussion ...... 36 3.1 Method Development ...... 36 3.1.1 HPLC Column and Method Selection ...... 36 3.1.2 Skin and Serum Extraction Methods Selection ...... 41 3.1.3 Application of Selected Methods ...... 43

3.2 Quantification of 25OH D3 and C3epi in a Mouse Model of Non-Melanoma Skin Cancer...... 46

3.2.1 Serum and Skin Levels of 25OH D3 ...... 46 3.2.2 Serum and Skin levels of C3epi ...... 48 3.2.3 Relationships between Metabolite Levels in Skin and Serum ...... 51

3.2.4 Levels of 25OH D3 and C3epi in Relation to NMSC Outcomes...... 53 Chapter 4: Conclusions and Future Directions ...... 56 References ...... 58

Appendix A: Development of a Vitamin D2 Fortified Soy Bread for Use in Clinical Trials...... 72 A-1.1 Introduction: Rationale for dietary delivery of vitamin D with soy to prevent prostate cancer ...... 73 A-1.1.1 Vitamin D Analysis in Foods ...... 78 A-1.1.2 Specific Aim A1 To develop a food product rich in vitamin D for use in cancer prevention clinical trials...... 79 A-2.2 Materials and Methods ...... 80 A-2.2.1 Bread Baking ...... 80 A-2.2.2 HPLC-MS/MS Analysis ...... 82 A-2.2.3 Fortification Curve ...... 83 A-2.2.4 Homogeneity Study ...... 83 x

A-2.2.5 Vitamin D Retention over Baking ...... 84 A-3.1 Results and Discussion ...... 85 A-3.1.1 Fortification Curve ...... 85 A-3.1.2 Homogeneity Study ...... 87 A-3.1.3 Vitamin D Stability over Baking ...... 88 A-4.1 Conclusions ...... 89

List of Figures

Figure 1: Structures of vitamins D2 and D3...... 2

Figure 2: Photobiology and metabolism of vitamin D3...... 6 Figure 3: The hedgehog cycle and its components that are regulated by vitamin D...... 15

Figure 4: Derivatization reaction mechanism of 25OH D3 with PTAD...... 21 Figure 5: Design of the study evaluating effects of dietary vitamin D on NMSC development...... 30 Figure 6: Improvement of chromatographic resolution of a mixed standard containing

PTAD-derivatized 25OH D3 and C3epi with C18 columns of varying length and particle size ...... 38

Figure 7: Resolution of 6(R)-PTAD-25OHD3 and 6(S)-PTAD-25OHD3 from co-eluting stereoisomers of PTAD-C3epi ...... 40

Figure 8: Dose-dependent levels of 25OHD3 in skin and serum of mice fed increasing doses of vitamin D3...... 45

Figure 9: Serum and skin levels of C3epi in mice fed increasing amounts of vitamin D3...... 50

Figure 10: Correlation between skin and serum 25OH D3 ...... 52 Figure 11: Dietary vitamin D supplementation effects on chronic UVB induced tumor grade. (102)...... 55

Figure 12: Interaction of genistein and 1,25(OH)2 D3 on the prostaglandin pathway. ... 77 Figure 13. Vitamin D Soy Bread Baking Process ...... 81

Figure 14: Bread fortification curve relating amount of added vitamin D2 rich yeast added to bread dough and achieved dose per slice of bread...... 86

xii

List of Tables

Table 1: Relevant MS/MS parameters used for the analysis of vitamin D metabolites .. 35

Table 2: Levels of 25OHD3 and C3epi from a variety of matrices ...... 42

Table 3: Homogeneity study. Average vitamin D2 content of breads made from partitions of larger dough masses which were fortified at different points in the bread making process...... 87

xiii

Chapter 1: Literature Review

1.1 What is Vitamin D?

In the early 20th century, rickets was a major health concern in children in the developed world. Vitamin D was initially discovered in 1922 as a dietary factor in cod liver oil that could cure this disease (1). The connection between light and production of vitamin D in foods and skin was made three years later by Steenbock and colleagues (2).

Over subsequent decades, landmark discoveries in the chemistry of vitamin D were made including the isolation of purified vitamin D2 (3) and D3 (4), and eventually the characterization of the photobiology of vitamin D production (5). Discoveries on the mechanisms by which vitamin D prevents rickets were soon to follow. The history of vitamin D has recently been comprehensively reviewed by Hector DeLuca (6). During the past few years, vitamin D has received a great amount of attention, but not for its role in bone health. New research has suggested that this important compound may play a role in several other diseases beyond rickets. While many aspects of this are debated, there is a need for insight on what role vitamin D may play in disease prevention and if this nutrient is able to contribute to new methods of medical treatment.

Figure 1: Structures of vitamins D2 and D3.

1.1.1 Dietary and Endogenous Vitamin D

Vitamin D is a unique nutrient, and some would even venture to call it a

“pseudovitamin,” because one does not need to obtain vitamin D from the diet as it can be produced endogenously in skin. The recommended dietary allowance for vitamin D has been established at 600 IU/day for adults and children and 800IU/day for adults over the age of 70 (7). This represents a considerable change from the adequate intake criteria previously established for vitamin D, which ranged from 200 to 600 IU/day depending on age (8). The new guidelines were set with regard to bone health, as there is not yet sufficient evidence to support recommendations for other health outcomes. However, several researchers believe the new guidelines are still too low to support an elevated vitamin D status, which they believe is necessary for the prevention of multiple diseases beyond rickets and osteomalacia (9). The upper limit for vitamin D is set at 4000 IU, which is twice as high as the previous recommendation (7,8). Some health care providers advise patients to take doses of vitamin D above the upper limit. Rather than taking these megadoses daily, however, patients take them weekly or monthly and sometimes only for a limited period of time (10).

Dietary sources that naturally contain appreciable amounts of vitamin D are quite uncommon. Fatty fish, fish liver oil, and egg yolks are the few natural dietary sources of vitamin D3 (cholecalciferol), and mushrooms and yeast, especially when treated with

UVB light, are the natural sources of vitamin D2 (ergocalciferol) (7). Some foods in the

United States, as well as some other countries, are fortified with vitamin D, including milk, other dairy products, cereal products, and infant formulas. Foods may be

3 supplemented with vitamin D2 or D3, while D3 fortification is more common. The two forms of vitamin D (figure 1) differ only in their side chain, and are considered equally effective in preventing rickets (7). Although a small variety of foods naturally contain or are fortified with vitamin D, many people are turning to dietary supplements as a source of the vitamin. Commercially, vitamin D2 is synthesized by UVB irradiation of ergosterol from yeast, and D3 is obtained from UVB irradiation of 7-dehydrocholesterol from lanolin, an oil obtained from sheep’s wool (10). Although rare, it is possible to achieve toxic levels of vitamin D resulting in a condition known as hypervitaminosis D.

This condition can lead to hypercalcemia and calcification of soft tissue, which can further result in damage to renal and cardiovascular tissue (11).

Like other fat-soluble vitamins, it is suggested that co-consuming vitamin D with a meal that includes some lipid content increases its absorption (12). Vitamin D is primarily absorbed into chylomicrons, and thus enters circulation from the lymph.

Consequently, some vitamin D is adsorbed and highly retained in adipose tissue before it can be further metabolized. A minor amount of vitamin D is absorbed through the portal system with more polar dietary components (7).

Endogenous production of vitamin D begins with UVB irradiation (290-320 nm) of ergosterol (provitamin D2) or 7-dehydrocholesterol (provitamin D3), which forces the

B-ring of these steroidal pro-vitamin D compounds to break at the ninth carbon (figure

2). Pre-vitamin D compounds are thermally isomerized to form vitamin D, which is then excreted into circulation (13). It is not possible to produce toxic levels of vitamin D from overexposure to sunlight. Excessive UVB irradiation of vitamin D and its precursors 4 results in the formation of biologically inert byproducts, such as suprasterols, tachysterol, and lumisterol (7).

A number of factors influence the amount of vitamin D that may be produced through exposure to sunlight. Lifestyle factors, such as the amount of clothing an individual wears greatly influences vitamin D status, with more coverage of the body resulting in lower amounts of vitamin D produced (14). Additionally, sunscreen, if correctly applied, inhibits vitamin D production as it blocks UVB light (15). A number of biological factors influence vitamin D production. Darker skin pigmentation seems to decrease vitamin D production after exposure to sunlight (7), and as an individual ages, the amount of 7-dehydrocholesterol in the skin decreases, which considerably reduces capacity for vitamin D production by age 70 (16). Finally, environmental factors such as season and latitude affect UVB exposure, which in turn affects vitamin D production.

However, the effects of latitude seem to be less severe than seasonal variation in UVB exposure (7). It is estimated that a person in a bathing suit receiving one minimal erythemal dose of UVB irradiation, the minimal amount of UV radiation needed to produce a slight redness in skin after 24 hours (17), can produce adequate amounts of vitamin D (13).

Figure 2: Photobiology and metabolism of vitamin D3. Adapted from Holick, 1994 and

DeLuca (7,18).

1.1.2 Vitamin D Metabolism and Status

A fairly comprehensive review of vitamin D metabolism was published in 1983 by DeLuca and Schnoes (19), and although dated, the information is still frequently referenced today. Vitamin D synthesized in the skin is circulated bound to the vitamin D binding protein, while dietary vitamin D circulates predominantly bound to chylomicrons. When it reaches the liver, vitamin D is hydroxylated by 25-hydroxylase

(CYP2R1) to 25-hydroxyvitamin D (25OH D). This metabolite quickly enters circulation bound to the vitamin D binding protein. 25OH D can be metabolized to 1,25- dihydroxyvitamin D (1,25(OH)2 D) by 1α-hydroxylase (CYP27B1) primarily in the kidney. The formation of 1,25(OH)2 D is tightly regulated, as this is the biologically active form of the vitamin. Renal production of 1,25(OH)2 D is influenced by serum calcium and phosphorous, fibroblast growth factor, and parathyroid hormone levels (13).

High levels of 1,25(OH)2 D increases transcription of 24-hydroxylase (CYP24A1), which rapidly metabolizes the active form to biologically inactive calcitroic acid. This enzyme also metabolizes 25OH D to 24,25(OH)2 D, preventing its activation. 24,25(OH)2 D is further oxidized to 24-oxo-25(OH)2 D, which is excreted in the bile along with calcitroic acid. Other minor metabolites of vitamin D exist, but they have shown to be rather biologically inert (figure 2).

The pharmacokinetics of vitamin D toxicity have been recently reviewed (20), providing pertinent information on several aspects of vitamin D metabolism. The half- lives of vitamin D metabolites in serum are quite variable and depend greatly on the affinity of the different metabolites for the vitamin D binding protein. Serum 25OH D

7 has the longest half-life, approximately 15 days, as it has the greatest affinity for the vitamin D binding protein. In contrast to this, vitamin D has a half-life of less than 24 hours in serum, while circulating 1,25(OH)2 D has a half-life between 4 and 20 hours

(20,21). Additionally, levels of 1,25(OH)2 D are highly regulated, and circulating concentrations of this metabolite in mice are not elevated when a toxic threshold is reached (20). The fact that this metabolite does not seem to correlate with deficiency as well as its short serum half-life, and highly-regulated and low serum concentration detract from its use as a marker of status. The best marker of vitamin D status is 25OH

D, as it has a long half-life, circulates at detectable levels, is not highly regulated by more than vitamin D concentration, and correlates well with intake and synthesis of vitamin D

(21). Most researchers and clinicians agree that 25OH D serum levels less than 30 nmol/L indicate deficiency and optimal status with respect to bone health rests around 50 nmol/L (7). Some advocate that 80 nmol/L is a more acceptable optimum level to promote benefits beyond bone health, such as the prevention of chronic diseases (22).

The Institute of Medicine committee concluded that the evidence for such a recommendation was not yet adequate.

1.2 Biological Functions and Implications of Vitamin D

1.2.1 Basic and Known Actions

The most basic and well-known function of vitamin D is the maintenance of serum calcium and phosphorus levels to promote normal bone mineralization. Thus,

8 deficiency of the vitamin manifests in bone mineralization disorders. Rickets, which is characterized by bony deformities in the skeletal structure of growing children, is the result of vitamin D deficiency in children, and osteomalacia, poor mineralization of bone, is characteristic of vitamin D deficiency in adults (23). 1,25(OH)2 D enters the nuclei of target cells and binds to the vitamin D receptor, which then dimerizes with the retinoid X receptor and is phosphorylated before it transcribes genes coded within the vitamin D response element. Many tissues possess the nuclear machinery to transcribe vitamin D- dependent genes, but the most notable for calcium and phosphorus maintenance are the small intestine and bone. In the small intestine 1,25(OH)2 D up-regulates production of calbindin and the epithelial calcium transporter to collectively increase calcium transport across the epithelial membrane. In bone, specifically in osteoblasts, 1,25(OH)2 D up- regulates production of osteocalcin, osteopontin, and alkaline phosphatase (23). These products are markers of osteoblast differentiation, and thus bone formation (24). Vitamin

D has also been shown to promote cell differentiation and prevent proliferation in other tissue types, and evidence has emerged suggesting that 1,25(OH)2 D is produced locally in many tissues for paracrine and autocrine signaling (13). These new findings have led to investigations into the function of vitamin D in other, non-bone-related, diseases.

Investigators have searched for a role of vitamin D in diseases ranging from cancer and diabetes to depression. Cancer will be the focus of this brief review. A recent meta-analysis that considered clinical trials with cancer outcomes, among others, concluded that there is no evidence that vitamin D has a protective effect against cancer

(25). Several investigators were quick to criticize the review citing several downfalls of

9 the analysis including the number of short term clinical trials included, the low doses of vitamin D used, and poor study selection criteria. Additionally, the selected studies did not always consider vitamin D status in addition to vitamin D intake (26–28). When individual cancers are considered separately, some clinical trials do in fact demonstrate an association between higher vitamin D status or intake and lower incidence. There is a fairly strong body of evidence supporting a protective role of vitamin D in colorectal cancer, and there also appears to be a similar relationship with breast cancer, although it is not as strong (29). Aside from these two cancers, the role of vitamin D in other cancers remains rather unclear. The remainder of this review will focus on associations between vitamin D and non-melanoma skin cancer, which is still highly debated due to conflicting evidence.

1.2.2 Vitamin D and Non-Melanoma Skin Cancer

Non melanoma skin cancer (NMSC) is the most prevalent type of cancer in the

United States with approximately 2.2 million people diagnosed each year (30). Although this disease is not particularly deadly, the cost incurred through treatment of the disease is a severe burden for those affected (31). Unlike melanomas, which originate from melanocyte cells in the skin, NMSC develops from keratinocytes. The two primary types of NMSC are basal and squamous cell carcinomas (BCC and SCC respectively). Basal cells are keratinocytes found at the lowest part of the epidermis, while squamous cells are keratinocytes from the outermost part of the epidermis. BCC is more common and less aggressive than SCC because the latter is more likely to metastasize, however metastases 10 are still uncommon (32). The most easily modifiable risk factor for NMSC is unprotected

UV exposure, yet some researchers suggest that individuals receive approximately 5-15 minutes of unprotected sun exposure three times a week for sufficient vitamin D production (33). Other researchers suggest that moderate unprotected sun exposure allows for sufficient production of vitamin D without significantly increasing a person’s risk for skin cancer (33). Still, it is reasonable to expect that individuals following such advice may remain in the sun for longer than necessary, increasing their risk of developing UV-induced skin cancer (34).

The role of vitamin D in NMSC is not well understood. Most of the epidemiological studies investigating the relationship between vitamin D and NMSC focus on BCC. Two prospective studies investigating the relationship of several dietary nutrients and BCC concluded that dietary vitamin D had no protective effect against BCC

(35,36). It should be noted that these studies did not measure serum 25OH D levels, and thus lack a proper biomarker to verify vitamin D intake and status. Asgari and colleagues

(37) reported an increased risk for BCC with higher prediagnostic serum 25OH D levels, adjusted for sun exposure, in a nested case control study. Epidemiological studies investigating the relationship between vitamin D status and general NMSC, including both subtypes BCC and SCC, have found conflicting results. Tang and colleagues (38) found that higher baseline 25OH D serum levels coincided with a decreased risk for self- reported NMSC, while Eide et al. (39) reported an increased risk of histologically confirmed NMSC with higher baseline 25OH D levels. Epidemiological studies on this subject are challenging in that it is difficult to control for sun exposure as a confounding

11 factor, since it contributes to both vitamin D status and NMSC risk, and it is difficult to assess. Additionally, NMSC is not required to be reported to cancer databases, limiting the amount of information available for this cancer as compared to others. These factors could contribute to some of the mixed results seen in these epidemiological studies.

Cell and animal studies provide insight into the role vitamin D may play in

NMSC. Much of the interest in this interaction stems from the fact that keratinocytes, melanocytes, and cancer cells express genes encoding the vitamin D receptor (VDR) and metabolizing enzymes (40,41). Cell studies suggest 1,25(OH)2 D has pro-differentiating and anti-proliferating activities in healthy keratinocytes, while cancerous cells lose sensitivity to the pro-differentiation signaling of 1,25(OH)2 D (41). Studies in mice have yet to adequately confirm results seen in cell studies. Most in vivo work is done with

VDR null mice, which is an indirect method of measuring vitamin D activity. These studies have shown that the VDR is intricately involved in the development of skin cancer, but these studies do not provide sufficient information to conclude if its effects are mediated solely by 1,25(OH)2 D activation of the VDR or other pathways (29). Some preliminary studies on topical application of 1,25(OH)2 D have provided conflicting evidence on whether or not this metabolite is beneficial for NMSC prevention (42,43).

Despite inconclusive evidence from animal studies, vitamin D has potential as a cell cycle regulator in skin, providing a possible basis for a role in NMSC.

Two cellular pathways involved in the development and progression of NMSC are thought to interact with vitamin D. The first, and perhaps most studied, is known as the hedgehog pathway. The signal cascade, shown in figure 3, begins with the binding of 12 hedgehog proteins (HH) to Patched (PTC), which suppresses inhibition of Smoothened

(SMO) by PTC. The release of SMO then allows GLI proteins to bind to nuclear transcription factors that escalate this signal cascade while also activating pro- proliferation genes and suppressing pro-differentiation genes, including the VDR (44,45).

It has been found that 1,25(OH)2 D in conjunction with the VDR suppresses several components of this pathway including HH, PTC, and GLI proteins (46). Interestingly, a limited amount of evidence suggests that vitamin D may regulate the hedgehog pathway through non-genomic means as well by binding directly to SMO (45). Dysregulation of vitamin D-mediated control of the hedgehog cycle could contribute to NMSC formation.

Vitamin D also interacts with the wnt/β-catechin pathway. Similar to the hedgehog pathway, 1,25(OH)2 D and the VDR seem to provide some key regulation that, when lost, leads to aberrant proliferation and less cell differentiation. The relationship between vitamin D and this pathway is less well-defined.

There is some evidence on the ability of 1,25(OH)2 D to protect skin cells against many types of UV-induced DNA damage (47). The modulation of the DNA damage response by 1,25(OH)2D has been recently reviewed by Bikle (45). Briefly, it has been proposed that 1,25(OH)2 D and the VDR play a role in removing UV-induced cyclobutane pyrimidine dimers and other pyrimidine photoproducts. If left unrepaired, these DNA mutations could lead to the development of NMSC. Evidence for a role of vitamin D in skin immunity also exists, and it is suspected that UV radiation may alter skin immunity, possibly contributing to cancer risk. However, it remains unclear whether vitamin D exhibits beneficial action in this case (45). Most investigators in this area focus

13 on the utility of pure 1,25(OH)2 D against NMSC rather than dietary vitamin D. This is not surprising, as epidemiological evidence has suggested that the effects of dietary vitamin D on NMSC are insignificant (41). However, no animal or human clinical trial has yet investigated the effects of dietary vitamin D on the development of NMSC.

HH *

- PTC* SMO

*GLI 1

*GLI 2 +

+ - Pro-proliferation Pro-differentiation genes genes including VDR

* Targets of genomic suppression by 1,25(OH) D and VDR 2

Figure 3: The hedgehog cycle and its components that are regulated by vitamin D. (HH, hedgehog protein; PTC, patched; SMO, smoothened; VDR, vitamin D receptor. Adapted from Varjosalo and Bikle (44,45))

1.3 Vitamin D Analysis

The most commonly measured metabolites of vitamin D are the 25-hydroxylated forms of vitamins D2 and D3. The additive concentration of these two metabolites yields the clinically relevant measurement of vitamin D status. Determination of vitamin D status has been a notable challenge in clinical laboratories, as there are inconsistencies between methods used for measurement and between laboratories performing the assays

(48). A number of programs have been developed to monitor and certify labs measuring serum 25OH D. In the Unites States, the National Institute of Standards and Technology

(NIST) and the National Institute of Health Office of Dietary Supplements (ODS) have collaboratively created the Vitamin D Metabolite Quality Assurance Program. This program does not test proficiency in 25OH D measurement but rather compiles measurements of standard reference materials from participating laboratories. A recent review of results from several standard reference materials measured in different laboratories nationwide indicated that, regardless of the assay used, concentrations of

25OH D tend to be over or under estimated compared to NIST standard reference materials (48). Another program coordinated through the Centers for Disease Control and Prevention, the vitamin D standardization certification program, does test proficiency in 25OH D measurement. This certification is obtained in two phases. The first phase allows laboratories to assess assay biases before analyzing serum with blinded 25OH D values in phase two (49). The ODS also coordinates the Vitamin D Standardization

Program, which represents an international collaboration of vitamin D quality assurance programs with the goal of improving status measurement worldwide.

The common methodology used to measure vitamin D status, as well as issues with the current assays, have recently been reviewed by Su and colleagues (50). The four most common methods used for 25OH D quantification include vitamin D binding protein-based competitive protein-binding assays, immunoassays, and high performance liquid chromatography (HPLC) assays with UV or mass spectral detection. Competitive binding protein assays and immunoassays are popular in clinical and research settings mainly due to their ease of use and affordability. However, these techniques are susceptible to overestimation of 25OH D3 due to interfering metabolites of vitamin D and underestimation of 25OH D2 due to lower binding affinity (50–52). Techniques involving HPLC can separate most of these interfering compounds from 25OH D. HPLC methods tend to perform better than immunoassays, reporting less variability and concentrations closer to predetermined 25OH D levels in standard reference materials

(48). The detection of 25OH D and other vitamin D metabolites by UV absorbance is possible, but this detection is still hindered by lower sensitivity and selectivity for these compounds that absorb at 265 nm, resulting in greater measurement variability (48,50).

Among all these methods, HPLC coupled with tandem mass spectrometry is the gold standard for vitamin D analysis (21).

1.3.1 HPLC-MS/MS Analysis

HPLC with tandem mass spectrometry (HPLC-MS/MS) allows for a specific and sensitive quantification of low-abundant compounds. Considerations that must be made in MS analysis of vitamin D compounds include their relatively poor ionization 17 efficiency, a number of isobaric interferences, and their low abundance in biological samples (53). Several methods for vitamin D analysis by HPLC-MS/MS exist, and have recently been reviewed by El-Khoury et al. (51). Electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI) are both used, operated in positive ion mode. However, a recent analysis of vitamin D methods used by different laboratories revealed increased variability in methods employing ESI versus those using APCI (54).

Thus, it appears that APCI should be the preferred ionization when possible. A variety of mass transitions have been reported for MS/MS of 25OH D, with water losses being the most common (51). Unfortunately, water loss transitions can be problematic, as up to approximately 200 other compounds isobaric with 25OH D could undergo similar transitions, confounding measurement. (53). As a result, it has been suggested that alternate transitions be used for quantification of vitamin D-related compounds (54).

Other pertinent MS settings, such as collision energy and source temperature vary, depending on the instruments used.

The most notable challenge to methods involving HPLC is the presence of 3-epi-

25 hydroxyvitamin D (C3epi). The only structural difference between this compound and 25OH D is the stereochemistry at the third carbon. Due to their structural similarity, chromatographic resolution of these two compounds is challenging. Additionally, this compound is isobaric with 25OH D with no notable unique fragmentation patterns, so the use of tandem mass spectroscopy can do little to alleviate this problem. Initially it was believed that C3epi would only confound measurements of 25OH D in children less than one year of age, but more recent work has shown that it is also a potential problem in

18 assessing the vitamin D status of adults. In children less than one year of age, the concentration of C3epi could average approximately 28% of total 25OH D concentration

(55), and in adults, two separate studies have reported concentrations of 4.6% and 4.8% of total 25OH D (56,57). C3epi does not bind to some of the commonly used immunoassays, but these do suffer from other cross reacting metabolites (58). A number of HPLC-MS/MS methods separating C3epi from 25OH D have been published, however many of the clinical laboratories participating in the Vitamin D Metabolite Quality

Assurance Program still do not employ these methods (48). Little is known about the biological relevance of C3epi, but it is thought that this metabolite is made endogenously via a C-3 epimerizing enzyme that has yet to be fully characterized (59). Like 25OH D3,

C3epi can be activated by CYP27B1 to C3epi-1,25(OH)2 D3, which binds to the vitamin

D receptor with less affinity than native 1,25(OH)2 D3. The biological responses elicited by C3epi-1,25(OH)2 D3 are weaker than those of the native form, with the exception of parathyroid hormone suppression (60). Since C3epi cannot produce equivalent responses to native vitamin D, the presence of C3epi should be considered in the development of any new methods for 25OH D measurement.

1.3.2 Advanced Techniques for Vitamin D Analysis in Biological Samples

Analysis of vitamin D metabolites using MS is challenging due to poor ionization efficiency, a number of interfering compounds, and low concentrations in biological samples. In order to overcome some of these obstacles and increase the sensitivity of

HPLC-MS/MS methods, several groups have begun employing chemical derivatization 19 with Cookson-type reagents. These reagents are 4-substituted 1,2,4-triazoline-3,5-dienes that react with conjugated diene systems in a Diels-Alder fashion (61). The resultant vitamin D derivatives contain functional groups that allow for increased ionization efficiency, and the fragmentation patterns of these derivatives are rather specific and predictable (61,62). Additionally, this mode of derivatization removes many interferences introduced by compounds that are isobaric with vitamin D metabolites. The number of these interfering compounds that are candidates for a Diels-Alder addition is likely very few, thus the new mass transitions used for derivatized vitamin D metabolites are more specific than their underivatized counterparts. A number of Cookson-type reagents have been suggested for use in vitamin D analysis (51), but the most prevalent, and the most commercially available, is 4-phenyl-3H-1,2,4-triazole-3,5(4H)-dione

(PTAD). The derivatization reaction mechanism is shown in figure 4.

Figure 4: Derivatization reaction mechanism of 25OH D3 with PTAD.

Aronov et al. first employed this derivatization agent for vitamin D metabolite profiling in human serum by HPLC-MS/MS (63). An emphasis was placed on sample preparation, including a comparison of solid phase extraction (SPE) and liquid liquid extraction (LLE) and optimization of derivatization conditions. LLE and SPE resulted in similar recoveries of metabolites, but the derivatization efficiency of SPE samples were lower than LLE. Despite this, the authors recommended the use of SPE as it allows for higher throughput than LLE. The optimized derivatization procedure proved to be otherwise effective, and demonstrated the applicability of this approach in a complicated sample matrix. Subsequent modifications to this method by other groups mainly focused on improving quantification of low abundant metabolites, especially 1,25(OH)2 D. One such improvement included the use of micro-flow liquid chromatography (64). This methodology, however, requires specialized equipment not present in many labs.

Another improvement utilizes methylamine as a solvent additive (65). Methylamine forms adducts with PTAD-derivatized vitamin D compounds, and these adducts have been shown to exhibit further enhanced response (65,66).

Although PTAD derivatization improves MS response, the process also has the potential to complicate the chromatography necessary for accurate vitamin D analysis.

The Diels-Alder reaction mechanism results in the formation of two isomeric products for every compound derivatized. For instance, derivatization of 25OH D3 results in the formation of 6-(R)-PTAD 25OH D3 and 6-(S)-PTAD 25OH D3 in a 1:4 ratio (66).

Depending on the analyte and chromatographic method used, these isomeric products

22 may be partially or fully resolved. In order to maximize the signal of the analytes, it would be advantageous to collapse the two isomers into a single peak.

Chemical derivatization changes the chromatography of vitamin D metabolites, and the resolution of C3epi from 25OH D remains a problem. No method separating

PTAD-derivatized C3epi from 25OH D has been published. Higashi and colleagues successfully separated the two compounds, but only after the PTAD derivatized analytes were acetylated (67). This same lab group later released a method utilizing 4-(4’- dimethylaminophenyl)-1,2,4-triazoline-3,5-dione (DAPTAD) as a novel derivatization agent which allows for resolution of 25OH D3 from C3epi (68). However, this reagent must be synthesized and is not commercially available. Thus, a method that separates

PTAD-derivatized C3epi and 25OH D would offer a significant improvement to current published methods.

Chromatographic separation is best described by two key HPLC concepts, separation efficiency and resolution. Separation efficiency can be defined in terms of

Height Equivalence to a Theoretical Plate (HETP). This is a theoretical measure of the number of equilibrations an analyte will make between the stationary and mobile phases in a given column, independent of column length. A small HETP value indicates that this theoretical plate height is small, so there would be more theoretical plates in a given column. Small HETP values indicate good efficiency (69). Resolution, a function of efficiency, describes how well two chromatographic peaks can be distinguished from each other. Peaks with a high degree of co-elution are poorly resolved and would have a 23 low calculated resolution value (69). Both HETP and resolution can be calculated using retention times and peak dimension from a chromatogram and should be considered in the development of new HPLC methods.

The use of internal standards is common in MS methods measuring vitamin D metabolites. Typically these standards contain three or six deuterium atoms, although

13 some C labeled vitamin D metabolites are commercially available. Isotope dilution is a powerful technique in mass spectrometry which uses these isotopically labeled internal standards. Briefly, samples are equilibrated with a known amount of labeled internal standard, and prepared for analysis. Analyte concentrations are quantified by calculating the ratio of labeled to unlabeled compounds. This ratio is compared to a standard curve constructed using a constant concentration of labeled standard and increasing amounts of unlabeled standard. The advantages of isotope dilution include its accuracy and precision, compensation for incomplete analyte recovery, and sensitivity. Disadvantages of this technique are mainly associated with the cost of the labeled standards, training of personnel, and additional necessary validation steps (70). Nonetheless, such methodology is extremely advantageous to analytical procedures with advanced sample preparation, such as chemical derivatization.

1.3.3 Vitamin D Analysis in Biological Tissues

The literature on vitamin D metabolite analysis in biological tissues is sparse.

Until recently, most tissue methods have focused on analysis of vitamin D and 25OH D in meat cuts from large animals, such as pigs and cows. In these cases, sample size was not a limiting factor. Typical sample sizes for these methods range from 10 to 60 g, 24 which is far beyond the sample size that could be obtained from smaller animal models such as rats and mice.

Among tissue methods, old and new, sample preparation tends to include lengthy saponification steps followed by labor intensive preparative HPLC (71,72), solid phase extraction (73,74), or a combination of both (75). The largest sample sizes and most complicated sample preparations are used in the earliest methods that employ HPLC-

PDA for separation and detection (71,72,75). HPLC-MS or HPLC-MS/MS have become the method of choice for quantification of vitamin D and 25OHD in tissues, but large sample sizes and complicated sample preparation are still used for analysis (73,74).

Among all HPLC methods, both normal and reversed phase chromatography are used. In some cases, both normal and reversed phase are used to separate vitamin D and 25OHD, respectively (71–73,75). It should be noted that radioimmunoassay, a very sensitive technique, has also been used to detect vitamin D metabolites in tissues (76,77).

However, these methods require expensive labeled radioisotopes and special handling of radioactive materials, making them undesirable for general use. Additionally, this methodology likely suffers from inherent problems similar to more commonly used immunoassays.

Recently, Lipkie and colleagues (78) published a method for analysis of vitamin

D2&3 and 25OH D2&3 in soft tissues from rats including liver, epididymal fat, and gastrocnemius muscle. This was the first method to utilize samples from an animal model with limited sample size. To improve signal and correct for matrix effects, PTAD derivatization, methylamine, and labeled internal standards were utilized, as discussed 25 above. This method was validated with standard reference material (SRM) 2972 from

NIST, but this SRM does not include a certified value for C3epi. This metabolite was also not mentioned in the paper by Lipkie et al., indicating it either remained undetected or unresolved from other 25OH D peaks.

1.4 Specific Aims

1.4.1 Specific Aim 1: To develop methodology suitable for the accurate evaluation of vitamin D status in murine skin.

As experimental evidence supports a role of vitamin D in the development of NMSC, methodology capable of accurately evaluating vitamin D metabolite concentrations in skin is necessary. Limited methodology has been developed for analysis of vitamin D metabolites in large tissue samples or soft tissues from rats, but none has been developed for the complex matrix and small sample size of murine skin.

I hypothesize that vitamin D metabolites will be present in low concentrations in murine skin, thus requiring the use of chemical derivatization to enhance signal in HPLC-

MS/MS analysis. Additionally, I predict that small amounts of C3epi will be present in skin, presenting a challenge for proper evaluation of vitamin D status.

1.4.2 Specific Aim 2: Apply the method developed in the first aim to samples from a study in which the effects of dietary vitamin D on NMSC were evaluated.

A group in the College of Medicine at The Ohio State University completed a study in which mice were fed escalating levels of vitamin D3 and either exposed or not exposed to

UV light. Serum and skin measurements of vitamin D status were not evaluated, yet these measurements could provide additional insights on the outcomes observed in this study.

I hypothesize that the method developed in specific aim 1 will be appropriate to detect differences in the dermal vitamin D statuses of the different treatment groups.

Furthermore, I expect to see dose-dependent increases in dermal and serum vitamin D status, as well as higher status in the mice exposed to UV light.

Chapter 2: Materials and Methods

2.1 Animals and Experimental Design

The outbred, albino Skh-1 mouse (Charles River Laboratories; Wilmington, MA) is considered the best mouse model for the study of UV-induced NMSC, as the lesions that develop in the dermis of this mouse are similar to SCC in humans (79). All mice used in this work were housed five to a cage in a vivarium at The Ohio State University according to standards established by the American Association for Accreditation of

Laboratory Animal Care. Food and water were provided ad-libitum, and fresh food was provided three times weekly. Prior to beginning all studies, procedures were approved by the appropriate Institutional Animal Care Utilization Committee. Dorsal skin and serum from Skh-1 mice fed normal AIN-93G diets were used to develop methodology for vitamin D metabolite analysis in these matrices. Human serum and pooled umbilical cord serum (Innovative Research; Novi, MI) were also used in the development of these methods.

To study the effects of dietary vitamin D on the development of NMSC, groups of

25 male or female Skh-1 mice were fed a modified AIN-93G diet containing 25, 150, or

1000 IU vitamin D3/kg chow (total n = 150). The dietary groups were further divided into treatment groups in which 15 mice received one minimal erythemal dose of UVB

28 radiation (2240 J/m2) three times per week for 25 weeks. As controls, 10 mice from each group received no UV treatment (figure 5). Mice were sacrificed 24 hours following the final UV treatment by lethal inhalation of CO2. UVB radiation (290-320 nm) was supplied by Phillips FS40UVB lamps (American Ultraviolet Company; Lebanon, IN) fitted with Kodacel filters (Eastman Kodak; Rochester, NY). Radiation was metered using a UVX radiometer (UVP Inc.; Upland, CA).

No UV (n = 10) 25 IU/kg food (n = 25) UV (n = 15)

No UV (n = 10) 75 male or female 150 IU/kg food mice (n = 150) (n = 25) UV (n = 15)

No UV (n = 10) 1000 IU/kg food (n = 25) UV (n = 15)

Week 0 Week 4 Week 29

Figure 5: Design of the study evaluating effects of dietary vitamin D on NMSC development.

2.2 Chemicals

Optima grade water, acetonitrile, methanol, and 99% formic acid (Fisher

Scientific; Pittsburgh, PA) were used for all experiments. 25OH D3 was obtained from

Isosciences (Trevose, PA) and d3-25OH D3, 25OH D3, C3epi and PTAD were obtained from Sigma Aldrich (St. Louis, MO).

2.3 Method Development

In order to resolve C3epi and 25OH D3, several HPLC columns were evaluated including Luna 5 µm C18(2) and Luna 3 µm C18(2) (4.6 x 250 mm; Phenomenex),

Zorbax Extend 1.8 µm C18 HT (2.1 x 50 mm; Agilent), Eclipse 5 µm XDB-C18 (4.6 x

150 mm; Agilent), and Symmetry 3.5 µm C18 (4.6 x 75 mm; Waters; Milford, MA).

Each column was evaluated using a binary mobile phase consisting of 0.1% aqueous formic acid solution (A) and either methanol or acetonitrile with 0.1% formic acid (B) on an Agilent 1200 series HPLC. Eluent from the HPLC was directed to a QTRAP 5500 mass spectrometer (AB Sciex; Framingham, MA) equipped with atmospheric pressure chemical ionization (APCI) operated in positive ion mode. MS conditions, including source temperature and collision energy, were optimized as well.

Existing methods for lipophilic compound extraction from skin were compared for vitamin D metabolite recovery. A protocol by Winegrath et al. (80) employing enzymatic digestion of skin by a combination of collagenase, lipase, and protease was compared to non-polar extraction protocols using ethyl acetate (81) and 5:1

31 hexane/dichloromethane (82) as well as the method described by Lipkie and collegues

(78), which involves a solid phase extraction (SPE) cleanup step. Using the optimized

HPLC-MS/MS method, these methods were evaluated on the quality of the mass chromatograms produced with respect to signal intensity, as well as signal to noise ratio.

Serum extraction methods were also evaluated. The SPE and liquid-liquid extraction (LLE) protocols described by Aronov and colleagues (63) were compared to a protocol which employs Phree SPE cartridges (Phenomenex) (83). Similar to the evaluation of skin extraction methods, these were mainly compared based on mass chromatogram quality. Recovery of a stable isotope internal standard was also assessed as a point of comparison. The methods described below were selected as best fit for the purpose of this work.

2.4 Sample Preparation

Standards of 25OH D3 ranging in concentration from 7.80 to 250 nmol/L were spiked with 20 µL of a d3-25OH D3 solution (200 nmol/L) and set aside. The dorsal skin and serum of 3 mice from each sex/diet/treatment group (total n=36) were analyzed for vitamin D metabolites. Skin samples (0.30 g) were frozen with liquid nitrogen, crushed to a fine powder, and transferred into pre-weighed glass vials. Samples were then spiked with 20 µL of the d3-25OH D3 solution, allowed to equilibrate for 15 min, and extracted with the solvent system used by Stahl and others (82) with slight modifications (84).

Briefly, 1 mL of ethanol containing 0.1% butylated hydroxytoluene, 1 mL of water, and 5

32 mL of 5:1 hexane/dichloromethane were added to the skin samples, and the mixture was probe sonicated for 30 s. The homogenized solutions were centrifuged at 2000 x g for 5 min to hasten phase separation, and the upper organic layer was decanted into clean glass vials. The extraction was repeated two times, with the addition of 5 mL 5:1 hexane/dicholormethane, and the pooled organic layers were dried under a stream of nitrogen gas.

Serum samples were extracted using Phree SPE cartridges (1 mL tubes) according to the Phenomenex technical protocol (83) with minor changes. 200 µL of serum was deposited into an SPE cartridge, spiked with 20 µL of the d3-25OH D3 solution, and allowed to equilibrate for 15 min. 500 µL of an 85:15 acetonitrile/methanol was then added to the cartridge, vortexed for 30 sec, allowed to rest for 1 min, and eluted under vacuum (-15” Hg) for 2 min. This extraction was repeated and the pooled eluents were dried under a stream of nitrogen gas.

Dried extracts and standards were derivatized using 4-Phenyl-1,2,4-triazole-3,5- dione (PTAD) according to Lipkie et al. (78). 100 µL of a 2 mg/mL PTAD solution was added to the extracts, which were then mixed on a vortex for 10 min. An additional 100

µL of PTAD was added to the solution and mixing was repeated for another 10 min. The reaction was quenched by adding 20 µL of water and mixing for 5 min, and the reaction mixture was dried under nitrogen gas. Skin extracts were reconstituted with 100 µL of acetonitrile, and centrifuged at 15,000 x g for 2 min prior to analysis. Serum extracts were reconstituted with 200 µL acetonitrile and passed through 0.22 µm, 4 mm nylon filters before injection onto the HPLC-MS/MS system. 33

2.5 HPLC-MS/MS Analysis

Vitamin D metabolites were separated by HPLC (Agilent 1200 Series; Santa

Clara, CA) using a Luna C18 reversed phase column (4.6x250 mm, 5 µm; Phenomenex,

Torrance, CA) maintained at 40 °C. The binary mobile phase consisted of water (A) and acetonitrile (B), both containing 0.1% formic acid. Optimal separation of vitamin D metabolites was achieved using a step gradient of 29% A for 3.5 min and an immediate switch to 20% A for 3 min, followed by a column wash with 100% B for two minutes and reconditioning at initial conditions for 3.5 min. The injection volume was 10 µL and the injection needle was submerged in acetonitrile between runs. Eluent from the HPLC was directed to the QTRAP system described above. MS parameters were as follows: source temperature, 450 °C; curtain gas, 30 psi; ion source gas 1, 60 psi; declustering potential, 185 V; entrance potential, 10 V; collision cell exit potential, 11 V. Mass transitions and collision energies are detailed in table 1.

25OH D3 and C3epi chromatographic peaks were integrated in Analyst 1.5.1

(ABSciex; Framingham, MA). Isotope dilution methodology was used to quantify metabolites. Briefly, derivatized standards with increasing concentrations of 25OH D3 and constant levels of d3-25OH D3 were used to construct a standard curve based on peak area ratios of labeled and unlabeled standards. The same ratios were calculated in the spiked biological samples. Peak areas of both isomeric forms of PTAD-25OH D3 and the labeled internal standard were summed to calculate the ratios used in the standard curve.

2.6 Statistical Analysis

Descriptive analysis was performed using Microsoft Excel (Redmond, WA).

Statistically significant differences (p<0.05) of 25OH D3 and C3epi levels in skin and serum were calculated using ANOVA. The analysis was modeled to detect differences by diet, UV exposure, or sex alone and interactions between these factors including diet x

UV exposure, diet x sex, UV exposure x sex, and diet x UV exposure x sex. Post-hoc analysis was performed using Tukey’s test in SPSS (IBM, Armonk, NY).

Table 1: Relevant MS/MS parameters used for the analysis of vitamin D metabolites.

Retention Collison Dwell Analyte Mass Transition Time (min) Energy (eV) Time (ms)

PTAD-25OH D3 5.9 558.3 > 298.3 25 280 PTAD-C3epi 6.2 558.3 > 298.3 25 280

PTAD-d3-25OH D3 6.3 561.3 > 301.3 25 280

Chapter 3: Results and Discussion

3.1 Method Development

3.1.1 HPLC Column and Method Selection

As shown in figure 4 (Section 1.3.2), PTAD derivatization of vitamin D metabolites results in the formation of two isomeric products, (R) and (S), for each compound. Thus, the chromatography of PTAD-derivatized vitamin D metabolites is more complicated than that of underivatized metabolites, but the signal enhancement imparted by PTAD well outweighs this cost. The HPLC method described in this work was optimized for analysis of PTAD-1,25(OH)2 D3, PTAD-25OH D3, and PTAD-C3epi.

One factor that was considered in the optimization process was collapsing (R) and (S) isomers of analytes into a single peak when possible to enhance signal and simplify data processing. Additionally, resolution of PTAD-C3epi from PTAD-25OH D3 was a major focus of this method development.

A variety of C18 columns are commonly used to chromatograph derivatized vitamin D metabolites (51), so C18 columns of varying dimensions and particle sizes were evaluated for the separation of PTAD-C3epi and PTAD-25OH D3 from a mixed standard. A binary mobile phase of water and acetonitrile, both with 0.1% formic acid, was found to outperform that of water and methanol. Figure 6 displays separations of

PTAD-C3epi and PTAD-25OH D3 using optimized compositions of this mobile phase on the selected columns. The UHPLC Zorbax Extend column (Figure 6, A) only partially resolved the isomers of PTAD-25OH D3, which likely co-eluted with PTAD-C3epi. As columns of increasing length were evaluated, both (R) and (S) PTAD-C3epi eluted as a single peak between the two PTAD-25OH D3 isomers. Column length is a key element of chromatographic separation, as longer columns allow for more interactions between solutes and the solid phase, resulting in enhanced separation (85).

Figure 6: Improvement of chromatographic resolution of a mixed standard containing

PTAD-derivatized 25OH D3 and C3epi with C18 columns of varying length and particle size (a: Zorbax Extend; b: Symmetry; c: Eclipse; d: Luna 5 µm; e: Luna 3 µm. Peak identities in chromatogram e include [1]. 6(R)-PTAD-25OHD3, [2]. (R)- and (S)-PTAD-

C3epi, and [3] 6(S)-PTAD-25OHD3.)

The three best performing columns were the Eclipse (150 mm, c) and both 250 mm Luna columns, one with 5 µm packing (d) and the other with 3 µm packing (e).

Among these three, the 3 µm Luna was superior in terms of resolution and efficiency, which were calculated by equations [1] and [2], respectively (69,86). Efficiency in this case was defined as height equivalent to a theoretical plate (HETP). This is a theoretical measure of the number of equilibrations an analyte will encounter in a given column, independent of column length. A low HETP is indicative of a large number of equilibrations and thus high separation efficiency, resulting in narrower peaks (69).

Figure 7 summarizes these calculations for the three best performing columns. The 5 µm

Luna and Eclipse columns differed minimally with respect to resolution, but the 5 µm

Luna did exhibit greater separation efficiency. However, compared to both of these, the

3 µm Luna showed drastically improved resolution and efficiency. Columns with smaller particle sizes allow for expedited equilibration of an analyte between the solid and mobile phases, resulting in greater efficiency (87). On the 3 µm Luna column, PTAD-C3epi and

(S)-PTAD-25OH D3 are almost baseline resolved at a value of 1.40. A resolution greater than 1.50 is considered baseline resolution (86).

[1]

[2]

Where: - tr is retention time - w is peak width at half maximum - l is column length

Figure 7: Resolution of 6(R)-PTAD-25OHD3 (1) and 6(S)-PTAD-25OHD3 (3) from co- eluting stereoisomers of PTAD-C3epi (2).

It should be noted that the 5 µm and 3 µm Luna columns were selected for the analysis of murine skin and serum, respectively, as described in a following section. The separation of PTAD-C3epi from PTAD-25OH D3 with the 5 µm Luna column was considered sufficient for analysis from skin.

3.1.2 Skin and Serum Extraction Methods Selection

Of all the extraction methods evaluated in murine skin, the one best fit for purpose of this analysis was with a 5:1 hexane/dichloromethane, according to Kopec

(84). The recovery and any matrix effects are both adjusted for using isotope dilution methodology. Among the serum extraction methods evaluated, the Phree SPE cartridge gave superior recovery compared to the SPE method described by Aronov and others

(63). It also allowed higher throughput than the liquid-liquid extraction method used by the same group. Thus, this method was selected as the serum extraction used in subsequent investigations.

Table 2: Levels of 25OHD3 and C3epi from a variety of matrices

25OHD3 C3epi C3epi Sample (nmol/L) (nmol/L) Proportion (%) Cord Serum 42.4 3.6 7.8 Adult Serum 1 90.2 5.5 5.8 Adult Serum 2 47.4 3.4 6.7 Adult Serum 3 57.2 4.6 7.4

(pmol/g) (pmol/g) Skin 1 3.3 1.4 29.8 Skin 2 5.6 1.3 18.8 Skin 3 8.9 4.7 34.6

3.1.3 Application of Selected Methods

The applicability of this methodology was assessed in different sample matrices including adult serum, pooled umbilical cord serum, and murine skin (table 2). These samples were selected not only because of their different physical and chemical matrices but also because they were expected to contain differing proportions of C3epi.

Detectability of PTAD derivatized 25OH D3 and C3epi was excellent in all samples. In three different skin samples assayed, C3epi represented 18.6% to 34.8% of total 25OH

D3, while in the three adult serums it only represented 5.8% to 7.4%. The level of C3epi in the pooled cord blood was 7.8%, similar to the adult serum.

Often, it is of interest to analyze 25OH D in biological microsamples. The use of dried blood spots in lieu of serum collection has been increasing in popularity for the measurement of various biomarkers (88). Some methods for vitamin D status measurement from dried blood spots have been developed (51,67,68), many using PTAD derivatization as a means of enhancing MS/MS response due to the small sample size associated with this methodology. It has been shown that measurement of vitamin D status from this matrix by HPLC-MS/MS is comparable to serum status measurement by the same means (89). The use of this sampling method has become especially popular in pediatrics, yet this population is the most likely to exhibit elevated levels of C3epi

(55,88). Unfortunately, current published methods utilizing PTAD for the analysis of vitamin D status in dried blood spots do not account for C3epi. Methodology that allows for proper separation of PTAD-C3epi from PTAD-25OH D3 will be useful as this mode of blood sampling becomes increasingly utilized.

Since it is likely that the tissue microenvironment has a role in cancer etiology

(90), determination of vitamin D status in tissues may provide pertinent information on the function of this nutrient in cancer development and progression. The concentration of

25OH D in tissues is likely rather low, as demonstrated in this work and by Lipkie and collegues (78), necessitating the use of PTAD or similar chemical derivatization. As seen in the skin samples discussed above, C3epi is present in relatively appreciable quantities.

Left unresolved, PTAD-C3epi would contribute significantly to PTAD-25OH D3 measurement, overestimating 25OH D3 concentrations. The presence of C3epi in other tissues has not been investigated, but should be considered in future work.

12 30 Skin Serum C

10 25 / g) g) / 8 B 20

b (nmol/L)

pmol

(

3 6 15 A a

25OH D 25OH 4 10

Skin Skin Serum 25OH D 25OHSerum 2 5

0 0 25 150 1000

Vitamin D3 dose (IU/kg diet)

Figure 8: Dose-dependent levels of 25OHD3 in skin (left axis) and serum (right axis) of mice fed increasing doses of vitamin D3. Letters represent significant differences

(P<0.05), with lower case letters pertaining to skin levels and upper case levels referring to serum levels. Error bars indicate standard error.

3.2 Quantification of 25OH D3 and C3epi in a Mouse Model of Non-Melanoma Skin

Cancer

3.2.1 Serum and Skin Levels of 25OH D3

Levels of 25OH D3 increased in a dose-dependent manner in both serum (P<0.01) and skin (P<0.01), as shown in figure 8. Serum levels were considerably higher than skin levels, as average serum 25OH D3 ranged from approximately 10 nmol/L to 25 nmol/L depending on dose, while average skin levels ranged from approximately 4 pmol/g to 9 pmol/g. Compared to other studies in which mice were fed 1000 IU vitamin

D3, the serum levels of the mice in this work fed the same dose are within the same order of magnitude. These studies reported serum 25OH D3 of 75 and 115 nmol/L (91,92). It is likely that differences in mouse breeds and study conditions account for the variability in 25OH D3 levels across studies. Few points of comparison exist for the skin 25OH D3 levels described here, but they seem reasonable when compared to values found in other species and tissues. Höller and colleagues reported 25OH D3 in the skin of pigs fed 2000

IU vitamin D3/kg diet at approximately 100 pmol/g (74). The liver, muscle, and fat tissue of rats fed 740 IU vitamin D3/kg diet analyzed by Lipkie and colleagues all contained approximately 1.25 pmol/g tissue (78). Dose-dependent increases in tissue 25OH D3 levels were also demonstrated in Lipkie’s study, but this increase was only shown over two doses. This study is the first to show differences in 25OH D3 content of skin across multiple dietary fortification levels of vitamin D3.

With respect to serum 25OH D3 levels, UV treatment was the only other factor to have a significant effect, as no gender effects or multifactorial interactions between dose, 46 gender, and/or UV treatment were detected. Contrary to what was expected, the mice exposed to UVB light tended to have lower serum levels of 25OH D3 than those left unexposed (P<0.01). Although this result is quite counterintuitive, as UVB exposure is known to elevate vitamin D status, no studies have evaluated vitamin D status response to dietary supplementation and controlled chronic UV exposure. It is possible that the chronic exposure administered to these mice influenced the homeostasis in serum 25OH

D3. Mice chronically exposed to UV light exhibit increased hepatic cytochrome P450 content (93). CYP24A1, the enzyme responsible for 25OH D3 degradation is a cytochrome P450 protein. Although its expression is normally relegated to the kidneys and target tissues of vitamin D (94), expression of CYP24A1 has been shown to be inducible through pharmacological intervention in human hepatocytes (95). It is plausible that chronic UV exposure could also induce expression of CYP24A1 or other vitamin D metabolizing enzymes in the liver and/or other tissues. Furthermore, it has been shown in sheep that exposure to UVB light increases the rate at which 25OH D3 is cleared from serum (96). Although hepatic cytochrome P450 content and 25OH D3 clearance were not measured here, the results from this study seem to be in agreement with this rationale. Binkley et al.. found that high UV exposure may not guarantee elevated vitamin D status in humans, citing genetic elements or enhanced cutaneous destruction of vitamin D as possible explanations (97). Similar factors may also be at play here. This finding raises the question of whether or not individuals taking vitamin D supplements who also obtain a large amount of UV exposure receive the full benefits of these supplements in terms of vitamin D status elevation.

No significant influences of UV exposure or gender on 25OH D3 levels in skin were detected. Additionally, no two-way interactions between these factors were detected.

3.2.2 Serum and Skin levels of C3epi

Serum levels of C3epi increased in a dose-dependent manner (P<0.01) and varied according to UV exposure (P=0.011). Additionally, a significant dose by UV exposure interaction was observed (P=0.028), as illustrated in figure 9A. At high doses of vitamin

D3, UV exposure appeared to attenuate C3epi serum response. The reasoning behind this could be similar to that discussed above, for the basic effect of UV exposure on serum

25OHD3, as CYP24A1 also metabolizes C3epi (59). Some epidemiological investigations into the levels of C3epi in human populations have been performed, with wide-ranging levels reported (60). Interestingly, some of these reports have ventured to identify determinants of serum C3epi levels. Most positive determinants of C3epi levels include those associated with increased serum 25OH D3 levels, such as supplementation with vitamin D3 and increased sun exposure (56,98).

In skin, a dose-dependent increase of C3epi was observed (P<0.001), but a significant interaction between dose and sex was also seen with respect to these levels

(P=0.045). Figure 9B shows skin C3epi levels for female and male mice. At the highest dietary level, the skin of male mice contained significantly higher levels of C3epi than female mice. Epidemiological work on serum C3epi levels found no evidence of gender effects (56,98), but whether or not this null effect extends to skin has not been previously 48 investigated. Additionally, the presence of C3epi in intact skin has never been evaluated in vivo. Gender effects in several aspects of this mouse model of NMSC have been documented, including cancer etiology and even skin structure (79,99). The presence of

C3epi in intact human skin has not been evaluated, but in this study this metabolite was rather significant in skin. If C3epi was included in the 25OH D3 skin measurement, it would have represented between 15% and 39% of the total reading.

12 A

No UV UV c

8 b 6

4 ab a

2 a a C3epi Serum Levels (nmol/L) Levels C3epiSerum

0 25 150 1000

Vitamin D3 Dose (IU/kg diet)

B 5 Female Male c

b 3

a 2 a a a

Skin C3epi (pmol/g) C3epiSkin 1

0 25 150 1000

Vitamin D3 Dose (IU/kg diet) Figure 9: Serum (A) and skin (B) levels of C3epi in mice fed increasing amounts of vitamin D3. Letters indicate significant differences (P<0.05), and error bars indicate standard error. 50

3.2.3 Relationships between Metabolite Levels in Skin and Serum

Highly significant correlations existed between skin and serum levels of both

25OH D3 and C3epi (P<0.01). The Pearson correlation between skin and serum 25OH

D3 was 0.715 indicating that serum vitamin D status reflected status in the skin of these mice (figure 10A). The relationship between skin and serum C3epi, however, was much stronger with a Pearson correlation of 0.916 (figure 10B). This strong relationship is quite intriguing, as C3-epimerization of vitamin D metabolites is thought to occur via a largely uncharacterized metabolic pathway specifically expressed in several cell types, including human keratinocytes (100). This data suggests that C3epi originating from skin contributed a significant proportion of circulating C3epi in these mice. Other cell types including rat and human osteosarcoma, human colon adenocarcinoma, porcine kidney cells, and human hepatoblastoma have been shown to express enzymes involved in C3- epimerization pathway to varying extents (59,101). The fact that the correlation of skin and serum C3epi was much stronger than that of 25OH D3, despite the structural similarities of these metabolites, supports the view that some additional biological mechanism specific to C3epi may have been involved. Correlations between serum

C3epi levels in serum and other organs were not evaluated here, but this information could provide more insight on the role of skin in C3epi production in future studies.

A 35

30 (nmol/L)

25 3 20 15 10

Serum 25OHDSerum 5 0 0 2 4 6 8 10 12 14

Skin 25OHD3 (pmol/g)

14 B

/L) 12

nmol 10 8

C3epi ( C3epi 6 4

Serum Serum 2 0 0 1 2 3 4 5 6 Skin C3epi (pmol/g)

Figure 10: Correlation between skin and serum 25OH D3 (A; Pearson correlation=0.715) and C3epi (B; Pearson correlation=0.918).

3.2.4 Levels of 25OH D3 and C3epi in Relation to NMSC Outcomes

The cancer outcomes of this study have been described previously (102). No significant differences in tumor size or burden were detected between the vitamin D doses. Male mice tended to develop fewer (P<0.01) but larger (P=0.021) tumors than female mice. When tumors were pathologically graded, it was observed that male mice developed numerically more cancerous lesions in a seemingly dose-dependent manner, while the female mice fed 25 IU and 1000 IU had similar levels of cancerous lesions and those fed 150 IU had elevated numbers (figure 11). This indicates that the development of NMSC in males and females may have been differentially affected by dietary vitamin

D3. As mentioned above, gender differences in these mice are common (99). At the very least, no protective effect of elevated vitamin D status was detected.

It is intriguing that male mice with higher skin levels of C3epi developed more cancerous lesions. Although this study was not designed to exclusively test the effects of

C3epi on NMSC, the data lends plausibility to a role of C3epi in this disease. 25OH D3 is metabolized by CYP27B1to form the active metabolite 1,25(OH)2 D3. Similarly, C3epi is processed by CYP27B1 resulting in the formation of a C-3 epimer of 1,25(OH)2 D3

(59). This activated C3epi-1,25(OH)2 D3 binds to the vitamin D receptor, but it elicits less potent responses than 1,25(OH)2 D3 (60). If higher levels of 3-epi-1,25(OH)2 D3 were also present in the skin of the male mice with elevated C3epi, then perhaps 3-epi-

1,25(OH)2 D3 was competing with 1,25(OH)2 D3 for binding with the vitamin D receptor, attenuating the anti-proliferative and differentiating effects of non-epimerized vitamin D.

Although the current study does not test this hypothesis, it does provide some evidence that C3epi may play a role in NMSC.

Figure 11: Dietary vitamin D supplementation effects on chronic UVB induced tumor grade. (102).

Chapter 4: Conclusions and Future Directions

In this work, existing methodology for the analysis of PTAD-derivatized 25OH

D3 was improved by successful chromatographic separation of PTAD-C3epi from PTAD-

25OH D3. This separation will allow researchers to more accurately assess vitamin D status in biological microsamples. Using the methodology described in this work, 25OH

D3 and C3epi levels were characterized in serum and, for the first time, in murine skin to provide insight in what role vitamin D may play in a mouse model of NMSC. Serum and skin levels of 25OH D3 were found to increase significantly with increasing dietary vitamin D3. Additionally, levels of C3epi serum and skin were elevated at the highest administered dose of vitamin D3 and varied according to UV treatment and gender, respectively. From this data it is clear that dietary administration of vitamin D3 can affect levels in the skin, despite the fact that a protective effect of vitamin D against NMSC was not detected in this study.

The role of C3epi in NMSC is yet to be evaluated. Our findings demonstrate that such an investigation may provide valuable insight to the role of dietary vitamin D in this disease. Levels of C3epi in skin and serum were found to be strongly correlative, which suggests that skin may be a contributor of circulating C3epi. Additionally, male mice with higher dermal levels of C3epi developed greater numbers of cancerous lesions.

Future work will focus on the determination of dermal contribution to global C3epi levels as well as the potential role of C3epi in NMSC incidence and progression.

References

1. McCollum E, Simmonds N, Becker E, Shipley P. An Experimental Demonstration of the Existence of a Vitamin which Promotes Calcium Deposition. J Biol Chem. 1922;53:293–312.

2. Steenbock H, Black A, Nelson MT, Hoppert CA, Riising BM. The Induction of Growth-promoting and Calcifying Properties in Fats and Their Unsaponifiable Constituents by Exposure to Light. J Biol Chem. 1925;64:263–98.

3. Askew F, Bourdillon R, Bruce H, Jenkins R, Webster T. The Distillation of Vitamin D. Proc R Soc London. 1930;107:76–90.

4. Windaus A, Bock F. Uber das provitamin aus dem sterin der schweineschwarte. Z Physiol Chem. 1937;245:168–70.

5. Holick MF, Frommer JE, McNeill SC, Richtand NM, Henley JW, Potts JT. Photometabolism of 7-Dehydrocholesterol to Pre-vitamin D3 in Skin. Biochem Biophys Res Commun. 1977;76:107–14.

6. Deluca HF. History of the discovery of vitamin D and its active metabolites. Bonekey Rep. Nature Publishing Group; 2014;3:479. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24466410

7. Institute of Medicine of the National Academies. Dietary Reference Intakes for Calcium and Vitamin D. Ross AC, Taylor CL, Yaktine AL, Del HB, editors. Washington. D.C.: The National Academies Press; 2011.

8. Institute of Medicine of the National Academies. Dietary Reference Intakes for Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride. Washington. D.C.: National Academy Press; 1997.

9. Kupferschmidt K. Uncertain Verdict as Vitamin D Goes on Trial. Science. 2012;337:1476–8.

10. Holick MF. Vitamin D deficiency. N Engl J Med. 2007;357:266–81.

11. Deluca HF. Vitamin D: the Vitamin and the Hormone. Fed Proc. 1974;33:2211–9.

12. Mulligan GB, Licata A. Taking vitamin D with the largest meal improves absorption and results in higher serum levels of 25-hydroxyvitamin D. J bone Miner Res. 2010;25:928–30.

13. Holick MF. Vitamin D and Health. In: Holick MF, editor. Vitamin D: Physiology, Molecular Biology, and Clinical Applications. Second. New York, NY: Springer Science+Business Media; 2010. p. 3–34.

14. Matsuoka L, Wortsman J, Dannenberg M, Hollis B, Lu Z, Holick M. Clothing prevents ultraviolet-B radiation-dependent photosynthesis of vitamin D. J Clin Endocrinol Metab. 1992;75:1099–103.

15. Misra M, Pacaud D, Petryk A, Collett-Solberg PF, Kappy M. Vitamin D deficiency in children and its management: review of current knowledge and recommendations. Pediatrics. 2008;122:398–417.

16. Holick MF, Matsuoka LY, Wortsman J. Age, Vitamin D, and Solar Ultraviolet. Lancet. 1989;2:1104–5.

17. Quinn A, Diffey B, Craig P, Farr P. Definition of the Minimal Erythemal Dose Used for Diagnostic Phototesting. Br J Dermatol. 1994;131:56.

18. Holick M. McCollum Award Lecture, 1994: Vitamin D- new horizons for the 21st century. Am J Clin Nutr. 1994;60:619–30.

19. DeLuca HF, Schnoes HK. Vitamin D: Recent Advances. Ann Rev Biochem. 1983;52:411–39.

20. Jones G. Pharmacokinetics of vitamin D toxicity. Am J Clin Nutr. 2008;88:582S– 586S.

21. Zerwekh JE. Blood biomarkers of vitamin D status. Am J Clin Nutr. 2008;87:1087S–1091S.

22. Vieth R, Bischoff-Ferrari H, Boucher BJ, Dawson-Hughes B, Garland CF, Heaney RP, Holick MF, Hollis BW, Lamberg-Allardt C, et al. The urgent need to recommend an intake of vitamin D that is effective. Am J Clin Nutr. 2007;85:649– 50.

23. Holick MF. Vitamin D. In: Stipanuk MH, editor. Biochemical, Physiological, and Molecular Aspects of Human Nutrition. Philidelphia, PA: Elsevier Health Sciences; 2006. p. 863–83.

24. Weinreb M, Shinar D, Rodan G a. Different pattern of alkaline phosphatase, osteopontin, and osteocalcin expression in developing rat bone visualized by in situ hybridization. J Bone Miner Res. 1990;5:831–42.

25. Bolland MJ, Grey A, Gamble GD, Reid IR. The effect of vitamin D supplementation on skeletal, vascular, or cancer outcomes: a trial sequential meta- analysis. Lancet Diabetes Endocrinol. 2014;8587:1–14.

26. Anagnostis P, Karras SN, Athyros VG, Annweiler C, Karagiannis A. Responses to: The effect of vitamin D supplementation on skeletal, vascular, or cancer outcomes. Lancet Diabetes Endocrinol. 2014;2:362–3.

27. Bischoff-Ferrari H, Orav EJ, Willett WC, Dawson-Hughes B. Responses to: The effect of vitamin D supplementation on skeletal, vascular, or cancer outcomes. Lancet Diabetes Endocrinol. 2014;2:363–4.

28. Grant WB. Vitamin D Status: Ready for Guiding Prostate Cancer Diagnosis and Treatment? Clin Cancer Res. 2014;20:2241–3.

29. Bikle DD. Vitamin D and cancer: the promise not yet fulfilled. Endocrine. 2014; Available from: http://www.ncbi.nlm.nih.gov/pubmed/24402695

30. The American Cancer Society. Available from: www.cancer.org

31. Chen J, Fleisher A, Smith E, Kancler C, Goldman N, Williford P, Feldman S. Cost of Nonmelanoma Skin Cancer Treatment in the United States. Dermatologic Surg. 2001;27:1035–8.

32. The American Cancer Society. Skin Cancer : Basal and Squamous Cell. 2013 p. 47.

33. Holick MF. Sunlight, UV-Radiation, Vitamin D, and Skin Cancer: How Much Sunlight Do We Need? In: Reichrath J, editor. Sunlight, Vitamin D and Skin Cancer. Austin, TX: Springer Science+Business Media; 2008. p. 1–15.

34. Oberyszyn TM. Non-melanoma skin cancer: importance of gender, immunosuppressive status and vitamin D. Cancer Lett. 2008;261:127–36.

35. Hunter DJ, Colditz GA, Stampfer MJ, Rosner B, Willett WC, Speizer FE. Diet and risk of basal cell carcinoma of the skin in a prospective cohort of women. Ann Epidemiol. 1992;2:231–9.

36. Van Dam RM, Huang Z, Giovannucci E, Rimm EB, Hunter DJ, Colditz GA, Stampfer MJ, Willett WC. Diet and basal cell carcinoma of the skin in a prospective cohort of men. Am J Clin Nutr. 2000;71:135–41.

37. Asgari MM, Tang J, Warton ME, Chren M-M, Quesenberry CP, Bikle D, Horst RL, Orentreich N, Vogelman JH, Friedman GD. Association of prediagnostic serum vitamin D levels with the development of basal cell carcinoma. J Invest Dermatol. 2010;130:1438–43.

38. Tang JY, Parimi N, Wu A, Boscardin WJ, Shikany JM, Chren MM, Cummings SR, Epstein EH, Bauer DC. Inverse association between serum 25(OH) vitamin D levels and non-melanoma skin cancer in elderly men. Cancer Causes Control. 2010;21:387–91.

39. Eide MJ, Johnson D a, Jacobsen GR, Krajenta RJ, Rao DS, Lim HW, Johnson CC. Vitamin D and nonmelanoma skin cancer in a health maintenance organization cohort. Arch Dermatol. 2011;147:1379–84.

40. Mason RS, Reichrath J. Sunlight vitamin D and skin cancer. Anticancer Agents Med Chem. 2013;13:83–97.

41. Bikle DD, Oda Y, Xie Z. Vitamin D and skin cancer: a problem in gene regulation. J Steroid Biochem Mol Biol. 2005;97:83–91.

42. Lutzow-Holm C, De Angelis P, Grosvik H, Clausen O. 1 , 25-dihydroxyvitamin D3 and the vitamin D analogue KH1060 induce hyperproliferation in normal mouse epidermis. Exp Dermatol. 1993;2:113–21.

43. Dixon KM, Norman AW, Sequeira VB, Mohan R, Rybchyn MS, Reeve VE, Halliday GM, Mason RS. 1α,25(OH)₂-vitamin D and a nongenomic vitamin D analogue inhibit ultraviolet radiation-induced skin carcinogenesis. Cancer Prev Res. 2011;4:1485–94.

44. Varjosalo M, Taipale J. Hedgehog: functions and mechanisms. Genes Dev. 2008;22:2454–72.

45. Bikle DD. Protective actions of vitamin D in UVB induced skin cancer. Photochem Photobiol Sci. 2012;11:1808–16.

46. Teichert AE, Elalieh H, Elias P, Welsh J, Bikle DD. Over-expression of hedgehog signaling is associated with epidermal tumor formation in vitamin D receptor null mice. J Invest Dermatol. 2011;131:2289–97.

47. Gordon-Thomson C, Gupta R, Tongkao-on W, Ryan A, Halliday GM, Mason RS. 1α,25 dihydroxyvitamin D3 enhances cellular defences against UV-induced oxidative and other forms of DNA damage in skin. Photochem Photobiol Sci. 2012;11:1837–47.

48. Bedner M, Lippa K a, Tai SS-C. An assessment of 25-hydroxyvitamin D measurements in comparability studies conducted by the Vitamin D Metabolites Quality Assurance Program. Clin Chim Acta. 2013;426:6–11.

49. Centers for Disease Control and Prevention. Vitamin D Standardization Certification Program Participant Protocol. Atlanta, GA; 2014 p. 1–5.

50. Su Z, Narla SN, Zhu Y. 25-Hydroxyvitamin D : Analysis and clinical application. Clin Chim Acta. 2014;433:200–5.

51. El-Khoury JM, Reineks EZ, Wang S. Progress of liquid chromatography-mass spectrometry in measurement of vitamin D metabolites and analogues. Clin Biochem. The Canadian Society of Clinical Chemists; 2011;44:66–76.

52. Wallace AM, Gibson S, de la Hunty A, Lamberg-Allardt C, Ashwell M. Measurement of 25-hydroxyvitamin D in the clinical laboratory: current procedures, performance characteristics and limitations. Steroids. 2010;75:477–88.

53. Volmer DA, Mendes LRBC, Stokes CS. Analysis of Vitamin D Metabolic Markers by Mass Spectrometry: Current Techniques, Limitations of the “Gold Standard” Method, and Anticipated Future Directions. Mass Spectrom Rev. 2013;1–22.

54. Couchman L, Benton CM, Moniz CF. Variability in the analysis of 25- hydroxyvitamin D by liquid chromatography-tandem mass spectrometry: the devil is in the detail. Clin Chim Acta. 2012;413:1239–43.

55. Singh RJ, Taylor RL, Reddy GS, Grebe SKG. C-3 epimers can account for a significant proportion of total circulating 25-hydroxyvitamin D in infants, complicating accurate measurement and interpretation of vitamin D status. J Clin Endocrinol Metab. 2006;91:3055–61.

56. Cashman KD, Kinsella M, Walton J, Flynn A, Hayes A, Lucey AJ, Seamans KM, Kiely M. The 3 Epimer of 25-Hydroxycholecalciferol Is Present in the Circulation of the Majority of Adults in a Nationally Representative Sample and Has Endogenous Origins. J Nutr. 2014;25:1–8.

57. Lensmeyer G, Poquette M, Wiebe D, Binkley N. The C-3 epimer of 25- hydroxyvitamin D(3) is present in adult serum. J Clin Endocrinol Metab. 2012;97:163–8.

58. Carter GD. 25-hydroxyvitamin D: a difficult analyte. Clin Chem. 2012;58:486–8.

59. Kamao M, Tatematsu S, Hatakeyama S, Sakaki T, Sawada N, Inouye K, Ozono K, Kubodera N, Reddy GS, Okano T. C-3 epimerization of vitamin D3 metabolites and further metabolism of C-3 epimers: 25-hydroxyvitamin D3 is metabolized to 3-epi-25-hydroxyvitamin D3 and subsequently metabolized through C-1alpha or C-24 hydroxylation. J Biol Chem. 2004;279:15897–907.

60. Bailey D, Veljkovic K, Yazdanpanah M, Adeli K. Analytical measurement and clinical relevance of vitamin D3 C3-epimer. Clin Biochem. 2013;46:190–6.

61. Shimada K, Oe T, Mizuguchi T. Cookson-type Reagents: Highly Sensitive Derivatization Reagents for Conjugated Dienes in High-performance Liquid Chromatography. Analyst. 1991;116:1393–7.

62. Higashi TH, Miura KM, Kitahori JK, Shimada KS. Usefulness of Derivatization in High-Performance Liquid Chromatography / Tandem Mass Spectrometry of Conjugated Vitamin D Metabolites. Anal Sci. 1999;15:619–23.

63. Aronov P a, Hall LM, Dettmer K, Stephensen CB, Hammock BD. Metabolic profiling of major vitamin D metabolites using Diels-Alder derivatization and ultra-performance liquid chromatography-tandem mass spectrometry. Anal Bioanal Chem. 2008;391:1917–30.

64. Duan X, Weinstock-Guttman B, Wang H, Bang E, Li J, Ramanathan M, Qu J. Ultrasensitive quantification of serum vitamin D metabolites using selective solid- phase extraction coupled to microflow liquid chromatography and isotope-dilution mass spectrometry. Anal Chem. 2010;82:2488–97.

65. Ding S, Schoenmakers I, Jones K, Koulman A, Prentice A, Volmer D a. Quantitative determination of vitamin D metabolites in plasma using UHPLC- MS/MS. Anal Bioanal Chem. 2010;398:779–89.

66. Higashi T, Shibayama Y, Fuji M, Shimada K. Liquid chromatography-tandem mass spectrometric method for the determination of salivary 25-hydroxyvitamin D3: a noninvasive tool for the assessment of vitamin D status. Anal Bioanal Chem. 2008;391:229–38.

67. Higashi T, Suzuki M, Hanai J, Inagaki S, Min JZ, Shimada K, Toyo’oka T. A specific LC/ESI-MS/MS method for determination of 25-hydroxyvitamin D3 in

neonatal dried blood spots containing a potential interfering metabolite, 3-epi-25- hydroxyvitamin D3. J Sep Sci. 2011;34:725–32.

68. Ogawa S, Ooki S, Morohashi M, Yamagata K, Higashi T. A novel Cookson-type reagent for enhancing sensitivity and specificity in assessment of infant vitamin D status using liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom. 2013;27:2453–60.

69. Ismail B, Nielsen SS. Basic Principles of Chromatography. In: Nielsen SS, editor. Food Analysis. New York, NY: Springer Science+Business Media; 2010. p. 473– 98.

70. Sargent M, Harte R, Harrington C. Guidelines for Achieving High Accuracy in Isotope Dilution Mass Spectrometry (IDMS). Cambridge: Royal Society of Chemistry; 2002.

71. Thompson JN, Plouffe L. Determination of cholecalciferol in meat and fat from livestock fed normal and excessive quantities of vitamin D. Food Chem. 1993;46:313–8.

72. Mattila PH, Piironen VI, Uusi-Rauva EJ, Koivistoinen PE. Contents of Cholecalciferol, Ergocalciferol, and Their 25-Hydroxylated Metabolites in Milk Products and Raw Meat and Liver As Determined by HPLC. J Agric Food Chem. 1995;43:2394–9.

73. Strobel N, Buddhadasa S, Adorno P, Stockham K, Greenfield H. Vitamin D and 25-hydroxyvitamin D determination in meats by LC – IT-MS. Food Chem. 2013;138:1042–7.

74. Höller U, Quintana AP, Gössl R, Olszewski K, Riss G, Schattner A, Nunes CS. Rapid determination of 25-hydroxy vitamin D3 in swine tissue using an isotope dilution HPLC-MS assay. J Chromatogr B Analyt Technol Biomed Life Sci. 2010;878:963–8.

75. Jakobsen J, Clausen I, Leth T, Ovesen L. A new method for the determination of vitamin D3 and 25-hydroxyvitamin D3 in meat. J Food Compos Anal. 2004;17:777–87.

76. Wertz AE, Knight TJ, Trenkle A, Sonon R, Horst RL, Huff-Lonergan EJ, Beitz DC. Feeding 25-hydroxyvitamin D3 to improve beef tenderness. J Anim Sci. 2004;82:1410–8.

77. Purchas R, Zou M, Pearce P, Jackson F. Concentrations of vitamin D3 and 25- hydroxyvitamin D3 in raw and cooked New Zealand beef and lamb. J Food Compos Anal. 2007;20:90–8.

78. Lipkie TE, Janasch A, Cooper BR, Hohman EE, Weaver CM, Ferruzzi MG. Quantification of vitamin D and 25-hydroxyvitamin D in soft tissues by liquid chromatography-tandem mass spectrometry. J Chromatogr B. 2013;932:6–11.

79. Benavides F, Oberyszyn TM, VanBuskirk AM, Reeve VE, Kusewitt DF. The hairless mouse in skin research. J Dermatol Sci. 2009;53:10–8.

80. Wingerath T, Sies H, Stahl W. Xanthophyll esters in human skin. Arch Biochem Biophys. 1998;355:271–4.

81. Okano T, Tsugawa N, Kobayashi T. Identification and Determination of 1a,25- dihydroxyvitamin D3 in Rat Skin by High Performance Liquid Chromatography and Radioreceptor Assay. J Chromatogr. 1989;493:63–70.

82. Stahl W, Schwarz W, Sundquist a R, Sies H. Cis-Trans Isomers of Lycopene and Beta-Carotene in Human Serum and Tissues. Arch Biochem Biophys. 1992;294:173–7.

83. Huq S, Safan EP. Rapid and Effective Cleanup of Vitamin D2 and D3 from Human Plasma using Phree Phospholipid Removal PLates and Kinetex Core-Shell HPLC/UPLC Columns. Phenomoenex Technical Note. Torrance, CA; p. 4.

84. Kopec RE. Bioavailability, Metabolism, and Bioefficacy of Tomato Carotenoids. The Ohio State University; 2012. p. 184.

85. Hanai T. HPLC, a practical guide. Smith, Rog. Cambridge, UK: The Royal Society of Chemistry; 1999.

86. Snyder LR, Kirkland JJ, Glajch JL. Practical HPLC Method Development. New York, NY: Wiley; 1997.

87. Reuhs BL, Rounds MA. High-Performance Liquid Chromatography. In: Nielsen SS, editor. Food Analysis. New York, NY: Springer Science+Business Media; 2010. p. 499–512.

88. Déglon J, Thomas A, Mangin P, Staub C. Direct analysis of dried blood spots coupled with mass spectrometry: concepts and biomedical applications. Anal Bioanal Chem. 2012;402:2485–98.

89. Heath AK, Williamson EJ, Ebeling PR, Kvaskoff D, Eyles DW, English DR. Measurements of 25-hydroxyvitamin D concentrations in archived dried blood spots are reliable and accurately reflect those in plasma. J Clin Endocrinol Metab. 2014;jc20141269. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24885629

90. Hanahan D, Weinberg R a. Hallmarks of cancer: the next generation. Cell. 2011;144:646–74.

91. Liu NQ, Ouyang Y, Bulut Y, Lagishetty V, Chan SY, Hollis BW, Wagner C, Equils O, Hewison M. Dietary vitamin D restriction in pregnant female mice is associated with maternal hypertension and altered placental and fetal development. Endocrinology. 2013;154:2270–80.

92. Sergeev IN, Song Q. High vitamin D and calcium intakes reduce diet-induced obesity in mice by increasing adipose tissue apoptosis. Mol Nutr Food Res. 2014;58:1342–8.

93. Goerz G, Merk H, Bolsen K, Tsambaos D, Berger H. Influence of chronic UV- light exposure on heparic and cutaneous monooxygenases. Experientia. 1983;39:385–6.

94. Prosser DE, Jones G. Enzymes involved in the activation and inactivation of vitamin D. Trends Biochem Sci. 2004;29:664–73.

95. Pascussi JM, Robert A, Nguyen M, Walrant-debray O, Garabedian M, Martin P, Pineau T, Saric J, Navarro F, et al. Possible involvement of pregnane X receptor – enhanced CYP24 expression in drug-induced osteomalacia. J Clin Invest. 2005;115:177–86.

96. Hidiroglou M. Kinetics of Intravenously Administered 25-Hydroxyvitamin D3 in Sheep and the Effect of Exposure to Ultraviolet Radiation. J Anim Sci. 1987;65:808–14.

97. Binkley N, Novotny R, Krueger D, Kawahara T, Daida YG, Lensmeyer G, Hollis BW, Drezner MK. Low Vitamin D Status despite Abundant Sun Exposure. 2014;92:2130–5.

98. Engelman C, Bo R, M Z, Steltenpohl H, Kirby T, Nieto F. Epidemiologic study of the C-3 epimer of 25-hydroxyvitamin D3 in a population-based sample. Clin Nutr. 2014;33:421–5.

99. Thomas-Ahner JM, Wulff BC, Tober KL, Kusewitt DF, Riggenbach JA, Oberyszyn TM. Gender differences in UVB-induced skin carcinogenesis, inflammation, and DNA damage. Cancer Res. 2007;67:3468–74.

100. Reddy GS, Muralidharan K, Okamura W, Tserng K, McLane J. Metabolism of 1alpha,25-dihydroxyvitamin D3 and its C-3 epimer 1alpha,25-dihydroxy-3-epi- vitamin D3 in neonatal human keratinocytes. Steroids. 2001;66:441–50.

101. Kamao M, Tatematsu S, Sawada N, Sakaki T, Hatakeyama S, Kubodera N, Okano T. Cell specificity and properties of the C-3 epimerization of Vitamin D3 metabolites. J Steroid Biochem Mol Biol. 2004;89-90:39–42.

102. Harper AR. The Effects of Dietary Administration of Vitamin D on UVB Induced Skin Inflammation and Carcinogenesis. The Ohio State University; 2010. p. 36.

103. Schwartz G, Hulka B. Is vitamin D deficiency a risk factor for prostate cancer? (Hypothesis). Anticancer Res. 1990;10:1307–11.

104. Hanchette CL, Schwartz GG. Geographic patterns of prostate cancer mortality: Evidence for a protective effect of ultraviolet radiation. Cancer. 1992;70:2861–9.

105. John EM, Koo J, Schwartz GG. Sun exposure and prostate cancer risk: evidence for a protective effect of early-life exposure. Cancer Epidemiol Biomarkers Prev. 2007;16:1283–6.

106. Ahonen MH, Tenkanen L, Teppo L, Hakama M, Tuohimaa P. Prostate cancer risk and prediagnostic serum 25-hydroxyvitamin D levels (Finland). Cancer Causes Control. 2000;11:847–52.

107. Corder EH, Guess HA, Hulka BS, Friedman GD, Sadler M, Vollmer RT, Lobaugh B, Drezner MK, Vogelman JH, Orentreich N. Vitamin D and prostate cancer: a prediagnostic study with stored sera. Cancer Epidemiol Biomarkers Prev. 1993;2:467–72.

108. Braun MM, Helzlsouer KJ, Hollis BW, Comstock GW. Prostate cancer and prediagnostic levels of serum vitamin D metabolites (Maryland, United States). Cancer Causes Control. 1995;6:235–9.

109. Nomura AM, Stemmermann GN, Lee J, Kolonel LN, Chen TC, Turner A, Holick MF. Serum vitamin D metabolite levels and the subsequent development of prostate cancer (Hawaii, United States). Cancer Causes Control. 1998;9:425–32.

110. Li H, Stampfer MJ, Hollis JBW, Mucci L a, Gaziano JM, Hunter D, Giovannucci EL, Ma J. A prospective study of plasma vitamin D metabolites, vitamin D

receptor polymorphisms, and prostate cancer. PLoS Med. 2007;4:e103. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1831738&tool=pmcen trez&rendertype=abstract

111. Gilbert R, Metcalfe C, Fraser WD, Donovan J, Hamdy F, Neal DE, Lane JA, Martin RM. Associations of Circulating 25-hydroxyvitamin D with prostate cancer diagnosis , stage and grade. Int J cancer. 2012;131:1187–96.

112. Faupel-Badger JM, Diaw L, Albanes D, Virtamo J, Woodson K, Tangrea J a. Lack of association between serum levels of 25-hydroxyvitamin D and the subsequent risk of prostate cancer in Finnish men. Cancer Epidemiol biomarkers Prev. 2007;16:2784–6.

113. Travis RC, Crowe FL, Allen NE, Appleby PN, Roddam AW, Tjønneland A, Olsen A, Linseisen J, Kaaks R, et al. Serum vitamin D and risk of prostate cancer in a case-control analysis nested within the European Prospective Investigation into Cancer and Nutrition (EPIC). Am J Epidemiol. 2009;169:1223–32.

114. Barnett CM, Nielson CM, Shannon J, Chan JM, Shikany JM, Bauer DC, Hoffman AR, Barrett-Connor E, Orwoll E, Beer TM. Serum 25-OH vitamin D levels and risk of developing prostate cancer in older men. Cancer Causes Control. 2010;21:1297–303.

115. Park SY, Cooney RV, Wilkens LR, Murphy SP, Henderson BE, Kolonel LN. Plasma 25-hydroxyvitamin D and prostate cancer risk: the multiethnic cohort. Eur J Cancer. 2010;46:932–6.

116. Ahn J, Peters U, Albanes D, Purdue MP, Abnet CC, Chatterjee N, Horst RL, Hollis BW, Huang W-Y, et al. Serum vitamin D concentration and prostate cancer risk: a nested case-control study. J Natl Cancer Inst. 2008;100:796–804.

117. Meyer HE, Robsahm TE, Bjørge T, Brustad M, Blomhoff R. Vitamin D, season, and risk of prostate cancer: a nested case-control study within Norwegian health studies. Am J Clin Nutr. 2013;25:147–54.

118. Tuohimaa P, Tenkanen L, Ahonen M, Lumme S, Jellum E, Hallmans G, Stattin P, Harvei S, Hakulinen T, et al. Both high and low levels of blood vitamin D are associated with a higher prostate cancer risk: a longitudinal, nested case-control study in the Nordic countries. Int J cancer. 2004;108:104–8.

119. Kristal AR, Till C a, Song X, Tangen CM, Goodman PJ, Neuhouser ML, Schenk JM, Thompson IM, Meyskens FL, et al. Plasma Vitamin D and Prostate Cancer Risk; Results from the Selenium and Vitamin E Cancer Prevention Trial. Cancer

Epidemiol Biomarkers Prev. 2014; Available from: http://www.ncbi.nlm.nih.gov/pubmed/24732629

120. Gandini S, Boniol M, Haukka J, Byrnes G, Cox B, Sneyd MJ, Mullie P, Autier P. Meta-analysis of observational studies of serum 25-hydroxyvitamin D levels and colorectal, breast and prostate cancer and colorectal adenoma. Int J Cancer. 2011;128:1414–24.

121. Gilbert R, Martin RM, Beynon R, Harris R, Savovic J, Zuccolo L, Bekkering GE, Fraser WD, Sterne JC, Metcalfe C. Associations of circulating and dietary vitamin D with prostate cancer risk: a systematic review and dose-response meta-analysis. Cancer Causes Control. 2011;22:319–40.

122. Schwartz GG, Whitlatch LW, Chen TC, Lokeshwar BL, Holick MF. Human prostate cells synthesize 1,25-dihydroxyvitamin D3 from 25-hydroxyvitamin D3. Cancer Epidemiol biomarkers Prev. 1998;7:391–5.

123. Welsh J. Cellular and molecular effects of vitamin D on carcinogenesis. Arch Biochem Biophys. Elsevier Inc.; 2012;523:107–14.

124. Moreno J, Krishnan A V, Peehl DM, Feldman D. Mechanisms of vitamin D- mediated growth inhibition in prostate cancer cells: inhibition of the prostaglandin pathway. Anticancer Res. 2006;26:2525–30.

125. Ben-Shoshan M, Amir S, Dang DT, Dang LH, Weisman Y, Mabjeesh NJ. 1alpha,25-dihydroxyvitamin D3 (Calcitriol) inhibits hypoxia-inducible factor- 1/vascular endothelial growth factor pathway in human cancer cells. Mol Cancer Ther. 2007;6:1433–9.

126. Kovalenko PL, Zhang Z, Cui M, Clinton SK, Fleet JC. 1,25 dihydroxyvitamin D- mediated orchestration of anticancer, transcript-level effects in the immortalized, non-transformed prostate epithelial cell line, RWPE1. BMC Genomics. 2010;11:1–15.

127. Chung I, Han G, Seshadri M, Gillard BM, Yu W, Foster B a, Trump DL, Johnson CS. Role of vitamin D receptor in the antiproliferative effects of calcitriol in tumor-derived endothelial cells and tumor angiogenesis in vivo. Cancer Res. 2009;69:967–75.

128. Mordan-McCombs S, Brown T, Wang W-LW, Gaupel A-C, Welsh J, Tenniswood M. Tumor progression in the LPB-Tag transgenic model of prostate cancer is altered by vitamin D receptor and serum testosterone status. J Steroid Biochem Mol Biol. 2010;121:368–71.

129. Kovalenko PL, Zhang Z, Yu J-G, Li Y, Clinton SK, Fleet JC. Dietary vitamin D and vitamin D receptor level modulate epithelial cell proliferation and apoptosis in the prostate. Cancer Prev Res. 2011;4:1617–25.

130. Zheng Y, Zhou H, Ooi LL, Snir AD, Dunstan CR, Seibel MJ. Vitamin D deficiency promotes prostate cancer growth in bone. Prostate. 2011;71:1012–21.

131. Kris-Etherton PM, Hecker KD, Bonanome A, Coval SM, Binkoski AE, Hilpert KF, Griel AE, Etherton TD. Bioactive Compounds in Foods : Their Role in the Prevention of Cardiovascular Disease and Cancer. Am J Med. 2002;113:71S–88S.

132. Castle EP, Thrasher JB. The role of soy phytoestrogens in prostate cancer. Urol Clin North Am. 2002;29:71–81.

133. Lamartiniere CA, Cotroneo MS, Fritz WA, Wang J, Mentor-Marcel R, Elgavish A. Genistein Chemoprevention : Timing and Mechanisms of Action in Murine Mammary and Prostate. J Nutr. 2002;132:552S–558S.

134. Badawi A. The role of prostaglandin synthesis in prostate cancer. BJU Int. 2000;85:451–162.

135. Swami S, Krishnan A V., Moreno J, Bhattacharyya RB, Peehl DM, Feldman D. Calcitriol and Genistein Actions to Inhibit the Prostaglandin Pathway: Potential Combination Therapy to Treat Prostate Cancer. J Nutr. 2007;137:205S–210S.

136. Rao A, Woodruff RD, Wade WN, Kute TE, Cramer SD. Genistein and Vitamin D Synergistically Inhibit Human Prostatic Epithelial Cell Growth. J Nutr. 2002;132:3191–4.

137. Wang JY, Swami S, Krishnan A V., Feldman D. Combination of calcitriol and dietary soy exhibits enhanced anticancer activity and increased hypercalcemic toxicity in a mouse xenograft model of prostate cancer. Prostate. 2012;72:1628– 37.

138. Farhan H, Wähälä K, Adlercreutz H, Cross HS. Isoflavonoids inhibit catabolism of vitamin D in prostate cancer cells. J Chromatogr B. 2002;777:261–8.

139. Byrdwell WC, Devries J, Exler J, Harnly JM, Holden JM, Holick MF, Hollis BW, Horst RL, Lada M, et al. Analyzing vitamin D in foods and supplements: methodologic challenges. Am J Clin Nutr. 2008;88:554S–557S.

140. Ahn-Jarvis JH, Riedl KM, Schwartz SJ, Vodovotz Y. Design and selection of soy breads used for evaluating isoflavone bioavailability in clinical trials. J Agric Food Chem. 2013;61:3111–20.

141. Huang M, LaLuzerne P, Winters D, Sullivan D. Measurement of vitamin D in foods and nutritional supplements by liquid chromatography/tandem mass spectrometry. J AOAC Int. 2009;92:1327–35.

142. Jakobsen J, Knuthsen P. Stability of vitamin D in foodstuffs during cooking. Food Chem. 2014;148:170–5.

Appendix A:

Development of a Vitamin D2 Fortified Soy Bread for Use in Clinical Trials.

A-1.1 Introduction: Rationale for dietary delivery of vitamin D with soy to prevent prostate cancer

Most of the data supporting a role of vitamin D in prostate cancer prevention originates from association studies in human populations. Schwartz and Hulka were the first to associate increased prostate cancer risk with factors that likely reduce cutaneous synthesis of vitamin D. These factors included advanced age, dark skin pigmentation, and residence in northern latitudes (103). Later studies associated low UVB exposure with greater prostate cancer risk and high early-life exposure to UVB light with lower risk (104,105). Other epidemiological studies have attempted to find a relationship between serum vitamin D status and prostate cancer risk. A study in Finnish men enrolled in the Helinski Heart Study concluded that low pre-diagnostic serum 25OH D was associated with an increased risk of prostate cancer (106). Another study of northern

Californian men inversely associated prostate cancer risk with higher serum 1,25(OH)2D levels, especially if serum 25OH D levels were low (107). Conversely, two studies involving men living in Maryland or Hawaii both demonstrated a lack of correlation between prediagnostic serum vitamin D and prostate cancer risk (108,109). Other studies have demonstrated null effects as well (110–115). Further contrasting these results, a study of the men enrolled in the Prostate, Lung, Colorectal, and Ovarian Cancer

Screening Trial proposed the possibility that higher baseline vitamin D status may be associated with higher risk of prostate cancer. However, the association between 73 vitamin D and prostate cancer in this study seemed non-linear (116). Meyer and colleagues recently observed similar results in a nested case-control study of Norwegian men, but a positive association with serum 25OH D levels and prostate cancer risk was only evident in the summer and autumn (117). Interestingly, two studies have suggested

U-shaped relationships for prediagnostic serum 25OH D and prostate cancer, suggesting an optimal vitamin D status for preventing the disease (118,119) Due to their conflicting nature, meta-analyses of these epidemiological studies have shown null effects of elevated vitamin D status in prostate cancer while highlighting the incredible amount of variability in the available data (120,121). As mentioned before, epidemiology of vitamin D is complicated due to the number of confounding factors including sun exposure and population heterogeneity.

Since the epidemiological evidence is so conflicted, investigators have begun to rely on laboratory studies to clarify the relationship between vitamin D and prostate cancer. Cell studies primarily evaluate 1,25(OH)2 D, as previous studies showed that prostate cells can produce this active metabolite of vitamin D, suggesting that local production may regulate prostate cancer (122). 1,25(OH)2 D3 is an agonist for the nuclear vitamin D receptor, and can act as a regulator of several genes related to prostate cancer progression including those involved in growth factor signaling, prostaglandin metabolism, apoptosis, inflammation, and tumor progression (123,124). Additional research has demonstrated that 1,25(OH)2 D3 also exhibits anti-angiogenic properties

(125,126) and reduces oxidative stress (126). Animal studies have shown that the anticancer effects of vitamin D are likely mediated through the VDR (127–129). Vitamin D

74 deficiency has also been shown to promote early prostate cancer progression (129) as well as the growth of bone metastases (130) in mouse models of prostate cancer. There is no clear explanation as to why a dichotomy exists between the laboratory and epidemiological data for the role of vitamin D in prostate cancer. Due to strong laboratory data and the problematic nature of vitamin D epidemiology, researchers consistently call for randomized, double-blind clinical trials to help shed light on this problem and provide more definitive answers that may be translated into recommendations for prostate cancer patients.

Emerging evidence has suggested that vitamin D may interact with isoflavones, phytoestrogens found in soy, to synergistically inhibit prostate cancer. Generally, an association has been drawn between soy consumption and lower prostate cancer incidence with isoflavones being implicated as the bioactive component (131). The potential anti-prostate cancer activity of soy has been demonstrated in pre-clinical models

(132,133). These effects may be due to isoflavone-mediated induction/inhibition of multiple cell signaling pathways, inhibition of angiogenesis, cell cycle interruption, and induction of apoptosis (132). Most of the research on isoflavones has specifically focused on the compound genistein.

Genistein and vitamin D are thought to act synergistically through the prostaglandin pathway. This pathway signals for cell growth, metastasis, angiogenesis, and apoptosis beginning with the conversion of arachidonic acid to prostaglandins by cyclooxygenase-2 (COX-2). Prostaglandins subsequently bind to a receptor, signaling a cascade of cellular events that have been shown to accentuate the growth of prostate 75 cancer (134). Genistein has been shown to inhibit the activity and expression of COX-2.

1,25(OH)2 D has also been shown to inhibit the expression of COX-2 as well as the prostaglandin receptor while also activating the enzyme responsible for prostaglandin degradation. Interestingly, genistein also inhibits CYP24A1, the enzyme responsible for the degradation of 1,25(OH)2 D (135). The inhibition of 1,25(OH)2 D degradation by genistein allows for prolonged action of the active vitamin D metabolite (figure 12).

Collectively, the interaction between these two compounds on the prostaglandin pathway allow genistein and 1,25(OH)2 D to exhibit greater anti-prostate cancer activity together than they do alone. This has been shown in vitro as well as with in vivo prostate cancer xenografts (136,137). Other isoflavones, such as daidzein, have shown some similar activity (138). This intersection of soy and vitamin D offers interesting possibilities for combined preventative therapy using dietary components.

degradation

Arachidonic Acid CYP24A GENISTEIN 1 expression, activity expression COX-2 1,25(OH)2 D

Prostaglandin 15PGD Prostaglandin E2 degradation H

Receptor

Angiogenesis Apoptosis Metastasis Proliferation

Figure 12: Interaction of genistein and 1,25(OH)2 D3 on the prostaglandin pathway.

Adapted from Swami (135).

A-1.1.1 Vitamin D Analysis in Foods

The analysis of vitamin D in foods is rather complex due to necessary adaptations of extraction methods based on the physical properties of the product. Most extraction protocols follow a common procedure including saponification, extraction with a non- polar solvent such as hexane or petroleum ether, sample preparation using HPLC and/or solid phase extraction, and finally, separation by HPLC and detection. Vitamin D is detectable by photodiode array, but when extracted from food, a number of compounds that interfere with UV measurement are present. Thus, detection by MS is optimal even though it is more costly and slightly less precise than UV detection. The Association of

Official Analytic Chemists has a number of validated methods for vitamin D analysis in foods, but these methods are labor intensive, time consuming, and only validated for a select few materials (139). The amount of vitamin D in foods is measured in

“international units” (IU). The IU is a measure of biological activity or effect of a given dose of a material. In the case of vitamin D, 1 µg is equivalent to 40 IU (7).

Concentrations of vitamin D in food are much higher than those in biological samples, so techniques used to enhance signal, such as chemical derivatization, are not used.

A-1.1.2 Specific Aim A1 To develop a food product rich in vitamin D for use in cancer prevention clinical trials.

Previous research at The Ohio State University has resulted in the development of soy bread for use in prostate cancer clinical trials. The fortification of this soy bread with vitamin D is a logical next step, since emerging evidence suggests that soy isoflavones interact with vitamin D to synergistically inhibit prostate cancer. Lallemand Inc. has developed a yeast product enriched in vitamin D2, meant for application in the baking industry. Characterization of the fortified bread is essential before the product is used in further animal experiments and human clinical trials.

I hypothesize that the vitamin D2 rich yeast, when added to soy breads, will provide consistent concentrations of vitamin D at varying levels of addition. Furthermore, I predict that a small loss of vitamin D content may occur through the baking process, and small quantities vitamin D2 rich yeast can be added to a large dough mass and provide a consistent dose of the nutrient.

A-2.2 Materials and Methods

A-2.2.1 Bread Baking

Bread was manufactured using a sponge-dough process as described by Ahn-

Jarvis and colleagues (140) and described in figure 13. Briefly, the ingredients constituting the sponge mixture were combined in the bowl of a Kitchen Aid mixer

(model KV25GO, Kitchen Aid, St. Joseph, MI) affixed with a paddle mixer attachment.

The sponge was then allowed to rest for 2 hours at room temperature in a polyethylene bag. The rested sponge was then combined with the remaining ingredients to form the dough, using a Kitchen Aid mixture affixed with a dough hook attachment. The dough was formed into loaves which were then panned and proofed in a proofing cabinet

(CM2000 combination module Flavor View Inter Metro Industries Corp., Milkes-Barre,

PA, USA) set at 40 °C and maximum humidity for 1 hour. Following proofing, loaves were baked for 50 min at 150 °C in a convection oven (model JA14, Doyon, Liniere,

Quebec, Canada).

Before further sample processing, breads were cooled, sliced, frozen with liquid nitrogen, and lyophilized (FreeZone 12 Plus, Labconco; Kansas City, MO). A moisture content of 33.3% was assumed when back-calculating vitamin D dosing.

Sponge

-80% of formula water - 65% of formula wheat flour -Wheat gluten -50% of formula salt -50% of formula yeast

Rest 2 hr

Dough

-20% of formula water -35% of formula wheat flour -Soymilk/soy flour mix -50% of formula salt -Sugar -Shortening -50% of formula yeast -Dough conditioner

Proof 1 hr Bake 50 min

Figure 13. Vitamin D Soy Bread Baking Process

A-2.2.2 HPLC-MS/MS Analysis

Vitamin D2 content was measured in bread using methodology similar to Huang and colleagues (141) with considerable changes. Approximately 1.0 g of freeze-dried bread was weighed into 22 mL glass tubes. A mixture of 90:5:5 ethanol/methanol/2- propanol (4 mL) was added to the sample followed by 2 mL of 50% aqueous potassium hydroxide. This mixture was then shaken for five hours on a rotary arm shaker set at maximum speed. Following saponification, 3 mL hexane containing 0.1% butylated hydroxyl toluene were added to saponified bread. The resulting mixture was probe sonicated for 8 sec and vortexed for 30 sec prior to centrifugation for 10 min at 2000 x g.

The upper organic layers were isolated, and the hexane extraction was repeated twice more, pooling the extracts from each round. Aliquots (3 mL) were dried under nitrogen gas. Dried extracts were then resolubolized in 600 µL of ethanol and filtered through a

0.2 µm nylon filter prior to injection onto the HPLC-MS/MS system.

Samples (10 µL) were processed by an Agilent 1200 series HPLC (Santa Clara,

CA) equipped with a Symmetry C18 column (4.6x75 mm, 3.5 µm; Waters, Milford,

MA). Solvent conditions consisted of an isocratic elution of a 92.5% aqueous methanol solution supplemented with 0.1% formic acid for 10 min followed by a was for 2 min with 100% MTBE and equilibration at initial conditions for 3 min. Eluent from the

HPLC was directed to a QTRAP 5500 mass spectrometer (AB Sciex; Framingham, MA) equipped with atmospheric pressure chemical ionization (APCI) operated in positive ion mode. MS parameters were as follows: source temperature, 450 °C; curtain gas, 30 psi; ion source gas, 40 psi; declustering potential, 185 V; entrance potential, 10 V; collision

82 cell exit potential, 11 V. The transitions that were monitored and summed for quantification included: 397379 (CE=20 eV), 397271 (CE=20 eV), 397125

(CE=20 eV), 397187 (CE=35 eV). A dwell time of 107 msec was used for all transitions.

A-2.2.3 Fortification Curve

Breads formulated with 0.4, 0.6, 0.8, and 1.2 g VDY per loaf (n=3 at each amount) were manufactured and assayed for vitamin D2 content.

A-2.2.4 Homogeneity Study

Two dough masses were made each with 0.48 g VDY. For the first mass, the vitamin D yeast was added in the sponging stage of the bread baking process by completely suspending the yeast in the formula water. The VDY and water were mixed for 15 seconds on the lowest setting of a Kitchen Aid mixer, and then the additional ingredients were added to complete the sponge (Dough 1). The second mass was supplemented with the VDY at the dough formation stage, immediately following fermentation of the sponge. Similarly, the yeast was completely suspended in the remaining formula water and added to the sponge with the other ingredients to complete the bread dough (Dough 2).

The completed dough masses were broken into three portions of approximately equal mass and placed into miniature loaf pans. The miniature loaves were then baked to an internal temperature of 92 °C. Following baking, the loaves were cooled to room

83 temperature. Each loaf was then individually frozen in liquid nitrogen, lyophilized, pulverized into a powder, and assayed in triplicate for vitamin D2 content.

A-2.2.5 Vitamin D Retention over Baking

The retention of vitamin D content of the bread after baking was also evaluated.

Retention here is defined as the percentage of vitamin D2 content remaining in the bread when compared to the content in the unbaked dough. Three breads with identical amounts of VDY added were prepared. A 10 g sample from each loaf was taken before the breads were baked. The dough samples were immediately frozen with liquid nitrogen and lyophilized alongside the finished bread. The three dough samples and corresponding bread loaves were assayed for vitamin D2 content.

A-3.1 Results and Discussion

A-3.1.1 Fortification Curve

In order to facilitate the manufacture of vitamin D soy breads with varying levels of vitamin D2, the dose achieved per slice of bread with different levels of formula VDY was investigated. This fortification curve is shown in figure 13. The achieved doses of vitamin D2 increased proportionally with the added VDY over a wide range (660-1980

IU/slice; R2 = 0.9999). This relationship is described in equation A1. Additionally, the achieved dose was extremely reproducible at each point with standard deviations only ranging from 10 to 50 IU/slice. It is possible that the slope of the fortification curve will change slightly between batches of VDY, as the concentration of vitamin D2 in the yeast may vary.

[A1]

Where:

2000

R2 = 0.9999

1500

(IU/slice)

2 1000

500 Vitamin D Vitamin

0 0 0.5 1 1.5

Vitamin D2 rich yeast (g)

Figure 14: Bread fortification curve relating amount of added vitamin D2 rich yeast added to bread dough and achieved dose per slice of bread.

A-3.1.2 Homogeneity Study

To evaluate whether or not the distribution of vitamin D within a bread loaf is affected by when the VDY is added, two dough masses were made each with identical amounts of VDY. The first dough mass was made by introducing VDY in the sponging stage, while in the second mass, VDY was added in the dough formation stage. Each dough mass was then partitioned three ways and these partitions were baked and assayed for vitamin D2 content in triplicate. From this analysis, described in table 3, it was clear that a homogeneous dose of vitamin D2 is better achieved when VDY is incorporated into the sponging stage of bread production. The amount of VDY included in the formulation is quite small, especially compared to the final mass of a loaf of bread. It is vital that each slice of bread from a loaf has identical doses of vitamin D2, as subjects consuming this product will be advised to eat only two slices per day. Additionally, these results should be considered in scale-up operations in which multiple loaves of bread are made from a single, large dough mass.

Table 3: Homogeneity study. Average vitamin D2 content (±SD) of breads made from partitions of larger dough masses which were fortified at different points in the bread making process.

Dough Mass 1* Dough Mass 2** Partition Number 1 2 3 1 2 3 Vitamin D2 (IU/g) 10.7 ± 0.63a 10.1 ± 0.22a 9.9 ± 0.16a 8.3 ± 0.18b 8.6 ± 0.22b 9.9 ± 0.27a

*Vitamin D2 yeast was added at the sponging stage ** Vitamin D2 yeast was added at the dough formation stage

A-3.1.3 Vitamin D Stability over Baking

The retention of vitamin D2 content was calculated as shown in equation A2

(142). The average retention of vitamin D2 in this soy bread system was 82.4%, indicating a loss of almost 18% of the vitamin D2 content. This is on par with observations of wheat and rye breads also fortified via VDY, 85% and 73%, respectively

(142). This loss should be considered if additional modifications are made to this formulation.

[A2]

A-4.1 Conclusions

The fortification of soy bread with vitamin D via vitamin D2 rich yeast was investigated in this work. Fortification was linear over a wide range of vitamin D doses, and vitamin D was stable over baking. In order to achieve a homogenous dose of vitamin

D throughout a loaf of bread, vitamin D2 rich yeast should be added early in the bread making process. The development of this product is expected to facilitate future research into the combined therapeutic effects of dietary vitamin D and soy.