Prevalence of two common vitamin D binding protein polymorphisms in a Swedish population

(1)

Örebro University

School of health sciences Department of clinical medicine

Biomedical Laboratory Science program Course: BL1701

Date: 160524

Prevalence of two common vitamin D

binding protein polymorphisms in a Swedish

population

Author: Ida Göthe Rosvall Supervisor: Lovisa Olsson, Clinical chemist, PhD Department of laboratory medicine, Molecular diagnostics

(2)

ABSTRACT

Vitamin D deficiency is not only associated with bone mineralization defects but also other disorders such as cancer, cardiovascular disease and diabetes. The vitamin D metabolite 25-hydroxyvitamin D (25(OH)D) has a longer half-life then the active vitamin D and is therefore the best indicator for evaluating the vitamin D status. While transported in the blood, the vitamin D is primarily bound to vitamin D binding protein (VDBP). The GC gene, coding for VDBP, contains two common single nucleotide polymorphisms (SNPs), rs7041T>G and rs4588C>A, affecting the concentration and binding affinity of the VDBP. The prevalence of the rs7041 and rs4588 alleles was examined in 386 Caucasian individuals using pyrosequencing. The allele frequencies were calculated from the genotypes using the Hardy-Weinberg equilibrium. The

correlation of the rs7041 and rs4588 to the 25(OH)D concentration was examined using Spearman’s test (α=0,01). The majority of the test population carried the rs7041G allele (61 %) and rs4588C allele (72 %) showing high similarity to the European population genetics. The rs7041 and rs4588 genotypes showed a weak correlation to the 25(OH)D level. Genotypes containing an increasing gradient of the rs7041T and rs4588A allele showed lower 25(OH)D concentrations, however, during wintertime the difference in 25(OH)D concentration between the genotypes of both SNPs seemed to decrease.

Key words: Vitamin D, Vitamin D binding protein, Single nucleotide polymorphisms,

(3)

ABBREVIATIONS

25(OH)D 25-hydroxyvitamin D

1,25(OH)2D 1,25-dihydroxyvitamin D

VDBP Vitamin D binding protein

GC GC gene (coding for VDBP)

DNA Deoxyribonucleic acid

T Thymine

G Guanine

C Cytosine

A Adenine

SNP Single nucleotide polymorphism

PCR Polymerase chain reaction

MgCl2 Magnesium chloride

dNTP deoxynucleotide triphosphate

APS adenosine-5´-phosphosulfate

PPi pyrophosphate

ATP adenosine triphosphate

dATPαS adenosine-α-thiotri-phosphate MilliQ water Ultrapure quality water of “type 1”

(4)

INTRODUCTIONS

Vitamin D

Vitamin D exists in two forms: ergocalciferol (D2) and cholecalciferol (D3). D3 is mainly synthesized under the influence of UVB light on 7-dehydrocholesterol in the skin but can also be obtained from animal products. D2 can only be obtained from the diet, however both from plants and animal products. The synthesized D3 and the vitamin D absorbed in jejunum is transported to the liver where the vitamin is

hydroxylated to the metabolite calcidiol or 25-hydroxyvitamin D (25(OH)D). 25(OH)D is then transported to the kidneys, where it is further hydroxylated to calcitriol, the active 1,25-dihydroxyvitamin D (1,25(OH)2D) (1). The 1,25(OH)2D promotes the

mobilization of calcium and phosphate to the circulation from both bowl and bone, and reabsorbs filtrated calcium ions (Ca2+) from the kidneys (figure 1) (1,2,3). The

hydroxylation process in the kidneys is stimulated by parathyroid hormone (1).

Secretion of parathyroid hormone is controlled by serum calcium concentration through negative feedback (2,3).

Because the 1,25(OH)2D level controls the mobilization of calcium, vitamin D

deficiency can contribute to bone loss (4) and the mineralization defects of the bone tissue can in severe cases lead to osteomalacia in adults and rickets in children (1, 4). Vitamin D deficiency can also be associated with other disorders like diabetes, cancer (1,5,6), cardiovascular disease (1,6), multiple sclerosis, rheumatoid arthritis and tuberculosis (6). 1,25(OH)2D has a short half-life (7) and only exists in a small

concentration in plasma (1,4). Measuring 1,25(OH)2D can lead to misinterpretation

regarding the vitamin D status as normal levels can occur due to elevated parathyroid hormone levels caused by the 25(OH)D deficiency (4). 25(OH)D is therefore a better indicator for evaluating vitamin D status (1,4).

Vitamin D concentration

Because the majority of vitamin D is synthesized in the skin, the plasma concentration of vitamin D alters depending on season in temperate climates. The highest

(7)

(2). However, how effective the sun exposure is also depends on clothing, sunscreen, pigmentation, time of day and age. Elderly people produce 75 % less cutaneous D3 than young adults (4). Individuals with darker skin demands sun exposure for a longer period of time to obtain sufficient vitamin D levels and therefore have a particular high risk of vitamin D deficiency living in a high altitude country like Sweden (8).

Pathological disorders like chronic kidney disease affects the conversion of 25(OH)D to 1,25(OH)2D and results in secondary vitamin D deficiency. With supplements, optimal

levels may be obtained (4). Obesity, type 1 diabetes mellitus, chronic liver diseases and different renal diseases also affect the vitamin D level negatively (5).

Pregnancy and oral contraceptive pills containing estrogen may lead to increased or unchanged 25(OH)D levels (5).

There is a debate about the sufficient level of 25(OH)D needed to obtain optimal calcium levels (9). The Endocrine Society´s clinical practice guideline recommends a concentration over the limit 75 nmol/L (30 ng/ml) for plasma 25(OH)D while the Institute of medicine states that 50 nmol/L (20 ng/ml) is sufficient (2,6). In Örebro county 75 nmol/L is used as a cut off value for 25(OH)D. Levels below 75 nmol/L are defined as insufficient (1), however it is a common finding in the healthy Swedish population during wintertime (4).

(8)

Figure 1. Vitamin D metabolism and physiological effects on target organs. The

majority of the vitamin D (D3) is synthesized by the UV effect on the skin. Vitamin D3 and vitamin D2 are hydroxylated to 25-hydroxyvitamin D in the liver and then

transported to the kidney where it can be further hydroxylated to 1,25-dihydroxyvitamin D. Parathyroid hormone levels regulate the kidney hydroxylation.

1,25-dihydroxyvitamin D is the active form of the vitamin and has the overall effect of increasing the plasma Ca2+_{level. Image from Molina. PE. Endocrine Physiology. 4}th_ed.

New York: McGraw-Hill Companies Inc; 2013 (3). Available from:

http://accessmedicine.mhmedical.com.db.ub.oru.se/ViewLarge.aspx?figid=42541007& gbosContainerID=0&gbosID=0.

Vitamin D binding protein

While transported in the blood the majority of the 25(OH)D and active 1,25(OH)2D is

bound to carrier proteins. 85-90 % of all 25(OH)D is bound to vitamin D binding protein (VDBP), 10-15 % is bound to albumin and 1 % is free. The albumin bound 25(OH)D and the free 25(OH)D are usually referred to as bioavailable vitamin D (6). VDBP is a 458 amino acid long protein belonging to the albumin family (10) and is coded for by the GC gene (Gene ID #2638) located on the q arm of chromosome 4

(9)

(4q11-13) (11). VDBP is expressed in the liver and can be found in plasma, ascites, cerebrospinal fluid and urine (10).

There are many polymorphic variations in the GC gene (10), two of them are single nucleotide polymorphisms (SNPs) rs7041 and rs4588 (6). Amino acid number 432 of the vitamin D binding protein may be aspartic acid (GAT) or glutamic acid (GAG). This variation is the result of rs7041, where the nitrogen base thymine (T) is exchanged to guanine (G) in the DNA sequence (1296T>G) (figure 2). Rs4588 is another SNP where nitrogen base cytosine (C) may be exchanged for adenine (A) (1307C>A), which results in the exchange of amino acid threonine (ACG) to lysine (AAG) in position 436 of the VDBP (figure 2). The combination of these two polymorphisms form three isoforms of the vitamin D binding protein entitled GC-1f (rs7041T, rs4588C), GC-1s (rs7041G, rs4588C) and GC-2 (rs7041T, rs4588A) (6).

These variants of the GC gene are known to alter the VDBP level and its binding affinity to 25(OH)D and therefore also the level of 25(OH)D (6,12). The SNPs rs7041 and rs4588 have been proven to be associated with race and ethnicity (6) and Europeans have in previous studies shown higher 25(OH)D levels than other ethnic groups (12).

Figure 2. DNA reference sequence for the GC gene (transcript GC-001:

ENST00000273951) showing the position of the rs7041 and rs4588 single nucleotide polymorphisms. The rs7041 may be T/G and is marked in yellow at position 1296 (NM_000583.3:c) of the transcript. The rs4588 may be C/A/T and is marked in turquoise at position 1307 (NM_000583.3:c) of the transcript. Image from Ensambl release 84, March 2016 © WTSI / EMBL-EBI and is available from:

http://www.ensembl.org/Homo_sapiens/Transcript/Sequence_cDNA?db=core;g=ENSG 00000145321;r=4:71741693-71804041;t=ENST00000273951.

(10)

VDBP concentration

Low levels of VDBP can be seen in obesity, type 1 diabetes mellitus, chronic liver diseases and renal diseases, same as for the level of 25(OH)D. The level of VDBP also decreases with age and is generally lower in the male gender (5).

The VDBP level is increased during pregnancy and during use of oral contraceptive pills. The free level of 25(OH)D however remains the same regardless of VDBP level. In healthy individuals with normal vitamin D levels, the biological effects of vitamin D on parathyroid hormone levels are mainly independent of VDBP concentration (5).

Prevalence of the VDBP variations

62 % of the world population carries the rs7041T allele, whereas 38 % carry the G allele (figure 3a). Homozygosity for the G allele is the least common genotype in the world population at only 19 % while 43 % of individuals are homozygote for the T allele and 39 % are heterozygote. In African populations the T allele is dominating with 91 %. In Europe the G allele is more common with 58 % against the 42 % for the T allele (figure 3b). The most common genotype in Europe is T/G heterozygosity (51 %) (13).

Figure 3a and 3b. 3a is showing the distribution of the rs7041 alleles T and G in the world population compared to the distribution in the European population in 3b. Image created by Göthe Rosvall I.

In the world population 79 % carry the rs4588C allele versus 21 % for the A allele (figure 4a). In the African population the C allele also dominates with 93 % C to 7 % A. In the European population the conditions are the same, the C allele dominates over the A allele with 75 % to 25 % (figure 4b). The most common genotype in Europe is homozygosity for the C allele (56 %) (14). Rs4588 can also result in the transversion of threonine (ACG) to methionine (ATG) but this transversion is very rare and there is no available data on allele frequency for the T allele (15).

62% 38%

rs7041 allele frequency - World

T

G

42% 58%

rs7041 allele frequency - Europe

T

G

(11)

Figure 4a and 4b. 4a is showing the distribution of the rs4588 alleles A and C in the world population compared to the distribution in the European population in 4b. Image created by Göthe Rosvall I.

Polymerase chain reaction

Polymerase chain reaction (PCR) is an amplification method used to create multiple copies of a specific DNA sequence for molecular analysis. The method was developed in 1983 by Kary Mullis and is based on DNA replication. The reaction in vitro uses single stranded DNA as template, primer sequences (short oligonucleotides)

complementary to the template DNA, the enzyme DNA polymerase, nucleotides, a PCR buffer providing optimal conditions for the enzyme activity and if necessary extra MgCl2. The double stranded DNA is first denatured to single stranded DNA at a high

temperature (95 °C). The temperature is then reduced to around 60 °C so that the primers can anneal specifically to the DNA template. The primers determine the specificity of the PCR, and it is therefore important to optimize the annealing

temperature. If the annealing temperature is too low the primers will bind unspecifically and if the temperature is too high the primers might not bind properly at all and there will be no amplification. Annealing temperature differs depending on the reaction conditions and the length and content of the primers and must therefore be carefully optimized. The annealing temperature is usually slightly lower then the melting

temperature (Tm) of the primer. After the annealing step the temperature is raised to the

optimal temperature (72 °C) for the DNA polymerase that adds nucleotides, extending the primer sequence, using the single stranded DNA as template. After polymerization the amplicons are denatured by once again raising the temperature and a new cycle begins. The generated DNA fragments are used as template in the next reaction cycle, which creates an exponential increase of the DNA fragment (16).

21%

79%

rs4588 allele frequency - World

A

C

25%

75%

rs4588 allele frequency - Europe

A

C

(12)

The salt concentration of the reaction affects the denaturing and annealing temperatures as well as the activity of the DNA polymerase. High salt concentrations promote the amplification of short DNA products. Too few Mg2+ ions makes enzyme activity slow while too many Mg2+ ions promotes misincorporated nucleotides (16). For

pyrosequencing purpose the DNA amplification is performed with one biotin labeled primer. The amplicons will then also be biotin labeled and can be washed and separated using streptavidin covered magnetic beads (17).

Pyrosequencing

Pyrosequencing is a technique today used for detection of SNPs, mutations and DNA methylation analysis (18). The method was developed at the Royal Institute of

Technology (KTH) in Stockholm, Sweden 1996 by Pål Nyrén and Mostafa Ronaghi (19). A sequencing primer, hybridized to the washed and denatured single stranded amplicons, is added to the pyrosequencing reaction along with nucleotides (dNTPs), enzymes and substrates. The enzymes used in the reaction are DNA-polymerase without exonuclease activity, ATP sulfurylase, luciferase and apyrase. The substrates used in the reaction are adenosine 5´phosphosulphate (APS) and luciferin. Nucleotides are added to the reaction in a predetermined order. When a nucleotide complementary to the

template is added to the reaction, the DNA polymerase will add it to the growing strand. When the nucleotide forms the phosphodiester bond to the previous nucleotide,

pyrophosphate (PPi) is released. PPi and APS will in the presence of ATP-sulfurylase create adenosine triphosphate (ATP). In the presence of ATP, luciferase can convert luciferin to oxyluciferin. The reaction generates a luminescent signal that is detected by a CCD camera and registered as a peak in a pyrogram (figure 5). The height of the peaks correspond to the number or nucleotides added and it is therefore possible to decide the sequence of the fragment. Excess nucleotides are degraded by apyrase (16). Deoxyadenosine-α-thiotri-phosphate (dATPαS) is used as a substitute for the nucleotide dATP since false luminescence signals can be generated by luciferase (19).

(13)

Figure 5. Enzymatic reaction in the presence of pyrophosphate (PPi). The enzyme sulfurylase coverts adenosine 5´phosphosulphate (APS) and PPi to adenosine triphosphate (ATP). The ATP allows the enzyme luciferase to convert luciferin to oxyluciferin while a luminescent reaction is generated. The luminescence is registered as a peak in a pyrogram. Image created by Göthe Rosvall I.

Aim

Genotyping of the vitamin D binding protein gene polymorphisms may be a tool for interpreting vitamin D status. The aim of this study is to examine the prevalence of the rs7041 and rs4588 alleles of the GC gene in a Swedish population using the

pyrosequencing method and to examine the genotype association with the 25(OH)D concentration.

(14)

MATERIAL AND METHOD

Participants

389 senior Caucasian individuals were recruited from central Sweden during 2003 and 2004. Out of the 389 individuals 127 were male and 262 were female, with a mean age of 74 ± 5 years (20). Olsson L. selected these individuals for the previous study about vitamin D levels of active seniors “Subjective well-being in old age and its association with biochemical and genetic biomarkers and with physical activity” from 2015 (20).

Ethical consideration

The Research Ethics Committee of the Örebro County Council and the Regional Ethical Board, Uppsala, Sweden approved of the active senior study (Dnr 2011/422). All

participants received written information about the studies and gave a specific written consent to the studies (20).

Clinical samples

Venous blood samples were in 2003 and 2004 gathered in serum tubes and 4,5 ml EDTA tubes (Becton Dickinson, Stockholm, Sweden) from each of the 389 individuals (20).

The 25(OH)D level was in 2014 measured from the serum, stored in -80 °C, using liquid chromatography tandem mass spectrophotometry (LC-MS/MS) at Örebro

University Hospital, Sweden. Records of the results from 2014 were used for this study. The genomic DNA used for this study was in 2004-2010 extracted and purified from the 389 active seniors from 200 µl whole blood anticoagulated with EDTA, using QIAamp DNA blood mini kit according to the manufacturer´s instructions (Qiagen Inc, Hilden, Germany) (20), stored in -20 °C and in 2011 to 2012 diluted to 5 ng/µl.

Optimization of the PCR

The PCR method was optimized using clinical samples of genomic DNA from unidentified individuals stored in -20 °C. The DNA concentration of the samples was after being measured, using Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, USA) according to the manufacturer´s instructions, diluted in water (MilliQ quality) to a concentration of 10 ng/µl.

(15)

The PCR reactions were during optimization performed in a 50 µl reaction mixture adding 50 ng of DNA to 1 x PCR buffer, 1,5- 4,5 mM MgCl2, 200 µM of each dNTP,

1,25 U (0,025 U/µl) Hot Star plus polymerase (HotStarTaq plus DNA polymerase kit, Qiagen Inc) and 0,11 µM of each primer of the following primer sequences;

5´-GGTATAGAATTTTCTTGAGACAGGCAAGTA-´3 (VDP2F, forward) and Biotin- 5´-AAGATTCTGCCATGTTAAGTGGA-´3 (VDP2R, reverse) (Biomers, Ulm, Germany), according to the manufacturers’ instructions (Qiagen Inc).

The primer sequences were designed in PyroMark Assay design 2.0 (2009).

The amplification took place in Veriti 96 well Thermal cycler (AB Applied Biosystems, Thermo Fisher Scientific, Waltham, USA) starting with initial denaturation at 95 °C for 5 seconds, followed by 40 cycles of denaturation at 94 °C for 30 seconds, annealing at 56-68 °C for 30 seconds, extension at 72 °C for 60 seconds and a final extension at 72 °C for 7 minutes and was then cooled down to 4 °C.

The optimal salt concentration was settled using reaction mixtures containing the following concentrations of MgCl2; 1,5 mM, 2,5 mM, 3,5 mM and 4,5 mM.

The optimal annealing temperature for the PCR reaction was settled using a temperature gradient of 56, 59, 60, 62, 65 and 68 °C.

The optimal concentration of DNA needed for proper amplification was settled by diluting DNA samples to concentrations of 5, 10, 20 and 50 ng/µl and running the PCR in a 50 µl reaction containing 25 ng (0,5 ng/µl), 50 ng (1 ng/µl), 100 ng (2 ng/µl) and 250 ng (5 ng/µl) DNA.

The amplification of 25 ng DNA in a 50 µl reaction (0,5 ng/µl) was later performed using a PCR reaction with an initial denaturation of 95 °C for 5 minutes according to Qiagens instructions.

The PCR products were after amplifications separated with gel electrophoresis on a 2 % agarose gel run at 76 volts and visualized with Gel-red (Biotium Inc, Hayward, USA). The bands were compared to DNA ladder O´GeneRuler 100 bp (0,1 µg/µl, Thermo

(16)

PCR

Amplification of the genomic DNA was after optimization carried out in a 25 µl

reaction mixture containing 4,5 mM MgCl2 using the HotStarTaq plus DNA polymerase

kit (Qiagen Inc) according to the manufacturers’ instructions, adding 10 ng of genomic DNA per reaction (0,4 ng/µl).

The amplification took place in Veriti 96 well Thermal cycler (AB Applied Biosystems, Thermo Fisher Scientific, Waltham, USA) starting with initial denaturation at 95 °C for 5 minutes, followed by 40 cycles of denaturation at 94 °C for 30 seconds, annealing at 60 °C for 30 seconds, extension at 72 °C for 60 seconds and a final extension at 72 °C for 7 minutes and was then cooled down to 4 °C.

Roughly a fifth of the clinical samples were run with an initial denaturation at 95 °C for only 5 seconds during DNA amplification.

Optimization of the pyrosequencing

The amplicons controlled by gel electrophoresis were after PCR optimization run through the pyrosequencing protocol.

25 µl DNA amplicons were mixed in 1100 shakings/min for 10 minutes with 3 µl Streptavidin Sepharos solution (GE Healthcare, BioScience AB, Uppsala, Sweden), 12 µl MilliQ water and 40 µl binding buffer (PyroMark, Qiagen) per well. The amplicons were denatured and washed using Iggy vacuum prep tool (Qiagen Inc) using 70 % ethanol, 0,2 mol/L NaOH denaturation solution and 5,47 x 10-6 mol/L Tris washing buffer solution (pH 7,6) according to the manufacturers’ instructions. The amplicons were then transferred to a PSQ plate containing 40 µl of 0,4µM sequencing primer VDP2S (Biomers, Ulm, Germany) with the following sequence:

5´-AAAGCAAATTGCCTG-3´ per well, diluted in annealing buffer (PyroMark, Qiagen). The single stranded amplicons were incubated with the sequencing primer for two minutes in 80 °C.

The primer sequence was designed in PyroMark Assay design 2.0 (2009).

To estimate any background reactions and self-priming, 40 µl of 0,4µM sequencing primer (VDP2S) diluted in annealing buffer, 0,4µM reverse primer (VDP2R) diluted in

(17)

annealing buffer and 0,4µM sequencing primer (VDP2S) together with 0,4µM reverse primer (VDP2R) diluted in annealing buffer, were added to wells of the PSQ plate and was also incubated in 80 °C for two minutes. No clinical DNA was transferred to these wells.

The reverse strands of the amplicons were then sequenced in PyroMark Q96 ID (Qiagen Inc) operating a standard pyrosequencing protocol from Qiagen using PyroMark Gold Q96 reagents kit (Qiagen Inc) including nucleotides dATPαS, dCTP, dGTP and dTTP, an enzyme mix containing DNA-polymerase, ATP sulphurylase, luciferase and apyrase, and a substrate mix of Adenosine 5´phosphosulphate (APS) and luciferin. The

nucleotides were added to the pyrosequencing reaction in the following order:

AT/GGCCACACCCACA/T/GGAACTGGCAAA, where the grey marked area marks the SNPs in question. The sequence was analyzed by the PSQ software PyroMark Q96 2.5.8 (Qiagen Inc).

The marked SNP area for rs4588 (A/T/G) was later modified and the updated SNP marked areas were as follows: AT/GGCCACACCCAC/ATGGAACTGGCAAA. As a part of quality assurance the amplicons was also sequenced using the Sanger sequencing method. The reverse primer used for amplification before Sanger sequencing was VBP20R (Biomers, Ulm, Germany) with the following sequence: 5´-AAGATTCTGCCATGTTAAGTGGA-´3, lacking biotin.

Pyrosequencing

The amplicons of the clinical samples were prepared usingStreptavidin Sepharos solution (GE Healthcare, BioScience AB, Uppsala, Sweden), washed and denatured using Iggy vacuum prep tool (Qiagen Inc) according to the manufacturers’ instructions and sequenced in PyroMark Q96 ID (Qiagen Inc) using 0,4µM sequencing primer VDP2S (Biomers, Ulm, Germany) with the following sequence:

5-AAAGCAAATTGCCTG-3 and the PyroMark Gold Q96 reagents kit (Qiagen Inc). The nucleotides were added to the pyrosequencing reaction in the following order:

(18)

A positive control sample, heterozygote for both SNPs and a negative control containing no DNA was run with the clinical samples every runtime.

A second sequencing primer: 0,4µM VBP2S (Sigma-Aldrich, Stockholm, Sweden), with the same sequence as previously mentioned (5-AAAGCAAATTGCCTG-3), was later used for sequencing the VDBP polymorphisms.

Statistical analysis

Using the interpreted results from the analyzing software PyroMark Q96 2.5.8 (Qiagen Inc) the genotype frequencies for rs7041 and rs4588 was calculated in Microsoft Excel, 2011, version 14.6.0.

Allele frequencies for rs7041 and rs4588 were determined using the Hardy-Weinberg (H-W) equation under the assumption that the population was in genetic equilibrium. The equation for H-W equilibrium is as follows; p2 + 2pq + q2 =1 where the p and q represent the two alleles, p2 and q2 represent the homozygote genotypes and 2pq the heterozygote genotype. The sum of all frequencies must always equal 1 (21).

The hypothesis that the population in fact was in genetic equilibrium was analyzed using a Chi-squared test (X2 test) with a 5 % significance level (α=0,05) investigating the resemblance between the observed (O) and the expected (E) frequencies using the equation !! ₌ (!!!)!

! (22).

The 25(OH)D concentration correlation to rs7041 and rs4588 was investigated using Spearman’s test at a significance level of 1 % (α=0,01) in IBM SPSS statistics, 2013, version 22.0. Spearman’s test is a correlation test that is applied when one of the variables cannot be graded. In the equation: !! = 1 −! !!

! ! !!!

! (!!_!!),the calculated rs lands

between -1 and +1. If the calculated value is close to -1 there is a negative correlation and if the calculated value is close to +1 there is a positive correlation between the variables (22).

(19)

RESULT

Optimization

All PCR optimization results were based on visual assessment of the agarose gel. The PCR reaction showed the highest efficiency between annealing temperatures 59 °C and 62 °C where strong, clear, single bands of DNA at 314 nt fragments were visualized (figure 6), therefore an optimal annealing temperature of 60 °C was chosen.

The PCR amplification was most efficient at a salt concentration of 4,5 mM MgCl2

(figure 6).

The DNA amplification generated amplicons in all DNA concentrations tested from 25 ng/50 µl reaction (0,5 ng/µl) up to 250 ng/50 µl reaction (5 ng/µl), however a

concentration of 10 ng/25 µl reaction (0,4 ng/µl) was chosen for sequencing of the clinical samples.

The later amplification of DNA with a concentration of 0,5 ng/µl, at the initial

denaturation temperature 95 °C for the optimal 5 minutes, and annealing temperatures ranging from 56-68°C, was assessed the same way. The 2% agarose gel showed strong single bands at 314 nt for all temperatures with the exception of the bands of DNA amplified at 68 °C which showed weak bands compared to other annealing

(20)

Figure 6. Results from the optimization of the PCR reaction. Gel electrophoresis of the DNA, amplified in a gradient of MgCl2 concentrations and annealing temperatures, run

in a 2 % agarose gel at 76 volt. Optimization with 4,5 mM MgCl2 was later run with

better-quality results.

The results from the pyrosequencing optimization were assessed based on correct interpretation by the software PyroMark Q96 2.5.8 (Qiagen Inc) and checking of peak height and luminescent intensity. The DNA samples showed good peak heights (figure 7a) and no background peaks with the sequencing primer (figure 7b).

The amplicons from the PCR amplification with annealing temperature 68 °C generated very low peak heights in the sequencing reaction.

The biotin labeled reverse primer created a small number of low height peaks both in the wells containing only reverse primer and the wells containing reverse and

(21)

Figure 7a. Double heterozygote individual sequenced from DNA amplified with annealing temperature 60 °C, showing good peak heights and no background peaks.

Figure 7b. Sequencing reaction of 0,4 µM sequencing primer diluted in annealing buffer containing no clinical DNA (NTC), showing background peaks <2,5 relative light units (RLU).

(22)

Figure 7c. Reaction of 0,4µM of reverse and sequencing primer diluted in annealing buffer containing no clinical DNA. A few relatively low height peaks, with the highest amplitude of 7 relative light units (RLU), can be visualized.

Genotypes and allele frequencies

386 individuals were successfully genotyped for the rs7041 and rs4588 single

nucleotide polymorphisms. Three out of the previous 389 individuals were excluded, due to failed sequencing and they were not further used in this study.

136 (35 %) individuals of the test population were genotyped as G/G, 198 (51 %) as G/T and 52 (14 %) was genotyped as T/T for rs7041 (figure 8a). The observed allele frequencies were for the G allele 0,609 (61 %) and for the T allele 0,391 (39 %) (figure 8b).

Figure 8a and 8b. 8a shows the distribution of the genotypes T/T, G/G and G/T for the rs7041 alleles within the test population. 8b shows the distribution of allele T and G within the test population.

39% 61% rs7041 allele frequency

T

G

14% 35% 51% rs7041 genotypes

T/T

G/G

G/T

(23)

201 individuals (52 %) of the test population were genotyped as C/C, 157 (41 %) as A/C and 28 (7 %) were genotyped as A/A for rs4588 (figure 9a). The observed allele frequencies were for the C allele 0,724 (72 %) and for the A allele 0,276 (28 %) (figure 9b).

Figure 9a and 9b. 9a shows the distribution of the genotypes A/A, C/C and A/C for the rs4588 alleles within the test population. 9b shows the distribution of allele A and C within the test population.

The expected (E) frequencies of each genotype for both SNPs, calculated with the H-W equilibrium, were statistically compared to the observed number of individuals with respective genotypes using two separate Chi-squared tests with significance levels of α=0,05. The chi-squared tests showed calculated values of 2,24 for rs7041 and 0,098 for rs4588 (critical value 3,84 according to X2 table for df=1). The null hypothesis, saying there is no significant difference between the measured and the expected frequencies, could not be rejected in either case.

Genotypes and 25(OH)D concentrations

The distribution of all 25(OH)D concentrations of the Swedish test population (n=351) shows the highest frequencies around concentrations of 51-60 nmol/L (figure 10).

347 genotyped individuals with corresponding 25(OH)D levels were available for this study. The mean vitamin D concentration for the genotyped population (n=347) was 58,0 nmol/L (Std. Dev. 19,57). 68,3 % of the individuals had a 25(OH)D level between 38,4-77,6 nmol/L (mean ±Std.Dev.). 95,4 % of the individuals had a 25(OH)D level between 18,8-97,1 nmol/L (mean ± 2 x Std.Dev.).

7% 52% 41% rs4588 genotypes

A/A

C/C

A/C

28% 72% rs4588 allele frequency

A

C

(24)

Figure 10. Distribution of vitamin D concentrations in the Swedish population (n=351).

The mean 25(OH)D concentrations increased gradually with the presence of the

rs7041G allele (table 1) and the rs4588C allele (table 2) while the genotypes containing an increasing presence of the rs7041T allele and rs4588A allele showed lower 25(OH)D concentrations.

Table 1. Central tendencies and standard deviations of the 25(OH)D concentrations for rs7041 genotypes. Rs7041 Mean (nmol/L) N Std. Deviation Median G/G 63,6126 123 19,45430 60,6100 G/T 56,8055 177 19,74700 56,7300 T/T 47,7000 47 13,72279 47,0800 Total 57,9851 347 19,57491 56,9900

Table 2. Central tendencies and standard deviations of the 25(OH)D concentrations for rs4588 genotypes. Rs4588 Mean (nmol/L) N Std. Deviation Median C/C 62,5971 181 19,46390 60,7200 A/C 54,2062 141 19,29092 52,5500 A/A 45,9068 25 10,75764 47,1600 Total 57,9851 347 19,57491 56,9900 0 10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80 90 100 110 120 130 F re que nc ie s

(25)

After excluding the individuals with a 25(OH)D concentration >100 nmol/L, 334 genotyped individuals with corresponding 25(OH)D concentrations were available. A two tailed Spearman’s test at a significance level of 1 % (α=0,01) showed that both the rs7041 (rs = -0,251) and rs4588 (rs= -0,257) were weakly correlated to the 25(OH)D

concentration.

Monthly variation of the 25(OH)D concentration

After excluding the individuals with 25(OH)D concentrations above 100 nmol/L, the mean concentrations for the genotyped population (n=334) were calculated. The

25(OH)D was measured in February, March, June, September, October, November and December and showed a mean concentration of 56,1 nmol/L (Std. Dev.=17,31) over the year. The highest mean concentration, 68,2 nmol/L (Std. Dev. 14,22), was found in September and the lowest mean concentration, 50,6 nmol/L (Std.Dev. 16,11), was found in February/March.

The rs7041 genotype with the lowest vitamin D concentrations all over the year was rs7041T/T with mean values ranging from 46,2 nmol/L (Std. Dev. 9,63) in

February/March to 61,9 nmol/L (Std. Dev. 16,33) in September. Rs7041G/G had mean concentrations of 54,5 nmol/L (Std. Dev. 15,96) in February/March and 72,0 nmol/L (Std. Dev. 16,67) in September. Rs7041G/T measured mean concentrations 49,4 nmol/L (Std. Dev. 17,43) in February/March and 67,5 nmol/L (Std. Dev. 12,87) in September (figure 11).

Among the rs4588 genotypes the A/A genotyped individuals showed the lowest mean vitamin D concentrations at 47,2 nmol/L (Std. Dev. 10,43) in February/March. In September there were no individuals with genotype rs4588A/A measured for 25(OH)D and therefore there is no mean concentration of 25(OH)D for this genotype in

September. The rs4588C/C genotype had mean concentrations ranging from 53,1 nmol/L (Std. Dev. 16,02) in February/March to 73,1 nmol/L (Std. Dev. 15,79) in September while the rs4588A/C genotype had mean concentrations ranging from 48,4 nmol/L (Std. Dev. 17,03) in February/March to 65,3 nmol/L (Std. Dev. 12,99) in

(26)

Figure 11. Vitamin D concentrations spread over the year showing the difference in levels depending on the genotype of single nucleotide polymorphism rs7041. 1-6 on the X-axis equals the months February/March, June, September, October, November and December. The mean values are based on the 25(OH)D levels correlated to genotype where the 25(OH)D levels >100 nmol/L are excluded (n=334).

Figure 12. Vitamin D concentrations spread over the year showing the difference in levels depending on the genotype of single nucleotide polymorphism rs4588. 1-6 on the X-axis equals the months February/March, June, September, October, November and December. In September there were no available 25(OH)D concentrations of

rs4588A/A individuals. The mean values are based on the 25(OH)D levels correlated to genotype and where the 25(OH)D levels >100 are excluded (n=334).

30 35 40 45 50 55 60 65 70 75 0 1 2 3 4 5 6 7 25(OH )D c on ce n tration ( nm ol /L) Months rs7041G/G rs7041G/T rs7041T/T 30 35 40 45 50 55 60 65 70 75 0 1 2 3 4 5 6 7 25(OH )D c on ce n tration (n mol/L ) Months rs4588C/C rs4588A/C rs4588A/A

(27)

DISCUSSION

Optimization

The DNA concentration used for the PCR reactions of the Swedish population was 0,4 ng/µl despite the fact that the lowest optimized DNA concentration was 0,5 ng/µl. The concentrations did not differ enough to matter, however, a low amount of start template increases the specificity of the PCR reaction.

A preheating period, used to denature the amplicons and activate the Taq-polymerase during PCR, of only 5 seconds instead of 5 minutes, as Qiagens suggests, seems to also be functional for amplification. The gel bands of the amplicons from the 5-second preheating period were as strong as the 5-minute preheating period gel bands. It is likely that the 5-second preheating period lead to a gradual activation of the Taq polymerase and that amplification was unaffected.

During the pyrosequencing optimization, the marked nucleotide area for rs4588 was modified due to trouble interpreting the results. The failed results were caused by

interpretation of the wrong nucleotides. The A/T/G nucleotides that were first marked as the rs4588 SNP are the same nucleotides as the rs4588 alleles of the forward strand, it is however the reverse strand that is sequenced and therefore C/A/T are the correct

nucleotides marking the rs4588 SNP of the reverse strand. The modification did not alter the order of the nucleotides added to the pyrosequencing reaction, simply the analyzed sequence. No modifications of the rs7041 nucleotides were made, since the T/G alleles of rs7041 are the nucleotides of the reverse strand and were correctly analyzed from the start.

The reverse primer generated a small number of low height peaks (figure 7c), which were most likely caused by self-priming of the reverse primer and may be resolved by using a smaller amount of reverse primer in the PCR reaction. However, due to the low height of the peaks this was not deemed necessary.

(28)

Genotypes and allele frequencies

Heterozygosity (G/T) for rs7041 and homozygosity for the C allele of rs4588 were found to be the most common genotypes of the Swedish population.As expected the most common rs7041 allele in the Swedish population was the G allele, included in the coding sequence, GAG, for glutamic acid promoting the GC-1s phenotype (rs7041G, rs4588C) of the vitamin D binding protein. The frequencies of the G and T allele were consistent with the distribution in Europe (compare figure 3b and 8b).

The most common rs4588 allele in the Swedish population was the C allele, also

expected when comparing to the European distribution (compare figure 4b and 9b). The C allele, included in the coding sequence ACG, for the amino acid threonine promotes both the GC-1s (rs7041G, rs4588C) and GC-1f (rs7041T, rs4588C) phenotypes of the vitamin D binding protein.

Critical value of the chi test

The critical value of the chi-square test can be found in any X2 test table with a set significance level and “degrees of freedom” (df), which in this case is 1 since there are two alleles. A calculated value close to 0 is a sign of close resemblance to the estimated frequency.

Neither of the rs7041 (x2= 2,24) or rs4588 (x2= 0,098) genotypes showed any

significant difference between the observed frequencies and the expected frequencies calculated with the H-W equation. This implies that the observed frequencies from the test population are representative of a genetically balanced population.

Genotypes and 25(OH)D concentrations

The rs7041 (rs= -0,251) and rs4588 (rs = -0,257)both showed weak, negative correlation

to the 25(OH)D concentration. Homozygous individuals for the rs7041T allele and the rs4588A allele showed lower 25(OH)D concentrations compared to other genotypes. Findings of these alleles during sequencing may therefore contribute to the

interpretation of the 25(OH)D level. Homozygous individuals for the rs7041G allele and the rs4588C allele on the other hand, seems to be associated with higher 25(OH)D concentrations.

(29)

Double heterozygote individuals (rs7041G/T, rs4588A/C) technically have the genetic ability to create four different haplotypes of the VDBP, however, one individual does not produce four different proteins. Unfortunately this study showed a great number of double heterozygote individuals, 51 % for rs7041 and 41% for rs4588, making it difficult to determine the association between the GC-haplotype and the 25(OH)D concentration.

Monthly variation of the 25(OH)D concentration

The fact that the highest 25(OH)D concentrations were found in September and the lowest in February/March is consistent with previous findings (2) and is almost

certainly caused by the lack of sufficient sunlight during the winter months in Sweden. During wintertime however, the concentration gap between the genotypes decreased and the 25(OH)D concentration seems to depend less on genotype than in summer time. In conclusion this makes it difficult to use genotyping as a diagnostic tool for vitamin D deficiency during wintertime in Sweden.

Worth noticing is that all mean concentrations, even during summer time, are below the cut off value (75 nmol/L) used at Örebro University hospital, indicating that a large portion of the test subjects had 25(OH)D concentrations below 75 nmol/L. These results might be caused by the high mean age of the test population.

Error sources

A sequencing primer (VDP2S) from Sigma-Aldrich was used to sequence the DNA fragments without first being optimized. This was not evaluated to be necessary since the same primer sequence was used to successfully sequence the DNA fragment in question earlier, only with a VDP2S sequencing primer from a different company. The result from the pyrosequencing reactions does not indicate any difference in the

function between these primers.

Low amplitudes occurred during pyrosequencing even after PCR optimization. This can’t be explained by insufficient amounts of or mistreated pyrosequencing reagents but

(30)

the use of the vacuum prep tool, transferring the DNA from the PCR plate to the PSQ plate.

Strengths and weaknesses

Real-time PCR using Taqman probes could be used for detection of rs7041 and rs4588 alleles without sequencing (16).However, this would demand the use of two allele-specific probes for each SNP, making pyrosequencing the more cost effective option. Using real-time PCR, the unusual T allele of rs4588 would not be detected unless a third probe for that allele was used. The benefit of using pyrosequencing is that both SNPs can be analyzed using the same sequencing primer, saving both time and cost. Using pyrosequencing in combination with a protein analysis would make it possible to correlate the 25(OH)D level to the VDBP haplotype, and not only the genotype. By using only pyrosequencing the understanding of the role of the vitamin D binding protein concentration and its affinity is lost.

Conclusion

The prevalence of the rs7041 alleles were 61 % G and 39 % T and of the rs4588 alleles 72 % C and 28 % A in the Swedish test population, showing distributions similar to the European population. The vitamin D concentrations were lowest in February/March and highest in September, as expected due to lack of sunlight during wintertime. The rs7041 and rs4588 genotypes were weakly correlated to the 25(OH)D concentration and the rs7041T allele and rs4588A allele seems to contribute to lower 25(OH)D

concentrations. However, during wintertime the difference in 25(OH)D mean

concentrations between the genotypes seem to decrease, making the genotyping method difficult to use for evaluation of the vitamin D status at that time of year.

(31)

REFERENCES

1. Nilsson-Ehle P, Berggren Söderlund M, Theodorsson E (editors). Laurells Klinisk kemi. 9th ed. Lund: Studentlitteratur AB; 2012.

2. Rodwell VW, Bender DA, Botham KM, Kennelly PJ, Well PA. Harper´s illustrated biochemistry. 30th ed. New York: McGraw Hill Education; 2015. 3. Molina. PE. Endocrine Physiology. 4th ed. New York: McGraw-Hill Companies

Inc; 2013.

4. Kennel Ka, Drake MT, Hurley DL. Vitamin D deficiency in adults: when to test and how to treat. Mayo Clin Proc. 2010 Aug; 85(8): 752-758.

5. Yousefzadeh P, Shapes S, Wang X. Vitamin D binding protein impact on 25-hydroxyvitamin D levels under different physiologic and pathologic conditions. International Journal of endocrinology. 2014; 2014 (2014): 6 pages.

6. Gozdzik A, Zhu J, Wong B, Fu L, Cole D, Parra E. Association of vitamin D binding protein (VDBP) polymorphisms and serum 25(OH) concentrations in a sample of young Canadian adults of different ancestry. J Steroid Biochem Mol Biol. 2011 Nov; 127 (3-5): 405-12.

7. Lips P. The relative value of 25(OH)D and 1,25(OH)2D measurements.

Presented at European society of endocrinology (ECE) 2009, Istanbul, Turkey. Endocrine abstracts (2009) 20 ME11.

8. Kasper D, Fauci A, Hauser S, Longo D, Jameson JL, Lascalzo J. Harrison´s principles of internal medicine. 19th ed. New York: McGraw-Hill education; 2015.

9. McKenna MJ, Murray BF. Vitamin D dose response is underestimated by Endocrine Society´s Clinical Practice Guideline. Endocr Connect. 2013 Jan; 2(2): 87-95.

10. Uniprot [Internet]. Washington DC: PIR; Hinxton: EMBL-EBI and Geneva: SIB. c2002 - [cited 2016 Apr 2]. Available from:

http://www.uniprot.org/uniprot/P02774

11. National center for biotechnology information (NCBI) [Internet]. Bethestda, MD: US national library of medicine; -. GC group-specific component (vitamin

(32)

12. Carpenter T, Zhang J, Parra E, Ellis B, Simpson C, Lee W, Balko J et al. Vitamin D binding protein is a key determinant of 25-hydroxyvitamin D levels in infants and toddlers. J Bone Miner Res. 2013 Jan; 28(1): 213-221.

13. Ensambl [Internet]. Hinxton: Wellcome Trust Sanger Institute

(WTSI)/European bioinformatics institute (MBL-EBI); c2000. rs7041; population genetics [cited 2016 Apr 2]. Available from:

http://www.ensembl.org/Homo_sapiens/Variation/Population?db=core;r=4:717 52117-71753117;v=rs7041;vdb=variation;vf=6710

14. Ensambl [Internet]. Hinxton: Wellcome Trust Sanger Institute

(WTSI)/European bioinformatics institute (MBL-EBI); c2000. rs4588; population genetics [cited 2016 Apr 2]. Available from:

http://www.ensembl.org/Homo_sapiens/Variation/Population?r=4:71752106-71753106;v=rs4588;vdb=variation;vf=4371

15. National center for biotechnology information (NCBI), dbSNP short genetic variations [Internet]. Bethestda, MD: US national library of medicine; -. rs4588 [cited 2016 Apr 2]. Available from:

https://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs=4588

16. Buckingham L. Molecular diagnostics. 2:nd ed. Philadelphia: F.A. Davis Company; 2012.

17. Hochstein N, Honsel D, Kappmeier C, Rütjes T, Andreou I, Kreutz M et.al. Pyrosequencing and its applications. QIAGEN GmbH, QIAGEN strasse 1, 40724 Hilden, Germany. 2010.

18. Royo JL, Hidalgo M, Ruiz A. Pyrosequencing protocol using a universal biotinylated primer for mutation detection and SNP genotyping. Nat Protoc. 2007; 2 (7): 1734-9.

19. Ronaghi M, Karamohamed S, Pettersson B, Uhlén M, Nyrén P. Real-time DNA sequencing using detection of pyrophosphate release. Anal biochem. 1996; 242 (1): 84-89.

20. Olsson L. Subjective well-being in old age and its association with biochemical and genetic biomarkers and with physical activity [doctorial dissertation]. Örebro: Örebro university; 2015.

(33)

21. Strachan T, Read A. Human molecular genetics. 4th ed. New York: Garland Science, Taylor & Francis Group, LCC; 2011.

22. Ejlertsson G. Statistik för hälsovetenskaperna. 2nd ed. Lund: Studentlitteratur AB; 2012