Article | Open

Circulating vitamin D binding protein, total, free and bioavailable 25-hydroxyvitamin D and risk of colorectal cancer

  • Scientific Reports 5, Article number: 7956 (2015)
  • doi:10.1038/srep07956
  • Download Citation
Received:
Accepted:
Published online:

Abstract

Epidemiological investigation have suggested that there is a significantly inverse association between circulating 25-hydroxyvitamin D (25(OH)D) and the risk for developing colorectal cancer (CRC) in humans. However, little is known about the role of vitamin D binding protein (VDBP) in colorectal carcinogenesis. Blood samples were collected from 212 CRC patients and 212 controls matched with age, gender and blood collection time. We used logistic regression to calculate the odds ratios and 95% confidence intervals for further estimation of the association of the quartiles of VDBP, total, free and bioavailable 25(OH)D with CRC risk. The results revealed that there was no significant association between circulating VDBP concentrations and CRC in the present study, and that a negative association existed between total 25(OH)D and the risk of CRC, which was unchanged after adjustment for VDBP. Higher levels of free and bioavailable 25(OH)D were significantly associated with decreased risk of CRC. After stratifying by VDBP, high levels of total, free and bioavailable 25(OH)D were associated significantly with decreased CRC risk among participants with circulating VDBP below the median. These findings indicate that VDBP is not directly associated with the risk of CRC, but it modulates circulating free and bioavailable 25(OH)D concentration.

Introduction

Colorectal cancer (CRC) is the third most prevalent cancer and the fourth most common cause of cancer mortality worldwide1. Accumulating epidemiological studies suggest that high vitamin D intake may influence the risk of CRC2,3, which is an interesting cancer prevention strategy that has raised public concerns globally. Although vitamin D can be obtained from diet and dietary supplements, most vitamin D is synthesized by human skin exposed to the sunlight. Due to reduced outdoor activity, the absence of adequate sun exposure and limited resource of vitamin D supplements from diet, vitamin D deficiency is a common phenomenon both in healthy individuals and in CRC cases4.

25-hydroxyvitamin D (25(OH)D), the precursor of the active form of vitamin D, is recognized as the optimal indicator of vitamin D metabolic status with a relatively long half-life and high concentration in plasma5. 25(OH)D is commonly bound to vitamin D binding protein (VDBP) and albumin(Alb), which can be converted into hormonally active 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) in the kidney, colon and several other tissues6,7,8,9. The majority of 25(OH)D and 1,25(OH)2D3 are primarily bound to VDBP, approximately 10–15% to Alb, free 25(OH)D and 1,25(OH)2D3 only account for less than 1%10. Since the affinity of Alb to 25(OH) D or 1,25(OH)2D3 is weaker than that of VDBP, the loosely binding fraction and the free fraction consist of bioavailable 25(OH)D11. Vitamin D receptor (VDR), a key nuclear receptor, can only be activated by the free form of 1,25(OH)2D3, and regulates the transcription and expression of numerous vitamin D targeted genes that are related to cell proliferation, differentiation, invasion and angiogenesis.

VDBP, which is synthesized and secreted primarily by the liver, includes three functional domains. Although 25(OH)D and 1,25(OH)2D3 are mainly bound to VDBP domain I, circulating VDBP and Alb levels considerably exceed the concentrations of 25 (OH)D and 1,25(OH)2D312. VDBP also plays an important role in the elimination of dead or injured cell, macrophage activation and neutrophil chemotaxis in inflammatory condition12. As inflammation is deemed to be an important contributor to carcinogenesis, while inflammatory-cells such as macrophage and neutrophil are the main components to defense against cancer cell13, we thus speculate that VDBP may be involved in carcinogenesis and cancer progression.

Recently, several studies have reported the relationship among VDBP, 25(OH)D and the risk of cancer. Two studies showed that circulating VDBP was inversely associated with the risk of pancreatic cancer and renal cell carcinoma14,15. Another study reported no association between VDBP and bladder cancer16. In addition, study by Stephanie et al17 indicated that VDBP might modulate the impact of vitamin D status on prostate cancer. However, there is no study reporting the influence of VDBP on the association between circulating total, free, and bioavailable 25(OH)D and the risk of CRC. Therefore, in this study, we measured the circulating VDBP, total, free and bioavailable 25(OH)D in 212 CRC patients and 212 well-matched healthy controls and further explored the relationship between total, free, bioavailable 25(OH)D and VDBP, and the risk of CRC.

Methods

All subjects in our study were from a health assessment cohort population in Nanjing First Hospital (Nanjing, Jiangsu, China). The health assessment cohort population comprised 25616 routine physical examination of individuals who performed annually in Nanjing First Hospital from 2010 to 2012. All participants were healthy, cancer-free individuals, didn′t show clinical CRC symptoms, rectal digital examination and fecal occult blood test were normal and tumor protein biomarkers (plasma CEA, CA199, CA50 and CA242) were lower than the reference values when they were enrolled in the cohort, with an average follow-up time of 3.4 years until diagnosis. The case group consisted of 212 CRC patients, including 115 colon and 97 rectal cancer patients, and ranging in age from 37 to 83 years. All CRC patients were all initially diagnosed by colonoscopy and CT detections, and confirmed with histological evidence from biopsies performed from 2013 to 2014 in Nanjing First Hospital. Patients with acute or chronic renal and liver failures, nutritional problem (malnutrition, anemia, hypoproteinemia, anorexia and inappetence), active infection, or without histological confirmation were excluded from the present study. The controls were composed of 212 blood collections from healthy individuals with matched gender, age (±5 year) and blood collection time. Due to unwillingness of the controls and its invasiveness, colonoscopy was not used to exclude the CRC in the controls. However, all the controls were clinical symptom-free of CRC, rectal digital examination, fecal occult blood test and tumor protein biomarkers (plasma CEA, CA199, CA50 and CA242) were normal. All participants enrolled in our study did not use vitamin D or calcium supplement and drugs that might influence circulating calcium, phosphorus and 25(OH)D concentrations in recent six months. Relative clinical data were obtained from the hospital medical record. Blood samples of enrolled subjects in our study were collected during the first time health check-up in Nanjing First Hospital between November and February from 2010 to 2012. Samples were drawn between 8:00 and 10:00 each morning, and plasma samples were stored at -80 °C until measurement. This study was approved by the Institution Ethics Commission of Southeast University, written informed consents were obtained from all participants and the methods were carried out in accordance with the approved guidelines.

Plasma total 25(OH)D concentration were measured by direct competitive enzyme-linked immunosorbent assay using 25(OH) D immunoassay kit (HCB Ltd, Vancouver, Canada) with the coefficient of variations of inter- and intra-batch to be 9.0% and 9.0%, respectively. Measurement of circulating plasma VDBP concentration was carried out using human vitamin D binding protein immunoassay kit (R&D Systems, Minneapolis, USA) according to the manufacturer's protocol. The inter- and intra-batch of the kit were 10.8% and 15.2%, respectively. Blinded quality controlled samples were included in each batch, and all the measurements were conducted in triplicate in all plasma samples. Plasma Alb concentration was measured by Beckman automatic biochemical analyzer (AU680) (Beckman Coulter lnc, Tokyo, Japan) using a bromocresol green dye assay. Applying the mathematical equations provided by Bhan et al18, free and bioavailable 25(OH)D were calculated based on the concentrations of plasma Alb, total 25(OH) D and VDBP. All measurements were carried out in April of 2014.

Descriptive statistics for baseline characteristics were calculated with the median and quartile for continuous variables or frequencies for categorical variables. Quartile cut-points for VDBP, total, free and bioavailable 25(OH)D were determined based on the distribution among controls. The trend across categories by modeling the median of each category as a continuous variable and Wald test was used to evaluate its statistical significance. Conditional logistic regression was used to estimate the odds ratio(OR) and 95% confidence interval (CI) of CRC by the quartiles of VDBP, total, free and bioavailable 25(OH)D. Factors which may influence plasma 25(OH)D, VDBP, or the risk of CRC were included in the multivariable logistic regression models. Therefore, multivariable model was adjusted for the matching factors (sex and age), body mass index (BMI), and the status of smoking, drinking, diabetes and hypertension. Moreover, the results were mutually adjusted for 25(OH)D or VDBP. We also analyzed the stratifying total, free and bioavailable 25(OH)D by VDBP (<median vs. ≥median) and VDBP by 25(OH)D (<median vs. ≥median) using unconditional logistic regression. Likehood ratio test was applied to assess statistical interaction in the multivariable-adjusted model. All these statistical calculations were performed using SPSS 17.0 software (SPSS Inc, Chicago, USA) and a p-value < 0.05 was recognized as statistically significance in all calculations.

Results

A total of 424 individuals including 212 newly diagnosed CRC patients and 212 healthy participants matched with gender and age (±5 year) as well as blood collection time were enrolled in this study. Baseline characteristics of cases and controls were described in Table 1. No significant difference was found in BMI value, frequencies of smoking, drinking, diabetes and hypertension as well as concentration of circulating VDBP when compared with two groups. However, the plasma total, free and bioavailable 25(OH)D in CRC cases were significantly lower in comparison with the controls (p-value < 0.001, = 0.018 and 0.001).

Table 1: Baseline characteristics of colorectal cancer patients and controlsa

The results of associations among plasma VDBP, total, free and bioavailable 25(OH) D and the risk of CRC were listed in Table 2. Plasma VDBP concentration was not observed to be directly associated with the risk of CRC, further adjustment for 25(OH)D did not alter the quartile risk estimates (p-trend = 0.944, OR = 0.93, 95%CI = 0.51–1.69 for Q4:Q1). Adjustment for VDBP did not alter the inverse association between plasma total 25(OH)D and the risk of CRC (p-trend = 0.027, OR = 0.53, 95%CI = 0.29–0.98 for Q4:Q1). Moreover, negative associations were also observed between free 25(OH)D (p-trend < 0.001, OR = 0.42, 95%CI = 0.22–0.82 for Q4: Q1), bioavailable 25(OH)D (p-trend = 0.006, OR = 0.29, 95%CI = 0.15–0.56 for Q4: Q1) and the risk of CRC. No significant risk association was found between VDBP and CRC stratifying by 25(OH)D (p-interaction = 0.280, p-trend = 0.490, OR = 0.72, 95%CI = 0.33–1.57). On the contrary, plasma VDBP modified the association between total, free and bioavailable 25(OH)D and CRC, with lower risk for elevated total, free and bioavailable 25(OH)D appeared to be restricted to individuals with lower concentrations of VDBP (total 25(OH)D Q4:Q1 p-interaction = 0.001, p-trend < 0.001, OR = 0.14, 95%CI = 0.05–0.41; free 25(OH)D Q4:Q1 p-interaction = 0.001, p-trend < 0.001, OR = 0.10, 95%CI = 0.04–0.22; bioavailable 25(OH)D Q4:Q1 p-interaction = 0.001, p-trend < 0.001, OR = 0.03, 95%CI = 0.01–0.08) (Table 3).

Table 2: Odds ratios and 95% confidence interval for the association of plasma VDBP, total, free and bioavailable 25(OH)D with the risk of colorectal cancer
Table 3: Odds ratios and 95% confidence interval for the association between plasma VDBP stratifying by total 25 (OH)D or total, free and bioavailable 25 (OH)D stratifying by VDBP and the risk of colorectal cancer

Discussion

In this study, we found strong inverse associations between total, free and bioavailable 25(OH)D and the risk of CRC. However, we did not find a direct association between circulating VDBP and the risk of CRC. Interestingly, higher circulating total, free and bioavailable 25(OH)D concentrations were inversely associated with CRC only when circulating VDBP below the median. Our results suggest that higher VDBP level may sequester more 25(OH)D leading to less free and bioavailable 25(OH)D, and free and bioavailable 25(OH)D may be more biologically relevant to the risk of CRC in comparison with total 25(OH)D.

VDBP, a glycosylated-globulin, plays an important role not only in transporting vitamin D and circulating metabolites, but also in other biological functions to affect carcinogenesis. It is a member of the circulating actin scavenger system that prevents formation of F-actin networks and harmful effects following from cell or tissue damage12. VDBP is involved in macrophage activation by converting as group-specific component protein derived macrophage activating factor (GcMAF) and neutrophil chemotaxis by enhancing the chemotactic effect of complement-derived peptides19. Several studies have investigated the association between plasma VDBP and cancer risk, including bladder, prostate and colorectal cancers16,17,20,21. A recent study conducted by Anic et al20 reported that circulating VDBP was not associated with CRC risk, but found positive associations between 25(OH)D, the 25(OH)D:DBP molar ratio and CRC among men with higher VDBP concentration. Another study reported by Weinstein et al21 indicated that circulating 25(OH)D was inversely associated with CRC and DBP did not modify the association between 25(OH)D and CRC. Consistent with these studies, our results did not conclude that VDBP was associated with CRC, but low levels of total, free and bioavailable 25(OH)D were associated with elevated CRC risk, whereas the negative associations between total, free, bioavailable 25(OH)D and CRC were found only when individual with lower level of VDBP (<median). These findings indicate that VDBP is not involved in colorectal carcinogenesis through such mechanisms, but support the “free hormone hypothesis” that lower concentration of VDBP may bind less 25(OH)D contributing to higher level of circulating free and bioavailable 25(OH)D, while free and bioavailable 25(OH)D have more biological effects on colorectal tissue. It has been reported that 1,25(OH)2D3 and its analogues could inhibit growth and promote apoptosis, proliferation and differentiation of colon cancer cell in a dose-dependent manner22,23. 25(OH)D can be transformed into 1,25(OH)2D3 in colon tissue, and bioavailable 25(OH)D may be taken up into colorectal cells via diffusion across cell membranes, or via endocytosis of bound to VDBP by interaction of VDBP with megalin and cubulin24,25. The metabolite of 1,25(OH)2D3 binds to VDR is expressed in colorectal cell and forms a 1,25(OH)2D3-VDR complex to regulate a series of genes that are involved in colorectal carcinogenesis26,27. Moreover, VDBP may only have its canonical role in transporting 25(OH)D, not directly relating to colorectal carcinogenesis. Furthermore, previous studies have indicated that VDBP was markedly deglycosylated in CRC patients28. A higher level of deglycosylated VDBP results in a lower level of GcMAF. Consequently, eliminated GcMAF cannot perform its biological activities to macrophages and phagocytic activation, anti-angiogenesis effect and antitumor activity29,30,31,32,33. In addition, more than 120 single nucleotide polymorphisms were identified in VDBP gene, rs7041 and rs4588 has been demonstrated to influence the affinity of VDBP for VD and were associated with the concentration of 25(OH)D, VDBP who carried allele C of rs4588 has been illustrated to be less able to convert as GcMAF34,35. Therefore, genetic variation of VDBP gene may affect its secondary construction and function, eliminating its functions of antitumorigenesis.

To our knowledge, it is the first investigation studying the association of plasma VDBP, free and bioavailable 25(OH)D with the risk of CRC. Our prospective measurement of circulating VDBP, Alb, 25(OH)D to assess the association of VDBP, total, free and bioavailable 25(OH)D with CRC risk is a strength of our study. Moreover, all plasma samples were collected between 8:00 and 10:00 each morning in winter season in Nanjing first hospital, which could maximize reduce the influence from sunlight, intensity of ultraviolet ray, and diurnal and postprandial fluctuations. However, several limitations of this study should be addressed. Due to limited available plasma samples, small sample size of our study could not represented the whole population and would reduce statistical power in our study. Although VDBP was supposed to be stable in adulthood36, each plasma sample of the individual was only collected and measured at one point, which could not reflect the comprehensive status of total circulating vitamin D and DBP in all participants. Furthermore, we did not obtain detailed information concerning multiple potential confounding factors of each participant, such as physical activity, dietary vitamin intake and exposure to sunshine.

In summary, although we did not observe a direct association between VDBP and the risk of CRC, high levels of circulating total, free, bioavailable 25(OH)D were associated with a decreased risk of CRC, particularly when VDBP concentration was below the median. These findings suggest that VDBP is just a VD transporter which may not affect colorectal carcinogenesis, but can modulate the impact of vitamin D status on CRC. Thus the simultaneous detection and investigation of plasma total, free, bioavailable 25(OH)D and VDBP may be important for studying the relationship between vitamin D and the risk of CRC.

References

  1. 1.

    et al. Plasma vitamin D metabolites and risk of colorectal cancer in women. Cancer Epidemiol Biomarkers Prev 13, 1502–1508 (2004).

  2. 2.

    et al. Association between vitamin D and risk of colorectal cancer: a systematic review of prospective studies. J Clin Oncol 29, 3775–3782 (2011).

  3. 3.

    et al. Calcium and vitamin D intake and risk of colorectal cancer: the Multiethnic Cohort Study. Am J Epidemiol 165, 784–793 (2007).

  4. 4.

    & Vitamin D deficiency: a worldwide problem with health consequences. Am J Clin Nutr 87, 1080S–1086S (2008).

  5. 5.

    et al. A nested case control study of plasma 25-hydroxyvitamin D concentrations and risk of colorectal cancer. J Natl Cancer Inst 99, 1120–112 (2007).

  6. 6.

    , & Vitamin D signalling pathways in cancer: potential for anticancer therapeutics. Nat Rev Cancer 7, 684–700 (2007).

  7. 7.

    , , & Vitamin D metabolism in human colon adenocarcinoma-derived Caco-2 cells: expression of 25-hydroxyvitamin D3-1alpha-hydroxylase activity and regulation of side-chain metabolism. J Steroid Biochem Mol Biol 62, 21–28 (1997).

  8. 8.

    , , , & 25-hydroxy-vitamin d metabolism in human colon cancer cells during tumor progression. Biochem Biophys Res Commun 285, 1012–1017 (2001).

  9. 9.

    et al. 25-hydroxyvitamin D-1alpha-hydroxylase in normal and malignant colon tissue. Lancet 357, 1673–1674 (2001).

  10. 10.

    et al. Assessment of the free fraction of 25-hydroxyvitamin D in serum and its regulation by albumin and the vitamin D-binding protein. J Clin Endocrinol Metab 63, 954–959 (1986).

  11. 11.

    & Bioavailable vitamin D in chronic kidney disease. Kidney Int 82, 5–7 (2012).

  12. 12.

    , , & The vitamin D axis in the lung: a key role for vitamin D-binding protein. Thorax 65, 456–462 (2010).

  13. 13.

    , & Immunity, inflammation, and cancer. Cell 140, 883–899 (2010).

  14. 14.

    et al. Impact of circulating vitamin D binding protein levels on the association between 25-hydroxyvitamin D and pancreatic cancer risk: a nested case-control study. Cancer Res 72, 1190–1198 (2012).

  15. 15.

    , , , & Vitamin D-binding protein, circulating vitamin D and risk of renal cell carcinoma. Int J Cancer 134, 2699–2706 (2014).

  16. 16.

    , , & Influence of vitamin D binding protein on the association between circulating vitamin D and risk of bladder cancer. Br J Cancer 107, 1589–1594 (2012).

  17. 17.

    et al. Circulating 25-hydroxyvitamin D, vitamin D-binding protein and risk of prostate cancer. Int J Cancer 132, 2940–2947 (2013).

  18. 18.

    et al. Bioavailable vitamin D is more tightly linked to mineral metabolism than total vitamin D in incident hemodialysis patients. Kidney Int 82, 84–89 (2012).

  19. 19.

    & Identification of a region in the vitamin D-binding protein that mediates its C5a chemotactic cofactor function. J Biol Chem 279, 53282–53287 (2004).

  20. 20.

    , , , & Serum vitamin D, vitamin D binding protein, and risk of colorectal cancer. Plos One 9, e102966 (2014).

  21. 21.

    et al. Serum 25-hydroxyvitamin D, vitamin D binding protein and risk of colorectal cancer in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial. Int J cancer, 10.1002/ijc.29157 (2014).

  22. 22.

    , , , & Apoptosis is induced by the active metabolite of vitamin D3 and its analogue EB1089 in colorectal adenoma and carcinoma cells: possible implications for prevention and therapy. Cancer Res 60, 2304–2312 (2000).

  23. 23.

    & Regulation of vitamin D receptor abundance and responsiveness during differentiation of HT-29 human colon cancer cells. Endocrinology 132, 1808–1814 (1993).

  24. 24.

    et al. An endocytic pathway essential for renal uptake and activation of the steroid 25-(OH) vitamin D3. Cell 96, 507–515 (1999).

  25. 25.

    et al. Cubilin dysfunction causes abnormal metabolism of the steroid hormone 25(OH) vitamin D(3). Proc Natl Acad Sci U S A 98, 13895–13900 (2001).

  26. 26.

    et al. 25-Hydroxyvitamin D(3)-1alpha-hydroxylase and vitamin D receptor gene expression in human colonic mucosa is elevated during early cancerogenesis. Steroids 66, 287–292 (2001).

  27. 27.

    et al. Regulation of CYP24 splicing by 1,25-dihydroxyvitamin D(3) in human colon cancer cells. J Endocrinol 212, 207–215 (2012).

  28. 28.

    , & Glycosylation status of vitamin D binding protein in cancer patients. Protein Sci 18, 2036–2042 (2009).

  29. 29.

    , , & Immunotherapy of metastatic colorectal cancer with vitamin D-binding protein-derived macrophage-activating factor, GcMAF. Cancer Immunol Immunother 57, 1007–1016 (2008).

  30. 30.

    , , , & Association of the macrophage activating factor (MAF) precursor activity with polymorphism in vitamin D-binding protein. Anticancer Res 24, 3361–3366 (2004).

  31. 31.

    , , , & Effects of vitamin D(3)-binding protein-derived macrophage activating factor (GcMAF) on angiogenesis. J Natl Cancer Inst 94, 1311–1319 (2002).

  32. 32.

    , & Antitumor effect of vitamin D-binding protein-derived macrophage activating factor on Ehrlich ascites tumor-bearing mice. Proc Soc Exp Biol Med 220, 20–26 (1999).

  33. 33.

    et al. Vitamin D binding protein-macrophage activating factor inhibits HCC in SCID mice. J Surg Res 172, 116–122 (2012).

  34. 34.

    , , & Biological and clinical aspects of the vitamin D binding protein (Gc-globulin) and its polymorphism. Clin Chim Acta 372, 33–42 (2006).

  35. 35.

    & Vitamin D3 binding protein (group-specific component) is a precursor for the macrophage-activating signal factor from lysophosphatidylcholine-treated lymphocytes. Proc Natl Acad Sci U S A 88, 8539–8543 (1991).

  36. 36.

    , , & Reproducibility of serum 25-hydroxyvitamin d and vitamin D-binding protein levels over time in a prospective cohort study of black and white adults. Am J Epidemiol 176, 615–621 (2012).

Download references

Acknowledgements

This study was supported by National Natural Science Foundation of China (no. 81172141), Nanjing Science and Technology Committee project (no. 201108025), Nanjing Medical Technology Development Project (no. ZKX11025), Nanjing Health Young Talent Project, Jiangsu Provincial Key Medical Talents to S.K.W., Nanjing Medical Science and Technique Development Foundation to Y.Q.P. (no. QRX11255) and B.S.H. (no. QRX11254).

Author information

Affiliations

  1. Medical college, Southeast University, Nanjing 210009, Jiangsu, China

    • Hou-Qun Ying
    •  & Shu-Kui Wang
  2. Nanjing First Hospital, Nanjing Medical University, Nanjing 210006, Jiangsu, China

    • Hou-Qun Ying
    • , Hui-Ling Sun
    • , Bang-Shun He
    • , Yu-Qin Pan
    • , Feng Wang
    • , Qi-Wen Deng
    • , Jie Chen
    • , Xian Liu
    •  & Shu-Kui Wang
  3. College of Life Sciences, Nanjing Normal University, Nanjing 210006, Jiangsu, China

    • Hui-Ling Sun
    •  & Jie Chen

Authors

  1. Search for Hou-Qun Ying in:

  2. Search for Hui-Ling Sun in:

  3. Search for Bang-Shun He in:

  4. Search for Yu-Qin Pan in:

  5. Search for Feng Wang in:

  6. Search for Qi-Wen Deng in:

  7. Search for Jie Chen in:

  8. Search for Xian Liu in:

  9. Search for Shu-Kui Wang in:

Contributions

H.Q.Y. performed the laboratory detection, wrote the manuscript. H.L.S. performed the laboratory detection. B.S.H. and Y.Q.P. collected the blood sample and clinical data. F.W. and Q.W.D. conducted the statistics. J.C. and X.L. prepared Table 13. S.K.W. designed the study and reviewed and approved the manuscript.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Shu-Kui Wang.

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Creative Commons BY-NC-NDThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder in order to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/