Vitamin D status is not associated with inflammatory cytokine levels during experimental human endotoxaemia

Abstract

Vitamin D has been shown to modulate innate immune responses in vitro and ex vivo; however, human in-vivo data are lacking. At high latitudes, seasonal vitamin D deficiency is common due to alternating ultraviolet (UV)-B radiation exposure. In the present study, we investigated whether levels of 25 hydroxyvitamin D₃ [25(OH)D₃] and its active metabolite 1,25 dihydroxyvitamin D₃ [1,25(OH)₂D₃] are subject to seasonal variation and whether plasma levels of these vitamin D metabolites correlate with the in-vivo cytokine response during experimental human endotoxaemia [administration of lipopolysaccharide (LPS) in healthy volunteers]. Plasma levels of 25(OH)D₃ and 1,25(OH)₂D₃ were determined in samples obtained just prior to administration of an intravenous bolus of 2 ng/kg LPS (derived from Escherichia coli O:113) in 112 healthy male volunteers. In the same subjects, plasma levels of the inflammatory cytokines tumour necrosis factor (TNF)-α, interleukin (IL)-6 and IL-10 were analysed serially after endotoxin administration. Plasma levels of 1,25(OH)₂D₃, but not 25(OH)D₃, were subject to significant seasonal variation, with lower levels in autumn and winter. 25(OH)D₃ and 1,25(OH)₂D₃ levels did not correlate with plasma cytokine responses. Furthermore, 25(OH)D₃ deficient subjects (< 50 nmol/l) displayed an identical cytokine response compared with sufficient subjects. In conclusion, plasma levels of vitamin D are not correlated with the LPS-induced TNF, IL-6 and IL-10 cytokine response in humans in vivo. These findings question the direct role of vitamin D in modulation of the innate immune response.

Introduction

Vitamin D has a plethora of functions in the human body, of which the regulation of calcium metabolism is the most well known [1,2]. Vitamin D₃ (cholecalciferol) is obtained through de-novo synthesis in the skin under influence of ultraviolet (UV)-B exposure and dietary intake [3,4]. Vitamin D₃ is hydroxylated in the liver into 25 hydroxyvitamin D₃ [25(OH)D₃], its main circulating metabolite that is used as a marker of vitamin D status [3,4]. Ultimately, 25(OH)D₃ is converted by 1α-hydroxylase into its biological active metabolite 1,25 dihydroxyvitamin D₃ [1,25(OH)₂D₃] that can bind to the vitamin D receptor (VDR). 1,25(OH)₂D₃ is formed in many tissues/cells of the body, of which the kidney is the most important source [4,5]. Especially at higher latitudes, vitamin D status has been described to be subject to seasonal variation, with lower levels during the winter months and early spring due to little UV-B radiation exposure [6–9].

Recently, an important role for vitamin D in the modulation of the innate immune response has been identified [10]. 1,25(OH)₂D₃ limits production of proinflammatory cytokines such as tumour necrosis factor (TNF)-α and interleukin (IL)-6 in lipopolysaccharide (LPS)- and Candida albicans-stimulated human peripheral blood mononuclear cells (PBMCs) in vitro [7,8]. Furthermore, seasonal variations in 25(OH)D₃ and 1,25(OH)₂D₃ levels (in healthy volunteers living in the Netherlands at 52·5° N latitude) correlated with alterations in cytokine production by ex-vivo LPS-stimulated PBMCs, with lower TNF-α and IL-6 production in winter [7]. Encouraged by the increasing evidence of immunomodulatory effects of vitamin D, a role for seasonality/UV exposure and vitamin D levels in the prevalence of autoimmune diseases such as multiple sclerosis and rheumatoid arthritis, and susceptibility towards influenza and viral upper respiratory tract infections during the winter season, has been implicated [11,12].

As outlined above, until now, the effects of vitamin D on the innate immune response have been established in vitro and ex vivo; however, in-vivo evidence in humans is lacking. In this retrospective study we investigated whether vitamin D plasma levels of healthy volunteers in the Netherlands are subject to seasonal variation and whether they correlate with the plasma cytokine response during experimental human endotoxaemia (LPS administration), a well-characterized, standardized model of the innate immune response in humans in vivo [13].

Materials and methods

Subjects

The study population consisted of 112 young healthy non-smoking male student volunteers who participated in seven experimental human endotoxaemia trials performed in our institution from 2005 to 2010 (NCT00246714, NCT00513110, NCT00783068, NCT00785018, NCT00916448, NCT01349699, NCT01091571). We included only subjects who, apart from LPS, received either no intervention, a placebo or an intervention that did not influence their cytokine response and of whom sufficient plasma samples were left to analyse vitamin D and plasma cytokine levels. Subjects from each separate trial were not distributed equally over the four seasons, although there was overlap between two or even three seasons within four of the seven trials. As the protocol of the different LPS trials was identical, bias is unlikely. All subjects were Caucasian residents of the Netherlands (52·5° N latitude). We studied only male volunteers because of considerable differences in the LPS-induced cytokine response between males and females, and possible effects of hormonal variations due to the menstrual cycle [14]. The study protocols were approved by the Ethics Committee of the Radboud University Nijmegen Medical Centre and complied with the Declaration of Helsinki including current revisions and the Good Clinical Practice guidelines. Written informed consent was obtained from all study participants. Screening of the subjects within 14 days before the test revealed no abnormalities in medical history or physical examination. Volunteers were not taking any prescription medications, and were tested negative for hepatitis B surface antigen and human immunodeficiency virus infection. Based on the day of LPS administration (which was also the day that plasma for vitamin D analysis was collected), we divided the study population into the four seasons: spring (April–June), summer (July–September), autumn (October–December) and winter (January–March).

Experimental human endotoxaemia experiments

Subjects refrained from food 12 h before the start of the experiment, and caffeine- or alcohol-containing substances 24 h before the start of the experiment. The experiments were performed at the research unit of the intensive care department, with subjects in the supine position. After local anaesthesia (lidocaine HCl 20 mg/ml) the radial artery was cannulated using a 20-gauge arterial catheter (Angiocath; Becton Dickinson, San Jose, CA, USA) for blood sampling and arterial blood pressure monitoring (Edwards Lifesciences LLC, Irvine, CA, USA), connected to a Phillips IntelliVue MP70 monitor (Philips Medical Systems, Amsterdam, the Netherlands). A cannula was placed in the antecubital vein to permit infusion of 2·5% glucose/0·45% saline solution; subjects received a bolus of 1·5 l for 1 h before LPS infusion (prehydration), followed by 150 ml/h until 6 h after LPS infusion and 75 ml/h until the end of the experiment. US reference Escherichia coli endotoxin (E. coli O:113, Clinical Center Reference Endotoxin; National Institutes of Health (NIH), Bethesda, MA, USA) was used. Ec-5 endotoxin, supplied as a lypophilized powder, was reconstituted in 0·9% NaCl and vortex-mixed for at least 10 min after reconstitution. The endotoxin solution was administered as an intravenous bolus injection at a dose of 2 ng/kg of body weight. For safety reasons vital signs, including heart rate (HR) and mean arterial pressure (MAP), were monitored continuously throughout the study protocol.

25(OH)D₃ and 1,25(OH)₂D₃ analysis

25(OH)D₃ and 1,25(OH)₂D₃ levels were determined in plasma samples obtained just prior to LPS administration. To this end, ethylenediamine tetraacetic acid (EDTA) anti-coagulated blood was collected from the arterial line and centrifuged immediately at 2000 g for 10 min at 4°C, after which plasma was stored at −80°C until analysis. 25 OHD₃ was measured in one batch using the Roche Cobas E601 analyser, according to the manufacturer's instructions [vitamin D₃ (25-OH) assay with polyclonal antibody; Roche Diagnostics, Burgess Hill, UK]. 1,25(OH)₂D₃ levels were analysed in one batch using a radioimmunoassay (1,25 dihydroxyvitamin D kit; Immunodiagnostic Systems Ltd, Boldon, UK). Vitamin D deficiency was defined as plasma 25(OH)D₃ levels of < 50 nmol/l, which is advocated by the American Institute of Medicine (2011).

Plasma cytokine analysis

EDTA anti-coagulated blood was collected from the arterial line and centrifuged immediately at 2000 g for 10 min at 4°C after which plasma was stored at −80°C until analysis. Concentrations of TNF-α, IL-6 and IL-10 were measured in one batch using a simultaneous Luminex Assay, according to the manufacturer's instructions (Milliplex; Millipore, Billerica, MA, USA). Detection range was 3·2–10 000 pg/ml for all cytokines.

Statistical analysis

Data are represented as mean or medians based on distribution (calculated by the Shapiro–Wilk test). Statistical tests used are indicated in the text and/or figure legends. Statistical analyses were performed using Graphpad Prism 5 (Graphpad Software, La Jolla, CA, USA).

Results

Seasonal variation in vitamin D plasma levels

Demographic characteristics of the study population of 112 healthy volunteers are listed in Table 1. Plasma levels of 25(OH)D₃ did not differ significantly between seasons (Fig. 1a). However, 1,25(OH)₂D₃ levels were subject to seasonal variation, with lower concentrations found in autumn and winter (Fig. 1b). There was a weak but significant correlation between plasma levels of 25(OH)D₃ and 1,25(OH)₂D₃ (Spearman's r = 0·23, P = 0·02).

Table 1.

Demographic characteristics of the study population

Parameter	All subjects (n = 112)	Spring (n = 32)	Summer (n = 21)	Autumn (n = 34)	Winter (n = 25)
Age (years)	22·2 ± 0·3	21·1 ± 0·3	22·9 ± 0·7	22·7 ± 0·4	22·5 ± 0·6
BMI (kg/m2)	22·8 ± 0·2	23·1 ± 0·5	22·2 ± 0·5	22·7 ± 0·4	22·9 ± 0·4

Data are presented as mean ± standard error of the mean. BMI: body mass index.

Fig. 1.

25 Hydroxyvitamin D₃ [25(OH)D₃] and 1,25 dihydroxyvitamin D₃ [1,25(OH)₂D₃] plasma levels, divided according to the four seasons: spring (April–June, n = 32), summer (July–September, n = 21), autumn (October–December, n = 34) and winter (January–March, n = 25). Data are represented as box and whiskers (range). P-values calculated using the Kruskal–Wallis test with Dunn's post-hoc test (* indicates P < 0·05).

Correlation of vitamin D levels with the LPS-induced plasma cytokine response

The pattern of LPS-induced plasma cytokine levels of the entire study population are depicted in Fig. 2. Plasma levels of 25(OH)D₃ and 1,25(OH)₂D₃ did not correlate with the LPS-induced plasma cytokine response (correlations between vitamin D levels and both area under the curve (AUC) and peak plasma cytokine levels are listed in Table 2).

Fig. 2.

Time–course of lipopolysaccharide (LPS)-induced plasma cytokines of the 112 subjects. Data are represented as box and whiskers (range) of 112 subjects.

Table 2.

Correlation between plasma levels of 25 hydroxyvitamin D₃ [25(OH)D₃₎] or 1,25 dihydroxyvitamin D₃ [1,25(OH)₂D₃] and AUC and peak plasma cytokines

	AUC TNF-α	Peak TNF-α	AUC IL-6	Peak IL-6	AUC IL-10	Peak IL-10
25(OH)D3	−0·06 (0·50)	−0·04 (0·69)	−0·03 (0·78)	−0·06 (0·52)	0·13 (0·17)	0·14 (0·16)
1,25(OH)2D3	0·06 (0·55)	0·11 (0·25)	−0·07 (0·50)	−0·05 (0·61)	0·13 (0·19)	0·09 (0·36)

Spearman's correlation coefficients (P-value) of 112 subjects are listed. AUC: area under the curve of plasma cytokines in time; IL: interleukin; TNF: tumour necrosis factor.

LPS-induced cytokine responses in vitamin D-deficient and -sufficient subjects

To investigate further the possible immunomodulatory effects of 25(OH)D₃ we compared LPS-induced cytokine responses of 25(OH)D₃-deficient subjects to those with sufficient levels. The majority of our study population was 25(OH)D₃-deficient [65 deficient (58%) versus 47 sufficient (42%)]. However, as depicted in Fig. 3, there were no differences in cytokine responses between deficient and sufficient subjects (area under curve of plasma cytokine depicted in Fig. 3, peak cytokine levels showed similar results and are not shown). Finally, we analysed the plasma cytokine responses (AUC and peak levels) based on a division of 25(OH)D₃ levels in quartiles, which would increase sensitivity in the case of a non-linear relation between 25(OH)D₃ levels and cytokine responses, but again no differences were found (data not shown). To evaluate the power of this study, we performed a calculation to determine the minimum detectable difference in AUC TNF-α levels between 25(OH)D₃-deficient and -sufficient subjects (based on the data from Fig. 3, left panel). Based on 65 deficient and 47 sufficient subjects and an alpha of 0·05, we would be able to detect a difference in TNF levels of 25% with a power of 80%.

Fig. 3.

Area under curve of lipopolysaccharide (LPS)-induced plasma cytokines between 25 hydroxyvitamin D₃ [25(OH)D₃]-deficient (< 50 nmol/l, n = 65) and sufficient (≥ 50 nmol/l, n = 47) subjects. Lines indicate median values. P-values calculated using Mann–Whitney U-tests.

Discussion

In the present study, we investigated the seasonal variation of vitamin D plasma levels in young healthy males in the Netherlands and its relation with the plasma cytokine response induced by LPS administration in vivo. We did not find a significant seasonal variation in levels of 25(OH)D₃, the main circulating metabolite. However, plasma concentrations of the active metabolite 1,25(OH)₂D₃ differed between seasons, with lower levels in autumn and winter. Furthermore, no relation between both vitamin D metabolites and the LPS-induced plasma cytokine response was found. The main finding of our study is that vitamin D status is not associated with the systemic inflammatory cytokine response during experimental human endotoxaemia.

In contrast to what others have found at both higher, lower and similar latitudes [6–9,15–17], there was no seasonal variation in levels of the main circulating vitamin D metabolite 25(OH)D₃ in our study. Moreover, the majority of the study subjects was deficient for 25(OH)D₃. The high prevalence of vitamin D deficiency in our healthy young male population was surprising with regard to the fact that deficiency is associated with low sun exposition, as seen at older age and for immigrant females. Interestingly, in a Norwegian study (60° N latitude) in elderly people, seasonal variation was found in a home-living group but not in a hospitalized group [18]. Furthermore, in the hospitalized group the prevalence of 25(OH)D₃ deficiency was much higher compared to the home-living group [18]. It appears that in our study population, consisting of young male volunteers, virtually all of them students, UV-B exposure as the main resource as well as additional dietary intake may be relatively low. Unfortunately, we did not record information regarding vitamin D supplementation, outdoor activities or holidays in our study population.

In accordance with other studies [6,18], we found a weak but nevertheless significant correlation between 25(OH)D₃ and1,25(OH)₂D₃ levels. The seasonal variation in 1,25(OH)₂D₃ levels as opposed to 25(OH)D₃ was unexpected, as 25(OH)D₃ is the main circulating metabolite of vitamin D and can be considered the source for 1,25(OH)₂D₃, and production of the latter is not influenced by UV-B radiation. The fact that 1,25(OH)₂D₃ levels were lower in autumn and winter and not in winter and spring is in agreement with an UV-B independent effect. Interindividual differences in 1α-hydroxylase activity, due possibly to variation in levels of parathyroid hormone (PTH) (which stimulates 1α-hydroxylase activity [19]), may represent a possible explanation for this discrepancy [6]. In this respect, low dietary calcium intake might be involved, as reduced calcium levels result in increased PTH levels and subsequent augmented 1α-hydroxylase activity [6]. Although we included only subjects who, besides LPS, received no intervention, a placebo or an intervention that did not influence their cytokine response, the interventions could potentially have influenced vitamin D levels. To address this, we also analysed seasonal variation in 25(OH)D₃ and 1,25(OH)₂D₃ levels in the group of subjects that received no intervention or a placebo (n = 58), which revealed similar results (data not shown).

Two recent publications have found distinct anti-inflammatory effects of vitamin D; 1,25(OH)₂D₃ was shown to inhibit LPS-induced proinflammatory cytokine production by human PBMCs in vitro, and seasonal variation in cytokine production of ex-vivo LPS-stimulated PBMCs of healthy volunteers was found, which correlated with their 25(OH)D₃ and 1,25(OH)₂D₃ levels [7,8]. In contrast, we did not find any correlations between 25(OH)D₃ or 1,25(OH)₂D₃ levels and inflammatory cytokines after LPS administration in vivo. In accordance, no differences in plasma cytokine levels were found between vitamin D-sufficient and -deficient subjects. An explanation might be that the anti-inflammatory effects of 1,25(OH)₂D₃ in vitro were found at concentrations of 100 nmol/l [7,8], corresponding to 38 000 pg/ml, whereas the highest plasma concentration found in our study population was 72 pg/ml. Furthermore, higher plasma concentrations of 1,25(OH)₂D₃ (up to twofold higher) and especially 25(OH)D₃ (up to fourfold higher) were found compared to our study [7,8]. In this respect, it might be speculated that 25(OH)D₃ vitamin D levels in our subjects were too low to exert any anti-inflammatory effects. An alternative explanation is compartmentalization of the innate immune response. We have shown earlier that the cytokine production of ex-vivo LPS-stimulated leucocytes does not correlate with the plasma cytokine response elicited by LPS administration in vivo in the same subject [20,21]. In this respect, vitamin D levels might influence the leucocyte compartment, but not tissue-resident cells of the immune system such as macrophages, which are suggested to be the major cytokine producers after LPS administration in vivo [21]. It could be argued that our study was not powered adequately to detect more subtle (< 25%) differences in cytokine responses. However, the actual difference in AUC TNF-α levels between deficient and sufficient subjects was less than 4%. Irrespective of the relevance of such a small difference, to reach statistical significance of this difference with a power of 80%, a study population of almost 4500 subjects would be needed.

To the best of our knowledge, there are no studies regarding the direct effects of vitamin D (supplementation) on systemic inflammation in vivo in humans or animals. However, several studies have shown that the human cationic anti-microbial peptide LL-37, one of the downstream targets of the VDR, decreases plasma cytokine levels and reduces mortality in rodent sepsis and endotoxaemia through neutralization of LPS [22–24]. As LL-37 also promotes innate immunity due to its direct antimicrobial effects, neutralization of LPS has been proposed as a compensatory mechanism to curtain excessive inflammation [23]. Because no correlations between 25(OH)D₃/1,25(OH)₂D₃ levels and cytokine responses were found, and LL-37 can be induced by many other compounds/stimuli apart from 1,25(OH)₂D₃ [25], a possible relation between LL-37 and the plasma cytokine response in our study would be independent from vitamin D status, and therefore outside the study's scope.

Several limitations of our study deserve attention. First, we studied only young healthy male student volunteers living at the same latitude, which is an advantage with regard to the homogeneity of the study population but limits extrapolation to the general population. However, to our knowledge there are no data that indicate differential effects of vitamin D on the innate immune response between sexes or age classes. Secondly, we studied only a limited set of inflammatory mediators. Vitamin D has been shown to modulate production of other cytokines in response to LPS, notably IL-1β [7]. However, IL-1β is barely detectable during human endotoxaemia and was therefore not analysed in this study [14]. Thirdly, the human experimental endotoxaemia model is limited to activation of the innate immune system, while vitamin D has also been shown to modulate the T cell compartment and therefore adaptive immune responses [26–28]. Therefore, vitamin D might exert immunomodulatory effects that do not emerge during experimental endotoxaemia. Furthermore, due to the use of LPS, the human endotoxaemia model is confined to Toll-like receptor (TLR)-4 signalling. While LPS is one of the most potent activators of innate immunity, vitamin D may exert immunomodulatory effects upon activation of the innate immune system with other stimuli, such as other TLR ligands or Candida albicans [8]. Finally, this model does not represent an innate immune response against live bacteria in which multiple other factors that are possibly influenced by vitamin D, such as direct anti-microbial effects of LL-37 and neutrophil function, may play a role. Neutrophils, vital in killing of pathogens, express the vitamin D receptor and vitamin D has been shown to directly modulate both neutrophil chemotaxis and activity [29–31].

In conclusion, our study reveals that plasma levels of 25(OH)D₃ and 1,25(OH)₂D₃ are not correlated to the systemic inflammatory cytokine response during experimental human endotoxaemia. These findings question the direct role of vitamin D in modulation of the innate immune system. Intervention studies in which vitamin D is supplemented before an inflammatory challenge could provide the definitive answer to the question of whether or not vitamin D exerts direct immunomodulatory effects in humans in vivo.

Disclosure

Authors declare that they have no conflicts of interest.