More than 6 million adults in Canada have an increased risk of type 2 diabetes (T2D)(Reference Rosella, Lebenbaum and Fitzpatrick1). Several observational studies have indicated that vitamin D deficiency may contribute to the development of T2D(Reference Bouillon, Manousaki and Rosen2). Vitamin D status is typically assessed by measuring the serum concentration of 25-hydroxyvitamin D (25(OH)D), the primary circulating form and precursor to the biologically active form, 1,25-dihydroxyvitamin D3 (Reference Hollis3). A systematic review and meta-analysis of fifty-four studies (n 38 016) reported a pooled prevalence of vitamin D deficiency among patients with T2D of 64·2 %(Reference Taderegew, Woldeamanuel and Wondie4). The highest prevalence was observed in African countries (70·9 %), and the lowest was found in Middle Eastern countries (57·1 %)(Reference Taderegew, Woldeamanuel and Wondie4). Although observational studies suggest an inverse relationship between vitamin D levels and T2D development, large studies on levels of vitamin D in Canadian subjects with T2D are lacking(Reference Grossman and Roscoe5). Nevertheless, data from the Canadian Health Measures Survey showed that 10 % of individuals had levels between 30 and 40 nanomoles per litre (nmol/l), which is considered insufficient, while 8 % had levels below 30 nmol/l, indicating deficiency in the Canadian population(6).
Vitamin D deficiency is associated with impaired Ca absorption as the active form of vitamin D, 1,25-dihydroxyvitamin D3, is essential for stimulating active transcellular Ca transport in the intestine(Reference Fleet7). Therefore, deficient vitamin D levels may reduce insulin secretion in response to increased glucose since insulin secretion is calcium dependent(Reference Mitri, Dawson-Hughes and Hu8). Vitamin D deficiency has also been associated with increased systemic inflammation, which contributes to T2D pathogenesis(Reference Wimalawansa9–Reference Szymczak-Pajor, Drzewoski and Śliwińska11). Mechanistically, vitamin D may regulate inflammatory cytokines by interacting with the vitamin D receptor that binds to vitamin D response elements in the promoter regions of cytokine-related genes or through modulation of transcription factors involved in immune signalling(Reference Wimalawansa9). Recent in vitro studies have shown that 25(OH)D regulates hundreds of genes involved in immune function and signalling pathways via vitamin D response element-mediated transcriptional regulation, many of which are implicated in the pathogenesis of T2D(Reference Ramagopalan, Heger and Berlanga12). Overall, vitamin D deficiency may promote insulin resistance and accelerate the development of T2D through impaired calcium signalling and increased systemic inflammation.
Vitamin D is produced through cutaneous synthesis following ultraviolet B exposure as well as obtained from dietary sources and supplementation(Reference Hawk13). Regular sun exposure is the most natural way to get enough vitamin D(Reference Hawk13). Seasonal variation affects serum 25(OH)D levels, with the lowest levels typically observed after winter and the highest at the end of summer(Reference Klingberg, Oleröd and Konar14). This fluctuation is important to consider when interpreting vitamin D status and its potential biological effects(Reference Klingberg, Oleröd and Konar14). Therefore, Health Canada emphasised the intake of vitamin D-fortified foods and supplements to meet vitamin D needs(Reference Gill and Kalia15).
Milk, yogurt, fatty fish, beef, mushrooms and eggs are sources of vitamin D, and these foods may interact synergistically within the human body(16). Since 2022, Vitamin D is added to milk at a standard level of approximately 2 μg/100 ml as mandated by Health Canada’s regulations(16). In a study with thirty-seven elderly women, the dairy product provided 17–25 % of the recommended dose of vitamin D (10–15 mg)(Reference Bonjour, Benoit and Pourchaire17). In addition, a review suggested that consuming at least three servings of low-fat dairy daily may reduce T2D risk, mainly due to beneficial effects of increased Ca and vitamin D ingestion on insulin sensitivity(Reference Cândido, Ton and Alfenas18).
Emerging evidence suggests that serum 25(OH)D concentrations is associated with alterations of the abundance of various plasma and serum proteins involved in lipid metabolism or inflammation(Reference Mousa, Razali and Zughaier19–Reference Garcia-Bailo, Joseph and Da Costa21). Additionally, hormonal contraceptive use has been shown to influence the relationship between 25(OH)D levels and positive acute-phase proteins(Reference Garcia-Bailo, Joseph and Da Costa21). However, the correlation between serum 25(OH)D levels and blood protein biomarkers in individuals at risk of developing T2D remains unclear. The objective of the present study was to investigate the relationship between serum protein profiles and 25(OH)D levels, categorised as insufficient (< 50 nmol/l), adequate (50–74·9 nmol/l) and optimal (≥ 75 nmol/l), in overweight hyperinsulinaemic individuals participating in a randomised crossover dairy intervention trial.
Methods
The present study represents a secondary proteomic analysis conducted within the randomised crossover dairy intervention trial. Participant recruitment was conducted at the Centre de recherche du CHU de Québec–Université Laval in Quebec City, Canada, between February 2017 and July 2018. A flow diagram of the study protocol is provided in Figure 1. A total of twenty-five French-Canadian adults with hyperinsulinaemia (nineteen males and eight females) were enrolled based on defined eligibility criteria. The average participant age was 55 (sd 14) years, and the mean BMI was 31·3 (sd 3·3) kg/m2. The proteomic data were obtained from twenty-five French-Canadian subjects (eighteen men and seven women).
Flow diagram of the study protocol.
Figure 1. Long description
Flow diagram showing participant progression through a randomized crossover dairy intervention study. A total of 396 individuals were assessed for eligibility, 33 were randomized, and participants were allocated to adequate dairy or high dairy intake interventions. Following intervention periods, washout, losses to follow-up, and exclusions, 25 participants were included in the final proteomic analysis.
Eligibility criteria included fasting insulin concentration exceeding 90 pmol/l, fasting plasma glucose in serum below 7·0 mmol/l HbA1c levels under 6·5 %. Participants were recruited through advertisements and institutional mailing lists.
The study protocol was reviewed and approved by the Research Ethics Committee of the CHU de Québec–Université Laval (approvals 2017–3228 and 2022–6092) and assigned to the principles of the Declaration of Helsinki. All participants received comprehensive information about the study’s purpose and procedures and provided written informed consent. The clinical trial was registered on ClinicalTrials.gov (accessed on 24 September 2023) under the registration number NCT02961179. This study aligns with the CONSORT 2010 extension for randomised crossover trials(Reference Dwan, Li and Altman22) and the updated CONSORT 2025 guidelines(Reference Hopewell, Chan and Collins23). Completed CONSORT checklists are provided in the online Supplementary Material.
Dietary intervention
Participants were randomly assigned to follow either a high-dairy (HD; ≥ 4 servings/d) or adequate-dairy (AD; ≤ 2 servings/d) for the first 6-week intervention phase. After a 6-week washout period, the groups switched diets in a crossover manner. Serving sizes and choices were based on national food-guide recommendations for dairy products, following the serving sizes outlined in Canada’s Food Guide for Healthy Eating (2007)(Reference Katamay, Esslinger and Vigneault24), where a serving is defined as one cup (250 ml) of milk, ¾ cup (175 g) of yogurt or kefir or 1·5 ounces (50 g) of cheese. Ice cream (125 ml per serving) was permitted up to three times per week. Products such as butter, whipped cream, higher-fat creams (> 15 %), modified milk ingredients, milk substitutes and derivatives were not counted towards the daily dairy total. AD intake was limited to no more than two servings per day.
During the HD phase, participants consumed 4–5 dairy servings per day. All volunteers were asked to keep their usual eating patterns, physical activity and lifestyle habits unchanged throughout the study. A trained dietitian provided written instructions during telephone screening, and adherence was checked at every visit using a validated FFQ. To support compliance, three reminder emails were sent during each intervention period to request a 24-h dietary recall. Metabolomic profiling confirmed changes in serum heptadecanoic acid levels, indicating adherence to the prescribed dairy intake(Reference O’Connor, Greffard and Leclercq25).
Dietary intake was assessed at each visit with a web-based FFQ comprising ninety-one items and thirty-three follow-up questions, linked to a nutrition data system for research(Reference Goulet, Nadeau and Lapointe26). Nutrient values were calculated with the 2015 Canadian nutrient database(27).
Anthropometric assessments
Participants’ anthropometric data, including body weight (kg) and height (cm), were collected during the study. Body weight was measured using a calibrated professional digital scale (Health O Meter Professional, Sunbeam Products, Inc.) with a precision of 0·1 kg. Height was determined using a fixed wall-mounted stadiometer (Easy-Glide Bearing Stadiometer, Perspective Enterprises), accurate to within 1 mm. To ensure standardisation, participants wore lightweight clothing and remained barefoot during measurements. BMI was then computed as weight in kilograms divided by height in metres squared (kg/m²).
Clinical measurements
The clinical assessments were performed over four scheduled visits: at the beginning and end of each dietary intervention phase, spaced 6 weeks apart. Fasting blood glucose levels were measured using a hexokinase enzymatic method on a Siemens clinical chemistry platform(Reference Passey, Gillum and Fuller28), while fasting insulin concentrations were determined using a Siemens chemiluminescent immunoassay system(Reference Kirk and Othmer29). Insulin resistance was estimated using homeostatic model assessment for insulin resistance formula: [fasting insulin (pmol/l) × fasting glucose (mmol/l)]/135(Reference Kirk and Othmer29).
Serum lipid profiles, including total cholesterol and TAG, were assessed using standard enzymatic colorimetric assays with Siemens reagents and instrumentation(Reference McNamara and Schaefer30,Reference Burstein31) . The concentration of HDL-cholesterol was obtained after precipitation of very low-density and low-density lipoproteins using a heparin-manganese chloride solution(Reference Ash and Hentschel32). Levels of LDL-cholesterol were calculated using the Friedewald equation(Reference Albers, Warnick and Wiebe33).
Identification and quantification of serum proteins using MS
Serum protein profiling was carried out using a previously established protocol(Reference Mahdavi, Leclercq and Bodein34). In brief, each sample was spiked with indexed retention time peptides (0·5X; Biognosys) and analysed via nanoLC-tandem MS. The analytical setup consisted of a Dionex UltiMate 3000 nanoRSLC chromatographic system (Thermo Fisher Scientific) coupled to an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific) equipped with a nanoelectrospray ion source.
Peptides were initially captured on a PepMap C18 pre-column (5 mm × 300 µm; Thermo Fisher Scientific) at a flow rate of 20 μl/min using a loading solvent composed of 2 % isopropanol and 0·05 % trifluoroacetic acid. After 5 min, the analytical separation was performed on a 50 cm × 75 µm PepMap Acclaim C18 column (Thermo Fisher Scientific). Peptides were eluted using a linear gradient of solvent B (80 % isopropanol with 0·1 % formic acid) against solvent A (0·1 % formic acid), with a flow rate of 300 nl/min over a 90-minute runtime.
Mass spectrometric data were acquired in Data-Independent Acquisition mode using Thermo Xcalibur software (v4.3.73.11). The system employed internal mass calibration using the lock mass feature and the siloxane ion (m/z 445.12003). The electrospray voltage was set at 2100 V, and the capillary temperature was maintained at 275°C. MS1 scans were recorded in the Orbitrap over a 350–1000 m/z range at 60 000 resolution (automatic gain control target: 4 × 105, injection time: 50 ms). Each full MS scan was followed by seventy-five tandem MS fragmentation scans, covering 350–950 m/z with an isolation window of 8 m/z.
Precursor selection used quadrupole isolation, and fragmentation was performed using higher energy collisional dissociation with a normalised collision energy of 35 %. tandem MS fragment ions were detected in the Orbitrap at a resolution of 15 000 (200–1500 m/z), with an automatic gain control target of 5 × 104 and injection time of 22 ms.
Raw MS files were converted to mass-to-charge markup language format using MSConvert (version 3.0.21188). Data extraction and quantification were performed using data-independent acquisition-NN (version 1.8)(Reference Mahdavi, Leclercq and Bodein34). The analysis was based on the Homo sapiens reference proteome from UniProt (Proteome ID: UP000005640; July 2020 release; 63 807 protein entries). Data-independent acquisition-NN’s deep learning-based spectral prediction was used. Trypsin/P was selected as the digestion enzyme, allowing up to two missed cleavages. One variable modification was permitted per peptide (N-terminal methionine excision or methionine oxidation), and carbamidomethylation of cysteine was considered a fixed modification. Peptides between 7 and 30 amino acids in length, and precursor charge states of +2 to +4 (m/z range 350–950) were included. The fragment mass range was set to 200–1500 m/z. Match between runs was enabled, with all other parameters kept at default.
All resulting MS data have been deposited in the ProteomeXchange repository (https://www.proteomexchange.org) and are publicly available under accession number PXD035386 (accessed on 12 December 2022).
Serum levels of 25-hydoxyvitamin D
Blood samples were collected in serum separator tubes without anticoagulant. Serum was obtained after coagulation and centrifugation and stored at 4°C for no more than 12 h before being transferred to –80°C for long-term storage. Total 25-hydroxyvitamin D (25(OH)D) concentrations were measured using an ADVIA Centaur XP analyser (Siemens Healthineers) with a competitive chemiluminescent immunoassay. Calibration and quality control were performed according to the manufacturer’s instructions at the CHU de Quebec- Université Laval Hospital (Quebec, Canada). The intra-assay coefficient of variation for measurement was below 10 %, indicating acceptable assay precision.
25(OH)D concentrations were analysed in three ways: as continuous values, by deciles and as categorical variables using thresholds of 30, 50 nmol/l and 75, in line with the Canadian Health Measures Survey. Vitamin D insufficiency was defined as a serum 25(OH)D level below 50 nmol/l, while a concentration between 50 and 74·9 nmol/l was considered adequate and above 75 nmol/l as optimal levels in the different subjects.
Statistical analysis
Descriptive statistics were calculated to compare characteristics of participants across three categories of serum 25(OH)D concentration: < 50 nmol/l, 50–74·9 nmol/l and ≥ 75 nmol/l. Continuous variables, including age, weight, BMI, lipid profile (HDL, LDL, total cholesterol and TAG), fasting glucose and insulin, homeostatic model assessment for insulin resistance index and dietary intakes (energy, fat and vitamin D), were summarised using means and sd. Differences across vitamin D categories evaluated using one-way ANOVA, and corresponding P values were reported to assess statistical significance.
To normalise the data, continuous variables were log-transformed prior to correlation analyses. Partial correlation analyses were then performed to assess the correlation between serum 25(OH)D and individual proteins, adjusting for key confounding variables such as age, sex, BMI, vitamin D intake and cholesterol levels. Spearman correlation analyses were also conducted to examine the overall correlations between 25(OH)D and other clinical variables. Given the exploratory nature of the proteomic analyses, no multiple testing correction (e.g., false discovery rate) was applied to the Spearman correlation P values. Unadjusted P values (P < 0·05) were used to identify candidate proteins for subsequent pathway enrichment analyses, which were evaluated using more stringent criteria at the pathway level. Pathway enrichment analysis was performed using Database for Annotation, Visualisation, and Integrated Discovery, and pathways were considered significant at a Bonferroni-adjusted P value < 0·05. All statistical analyses were performed using R software.
Results
Characteristics of participants
The clinical and dietary characteristics of twenty-five hyperinsulinaemic subjects categorised by vitamin D status are presented in Table 1. Participants were classified into three groups based on serum 25(OH)D concentration: insufficient (< 50 nmol/l, n 9), adequate (50–74·9 nmol/l, n 6) and optimal (≥ 75 nmol/l, n 10). There were no significant differences in age, body weight, BMI, waist circumference, fasting glucose, fasting insulin, homeostatic model assessment for insulin resistance, HDL, LDL, total cholesterol, TAG levels or cholesterol/HDL and LDL/HDL ratios across vitamin D categories. However, as expected, both serum 25(OH)D concentration and vitamin D intake from the diet differed between different categories (P < 0·001 and P = 0·04, respectively). Participants in the optimal group had the highest vitamin D intake (43·7 (sd 34·9) µg/d) and serum 25(OH)D levels (91·8 (sd 20·97) nmol/l). Subjects in the adequate group had moderate dietary vitamin D intake (19·1 (sd 24·6) µg/d) and serum 25(OH)D levels (63·0 (sd 8·1) nmol/l), while participants in the insufficient group had the lowest intake and serum 25(OH)D concentrations (8·2 (sd 4·8) µg/d and 38·7 (sd 9·5) nmol/l, respectively).
Characteristics of individuals with hyperinsulinaemia across different categories of vitamin D status
Table 1. Long description
Table presenting demographic, anthropometric, and metabolic characteristics of individuals with hyperinsulinemia categorized by vitamin D status. Variables include participant characteristics and clinical measurements, with comparisons between vitamin D status groups.
The number of participants was 25, stratified into three groups based on vitamin D levels: insufficient (< 50 nmol/l, n 9), adequate (50–74·9 nmol/l, n 10) and optimal (≥ 75 nmol/l, n 6). Information displayed as the mean (standard deviation). Significance was defined as a P value of less than 0·05.
Serum 25-hydroxyvitamin D levels in different seasons
The mean level of 25(OH)D was 69·57 nmol/l (sd = 31·25) in winter, 73·65 nmol/l (sd = 28·66) in spring, 70·10 nmol/l (sd = 20·26) in summer and 71·30 nmol/l (sd = 22·20) in fall. There were no significant differences in 25(OH)D levels across the seasons.
Serum 25-hydroxyvitamin D levels before and after high dairy intake
The mean levels of 25(OH)D in the 25 subjects were measured under four conditions: pre-AD (67·52 (sd 24·39)), post-AD (72·41 (sd 25·77)), pre-HD (75·08 (sd 26·53)) and post-HD (70·54 (sd 27·83)). No statistically significant differences were observed between these groups.
Correlations between serum 25-hydroxyvitamin D and serum proteins in the three stratified categories
Based on the Canadian Health Measures Survey, serum 25(OH)D concentrations are influenced by several cofactors, including age, sex, BMI, dietary vitamin D intake, cholesterol levels, supplementation and season(Reference Wimalawansa9). These factors are commonly adjusted for in various statistical analyses. A partial correlation analysis, adjusted for age, sex, BMI, dietary vitamin D intake, cholesterol, supplementation and season, was conducted to identify serum proteins correlated with 25(OH)D levels, stratified into three categories: < 50 nmol/l, 50–74·9 nmol/l and ≥ 75 nmol/l. Table 2 presents the proteins that showed significant positive or negative correlations with serum 25(OH)D in each category. Among participants with 25(OH)D levels < 50 nmol/l, sixteen proteins showed positive correlations, while four proteins showed negative correlations with 25(OH)D. Among those with adequate levels (50–74·9 nmol/l), twenty-six and fourteen proteins exhibited positive and negative correlations, respectively. In the ≥ 75 nmol/l category, thirty-five proteins showed positive correlations, while twenty proteins were negatively correlated with serum 25(OH)D.
Serum proteins significantly correlated with 25(OH)D levels in individuals with hyperinsulinemia, stratified by vitamin D status
Table 2. Long description
Table listing serum proteins significantly associated with circulating 25-hydroxyvitamin D concentrations in individuals with hyperinsulinemia. Results are presented separately according to vitamin D status and include correlation coefficients and statistical significance values.
Adjusted for age, sex, BMI, dietary vitamin D intake, cholesterol, supplementation and season; 25(OH)D status: insufficient (< 50 nmol/l, n 9), adequate (50–74·9 nmol/l, n 10) and optimal (≥ 75 nmol/l, n 6).
To investigate the biological pathways associated with proteins correlating with 25(OH)D levels in the three different groups, gene ontology enrichment analysis was performed(Reference Dennis, Sherman and Hosack35). As shown in Table 3, there was no significant biological pathway identified for proteins that correlated with 25(OH)D levels in subject exhibiting < 50 nmol/l of 25(OH)D. In contrast, several immune-related pathways were significantly enriched among proteins positively correlating with 25(OH)D in the 50–74·9 nmol/l group, including complement activation, acute-phase response, innate immune response, immune response, neutrophil chemotaxis, positive regulation of interleukin-1 production, inflammatory response and killing of cells of another organism. The proteins that correlated negatively with 25(OH)D in the 50–74·9 nmol/l group were involved in adaptive immune response. For the subject exhibiting levels of 25(OH)D ≥ 75 nmol/l, several proteins were positively associated to immunoglobulin-mediated immune response, negative regulation of blood coagulation, immune response, acute-phase response and adaptive immune response. Conversely, proteins that negatively correlated with 25(OH)D in this group were significantly enriched in complement activation, classical pathway, innate immune response, acute-phase response and proteolysis.
Biological pathways associated with proteins significantly correlated with 25(OH)D levels, stratified by vitamin D status
Table 3. Long description
Table summarizing biological pathways identified from proteins significantly correlated with 25-hydroxyvitamin D concentrations. Pathways are presented according to vitamin D status and include pathway names together with enrichment statistics.
Adjusted for age, sex, BMI, dietary vitamin D intake, cholesterol, supplementation and season; 25(OH)D status: insufficient (< 50 nmol/l, n 9), adequate (50–74·9 nmol/l, n 10) and optimal (≥ 75 nmol/l, n 6).
Discussion
This proteomic study revealed that the correlation between serum 25(OH)D levels and serum proteins vary across different levels of vitamin D sufficiency. Although no biological pathways were identified among the proteins correlated with 25(OH)D in individuals with insufficiency (< 50 nmol/l), innate immune pathways were enriched at adequate levels (50–74·9 nmol/l), while proteins of the adaptive immune pathways were correlated with 25(OH)D in subjects with optimal levels (> 75 nmol/l).
In the current study, mean serum 25(OH)D concentrations remained above 50 nmol/l before and after both AD and HD interventions, indicating that participants were generally vitamin D sufficient throughout the study period. This may have limited the potential for further increases in circulating 25(OH)D levels following HD intake, consistent with a ceiling effect commonly observed in individuals with adequate baseline vitamin D status. In addition, the 6-week intervention period may not have been long enough to induce detectable differences between the AD and HD conditions. Since participants in the AD arm were already consuming adequate amounts of dairy products, the contrast between the two intervention conditions may also have been insufficient to produce a statistically significant effect. A more pronounced response might have been observed in individuals with lower habitual dairy intake or lower baseline vitamin D status. A review based on twenty studies across diverse populations reported a positive association between vitamin D-fortified milk consumption and serum vitamin D status(Reference Itkonen, Erkkola and Lamberg-Allardt36). In Canada, the 2007 Food Guide recommended two servings of low-fat dairy products per day for adults, which is considered sufficient to help maintain adequate 25(OH)D levels(Reference Katamay, Esslinger and Vigneault24). In 2024, the Food and Drug Regulations in Canada developed a new vitamin D fortification strategy to increase 25(OH)D in dairy products that also included yogurt and kefir(16). Overall, serum 25(OH)D levels remained within a similar range across all groups, suggesting that regular consumption of fortified milk, whether adequate or high intake maintained stable vitamin D status.
Previous proteomic studies have shown that serum vitamin D deficiency can alter the abundance of various serum proteins in both healthy individuals and those with metabolic syndrome. For instance, a proteomic analysis of 274 individuals with dyslipidaemia revealed that participants with vitamin D deficiency (< 12 ng/ml) exhibited elevated levels of the inflammatory marker serum amyloid A1 (SAA1) and superoxide dismutase 1, along with reduced levels of apolipoproteins M and D(Reference Mousa, Razali and Zughaier19). Further pathway analysis indicated that vitamin D deficiency was associated with the upregulation of inflammatory signalling pathways, including mitogen-activated protein kinase, Janus kinase–signal transducer and activator of transcription, Ras and cytokine–receptor interactions(Reference Mousa, Razali and Zughaier19). In a case–control study comparing obese and non-obese individuals, all with vitamin D deficiency (< 50 nmol/l), obese participants showed increased serum levels of apolipoprotein A1 and lipoprotein lipase, while apolipoprotein CIII and complement C3 levels were decreased(Reference Al-Daghri, Torretta and Capitanio20). In the present study, twenty proteins were found to correlate either positively or negatively with serum 25(OH)D levels in individuals with concentrations < 50 nmol/l. Although some apolipoproteins and complement factors such as Apolipoprotein A-I, Apolipoprotein M, Complement C4-A, Complement Factor I showed significant correlations (Table 2), gene ontology analysis did not identify any significantly enriched biological pathways in this category. The mean 25(OH)D concentration in this group was 38·73 (sd 9·52) nmol/l, indicative of mild insufficiency rather than severe vitamin D deficiency. According to Canadian guidelines, serum levels < 30 nmol/l are classified as deficient and are more likely to trigger proinflammatory responses and activate immune-related pathways(Reference Szymczak-Pajor and Śliwińska10,Reference Szymczak-Pajor, Drzewoski and Śliwińska11) . It is therefore possible that the average serum 25(OH)D concentration of 38 nmol/l in the < 50 nmol/l group of this study may not have been sufficiently low to show a detectable inflammatory pathway enrichment at the proteomic level.
In contrast, for subjects exhibiting 25(OH)D adequate levels (50–74·9 nmol/l), 25(OH)D was positively correlated with proteins involved in innate immune activation and negatively correlated to proteins of the adaptive immune response. Innate immunity, the body’s first line of defense, provides a rapid but non-specific response to pathogens through mechanisms such as physical barriers, complement activation and the acute-phase response(Reference Wei and Christakos37,Reference Jones, D’Vaz and Meldrum38) . In this context, complement activation leads to a cascade of protein interactions that promote pathogen elimination and inflammation. Additional innate pathways, such as neutrophil chemotaxis and the acute-phase response, are also activated resulting in the recruitment of immune cells such as monocytes and macrophages and increased production of pro-inflammatory proteins(Reference Jones, D’Vaz and Meldrum38). Conversely, the adaptive immune response pathways dependent on T and B cell activation and the development of immunologic memory correlated negatively with serum 25(OH)D levels. Recent evidence from healthy adults has shown that 25(OH)D levels ≥ 50 nmol/l are associated with reduced interleukin-6 (IL-6) and macrophage-derived chemokine, while levels near 70 nmol/l enhance regulatory T cell activity and suppress interferon-γ-inducible protein 10, indicating changes in the adaptive immune regulation(Reference Garcia-Bailo, El-Sohemy and Haddad39). Overall, the findings suggest that even at adequate levels, 25(OH)D did not correlate with proteins associated with adaptive immunity in hyperinsulinaemic subjects.
Finally, in subjects exhibiting serum 25(OH)D levels above 75 nmol/l, several serum proteins of the adaptive immune response correlated positively with this vitamin, while proteins involved in the innate immune response and the coagulation pathways were correlating negatively with 25(OH)D. Although the exact serum threshold of vitamin D for optimal immune function is still uncertain, recommendations generally suggest maintaining 25(OH)D levels at or above 75 nmol/l with optimal benefits observed between 100 and 150 nmol/l(Reference Dennis, Sherman and Hosack35). A prospective study involving infants at high risk of atopy found that higher serum 25(OH)D levels at birth (≥ 75 nmol/l) were associated with lower T-helper cell type 2 cytokine responses as an adaptive immune response and a reduced incidence of eczema by 6–12 months of age(Reference El-Hajj Fuleihan, Bouillon and Clarke40). The positive correlation of protein involved in adaptive immune with serum 25(OH)D levels in the groups of individual exhibiting ≥ 75 nmol/l of 25(OH)D may reflect a shift from immediate innate inflammatory defenses towards more adapted and specific immune mechanisms.
The negative regulation of blood coagulation pathway, an important defense mechanism to prevent excessive clotting during immune responses, was positively associated with 25(OH)D in the group of individuals with ≥ 75 nmol/l of this vitamin. This aligns with previous studies reporting that optimal 25(OH)D levels reduce pro-thrombotic activity and support vascular protection(Reference Gasque41,Reference Farooqui, Chavis and Ghouse42) . These prior reports associated higher 25(OH)D levels (more than 60 nmol/l) with reduced acute-phase proteins such as C-reactive-protein and retinol-binding-protein 4(Reference Riazati, Engle-Stone and Stephensen43,Reference Bellia, Garcovich and D’Adamo44) . The active form of vitamin D, 1,25-dihydroxyvitamin D3, binds to the vitamin D receptor, which interacts with specific regions in cytokine-encoding genes (vitamin D response elements), leading to reduced gene transcription and suppression of proinflammatory pathways that stimulate the production of acute-phase proteins by the liver(Reference Garcia-Bailo, El-Sohemy and Haddad39). Thus, optimal levels of 25(OH)D (≥ 75 nmol/l) may be required to promote long-term immune health.
To the best of our knowledge, this is the first study examining the correlation between serum levels of 25(OH)D and proteins after high dairy intake compared to adequate dairy intake. The strengths of the study included adjustment for essential confounders and stratifications by serum 25(OH)D. However, there are some limitations that need to be considered. First, small sample size within subgroups, particularly in the ‘Optimal’ (n 6) and ‘Insufficient’ (n 9) categories, may limit statistical power and increase the risk of false-positive findings in pathway enrichment analyses. Therefore, the identified biological pathways should be considered exploratory, requiring validation in larger cohorts. Second, no significant seasonal variation in serum 25(OH)D concentrations was observed, despite the Canadian setting where pronounced fluctuations in sunlight exposure are typically expected(Reference Shoben, Kestenbaum and Levin45). The widespread fortification of foods with vitamin D in Canada, along with potential use of vitamin D supplements, may have attenuated seasonal differences in circulating 25(OH)D levels(Reference Greene-Finestone, Berger and de Groh46). The relatively modest sample size may also have limited the statistical power to detect subtle seasonal fluctuations. Nevertheless, the absence of marked seasonal variation may limit the generalisability of these findings to populations with greater heterogeneity or more pronounced seasonal variation in vitamin D status. Third, all participants were overweight or obese adults with hyperinsulinaemia, which limits how the results can be applied to other groups, especially people with different health conditions involving abnormal oxidative stress or chronic inflammation. Further, the duration of the study was relatively short. Since 25(OH)D serum levels may take more time to change, a longer period of higher dairy intake might be needed to see clearer effects on vitamin D status and its relationship with serum proteins. Finally, although the abundance of several proteins may correlate with serum 25(OH)D levels, it remains unclear whether the functions or enzymatic activities of these proteins are similarly upregulated or downregulated in association with vitamin D status.
Conclusion
Findings from this exploratory proteomic analysis suggest that higher serum 25(OH)D levels were associated with differential enrichment of proteins involved in immune and inflammatory pathways. Although these associations are correlative and do not imply causality, the identified proteins and pathways may represent potential candidates for future validation studies investigating their relevance to T2D-related metabolic and inflammatory processes.
Supplementary material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114526108009
Acknowledgements
The authors sincerely thank all study participants for their time and commitment to this research. We are also grateful to Sarah O’Connor, Andréa Taschereau-Charron, Sarah Chouinard-Castonguay, Élise Cant, Valérie-Ève Julien and Camille Lambert for their valuable assistance with participant visits. We further thank Prof. Arnaud Droit and his team from the Proteomics Platform at the Centre de recherche du CHU de Québec–Université Laval, Faculty of Medicine, for their expertise in proteomics laboratory work.
The funding for this study was provided by various sources, including the Canadian Institutes of Health Research (CIHR), grant numbers PJT-364008 (awarded to I.R.) and PJT-173398 (awarded to M. L.). Additional financial support was obtained from the New Initiative Funds allocated by the Endocrinology/Nephrology axis at the Centre de Recherche du CHU de Québec (M. L., I. R.) A. M. received a Pierre J. Durand Doctoral Recruitment Scholarship from Université Laval, Grand Bâtisseur Fernand-Labrie from Fondation du CHU de Québec, and Fonds de recherche du Québec – Nature et technologies (FRQ-NT). This work was supported by the FRQ through the research centre grant for the CHU de Québec-Université Laval Research Center (reference: 30641).
A. M.: Formal analysis-Equal, Writing – original draft-Equal, Writing – review & editing-Equal; M. L.: Conceptualisation-Equal, Data curation-Equal, Formal analysis-Equal, Funding acquisition-Equal, Methodology-Equal, Resources-Equal, Supervision-Equal, Visualisation-Equal, Writing – original draft-Equal, Writing – review & editing-Equal; I. R.: Conceptualisation-Equal, Data curation-Equal, Formal analysis-Equal, Funding acquisition-Equal, Methodology-Equal, Resources-Equal, Supervision-Equal, Writing – original draft-Equal, Writing – review & editing-Equal.
The authors declare no conflicts of interest.
Mass spectrometry data have been deposited in the ProteomeXchange repository (https://www.proteomexchange.org) and are publicly available under accession number PXD035386. The clinical and anthropometric data that support the findings of this study, however, are not publicly available due to confidentiality and privacy concerns. As the dataset contains sensitive information about individuals, such as medical or personal data, it is essential to protect the privacy of participants. Additional inquiries can be directed to the corresponding authors (iwona.rudkowska@crchudequebec.ulaval.ca, michel.lebel@crchudequebec.ulaval.ca).
Open access
