Bolus Weekly Vitamin D₃ Supplementation Impacts Gut and Airway Microbiota in Adults With Cystic Fibrosis: A Double-Blind, Randomized, Placebo-Controlled Clinical Trial

Abstract

Context:

Disruption of gut microbiota may exacerbate severity of cystic fibrosis (CF). Vitamin D deficiency is a common comorbidity in patients with CF that may influence composition of the gut microbiota.

Objectives:

Compare microbiota of vitamin D-sufficient and -insufficient CF patients and assess impact of a weekly high-dose vitamin D₃ bolus regimen on gut and airway microbiome in adults with CF and vitamin D insufficiency (25-hydroxyvitamin D < 30 ng/mL).

Design:

Forty-one subjects with CF were classified into two groups: vitamin D insufficient (n = 23) and vitamin D sufficient (n = 18). Subjects with vitamin D insufficiency were randomized to receive 50,000 IU of oral vitamin D₃ or placebo weekly for 12 weeks. Sputum and stool samples were obtained pre- and postintervention and 16S ribosomal RNA genes sequenced using Illumina MiSeq technology.

Results:

Gut microbiota differed significantly based on vitamin D status with Gammaproteobacteria, which contain numerous, potentially pathogenic species enriched in the vitamin D-insufficient group. Principal coordinates analysis showed differential gut microbiota composition within the vitamin D-insufficient patients following 12 weeks treatment with placebo or vitamin D₃ (permutation multivariate analysis of variance = 0.024), with Lactococcus significantly enriched in subjects treated with vitamin D₃, whereas Veillonella and Erysipelotrichaceae were significantly enriched in patients treated with placebo.

Conclusion:

This exploratory study suggests that vitamin D insufficiency is associated with alterations in microbiota composition that may promote inflammation and that supplementation with vitamin D has the potential to impact microbiota composition. Additional studies to determine the impact of vitamin D on microbiota benefit clinical outcomes in CF are warranted.

This randomized, double-blind, placebo-controlled study in adults with cystic fibrosis demonstrates that correction of vitamin D deficiency favorably alters gut and airway microbiota.

Cystic fibrosis (CF) is a chronic disease, principally characterized by abnormal mucosa in the digestive and respiratory systems as a result of a mutation in the CF transmembrane conductance regulator (CFTR) gene. The CFTR gene encodes an ion channel that regulates chloride transport on the epithelial surface of the airway, pancreas, and gastrointestinal tract; hence, CFTR mutations result in impaired clearance of secretions in the lungs, decreased bicarbonate secretion from the pancreas, and malabsorption in the gastrointestinal tract (1). As a result, patients with CF develop recurrent sinus and pulmonary infections, CF-related diabetes as a result of chronic scarring and destruction of the pancreas, fat malabsorption, and eventually, respiratory decline and failure (2, 3). There is evolving evidence that the chronic proinflammatory intestinal environment in individuals with CF leads to and/or is a result of intestinal dysbiosis (disruption of the microbiota) with a substantial decrease in microbial diversity compared with healthy controls (4–6). Likewise, studies show that the airway microbiota is also affected in CF, with alterations beginning in childhood (7), showing a decrease in microbial diversity associated with advancing age and declining lung function (8, 9). In addition, there appear to be changes in specific bacterial groups that herald the onset of pulmonary exacerbations and initial Pseudomonas species colonization (10). Therefore, intestinal dysbiosis may have important implications for clinical symptoms and management in CF. Burke et al. (6) found a correlation among percent-predicted forced expiratory volume in 1 second (FEV_1%) and reduced gut microbial diversity, reduced α diversity (a measure of species richness and even-ness and a marker of a healthy gut microbiota), and lower relative abundance of beneficial bacteria, such as Roseburia, in those with most severe lung disease.

Patients with CF commonly develop vitamin D insufficiency for myriad reasons, including fat malabsorption, decreased vitamin D intake, and decreased sunlight exposure (11). An important extraskeletal function of vitamin D is its role in immune regulation, such as upregulation of antimicrobial peptide expression (cathelicidins and B-defensins) and inhibition of effector T cell Th1 and Th17 function (12–14). Recent studies in murine models identified vitamin D metabolism as an important factor influencing the gut microbiota, through the vitamin D receptor, which is abundantly expressed in the ileum (15, 16). Vitamin D receptor is proposed to maintain the integrity of the intestinal mucosal barrier through the inhibition of inflammation-induced epithelial cell apoptosis and enhancement of intercellular junctions (17, 18). Recent studies have demonstrated that vitamin D supplementation may be important to maintain normal gut microbial homeostasis in healthy subjects with a reduction in typical opportunistic pathogens and an increase in species richness (a parameter associated with a healthy gut microbiota) (19, 20).

Although there is robust evidence to support the hypothesis that microbial dysbiosis is predominant in individuals with CF, clinical significances of such dysbiosis are not well understood. Moreover, there have been no studies, to date, to evaluate if vitamin D insufficiency predisposes CF patients to a greater dysbiosis. The aim of the current study was to evaluate the role of vitamin D insufficiency on the airway and gut microbiota in patients with CF and explore the association with pulmonary outcomes. Furthermore, with the use of a randomized placebo-controlled trial, we sought to explore our hypothesis that supplementation with high-dose vitamin D might restore a more healthful microbiota and thus, improve clinical outcomes in patients with CF and vitamin D insufficiency.

Materials and Methods

The study was approved by the Emory University Institutional Review Board. Patient enrollment began in November 2015; follow-up for the last patient was completed in February 2017. The study was registered at www.clinicaltrials.gov (NCT02589444). Potential study subjects were identified among those presenting for routine visits at the CF Clinic at the Emory University Cystic Fibrosis Center, and interested subjects were recruited after written, informed consent. Subject enrollment and allocation are outlined in the Consolidated Standards of Reporting Trials (CONSORT) diagram (Fig. 1).

Figure 1.

CONSORT diagram showing recruitment and allocation of subjects. IBD, inflammatory bowel disease.

Trial design

This was a double-blind, randomized, placebo-controlled, interventional pilot study in 41 adults with CF. Patients with CF who were ≥18 years of age without contraindication to oral high-dose vitamin D met inclusion criteria. Exclusion criteria included the following: 1) use of immunosuppressants, 2) pregnancy or plans to become pregnant in the next 3 months, 3) disorders associated with hypercalcemia, 4) current hypercalcemia (albumin-corrected serum calcium >10.8 mg/dL, or ionized calcium >5.2 mg/dL), 5) history of nephrolithiasis with active symptoms within the past 2 years, 6) chronic kidney disease worse than stage III (<60 ml/min), 7) current substantial hepatic dysfunction total bilirubin >2.5 mg/dL, with direct bilirubin >1.0 mg/dL, 8) history of HIV/AIDS, 10) history of illicit drug abuse. In addition, we amended our study with three additional criteria: 11) systemic antibiotic use in the last 4 weeks, 12) use of probiotics, and 13) inflammatory bowel disease, 4 months after the start of the study and after 12 subjects were randomized, as we considered that these factors may also influence our study endpoints. Of the 12 subjects who were randomized, only 4 would have been excluded.

Intervention

Subjects who were screened and enrolled were classified into two groups: those with serum 25-hydroxyvitamin D [25(OH)D] level < 30 ng/mL (vitamin D insufficient) and those with serum 25(OH)D levels ≥ 30 ng/mL (vitamin D sufficient), based on levels obtained at baseline and per the CF Foundation Vitamin D Guidelines (21). Baseline assessment included demographic data, weight, height, CF mutation, history of current medications, history of pulmonary exacerbations (defined as new or worsening respiratory signs or symptoms reported by the subject or picked up by the clinician, requiring start of oral or intravenous antibiotics), hospitalizations in the past 3 months, and spirometry, including FEV₁% and forced vital capacity for measurement of lung function, performed as routine standard of care for CF outpatients.

Subjects who were vitamin D sufficient (n = 18 subjects) were followed longitudinally for 12 weeks and did not receive any further intervention. Subjects who were vitamin D insufficient (n = 23 subjects) were randomly assigned to either receive placebo or 50,000 IU of oral vitamin D₃, once weekly for 12 weeks, in random blocks of two or four by the Investigational Drug Pharmacy. The study drug and placebo were dispensed by the Investigational Drug Pharmacy and mailed to the subjects’ residence. The study team, clinician, and subjects were blinded to the treatment assignment until trial completion. Randomized subjects were asked to limit any additional supplemental vitamin D to <3000 IU/day and could continue their usual vitamin D supplementation at the time of screening and randomization. Participants were seen for study follow-up visits at 12 weeks in the clinic or at the Clinic Research Center at Emory University Hospital of the Atlanta Clinical and Translational Science Institute (±4 weeks). Three subjects had scheduled follow-up visits sooner than 12 weeks. We asked those subjects to take an additional vitamin D or placebo pill per week to ensure that all 12 study pills were completed before the follow-up visit. Follow-up data collection included history of medications and antibiotic use, in addition to any pulmonary exacerbations and hospitalizations in the interim.

Weekly phone calls were made to the randomized participants after week 1 as a reminder to take weekly study pills. Symptoms of vitamin D toxicity (excessive thirst, frequent urination, constipation, and confusion) were assessed by patient questionnaire serially and at the final study visit.

Sample collection and storage

Collection of stool and sputum samples for characterization of gut and airway microbiota was done at baseline and at the follow-up visit at 12 weeks (±4 weeks) for the entire cohort. Participants were asked to provide noninduced expectorated sputum into a collection container on the day of the study visit, and it was immediately stored at −80°C for future airway microbiota analysis. For collection of stool samples, subjects were provided with an OMNIgene-GUT stool kit (DNA Genotek, Ottawa, ON, Canada) and a prepaid envelope to have the sample mailed directly to the study laboratory within the next 48 hours. Each sample was homogenized, and three, 200 mg aliquots were placed in sterile Eppendorf tubes and stored at −80°C until future microbiota analysis. OMNIgene-GUT has been shown to stabilize DNA for microbiome profiling and minimize bias in microbial composition compared with methods requiring immediate refrigeration at −20°C (22).

Analytical methods

Serum 25(OH)D was measured using a chemiluminescent-based, automated method (IDS-iSYS; Immunodiagnostic Systems, Scottsdale, AZ). To ensure accuracy of the serum 25(OH)D measurements, our laboratory participates in the Vitamin D External Quality Assessment Scheme (site 606) and the National Institute of Standards and Technology/National Institutes of Health Vitamin D Metabolites Quality Assurance Program.

16S Ribosomal RNA gene sequencing and processing

OMNIgene-GUT-preserved samples were prepared for 16S ribosomal RNA (rRNA) gene amplification and sequencing using MiSeq technology (Illumina, San Diego, CA), following the protocol of the Earth Microbiome Project (www.earthmicrobiome.org/emp-standard-protocols), with its modifications to the PowerSoil DNA Isolation Kit (Qiagen, Germantown, MD) procedure for extracting DNA, as previously done (23, 24). DNA was extracted from frozen feces using a PowerSoil-htp Kit (Qiagen) with mechanical disruption (bead beating). The V4 region of the 16S rRNA genes was amplified with polymerase chain reaction (PCR) from each sample using the composite forward primer and the reverse primers 515FB-806RB (23, 25). PCR reactions consisted of 5PRIME HotMasterMix (Quantabio, Beverly, MA), 0.2 μM of each primer, and 10 to 100 ng template, and reaction conditions were 3 minutes at 95°C, followed by 30 cycles of 45 seconds at 95°C, 60 seconds at 50°C, and 90 seconds at 72°C on a thermocycler (Bio-Rad Laboratories, Hercules, CA). Two independent PCRs were performed for each sample and then combined and purified with Agencourt AMPure magnetic purification beads, and products were visualized by gel electrophoresis. Products were then quantified (fluorescence spectrophotometer; BioTek, Winooski, VT), and a master DNA pool was generated from the purified products in equimolar ratios. The pooled products were quantified and then sequenced using an Illumina MiSeq sequencer (paired-end reads, 2 × 250 bp) at Cornell University (Ithaca, NY). Sequences were demultiplexed and quality filtered using the Quantitative Insights Into Microbial Ecology (QIIME; version 1.8.0) software package (26). Forward and reverse Illumina reads were joined using the fastq-join method. We used the QIIME default parameters for quality filtering, as described in detail by Caporaso et al. (26) [reads truncated at the first low-quality base and excluded if 1) there were more than three consecutive low-quality base calls, 2) <75% of read length was consecutive high-quality base calls, 3) at least one uncalled base was present, 4) >1.5 errors were present in the bar code, 5) any Phred qualities were below 20, or 6) the length was <75 bases]. Sequences were clustered using the UCLUST algorithm (27) (i.e., a 97% threshold of pairwise identity) and assigned to operational taxonomic units using the Greengenes reference database 13_8. (28). New clusters were created with sequences that did not match any reference sequences. A single representative sequence for each operational taxonomic unit was aligned, and a phylogenetic tree was built using FastTree (29). The phylogenetic tree described previously was used to assess beta and alpha diversity. Unweighted UniFrac distances between samples were computed, as done previously, to measure beta diversity (30, 31). Principal coordinates analysis plots were used to assess and visualize beta diversity further. Linear discriminate analysis (LDA) effect (LEfSE) was used to investigate bacterial members that drive differences between groups by comparing the abundance of specific taxa between each experimental group (32). Groups were compared for distinct clustering using the permutation multivariate analysis of variance (permanova) method through QIIME. Unprocessed sequencing data are deposited in the European Nucleotide Archive under Accession Number PRJEB23120.

Outcome measures

The primary outcome was the change in phylotype richness of the airway and gut microbiota before and after supplementation with high-dose vitamin D₃. Secondary endpoints included serum 25(OH)D levels and lung function, as measured by FEV₁% in the 3 months preceding treatment and after completion of vitamin D₃ supplementation or placebo.

Statistical analysis

Patients were analyzed according to the study group to which they were assigned. Continuous variables were presented as the means with the standard deviation (SD). Normality assumptions were checked by Shapiro-Wilk tests. Differences among the control, placebo, and vitamin D groups, respectively, for continuous variables were assessed with one-way analysis of variance with Tukey methods or Kruskal-Wallis, Student’s t, or Mann-Whitney U tests, as appropriate. Categorical variables were presented as the percentage of frequency and differences among groups were assessed using the Fisher exact test. An unadjusted general linear regression model and a linear model adjusted by baseline serum 25(OH)D levels were used to evaluate the association between vitamin D treatment and the absolute change in serum 25(OH)D concentrations.

Results

Subject protocol and recruitment

Out of the 211 subjects screened, a total of 52 subjects were eligible and provided consent for participation in the study (Fig. 1). Of the 52 subjects who provided consent, 11 subjects were excluded, as a result of lack of collection of a baseline stool sample. Of the total of 41 eligible subjects, 23 were classified as vitamin D insufficient, and 18 subjects were classified as vitamin D sufficient. The 23 vitamin D-insufficient subjects were randomized to receive vitamin D₃ or placebo. One patient withdrew from the study in the placebo group, and one patient was lost to follow-up in each of the placebo and vitamin D₃ groups, respectively. All other randomized patients received the study medication, and 38 patients were included in the final analysis (Fig. 1; CONSORT diagram).

Baseline characteristics and vitamin D status

Baseline clinical and demographic characteristics were comparable across all three groups (Table 1). The majority of vitamin D-insufficient patients were homozygous for the ΔF508Del allele, and all had pancreatic insufficiency. By study design, the vitamin D-sufficient group had higher mean serum 25(OH)D concentration compared with the vitamin D-insufficient group. There was no difference in baseline mean serum 25(OH)D levels in the vitamin D-insufficient groups randomized to vitamin D or placebo (Table 1). Vitamin D-insufficient subjects who were randomized to vitamin D had a higher absolute change of 25(OH)D concentrations compared with the subjects randomized to placebo [P = 0.02 (unadjusted)], and this remained significant (P = 0.03) in a multivariate linear regression mode adjusted for baseline 25(OH)D level (Fig. 2).

Table 1.

Baseline Characteristics of Study Participants

Variable Name	Vitamin D Sufficient Control (n = 18)	Vitamin D Insufficient Randomized to Placebo (n = 10)	Vitamin D Insufficient Randomized to Vitamin D3 (n = 10)	P Valuea	P Valueb
Age, years, means ± SD	43 ± 17	32 ± 11	34 ± 10	0.18c	0.35d
Race, n (%)
White	17 (94%)	7 (70%)	10 (100%)	0.16e	0.21e
Hispanic	0 (0%)	1 (10%)	0 (0%)
Black	1 (6%)	1 (10%)	0 (0%)
Others	0 (0%)	1 (10%)	0 (0%)
Sex, n (%)
Male	9 (50%)	8 (80%)	6 (60%)	0.31e	0.58e
BMI, means ± SD; n,%	23 ± 3	22 ± 3	23 ± 6	0.99f	0.99d
Meet CF nutrition target (BMI ≥ 22 in female; ≥ 23 in male)	10 (56%)	5 (50%)	4 (40.00%)
CFRD status, n (%)
CFRD	2 (11%)	3 (30%)	1 (10%)	0.51g	0.58e
Pancreatic insufficiency, n (%)
Yes	18 (100%)	10 (100%)	10 (100%)	NA	NA
CF mutation
ΔF508 homo	5 (28%)	7 (70%)	4 (40%)	0.01g	0.006e
ΔF508 hetero	8 (44%)	2 (20%)	1 (10%)
Others	1 (56%)	1 (10%)	5 (50%)
Unknown	4 (22%)	0 (0%)	0 (0%)
Serum 25(OH)D level, ng/ml, means ± SD	37 ± 6	22 ± 6	25 ± 5	<0.001f	0.22d
Vitamin D supplementation, IU/day, means ± SD	3519 ± 2418	1770 ± 1643	1100 ± 849	0.001c	0.56d

Abbreviation: CFRD, CF-related diabetes.

^aP value comparing all three groups.

^bP value comparing the 25(OH)D-insufficient group randomized to placebo with weekly high-dose vitamin D₃.

^cKruskal-Wallis test.

^dMann-Whitney U test.

^eFisher exact test.

^fOne-way analysis of variance Tukey method.

^gFreeman-Halton test.

Figure 2.

Mean serum 25(OH)D in response to weekly vitamin D₃ treatment or placebo in adults with CF and vitamin D insufficiency. Means with standard error bars of serum 25(OH)D concentrations between the placebo group (red; n = 10) and treatment group (blue; n = 10) at baseline and final visit at 12 weeks. Subjects who were randomized to oral vitamin D₃ 50,000 IU had higher serum 25(OH)D concentrations compared with the subjects randomized to placebo [P = 0.02 (β = 16.81)], and this remained significant when adjusted by baseline serum 25(OH)D (β = 17.08, P = 0.03).

As expected, the mean daily oral habitual vitamin D intake in the vitamin-insufficient group was lower [1435 IU/d ± 1300 compared with 3500 ± 200 IU/d in the vitamin D-sufficient group (P = 0.0002)]. All but two subjects in the vitamin D-insufficient group, who were randomized to placebo, maintained serum 25(OH)D concentrations <30 ng/mL at the end of 12 weeks of intervention.

Lung function

Before randomization, the vitamin D-insufficient subjects receiving placebo or vitamin D₃ had similar FEV₁% values (Supplemental Table 1 ^{(574.2KB, docx)} ). There was no statistically significant change in the pulmonary outcomes (as measured by FEV1%) postintervention of adult CF patients receiving 50,000 IU vitamin D₃ weekly vs placebo. When adjusted for age, sex, race, CF genetic mutations, and smoking, there continued to be no significant difference in the two groups.

Gut and airway microbiota composition of individuals with CF based on vitamin D status before intervention

Gut microbiota composition analysis of the subjects with CF revealed substantial alterations between subjects with vitamin D sufficiency and vitamin D insufficiency. Taxa belonging to the class Gammaproteobacteria were substantially enriched in subjects with vitamin D insufficiency compared with subjects with vitamin D sufficiency; LDA score = 4.96 by LEfSE, whereas Bacteroidia class was enriched in subjects with vitamin D sufficiency patients (Fig. 3A and 3B). The analysis of upper-airway microbiota composition showed that initial serum 25(OH)D concentration level correlated with sputum microbiota alterations with differential clustering based on vitamin D status (Supplemental Fig. 1A ^{(574.2KB, docx)} ). Several taxa were enriched in the sputum samples of subjects with vitamin D insufficiency compared with samples from subjects with vitamin D sufficiency at baseline. Of particular interest, the sputum samples of subjects with vitamin D insufficiency were enriched in members of the genus Bacteroides (Supplemental Fig. 1A and 1B ^{(574.2KB, docx)} ).

Figure 3.

Differentially abundant taxa, according to vitamin D status of the subject at baseline. (A) LEfSE showing genera and species that were significantly altered in the stool samples of vitamin D-sufficient subjects (red) compared with vitamin D-insufficient subjects (green) at baseline. (B) Potentially pathogenic taxa belonging to the class Gammaproteobacteria were significantly more abundant in the stool samples of vitamin D-insufficient subjects compared with vitamin D-sufficient subjects at baseline (LDA score = 4.96).

Changes in gut microbiota in response to vitamin D supplementation in vitamin D-insufficient subjects with CF

Subjects who were randomized to once weekly 50,000 IU of oral vitamin D₃ had significantly increased serum 25(OH)D concentrations compared with subjects randomized to placebo at the end of the study period (Fig. 2). This increase in serum 25(OH)D levels in the vitamin D₃ intervention group concentration was associated with a shift in the gut microbiota with differential clustering at the end of 12 weeks of intervention compared with the placebo group (Fig. 4A and 4B). Permanova = 0.723 for Fig. 4A, and permanova = 0.024 for Fig. 4B, indicating statistically significant clustering postintervention. To determine microbial species driving the difference between the two groups, we performed LEfSE (Supplemental Fig. 2A ^{(574.2KB, docx)} ). The change in species abundance following 12 weeks of vitamin D or placebo groups presented in Fig. 5 demonstrate that Lactococcus was substantially increased, whereas Veillonella and Erysipelotrichaceae were substantially decreased after 12 weeks of vitamin D₃ supplementation (Fig. 5).

Figure 4.

Principal coordinate (PC) analysis based on unweighted UniFrac distance matrices generated with stool 16S rRNA gene sequencing. Stool samples from adults with CF and vitamin D insufficiency [25(OH)D < 30 ng/mL] receiving placebo (red) or receiving vitamin D₃ (blue) were analyzed at (A) baseline and (B) the end of 12 weeks of intervention of weekly 50,000 IU of vitamin D₃.

Figure 5.

Differentially abundant taxa in vitamin D-insufficient, randomized subjects based on treatment assignment. LEfSE showing genera and species substantially, differentially abundant in the stool samples from adults with CF and vitamin D insufficiency [25(OH)D < 30 ng/mL] after receiving placebo (red) or after receiving 50,000 IU of vitamin D₃ (green).

Changes in airway microbiota in response to vitamin D supplementation in vitamin D-insufficient subjects with CF

There was differential clustering of microbiota in subjects who were randomized to once weekly 50,000 IU of oral vitamin D₃ IU compared with subjects randomized to placebo at the end of the study period (Supplemental Fig. 2A ^{(574.2KB, docx)} ). The genera and species with significantly different abundance between the subjects with vitamin D insufficiency receiving placebo vs 50,000 IU vitamin D₃ are presented in Supplemental Fig. 2B ^{(574.2KB, docx)} .

Safety and adverse events

The maximum 25(OH)D concentration in an individual subject at 12 weeks was 72 ng/mL and 31 ng/mL in vitamin D and placebo groups, respectively. There were no reported symptoms of vitamin D toxicity, as assessed by the patient questionnaire at the final study visit. There were also no clinical signs of hypercalcemia.

Discussion

Prior limited data suggest that airway and gut microbiota in CF are disrupted as a result of multiple factors, including chronic inflammation in the respiratory and digestive tracts, malabsorption, and recurrent lung infection, and effects as a result of systemic and local antibiotics (33–35). Such dysbiosis may play a role in CF morbidity with an impact on nutrition and pulmonary outcomes (9, 10, 36).

We found that the gut microbiota of CF subjects was altered based on vitamin D status at baseline, and taxa belonging to the class Gammaproteobacteria were significantly enriched in the vitamin D-insufficient group compared with the vitamin D-sufficient group. Gammaproteobacteria are gram-negative bacteria comprising several pathogenic bacterial species, including Salmonella enterica, Yersinia pestis, Vibrio cholera, Escherichia coli, and Pseudomonas aeruginosa, the latter being of particular interest given its role in CF pulmonary infection (37). Similar observations were made by Bashir et al. (19), who found a decrease in the relative abundance of Gammaproteobacteria, particularly P. aeruginosa and Escherichia/Shigella species in the upper gastrointestinal tract following high-dose vitamin D₃ supplementation in healthy volunteers. In patients with chronic HIV infection, Gammaproteobacteria is enriched within the gut microbiota, and its relative abundance is positively associated with plasma interleukin-1β levels, an important mediator of the inflammatory response, suggesting that Gammaproteobacteria may have proinflammatory effects (38). Our finding of potentially pathogenic species, such as Gammaproteobacteria being enriched in vitamin D deficiency, support a role for vitamin D in ameliorating the chronic inflammation present in the CF gut.

Moreover, our study was designed to explore the effects of a high-dose vitamin D₃ bolus regimen on airway and gut microbiota in adult CF patients matched for age, sex, body mass index (BMI), and lung function. We found that treatment with vitamin D₃ influenced gut microbiota with differential clustering based on treatment assignment and a shift toward a potentially beneficial gut microbiota composition, whereas placebo did not lead to microbiota alterations. In addition, we found that homozygous ΔF508 mutation was significantly more prevalent in those subjects with vitamin D insufficiency, which is in contrast with the findings of Vanstone et al. (39), who found that CFTR mutations did not impact the serum 25(OH)D level.

In a cross-sectional study, looking specifically at the gut microbiota in children with CF, a lower abundance and temporal stability of Bifidobacterium species was demonstrated in patients with CF compared with their healthy siblings (5). Members of the Bifidobacterium species are considered as a marker of intestinal health and are commonly used in probiotic supplements. Likewise, in a study on adults with CF and stable lung disease, Burke et al. (6) observed a decrease in gut microbial diversity and suppression of potentially beneficial bacteria—Bifidobacterium and Akkermansia in CF patients compared with non-CF controls. Moreover, they found that when stratified by percent-predicted FEV1, individuals with CF with severe lung dysfunction (% predicted FEV₁ ≤ 40%) exhibited significantly reduced gut microbial diversity compared with patients with CF with mild or moderate dysfunction (6). In a longitudinal study examining gut microbial communities in infants with CF, Madan et al. (35) found statistically significant changes in the gut microbiota before the onset of the first pulmonary exacerbation and chronic P. aeruginosa colonization. These studies suggest that there is role for gut dysbiosis in CF to impact pulmonary outcomes. In contrast, we did not find any association between microbiota alterations and pulmonary function, as measured by FEV1%.

We found shifts in specific bacterial taxa in the gut microbiota from vitamin D-insufficient subjects receiving D₃ compared with the vitamin D-insufficient subjects receiving placebo at the end of 12 weeks of intervention. Of note, members of the genus Lactococcus, which have been positively associated with gut health, were enriched after 3 months of vitamin D₃ treatment compared with the baseline. Moreover, Veillonella at the genus level and Erysipelotrichaceae at the family level are potentially pathogenic bacteria and were found enriched in the vitamin D-insufficient patients receiving placebo compared with those receiving weekly 50,000 IU vitamin D₃. Specific taxa within Erysipelotrichaceae have been correlated to metabolic disorders and inflammation (38, 40). Veillonella species have been reported as a rare cause of serious infections, including bacteremia, meningitis, and pleuropulmonary infection (41). In a recent cross-sectional study looking at fecal microbiota based on vitamin D status in healthy subjects, Luthold et al. (20) also found Veillonella species to be relatively more abundant in individuals with the lowest intake and concentration of vitamin D compared with those with the highest intake and concentration of vitamin D. In our study, treatment with vitamin D₃ compared with placebo appeared to drive changes in both airway and gut bacterial communities with differential clustering compared with baseline based on treatment assignment. These findings suggest a potential role for vitamin D in altering the balance of symbionts to pathobionts and influencing the predominance of specific bacterial taxa, such as Gammaproteobacteria, which have a role in pulmonary outcomes in CF.

Analysis of airway microbiota showed that members of the genus Bacteroides were enriched at baseline in subjects with CF and vitamin D insufficiency. Bacteroides are anaerobic bacteria that are commensals in the gut and generally have a beneficial relationship with the host. However, they can cause substantial pathology, including bacteremia and abscess formation outside of the gastrointestinal tract (42). Given potential gut-barrier dysfunction in CF, it is possible that this finding is related to translocation of microbes from the gut lumen to the systemic circulation (43). There were many differences in the taxa based on treatment assignment in the placebo vs high-dose vitamin D₃ supplementation groups in the subjects with vitamin D insufficiency, as observed by LEfSe analysis. Of interest, Staphylococcus aureus and Staphylococcus epidermidis species were most significantly enriched in the placebo group with S. aureus, which has been implicated in pulmonary infection in CF and associated with a substantial inflammatory response in patients with CF (44). We also found that bacteria belonging to the genera Corynebacterium were significantly more abundant in the placebo group, consistent with a previous study showing Corynebacterium to be more abundant in the CF airway and as a cause of respiratory infections in children with CF (7, 45).

Microbial dysbiosis is a hallmark of the CF gut underlying to chronic mucosal inflammation. It is reasonable to postulate that vitamin D may mitigate the dysbiosis seen in the CF through its effect on intestinal mucosal inflammation (46). This theory is supported by the work of Morin et al. (47), who studied the effects of vitamin D₃ and its metabolites in CFTR knockdown intestinal epithelial cells. They observed that 1,25(OH)₂D₃ leads to an inhibition of interleukin-8 and reduces cytokine-induced nuclear factor κB nuclear translocation, thus resulting in a suppression of inflammatory mediators (47). Vitamin D has also been proposed to decrease intestinal inflammation through the inhibition of inflammation-induced epithelial cell apoptosis and reducing cytokine-induced nuclear factor κB nuclear translocation, thus resulting in a suppression of inflammatory mediators (48). Given these effects of vitamin D on the immune response, vitamin D deficiency creates an environment that favors the predominance of potentially pathogenic, pathobiont bacteria, such as Gammaproteobacteria, which were significantly enriched in the vitamin D-insufficient group compared with the vitamin D-sufficient group in our CF population. Thus, we propose that repletion of vitamin D insufficiency in CF patients may result in a shift toward potentially beneficial microbial communities by decreasing the competitive advantage of pathogenic bacteria.

Although our study cohort was relatively small, we were able to characterize the gut microbiota in individuals with CF, based on baseline vitamin D status, and show a substantial impact of vitamin D treatment on the gut microbiota in CF. Whereas our study was not designed to decipher cause-effect inter-relationships, our findings suggest a role for vitamin D in mitigating dysbiosis in CF by the enrichment or depletion of specific taxa. As this study was conducted in those with vitamin D insufficiency, it is uncertain whether vitamin D would continue to have a beneficial effect on the microbiota in patients who are considered vitamin D replete by current definitions.

The data from this study suggest a possible role for vitamin D in modulating the gut microbiota in CF. Vitamin D deficiency may predispose to a predominance of gram-negative bacteria of the class Gammaproteobacteria in the gut and Bacteroides in the airway microbiota. We found that the repletion of vitamin D insufficiency in CF patients may result in a shift toward commensal microbial communities in the gut, such as those belonging to the group Lactococcus. As this was an exploratory study, long-term studies are needed to confirm our findings and determine if they affect clinical outcomes in CF.