Abstract
Context:
Preclinical studies suggest that TNF-α suppresses PTH synthesis, inhibits renal 1α-hydroxylase activity, and impairs fibroblast growth factor 23 (FGF23) degradation. The impact of inflammation on vitamin D and mineral metabolism has not been well-characterized in Crohn's disease (CD).
Objective:
The objective of the study was to assess short-term changes in vitamin D-related mineral metabolism in CD after anti-TNF-α induction therapy.
Design/Participants:
Eighty-seven CD participants, aged 5–39 years, were assessed at the initiation of anti-TNF-α therapy and 10 weeks later.
Outcomes:
Indices of clinical disease activity and serum concentrations of vitamin D metabolites, vitamin D-binding protein (DBP), calcium, PTH, FGF23, IL-6, and TNF-α were measured at each visit. A multivariable generalized estimating equation (GEE) regression analysis was used to examine the correlates of PTH and 1,25-dihydroxyvitamin D [1,25(OH)2D] concentrations at each visit.
Results:
After anti-TNF-α therapy, cytokines and inflammatory markers [IL-6, TNF-α, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP)] concentrations decreased (all P < .0001), and PTH and 1,25(OH)2D concentrations increased (median 21 vs 30 pg/mL, P < .0001, and median 41.7 vs 48.1 pg/mL, P = .014, respectively). Levels of 25-hydroxyvitamin D [25(OH)D], 24,25-dihydroxyvitamin D, DBP, and FGF23 did not change. In GEE analyses, higher IL-6, TNF-α, ESR, and CRP were associated with lower PTH concentrations (all P < .001), adjusted for corrected calcium and 25(OH)D levels. Higher PTH was associated with higher 1,25(OH)2D concentrations (P < .001) at each visit, independent of 25(OH)D concentrations. Higher levels of all inflammatory markers were associated with lower 1,25(OH)2D concentrations (all P < .05). However, when PTH was added to these models, the inflammatory markers (with the exception of CRP) were no longer significantly associated with 1,25(OH)2D.
Conclusions:
Greater inflammation was associated with lower PTH and 1,25(OH)2D concentrations. After anti-TNF-α induction, PTH and 1,25(OH)2D concentrations increased without concomitant changes in 25(OH)D and FGF23, consistent with effects of inflammation on PTH and thereby renal conversion of 25(OH)D to 1,25(OH)2D.
Crohn's disease (CD) is an autoimmune condition of the gastrointestinal tract characterized by chronic inflammation and defective innate immune regulation of the gut microbiome. Most studies of vitamin D metabolism in CD focused on nutritional vitamin D deficiency (1–4). However, animal studies demonstrated myriad effects of inflammatory cytokines on vitamin D metabolism. For example, TNF-α, IL-6, and IL-1β activated the parathyroid calcium-sensing receptor (5, 6) and inhibited renal expression of the 1α-hydroxylase responsible for converting 25-hydroxyvitamin D [25(OH)D] to 1,25-dihydroxyvitamin D [1,25(OH)2D] (7). Furthermore, TNF-α inhibited Phex gene expression in a mouse model of colitis. Although not reported in this study, decreased fibroblast growth factor 23 (FGF23) proteolysis by the Phex endopeptidase could increase FGF23 levels (8). FGF23 is a key regulator of vitamin D metabolism: it inhibits PTH synthesis and the renal 1α-hydroxylase and induces the renal 24-hydroxylase enzyme responsible for catabolism of 25(OH)D and 1,25(OH)2D to 24,25-dihydroxyvitamin D [24,25(OH)2D] and 2,24,25-trihydroxyvitamin D respectively (9). Therefore, these multifactorial perturbations may result in reduced concentrations of circulating PTH and 1,25(OH)2D in systemic inflammatory diseases.
The majority of 25(OH)D and 1,25(OH)2D circulate bound to vitamin D-binding protein (DBP) with 10%–15% bound to albumin and less than 1% in their free forms. DBP not only transports vitamin D metabolites but also plays a key role in regulating the availability of 25(OH)D to monocytes (10) and dendritic cells (11). To our knowledge, DBP levels have not been reported in inflammatory bowel disease. We recently examined changes in vitamin D and PTH levels over a 3- to 4-year interval after CD diagnosis in 52 children and adolescents (12): CD was associated with low 25(OH)D and 1,25(OH)2D levels and a relative hypoparathyroidism at the time of diagnosis, compared with controls. As disease activity improved on therapy, PTH and 1,25(OH)2D levels increased significantly. More recently El-Hodhod et al (13) reported that FGF23 levels were elevated in children with inflammatory bowel disease during flares and decreased during remission. These studies were limited by heterogeneity in therapy and follow-up interval as well as a lack of concurrent measures of cytokines, PTH, FGF23, calcium, DBP, and vitamin D metabolites. Monoclonal antibodies targeting TNF-α are now a cornerstone of therapy for CD, resulting in rapid improvements in disease activity. The objectives of this study were to examine short-term changes in vitamin D and mineral metabolism in children and young adults after induction with anti-TNF-α therapy and to examine associations among measures of inflammation and vitamin D and mineral metabolism.
Materials and Methods
Study participants
CD patients, aged 5–40 years, who were initiating anti-TNF-α therapy were recruited from the inflammatory bowel disease centers at the Children's Hospital of Philadelphia and the Hospital of the University of Pennsylvania. Study visits were completed at the time of the first infusion of anti-TNF-α therapy (87% on the day of infusion, 13% 1–18 d prior) and 10 weeks later [median 72 d (interquartile range [IQR] 69, 80)]. A total of 97 participants enrolled. Three participants missing 25(OH)D, 1,25(OH)2D, and PTH at baseline were excluded from analysis. This report is limited to the remaining 87 that completed the 10-week visit; 85 were treated with infliximab and two with adalimumab. Participants were excluded for prior anti-TNF-α therapy; pregnancy; cognitive/developmental disorders that impacted their ability to complete the study procedures; and medical illness or therapies potentially affecting bone, nutrition or growth, including kidney disease, seizure disorder, and liver disease. Patients who were nonambulatory or planning to relocate in the coming year were also excluded. The study protocol was approved by the Institutional Review Board of the Children's Hospital of Philadelphia and the University of Pennsylvania. Informed consent was obtained from participants 18 years of age and older and from a parent or guardian of participants under the age of 18 years. Assent was obtained from participants 7–18 years of age.
CD characteristics
CD was confirmed by endoscopic, histological, and clinical parameters. CD activity was assessed at each study visit for those less than 21 years of age using the Pediatric Crohn's Disease Activity Index (PCDAI; score range 0–100) (14). Information on disease characteristics, medications, and sources of vitamin D supplementation were obtained by questionnaire and confirmed in the medical record. Participants were asked to report all nutrition supplements and to provide supplement bottles and/or brand names. A worksheet was used to determine the quantity of ergocalciferol or cholecalciferol (international units per day).
Anthropometry and race
Height was measured with a stadiometer and weight with a digital scale. Height and body mass index (BMI) (kilograms per square meter) Z-scores were calculated using national reference data (latter in participants ≤ 20 y of age) (15). Participant race was identified by the participant or guardian according to National Institutes of Health categories.
Laboratory measurements
Serum 25(OH)D and 1,25(OH)2D concentrations were measured by I125 RIA with a sensitivity of 1.5 ng/mL and less than 2 pg/mL, respectively (RIA; DiaSorin Inc) (16). Intraassay coefficients of variation (CVs) were 2.2% and 7%–11%, respectively (17). Serum 24,25(OH)2D was measured by combined tandem mass spectrometry (18) with an interassay CV of 8.6% and a sensitivity of 0.06 ng/mL. Serum DBP was measured in duplicate (CV < 10%) using an ELISA with a sensitivity of 3.74 ng/mL (R&D Systems). The interassay CV was 1.6%–3.6%, and recovery was 98%–103%. Serum intact PTH and calcium concentrations were measured by Quest Diagnostics (Nichols Institute) using a chemiluminescence assay (sensitivity 3 pg/mL and interassay CV of 7%–9%) and spectrophotometry, respectively; calcium was corrected for albumin (19). Intact FGF23 was measured via an ELISA (Kainos Laboratories Inc; sensitivity of 3 pg/mL and CV 6.7%–12.4%). Serum TNF-α and IL-6 were measured by the Luminex platform and the human cytokine six-plex high-sensitivity antibody bead kit (Millipore) with a sensitivity of 0.08 and 0.10 pg/mL and interassay variation of 8.3% and 7%, respectively. Other laboratory studies included serum hematocrit, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), albumin, and creatinine, performed using standard clinical techniques. For participants under 18 years of age, estimated glomerular filtration rate (eGFR) was calculated based on height and serum creatinine using the Schwartz formula (20). For participants 18 years of age and older, eGFR was calculated based on serum creatinine according to the Chronic Kidney Disease Epidemiology Collaboration formula (21).
Serum free 25(OH)D concentrations were calculated using total 25(OH)D, DBP, and albumin concentrations as previously described: free 25(OH)D = 25(OH)D/[1 + (6 × 105 × albumin) + (7 × 108 × DBP)] (22). Estimates using this method were highly correlated (R = 0.93) with free 25(OH)D concentrations measured by centrifugal ultrafiltration (22). Serum free 1,25(OH)2D concentrations were calculated according to the same method using specific affinity constants determined for 1,25(OH)2D (23).
Statistical analysis
All analyses were performed using STATA 13.0 (Stata Corp). A value of P < .05 was the criterion for statistical significance. Distributions of all continuous variables were assessed for normality. For normally distributed data, mean ± SD was reported and changes between baseline and 10 weeks assessed using the paired t test. For skewed data, median and IQR were reported and differences assessed using the Wilcoxon signed-rank test. Vitamin D deficiency was defined as a serum 25(OH)D concentration less than 20 ng/mL. Univariate and multivariable generalized estimating equation (GEE) regression analysis was used to evaluate correlates of PTH at each visit, including age, sex, race (black vs other), corrected calcium, vitamin D metabolites, FGF23, inflammatory markers (ESR and CRP), and cytokines (TNF-α and IL-6). Univariate and multivariable GEE regression analysis was used to examine correlates of 1,25(OH)2D concentrations, including age, sex, race, albumin, corrected calcium, 25(OH)D, PTH (in a univariate model), FGF23, inflammatory markers, and cytokines. PTH was subsequently added to the multivariable models for 1,25(OH)2D to determine whether adjustment for PTH attenuated the associations between inflammation and 1,25(OH)2D levels. Although this approach does not determine causality, substantial attenuation would suggest that inflammation decreases 1,25(OH)2D levels, at least in part, through inhibition of PTH. Skewed data were natural log transformed for the models.
Results
Participant characteristics and treatment (Table 1)
Table 1.
Participant Characteristics at Initiation of Anti-TNF-α Therapy
| Characteristic | Distribution |
|---|---|
| Age, y, n, % | |
| 5–9 | 12 (14) |
| 10–14 | 33 (38) |
| 15–20 | 35 (40) |
| ≥21 | 7 (8) |
| Male, n, % | 53 (61) |
| Black race, n, % | 15 (17) |
| Winter season (November-April), n, % | 56 (64) |
| Height Z-score, mean (SD) | −0.43 (1.01) |
| BMI Z-score, mean (SD)a | −0.17 (1.09) |
| BMI, median (IQR) | 23.4 (20.9, 24.4) |
| PCDAI, mean (SD)b | 28 (16) |
| PCDAI, n, % | |
| 0–10 (no disease) | 16 (21) |
| 11–30 (mild disease) | 34 (44) |
| >30 (moderate to severe disease) | 28 (36) |
| Site of disease, n, % | |
| Isolated upper gastrointestinal disease | 71 (82) |
| Isolated ileal disease | 5 (6) |
| Ileocolonic disease | 59 (68) |
| Isolated colonic disease | 22 (25) |
| Perianal involvement | 28 (32) |
| On systemic corticosteroid therapy, n, % | 26 (30) |
| On vitamin D supplement, n, % | 63 (74) |
| Total vitamin D dose, IU, among participants on supplement, median (IQR) | 462 (400, 800) |
| Serum creatinine, mg/dL, mean (SD) | 0.60 (0.16) |
| eGFR, mL/min per 1.73 m2, median (IQR) | 160 (133, 183) |
BMI Z-score reported for the 78 participants 20 years of age or younger and BMI reported for nine older than 20 years.
PCDAI was limited to participants younger than 21 years of age (n = 78; 2 missing).
Among the 85 participants treated with infliximab, the median dose was 5.0 mg/kg, and all participants received at least their second infusion at the time of the follow-up visit. Consistent with the standard dosing schedule, 79 participants (93%) received their third infusion of infliximab prior to or at the follow-up visit. One individual had a fourth infusion prior to follow-up due to a reaction during the third infusion. Eighty-three participants (95%) had a serum creatinine measured within 6 months of baseline, and all participants had normal renal function. At baseline 36% of the participants were deficient in 25(OH)D, and 74% of participants were receiving vitamin D supplementation.
Changes in disease activity, vitamin D-related mineral metabolism, and inflammation
Disease activity improved significantly over the 10-week interval. At baseline, 36% of the 78 pediatric participants were classified as moderate to severe, as compared with 1% at 10 weeks (P < .001), and the PCDAI score decreased significantly (P < .0001). Changes in PCDAI scores and laboratory parameters are summarized in Figure 1 and Table 2. The concentrations of cytokines and inflammatory markers all decreased significantly, whereas PTH and total and free 1,25(OH)2D concentrations increased significantly. Serum 25(OH)D (total and free), 24,25(OH)2D, DBP, and FGF23 concentrations did not change significantly. Serum albumin and total calcium increased significantly (P < .0001), but corrected calcium did not change. Findings were similar when limited to the 61 participants not on corticosteroid therapy at baseline: PTH increased (median 21 to 30 pg/mL, P < .0001), 1,25(OH)2D increased (median 42.2 to 48.6 pg/mL, P = .028), and all markers of inflammation and disease activity decreased (all P < .0001). The only difference in this subanalysis was that FGF23 increased (median 29.4 to 32.3 pg/mL, P = .041). Findings were also similar when limited to the 51 participants for whom season of visit (winter vs other) did not change over the study period: PTH increased significantly (21 to 31.5 pg/mL, P < .0001), and all markers of inflammation/disease activity decreased (all P ≤ .0001). The only difference was that although the magnitude of the increase in 1,25(OH)2D was comparable (43.6 to 48.6 pg/mL, P = .088), it was no longer statistically significant likely due to the smaller sample size. Limiting the study to pediatric participants did not impact our findings.
Figure 1.
Measures of inflammation and mineral metabolism before and after anti-TNF-α induction.
Table 2.
Measures of Disease Activity, Inflammation, and Mineral Metabolism at Baseline and After Anti-TNF-α Induction Therapy
| Baseline | 10 Weeks | P Valuea | |
|---|---|---|---|
| Inflammation and disease activity | |||
| PCDAIb | 27.5 (15, 40) | 10 (5, 15) | <.0001 |
| Erythrocyte sedimentation rate, mm/h | 24.5 (12, 40) | 9 (4, 15) | <.0001 |
| CRP, mg/dL | 1.1 (0.5, 2.6) | 0.5 (0.3, 0.5) | <.0001 |
| TNF-α, pg/mL | 7.2 (5.1, 10.3) | 1.7 (1.1, 2.6) | <.0001 |
| IL-6, pg/mL | 12.4 (6.0, 24.9) | 4.9 (2.6, 9.4) | <.0001 |
| Mineral metabolism | |||
| 25(OH)D, ng/mL | 25.2 (17.1, 35.0) | 26.2 (19.6, 33.1) | .50 |
| 1,25(OH)2D, pg/mL | 41.7 (31.6, 58.0) | 48.1 (38.2, 60.1) | .014 |
| 24,25(OH)2D, ng/mL | 3.1 (1.9, 4.5) | 3.1 (2.0, 4.4) | .72 |
| DBP, mg/dL | 18.9 (12.4, 25.4) | 19.8 (11.9, 24.2) | .20 |
| Albumin, g/dL | 3.8 ± 0.6 | 4.2 ± 0.4 | <.0001 |
| Free 25(OH)D, pg/mL | 10.5 (6.5, 14.4) | 10.3 (7.0, 14.8) | .88 |
| Free 1,25(OH)2D, pg/mL | 0.26 (0.21, 0.42) | 0.32 (0.25, 0.50) | .008 |
| Corrected calcium, mg/dL | 9.1 ± 0.4 | 9.1 ± 0.3 | .59 |
| PTH, pg/mL | 21 (13, 33) | 30 (20, 40) | <.0001 |
| FGF23, pg/mL | 29.4 (23.8, 37.2) | 31.8 (23.8, 40.8) | .22 |
Results presented as median (IQR) or mean ± SD as appropriate. Bold numbers are significant P values.
Paired t test or Wilcoxon signed-rank test as appropriate.
Limited to 78 participants younger than 21 years of age.
GEE regression analyses of PTH levels
In univariate GEE analysis, higher IL-6, TNF-α, ESR, and CRP (all P < .001) and greater PCDAI (P = .015 for PCDAI of 11–30 vs 0–10 and P < .001 for PCDAI > 30 vs 0–10) were associated with lower PTH concentrations. As expected, 25(OH)D (P = .026) and corrected calcium (P < .001) were inversely associated with PTH. In multivariable GEE analyses (Table 3), these inverse relationships between PTH and markers of inflammation and disease activity persisted, adjusted for corrected calcium and 25(OH)D concentration. A dose response was noted for PCDAI categories as more severe disease activity was associated with lower PTH concentrations. Being on vitamin D supplementation (≥400 IU/d) was not associated with PTH in multivariable GEE regression analysis. Adjustment for season did not impact model findings. To address whether the associations observed were influenced by corticosteroid exposure, the multivariable analyses were repeated limited to the 61 participants who were not on systemic corticosteroid therapy at baseline as well as to the 26 participants on corticosteroids at baseline: higher IL-6, TNF-α, ESR, and CRP, and PCDAI greater than 30 remained independent determinants of PTH (all P ≤ .014).
Table 3.
Multivariable GEE Regression Analyses of the Association Between Measures of Inflammation and PTH Concentrations
| Variablea | Difference in PTH,%b | 95% CI | P Value |
|---|---|---|---|
| Corrected calcium per 1 mg/dL | −40.9 | (−51.3, −28.4) | <.001 |
| 25(OH)D per 10% | −1.8 | (−3.2, −0.4) | .011 |
| TNF-α per 10% | −1.6 | (−2.1, −1) | <.001 |
| Corrected calcium per 1 mg/dL | −38.9 | (−51.3, −23.3) | <.001 |
| 25(OH)D per 10% | −2.1 | (−3.5, −0.7) | .003 |
| IL-6 per 10% | −1.6 | (−2.2, −0.9) | <.001 |
| Corrected calcium per 1 mg/dL | −30.5 | (−42.2, −16.4) | <.001 |
| 25(OH)D per 10% | −2.3 | (−3.6, −1) | .001 |
| ESR per 1 mm/h | −1.4 | (−1.8, −1.1) | <.001 |
| Corrected calcium per 1 mg/dL | −32.1 | (−44.1, −17.6) | <.001 |
| 25(OH)D per 10% | −1.7 | (−3.1, −0.3) | .015 |
| CRP per 1 mg/dL | −10.9 | (−15.2, −6.4) | <.001 |
| Corrected calcium per 1 mg/dL | −43.4 | (−53.9, −30.4) | <.001 |
| 25(OH)D per 10% | −2.0 | (−3.7, −0.2) | .027 |
| PCDAI vs 0–10 (no disease) | |||
| 11–30 (mild disease) | −18.5 | −31.2, −3.6) | .017 |
| >30 (moderate to serve disease) | −46.3 | −57.3, −32.5) | <.001 |
Bold values highlight main variables of interest in each model. Abbreviation: CI, confidence interval.
Each variable is expressed as a 10% difference for variables that are natural log transformed, or a 1 unit difference for variables that are untransformed in the GEE model.
PTH was natural log transformed in all models. The percentage difference in PTH was calculated as (1.1β − 1) × 100 for log-transformed variables and (eβ − 1) × 100 for untransformed variables, where β is the regression parameter for the variable of interest.
GEE regression analyses of 1,25(OH)2D Levels
In univariate GEE analysis, higher markers of inflammation (IL-6, TNF-α, ESR, and CRP) and moderate to severe disease activity (PCDAI > 30 vs 0–10) were associated with lower 1,25(OH)2D concentrations at each visit (all P ≤ .02), whereas higher concentrations of PTH (P = .008), 25(OH)D (P < .001), and albumin (P = .04) were associated with higher 1,25(OH)2D. Corrected calcium was not associated with 1,25(OH)2D. The positive association between PTH and 1,25(OH)2D concentrations at each visit persisted after adjustment for 25(OH)D concentrations (P < .001), whereas the association between albumin and 1,25(OH)2D concentrations did not. In models adjusted for 25(OH)D, higher levels of all inflammatory markers and moderate to severe disease activity were associated with lower 1,25(OH)2D concentrations at each visit (all P < .05). Being on vitamin D supplementation (≥400 IU/d) and season were not associated with 1,25(OH)2D in these models. However, with the exception of CRP, when PTH was added to these models, the inflammatory markers and PCDAI were no longer significantly associated with 1,25(OH)2D (Table 4).
Table 4.
Multivariable GEE Regression Analyses of Associations Among Measures of Inflammation, PTH, and1,25(OH)2D Concentrations
| Variablea | Difference in 1,25(OH)2D, %b | 95% CI | P Value | Difference in 1,25(OH)2D, % | 95% CI | P Value |
|---|---|---|---|---|---|---|
| 25(OH)D per 10% | 2.9 | (1.4, 4.4) | <.001 | 3.2 | (1.7, 4.7) | <.001 |
| PTH per 10% | PTH not included | 1.6 | (0.5, 2.7) | .005 | ||
| TNF-α per 10% | −0.7 | (−1.3, −0.2) | .013 | −0.5 | (−1.1, 0.1) | .098 |
| 25(OH)D per 10% | 2.7 | (1.2, 4.2) | <.001 | 3.1 | (1.5, 4.6) | <.001 |
| PTH per 10% | PTH not included | 1.5 | (0.4, 2.6) | .007 | ||
| IL6 per 10% | −0.8 | (−1.4, −0.2) | .008 | −0.6 | (−1.2, 0) | .061 |
| 25(OH)D per 10% | 2.8 | (1.1, 4.5) | .001 | 3.2 | (1.6, 4.9) | <.001 |
| PTH per 10% | PTH not included | 1.7 | (0.6, 2.8) | .002 | ||
| ESR per 1 mm/h | −0.4 | (−0.8, 0) | .042 | −0.2 | (−0.6, 0.3) | .44 |
| 25(OH)D per 10% | 3.0 | (1.4, 4.6) | <.001 | 3.3 | (1.7, 4.9) | <.001 |
| PTH per 10% | PTH not included | 1.4 | (0.3, 2.5) | .01 | ||
| CRP per 1 mg/dL | −5.8 | (−8.4, −3.2) | <.001 | −4.0 | (−7.3, −0.6) | .02 |
| 25(OH)D per 10% | 3.4 | (1.6, 5.4) | <.001 | 3.8 | (1.9, 5.7) | <.001 |
| PTH per 10% | PTH not included | 1.7 | (0.6, 2.9) | .004 | ||
| PCDAI vs 0–10 (no disease) | ||||||
| 11–30 (mild disease) | −2.2 | (−15.8, 13.6) | .77 | 0.4 | (−14.3, 17.7) | .96 |
| >30 (moderate to severe disease) | −26.3 | (−41.2, −7.6) | .008 | −16.9 | (−34.3, 5.2) | .12 |
Bold values highlight main variables of interest in each model. Abbreviation: CI, confidence interval.
Each variable is expressed as a 10% difference for variables that are natural log transformed, or a 1 unit difference for variables that are untransformed in the GEE model.
1,25(OH)2D was natural log transformed in all models. The percentage difference in 1,25(OH)2D was calculated as (1.1β − 1) × 100 for log-transformed variables and (eβ − 1) × 100 for untransformed variables, where β is the regression parameter for the variable of interest.
Discussion
This study of 87 pediatric and adult CD participants is the first to report acute changes in comprehensive measures of vitamin D-related mineral metabolism after anti-TNF-α induction therapy. As demonstrated by the dramatic declines in cytokines, inflammatory markers, and CD activity over the study period, TNF-α blockade has potent antiinflammatory actions in CD. Serum PTH and 1,25(OH)2D (total and free) concentrations increased significantly in the absence of changes in 25(OH)D, 24,25(OH)2D, DBP, or FGF23 concentrations. The stable 25(OH)D levels indicate that the increases in 1,25(OH)2D levels were not due to improved gut absorption of vitamin D. Furthermore, the demonstration that DBP levels did not change, that albumin was not independently associated with 1,25(OH)2D, and that both the total and free 1,25(OH)2D levels increased indicates that the increase in total 1,25(OH)2D levels was not due to increases in serum albumin and/or DBP levels. The observation that the associations of greater disease activity, cytokines, and inflammatory makers with lower 1,25(OH)2D levels were largely eliminated with an adjustment for PTH levels suggest that increases in PTH levels mediated the increases in 1,25(OH)2D by stimulating the renal 1α-hydroxylase. Together these findings confirm the suppressive effect of inflammation on PTH and renal activation of vitamin D, which was relieved after anti-TNF-α induction.
Our findings are supported by several preclinical studies indicating that inflammatory cytokines may perturb vitamin D metabolism by up-regulating the calcium-sensing receptor and inhibiting expression of the renal 1α-hydroxylase. Nuclear factor-κB, a key transcription factor in the signaling of TNF-α and IL-1β, was shown to suppress the 1α-hydroxylase in vitro in human embryonic kidney cells (24). Injection of IL-6 and IL-1β increased expression of the calcium-sensing receptor in the rat parathyroid, thyroid, and kidney with an associated decrease in circulating concentrations of PTH, calcium, and 1,25(OH)2D (5, 6). Although colonic expression of the 1α-hydroxylase was increased in a mouse model of colitis, renal expression of the 1α-hydroxylase was suppressed 5-fold, with consequently lower circulating 1,25(OH)2D levels, compared with control mice (7).
Our findings corroborate and expand on our prior observational study showing that compared with healthy controls, children with incident CD had a deficiency of 25(OH)D and 1,25(OH)2D and a relative hypoparathyroidism, which resolved over long-term follow-up after various types of treatment (12). Our findings also build upon data from other inflammatory disease states. A cross-sectional analysis of adults with rheumatoid arthritis showed that PTH and 1,25(OH)2D were negatively correlated with disease activity assessed by CRP (Spearman's rho −0.26 and −0.52, respectively); the inverse association between 1,25(OH)2D and CRP was present in both those who did and those who did not receive glucocorticoid therapy, and concentrations of vitamin D metabolites and PTH did not differ by glucocorticoid therapy (25). Similarly, using analyses restricted to those not on glucocorticoids at baseline, we confirmed that the results observed here were not attributable to the discontinuation of glucocorticoid therapy. More recently a longitudinal study of persons with type 1 diabetes found that worse baseline periodontal health was associated with lower serum 1,25(OH)2D concentrations and that serum 1,25(OH)2D increased significantly after antiinfective periodontal therapy. Serum 25(OH)D did not change after therapy, and PTH was not measured (26).
Most circulating 1,25(OH)2D is derived from renal 1α-hydroxylation of circulating 25(OH)D. Inflammatory cytokines have been shown to stimulate the colonic 1α-hydroxylase, and higher serum 1,25(OH)2D concentrations have been attributed to inflammation-induced intestinal overexpression of intestinal 1α-hydroxylase in CD patients (27). Therefore, the increases in serum 1,25(OH)2D in our study after anti-TNF induction and concurrent declines in inflammatory cytokines are most likely due to increased renal 1α-hydroxylase activity. Indeed, seven of the eight participants with baseline 1,25(OH)2D concentrations greater than 70 pg/mL, presumably secondary to intestinal inflammation, experienced declines in their 1,25(OH)2D after induction.
The putative explanation that the lower PTH and 1,25(OH)2D levels observed in our prior study were mediated by TNF-α induced increases in FGF23 levels was not substantiated here. FGF23 levels were not associated with disease activity at baseline and did not change during anti-TNF-α therapy despite dramatic decreases in disease activity and inflammatory markers. Furthermore, if FGF23 excess contributed to low vitamin D levels through stimulation of 24-hydroxylase, one would expect to see decreases in 24,25(OH)2D with anti-TNF-α therapy; this was not observed here. In the subanalysis of participants not on corticosteroid therapy, FGF23 increased, which would not explain changes in PTH and 1,25(OH)2D because FGF23 is inhibitory on both. Similarly, another recent study of CD patients not only found no association between FGF23 and disease activity but also demonstrated lower FGF23 concentrations compared with controls (28). The lack of a decline in 24,25(OH)2D supports the hypothesis that the effect of PTH on 1,25(OH)2D is mostly through the stimulation of the 1α-hydroxylase rather than through the inhibition of the 24-hydroxylase.
The primary limitations of this study are as follows: 1) it was uncontrolled; however, our observations were unlikely attributable to natural history, given the magnitude of changes over a short duration of follow-up; 2) mechanism cannot be fully delineated using biomarkers in a human study because tissue for gene expression would be needed; 3) free vitamin D levels were calculated based on DBP and albumin levels, rather than measured directly; 4) measures of serum ionized calcium, magnesium, phosphorus, and urine calcium were not performed, although the anticipated change in serum calcium, magnesium, and phosphate, if any, would be an increase due to improved intestinal absorption, all of which would suppress PTH; and 5) because the PCDAI is a pediatric index, we were unable to apply it among the adult participants, although we did analyze complete measures of inflammatory markers among adult participants.
This study has several important strengths. To our knowledge, it is the first study to examine changes in vitamin D metabolism after anti-TNF induction in any disease state. It corroborates and builds upon the recent study of changes in vitamin D metabolism after treatment of periodontal inflammation in diabetic patients, suggesting our findings are generalizable to other inflammatory states. It is unique in its comprehensive assessment of inflammation and vitamin D-related mineral metabolism, with longitudinal measures of total and free vitamin D metabolites, PTH, FGF23, and calcium. The study of a targeted biological therapy in a single disease, uniformity of dosing, and consistent timing of short-term follow-up definitely establishes the relationships between inflammation and vitamin D metabolism.
In summary, we demonstrated that inflammation was associated with lower PTH and 1,25(OH)2D concentrations in CD patients. After anti-TNF induction, PTH and 1,25(OH)2D concentrations increased without concomitant changes in 25(OH)D and FGF23, confirming the suppressive effects of inflammation on PTH and thereby renal activation of vitamin D. Future studies are needed to understand how TNF blockade and consequent increases in PTH and 1,25(OH)2D relate to bone health, particularly in the growing skeleton.
Acknowledgments
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
This work was supported by National Institutes of Health Grants K23 DK082012 (to M.T.), K23 DK093556 (to M.R.D.), and K24 DK076808 (to M.B.L.); Grant 5K08HD060739–05 (to M.D.D.); by the National Center for Research Resources Grant UL1RR024134, which is now at the National Center for Advancing Translational Sciences; Grant UL1TR000003; and by the Penn Joint Center for Inflammatory Bowel Diseases. M.R.D. is also supported by The Nephcure Foundation-American Society of Nephrology Research Grant.
Current address for M.T.: Merck, Sharp, & Dohme Corp, North Wales, Pennsylvania. Her research contribution was completed while she was affiliated with the Children's Hospital of Philadelphia.
Disclosure Summary: I.H.d.B. receives research funding from Abbott Laboratories. M.R.D. receives research funding from Genentech and is a consultant for Infiniti Medical. R.N.B receives funding from Janssen Biotech, Inc. M.B.L. has consultancy agreements with Amgen, Inc, Johnson & Johnson, and Novartis. The remaining authors have nothing to disclose.
Footnotes
- BMI
- body mass index
- CD
- Crohn's disease
- CRP
- C-reactive protein
- CV
- coefficient of variation
- DBP
- vitamin D-binding protein
- eGFR
- estimated glomerular filtration rate
- ESR
- erythrocyte sedimentation rate
- FGF23
- fibroblast growth factor 23
- GEE
- generalized estimating equation
- IQR
- interquartile range
- 24,25(OH)2D
- 24,25-dihydroxyvitamin D
- 25(OH)D
- 25-hydroxyvitamin D
- 1,25(OH)2D
- 1,25-dihydroxyvitamin D
- PCDAI
- Pediatric Crohn's Disease Activity Index.
References
- 1. Levin AD, Wadhera V, Leach ST, et al. Vitamin D deficiency in children with inflammatory bowel disease. Digest Dis Sci. 2011;56:830–836 [DOI] [PubMed] [Google Scholar]
- 2. Pappa HM, Gordon CM, Saslowsky TM, et al. Vitamin D status in children and young adults with inflammatory bowel disease. Pediatrics. 2006;118:1950–1961 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Sentongo TA, Semaeo EJ, Stettler N, Piccoli DA, Stallings VA, Zemel BS. Vitamin D status in children, adolescents, and young adults with Crohn disease. Am J Clin Nutr. 2002;76:1077–1081 [DOI] [PubMed] [Google Scholar]
- 4. Sylvester FA, Wyzga N, Hyams JS, et al. Natural history of bone metabolism and bone mineral density in children with inflammatory bowel disease. Inflamm Bowel Dis. 2007;13:42–50 [DOI] [PubMed] [Google Scholar]
- 5. Canaff L, Hendy GN. Calcium-sensing receptor gene transcription is up-regulated by the proinflammatory cytokine, interleukin-1β. Role of the NF-κB pathway and κB elements. J Biol Chem. 2005;280:14177–14188 [DOI] [PubMed] [Google Scholar]
- 6. Canaff L, Zhou X, Hendy GN. The proinflammatory cytokine, interleukin-6, up-regulates calcium-sensing receptor gene transcription via Stat1/3 and Sp1/3. J Biol Chem. 2008;283:13586–13600 [DOI] [PubMed] [Google Scholar]
- 7. Liu N, Nguyen L, Chun RF, et al. Altered endocrine and autocrine metabolism of vitamin D in a mouse model of gastrointestinal inflammation. Endocrinology. 2008;149:4799–4808 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Uno JK, Kolek OI, Hines ER, et al. The role of tumor necrosis factor α in down-regulation of osteoblast Phex gene expression in experimental murine colitis. Gastroenterology. 2006;131:497–509 [DOI] [PubMed] [Google Scholar]
- 9. Inoue Y, Segawa H, Kaneko I, et al. Role of the vitamin D receptor in FGF23 action on phosphate metabolism. Biochem J. 2005;390:325–331 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Chun RF, Lauridsen AL, Suon L, et al. Vitamin D-binding protein directs monocyte responses to 25-hydroxy- and 1,25-dihydroxyvitamin D. J Clin Endocrinol Metab. 2010;95:3368–3376 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Jeffery LE, Wood AM, Qureshi OS, et al. Availability of 25-hydroxyvitamin D(3) to APCs controls the balance between regulatory and inflammatory T cell responses. J Immunol. 2012;189:5155–5164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Prosnitz AR, Leonard MB, Shults J, et al. Changes in vitamin D and parathyroid hormone metabolism in incident pediatric Crohn's disease. Inflamm Bowel Dis. 2013;19:45–53 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. El-Hodhod MA, Hamdy AM, Abbas AA, Moftah SG, Ramadan AA. Fibroblast growth factor 23 contributes to diminished bone mineral density in childhood inflammatory bowel disease. BMC Gastroenterol. 2012;12:44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hyams JS, Ferry GD, Mandel FS, et al. Development and validation of a pediatric Crohn's disease activity index. J Pediatr Gastroenterol Nutr. 1991;12:439–447 [PubMed] [Google Scholar]
- 15. Ogden CL, Kuczmarski RJ, Flegal KM, et al. Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version. Pediatrics. 2002;109:45–60 [DOI] [PubMed] [Google Scholar]
- 16. Hollis BW. Quantitation of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D by radioimmunoassay using radioiodinated tracers. Methods Enzymol. 1997;282:174–186 [DOI] [PubMed] [Google Scholar]
- 17. Clive DR, Sudhaker D, Giacherio D, et al. Analytical and clinical validation of a radioimmunoassay for the measurement of 1,25 dihydroxy vitamin D. Clin Biochem. 2002;35:517–521 [DOI] [PubMed] [Google Scholar]
- 18. Bosworth CR, Levin G, Robinson-Cohen C, et al. The serum 24,25-dihydroxyvitamin D concentration, a marker of vitamin D catabolism, is reduced in chronic kidney disease. Kidney Int. 2012;82:693–700 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Serum-calcium. Lancet. 1979;1:858–859 [PubMed] [Google Scholar]
- 20. Schwartz GJ, Haycock GB, Edelmann CM, Jr, Spitzer A. A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics. 1976;58:259–263 [PubMed] [Google Scholar]
- 21. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Bikle DD, Gee E, Halloran B, Kowalski MA, Ryzen E, Haddad JG. Assessment of the free fraction of 25-hydroxyvitamin D in serum and its regulation by albumin and the vitamin D-binding protein. J Clin Endocrinol Metab. 1986;63:954–959 [DOI] [PubMed] [Google Scholar]
- 23. Bikle DD, Siiteri PK, Ryzen E, Haddad JG. Serum protein binding of 1,25-dihydroxyvitamin D: a reevaluation by direct measurement of free metabolite levels. J Clin Endocrinol Metab. 1985;61:969–975 [DOI] [PubMed] [Google Scholar]
- 24. Ebert R, Jovanovic M, Schneider D, et al. Down-regulation by nuclear factor κB of human 25-hydroxyvitamin D3 1α-hydroxylase promoter. Mol Endocrinol. 2004;18:2440–2450 [DOI] [PubMed] [Google Scholar]
- 25. Oelzner P, Muller A, Deschner F, et al. Relationship between disease activity and serum levels of vitamin D metabolites and PTH in rheumatoid arthritis. Calcif Tissue Int. 1998;62:193–198 [DOI] [PubMed] [Google Scholar]
- 26. Antonoglou G, Knuuttila M, Niemela O, et al. Serum 1,25(OH)D level increases after elimination of periodontal inflammation in T1DM subjects. J Clin Endocrinol Metab. 2013;98(10):3999–4005 [DOI] [PubMed] [Google Scholar]
- 27. Abreu MT, Kantorovich V, Vasiliauskas EA, et al. Measurement of vitamin D levels in inflammatory bowel disease patients reveals a subset of Crohn's disease patients with elevated 1,25-dihydroxyvitamin D and low bone mineral density. Gut. 2004;53:1129–1136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Oikonomou KA, Orfanidou TI, Vlychou MK, et al. Lower fibroblast growth factor 23 levels in young adults with Crohn Disease as a possible secondary compensatory effect on the disturbance of bone and mineral metabolism. J Clin Densitom. 2014;17(1):177–184 [DOI] [PubMed] [Google Scholar]
