1 Vitamin D and Asthma Health DisparitiesScott T. Weiss MD, MS Professor of Medicine, Harvard Medical School Associate Director, Channing Division of Network Medicine Brigham and Women’s Hospital Director, Partners Personalized Medicine Partners HealthCare Boston MA BRIGHAM AND WOMEN’S HOSPITAL HARVARD MEDICAL SCHOOL

2 Disclosures NIH Funding UpToDate, Inc.NHLBI Funding for “Vitamin D Antenatal Asthma Reduction Trial (VDAART)” NIGMS Funding for “eMERGE” NIH Directors Funding for New England All of Us site NHLBI Funding for Systems Biology of Airways Disease PPG NHLBI Funding for Genetics of Asthma in Costa Rica and TopMed supplement NHLBI Funding for Physical Characteristics of Cells PPG core A PI UpToDate, Inc. Author royalties Industry Collaborations Novartis Genome Network Sciences

3 Outline Importance of vitamin D from an evolutionary perspectiveThe vitamin D compromise Vitamin D deficiency is widespread greater in AA Studies of vitamin D deficiency and asthma in minority populations Effect of race on VDAART results Conclusions This is a brief outline of what I want to discuss today. There are few studies that have investigated vitamin D as a cause of asthma disparities, so I hope to convince you of the importance of vitamin D in human health through the adaptations that have occurred (the vitamin D compromise), that because of breaking of this vitamin D compromise, vitamin D deficiency is now widespread. I will then review some studies that have specifically investigated vitamin D deficiency in minority populations.

4 Studies in Minority PopulationsCase-control study of 92 AA asthmatics and 21 AA controls from Washington, DC This was a small study of 92 AA asthmatics and 21 AA controls. The authors found that most of the subjects had vitamin D insufficiency (<30 ng/ml) or deficiency (<20) (left panel). Cases also had significantly lower levels than non asthmatics. This was a small study and was cross sectional, and did not address symptoms or severity of disease. P = 0.002 Freishtat et al. J Pediatr 2010

5 Studies in Minority PopulationsCase-control study of 287 PR asthmatics and 273 controls The next study was a case control study from Juan Celedon’s group. In this study of over 500 Puerto Rican children, children with insufficient vitamin D levels had greater risks for severe asthma exacerbations, lower indices of lung function, and greater risks for allergic sensitization. These effects persisted despite adjusting for potential confounders (increasing african ancestry, time outdoors, and use of supplements). Brehm et al. AJRCCM 2012

6 Studies in Minority Populations2 samples of adolescents NHANES (n=6487) Asthma Control Evaluation (ACE) trial (n=226) No significant and consistent associations with markers of asthma severity, spirometric indices, or atopy/inflammation Finally, the largest study comes from the NHANES study. The authors found that AA children had lower levels of 25OHD than non-AA children, but that there were no differences by asthma status. Finally, they did not find any associations between vitamin D and markers of asthma severity, lung function, or atopy. Gergen et al. Am J Clin Nutr 2013

7 VDAART Vitamin D Antenatal Asthma Reduction TrialU01HL (NHLBI) Data Coordinating Center Channing Laboratory, BWH, Boston, MA (Dual-PI: Scott Weiss, Gus Litonjua) 3 Clinical Centers BU/BMC, Boston, MA (PI: George O’Connor) Washington University, St. Louis, MO (PI: Robert Strunk) Kaiser Permanente/UCSD (PI: Robert Zeiger)

8 Vitamin D levels at weeks of pregnancy predict Preeclamsia at weeks. A gene network in blood At weeks encompasses the Vitamin D pathway and also predicts PE at weeks Mirzakhani H et. al. JCI 2016; 126:

9 PRIMARY INTENT-TO-TREAT ANALYSISChawes B. JAMA 2016; Litonjua A. JAMA 2016.

10 Post-hoc: Stratification by 3rd trimester levelGroup E: 4,400 IU, 3rd trimester 25OHD > 75 nmol/L Group B: 400 IU, 3rd trimester 25OHD < 75 nmol/L Figure S2. Kaplan-Meier curves of the ITT analysis stratified by whether 3rd trimester 25(OH)D3 levels in the mother were ≥ 75 nmol/L (30 ng/ml). In this analysis, the children of mothers who were on the 4,400 IU/day arm and whose 3rd trimester 25(OH)D3 levels were ≥ 75 nmol/l (Group E) had the lowest risk for developing asthma compared with the other groups (p=0.02). The other strata were: Group A = children whose mothers were in the 4,400 IU/day arm and whose 3rd trimester 25(OH)D3 levels were < 75 nmol/l, Group B = children whose mothers were in the 400 IU/day arm and whose 3rd trimester 25(OH)D3 levels were < 75 nmol/L, Group C = children whose mothers were in the 4,400 IU/day arm and unknown 3rd trimester 25(OH)D3 level, Group D = children whose mothers were in the 400 IU/day arm and unknown 3rd trimester 25(OH)D3 levels, Group F = children whose mothers were in the 400 IU/day arm and whose 3rd trimester 25(OH)D3 levels were ≥ 75 nmol/l. p=0.02

11 Risk of asthma or recurrent wheeze by treatment arm and inital maternal 25(OH)D level (reference group < 20 25(OH)D, placebo ) Odds ratios adjusted for: maternal age, race, center, BMI JACI in press

12

13 Initial 25OHD levels of pregnant women in VDAARTPlot of 25OHD levels at randomization in VDAART women, stratified by treatment arm. To show that randomization worked.

14 This is another plot that can replace the first slideThis is another plot that can replace the first slide. This is not stratified by treatment arm.

15 Baseline Characteristics of the AA and nonAA participantsNon-AA (N = 400) P Treatment arm 49% (154) 51% (202) 0.82 Maternal asthma 42% (131) 39%(155) 0.43 Gender of child(Male) 54% (167) 52% (209) 0.79 C-section 33% (102) 27% (108) 0.12 Married 15% (46) 70% (279) < 0.001 Asthma/wheeze (child) 36% (111) 23% (91) BMI > 25 70% (217) 54% (214) College 10% (30) 55% (219) Adherence > 0.80 26% (81) 77% (309) Maternal age > 27 31% (98) 71% (285) Birth weight > 3500 21% (66) 47% (190) Gestational age < 37 5% (20) 0.02 Siblings 74% (231) 59% (236) We conducted an analysis to see whether vitamin D supplementation had differential effects between AA and nonAA women. This table presents the baseline characteristics. AA women had greater rates of pre-pregnancy obesity than nonAA women, and had greater proportion of indices of lower SES. There was also more asthma/wheeze in the offspring of the AA women compared with the nonAA women. 202 children developed asthma/wheeze by age 3, with 111 (36%) in AA vs 92 (23%) in nonAA participants (p<0.001).

16 25OHD levels and race Non-AA participants had higher levels of 25OHD in the trials Initial level 27.1 ng/mL [9.73 ng/mL] for nonAA vs ng/mL [8.25 ng/mL] for AA (p<0.001) 3rd trimester 36.33 ng/mL [13.84 ng/mL] for non-AA vs ng/mL [14.86 ng/mL] for AA (p<0.001) However, AA women had a greater increase in 25OHD levels than nonAA women 11.55 ng/mL [13.86 ng/mL] in AA women vs 9.23 ng/mL [12.85] in nonAA women (p=0.02) The greater increase in 25OHD levels in AA is probably partly due to a lower starting level. However, this puts into question the effect of the binding protein (GC) polymorphisms. Perhaps the effect of the GC polymorphisms do not have as big an effect in adequately supplemented populations.

17 Despite lower 25OHD levels in AA participants, there was no difference in the effect of vitamin D supplementation on asthma/wheeze in children Supplementation alone (without regard for race) had an effect on asthma/wheeze in the children. This effect was identical in AA and nonAA participants, despite that fact that AA women (and their children) had lower levels of 25OHD, and AA children had more asthma/wheeze.

18 Acknowledgements VDAART DCC: Gus Litonjua, Vince Carey, Nancy Laranjo, BJ Hartsfield VDAART Clinical Centers: ST Louis: Robert Strunk, Len Bacharier, Avi Beigelman San Diego: Bob Zeiger Boston: George O’Connor COPSAC study: Hans Bisgaard, Bo Chawes, Helene Wolsk Channing Division of Network Medicine: Benj Raby, Jessica Su, Rachel Kelly, Mike McGeachie, Hooman Mirzakani, Ed Silverman, Mike Cho, Peter Castaldi, Sunita Sharma, Kelan Tantisira, Alvin Koh, Bo Chawes, Helene Wolsk

19 Questions?

20 17q12: first asthma-susceptibility locus identified by GWASMoffatt MF et al. Nature : Common SNPs (~40% frequency) on 17q associated with asthma (OR 1.11 – 1.52) Associated with exacerbations, smoking, severe asthma, and asthma associated with respiratory infections (Tavendale et al. 2008; Bisgaard et al. 2009; Halapi et al. 2010; Smit et al. 2010) Slide courtesy of Benj Raby

21 17q12 most reproducible asthma locusEVE 12,976 subjects (5,416 asthma) GABRIEL 26,475 subjects (10,365 asthma) Torgenson et al. Nature Genetics 2011 Moffatt et al. New Engl J Med 2010 Slide courtesy of Benj Raby

22 VDAART, rs12936231: Intervention effect in genotype strataGG CG CC HR=0.55 ( ), p=0.14 HR=0.80 ( ), p=0.27 HR=1.06 ( ), p=0.86 Slide courtesy of Bo Chawes

23 Mediation analysis Can we identify the mechanisms through which the vitamin D intervention acts on asthma risk by looking at common metabolites? Is this risk mediated through these metabolites? R package: Mediation TOTAL EFFECT =  AVE. CAUSAL MEDIATION EFFECT +  AVE. DIRECT EFFECT Total effect: change in risk of asthma due to intervention ACME: Average change in the risk of asthma among those with sufficient vitamin D levels that arrives as a result of the metabolites rather than 'directly' from the effects of vitamin D (i.e. expected difference in the potential outcome when the mediator took the value that would realize under the treatment condition as opposed to the control condition) ADE: change in risk of asthma actually due to vitamin D status (unrelated to the metabolites of interest)    Slide courtsey of Rachel Kelly

24 Results Slide courtesy of Rachel Kelly Metabolite Estimate lower 95%CIupper 95% CI p-value oleoyl-linoleoyl-glycerol (18:1/18:2) [1] 0.025 0.003 0.057 0.010 oleoyl-oleoyl-glycerol (18:1/18:1) [2]* 0.026 0.002 0.058 0.020 palmitoyl-oleoyl-glycerol (16:0/18:1) [2]* 0.022 0.001 0.055 0.040 sphingomyelin 0.024 0.056 1-palmitoyl-GPE (16:0) 0.021 -0.001 0.054 0.060 2-hydroxybutyrate/2-hydroxyisobutyrate -0.018 -0.047 0.070 oleoyl-oleoyl-glycerol (18:1/18:1) [1]* 0.016 0.000 0.043 palmitoyl-oleoyl-glycerol (16:0/18:1) [1]* 0.080 palmitoleoyl-linoleoyl-glycerol (16:1/18:2) [1]* 0.012 -0.004 0.039 0.170 glycerophosphoinositol* -0.007 -0.029 0.400 glycosyl-N-palmitoyl-sphingosine (d18:1/16:0) 0.008 -0.015 0.540 Slide courtesy of Rachel Kelly

25 ORMDL3 is a negative regulator of Sphingolipid SynthesisSPT

26 Sphingomyelin, Asthma risk and Vitamin D InterventionPositively associated with vitamin D, inversely associated with asthma In individuals with the risk allele sphingomyelin levels – a downstream product of sphingolipid metabolism- are lower than in wild type individuals Overexpression of ORMDL3 inhibits SPT SPT (serine palmitoyltransferase) - regulator of sphingolipid biosynthesis Impaired Sphingolipid Synthesis in the Respiratory Tract Induces Airway Hyperreactivity (Worgall et al(2013)) Vitamin D Slide ocurtesy of Rachel Kelly

27 Sphingomyelin, Asthma risk and Vitamin D InterventionPositively associated with intervention, inversely associated with asthma In individuals with the risk allele sphingomyelin levels – a downstream product of sphingolipid metabolism- are lower than in wild type individuals Overexpression of ORMDL3 inhibits SPT SPT (serine palmitoyltransferase) - regulator of sphingolipid biosynthesis Impaired Sphingolipid Synthesis in the Respiratory Tract Induces Airway Hyperreactivity (Worgall et al(2013)) Vitamin D Slide courtesy of Rachel Kelly

28 Sphingomyelin levels according to genotypeSlide courtesy of Rachel Kelly