TOLLIP, MUC5B and the Response to N-acetylcysteine among Individuals with Idiopathic Pulmonary Fibrosis 20 minutes Justin Oldham, MD MS Pulmonary and Critical Care Medicine University of Chicago Pulmonary Fibrosis Foundation Summit 2015

Disclosures No relevant commercial interests

Background rs5743890 (TOLLIP) rs35705950 (MUC5B) AG GG GA AA Contemporaneously several studies were published identifying gene polymorphisms associated with IPF susceptibilty as well as survival. These genes of course are TOLLIP and MUC5B Noth et al. Lancet RM 2013 Fingerlin et al. Nature Genetics 2013 Peljto et al. JAMA 2013

TOLLIP encodes toll-like interacting protein (tollip) Inhibitory adaptor protein acting downstream of toll-like receptors (TLR) TLR-2 and TLR-4 are active in the lung After interaction with TLR-2 and TLR-4 ligands, tollip protein increases anti-inflammatory IL-10 production and suppresses pro-inflammatory TNF-α and IL-6 production TLR-4 is an oxidant dependent pathway Can be modulated by the presence of anti-oxidants, including N-acetylcysteine (NAC) Shah et al.. J Immunol 2012 Saito et al. Cell and tissue research 2005 Janardhan et al. Histology and histopathology 2006

MUC5B encodes a mucin-producing protein (Mucin-5b) Mucin-5b Contributes to airway mucus production Influences lower airway microbiome Is necessary for airway defense Molyneaux et al. AJRCCM 2014. Roy et al. Nature 2014.

Research Question Because TOLLIP and MUC5B play critical roles in lung host defense, do variants within these genes influence the effect of immunsuppressive and anti-oxidant therapy?

Randomized controlled trial comparing 3 treatment arms Background Randomized controlled trial comparing 3 treatment arms Prednisone/Azathioprine/N-acetylcysteine combination (PAN) N-acetylcysteine (NAC) monotherapy Placebo Trial stopped and PAN arm terminated after interim analysis showed an increased risk of death and hospitalization Trial then resumed with NAC and placebo arms So we are all of course familiar with the PANTHER clinical trial which compared 3 treatment arms – placebo, NAC and combination prednisone, aza and NAC. The combination arm was stopped early due to increased risk of death and hospitalization in this group and the placebo and NAC arms were continued until completion. Analysis showed that NAC was no different than placebo when comparing FVC decline over time. This figure from the paper showed something interesting however. Prior to the clinical alert that was issued for the combination arm, NAC was performing quite well compared to placebo, whereas after the clinical alert the opposite was true. This went for several other metrics as well. Raghu et al. NEJM 2012 Martinez et al. NEJM 2014

Martinez et al. NEJM 2014

Our Goal Determine whether single nucleotide polymorphisms (SNPs) within TOLLIP and MUC5B modify the effect of PANTHER interventions

Methods 5 SNPs genotyped in patients consenting to genetic analysis rs5743890 (TOLLIP) (intronic) rs5743894 (TOLLIP) (intronic) rs3750920 (TOLLIP) (exonic) rs5743854 (TOLLIP) (promoter) rs35705950 (MUC5B) (promoter) Drug-gene interaction tested with multivariable Cox regression model with interaction term pinteraction<0.01 considered significant based on Bonferroni correction Composite endpoint used death, hospitalization, transplant or ≥10% decline in forced vital capacity (FVC)

Result 341 patients Enrolled in PANTHER 26 patients identifying as Hispanic ethnicity or non-white race excluded 315 patients eligible for analysis 154 patients enrolled in genetics sub-study 161 patients did not enroll in genetics sub-study Figure S1. PANTHER Genetic Analysis 54 assigned to Placebo arm 60 assigned to NAC arm 40 assigned to PAN arm

Results – Baseline Characteristics and Outcomes Genotyped (n=154) p-value Placebo Arm (n=54) NAC Arm (n=60) PAN Arm (n=40) Age, mean (±SD) 66.1 (7.9) 67.9 (8.7) 69.7 (6.8) 0.11 Male, n (%) 39 (72.2) 47 (78.3) 31 (77.5) 0.72 Ever Smoker, n (%) 41 (75.9) 44 (73.3) 28 (70) 0.81 FVC, % predicted (±SD) 73.2 (14.7) 73 (15.7) 71.2 (15.2) DLCO, % predicted (±SD) 46.4 (11.7) 43.9 (10.9) 43 (10.6) 0.3 Death, n (%) 2 (3.7) 1 (1.7) 4 (10) 0.14 FVC Decline ≥ 10%, n (%) 12 (22.2) 11 (18.3) 6 (15) 0.67 Hospitalization, n (%) 8 (14.8) 8 (13.3) 13 (32.5) 0.04 Transplant, n (%) 1 (1.9) 3 (5) 1 (2.5) 0.74 Composite Endpoint**, n (%) 17 (31.5) 19 (31.7) 19 (47.5) 0.18

Results - Drug-Gene Interaction Multivariable Cox Interaction Model Estimates* Variable Coefficient (95% CI) p-value rs5743890 (TOLLIP) 0.84 (-0.32 - 1.99) 0.16 NAC Therapy 0.17 (-0.58 - 0.92) 0.66 PAN Therapy 1.36 (0.56 - 2.17) 0.001 rs5743890*NAC interaction -1.17 (-2.91 - 0.56) 0.19 rs5743890*PAN interaction -1.08 (-2.70 - 0.55) rs5743894 (TOLLIP) 0.05 (-0.91 - 1.00) 0.93 0.45 (-0.40 - 1.30) 0.30 0.74 (-0.30 - 1.78) 0.17 rs5743894*NAC interaction -1.42 (-2.89 - 0.06) 0.06 rs5743894*PAN interaction 0.58 (-0.82 - 1.97) 0.42 rs3750920 (TOLLIP) 0.39 (-0.32 - 1.10) 0.28 1.34 (0.17 - 2.52) 0.03 0.89 (-0.49 - 2.27) 0.20 rs3750920*NAC interaction -1.47 (-2.45 - -0.48) rs3750920*PAN interaction 0.19 (-0.81 - 1.19) 0.71 rs35705950 (MUC5B) 0.61 (-0.53 - 1.75) 0.29 0.91 (-0.34 - 2.51) 0.15 0.98 (-0.42 - 2.37) rs35705950*NAC interaction -1.40 (-2.89 - 0.09) 0.07 rs35705950*PAN interaction 0.13 (-1.48 - 1.74) 0.87 Significant interaction detected between NAC and rs3750920 (TOLLIP) Suggested interaction observed between NAC and rs5743894 (TOLLIP) and rs35705950 (MUC5B) No interaction observed between these SNPs and PAN combination therapy One of the TOLLIP SNPs did not meet the proportional hazard assumption needed to conduct Cox modeling so only 4 SNPs were tested here. And what we found was significant interaction between NAC therapy and the TOLLIP exonic SNP. We also see that there was a suggestion of interaction between the 894 intronic TOLLIP SNP as well as the MUC5B promoter SNP. We did not observe any interaction between these SNPs and PAN therapy.

Results – rs3750920 (TOLLIP) genotype-stratified composite endpoint-free survival between NAC and placebo arms NAC Placebo logrank p=0.01 HR 3.23; 95% CI 0.79-13.16; p=0.10 logrank p=0.78 HR 0.76; 95% CI 0.27-2.19; p=0.62 logrank p=0.06 HR 0.14 ; 95% CI 0.02-0.83; p=0.03 Placebo So based on this interaction analysis we constructed genotype stratified Cox models and KMs for the exonic TOLLIP SNP and found that those with a CC genotype had a trend towards worse outcomes compared to those receiving placebo with this genotype. We observed no difference between groups with a CT genotype and found that those in the NAC arm with a TT genotype did significantly better than their counterparts in the Placebo arm with a HR of 0.14 and p-value of 0.03. CC: Unadjusted logrank p=0.01 CT: p=0.82 TT: p=0.06

Results – Sensitivity Analysis of genotype-stratified endpoint risk Table 2. Sensitivity Analysis of NAC-associated endpoint risk after rs3750920 genotype stratification*   SNP Genotype (Gene) Primary Composite Endpoint Hospitalization-Free Survival Progression-Free Survival HR (95% CI) p-value rs3750920 (TOLLIP) CC 3.22 (0.79-13.16) 0.1 3.74 (0.35-39.80) 0.27 1.27 (0.21-7.55) 0.79 rs3750920 (TOLLIP) CT 0.76 (0.27-2.19) 0.62 0.45 (0.10-2.15) 0.32 0.63 (0.17-2.30) 0.48 rs3750920 (TOLLIP) TT 0.14 (0.02-0.83) 0.03 ** 0.09 (0.01-0.76) * Models adjusted for age, gender, FVC (% predicted) and DLCO (% predicted) ** No events observed in this group so HR could not be estimated Primary Composite Endpoint = death, transplant, hospitalization or 10% FVC decline Hospitalization-Free Survival = death, transplant or hospitalization Progression-Free Survival = death, transplant or 10% FVC decline Harm associated with NAC therapy in those with rs3750920 (TOLLIP) CC genotype driven primarily by hospitalizations Benefit associated with NAC therapy in those with rs3750920 (TOLLIP) TT genotype consistent across all endpoints You’ll notice that the heterozygotes for the 94 TOLLIP SNP and MUC5B promoter SNP are lumped together. This was because of the small number of individuals who were homozgyous minor for these SNPs so a dominant model was used for them whereas an additive model was used for the 920 TOLLIP SNP.

Conclusion NAC may be an efficacious therapy for those with rs3750920 (TOLLIP) TT genotype (~25% of IPF population) NAC may be harmful for those with an rs3750920 CC genotype (~25% of IPF population) Replication needed

Replication Cohort

NAC (n=47) Non-NAC (n=358) p-value Replication Cohort (n=405) Baseline Characteristics and Genotypes* Characteristic NAC (n=47) Non-NAC (n=358) p-value Age, mean (±SD) 68.3 (8.3) 66.4 (7.8) 0.12 Male, n (%) 33 (70.2) 267 (74.6) 0.52 Ever Smoker, n (%) 235 (70.4) 0.98 Azathioprine therapy, n (%) 5 (10.9) 10 (2.8) 0.02 Prednisone therapy, n (%) 23 (50.0) 132 (36.9) 0.09 FVC, % predicted (±SD) 62.6 (18.6) 71.1 (14.2) <0.001 DLCO, % predicted (±SD) 39.1 (10.5) 48.0 (11.9) 0.003 Death, n (%)   19 (40.4) 78 (21.8) 0.01 FVC Decline ≥ 10%, n (%) 17 (36.2) 0.93 Transplant, n (%) 2 (4.3) 11 (3.1) 0.46 Composite Endpoint**, n (%) 27 (57.5) 167 (46.7) 0.6 Abbreviations: NAC = N-acetylcysteine; FVC = forced vital capacity; DLCO = diffusion capacity of the lung for carbon monoxide; SNP = single nucleotide polymorphism *Non-Hispanic white individuals by self-report ** Death, transplant or 10% FVC decline So we identified patients in each cohort who received NAC therapy then combined cohorts to perform the same analysis as we performed with the PANTHER trial. As opposed to the PANTHER trial, which compared equivalent groups, with these cohorts we see that patients who received NAC were substantially sicker than those who did not. So it’s perhaps not surprising that we saw a much higher mortality rate among those receiving NAC. Some of this may have been due to the fact that the patients who received NAC also received prednisone and azathioprine, which we now know is harmful in IPF. So not an ideal dataset with which to validate our findings but it was the best we had. As for the SNPs, we limited the analysis to those SNPs which had significant or near significant interaction in our previous modeling, specifically one of the intronic TOLLIP SNPs, the exonic TOLLIP SNP and the MUC5B promoter SNP.

Replication Cohort Interaction Model Estimates Variable Coefficient (95% CI) p-value rs5743894 (TOLLIP) -0.06 (-0.34 - 0.23) 0.70 NAC Therapy 0.46 (0.06 - 0.98) 0.08 rs3750920*NAC interaction -0.33 (-1.26 - 0.60) 0.48 rs3750920 (TOLLIP) 0.01 (-0.19 - 0.22) 0.90 1.12 (0.47 - 1.79) 0.001 -0.73 (-1.3 - -0.16) 0.012 rs35705950 (MUC5B) -0.23 (-0.54 - 0.09) 0.31 1.0 (0.38 - 1.62) rs35705950*NAC interaction -1.07 (-1.95 - -0.19) 0.02 * Adjusted for azathioprine use, prednisone use, FVC (% predicted) and DLCO (% predicted) What we found was that significant interaction was again present with the TOLLIP exonic SNP. Interaction present when pinteraction < 0.017 - Adjusts for multiple testing

Genotype Log rank p-value HR 95% CI p-value Replication Cohort Genotype-stratified Composite Endpoint Risk* Associated with NAC Therapy Genotype Combined Cohort Log rank p-value HR 95% CI p-value rs3750920 CC 0.004 3.11 1.25-7.72 0.01 rs3750920 CT 0.002 2.18 1.14-4.13 0.02 rs3750920 TT 0.51 0.23 0.06-0.94 0.04 * Adjusted for age, gender,prednisone use, azathioprine use, FVC (% predicted), DLCO (% predicted) and trial/center Conclusion NAC may be beneficial for genetically susceptible individuals, but……. Recommendation regarding the use of NAC to treat IPF cannot be made until a prospective, genotype-stratified clinical trial can be conducted And so we again calculated composite endpoint risk stratified by the TOLLIP exonic SNP genotype and we again found that those with a TT genotype had a significant reduction in endpoint risk when receiving NAC therapy.

Secondary post-hoc analysis Limitations Secondary post-hoc analysis treatment arms were not randomized on genotype Relied upon self-reported ancestry ~ 50% did not consent to genetic analysis Small sample size Unclear whether those analyzed are representative of those who were not

The University of Chicago Weill Cornell Medical College Acknowledgments The University of Chicago Shwu Fan Ma, PhD Yong Huang, MD Rekha Vij, MD Eleanor Valenzi, MD Cathryn Lee, MD Leah Witt, MD Mary Strek, MD Imre Noth, MD Weill Cornell Medical College Fernando J. Martinez, MD Duke University Kevin Anstrom, PhD The University of Washington Ganesh Raghu, MD The University of Colorado David Schwartz, MD

References Noth I, Zhang Y, Ma S-F, Flores C, Barber M, Huang Y, et al. Genetic variants associated with idiopathic pulmonary fibrosis susceptibility and mortality: a genome-wide association study. The Lancet. 2013;1(4):309-317. Fingerlin TE, Murphy E, Zhang W, Peljto AL, Brown KK, Steele MP, et al. Genome-wide association study identifies multiple susceptibility loci for pulmonary fibrosis. Nat Genet. 2013;45(6):613-620. Peljto AL, Zhang Y, Fingerlin TE, Ma SF, Garcia JG, Richards TJ, et al. Association between the MUC5B promoter polymorphism and survival in patients with idiopathic pulmonary fibrosis. JAMA : the journal of the American Medical Association. 2013;309(21):2232-2239. Idiopathic Pulmonary Fibrosis Clinical Research N, Martinez FJ, de Andrade JA, Anstrom KJ, King TE, Jr., Raghu G. Randomized trial of acetylcysteine in idiopathic pulmonary fibrosis. N Engl J Med. 2014;370(22):2093-2101. Raghu G, Anstrom KJ, King TE, Jr., Lasky JA, Martinez FJ. Prednisone, azathioprine, and N-acetylcysteine for pulmonary fibrosis. N Engl J Med. 2012;366(21):1968-1977. Roy MG, Livraghi-Butrico A, Fletcher AA, McElwee MM, Evans SE, Boerner RM et al. Muc5b is required for airway defence. Nature 2014;505:412-416. Molyneaux PL, Cox MJ, Willis-Owen SA, Mallia P, Russell KE, Russell AM et al. The role of bacteria in the pathogenesis and progression of idiopathic pulmonary fibrosis. American journal of respiratory and critical care medicine 2014;190:906-913. Shah JA, Vary JC, Chau TT, Bang ND, Yen NT, Farrar JJ et al. Human tollip regulates tlr2 and tlr4 signaling and its polymorphisms are associated with susceptibility to tuberculosis. J Immunol 2012;189:1737-1746. Saito T, Yamamoto T, Kazawa T, Gejyo H, Naito M. Expression of toll-like receptor 2 and 4 in lipopolysaccharide-induced lung injury in mouse. Cell and tissue research 2005;321:75-88. Janardhan KS, McIsaac M, Fowlie J, Shrivastav A, Caldwell S, Sharma RK, Singh B. Toll like receptor-4 expression in lipopolysaccharide induced lung inflammation. Histology and histopathology 2006;21:687-696.

Thank You! 20 minutes Pulmonary Fibrosis Foundation Summit 2015