A Pharmacogenetic Study of ADRB2 Polymorphisms and Indacaterol Response in COPD Patients

Supplementary Information

A pharmacogenetic study of ADRB2 polymorphisms and indacaterol response in COPD patients

Roman Yelensky1, Yali Li1, Steve Lewitzky1, Elisabeth Leroy1, Carlyn Hurwitz1, David Rodman1, Alexandre Trifilieff2, and Charles A. Paulding1

1Novartis Institutes for BioMedical Research, Massachusetts Avenue

Cambridge, MA 02139

2Respiratory Disease Area, Novartis Institutes for BioMedical Research, Basel, Switzerland

Correspondence should be addressed to Yali Li ()

Phone 617-871-3045

Fax 617-871-8508

Genotyping Methods

TaqMan

TaqMan® genotyping was performed using TaqMan Assays-by-Design and Assays-on-Demand (Applied Biosystems, Foster City, CA) on an ABI 7900 sequencer. Up to 20 ng of genomic DNA was used in the experiment according to the manufacturer's instructions.

Sequencing

Sequencing was performed on M13 tagged PCR product, in both forward and reverse directions in separate 10µl reactions containing 0.2µl Big Dye 3.1 (Applied Biosystems, Foster City, CA), 2µl 5X Sequencing Buffer (Applied Biosystems), 0.75µl DMSO (Acros Organics, through Thermo Fisher Scientific), 4.95µl PCR grade water, 0.1µl of forward or reverse M13 sequencing primer (10µM), and 2.0µl of purified PCR product. Reactions were carried out in 384-well GeneAmp® 9700 thermocyclers (Applied Biosystems) with a standard protocol (25 cycles of 96°C for 30sec, 50°C for 5 sec, 60°C for 1min followed by a 4°C hold). All sequencing products were cleaned up using CleanSeq magnetic beads (Beckman Coulter Genomics, Danvers, MA) on a Biomek FX (Beckman Coulter) and eluted in 20µl of PCR grade water prior to being run on a 3730XL (Applied Biosystems).

Statistical models

Efficacy variables Forced Expiratory Volume (FEV1), Transtional Dyspnea Index (TDI), and Peak Expiratory Flow (PEF) were analyzed using a linear mixed model (SAS 9.1.3 PROC MIXED), with the efficacy endpoint as the dependent variable, SNP (coded 0, 1 or 2 according to the number of copies of the minor allele that an individual carries or 1/0 for Gly16Gly individuals vs. others and Arg16Arg individuals vs. others) as the independent variable (fixed effect), and the following covariates:

Fixed effect covariates:

· FEV1 prior to inhalation and FEV1 30-min post-inhalation of salbutamol/albuterol (components of SABA reversibility)

· FEV1 prior to inhalation and FEV1 one hour post-inhalation of ipratropium (components of anti-cholinergic reversibility)

· One of the following, based on the efficacy variable being considered:

FEV1: baseline FEV1 measurement (distinct assessment from reversibility testing)
TDI: baseline BDI (baseline dyspnea index) focal score
PEF: baseline morning PEF measurement

· treatment arm (150µg or 300µg of indacaterol)

· smoking status (except in analysis of smoker/non-smoker subgroups)

· study (QAB1492335S or QAB1492336)

· country, crossed with race category

Random effects:

· center nested within country, nested within study

Similarly, the number of COPD exacerbations during the 26 week treatment period was analyzed using a generalized linear mixed model (Poisson regression using function lmer in R version 2.10.1, lme4 package version 0.999375-32). Covariates included:

Fixed effect covariates:

· baseline percentage of days of poor COPD control

· baseline COPD exacerbation history (had prior exacerbations or not – yes/no)

· FEV1 prior to inhalation and FEV1 30-min post-inhalation of salbutamol/albuterol

· FEV1 prior to inhalation and FEV1 one hour post-inhalation of ipratropium

· treatment arm (150µg or 300µg of Indacaterol)

· smoking status (except in analysis of smoker/non-smoker subgroups)

· study (QAB1492335S or QAB1492336)

· country, crossed with race category

· “exposure” time – log(# of days patient spent in trial)

Random effects:

· center nested within country, nested within study

Haplotype testing

The haplotype analysis in Supplementary Table 4 was performed with haplotypes imputed for each individual using the software PLINK v1.07. Haplotypes were tested individually as SNPs (tested haplotype vs. others) using an additive model. Same statistical models with corresponding covariates for SNP testing were applied.

Additional pharmacogenetic association results

Supplementary Table 1: Pharmacogenetics results (p-values) for morning PEF over 26 weeks

SNP model
Patient subset / Number of patients used in analysis / Gly16Arg minor allele additive / Arg16Arg homozygotes vs. other genotype classes / Gly16Gly homozygotes vs. other genotype classes / Gln27Glu minor allele additive / Thr164Ile minor allele additive / 5’ UTR minor allele additive
All indacaterol recipients / 610 / 0.66 / 0.42 / 0.22 / 0.93 / 0.89 / 0.89
Ex-smokers / 327 / 0.69 / 0.07 / 0.46 / 0.94 / 0.70 / 0.95
Current smokers / 283 / 0.49 / 0.77 / 0.45 / 0.73 / 0.52 / 0.70
Prior regular beta-agonist users / 267 / 0.84 / 0.73 / 0.59 / 0.68 / 0.33 / 0.67
No regular beta-agonist use / 343 / 0.68 / 0.15 / 0.10 / 0.67 / 0.62 / 0.73
Inhaled corticosteroid recipients / 253 / 0.50 / 0.59 / 0.59 / 0.72 / 0.59 / 0.68
No inhaled corticosteroids / 357 / 0.77 / 0.31 / 0.22 / 0.96 / 0.60 / 0.97
Reversible airflow limitation / 328 / 0.05 / 0.48 / 0.02 / 0.07 / 0.90 / 0.06
Non-reversible limitation / 282 / 0.24 / 0.11 / 0.60 / 0.15 / 0.56 / 0.15

Supplementary Table 2: Pharmacogenetic results (p-values) for TDI at 26 weeks

SNP model
Patient subset / Number of patients used in analysis / Gly16Arg minor allele additive / Arg16Arg homozygotes vs. other genotype classes / Gly16Gly homozygotes vs. other genotype classes / Gln27Glu minor allele additive / Thr164Ile minor allele additive / 5’ UTR minor allele additive
All indacaterol recipients / 560 / 0.06 / 0.08 / 0.18 / 0.02 / 0.53 / 0.05
Ex-smokers / 301 / 0.07 / 0.32 / 0.08 / 0.01 / 0.17 / 0.03
Current smokers / 259 / 0.50 / 0.31 / 0.86 / 0.40 / 0.94 / 0.47
Prior regular beta-agonist users / 249 / 0.15 / 0.25 / 0.21 / 0.12 / 0.10 / 0.29
No regular beta-agonist use / 311 / 0.31 / 0.47 / 0.37 / 0.25 / 0.38 / 0.28
Inhaled corticosteroid recipients / 234 / 0.01 / 0.0012 / 0.15 / 0.01 / 0.73 / 0.02
No inhaled corticosteroids / 326 / 0.95 / 0.54 / 0.58 / 0.45 / 0.75 / 0.49
Reversible airflow limitation / 301 / 0.0007 / 0.03 / 0.0013 / 0.0013 / 0.99 / 0.0036
Non-reversible limitation / 259 / 0.33 / 0.54 / 0.35 / 0.84 / 0.18 / 0.97

Although several smaller p-values were observed in the pharmacogenetic analysis of TDI (though not strictly statistically significant at the pre-determined threshold of 0.0002), further investigation revealed that these likely represent spurious associations and not a true effect of ADRB2 genetic variation on indacaterol response:

· In the reversible airflow subgroup, effect size estimates (the fit parameters for the SNP variable in the linear model) were 0.9 and 1.2 in Gly16Arg minor allele additive and minor allele dominant models, respectively. These results indicate improvement in dyspnea with presence of the minor Arg16 allele, an effect inconsistent with published reports suggesting that presence of the Arg allele worsens beta-agonist response.

· Similarly, in the inhaled corticosteroid recipients, Arg16Arg homozygotes were estimated to have a mean TDI score 2.2 points higher than that of patients in the other genotype classes, a result inconsistent with published reports of worse beta-agonist response in these individuals.

Supplementary Table 3: Trial-specific pharmacogenetic results (p-values, coefficients) for putative TDI associations in the reversible airflow limitation subgroup

Study / Number of patients used in analysis / SNP model
Gly16Arg minor allele additive / Arg16Arg homozygotes vs. other genotype classes / Gly16Gly homozygotes vs. other genotype classes / Gln27Glu minor allele additive / Thr164Ile minor allele additive / 5’ UTR minor allele additive
Beta / p-value / Beta / p-value / Beta / p-value / Beta / p-value / Beta / p-value / Beta / p-value
QAB149B2335S / 217 / 0.71 / 0.03 / 0.85 / 0.20 / 0.93 / 0.03 / -0.69 / 0.02 / 1.07 / 0.44 / -0.64 / 0.03
QAB149B2336 / 84 / 1.29 / 0.0067 / 1.29 / 0.14 / 2.14 / 0.0036 / -1.13 / 0.02 / -1.81 / 0.27 / -1.10 / 0.05
Pooled / 301 / 0.90 / 0.0007 / 1.13 / 0.03 / 1.20 / 0.0013 / -0.81 / 0.0013 / 0.009 / 0.99 / -0.76 / 0.0036

Supplementary Table 4: Pharmacogenetic result (p-values) for Gly16Arg/Gln27Glu haplotypes

Clinical endpoint / Number of patients used in analysis / Haplotype
GG (39.67%) / AC (39.67%) / GC (20.67%)
FEV1 / 626 / 0.35 / 0.89 / 0.22
exacerbations / 648 / 0.73 / 0.64 / 0.89
PEF / 610 / 0.93 / 0.66 / 0.69
TDI / 560 / 0.02 / 0.06 / 0.52

Supplementary Table 5: Pharmacogenetic result (p-values) for Caucasian study subjects

SNP model
Clinical endpoint / Number of patients used in analysis / Gly16Arg minor allele additive / Arg16Arg homozygotes vs. other genotype classes / Gly16Gly homozygotes vs. other genotype classes / Gln27Glu minor allele additive / Thr164Ile minor allele additive / 5’ UTR minor allele additive
FEV1 / 565 / 0.72 / 0.66 / 0.85 / 0.47 / 0.87 / 0.54
exacerbations / 585 / 0.82 / 0.57 / 0.93 / 0.92 / 0.16 / 0.99
PEF / 552 / 0.40 / 0.68 / 0.14 / 0.83 / 0.89 / 0.79
TDI / 504 / 0.08 / 0.07 / 0.22 / 0.02 / 0.56 / 0.06