Analyses of nicotine metabolism biomarker genetics stratified by sex in African and European Americans

Abstract

Nicotine is inactivated by the polymorphic CYP2A6 enzyme to cotinine and then to 3'hydroxycotinine. The Nicotine Metabolite Ratio (NMR; 3'hydroxycotinine/cotinine) is a heritable nicotine metabolism biomarker, varies with sex and ancestry, and influences smoking cessation and disease risk. We conducted sex-stratified genome-wide association studies of the NMR in European American (EA) and African American (AA) smokers (NCT01314001, NCT00666978). In EA females (n = 389) and males (n = 541), one significant (P < 5e-8) chromosome 19 locus was found (top variant: rs56113850, CYP2A6 (intronic), for C vs. T: females: beta = 0.67, P = 7.5e-22, 21.8% variation explained; males: beta = 0.75, P = 1.2e-37, 26.1% variation explained). In AA females (n = 503) and males (n = 352), the top variant was found on chromosome 19 but differed by sex (females: rs11878604, CYP2A6 (~ 16 kb 3'), for C vs. T: beta = - 0.71, P = 6.6e-26, 16.2% variation explained; males: rs3865454, CYP2A6 (~ 7 kb 3'), for G vs. T: beta = 0.64, P = 1.9e-19, 18.9% variation explained). In AA females, a significant region was found on chromosome 12 (top variant: rs12425845: P = 5.0e-9, TMEM132C (~ 1 Mb 5'), 6.1% variation explained) which was not significant in AA males. In AA males, significant regions were found on chromosomes 6 (top variant: rs9379805: P = 4.8e-9, SLC17A2 (~ 8 kb 5'), 8.0% variation explained) and 16 (top variant: rs77368288: P = 3.5e-8, ZNF469 (~ 92 kb 5'), 7.1% variation explained) which were not significant in AA females. Further investigation of these associations outside of chromosome 19 is required, as they did not replicate. Understanding how sex and ancestry influence nicotine metabolism genetics may improve personalized approaches for smoking cessation and risk prediction for tobacco-related diseases.

Figure 1

The top signal associated with the Nicotine Metabolite Ratio in European ancestry females and males was found on chromosome 19 in the region containing CYP2A6. The nicotine metabolite ratio was rank-transformed for analysis. Regional plots, generated using LocusZoom (freely available at locuszoom.org), show the top overall SNP in European American (EA) females (rs56113850) (a). The influence of rs56113850 on the Nicotine Metabolite Ratio (NMR) in EA females is shown in (b). The rs56113850 SNP was also the top SNP in EA males (c), and the influence of rs56113850 on the NMR in EA males is shown in d). Linkage disequilibrium patterns in (a) and (c) are based upon the hg19/1000 Genomes November 2014 release European reference population. The plots in (b) and (d) were created using SPSS version 23 (can be purchased from IBM, Armonk, New York, USA). The black horizontal line represents the mean rank-transformed NMR in each group, and the P-value is from the GWAS after adjusting for population substructure and NMR covariates.

Figure 2

The top signal associated with the Nicotine Metabolite Ratio in African ancestry females and males was found on chromosome 19 in the region containing CYP2A6. The nicotine metabolite ratio was rank-transformed for analysis. Regional plots were generated using LocusZoom (freely available at locuszoom.org). In African American (AA) females, the top variant was rs11878604 (a), which was also significant in AA males. The influence of rs11878604 on the Nicotine Metabolite Ratio (NMR) in AA females is shown in (b). In AA males, the top variant was rs3865454 (c), which was also significant in AA females. The influence of rs3865454 on the NMR in AA males is shown in (d). Note: in AA females, there was a second independent signal on chromosome 19: the top variant was an indel at position 41352257 (P = 8.3e−9, after conditioning on rs11878604); this variant was also GWS in AA males. Linkage disequilibrium patterns in (a) and (c) are based upon the hg19/1000 Genomes November 2014 release African reference population. The plots in (b) and (d) were created using SPSS version 23 (can be purchased from IBM, Armonk, New York, USA). The black horizontal line represents the mean rank-transformed NMR in each group, and the P-value is from the GWAS after adjusting for population substructure and NMR covariates.

Figure 3

There was a chromosome 12 signal associated with the Nicotine Metabolite Ratio in African American females which was not significant in African American males. The nicotine metabolite ratio was rank-transformed for analysis. A regional plot, generated using LocusZoom (freely available at locuszoom.org), shows the top overall chromosome 12 variant (rs12425845) in AA females (a); the influence of rs12425845 on the Nicotine Metabolite Ratio (NMR) in AA females is shown in (b). The chromosome 12 locus containing rs12425845 was not significant in AA males (c); the lack of influence of rs12425845 on the NMR in AA males is shown in (d). Linkage disequilibrium patterns in (a) and (c) are based upon the hg19/1000 Genomes November 2014 release African reference population. The plots in (b) and (d) were created using SPSS version 23 (can be purchased from IBM, Armonk, New York, USA). The black horizontal line represents the mean rank-transformed NMR in each group, and the P-value is from the GWAS after adjusting for population substructure and NMR covariates.

Figure 4

There was a chromosome 6 signal associated with the Nicotine Metabolite Ratio in African American males which was not significant in African American females. The nicotine metabolite ratio was rank-transformed for analysis. A regional plot, generated using LocusZoom (freely available at locuszoom.org), shows the top overall chromosome 6 variant (rs9379805) in AA males (a); the influence of rs9379805 on the Nicotine Metabolite Ratio (NMR) in AA males is shown in (b). The chromosome 6 locus containing rs9379805 was not significant in AA females (c); the lack of influence of rs9379805 on the NMR in AA females is shown in (d). Linkage disequilibrium patterns in (a) and (c) are based upon the hg19/1000 Genomes November 2014 release African reference population. The plots in (b) and (d) were created using SPSS version 23 (can be purchased from IBM, Armonk, New York, USA). The black horizontal line represents the mean rank-transformed NMR in each group, and the P-value is from the GWAS after adjusting for population substructure and NMR covariates.

Figure 5

There was a chromosome 16 signal associated with the Nicotine Metabolite Ratio in African American males which was not significant in African American females. The nicotine metabolite ratio was rank-transformed for analysis. A regional plot, generated using LocusZoom LocusZoom (freely available at locuszoom.org), shows the top overall chromosome 16 variant (rs77368288) in AA males (a); the influence of rs77368288 on the Nicotine Metabolite Ratio (NMR) in AA males is shown in (b). The chromosome 16 locus containing rs77368288 was not significant in AA females (c); the lack of influence of rs77368288 on the NMR in AA females is shown in (d). Linkage disequilibrium patterns in (a) and (c) are based upon the hg19/1000 Genomes November 2014 release African reference population. The plots in (b) and (d) were created using SPSS version 23 (can be purchased from IBM, Armonk, New York, USA). The black horizontal line represents the mean rank-transformed NMR in each group, and the P-value is from the GWAS after adjusting for population substructure and NMR covariates.