Genome-Wide Interaction Analysis of Genetic Variants With Menopausal Hormone Therapy for Colorectal Cancer Risk

Abstract

Background

The use of menopausal hormone therapy (MHT) may interact with genetic variants to influence colorectal cancer (CRC) risk.

Methods

We conducted a genome-wide, gene-environment interaction between single nucleotide polymorphisms and the use of any MHT, estrogen only, and combined estrogen-progestogen therapy with CRC risk, among 28 486 postmenopausal women (11 519 CRC patients and 16 967 participants without CRC) from 38 studies, using logistic regression, 2-step method, and 2– or 3–degree-of-freedom joint test. A set-based score test was applied for rare genetic variants.

Results

The use of any MHT, estrogen only and estrogen-progestogen were associated with a reduced CRC risk (odds ratio [OR] = 0.71, 95% confidence interval [CI] = 0.64 to 0.78; OR = 0.65, 95% CI = 0.53 to 0.79; and OR = 0.73, 95% CI = 0.59 to 0.90, respectively). The 2-step method identified a statistically significant interaction between a GRIN2B variant rs117868593 and MHT use, whereby MHT-associated CRC risk was statistically significantly reduced in women with the GG genotype (OR = 0.68, 95% CI = 0.64 to 0.72) but not within strata of GC or CC genotypes. A statistically significant interaction between a DCBLD1 intronic variant at 6q22.1 (rs10782186) and MHT use was identified by the 2–degree-of-freedom joint test. The MHT-associated CRC risk was reduced with increasing number of rs10782186-C alleles, showing odds ratios of 0.78 (95% CI = 0.70 to 0.87) for TT, 0.68 (95% CI = 0.63 to 0.73) for TC, and 0.66 (95% CI = 0.60 to 0.74) for CC genotypes. In addition, 5 genes in rare variant analysis showed suggestive interactions with MHT (2-sided P < 1.2 × 10⁻⁴).

Conclusion

Genetic variants that modify the association between MHT and CRC risk were identified, offering new insights into pathways of CRC carcinogenesis and potential mechanisms involved.

The use of menopausal hormone therapy (MHT) has been identified to be associated with a reduced risk of colorectal cancer (CRC) (1-4). In a meta-analysis including 20 studies, ever use of estrogen-only MHT (relative risk [RR] = 0.79, 95% confidence interval [CI] = 0.69 to 0.91) and ever use of combined estrogen-progestogen MHT (RR = 0.74, 95% CI = 0.68 to 0.81) were associated with a reduced CRC risk (1).

Previous gene-environment (GxE) interaction studies that investigated the association of MHT use with CRC risk according to genetic variants (5-10) have reported a few potential genetic modifiers of CRC risk associated with the use of MHT; however, these studies were based on limited candidate genes and/or pathways or limited sample size. We conducted a comprehensive genome-wide GxE analysis of common and rare genetic variants, using the largest known study sample to date, on one hand, to identify novel genetic variants that may modify the beneficial influence of MHT on CRC risk to obtain insight into potential mechanisms behind the association between MHT and CRC risk. On the other hand, the analysis can yield novel genetic susceptibility alleles for CRC risk, which may not be identified without accounting for the GxE component.

Methods

Study Participants

We included 38 studies from North America, Australia, and Europe participating in the multicentered Colon Cancer Family Registry, the Colorectal Transdisciplinary Study, the Genetics and Epidemiology of Colorectal Cancer Consortium, and the United Kingdom Biobank, which were included in genome-wide association studies (GWAS) as described previously (11-13). Study details and descriptions can be found in the Supplementary Methods (available online). All studies were approved by their respective institutional review boards, and study participants provided informed consent.

Exposure Assessment

Information on demographics and environmental risk factors were collected by interviews and/or structured questionnaires. We carried out a multistep data-harmonization procedure at the Genetics and Epidemiology of Colorectal Cancer Consortium coordinating center (Fred Hutchinson Cancer Research Center) as described previously (10,14,15).

Postmenopausal status was defined by using 1) menopausal status derived from studies, if available; 2) self-reported menopausal status, if study derived was not available; or 3) aged older than 55 years, if neither study derived nor self-report were available (Supplementary Table 1, available online). MHT use was considered as any MHT use or estrogen-only use or estrogen-progestogen use at or up to the reference time. Nonusers of any MHT at or up to reference time were used as the reference group.

Genotyping, Quality Control, and Imputation

Details on genotyping, imputation, and quality control have been reported previously (16). In brief, genotyped single nucleotide polymorphisms (SNPs) were excluded on the basis of call rate (<98%), evidence of departure from Hardy-Weinberg equilibrium in controls (P < 1 × 10⁻⁴). All autosomal SNPs in all studies were imputed to the Haplotype Reference Consortium r1.1 (2016) reference panel via the Michigan Imputation Server (17) and converted into a binary format for data management and analyses using R package BinaryDosage (18). Imputed common SNPs were restricted based on a pooled minor allele frequency (MAF) of at least 1% and imputation accuracy (R² > 0.8). After imputation and quality control analyses, more than 7.2 million common SNPs were included. All analyses were restricted to samples clustering with the Utah residents of Northern and Western European ancestry (the CEU population) in principal component analysis.

Statistical Methods

Statistical analyses of all data were conducted centrally on individual-level data. All tests of statistical significance were 2-sided. Unless otherwise indicated, we adjusted for age at the reference time, study center, and the first 3 principal components (Plink2) to account for potential population substructure. SNPs were treated as continuous variables (ie, log-additive effects). To evaluate MHT main effects, each study was analyzed separately using logistic regression models, and study-specific results were combined using fixed- and random-effects meta-analysis methods to obtain summary odds ratios (ORs) and 95% confidence intervals across studies. We calculated the heterogeneity P values using Cochran Q statistics (19). Quantile-quantile plots were used to assess whether the distribution of the P values was consistent with the null distribution (except for the extreme tail).

Genome-wide interaction scans of common markers were conducted using R package GxEScanR (20), which implements several interaction testing methods. To test for multiplicative statistical interactions between each SNP and environmental risk factors (MHT, estrogen only, estrogen-progestogen), we primarily used conventional case-control logistic regression analysis and 2-step methods (21-23) to test the GxE interaction term. Additionally, we also used a 2–degree-of-freedom (2-df) joint test (24) and 3-df joint test (25) to test GxE interaction in the context of simultaneously testing for the association between SNPs and CRC, and the association between SNPs and environmental risk factors (G|E) (MHT, estrogen only, estrogen-progestogen) associations. For the 2- and 3-df test, we do not report on known loci (16). For all novel findings, we examined the odds ratios of MHT, estrogen only, and estrogen-progestogen stratified by genotypes of statistically significant SNPs. More details in these testing methods can be found in the Supplementary Methods (available online).

For interaction analysis of rare genetic risk variants (MAF < 1%) and MHT, we conducted the Mixed effects Score Tests for interaction (MiSTi) (26), a set-based statistical framework providing mixed effects score tests for GxE interaction and addressing issues of power and low effect sizes, to discover genes that interact with MHT in relation to CRC risk (see the Supplementary Methods, available online). Because more than 20 000 genes were tested (22 476 genes for any MHT use, 20 609 for estrogen only, and 20 360 for estrogen-progestogen), interactions with a P value less than 2.5 × 10⁻⁶ were considered statistically significant, whereas those with a P value less than  1.2 × 10⁻⁴ were considered as suggestive.

Functional Annotation

We performed bioinformatic follow-up for genome-wide interaction study (GWIS) variants that were deemed statistically significant for downstream analysis (for more details, see the Supplementary Methods, available online). Relevant regional plots were generated using the command line version (Standalone) of LocusZoom v1.3 (27). Measures of linkage disequilibrium (LD) were estimated using study population controls.

Results

Detailed descriptive characteristics of the participants are shown in Table 1. MHT use was associated with reduced CRC risk both in cohort studies and case-control studies (Figures 1-3).

Figure 1.

Association of any menopausal hormone therapy use with the risk of colorectal cancer. CI = confidence interval; OR = odds ratio.

Figure 2.

Association of use of estrogen only with the risk of colorectal cancer. CI = confidence interval; OR = odds ratio.

Figure 3.

Association of use of combined estrogen-progestogen with the risk of colorectal cancer. CI = confidence interval; OR = odds ratio.

Table 1.

Descriptive characteristics of study participants included in the genome-wide interaction analysis between common variants and menopausal hormone therapy for risk of colorectal cancer^a

Study	CRC patients						Participants without CRC
	Total No.	No use of MHT	Any MHT	E-only	E + P	Age at diagnosis Mean (SD), y	Total No.	No use of MHT	Any MHT	E-only	E + P	Age at enrollment Mean (SD), y
		No. (%)	No. (%)	No.	No.
								No. (%)	No. (%)	No.	No.
CCFR Set 1	259	183 (70.7)	76 (29.3)	35	34	58.5 (9.9)	372	206 (55.4)	166 (44.6)	93	65	61.7 (7.9)
CCFR Set 3	427	292 (68.4)	135 (31.6)	68	42	61.8 (7.6)	250	155 (62.0)	95 (38.0)	54	27	62.7 (7.4)
CCFR Set 4	383	259 (67.6)	124 (32.4)	78	41	62.1 (9.3)	118	77 (65.3)	41 (34.7)	16	18	61.7 (9.3)
CLUEII	114	98 (86)	16 (14.0)	—	—	74.9 (9.5)	108	97 (89.8)	11 (10.2)	—	—	65.0 (9.4)
Colo 2&3	37	18 (48.6)	19 (51.4)	—	—	66.5 (11.4)	44	17 (38.6)	27 (61.4)	—	—	67.3 (9.2)
CPSII_1	263	176 (66.9)	87 (33.1)	50	37	74.8 (5.9)	255	142 (55.7)	113 (44.3)	83	30	74.3 (5.8)
CPSII_2	172	116 (67.4)	56 (32.6)	35	21	79.4 (6.1)	177	101 (57.1)	76 (42.9)	46	30	79.0 (6.0)
CRCGEN	274	266 (97.1)	8 (2.9)	—	—	68.8 (9.9)	394	377 (95.7)	17 (4.3)	—	—	65.9 (9.2)
DACHS_1	630	416 (66.0)	214 (34.0)	—	—	71.1 (9.5)	630	294 (46.7)	336 (53.3)	—	—	70.4 (8.7)
DACHS_2	229	161 (70.3)	68 (29.7)	—	—	72.2 (9.8)	162	88 (54.3)	74 (45.7)	—	—	72.3 (9.0)
DACHS_3	420	297 (70.7)	123 (29.3)	—	—	71.4 (9.5)	195	113 (57.9)	82 (42.1)	—	—	70.3 (10.1)
DALS_1	267	204 (76.4)	63 (23.6)	—	—	68.0 (7.7)	270	189 (70.0)	81 (30.0)	—	—	67.7 (7.9)
DALS_2	159	127 (79.9)	32 (20.1)	—	—	67.5 (7.5)	194	137 (70.6)	57 (29.4)	—	—	67.5 (8.2)
EPIC	771	544 (70.6)	227 (29.4)	—	—	67.2 (6.6)	865	619 (71.6)	246 (28.4)	—	—	72.5 (5.9)
ESTHER_VERDI	70	52 (74.3)	18 (25.7)	—	—	68.4 (6.8)	70	49 (70.0)	21 (30.0)	—	—	65.8 (6.7)
Kentucky	397	184 (46.3)	213 (53.7)	100	56	64.4 (8.9)	525	150 (28.6)	375 (71.4)	166	86	66.7 (6.6)
LCCS	116	90 (77.6)	26 (22.4)	—	—	66.1 (6.9)	108	88 (81.5)	20 (18.5)	—	—	65.7 (5.5)
MCCS_1	211	159 (75.4)	52 (24.6)	—	—	72.0 (7.1)	184	132 (71.7)	52 (28.3)	—	—	71.1 (7.2)
MCCS_2	85	65 (76.5)	20 (23.5)	—	—	74.3 (8.4)	86	65 (75.6)	21 (24.4)	—	—	73.8 (8.0)
MEC_1	99	55 (55.6)	44 (44.4)	27	—	70.3 (7.9)	115	42 (36.5)	73 (63.5)	37	—	70.3 (7.6)
MEC_2	15	2 (13.3)	13 (86.7)	5	—	80.1 (6.2)	30	4 (13.3)	26 (86.7)	12	—	74.6 (6.1)
MECC_3	309	260 (84.1)	49 (15.9)	—	—	69.5 (10.3)	367	290 (79.0)	77 (21.0)	—	—	73.0 (10.0)
NCCCSII	219	128 (58.4)	91 (41.6)	—	—	63.8 (9.8)	221	89 (40.3)	132 (59.7)	—	—	65.4 (9.4)
NFCCR_2	60	51 (85.0)	9 (15.0)	—	—	61.1 (7.9)	130	104 (80.0)	26 (20.0)	—	—	60.2 (7.2)
NHS_1_2	328	174 (53.0)	154 (47.0)	23	7	68.0 (7.4)	673	321 (47.7)	352 (52.3)	42	7	68.5 (6.9)
NHS_3_AD	410	187 (45.6)	223 (54.4)	21	10	68.1 (6.7)	335	133 (39.7)	202 (60.3)	15	7	67.9 (6.7)
PLCO_1_Rematch	216	125 (57.9)	91 (42.1)	—	—	68.8 (6.0)	123	61 (49.6)	62 (50.4)	—	—	67.5 (6.2)
PLCO_2	196	110 (56.1)	86 (43.9)	—	—	70.6 (6.6)	163	90 (55.2)	73 (44.8)	—	—	70.6 (6.3)
PLCO_3	295	157 (53.2)	138 (46.8)	—	—	67.0 (7.2)	1964	900 (45.8)	1064 (54.2)	—	—	62.1 (5.3)
PLCO_4_AD	434	241 (55.5)	193 (44.5)	—	—	64.0 (5.9)	587	274 (46.7)	313 (53.3)	—	—	61.9 (5.3)
REACH_AD	9	7 (77.8)	2 (22.2)	—	—	62.9 (4.0)	75	47 (62.7)	28 (37.3)	—	—	62.3 (5.6)
SMC_COSM	179	90 (50.3)	89 (49.7)	—	—	69.7 (9.7)	330	145 (43.9)	185 (56.1)	—	—	64.6 (7.8)
UKB_1	1073	996 (92.8)	77 (7.2)	—	—	65.4 (5.3)	4254	3928 (92.3)	326 (7.7)	—	—	65.4 (5.3)
USC_HRT_CRC	296	127 (42.9)	169 (57.1)	75	67	66.3 (5.5)	400	150 (37.5)	250 (62.5)	116	82	65.0 (6.8)
VITAL	114	61 (53.5)	53 (46.5)	—	—	70.5 (6.4)	126	60 (47.6)	66 (52.4)	—	—	71.5 (6.6)
WHI_1	450	297 (66.0)	153 (34.0)	95	58	71.0 (7.1)	519	282 (54.3)	237 (45.7)	137	100	71.2 (7.0)
WHI_2	977	576 (59.0)	401 (41.0)	202	199	72.2 (7.4)	990	512 (51.7)	478 (48.3)	260	217	72.0 (7.2)
WHI_3	556	313 (56.3)	243 (43.7)	117	126	78.6 (6.9)	558	267 (47.8)	291 (52.2)	148	142	78.5 (6.9)
Total	11519	7664 (66.5)	3855 (33.5)	931	698	—	16967	10795 (63.6)	6172 (36.4)	1225	811	—

^a

CCFR = Colon Cancer Family Registry; CLUEII = Campaign against Cancer and Heart Disease II; Colo 2&3 = Hawaii Colorectal Cancer Studies 2 & 3; CPSII = Cancer Prevention Study-II; CRCGEN = Colorectal Cancer Genetics & Genomics; DACHS = Darmkrebs: Chancen der Verhütung durch Screening; DALS = Diet, Activity, and Lifestyle Study; E + P = combined estrogen-progestogen; E-only = estrogen only; EPIC = European Prospective Investigation into Cancer; ESTHER_VERDI = Epidemiologische Studie zu Chancen der Verhütung = Früherkennung und optimierten THerapie chronischer ERkrankungen in der älteren Bevölkerung; Kentucky = Kentucky Case-Control Study; LCCS = Leeds Colorectal Cancer Study; MCCS = Melbourne Collaborative Cohort Study; MEC = Multiethnic Cohort Study; MECC = Molecular Epidemiology of Colorectal Cancer Study; MHT = menopausal hormone therapy; NCCCSII = The North Carolina Colon Cancer Study II; NHS = Nurses’ Health Study; NFCCR = Newfoundland Case-Control Study; PLCO = Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial; REACH = Colon Cancer Pathways: Hyperplastic Polyps and Adenomas; SMC_COSM = Swedish Mammography Cohort and Swedish Men Cohort; UKB = UK Biobank; USC_HRT_CRC = University of Southern California Hormone Replacement Therapy Colorectal Cancer Study; VITAL = Cancer Screening Trial VITamins And Lifestyle cohort; WHI = Women’s Health Initiative. The sign "—" = Not available

Genome-Wide MHT-Interaction Scans for CRC Risk

Statistical interaction results for genetic variants are summarized in Table 2. Although conventional case-control logistic regression models with a Bonferroni correction for multiple testing did not identify any statistically significant interactions between the use of any MHT, estrogen only, or estrogen-progestogen, and genetic variants (data not shown), we identified 2 interactions with common genetic variants reaching statistical significance for the 2-step method and 2-df joint test. The 2-step method (with G|E in step 1) identified a statistically significant interaction for any MHT use with SNP rs117868593 located 20 kb downstream of GRIN2B (Glutamate Ionotropic Receptor N-methyl D-aspartate Type Subunit 2B) variant at 12p13.1 (P_observed = .003, P_threshold = .005; Supplementary Figures 1 and 2, available online). The 2-df joint test identified a further statistically significant interaction for any MHT use with a DCBLD1 (Discoidin, CUB [Complement C1r/C1s, Uegf, Bmp1] And LCCL [Limulus factor C, Coch-5b2 and Lgl1] Domain Containing 1) intronic variant at 6q22.1 (rs10782186; joint P_observed = 4.23 × 10⁻⁸, P_threshold = 5 × 10⁻⁸; Supplementary Figures 3 and 4, available online). Several DCBLD1 intronic variants at 6q22.1 (rs4945586, rs9320604, rs4946260), which were in LD with rs10782186, also yielded low P values using the 2-df joint test although not genome-wide significant (5.28 × 10⁻⁸, 5.60 × 10⁻⁸, and 5.70 × 10⁻⁸; Supplementary Figures 3 and 4, available online). We did not identify any genome-wide statistically significant interactions between estrogen-only use or estrogen-progestogen use and common genetic variants for CRC risk. Common variants that reached the suggestive interaction level (P < 5 × 10⁻⁶) with MHT use for CRC risk are shown in Supplementary Tables 2-4 (available online), which included 87 SNPs with any MHT use, 80 with estrogen-only use, and 137 with estrogen-progestogen use. We also performed GWIS stratified by colon and rectal cancer, but the common variant analysis did not yield any statistically significant interactions for the MHT variables, respectively (data not shown).

Table 2.

Results of genome-wide interaction analyses with menopausal hormone therapy for colorectal cancer risk among postmenopausal women^a

MHT type	SNP	Chr	BP position	Locus	Gene	Count allele	Count allele frequency	Statistical method used to detect the GxMHT interaction	Pthreshold for GxMHT interaction	Pobserved for GxMHT interaction	Pheterogeneity	No. of studies included
Any MHT	rs117868593	12	13670508	12p13.1	GRIN2B	C	0.05	2-step method (by G|E in step 1)	5 × 10−3	.003	.98	38
Any MHT	rs10782186	6	117823508	6q22.1	DCBLD1	C	0.50	2-df joint test	5 × 10−8	4.23 × 10−8	.56	38

^a

Directly genotyped SNPs were coded as 0, 1, or 2 copies of the count allele. Imputed SNPs were coded as expected gene dosage. Multiplicative interaction terms were modeled as the product of MHT and each SNP of interest. All statistical tests were 2-sided. 2-df = 2–degree-of-freedom; Chr = chromosome; BP position = base pair position based on NCBI Build37; MHT = menopausal hormone therapy; SNP = single nucleotide polymorphism; GxMHT = interaction between SNPs and MHT; G|E = associations between SNPs and environmental risk factors in the combined case-control population.

Table 3 presents associations of MHT use with CRC risk by the genotype of the 2 SNPs that were found to be statistically significant. For rs117868593, there was a statistically significant protective effect of any MHT use only among women with the GG homozygotes (OR = 0.68, 95% CI = 0.64 to 0.72; P = 4.3 × 10⁻³⁷) but not in women with the GC genotype (OR = 0.91, 95% CI = 0.77 to 1.09; P = .31) or with the CC genotype (OR = 0.64, 95% CI = 0.22 to 1.85; P = .41). When stratified by MTH use, there was a statistically significant per-minor allele association with CRC risk in users of any MHT (OR = 1.20, 95% CI = 1.05 to 1.37) but not in nonusers (OR = 0.93, 95% CI = 0.83 to 1.03). For rs10782186, the protective effect of any MHT use compared with women not using any MHT was increasingly stronger for women with an increasing number of C alleles: TT (OR = 0.78, 95% CI = 0.70 to 0.87; P = 4.3 × 10⁻⁶), TC (OR = 0.68, 95% CI = 0.63 to 0.73; P = 1.4 × 10⁻²²), and CC (OR = 0.66, 95% CI = 0.60 to 0.74; P = 5.7 × 10⁻¹⁴). When rs10782186 was investigated in relation to CRC risk among strata of MTH use, the per-minor allele odds ratio for CRC risk was attenuated in users of any MHT (OR = 1.05, 95% CI = 0.99 to 1.11) compared with nonusers (OR = 1.14, 95% CI = 1.09 to 1.19).

Table 3.

Associations with colorectal cancer risk stratified by use of any menopausal hormone therapy and genotypes of SNPs of interest^a

SNP	MHT use	Genotype of SNP
		Homozygous noncarriers			Heterozygous			Homozygous carries of the minor allele			Per minor allele within strata of MHT use
		N Ca/Co	OR (95% CI)	P	N Ca/Co	OR (95% CI)	P	N Ca/Co	OR (95% CI)	P	OR (95% CI)	P
rs117868593		GG			GC			CC			Per C allele within strata of MHT use
	No	6991.4/9745.2	1.00 (Referent)	—	652.6/1026.1	0.91 (0.81 to 1.03)	.13	20/23.7	1.09 (0.53 to 2.22)	.81	0.93 (0.83 to 1.03)	0.17
	Yes	3390.5/5537.1	0.68 (0.64 to 0.72)	4.3 × 10−37	450.7/614.9	0.83 (0.72 to 0.96)	.011	13.8/20.1	0.69 (0.31 to 1.54)	.37	1.20 (1.05 to 1.37)	.008
	OR (95% CI)	—	0.68 (0.64 to 0.72)	4.3 × 10−37	—	0.91 (0.77 to 1.09)	.31	—	0.64 (0.22 to 1.85)	.41
rs10782186		TT			TC			CC			Per C allele within strata of MHT use
	No	1861.2/2936.4	1.00 (Referent)	—	3806.9/5361.8	1.15 (1.07 to 1.24)	3.5 × 10−4	1995.9/2496.9	1.29 (1.18 to 1.41)	1.9 × 10−8	1.14 (1.09 to 1.19)	1.8 × 10−8
	Yes	993/1624.8	0.78 (0.70 to 0.87)	4.3 × 10−6	1861.9/3068.3	0.78 (0.71 to 0.85)	3.8 × 10−8	1000/1478.9	0.85 (0.77 to 0.95)	.004	1.05 (0.99 to 1.11)	.14
	OR (95% CI)	—	0.78 (0.70 to 0.87)	4.3 × 10−6	—	0.68 (0.63 to 0.73)	1.4 × 10−22	—	0.66 (0.60 to 0.74)	5.7 × 10−14

a

Case/control counts were calculated by imputed genotype probabilities. Ca/Co = case/control; CI = confidence interval; MHT = menopausal hormone therapy; OR = odds ratio; P = probability value; SNP = single nucleotide polymorphism.

The GxE interactions between rs117868593 or rs10782186 and any MHT were not heterogeneous across studies overall (P = .98, P = .56, respectively) or stratified by study regions (North America, Australia, and Europe). The corresponding forest plots are shown in Supplementary Figures 5 and 6 (available online).

Rare Variants for CRC Risk

The rare variant analysis did not yield any statistically significant interactions (P < 2.5 × 10⁻⁶) for the MHT variables. However, several genes were found to reach the suggestive level for interaction (P < 1.2 × 10⁻⁴) for CRC risk: PREX1 (Phosphatidylinositol-3,4,5-Trisphosphate Dependent Rac Exchange Factor 1) with any MHT use (P = 5.02 × 10⁻⁵), SOS2 (SOS Ras/Rho Guanine Nucleotide Exchange Factor 2) with estrogen-only therapy (P = 9.23 × 10⁻⁵), as well as TMEM189-UBE2V1 (Transmembrane protein 189 - Ubiquitin Conjugating Enzyme E2 V1) (P = 2.46 × 10⁻⁵), FAM149A (Family With Sequence Similarity 149 Member A) (P = 9.67 × 10⁻⁵), and RPS13 (Ribosomal Protein S13) (P = 1.02 × 10⁻⁵) with estrogen-progestogen therapy (Table 4; quantile-quantile plots shown in Supplementary Figures 7-9, available online).

Table 4.

Suggestive association (P < 1.2 × 10⁻⁴) of genes from rare variants analyses of G×E with menopausal hormone therapy for colorectal cancer risk among postmenopausal women^a

MHT type	Gene	Gene name	Chr	No. of SNPs	P
Any MHT	ENSG00000124126	PREX1	20	45	5.02 × 10−5
E-only	ENSG00000100485	SOS2	14	15	9.23  × 10−5
E + P	ENSG00000124208	TMEM189-UBE2V1	20	57	2.46 × 10−5
E + P	ENSG00000109794	FAM149A	4	8	9.67 × 10−5
E + P	ENSG00000110700	RPS13	11	5	1.02 × 10−4

^a

Chr = chromosome; E + P = combined estrogen-progestogen; E-only = estrogen only; MHT = menopausal hormone therapy; P = Fisher P value by the set-based score (MiSTi) test; SNP = single nucleotide polymorphism.

Functional Annotations of Genetic Loci

We performed bioinformatic analysis of the 2 loci showing statistically significant interactions with MHT use (rs117868593 located 20 kb downstream of GRIN2B variant at 12p13.1 and a DCBLD1 intronic variant rs10782186 at 6q22.1). Annotation was performed for all variants tagged by the most statistically significant SNPs (r² > 0.5) using our novel functional annotation analyses. The GRIN2B rs117868593 locus is in LD with rs17822202 (D’ = 0.93 and r² = 0.85 in 1000 Genomes Project CEU), which is downstream of the GRIN2B gene. We noted that this SNP was associated with more pronounced enhancer activity in colon tumor and cancer cell lines than in normal colon tissues (Supplementary Figure 10, available online). The DCBLD1 rs10782186 is in high LD with rs9320604 (D’ = 0.99 and r² = 0.98 in 1000 Genomes Project CEU); a SNP overlapping histone methylation patterns with enhancer activity in normal colon tissues, colon tumor, and cancer cell lines, and associated with strong DNase hypersensitivity in tumor tissues (Supplementary Figure 11, available online).

Based on BarcUVa-Seq expression quantitative trait loci (e QTL) analysis (Supplementary Methods, available online), we identified 4 genes—EMP1(Epithelial membrane protein 1), RPL13AP20 (Ribosomal Protein L13a Pseudogene 20), FAM234B (Family With Sequence Similarity 234 Member B), and CDKN1B (Cyclin Dependent Kinase Inhibitor 1B)—whose expression in normal colon tissue was statistically significantly associated with the SNP rs117868593 or the SNPs in LD (R² > 0.5) (P < .05) (Supplementary Table 5 and Figure 12, available online), as well as 2 genes—ROS1 (ROS Proto-Oncogene 1, Receptor Tyrosine Kinase) and GOPC (Golgi Associated PDZ And Coiled-Coil Motif Containing)—with the SNP rs10782186 or the SNPs in LD (P < .05) (Supplementary Table 6 and Figure 13, available online). These eQTL effects persisted when restricting the sample to postmenopausal women although statistically significant for rs10782186_ROS1, rs117868593_RPL13AP20, and rs1806217_FAM234B.

Discussion

We identified novel GxE interactions between the use of any MHT and common variants at 2 loci for CRC risk among postmenopausal women. The putative target genes underlying these interactions include EMP1, RPL13AP20, FAM234B, CDKN1B, ROS1, and GOPC. In addition, we found suggestive interactions between the use of MHT and rare variants in PREX1, SOS2, TMEM189-UBE2V1, FAM149A, and RPS13. Using independent samples in the current study, the previously found SNPs for GxE interactions (Supplementary Table 7, available online) (7,10) did not show statistically significant interaction with MHT with respect to CRC risk. These earlier studies used a candidate gene approach, different covariable adjustment, or different exposure and nonexposure definitions compared with our GWAS study. Additionally, power could be further reduced by variations in the underlying distribution of MHT as new studies were introduced to the larger cohort.

Currently, the underlying etiologic mechanisms by which MHT affects CRC are not yet well understood. It is likely that protective cellular effects of estrogen and progesterone in the development of CRC are mediated through estrogen receptor α, estrogen receptor β (ESR2), and progesterone receptor (28-30). Estrogen and progestin may play a role in the pathway leading to DNA hypermethylation (31,32), which regulates gene expression including that of tumor suppressor genes and thereby play a crucial role in tumorigenesis of CRC. Estrogen has also been found to have an impact on a large number of serum proteome, which plays a role in mucosal protection and repair in the gastrointestinal tract (33) as well as colon transcriptome (34). In addition, estrogen may contribute to maintaining the genomic stability in colonic epithelial cells by upregulation of mismatch repair genes (35). MHT use has also been reported to have growth-inhibiting effects on colon cancer cells through upregulating cell cycle regulators (eg, TP53) (36). Consortium efforts that are powered to explore the relationships of MHT with specific subtypes of CRC may yield further insights to GxE interactions with respect to hormonal contributions to the pathogenesis of CRC (37).

The SNP rs117868593 located about 20 kb downstream from GRIN2B was not found to be associated with expression of the nearest gene GRIN2B but with EMP1, RPL13AP20, FAM234B, and CDKN1B. Expression of EMP1 has been found to be lower in human CRC than normal adjacent colorectal tissues (38), and overexpression of EMP1 was observed to reduce proliferation and induce apoptosis of CRC cells (39), which are consistent with our findings, that is, lower expression of EMP1 and higher risk of CRC associated with G allele of rs117868593. We found the MHT users with GG have a stronger statistically significant reduction of CRC risk, suggesting that EMP1 may function as an oncogene in hormone-dependent epithelium, which has been observed for EMP2, a paralog of EMP1 (40). Downregulation of CDKN1B, which mainly results from increased ubiquitin-mediated proteasomal degradation, has been associated with tumor progression in CRC (41), and CDKN1B could be induced through ESR2-mediated repression of the F-box protein p45 (SKP2), which has been identified as the substrate recognition component that targets and binds CDKN1B for ubiquitination and subsequent degradation (41-43). The link between CDKN1B and ESR2 might explain the observed interaction of CDKN1B with MHT. Potential mechanisms through which RPL13AP20 and FAM234B act in modifying MHT-associated CRC risk are unknown.

The region in which DCBLD1 is located, chromosome 6q22.1, has been reported as one of the suggestive susceptibility regions (P = 3.20 × 10⁻⁶) in a GWAS meta-analysis on CRC risk (12). Association estimates for the index SNP rs10782186 and correlated SNPs (rs4945586, rs9320604, and rs4946260) reported in the above-mentioned GWAS paper. The significance (P = 4.23 × 10⁻⁸) of the interaction in our GWIS using the 2-df joint test was mainly driven by the genetic association (P = 6.79 × 10⁻⁸) and was further strengthened by the GxE product term (P = .03). Thus, incorporating the GxE component helped uncover genetic susceptibility variants for CRC risk, which did not reach genome-wide significance level in GWAS. Analyses of associated gene expression indicated the involvement of ROS1 and GOPC. ROS1 is a transmembrane receptor tyrosine kinase that often shows genetic rearrangements in colorectal tumor tissue, such as intrachromosomal fusion with GOPC because of microdeletions at 6q22.1, which is highly prevalent in CRC (44,45). GOPC-ROS fusion proteins have been shown to activate the downstream signaling pathway, signal transducers, and activators of transcription-3 that play a important role in progression of CRC (45,46). The transcription factor of signal transducers and activators of transcription-3 in epithelial cells is activated by interleukin–6, promoting CRC tumorigenesis (29,47), whereas ESR2 mediates the downregulation of the inflammatory cytokine interleukin–6 network (48), which may explain the observed interaction with MHT.

There are still considerable challenges in investigating GxE interaction of rare genetic variants because of the scarcity of subjects with data on both these variants and the relevant environmental and lifestyle exposures. Therefore, the role that rare predisposition alleles play in modifying the association between environmental factors and CRC risk remains poorly understood. Our study used MiSTi to tackle the challenge for GxE interaction analysis of rare variants, which strengthened statistical power to robustly uncover potential rare variant GxE association signals. Through this method, we found suggestive interaction for MHT use with rare variants in 5 genes for CRC risk. Despite their as yet unknown mechanisms in modifying CRC risk associated with MHT use, our application of GxE interaction analysis for CRC risk to rare variants alongside common variants represents a novel and rigorous approach. GxE interaction studies of rare genetic variants that incorporate functional genomic information ideally accounting for MHT effects and studies with larger sample sizes and hence with greater statistical power may contribute to understanding any missing heritability of cancer that remains unexplained by common variants.

Our study has several strengths. First, our large sample size, including more than 28 000 participants, facilitated the most powerful scan for gene-MHT interaction to date. Second, we used recently developed statistical approaches that can provide greater statistical power than conventional case-control logistic regression (49). Because no single approach provides the best power across all possible patterns of GxE interaction, we used a combination of approaches to maximize the chance of identifying novel loci in this discovery analysis. MiSTi, used for rare variant analysis, helped identify suggestive associations with CRC risk through interaction with MHT for 5 genes that warrant further follow-up. Third, we carefully harmonized environmental data on MHT use and other covariates across studies to minimize between-study heterogeneity bias as previously described (11). We acknowledge, however, that our analysis was limited to populations of European ancestry; thus the results might not be generalizable to other race and ethnicity groups. Measurement error of the primarily self-reported exposure assessment might also have contributed to reduced power; however, previous studies have found the high validity for self-reported MHT use when compared with population-based prescription databases as references (50) and a high concordance between self-reported MHT use and that of physicians’ reports (51). Despite our sizable sample size and use of advanced statistical methods, we acknowledge that statistical power remains limited to detect small to modest-sized interaction effects in a genome-wide scan setting. This might explain the relatively small number of novel findings. To overcome these issues, it will be critical to expand sample sizes of well-characterized studies as well as incorporated functional genomic data relevant to CRC and MHT use, such as multi-omics data of normal and tumor colon tissue exposed and unexposed to MHT.

From a comprehensive genome-wide GxE interaction investigation, we identified 2 common loci, which were statistically significantly associated with CRC risk in conjunction with MHT use, as well as 5 genes, which showed suggestive evidence of GxE interaction through rare variant set analysis. The putative target genes of the 2 identified loci (EMP1, RPL13AP20, FAM234B, CDKN1B, ROS1, and GOPC) may explain the GxE interactions with MHT and offer new insights into CRC etiological mechanisms and pathways of CRC carcinogenesis. Further downstream, follow-up studies for exploring potential genetic functions are warranted to confirm the involvement of these genetic variants or genes in CRC risk associated with MHT use.

Funding

Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO): National Cancer Institute, National Institutes of Health, U.S. Department of Health and Human Services (U01 CA137088, R01 CA059045, U01 CA164930, R01201407). Genotyping/services were provided by the Center for Inherited Disease Research (CIDR) contract number HHSN268201200008I. This research was funded in part through the NIH/NCI Cancer Center Support Grant P30 CA015704. Scientific Computing Infrastructure at Fred Hutch funded by ORIP grant S10OD028685.

CLUE II: National Cancer Institute (U01 CA86308, Early Detection Research Network; P30 CA006973), National Institute on Aging (U01 AG18033), and the American Institute for Cancer Research. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.

The Colon Cancer Family Registry (CCFR, www.coloncfr.org) is supported in part by funding from the National Cancer Institute (NCI), National Institutes of Health (NIH) (award U01 CA167551). Support for case ascertainment was provided in part from the Surveillance, Epidemiology, and End Results (SEER) Program and the following U.S. state cancer registries: AZ, CO, MN, NC, NH; and by the Victoria Cancer Registry (Australia) and Ontario Cancer Registry (Canada). The CCFR Set-1 (Illumina 1M/1M-Duo) scan was supported by NIH awards U01 CA122839 and R01 CA143247 (to GC). The CCFR Set-3 (Affymetrix Axiom CORECT Set array) was supported by NIH award U19 CA148107 and R01 CA81488 (to SBG). The CCFR Set-4 (Illumina OncoArray 600K SNP array) was supported by NIH award U19 CA148107 (to SBG) and by the Center for Inherited Disease Research (CIDR), which is funded by the NIH to the Johns Hopkins University, contract number HHSN268201200008I. Additional funding for the OFCCR/ARCTIC was through award GL201-043 from the Ontario Research Fund (to BWZ), award 112746 from the Canadian Institutes of Health Research (to TJH), through a Cancer Risk Evaluation (CaRE) Program grant from the Canadian Cancer Society (to SG), and through generous support from the Ontario Ministry of Research and Innovation. The SFCCR Illumina HumanCytoSNP array was supported in part through NCI/NIH awards U01/U24 CA074794 and R01 CA076366 (to PAN). The content of this manuscript does not necessarily reflect the views or policies of the NCI, NIH, or any of the collaborating centers in the Colon Cancer Family Registry (CCFR), nor does mention of trade names, commercial products, or organizations imply endorsement by the US government, any cancer registry, or the CCFR.

COLO2&3: National Institutes of Health (R01 CA60987).

Colorectal Cancer Transdisciplinary (CORECT) Study: National Cancer Institute, National Institutes of Health (NCI/NIH), U.S. Department of Health and Human Services (grant numbers U19 CA148107, R01 CA81488, P30 CA014089, R01 CA197350; P01 CA196569; R01 CA201407) and National Institutes of Environmental Health Sciences, National Institutes of Health (grant number T32 ES013678).

CPS-II: The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study-II (CPS-II) cohort. This study was conducted with institutional review board approval.

CRCGEN: Colorectal Cancer Genetics & Genomics, Spanish study was supported by Instituto de Salud Carlos III, co-funded by FEDER funds -a way to build Europe- (grants PI14-613 and PI09-1286), Agency for Management of University and Research Grants (AGAUR) of the Catalan Government (grant 2017SGR723), and Junta de Castilla y León (grant LE22A10-2). Sample collection of this work was supported by the Xarxa de Bancs de Tumors de Catalunya sponsored by Pla Director d’Oncología de Catalunya (XBTC), Plataforma Biobancos PT13/0010/0013 and ICOBIOBANC, sponsored by the Catalan Institute of Oncology.

DACHS: German Research Council (BR 1704/6-1, BR 1704/6-3, BR 1704/6-4, CH 117/1-1, HO 5117/2-1, HE 5998/2-1, KL 2354/3-1, RO 2270/8-1 and BR 1704/17-1), the Interdisciplinary Research Program of the National Center for Tumor Diseases (NCT), Germany, and the German Federal Ministry of Education and Research (01KH0404, 01ER0814, 01ER0815, 01ER1505A and 01ER1505B).

DALS: National Institutes of Health (R01 CA48998 to M L Slattery).

EPIC: The coordination of EPIC is financially supported by International Agency for Research on Cancer (IARC) and also by the Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, which has additional infrastructure support provided by the NIHR Imperial Biomedical Research Centre (BRC). The national cohorts are supported by Danish Cancer Society (Denmark); Ligue Contre le Cancer, Institut Gustave Roussy, Mutuelle Générale de l’Education Nationale, Institut National de la Santé et de la Recherche Médicale (INSERM) (France); German Cancer Aid, German Cancer Research Center (DKFZ), German Institute of Human Nutrition Potsdam- Rehbruecke (DIfE), Federal Ministry of Education and Research (BMBF) (Germany); Associazione Italiana per la Ricerca sul Cancro-AIRC-Italy, Compagnia di SanPaolo and National Research Council (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF), Statistics Netherlands (The Netherlands); Health Research Fund (FIS) - Instituto de Salud Carlos III (ISCIII), Regional Governments of Andalucía, Asturias, Basque Country, Murcia and Navarra, and the Catalan Institute of Oncology—ICO (Spain); Swedish Cancer Society, Swedish Research Council and County Councils of Skåne and Västerbotten (Sweden); Cancer Research UK (14136 to EPIC-Norfolk; C8221/A29017 to EPIC-Oxford), Medical Research Council (1000143 to EPIC-Norfolk; MR/M012190/1 to EPIC-Oxford) (United Kingdom).

ESTHER_VERDI: This work was supported by grants from the Baden-Württemberg Ministry of Science, Research and Arts and the German Cancer Aid.

Harvard cohort (NHS): National Institutes of Health (R01 CA137178, P01 CA087969, UM1 CA186107, K24 DK098311, R01 CA151993, and R35 CA197735).

Kentucky: Clinical Investigator Award from Damon Runyon Cancer Research Foundation (CI-8); NCI R01CA136726.

LCCS: The Leeds Colorectal Cancer Study was funded by the Food Standards Agency and Cancer Research UK Programme Award (C588/A19167).

MCCS: The cohort recruitment was funded by VicHealth and Cancer Council Victoria. The MCCS was further supported by Australian NHMRC grants 509348, 209057, 251553, and 504711 and by infrastructure provided by Cancer Council Victoria. Cases and their vital status were ascertained through the Victorian Cancer Registry (VCR) and the Australian Institute of Health and Welfare (AIHW), including the National Death Index and the Australian Cancer Database.

MEC: National Institutes of Health (R37 CA054281, P01 CA033619, and R01 CA063464).

MECC: National Institutes of Health, U.S. Department of Health and Human Services (R01 CA081488, R01 CA197350).

NCCCS I and II: National Institutes of Health (R01 CA66635 and P30 DK034987).

NFCCR: This work was supported by an Interdisciplinary Health Research Team award from the Canadian Institutes of Health Research (CRT 43821); the National Institutes of Health, U.S. Department of Health and Human Services (U01 CA74783); and National Cancer Institute of Canada grants (18223 and 18226). The authors wish to acknowledge the contribution of Alexandre Belisle and the genotyping team of the McGill University and Génome Québec Innovation Centre, Montréal, Canada, for genotyping the Sequenom panel in the NFCCR samples. Funding was provided to Michael O. Woods by the Canadian Cancer Society Research Institute.

NSHDS: The research was supported by Biobank Sweden through funding from the Swedish Research Council (VR 2017-00650, VR 2017-01737), the Swedish Cancer Society (CAN 2017/581), Region Västerbotten (VLL-841671, VLL-833291), Knut and Alice Wallenberg Foundation (VLL-765961), and the Lion’s Cancer Research Foundation (several grants) and Insamlingsstiftelsen, both at Umeå University.

PLCO: Intramural Research Program of the Division of Cancer Epidemiology and Genetics and supported by contracts from the Division of Cancer Prevention, National Cancer Institute, NIH, DHHS. Funding was provided by National Institutes of Health (NIH), Genes, Environment and Health Initiative (GEI) Z01 CP 010200, NIH U01 HG004446, and NIH GEI U01 HG 004438.

REACH: National Cancer Institute (grant P01 CA074184 to JDP and PAN, grants R01 CA097325, R03 CA153323, and K05 CA152715 to PAN, and the National Center for Advancing Translational Sciences at the National Institutes of Health (grant KL2 TR000421 to ANB-H)

SMC_COSM: This work is supported by the Swedish Research Council/Infrastructure grant, the Swedish Cancer Foundation, and the Karolinska Institutés Distinguished Professor Award to Alicja Wolk.

UK Biobank: This research has been conducted using the UK Biobank Resource under Application Number 8614

VITAL: National Institutes of Health (K05 CA154337).

WHI: The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201100046C, HHSN268201100001C, HHSN268201100002C, HHSN268201100003C, HHSN268201100004C, and HHSN271201100004C.

Notes

Role of the funders: The funders had no role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.

Disclosures: All authors have no conflicts of interest to disclose. KV, who is a JNCI Associate Editor and co-author on this paper, was not involved in the editorial review or decision to publish the manuscript.

Author contributions: Writing—original draft: YT, AEK. Analysis—statistical methods: YT, AEK, YL, CQ, VDO. Supervising: UP, WJG, LH, JCC. Writing- reviewing & editing: YT, AEK, SAB, YL, CQ, TH, RCT, VDO, ND, DAD, AH, JRH, KMJ, JM, NM, MOS, CMU, JO, ARP, EAR, AS, MS, YS, FJD, VA, JB, SIB, DTB, HB, DDB, ATC, JCF, SG, SBG, SH, MH, MAJ, ADJ, TOK, SCL, LLM, LL, GGG, RLM, HN, RN, SO, AB, EAP, JDP, RLP, GR, LCS, RES, MLS, SNT, BVG, KV, EW, AW, MOW, AHW, PTC, GC, DVC, MJG, AK, JPL, VM, PAN, BP, DCT, KKT, UP, WJG, LH, JCC.

Acknowledgments: We thank Ferran Moratalla Navarro (Oncology Data Analytics Program, Catalan Institute of Oncology [ICO]). L’Hospitalet de Llobregat, Barcelona, Spain. Colorectal Cancer Group, ONCOBELL Program, Bellvitge Biomedical Research Institute (IDIBELL). L’Hospitalet de Llobregat, Barcelona, Spain. Consortium for Biomedical Research in Epidemiology and Public Health, (CIBERESP), Spain. Department of Clinical Sciences, Faculty of Medicine, University of Barcelona, Barcelona, Spain for providing a further eQTL analysis for the revision of our manuscript.

CCFR: We graciously thank the generous contributions of CCFR study participants, dedication of study staff, and the financial support from the U.S. National Cancer Institute, without which this important registry would not exist. The authors would like to thank the study participants and staff of the Seattle Colon Cancer Family Registry and the Hormones and Colon Cancer study (CORE Studies).

CLUE II: We thank the participants of Clue II and appreciate the continued efforts of the staff at the Johns Hopkins George W. Comstock Center for Public Health Research and Prevention in the conduct of the Clue II Cohort Study. Cancer data were provided by the Maryland Cancer Registry, Center for Cancer Prevention and Control, Maryland Department of Health, with funding from the State of Maryland and the Maryland Cigarette Restitution Fund. The collection and availability of cancer registry data is also supported by the Cooperative Agreement NU58DP006333, funded by the Centers for Disease Control and Prevention. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention or the Department of Health and Human Services.

CPS-II: The authors thank the CPS-II participants and Study Management Group for their invaluable contributions to this research. The authors would also like to acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention National Program of Cancer Registries, and cancer registries supported by the National Cancer Institute Surveillance Epidemiology and End Results program.

DACHS: We thank all participants and cooperating clinicians, and everyone who provided excellent technical assistance.

EPIC: We acknowledge the contributions of EPIC investigators, staff, and all participants.

Harvard cohorts (NHS): The study protocol was approved by the institutional review boards of the Brigham and Women’s Hospital and Harvard T.H. Chan School of Public Health, and those of participating registries as required. We would like to thank the participants and staff of the HPFS, NHS, and PHS for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY. The authors assume full responsibility for analyses and interpretation of these data.

Kentucky: We acknowledge the staff at the Kentucky Cancer Registry.

LCCS: We acknowledge the contributions of Jennifer Barrett, Robin Waxman, Gillian Smith, and Emma Northwood in conducting this study.

NCCCS I and II: We thank the study participants, and the NC Colorectal Cancer Study staff.

NSHDS: We thank the Västerbotten Intervention Programme, the Northern Sweden MONICA study, the Biobank Research Unit at Umeå University, and Biobanken Norr at Region Västerbotten for providing data and samples and acknowledge the contribution from Biobank Sweden, supported by the Swedish Research Council.

PLCO: The authors thank the PLCO Cancer Screening Trial screening center investigators and the staff from Information Management Services Inc and Westat Inc. Most importantly, we thank the study participants for their contributions that made this study possible. Cancer incidence data have been provided by the District of Columbia Cancer Registry, Georgia Cancer Registry, Hawaii Cancer Registry, Minnesota Cancer Surveillance System, Missouri Cancer Registry, Nevada Central Cancer Registry, Pennsylvania Cancer Registry, Texas Cancer Registry, Virginia Cancer Registry, and Wisconsin Cancer Reporting System. All are supported in part by funds from the Center for Disease Control and Prevention, National Program for Central Registries, local states or by the National Cancer Institute, Surveillance, Epidemiology, and End Results program. The results reported here and the conclusions derived are the sole responsibility of the authors.

WHI: The authors thank the WHI investigators and staff for their dedication and the study participants for making the program possible. A full listing of WHI investigators can be found at http://www.whi.org/researchers/Documents%20%20Write%20a%20Paper/WHI%20Investigator%20Short%20List.pdf.

Disclaimer: Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/World Health Organization.

Prior presentations: The content in the manuscript was partly reported as poster presentation in the 2nd International DKFZ Conference on Cancer Prevention on September 17-18, 2020, Heidelberg, Germany.

Data Availability

The data underlying this article are available in dbGaP at https://www.ncbi.nlm.nih.gov/gap/, and can be accessed with accession numbers phs001415.v1.p1, phs001315.v1.p1, phs001078.v1.p1, phs001499.v1.p1, phs001903.v1.p1, phs001856.v1.p1, phs001045.v1.p1, and phs001499.v1.p1.

Author notes

†

Yu Tian and Andre E. Kim contributed equally to this work.

‡

Ulrike Peters, W. James Gauderman, Li Hsu and Jenny Chang-Claude jointly supervised this work.