Common and rare sequence variants influencing tumor biomarkers in blood

Abstract

Background

Alpha-fetoprotein, cancer antigens 15.3, 19.9, 125, carcinoembryonic antigen and alkaline phosphatase are widely measured in attempts to detect cancer and to monitor treatment response. However, due to lack of sensitivity and specificity, their utility is debated. The serum levels of these markers are affected by a number of non-malignant factors, including genotype. Thus, it may be possible to improve both sensitivity and specificity by adjusting test results for genetic effects.

Methods

We performed genome-wide association studies of serum levels of alpha-fetoprotein (N = 22,686), carcinoembryonic antigen (N = 22,309), cancer antigens 15.3 (N = 7,107), 19.9 (N = 9,945) and 125 (N = 9,824), and alkaline phosphatase (N = 162,774). We also examined the correlations between levels of these biomarkers and the presence of cancer, using data from a nation-wide cancer registry.

Results

We report a total of 84 associations of 79 sequence variants with levels of the six biomarkers, explaining between 2.3 and 42.3% of the phenotypic variance. Among the 79 variants, 22 are cis (in- or near the gene encoding the biomarker), 18 have minor allele frequency less than 1%, 31 are coding variants and 7 are associated with gene expression in whole blood. We also find multiple conditions associated with higher biomarker levels.

Conclusions

Our results provide insights into the genetic contribution to diversity in concentration of tumor biomarkers in blood.

Impact

Genetic correction of biomarker values could improve prediction algorithms and decision-making based on these biomarkers.

Introduction

Tumor biomarkers are substances or processes that can indicate the presence of cancer [1]. Several tumor biomarkers are in clinical use for monitoring therapy but all lack the sensitivity and specificity to be used for screening. However, recent advances in the detection of circulating tumor DNA suggest that multi-analyte blood tests that combine an assay of somatically mutated DNA (“liquid biopsy”) and protein and carbohydrate biomarkers in serum have the potential to both find early cancer and to help determine its site of origin [2].

In this work, we focused on six commonly measured biomarkers, namely alpha-fetoprotein (AFP), carcinoembryonic antigen (CEA), cancer antigens (CA) 15.3, 19.9 and 125, and alkaline phosphatase (ALP). Measured in serum, these biomarkers are frequently used to monitor status of disease, response to therapy and recurrence [1]. AFP is used as a biomarker of hepatocellular carcinoma (HCC), endodermal sinus tumor of the ovary and non-seminoma testicular germ cell tumors (TGCT) [3]. CEA has been used as a biomarker for colorectal cancer [4]. CA-15.3 and CA-125 are mainly used as biomarkers of cancers of the breast and ovary, respectively [5, 6], and CA-19.9 is used as a biomarker for pancreatic cancer [7]. We also include ALP in our analysis because its levels are commonly elevated in cancers of the liver and bone and when other cancers metastasize to these tissues [8]. However, the measurement of ALP in serum is one of the most common blood tests ordered and we recognize that there are many reasons for ALP measurements other than suspicion of- or monitoring of neoplasms.

Despite widespread use of these biomarkers in clinical practice, their low sensitivity and specificity continue to cause controversy over their use [9–11]. As their levels are partially determined by genetic factors, one approach to improve their sensitivity and specificity would be to define “normal” values based on age, sex and genotype [2, 9]. We have previously reported how genetic correction for variants affecting levels of prostate specific antigen (PSA) results in personalized PSA cutoff value, which is more informative than a general cutoff value when deciding to perform a prostate biopsy [12].

The main goal of this study is to perform a genome-wide association study (GWAS) of the levels of all six tumor biomarkers to identify sequence variants that affect baseline biomarker levels, regardless of cancer diagnosis. We also describe the associations of the six tumor biomarker levels with various cancer diagnoses, obtained from a nation-wide cancer registry, and for comparison, with four non-neoplastic diseases.

Materials and Methods

Cancer diagnoses, including the date of diagnosis, were extracted from the Icelandic cancer registry (ICR) (http://www.cancerregistry.is), which contains all diagnoses of solid cancers made in in the country from January 1st 1955 to December 31st 2015 [13]. We also assessed four other diseases; inflammatory bowel disease (IBD), liver cirrhosis and pancreatitis, because these diseases are associated with inflammation in the gastrointestinal organs and fibromyalgia, because patients present with diverse symptoms and often undergo measurements for tumor biomarkers as part of a lengthy diagnosis journey. Tumor biomarker measurements were made in Icelandic laboratories from 1990 to 2015 and linked with disease diagnoses on the basis of encrypted social security numbers.

This study was approved by the National Bioethics Committee of Iceland (reference numbers VSNb2006010014/03.12, 06–007-V3 and VSNb201501033/03.12). A further description of subject recruitment, phenotyping, genotyping and imputation is available in the supplementary methods.

Genome wide association study

To identify genetic variants associated with baseline biomarker values, we performed a GWAS of all available data, including both cancer patients and individuals without cancer diagnosis. When multiple measurements of a biomarker were available for a subject, we used the earliest value recorded. We found this approach to be the most powerful as exclusion of cancer patients resulted in great loss of power, for residual associations (secondary, tertiary variants etc) in particular. The first measurement was used as this meant that measurements taken months or years before cancer diagnosis were available for a subset of the cancer patients.

As indicated in Table 1 the biomarkers all show extremely right-skewed distributions. The data contains a number of extreme outliers but no trends were observed to link these with date of measurement, age of the subject or even cancer type (Supplementary figure 1).

Table 1:

Tumor markers and subjects used in this study. AFP: Alpha-fetoprotein; CA: Cancer antigen; CEA: Carcinoembryonic antigen; ALP: Alkaline phosphatase.

Biomarker	Total subjects	N (% Female)	Measurement unit	Avg. age at first measurement (range)	Median first value (range)	Median largest value (range)	Reference valuea
AFP	22,686	12,886 (56.8)	U/ml	57 (0 – 100)	3.0 (0.4 – 385,725)	3.2 (0.4 – 385,725)	<5.8
CA-15.3	7,107	6,304 (88.7)	U/ml	62 (1 – 99)	17.7 (1.0 – 17,340)	20.8 (1.0 – 47,600)	<25
CA-125	9,824	9,087 (92.5)	U/ml	58 (1 – 103)	16.6 (0.6 – 56,820)	18.0 (0.6 – 63,062)	<35
CA-19.9	9,945	5,708 (57.4)	U/ml	66 (7 – 101)	16.3 (0.0 – 1,769,950)	18.9 (0.0 – 16,571,700)	<31
CEA	22,309	13,095 (58.7)	ng/ml	63 (0 – 103)	2.3 (0.0 – 63,962)	2.7 (0.0 – 116,069)	<4.6
ALP	162,774	87,897 (54.0)	U/ml	46 (0 – 114)	119.0 (0.0 – 15,825)	134.0 (5.0 – 21,668)	<105

^aThe reference value is that generally considered based on the respective tests used in Landspitali, the National University Hospital in Iceland.

We performed a rank-based inverse normal transformation adjusting for age at measurement and time to death for deceased subjects for each gender separately. Adjustment for time to death was performed as we have observed large changes close to the time of death for many quantitative measurements. In this case, a high biomarker value shortly before death might indicate a high tumor burden. We tested the association between biomarker value and genotype by a generalized form of linear regression. To assess significance of primary associations, we used different P value thresholds depending on the annotation class of the variant as described in Sveinbjornsson et al [14]. We consider loss-of-function variants (frameshifts, stop codon gained/lost, initiator codon variants and splice acceptor/donor variants) significant at 3.6 × 10⁻⁷, in-frame insertions/deletions, missense and splice region variants at 7.4 × 10⁻⁸, synonymous variants, up/downstream variants, variants that resulted in a stop-codon being retained, and variants in 3’/5’ untranslated regions at 5.3 × 10⁻⁹, intronic and intergenic variants within DNase hypersensitivity sites at 3.3 × 10⁻⁹ and intergenic and intronic variants outside DNase hypersensitivity sites at 1.1 × 10⁻⁹. We used the Variant Effect Predictor (VEP) release 80 [15] to annotate variants, considering only protein-coding transcripts from RefSeq release 67 [16].

Loci harboring variants associated by these criteria underwent further analyses to check for the presence of other, independent variants affecting the trait. While any variant passed the significance threshold at a locus, we tested all variants flanking the primary signal by 1–13 Mb, depending on the strength of association and the recombination pattern at the locus, by sequentially adding the top variant from previous steps as covariate in the regression. The use of wide windows was occasionally necessary to avoid spurious associations arising from very low, but not negligible, LD between an extremely significant variant and distant variants. Residual associations were generally close to the primary signal, as can be seen in Tables 2 and 3. We considered significant variants that passed a simple Bonferroni correction for the number of variants in the respective region. This process was repeated, adding the top variant as covariate for the next round until no significant associations remained. The gene expression and co-localization analyses are described in the supplementary methods.

Table 2:

Primary variants within loci associating with tumor biomarker levels.

rsID	Chromosome: Positiona	MAF [%]	Amin	Amaj	Gene	Annotation	P	PCorrectedb	Effectc	R2
Alpha-fetoprotein
rs28929474	chr14:94378610	0.8	T	C	SERPINA1	Glu366Lys	1.2 × 10−47	8.3 × 10−42	−0.85 (−0.96, −0.74)	0.012
Cancer antigen 15.3
rs760077	chr1:155208991	35.4	A	T	MTX1	Ser63Thr	< 1 × 10−300	< 1 × 10−300	0.86 (0.81, 0.90)	0.34
NA	chr9:133264504	19.5	G	GAAA CTGCC	ABO	intron	1.8 × 10−47	8.0 × 10−40	−0.33 (−0.38, −0.29)	0.035
Cancer antigen 125
rs62193080	chr2:241800675	20.6	G	C	GAL3ST2	intron	2.9 × 10−57	1.3 × 10−49	−0.32 (−0.36, −0.28)	0.033
rs3764246	chr16:760143	23.6	G	A	MSLN	upstream	3.2 × 10−15	3.0 × 10−8	−0.15 (−0.18, −0.11)	0.008
rs73005873	chr19:8896954	39.7	A	G	MUC16	intron	1.8 × 10−17	8.2 × 10−10	0.14 (0.11, 0.17)	0.009
Cancer antigen 19.9
rs708686	chr19:5840608	23.1	T	C	FUT6	upstream	1.9 × 10−179	1.8 × 10−172	−0.52 (−0.56, −0.49)	0.097
rs601338	chr19:48703417	39.3	G	A	FUT2	Trp154Ter	1.3 × 10−291	1.8 × 10−286	−0.57 (−0.61, −0.54)	0.16
rs34262244	chr19:17795098	28.5	A	G	B3GNT3	upstream	3.0 × 10−13	2.8 × 10−6	−0.13 (−0.17, −0.10)	0.0070
Carcinoembryonic antigen
rs7041150	chr9:106732343	37.8	A	C	—	intergenic	1.8 × 10−12	2.7 × 10−5	−0.08 (−0.10, −0.06)	0.0030
rs635634	chr9:133279427	13.0	T	C	ABO	upstream	9.4 × 10−16	8.8 × 10−9	−0.13 (−0.16, −0.10)	0.0039
rs708686	chr19:5840608	23.1	T	C	FUT6	upstream	4.6 × 10−16	4.3 × 10−9	0.10 (0.08, 0.13)	0.0039
rs9621	chr19:41727239	5.6	A	G	CEACAM5	Gly678Arg	5.2 × 10−201	3.5 × 10−195	0.71 (0.67, 0.76)	0.054
rs601338	chr19:48703417	39.3	G	A	FUT2	Trp154Ter	5.1 × 10−130	6.9 × 10−125	−0.27 (−0.29, −0.25)	0.035
Alkaline phosphatase
rs149344982	chr1:21563267	1.5	A	G	ALPL	Arg75His	1.4 × 10−157	9.5 × 10−146	−0.46 (−0.49, −0.43)	0.0061
rs1862069	chr2:169077231	49.6	A	G	DHRS9	upstream	6.1 × 10−10	0.0057	−0.03	0.00034
rs1260326	chr2:27508073	34.1	T	C	GCKR	Leu446Pro	4.6 × 10−11	3.1 × 10−5	0.03	0.00038
rs573778305	chr6:24429112	0.8	C	CT	GPLD1	Frameshift Val815	1.8 × 10−111	2.4 × 10−106	−0.53 (−0.58, −0.48)	0.0042
rs62621812	chr7:127375029	2.5	A	G	ZNF800	Pro103Ser	6.7 × 10−9	0.0045	−0.08 (−0.10, −0.05)	0.00029
rs6984305	chr8:9320758	8.4	A	T	—	intergenic	1.7 × 10−16	2.5 × 10−9	0.06 (0.05, 0.08)	0.00057
rs4242592	chr8:118956736	47.9	T	G	TNFRSF11B	upstream	2.5 × 10−10	0.0023	−0.03 (−0.03, −0.02)	0.00034
rs28601761	chr8:125487789	43.0	G	C	—	intergenic	4.0 × 10−15	1.8 × 10−7	−0.03 (−0.04, −0.02)	0.00053
rs41282145	chr9:101487225	4.3	A	T	TMEM246	upstream	1.4 × 10−14	1.3 × 10−7	0.08 (0.06, 0.10)	0.00050
NA	chr9:133264504	19.5	G	GAAA CTGCC	ABO	intron	< 1 × 10−300	< 1 × 10−300	−0.20 (−0.21, −0.19)	0.013
rs1935	chr10:63168063	47.6	C	G	JMJD1C	Glu2353Asp	1.7 × 10−29	1.2 × 10−23	−0.05 (−0.06, −0.04)	0.0011
rs10790256	chr11:118663373	22.2	T	C	TREH	synonymous	1.4 × 10−9	0.013	−0.03 (−0.04, −0.02)	0.00031
rs10893507	chr11:126416693	48.1	A	C	ST3GAL4	downstream	9.2 × 10−18	8.6 × 10−11	0.04 (0.03, 0.04)	0.00065
rs7955258	chr12:461781	44.5	A	G	B4GALNT3	intron	1.3 × 10−14	5.9 × 10−7	−0.03 (−0.04, −0.02)	0.00051
rs10849087	chr12:4540899	27.1	T	C	C12orf4	upstream	1.1 × 10−9	0.0010	−0.03 (−0.04, −0.02)	0.00031
rs2393791	chr12:120986153	35.4	C	T	HNF1A	intron	3.4 × 10−13	5.0 × 10−6	0.03 (0.02, 0.04)	0.00044
rs9533095	chr13:42394913	43.5	G	T	—	intergenic	7.6 × 10−10	0.011	−0.03 (−0.03, −0.02)	0.00033
rs28929474	chr14:94378610	0.8	T	C	SERPINA1	Glu366Lys	3.4 × 10−17	2.3 × 10−11	0.20 (0.15, 0.24)	0.00062
rs2297066	chr14:103100498	21.9	G	C	EXOC3L4	Asp93Glu	7.7 × 10−9	0.0052	0.03 (0.02, 0.04)	0.00029
rs71391445	chr16:72171122	18.1	G	GA	PMFBP1	intron	4.6 × 10−13	6.9 × 10−6	0.04 (0.03, 0.05)	0.00045
rs186021206	chr17:7166093	0.4	A	G	—	intergenic	7.3 × 10−89	3.3 × 10−81	0.63 (0.57, 0.69)	0.0034
rs5112	chr19:44927023	48.6	G	C	—	intergenic	1.7 × 10−16	7.7 × 10−9	−0.03 (−0.04, −0.03)	0.00058
rs8736	chr19:54173495	41.6	T	C	TMC4	upstream	8.1 × 10−16	7.6 × 10−9	0.03 (0.03, 0.04)	0.00056
rs2500430	chr20:25298327	49.2	G	A	ABHD12 PYGB	downstream	6.95 × 10−10	0.0065	−0.03 (−0.03, −0.02)	0.00034
rs41302559	chr20:57565383	0.9	A	G	PCK1	Arg483Gln	2.90 × 10−8	0.020	−0.12 (−0.17, −0.08)	0.00026

^aThe position is given in hg38

^bThe P-value after a weighted Bonferroni adjustment where a different threshold was used for each functional class (see methods).

^cEffect of the minor allele (Amin) is reported as standard deviations of rank-based inverse normally transformed data

Table 3:

Residual associations within loci associating with tumor biomarker levels.

rsID	Chromosome: Positiona	MAF [%]	Amin	Amaj	Gene	Annotation	P	Effect†	PCorrectb	Effectc	R2	Covariate(s)
Alpha-fetoprotein
rs17580	chr14:94380925	3.3	A	T	SERPINA1	Glu288Val	8.3 × 10−38	−0.38 (−0.43, −0.32)	2.3 × 10−33	−0.38 (−0.43, −0.32)	0.0091	rs28929474
rs2402446	chr14:94399800	47.6	T	A	—	intergenic	7.5 × 10−9	0.06 (0.04, 0.08)	2.1 × 10−5	0.06 (0.04, 0.08)	0.0020	rs28929474, rs17580
Cancer antigen 15.3
rs41264915	chr1:155197995	14.2	G	A	THBS3	intron	2.5 × 10−63	0.38 (0.34, 0.43)	5.8 × 10−59	0.38 (0.34, 0.43)	0.035	rs760077
rs72704117	chr1:155205298	3.0	T	C	THBS3	Arg102Gln	1.4 × 10−14	−0.38 (−0.48, −0.28)	3.2 × 10−10	−0.38 (−0.48, −0.28)	0.0083	rs760077, rs41264915
rs564968560	chr1:155614403	2.0	A	G	MSTO1	3’UTR	1.4 × 10−11	−0.37 (−0.48, −0.27)	3.2 × 10−11	−0.37 (−0.48, −0.27)	0.0055	rs760077, rs41264915 rs72704117
rs822493	chr1:155855602	3.3	T	C	SYT11	upstream	2.5 × 10−8	0.25 (0.16, 0.33)	5.8 × 10−4	0.25 (0.16, 0.33)	0.0038	rs760077, rs41264915 rs72704117, rs564968560
Cancer antigen 125
rs141828605	chr2:241803397	0.9	T	C	GAL3ST2	Pro143Leu	9.6 × 10−29	0.91 (0.75, 1.07)	2.7 × 10−24	0.91 (0.75, 1.07)	0.015	rs62193080
rs150107870	chr2:241804117	1.9	C	T	GAL3ST2	Leu383Pro	2.2 × 10−23	0.61 (0.49, 0.73)	6.2 × 10−19	0.61 (0.49, 0.73)	0.014	rs62193080, rs141828605
rs139344622	chr2:241803427	0.9	G	A	GAL3ST2	Tyr153Cys	8.4 × 10−16	0.68 (0.51, 0.84)	2.4 × 10−11	0.68 (0.51, 0.84)	0.008	rs62193080, rs141828605 rs150107870
rs5839764	chr2:241764203	38.6	G	C	D2HGDH	intron	2.2 × 10−11	−0.12 (−0.16, −0.09)	6.1 × 10−7	−0.12 (−0.16, −0.09)	0.007	rs62193080, rs141828605 rs150107870, rs139344622
rs9927389	chr16:764058	1.5	T	C	MSLN	Ala72Val	2.6 × 10−13	−0.46 (−0.58, −0.34)	6.7 × 10−9	−0.46 (−0.58, −0.34)	0.006	rs3764246
rs150425699	chr16:768452	0.4	A	G	MSLN	Arg557His	4.3 × 10−9	−0.74 (−0.98, −0.50)	7.0 × 10−5	−0.74 (−0.98, −0.50)	0.005	rs3764246, rs9927389
Cancer antigen 19.9
rs2608894	chr19:5847989	17.1	T	C	FUT3	intron	3.7 × 10−10	−0.21 (−0.27, −0.14)	1.4 × 10−5	−0.21 (−0.27, −0.14)	0.012	rs708686
Carcinoembryonic antigen
rs10901252	chr9:133252613	6.5	C	G	ABO	downstream	9.4 × 10−8	0.12 (0.07, 0.16)	3.0 × 10−3	0.12 (0.07, 0.16)	0.0017	rs635634
rs59654817	chr19:41709489	33.4	A	G	CEACAM5	intron	3.8 × 10−20	0.10 (0.08, 0.13)	6.3 × 10−15	0.10 (0.08, 0.13)	0.0049	rs9621
rs12985771	chr19:41725256	36.3	C	A	CEACAM5	intron	1.8 × 10−18	0.10 (0.08, 0.13)	3.0 × 10−13	0.10 (0.08, 0.13)	0.0051	rs9621, rs59654817
rs770162662	chr19:41755473	0.3	T	G	CEACAM6	upstream	4.5 × 10−10	0.63 (0.43, 0.83)	7.6 × 10−5	0.63 (0.43, 0.83)	0.0026	rs9621, rs59654817, rs12985771
rs7247317	chr19:41712945	49.0	T	G	CEACAM5	intron	6.9 × 10−9	0.10 (0.06, 0.13)	1.2 × 10−3	0.10 (0.06, 0.13)	0.0046	rs9621, rs59654817, rs12985771, rs770162662
rs757625335	chr19:48505357	0.3	A	G	LMTK3	intron	7.4 × 10−8	0.52 (0.33, 0.71)	4.2 × 10−3	0.52 (0.33, 0.71)	0.0014	rs601338
Alkaline phosphatase
rs138587317	chr1:21563248	0.12	A	G	ALPL	Glu69Lys	4.3 × 10−110	−1.12 (−1.22, −1.03)	4.2 × 10−105	−1.12 (−1.22, −1.03)	0.0043	rs149344982
rs121918007	chr1:21564139	0.06	A	!A	ALPL	Glu114Lys	2.8 × 10−51	−1.27 (−1.43, −1.10)	2.8 × 10−46	−1.27 (−1.43, −1.10)	0.0019	rs149344982, rs138587317
rs773257111	chr1:21563143	0.01	A	G	ALPL	Ala34Thr	6.9 × 10−27	−2.05 (−2.42, −1.67)	5.9 × 10−22	−2.05 (−2.42, −1.67)	0.0011	rs149344982, rs138587317, rs121918007
rs121918019	chr1:21564094	0.009	A	G	ALPL	Ala99Thr	3.6 × 10−12	−1.54 (−1.98, −1.11)	3.1 × 10−7	−1.54 (−1.98, −1.11)	0.00043	rs149344982, rs138587317, rs121918007, rs773257111
rs4654748	chr1:21459575	43.4	T	C	NBPF3	intron	1.5 × 10−118	−0.10 (−0.10, −0.09)	1.3 × 10−113	−0.10 (−0.10, −0.09)	0.0045	rs149344982, rs138587317, rs121918007, rs773257111, rs121918019
rs1780329	chr1:21576457	16.5	A	G	ALPL	intron	5.9 × 10−52	−0.08 (−0.09, −0.07)	5.0 × 10−47	−0.08 (−0.09, −0.07)	0.0019	rs149344982, rs138587317, rs121918007, rs773257111, rs121918019, rs4654748
rs11463187	chr1:21540322	21.2	TG	T	ALPL	intron	1.4 × 10−35	−0.07 (−0.08, −0.05)	1.2 × 10−30	−0.07 (−0.08, −0.05)	0.0014	rs149344982, rs138587317, rs121918007, rs773257111, rs121918019, rs4654748, rs1780329
rs115257434	chr1:21570934	2.3	A	G	ALPL	Intron	1.9 × 10−20	−0.14 (−0.16, −0.11)	1.6 × 10−15	−0.14 (−0.16, −0.11)	0.00083	rs149344982, rs138587317, rs121918007, rs773257111, rs121918019, rs4654748, rs1780329, rs11463187
rs1697405	chr1:21577713	40.5	!C	C	ALPL	3’ UTR	2.7 × 10−19	−0.042 (−0.05, −0.03)	1.6 × 10−15	−0.042 (−0.05, −0.03)	0.00085	rs149344982, rs138587317, rs121918007, rs773257111, rs121918019, rs4654748, rs1780329, rs11463187, rs115257434
rs1318236	chr1:21625531	43.9	C	T	RAP1GAP	intron	6.8 × 10−18	0.037 ( 0.03, 0.05)	2.3 × 10−14	0.037 ( 0.03, 0.05)	0.00067	rs149344982, rs138587317, rs121918007, rs773257111, rs121918019, rs4654748, rs1780329, rs11463187, rs115257434, rs1697405
rs17300770	chr6:24462792	11.6	C	G	GPLD1	Asp275Glu	3.1 × 10−43	−0.09 (−0.10, −0.08)	1.5 × 10−37	−0.09 (−0.10, −0.08)	0.0016	rs573778305
rs9467148	chr6:24435774	27.0	A	G	GPLD1	intron	1.8 × 10−29	0.05 (0.04, 0.06)	8.4 × 10−24	0.05 (0.04, 0.06)	0.0011	rs573778305, rs17300770
rs146221974	chr6:24473633	0.1	A	G	GPLD1	Ser159Leu	9.4 × 10−17	−0.51 (−0.63, −0.39)	4.4 × 10−11	−0.51 (−0.63, −0.39)	0.00058	rs573778305, rs17300770, rs9467148
rs116287860	chr6:24456679	8.5	C	A	GPLD1	intron	9.6 × 10−14	0.06 ( 0.04, 0.08)	4.5 × 10−8	0.06 ( 0.04, 0.08)	0.00049	rs573778305, rs17300770, rs9467148, rs146221974
rs183821586	chr6:24473963	0.1	G	A	GPLD1	intron	3.7 × 10−11	−0.42 (−0.54, −0.30)	1.7 × 10−5	−0.42 (−0.54, −0.30)	0.00040	rs573778305, rs17300770, rs9467148, rs146221974, rs116287860
rs6993155	chr8:125496809	4.4	G	A	—	intergenic	1.4 × 10−8	−0.06 (−0.08, −0.04)	7.9 × 10−4	−0.06 (−0.08, −0.04)	0.00028	rs28601761
rs2183745	chr9:101456893	27.8	T	A	—	intergenic	1.4 × 10−10	0.03 (0.02, 0.04)	3.8 × 10−6	0.03 (0.02, 0.04)	0.00036	rs41282145
rs56392308	chr9:133255669	6.5	C	CG	ABO	Frameshift Pro354	1.3 × 10−21	0.10 (0.08, 0.12)	4.2 × 10−18	0.10 (0.08, 0.12)	0.0011	chr9:133264504
rs527478501	chr11:62072649	0.6	G	A	—	intergenic	3.8 × 10−8	0.14 (0.09, 0.19)	1.1 × 10−3	0.14 (0.09, 0.19)	0.00026	rs174564
rs17145892	chr11:62432797	18.8	T	A	AHNAK	downstream	2.7 × 10−7	−0.03 (−0.04, −0.02)	8.0 × 10−3	−0.03 (−0.04, −0.02)	0.00022	rs174564, rs527478501
rs78689694	chr11:126364925	10.5	C	G	ST3GAL4	intron	1.9 × 10−12	−0.05 (−0.06, −0.03)	5.5 × 10−8	−0.05 (−0.06, −0.03)	0.00043	rs10893507
rs200173452	chr12:552099	0.4	T	C	B4GALNT3	Arg382Cys	3.1 × 10−8	0.18 (0.11, 0.24)	9.3 × 10−4	0.18 (0.11, 0.24)	0.00027	rs7955258
rs77303550	chr16:72045758	18.2	T	C	—	intergenic	2.4 × 10−7	0.03 (0.02, 0.04)	6.9 × 10−3	0.03 (0.02, 0.04)	0.00023	rs71391445
NA	chr17:7156651	17.5	!AAG AGAA AGAG	AAGA GAAA GAG	—	intergenic	4.0 × 10−8	0.03 (0.02, 0.04)	1.5 × 10−3	0.03 (0.02, 0.04)	0.00026	rs186021206, rs55714927
rs55714927	chr17:7176997	21.0	T	C	ASGR1	synonymous	2.6 × 10−48	0.07 (0.06, 0.09)	9.5 × 10−44	0.07 (0.06, 0.09)	0.0019	rs186021206

^aThe position is given in hg38

^bThe corrected P-value is the P-value after Bonferroni correction for the number of variants in the locus.

^cEffect of the minor allele (Amin) is reported as standard deviations of rank-based inverse normally transformed data

Comparison of biomarkers across diseases

Many of the subjects have undergone several biomarker measurements, prior to, during and/or after being diagnosed with cancer, with most measurements being from a few months before official diagnosis date or later. Other subjects have undergone measurements without being diagnosed with cancer. The latter subjects show a trend for being younger than the cancer patients. Supplementary figure 2 shows the age distribution of each subject group for each biomarker. Some subjects have had more than one type of cancer. In this case, the first diagnosis was used.

We used a two-sided Wilcoxon rank-sum test to compare the highest values of biomarkers recorded for individuals with specific diseases to the highest values of subjects without cancer diagnosis (or diagnosis of any of the other diseases, Figure 2). We considered significant P-values that passed Bonferroni correction for 144 tests (6 biomarkers x 24 disease conditions).

Figure 2:

Biomarker elevation across cancer types.

Figure 2 shows the median of the largest value ever recorded for each individual across cancers and diseases for a) Alpha fetoprotein; b) Cancer antigen 15.3; c) Cancer antigen 125; d) Cancer antigen 19.9; e) Carcinoembryonic antigen; f) Alkaline phosphatase. In each plot, International Statistical Classification of Diseases and Related Health Problems (ICD) 10 codes categorize cancers. The ICD-10 code C.80 represents cancers of unknown primary site. These cancers have high metastatic potential as they have spread before the primary tumor has grown large enough to be detected. Red line indicates the median for individuals not diagnosed with cancer (or any of the other diseases listed) at the end of 2015. Asterisk indicates diseases that differ significantly from this group by a two-sided Wilcoxon rank-sum test after Bonferroni correcting for 144 (6*24) tests. IBD: Inflammatory bowel disease.

The cohorts for each biomarker moderately overlap. Supplementary figure 3A shows the overlap among subjects measured for AFP, CA-15.3, CA-19.9, CA-125 and CEA. Almost all subjects also had ALP measured. Supplementary figure 3B shows the overlap among extreme outliers, defined as subjects having biomarker values >1000 for AFP, CA-15.3, CA-125 and CEA and >5000 for CA-19.9.

Results

Sequence variants associated with biomarker levels

We performed a GWAS on biomarker levels with subjects ranging from 7,107 to 162,774 (Table 1) and identified 84 associations between sequence variants and biomarker levels. We found 3 sequence variants associated with AFP levels, 6 with CA-15.3, 9 with CA-125, 4 with CA-19.9, 11 with CEA and 51 with ALP (Figure 1, Tables 2 and 3). To assess if any of the identified variants associates with risk of developing cancer, we tested the association between all the variants and the diagnosis of 20 cancer types in the Icelandic material. With the exception of the association of rs760077 with gastric cancer described below, none of the variants associates with the risk of developing cancer in our cohorts (P > 1 × 10⁻⁶, Supplementary figure 6). We also assessed the effect of the variants on gene expression in whole blood in a sample of 2,528 Icelanders who were whole-genome RNA sequenced (Table 3). Lastly, we performed a co-localization analysis to link the identified variants with a range of biochemical traits for which summary statistics are available at deCODE Genetics (Supplementary Table 2). We summarize the results for each biomarker below.

Figure 1:

A Manhattan plot

Figure 1 contains a Manhattan plot for A) Alpha fetoprotein; B) Cancer antigen 15.3; C) Cancer antigen 125; D) Cancer antigen 19.9; E) Carcinoembryonic antigen; F) Alkaline phosphatase.

Alpha-fetoprotein (AFP)

We found three sequence variants associated with AFP levels that collectively explain 2.3% of its variance. Two are independent missense variants in the SERPINA1 gene on chromosome 14 (rs28929474-T (Glu366Lys, allele Z) and rs17580- A (Glu288Val, allele S)) and one is a common intergenic variant at the same locus. SERPINA1 encodes the serine protease inhibitor α-1-antitrypsin and we found the minor alleles of both missense variants to be associated with decreased quantity of α-1-antitrypsin in blood in 6,452 Icelandic subjects for which this measurement was available (β = −1.71 SD, P = 1.0 × 10⁻⁸³ and β = -0.61 SD, P = 2.0 × 10⁻³⁸ for rs28929474-T and rs17580-A, respectively). Reduced levels of this protein are known to cause emphysema, cirrhosis and hepatocellular carcinoma [17]. The same alleles that reduce SERPINA1 also associate with reduced levels of AFP. Additionally, rs28929474-T associates with higher levels of three liver enzymes [18] (Supplementary Table 2).

It is of note that we find no association of variants within the AFP gene or its promoter with serum levels of AFP. We replicate neither of the two variants reported to associate with AFP levels in the Chinese population [19] (P = 0.84 for rs12506899 and P = 0.01 with effect in the opposite direction for rs2251844).

Cancer Antigen 15.3 (CA-15.3)

We found six sequence variants associated with CA-15.3 levels that collectively explain 42.3% of its variance. Five of these variants are within 1 Mb from the MUC1 gene on chromosome 1, which encodes CA-15.3, and one lies within an intron of the ABO gene on chromosome 9. A single missense variant in MTX1, rs760077-A, explains 33.5% of the trait variance. We have previously shown that rs760077-A is associated with higher number of tandem repeats within exon 2 of MUC1, and protection against gastric cancer [20]. It is likely that the increased number of epitopes on Muc1 leads to higher apparent CA-15.3 levels in the carriers. A second variant at the locus, rs41264915-G, is associated with decreased levels of MUC1 mRNA in blood (β = −0.33 SD, P = 1.5 × 10⁻¹⁶), a surprising finding given that the same allele is associated with elevated CA-15.3 levels.

Cancer Antigen 125 (CA-125)

We found nine sequence variants associated with CA-125 levels that collectively explain 10.5% of its variance. One variant is within an intron of the MUC16 gene on chromosome 19, which encodes CA-125. One of the variants lies upstream of the MSLN gene on chromosome 16, which encodes mesothelin, a cell surface molecule known to bind CA-125 on the mesothelial lining [21]. We further found two low frequency missense variants rs9927389 and rs150425699 (Ala72Val and Arg557His, respectively) in MSLN, suggesting that the observed effect on CA-125 levels is mediated through structural changes in mesothelin.

Five variants are in- or close to- GAL3ST2 on chromosome 2. Of these, three are low frequency (MAF<2%) missense variants (Pro143Leu, Leu383Pro and Tyr153Cys), implicating the encoded galactose-3-O-sulfotransferase in regulation of CA-125 levels. It is presently not clear how this enzyme affects CA-125 levels. We note that the binding site of the antibody used for CA-125 detection has not been reported by the antibody’s producer and speculate that variants in GAL3ST2 may influence the glycosylation of Mucin16 and thus affect the outcome of CA-125 measurements but not necessarily the quantity of the protein in blood. Notably, these variants explain a larger fraction of the variance than does the cis-variant in MUC16 (Table 2).

Cancer Antigen 19.9 (CA-19.9)

We found four sequence variants associated with CA-19.9 levels that collectively explain 27.4% of its variance. Some are moderately correlated with variants found to associate with CA-19.9 levels in the Chinese population [19]. CA-19.9 levels are defined by antibody binding to the cell surface carbohydrate Sialyl-Lewis A and all the associated variants implicate genes known to function in the production and secretion of this molecule. We found a stop-gained variant, rs601338, in FUT2 (secretor gene of the Lewis antigen pathway), a variant upstream of FUT6, rs708686, and a variant in an intron of the FUT3 (Lewis gene), rs2608894, to associate with CA-19.9 levels. Additionally, a variant upstream of B3GNT3, rs34262244-A, which greatly affects the expression of the gene (β = −1.04, P = 1.3 × 10⁻²⁹⁶), associate with CA-19.9. B3GNT3 encodes UDP-GlcNAc-β-Gal β-1,3-N-acetlyglucosaminyltransferase 3, which plays a role in the formation of the backbone of sulfo-sialyl Lewis X tetrasaccaride structures [22].

The signals characterized by rs601338 (stop-gained in FUT2) and rs708686 (upstream of FUT6) co-localize with biochemical traits in blood (Supplementary Table 2). Among these are associations of rs708686-T with levels of Galactoside 3(4)-L-fucosyltransferase (β = −0.84 SD, P = 2.5 × 10⁻²²), an effect consistent with reduced levels of CA19.9), and replicated associations of rs601338 with lipase levels previously reported [23].

Carcinoembryonic antigen (CEA)

Eleven sequence variants were associated with CEA levels in our study, explaining 11.9% of its variance. Again, some show moderate correlations with variants associated with CEA levels in the Chinese population [19]. Five of the variants are within 250kb from CEACAM5 gene on chromosome 9, which encodes CEA. The variants rs601338 and rs708686 in FUT2 and upstream of FUT6, respectively, that associate with CA-19.9 levels also associate with CEA but the effect of rs708686 is in the opposite direction to that on CA-19.9.

We also observed an association with an intergenic variant, rs7041150, and two independent variants upstream and downstream of ABO. One of these, rs10901252-C, associates with increased expression of ABO in whole blood (β = 1.61 SD, P = 1.8 × 10⁻²¹⁰). The variants in ABO co-localize with several blood-related traits such as hematocrit, hemoglobin and granulocyte levels [24] and with cholesterol levels [25] (Supplementary Table 2).

Alkaline phosphatase (ALP)

We identified 51 variants associated with ALP levels, collectively explaining 6.9% of its variance. We confirm 13 of 14 associations reported in a recent study of ALP in European populations [26], which partially overlaps with our material. Among these are several associations with the ALPL gene on chromosome 1, driven by familial clusters carrying very rare coding variants (Table 2). The remaining variants are scattered throughout the genome, with variants in - or near genes involved in glycoprotein biology (ABO, FUT2, ST3GAL4, GPLD1, ASGR1) and carbohydrate and lipid metabolism (JMJD1C, TREH, HNF1A, B4GALNT3, PCK1, DHRS9, GCKR) being prominent. Many of these variants also co-localize with other biochemical traits, such as triglyceride concentrations [25] (Supplementary Table 2).

Out of the 51 variants, 19 are coding and 3 associate with changes in gene expression (Table 4). We found associations between rs8736-T and decreased expression of TMC4 (β = −0.97 SD, P < 10⁻³⁰⁰), between rs4654748-T and decreased expression of NBPF3 (β = −0.94 SD, P = 8.6 × 10⁻³¹¹) and between rs174564-G and increased expression of FADS2 in whole blood (β = 0.23 SD, P = 7.9 × 10⁻²³).

Table 4:

Variants associated with biomarker levels and with gene expression in whole blood. EA: Effect allele; ALP: Alkaline phosphatase; CA: Cancer Antigen; CEA: Carcinoembryonic antigen.

rsID	Chromosome:positiona	EAb	Biomarker(s)	Gene	β [SD]	P	Covariatec
rs760077	chr1:155208991	A	CA-15.3	THBS3	0.64	3.4 × 10−114	—
rs41264915	chr1:155197995	G	CA-15.3	MUC1	−0.33	1.5 × 10−16	—
rs34262244	chr19:17795098	A	CA-19.9	B3GNT3	−1.04	1.3 × 10−296	—
rs10901252	chr9:133252613	C	CEA	ABO	1.61	1.8 × 10−210	—
rs4654748	chr1:21459575	C	ALP	NBPF3	−0.94	8.6 × 10−311	—
rs174564	chr11:61820833	G	ALP	FADS2	0.23	7.9 × 10−23	rs7943728
rs8736	chr19:54173495	T	ALP	TMC4	−0.97	1.0 × 10−300	—

^aThe position is given for hg38.

^bThe effect allele is the minor allele.

^cWe condition the expression on the covariate when there is more than one variant in the region affecting the expression of the gene and the variant affecting biomarker value is not the most significant variant.

Diseases affecting biomarker levels

Given the lack of specificity of the cancer biomarkers, it is of interest to explore which conditions are likely to affect the levels of a particular marker. To search for conditions associated with elevated levels of the selected biomarkers, we compared the highest values for all individuals in a patient group with subjects without cancer diagnosis and identified several diseases where patients showed elevations in biomarker levels (Figure 2, Supplementary Table 1). We also compared the first measurements done on each individual (Supplementary Figure 4) as well as the proportion of the highest values falling into bins (Supplementary Figure 5) defined using the reference values used at Landspitali, the National University Hospital in Iceland (Table 1).

In general, the biomarker levels agree with the indicated clinical use of the respective biomarker but there are some interesting departures from the expected. High AFP levels are overwhelmingly found in HCC with increased levels also seen in TGCT and cirrhosis, a risk factor for HCC. The largest CA-125 levels are strongly associated with ovarian/peritoneal and pancreatic cancers as expected, but a highly significant association is also found between CA-125 levels and cancers of unknown primary site and a more moderate increase in a number of other malignancies. In addition to pancreatic cancer, CA-19.9 levels are also high in cholangiocarcinoma (CCA), but the increase in other cancers is much less. CA-15.3 shows a moderate increase in several cancer types in addition to breast cancer, which does not stand out with respect to this biomarker. The levels of CEA and ALP are highest in cancers of the gastrointestinal tract, but both markers show increased levels in many cancer types and ALP levels are increased in all the non-cancer phenotypes tested as well.

Discussion

Tumor antigens are often measured in individuals with unknown ailments in search of diagnostic clues, as well as being used for monitoring the progression of tumors and response to treatment. In our study, we sought to identify variants that influence tumor biomarker values regardless of cancer diagnosis. We did not remove cancer patients from the cohort before association testing. Their inclusion adds variance to the dataset and results in more conservative estimates of the variance explained by each variant. We repeated the GWAS of the cancer antigens, excluding breast cancer, peritoneal and ovarian cancer, pancreatic cancer and pancreatic, colorectal and gastric cancer patients for CA-15.3, CA-125, CA-19.9 and CEA, respectively. We observed all the same primary associations but generally larger effect estimates and higher P-values, as a result of reduced sample size. This, together with the observation that only rs760077 is associated with cancer status in the Icelandic cohorts (Supplementary Figure 6) suggests that these variants generally do not act through effect on cancer risk or aggressiveness.

We report 84 associations between sequence variants and tumor biomarker levels in the Icelandic population. While we are unaware of a study having been published on CA-15.3 and CA-125, variants affecting the other biomarkers have been reported [19, 27]. We confirmed most of those associations and discover many additional variants, some of which are rare variants with large effects. In an attempt to shed light on their biology, we tested all the variants for association with expression of genes in their vicinity and for co-localizations with biochemical traits in blood. Blood is not the most relevant tissue for any of the biomarkers and we would not be able to detect tissue or cell type specific effects. Some of the variants may also assert their influence on genes farther away than the 250 kb cut-off.

Our comparisons across cancer types show that tumor antigens are strongly associated with diagnosis of several cancers while also highlighting the lack of specificity of these tests. As demonstrated in Figure 2, many cancers other than those for which the biomarkers are most commonly used showed significant elevation of biomarker values.

A limitation and a potential source of bias in our study is that subjects were not randomly selected for biomarker assaying but rather measurements were done because of suspicion of a particular ailment, to monitor the progression of an existing disease or for some other clinical reason. In particular, the subject labeled as ‘Without cancer diagnosis’ in Figure 2 are individuals seeking medical assistance and biomarker levels may not reflect those found in healthy individuals.

The utility of genetic correction in prediction models depends on the fraction of the trait’s variance explained by sequence variants. The relatively high variance explained by identified genetic factors, for some of the cancer antigens in particular, suggests that improvements could be achieved by the inclusion of these variants in such models. A single variant, rs760077, explains >33% of the variance in CA-15.3. This biomarker is generally considered of low specificity but this high fraction of variance explained highlights the potential for improvement by correcting values for genotype. However, CA-125 is perhaps of most interest in this regard, as effective screening tools for ovarian cancer are currently lacking. Recent screening trials of ovarian cancer reported no mortality benefit from screening with CA-125 and trans-vaginal ultrasound [9, 28]. A retrospective study of these cohorts based on genotypically corrected CA-125 levels may show greater benefit of CA-125 screening. Furthermore, corrected biomarker values should be useful when combining liquid biopsies and more traditional biomarkers to detect early tumors [2].