The HABP2 G534E Variant Is an Unlikely Cause of Familial Nonmedullary Thyroid Cancer

Abstract

Context:

A recent study reported the nonsynonymous G534E (rs7080536, allele A) variant in the HABP2 gene as causal in familial nonmedullary thyroid cancer (NMTC).

Objective:

The objective of this study was to evaluate the causality of HABP2 G534E in the TCUKIN study, a multicenter population-based study of NMTC cases from the British Isles.

Design and Setting:

A case-control analysis of rs7080536 genotypes was performed using 2105 TCUKIN cases and 5172 UK controls.

Participants:

Cases comprised 2105 NMTC cases. Patient subgroups with papillary (n = 1056), follicular (n = 691), and Hürthle cell (n = 86) thyroid cancer cases were studied separately. Controls comprised 5172 individuals from the 1958 Birth Cohort and the National Blood Donor Service study. The controls had previously been genotyped using genome-wide single nucleotide polymorphism arrays by the Wellcome Trust Case Control Consortium study.

Outcome Measures:

Association between HABP2 G534E (rs7080536A) and NMTC risk was evaluated using logistic regression.

Results:

The frequency of the HABP2 G534E was 4.2% in cases and 4.6% in controls. We did not detect an association between this variant and NMTC risk (odds ratio [OR] = 0.896; 95% confidence interval, 0.746–1.071; P = .233). We also failed to detect an association between the HABP2 G534E and cases with papillary (1056 cases; G534E frequency = 3.5%; OR = 0.74; P = .017), follicular (691 cases; G534E frequency = 4.7%; OR = 1.00; P = 1.000), or Hürthle cell (86 cases; G534E frequency = 6.3%; OR = 1.40; P = .279) histology.

Conclusions:

We found that HABP2 G534E is a low-to-moderate frequency variant in the British Isles and failed to detect an association with NMTC risk, independent of histological type. Hence, our study does not implicate HABP2 G534E or a correlated polymorphism in familial NMTC, and additional data are required before using this variant in NMTC risk assessment.

Nonmedullary thyroid cancer (NMTC) is the most common endocrine malignancy. Because incidence rates are growing at an annual rate of 5% in the United States (1), this malignancy will soon become the third most commonly diagnosed cancer among American women, and it is now the second most common cancer among U.S. Hispanic women (1–3). NMTC is also one of the few cancer types for which a genetic risk is higher than the risk conferred by lifestyle and environmental exposures (4). Genome-wide association and candidate studies have identified low penetrance NMTC variants on chromosomes 9q22, 14q13, 2q35, 8p12, 8q24, and 14q13 (5–8), whereas linkage studies of familial NMTC have suggested that genomic regions on 1q21, 2q21, and 19p13 harbor highly penetrant variants (9–11), although no causal mutations have been identified conclusively in the latter regions.

In a recent study, Gara et al (12) identified a missense variant (G534E or rs7080536 allele A, rs7080536A) in the HABP2 gene, on chromosome 10q25, that showed complete cosegregation in a kindred of seven affected NMTC cases from unknown ethnicity, including six with a papillary histology and one with a follicular adenoma. After several functional experiments, which included enhanced colony formation and increased cell migration and reporting a significant frequency difference, at P < .001 between the data from the thyroid cancer The Cancer Genome Atlas study and a “multiethnic” database, the authors concluded that HABP2 G534E was causal of familial NMTC. The HABP2 gene was identified as a plasma-hyaluronan-binding protein with serine protease activity that plays a role in coagulation and fibrinolysis (13–15). The HABP2 G534E variant is also known as Marburgh I polymorphism, which has been shown to reduce HABP2 activity (16) and has been implicated as a risk factor in cardiovascular diseases (17), in the progression of carotid stenosis (18) and in venous thromboembolism in some but not all studies (19, 20). Gara et al (12) thus suggested a new role for HABP2 as a familial NMTC gene, a finding that could be of potentially great importance for risk assessment and for personalized prevention and treatment of this increasingly common malignancy. Intrigued by the potential importance of HABP2 G534E for NMTC risk, we decided to investigate this variant in a large multicenter population-based study of NMTC in the British Isles (21, 22).

Subjects and Methods

Study samples

A total of 2105 NMTC cases were recruited through a multicenter Thyroid Cancer genetics UK and Ireland (TCUKIN) study (21, 22). All cases had histologically confirmed NMTC and were of northern European ancestry. After completion of a brief questionnaire, cases donated approximately 10 mL of blood for DNA isolation. The Southampton and South West Hampshire Research Ethics Committee approved the TCUKIN research protocol. For the present study, we also used previously published genotype data from 5172 UK population controls, which included 2673 participants in the 1958 Birth Cohort (58C) and 2499 donors of the National Blood Donor Service (NBS) (21).

Genotyping

All cases were genotyped for HABP2 G534E/rs7080536A using KASP chemistry (LGC Genomics) following the manufacturer's protocol (genotyping probes are shown in Supplemental Table 1). Two of the HABP2 G534E homozygous cases and two of the heterozygous cases detected by KASP genotyping were verified by Sanger sequencing (Supplemental Figure 1). Call rates for genotyping were >99%; concordance between KASP genotyping and Sanger sequencing was 100%, and the visual inspection of the genotype clusters did not reveal obvious technical issues. HABP2 G534E/rs7080536A was in Hardy-Weinberg equilibrium in both cases and controls (data not shown).

Statistical analysis

Association testing

Logistic regression methods implemented in PLINK (23) and R (24) were used to obtain association statistics (odd ratios [ORs] and two-sided P values) as reported previously (21, 22). We also carried out case-control and case-only analyses between HABP2 G534E and clinical characteristics including age of onset and histological subtype (available in 1833 cases). These analyses were carried out using the χ² test and R.

Haplotype analysis

To investigate the origin of the HABP2 G534E variant, we genotyped four HABP2 G534E homozygous and 18 heterozygous carriers at four closely linked single nucleotide polymorphisms (rs10787491, rs932650, rs10885478, and rs1885434), which covered a 6-kb region (chr10:113,584,808–113,590,838) centered around HABP2 G534E. Haplotype reconstruction was carried out using Haploview (25).

Results

Association between HABP2 G534E and NMTC risk

We genotyped 2105 cases from the British Isles for the HABP2 G534E variant and used publicly available genotype data from 5172 population-matched controls as reported before (21, 22, 26–28). Allele counts, ORs, and two-sided allelic P values are shown in Table 1. The population frequency of HABP2 G534E was found to be 4.6%, which suggests that this is a low-to-moderate frequency variant in the general population of the British Isles. As shown in Table 1, HABP2 G534E was more common in controls than in cases, and we failed to detect a significant association between this variant and NMTC risk in our study (OR = 0.896; 95% confidence interval [CI], 0.746–1.071; P = .233; Table 1), despite having considerable statistical power to detect variants associated with the expected familial effect of HABP2 G534E (data not shown).

Table 1.

Genotype Frequencies and Association Statistics Between NMTC Risk and rs7080536 (HABP2 G534E) in Our Study

Genotypes/Risk Allele	No. of Cases (Frequency)	No. of Controls (Frequency)	OR (95% CI)	P Value
GG	1924 (0.918)	4704 (0.910)	Reference
GA	167 (0.080)	456 (0.088)	0.887 (0.733–1.070)	.212
AA	4 (0.002)	12 (0.002)	0.807 (0.190–2.668)	1.000
Allele A	175 (0.042)	480 (0.046)	0.896 (0.746–1.071)	.233

Association between HABP2 G534E and histological subtype and age of diagnosis

Histologically, NMTC can be divided into main subtypes: papillary thyroid cancer, the most common subtype, which accounts for approximately 80% of all the cases; and follicular thyroid cancer, which accounts for approximately 10–20% of cases (29, 30) and also includes cases with follicular variant histology (30). Hürthle cell carcinoma is also a rare histological subtype commonly found in NMTC families (31). In our study, we evaluated the association between the HABP2 G534E and these three NMTC subtypes. As shown in Table 2, we failed to detect significant associations between HABP2 G534E and increased risk of any of the three NMTC subtypes tested in our study. Unexpectedly, we detected a protective effect of this variant in papillary thyroid cancer (P = .017; OR = 0.74; Table 2), which further disagrees with the findings of Gara et al (12). Next, we examined whether the HABP2 G534E heterozygotes or homozygotes had an earlier age of onset because the expectation for highly penetrant mutations is an anticipation of the disease. We did not detect differences in HABP2 G534E carriers and noncarriers in our study (average age for heterozygous, 48.3 y; G534E homozygous, 46.2 y; noncarriers, 46.7 y; P > .273 for all comparisons; Table 3). Therefore, the stratification of our study by histological type and age of onset also failed to detect an effect of HABP2 G534E on NMTC risk.

Table 2.

Association Statistics for NMTC Subtypes (Papillary, Follicular, and Hürthle Cell) in Our Study

Subtype	Genotypes/Risk Allele	No. of Cases (Frequency)	No. of Controls (Frequency)	OR (95% CI)	P Value
Papillary TC	GG	979 (0.931)	4704 (0.910)	Reference
	GA	71 (0.068)	456 (0.088)	0.75 (0.58–0.97)	.0278
	AA	1 (0.001)	12 (0.002)	0.40 (0.05–3.08)	.7103
	Allele A	73 (0.035)	480 (0.046)	0.74 (0.58–0.95)	.017
Follicular TC	GG	625 (0.910)	4704 (0.910)	Reference
	GA	60 (0.087)	456 (0.088)	0.99 (0.75–1.31)	1.000
	AA	2 (0.003)	12 (0.002)	1.25 (0.28–5.62)	.6756
	Allele A	64 (0.047)	480 (0.046)	1.00 (0.77–1.31)	1.000
Hürthle cell TC	GG	75 (0.872)	4704 (0.910)	Reference
	GA	11 (0.128)	456 (0.088)	1.51 (0.80–2.87)	.2017
	AA	0 (0)	12 (0.002)	NA	NA
	Allele A	11 (0.063)	480 (0.046)	1.40 (0.76–2.60)	.2794

Abbreviation: TC, thyroid cancer.

Table 3.

Average Age of Diagnosis for HABP2 G534E Carriers (Homozygous and Heterozygous) and Noncarriers in Our Study

Genotype	Total No. of Cases	Average Age at Diagnosis (SD), y	P Value
Noncarriers	1685	46.7 (15.2)	Reference
Heterozygous	138	48.3 (16.1)	.273
Homozygous	4	46.2 (18.9)	.959

Haplotype analysis

An intriguing possibility that could explain our failure to replicate the findings of Gara et al (12) is that HABP2 G534 could have multiple independent origins and that one of the HABP2 G534-bearing haplotypes could have a cryptic NMTC casual mutation in a second gene in the same region. To assess whether HABP2 G534 had single or multiple origins, we evaluated closely linked markers in the region. Table 4 shows the haplotypes reconstructed in the 22 individuals (four homozygous and 18 heterozygous) with this variant in our study. We found that all HABP2 G534E carriers shared the same 2.5-kb core haplotype defined by three markers (HABP2 G534E/rs7080536A, rs10885478G, rs1885434G; Table 4). This finding is important because it suggests that the G534E allele has a single origin since it only happens in a unique haplotype.

Table 4.

Haplotypes Surrounding the HABP2 G534E Variant in Four Homozygous and 18 Heterozygous NMTC Cases

Individual	G534E Genotype	rs10787491	rs932650	rs7080536 (G534E)	rs10885478	rs1885434
S1	Homozygous	GG	TT	AA	GG	GG
S2	Homozygous	GG	TT	AA	GG	GG
S3	Homozygous	GG	TT	AA	GG	GG
S4	Homozygous	GG	00	AA	GG	GG
S9	Heterozygous	GG	TT	GA	GG	GG
S10	Heterozygous	GG	TT	GA	GG	GG
S11	Heterozygous	GG	TT	GA	GG	GG
S15	Heterozygous	GG	TT	GA	GG	GG
S16	Heterozygous	GG	TT	GA	GG	GG
S17	Heterozygous	GG	TT	GA	GG	GG
S22	Heterozygous	GG	TT	GA	GG	GG
S24	Heterozygous	GG	TT	GA	GG	GG
S23	Heterozygous	GT	TT	GA	GG	GG
S8	Heterozygous	GT	TT	GA	GG	GG
S19	Heterozygous	GT	CC	GA	GG	AG
S20	Heterozygous	GT	CC	GA	GG	AG
S13	Heterozygous	GT	CC	GA	GG	AG
S14	Heterozygous	GT	CC	GA	GG	AG
S7	Heterozygous	GT	CC	GA	GG	AG
S5	Heterozygous	GG	CC	GA	AG	AG
S6	Heterozygous	GG	CC	GA	AG	AG
S12	Heterozygous	GT	TC	GA	GG	AG

Alleles in the shared core haplotype defined by rs7080536/HABP2 G534E, rs10885478, and rs1885434 are in boldface.

Discussion

NMTC is the most common endocrine malignancy and one of the few cancers where the variance in the risk explained by a genetic predisposition (53%) is more important than the risk explained by lifestyle, the environment, or chance (32). Although several regions harboring highly penetrant NMTC variants have been identified so far, no mutated genes have been conclusively found in such regions (5–11). Only the recent study by Gara et al (12) reported HABP2 G534E as a causal variant in an NMTC family. Because these findings can have potentially important implications in NMTC risk assessment and management, we investigated the role of this HABP2 variant in our large population-based study but failed to confirm a causal role of HABP2 G534E in NMTC.

The frequency of HABP2 G534E has been reported to range from 1.6 to 4.3% in European populations (33). Our study confirmed these previous findings because we observed a population frequency of 4.6% in the British Isles controls. Our data are also consistent with a previous British study on cardiovascular disease where HABP2 G534E population frequency was found to be 4.3% (17). We can therefore conclude that HABP2 G534E is a low-to-moderate frequency variant that segregates in populations of European ancestry. This finding may explain why Gara et al (12) reported a frequency of approximately 4% in cases from the NMTC. The Cancer Genome Atlas study, most of whom were likely to be white/European Americans (30). As a control group, Gara et al (12) used data from a “multiethnic database” where HABP2 G534E frequency was approximately 0.07%. This database was most likely the 1000 Genomes Study, which contains populations of both European and non-European ancestry and where HABP2 G534E frequencies are affected by the very different ancestries in the sample. Thus, the significant P value reported by Gara et al (6) could have been the result of population stratification rather than of an association with the disease. The population frequency analyses from our study therefore suggest that HABP2 G534E is a low-to-moderate frequency variant that is highly unlikely to be a highly penetrant, cancer-causing mutation. This conclusion is further supported by our association analyses, where we failed to detect associations between HABP2 G534E and increased NMTC risk using both case-control and cases-only analyses, showing that it is not even a low-penetrance thyroid cancer gene. Although we cannot exclude the possibility that the G534E variant reported as causal in the Gara et al (12) study resides in a very rare haplotype harboring additional cryptic mutations that cause NMTC, our haplotype analyses suggested a single origin of this variant in populations of European ancestry. Although further haplotype analyses should also be carried out in the family of Gara et al (12), our finding of a single haplotype does not support the notion of multiple independent origins of G534E.

In summary, our study suggests that the HABP2 G534E is very unlikely to cause familial NMTC. At the population level, we also found that this variant does not increase NMTC risk. We therefore suggest that careful replication studies should be performed and great caution should be taken when assessing NMTC risk among carriers of any HABP2 variant.