Cancer Diagnosis & Prognosis

Abstract

Background/Aim: Several studies have reported on the relationship between HOXB13 variants and an increased prostate cancer (PC) risk. To our knowledge there are not many studies on HOXB13 mutations in Japanese patients with prostate cancer, and there many issues remain uninvestigated. We herein clarified the association between HOXB13 genetic variants and PC risk in a Japanese population. Patients and Methods: PC patients were diagnosed at the Gunma University Hospital and affiliated hospitals from 1994 to 2016. Sanger sequencing was performed on the coding regions of the HOXB13 gene in 152 familial PC (FPC) patients. Genotyping was performed on single nucleotide variants (SNVs) found in Sanger sequencing in 230 FPC patients from 152 pedigrees and 197 sporadic PC (SPC) patients and 144 controls. Allelic frequency and clinical data for each variant were studied in cases and controls. Results: G132E and F127C were identified in FPC patients. The frequencies of G132E and F127C were significantly higher compared to the control group (p=0.039). In three families, seven PC patients shared the G132E variant, within second-to-third-degree relatives. It was not possible to clarify to pathogenicity of each SNV alone. Conclusion: We found two significant variants of the HOXB13 gene, G132E, F127C by analyzing and comparing gene samples from PC and non-PC patients. Furthermore, the HOXB13 G132E variant was found significantly increased in the FPC group.

Keywords: prostate cancer, familial prostate cancer, HOXB13

In 2020, 191,930 new cases of prostate cancer (PC) and 33,330 deaths were anticipated in the US, where PC is the most frequent non-dermatological cancer in men. PC is the second-leading cause of cancer-related death in men in the US, exceeded only by lung cancer (1). The incidence of PC varies greatly among populations: the PC-specific death rate of African-American men is twice that of non-Hispanic white men in the US. Asian men have a very low incidence of PC, although both the incidence and mortality have gradually increased, attributable to both environmental factors (particularly a Western diet) and genetic factors (2,3). In Japan, the incidence of PC has recently increased; according to the National Cancer Center (2017), PC is the most common cancer in Japan and the sixth-leading cause of cancer death in Japanese men (4).

PC exhibits familial clustering. The first-degree relatives of an affected individual are two-to-three-fold more likely to develop PC than others. Independent segregation analyses revealed that both an early age of PC onset and the existence of multiple affected family members were strong predictions of risk (5-9). Of all PC cases, 5-10% are believed to be primarily caused by high-risk (inherited) genetic factors or PC susceptibility genes. The results of several large case-control and cohort studies on various populations suggest that a family history is a major PC risk factor (6,10). Hereditary PC is distinguished by early age at onset and autosomal-dominant inheritance (11).

Genome-wide association studies have identified more than 170 common genetic variants [including single nucleotide variants (SNVs)] associated with PC (12). Several rare variants increase the risk of morbidity. A study conducted in the US found that HOXB13 G84E was pathogenic (13). A study conducted in China found that G84E was absent, but G135E was present (14). In a study conducted in Japan, Hayano et al. identified BRCA2, HOXB13, and TRRAP genes as susceptibility genes (15).

We collected clinical and genomic data on PC patients living in the suburbs of the Gunma Prefecture, Japan. We performed targeted HOXB13 sequencing and genotyping and explored the clinical characteristics of PC patients.

Patients and Methods

Study subjects. We defined PC patients with a family history of PC as familial PC (FPC) patients and those without such history as sporadic PC (SPC) patients. We included 571 men: 427 PC patients (230 FPC patients from152 pedigrees and 197 SPC patients) and 144 controls. All PC patients were diagnosed at the Gunma University Hospital and affiliated hospitals between 1994-2016. We screened HOXB13 sequence variants in germline DNA samples of one affected proband from each of 152 pedigrees affected by FPC. Controls were recruited from the outpatient clinics of Gunma University Hospital; the most common diagnosis was benign prostate hypertrophy (BPH). Controls were excluded if they evidenced an abnormal prostate-specific antigen level (>4.0 ng/ml), an abnormal digital rectal examination, or a previous PC diagnosis.

Study design. In the first experiment,152 FPC probands underwent Sanger sequencing of the HOXB13 coding regions. We then genotyped about the SNVs found in the first sequencing in 230 FPC patients, 197 SPC patients, and 144 controls. The frequencies and clinical features of all variants were studied in 152 FPC probands, 197 SPC patients, and 144 controls.

Sequence analysis and genotyping. Genomic DNA was isolated from whole blood cells using a GENOMIX kit (Talent srl, Trieste, Italy). Sanger sequencing of HOXB13 coding regions employed gene-specific primers. PCR was performed in a reaction volume of 10 μl containing 10 ng genomic DNA, 2.0 pmol of each primer, 10×EX Taq buffer, 2.5 mmol/l deoxynucleoside triphosphates, and 0.25 U TaKaRa Ex Taq DNA polymerase (Hot Start Version) (Takara Bio Inc., Shiga, Japan). PCR amplification employed a GeneAmp PCR System 9700 thermal cycler (Applied Biosystems, Foster City, CA, USA). The conditions were 94˚C for 1 min, followed by 35 cycles of 30 s at 94˚C, 30 s at 60˚C, and 30 s at 72˚C, followed by elongation at 72˚C for 5 min. The PCR products were mixed with Big Dye Terminator ver. 3.1 Matrix Standard Kit reagents (Thermo Fisher Scientific, Waltham, MA, USA) and formamide, and analyzed on an ABI PRISM 310 Genetic Analyzer (Thermo Fisher Scientific). Data were analyzed using SEQUENCHER (Gene Codes) software. SNP markers were genotyped via real-time PCR using the Taq Man SNP genotyping assay (ID C_2160574_30, Applied Biosystems). Each reaction mixture (5 μl) contained 10 ng genomic DNA, 2.5 μl Premix Ex Taq, and 0.25 μl Allelic Discrimination Assay Mix (Applied Biosystems). PCR amplification proceeded in an C1000 thermal cycler (a CFX96 real-time PCR detection system; Bio-Rad Laboratories Inc., Hercules, CA, USA). The cycling conditions were 95˚C for 30 s, followed by 40 cycles of 5 s at 95˚C and 30 s at 60˚C. Genotypes were automatically called and manually verified.

Allele frequency database. Allele frequency data from more than 60,000 individuals stored in the genome Aggregation Database (gnomAD) served as references (https://gnomad.broadinstitute.org). To gather Japanese data, we accessed the database of the Japanese Multi Omix Reference Panel (jMORP) (16) and the Human Genetic Variation Database (HGVD, ver. 2.3) (17). jMORP was constructed using whole-genome sequencing data from more than 8000 healthy individuals. HGVD collected exome variant data on more than 1000 Japanese individuals.

Statistical analysis. Case-control association analysis was performed using the Fisher’s exact test. The two-sided Fisher’s exact test was used for all comparisons except age at diagnosis. Welch’s t-test was used to compare ages at diagnosis. A p-value <0.05 was considered significant. To eliminate selection bias in the statistical analysis of genotyping, excluding duplication within the same family, 152 probands were used in genotyping analysis. However, clinical data was evaluated for all PC patients. All analyses were performed with the aid of SPSS ver. 25 (IBM, Chicago, IL, USA).

Ethics statement. All patients provided written informed consent. The study protocol was approved by the Institutional Review Board of Gunma University Hospital (approval no. 5.2013.12.2, HS2017-263) and adhered to the standards of the International Conference on Harmonization/Good Clinical Practice and the 1964 Helsinki declaration and later amendments, or comparable ethical standards.

Results

The mean age at PC diagnosis was 69.1 years (standard deviation 7.4 years) in all PC patients. Table I shows the clinical characteristics of all PC patients and controls in this study. We sequenced the entire HOXB13 coding regions of 152 FPC patients; analyzed the data using SEQUENCHER; and detected five SNVs, of which two (rs8556 and rs9900627) were synonymous substitutions (no amino acid change) and three non-synonymous. G132E refers to a change in the second position of codon 132 (GGA to GAA; glycine to glutamic acid; p. Gly132Glu). F127C refers to a change in the second position of codon 127 (TTT to TGT; phenylalanine to cysteine; p. Phe127Cys). A212T(rs145059285) refers to a change in the first position of codon 212 (GCC to ACC; alanine to threonine; p. Ala212Thr). We genotyped the three non-synonymous variants (G132E, F127C, and A212T) in 427 PC patients and 144 controls. We found that 20 PC patients and 1 control carried variants: G132E in 11 FPC patients; F127C in 4 FPC patients; and A212T in 3 FPC patients, 2 SPC patients, and 1 control. We compared the minor allelic frequencies of all five SNVs in 349 unrelated PC patients (152 FPC probands and 197 SPC patients) and 144 controls using the available databases (Figure 1). Non-synonymous variants (G132E, F127C, and A212T) were rare. G132E was found only in East Asia and was more common in our FPC patients than in the general populations. We identified only one homozygous G132E PC patient; homozygosity has been not reported in the public databases. F127C was not found in any database, including dbSNP of the National Center for Biotechnology Information. The A212T frequency did not differ between our PC patients and general populations. We drew family trees to place the various carriers (Figure 2). In three families, seven PC patients shared G132E with second-to-third-degree relatives. In pedigrees 2 and 9, this was the dominant mode of inheritance. In pedigree 66, three PC patients with F127C were observed in one family. In pedigrees 52 and 59, both families had two FPC patients, but they did not share A212T. Table II lists the genotypes of all variants along with significant differences between cases (152 FPC probands and 197 SPC patients) and 144 controls with and without variants. Patients were significantly more likely to host either of the two HOXB13 variants than controls (p=0.039). FPC patients were significantly more likely to carry G132E than SPC patients (p=0.001) (Table III). Table IV lists the clinical characteristics of PC patients with and without the two variants. The age of onset was significantly lower among G132E carriers. None of SPC patients carried G132E and F127C variant. No other clinical findings differed significantly between the patient groups.

Discussion

The homeobox genes encode a highly conserved family of transcription factors that play essential role in embryonic development and tissue homeostasis. Humans have four HOX gene clusters on different chromosomes (18). HOX genes also play increasingly important roles in terms of genetic predispositions toward many types of cancer. A large-scale twin study revealed that hereditary factors contributed significantly to colorectal, breast, and PC development (19). HOXB13 (located on chromosome 17q21.2) encodes a homeobox B transcription factor playing important roles in prostate development and tumorigenesis. Many studies have reported significant associations between HOXB13 and various cancers (20-24). Remarkably, inherited HOXB13 variants have not been widely assessed in terms of PC risk.

We focused on HOXB13 variants in PC patients living in the suburbs of the Gunma Prefecture, Japan. We identified five SNVs and three missense variants (G132E, F127C, and A212T). Ewing et al. (13) reported that HOXB13 G84E was associated with a significantly increased risk of hereditary PC patients. The Genome Aggregation Database (25) indicates that G84E has been found in African, Ashkenazi Jewish, and Latino populations; but G132E and G135E only in East Asians. However, a similar analysis of a Chinese cohort reported G135E but not G132E (14).

Momozawa et al. sequenced eight genes associated with hereditary PC in 7636 unselected Japanese patients with PC and 12,366 male controls; they categorized the variants according to the ACMG/AMP guidelines, G132E was classified as a pathogenic variant in PC patients (26). In that study, the proportion of G132E carriers was 0.73% in the prostate cancer group and 0.1% in the control group. In our study, the proportion of G132E carriers was 2.6% in the PC group and no carriers in control. Higher results of in our study may be biased by the limited number of cases in the suburbs of Gunma. However, as examined by Momozawa et al., it seems that the proportion of G132E rate was significantly higher in the PC group compared to the control group.

Hayano et al. (15) performed exome sequencing (followed by Sanger sequencing) when seeking putative, causative germline variants in 140 Japanese PC patients from 66 families. They found BRCA2, HOXB13, or TRRAP variants in seven large PC families (three or four patients per family; one variant per family). They also identified two or more variants per family in another 59 small PC families (two patients per family). They concluded that HOXB13 variants F127C and G132E were deleterious and were significantly more common in PC patients than controls. However, one limitation of that study was the absence of healthy controls.

We compared PC samples from both FPC and SPC patients, as well as BPH samples from healthy controls. In addition to F127C and G132E, we found an A212T variant, which was less common than the other variants and thus may not be associated with PC. Generally, hereditary tumors develop at a young age and are highly malignant. The age of onset was significantly lower among G132E carriers, but no other clinical findings differed significantly between the patient groups.

Our study had several limitations. First, we included a relatively small number of participants whose clinical characteristics varied. Our results differ from those of Momozawa et al., who enrolled many more patients. We considered the results published by Momozawa et al. are very useful because they were obtained from many patient samples. However, our study was also meaningful because it was a conclusion drawn from the detailed follow-up data. Second, our work was retrospective in nature, and the FPC and SPC groups differed significantly in terms of prognostic indicators such as the Gleason score. Third, the analysis restricted to a limited regional area in Japan with concerns as to generalizability of the findings, and the overall limited implications for the clinical practice represent limitations to this contribution. There are some reports comparing prostate cancer tissue removed after diagnosis with non-cancer tissue by immunohistological examination. Jae HC et al. reported that the expression level of sirtuins (SIRT) 2 in the SIRT family affects the incidence of cancer (27). In this study, we could not investigate the pathological examination using the excised specimen, the immunohistological examination, and the association with the presence or absence of the mutation of HOXB13. I would like to make it an issue for future study.

In conclusion, we found two significant variants of the HOXB13, G132E and F127C, by comparing PC and non-PC patients as in the previous report. G132E is a very rare variant found only in East Asia. It might be a unique Japanese variant that is also different from the G135E seen in the Chinese. Compared to a healthy Japanese population, G132E was significantly more common in the FPC group. It may be associated with Japanese PC. G132E and F127C were shared in a hereditary PC pedigree. Two of them met Carter’s definition of hereditary prostate cancer (3 PC patients in a nuclear family). Only one family shared one variant, it cannot be concluded that it is a deleterious variant of hereditary PC in this study. Although it is a small study, it is valuable accumulation of prostate cancer families in the Gunma suburbs. If follow-up surveys are possible, we would like to plan further evaluation of the family line.

Conflicts of Interest

Kazuhiro Suzuki has potential financial conflicts of interest as below. Consultancies: Takeda Pharmaceutical, Astellas Pharma, Daiichi-Sankyo, AstraZeneca, Sanofi, Janssen, Bayer; Grants received: Takeda Pharmaceutical, Astellas Pharma, Daiichi-Sankyo, Ono Pharmaceutical. All other Authors have declared that no conflicts of interest exist.

Authors’ Contributions

Concept and design: SK, HM, KS, YM; Acquisition of data: SK, HM, NO, MA, YM; Analysis of interpretation of data: YS, SA, SK, HM, HK, YM; Drafting of the manuscripts: SK, YM; Critical revision of the manuscript for important intellectual content: KS, YM; Statistical analysis: KS, YM; Obtaining funding: none; Administrative, technical, or material support: KS; Supervision: KS.

Acknowledgements

The Authors are grateful to Atsuko Ohyama, Hayumi Ohyama and Rie Suzuki (Gunma University Graduate School of Medicine, Department of Urology) for their technical supports and collecting samples.

Funding

This work was supported by Grants-in-Aid for Scientific Research from Japan Society for the Promotion of Science (JSPS) (JSPS KAKENHI Grant Number 21K09366).

Variants in HOXB13, G132E and F127C, Are Associated With Prostate Cancer Risk in Japanese Men

Abstract

Patients and Methods

Results

Discussion

Conflicts of Interest

Authors’ Contributions

Acknowledgements

Funding

References

Tools

Similar articles

Table of contents