TBX15 rs98422, DNM3 rs1011731, RAD51B rs8017304, and rs2588809 Gene Polymorphisms and Associations With Pituitary Adenoma

Abstract

Background: Pituitary adenoma (PA) is a benign tumor of parenchymal cells in the adenohypophysis, and it’s development is strongly associated with genetic factors.This study aim was to find whether TBX15 rs98422, DNM3 rs1011731, RAD51B rs8017304, and rs2588809 single nucleotide polymorphisms can be associated with pituitary adenoma. While the TBX15 gene belongs to the T-box family of genes and is a transcription factor involved in many developmental processes, the DNM3 encodes a protein that is a member of the dynamin family with mechanochemical properties involved in actin-membrane processes, predominantly in membrane budding, and the RAD51B gene plays a significant role in homologous recombination in DNA repair for genome stability. Materials and Methods: The study enrolled 113 patients with pituitary adenoma and 283 healthy control subjects. DNA samples were extracted and purified from peripheral blood leukocytes. Genotyping was carried out using real-time polymerase chain reaction. The results were assessed using binomial logistic regression. Results: Our study revealed that RAD51B rs2588809 TT genotype could be associated with PA development in the co-dominant (OR=6.833; 95% CI=2.557-18.262; p<0.001) and recessive (OR=7.066; 95% CI=2.667-18.722; p<0.001) models. The same results were observed in females but not in males and PA without recurrence, while in PA with recurrence, no statistically significant results were obtained. Conclusion: RAD51B rs2588809 TT genotype may increase the odds of PA development in women; it may also be associated with non-recurrent PA development.

Pituitary adenoma (PA) is an intracranial tumor localized in the bone cavity (sella turcica) surrounded by multiple neural, vascular, endocrine, and bone structures, which further may contribute to an assortment of tumor types (1-7). PA accounts for approximately 15 to 20 percent of primary brain tumors with a prevalence of 77.6-97.6 PA cases per 100,000 individuals. Clinically significant PAs occur in one out of 1064 individuals (5-11). PA can occur insidiously – most patients do not realize they have it until specifically investigated. This tumor can manifest in two ways: an endocrine imbalance or pressure on the surrounding structures. The latter is the most common form of macroadenoma manifestation (12). Six to ten percent of all PAs expand into the cavernous sinus (13,14). The optic chiasm is directly above the pituitary gland, so a prolonged compression of the chiasm can cause primary optic nerve atrophy and result in visual function defects, such as decreased visual acuity and visual field defects or impaired color vision (15). The earlier the tumor is diagnosed, the more likely it is to be removed and the visual function to be preserved. Endocrine changes may be due to the overexpression of tumor hormones or hypoexpression, when the tumor compresses the pituitary gland (16).

The etiology and pathogenesis of PA are complex and still poorly understood. PA represents a heterogeneous disease whose pathogenesis is a multifactorial process that involves both environmental and genetic factors. Therefore, a better understanding of the PA pathogenesis requires a comprehensive research of this disease’s biological and genetic markers. Plenty of possible molecular markers, as well as intelrleukin 9 variant rs1859430, which might be incorporated in the tumorogenesis of PAs, are currently under investigation (17). Recent studies focus on genetic markers for cancer development, so we aimed to elucidate the role of four TXB15, DNM3 and RAD51B single nucleotide variants in PA development. The elevated genes has been reported in a variety of cancers, including prostate, ovarian cancers. However, the data regarding TBX15, DNM3 and RAD51B genes and PA is still lacking (18-20).

TBX15, a T-box family member, is possibly involved in cancer cell transformation because of its antiapoptotic function (18). It also is known that T-box genes are involved in in carcinogenesis (21-23). TBX15 is associated with prostate (24), thyroid cancer (19,25,26), ovarian carcinoma (20).

The other marker dynamin 3 (DNM3) is a candidate tumor suppressor gene. This gene encodes a member of the dynamin family, which possesses mechanochemical properties to tabulate and sever membranes (27). However, few reports describe the relationship between DNM3 and malignant diseases (28,29). DNM3 has been found mainly in the brain (at a lower level than DNM1) and testicles, and less frequently in the lungs and heart (30). The importance of the DNM3 gene has been investigated in gliomas (31,32), hepatocellular carcinoma (33-35), colon cancer (36), and papillary thyroid carcinoma (37). Few studies have investigated the association between DNM3 and hepatocellular carcinoma, breast cancer, T-cell lymphoma, colon cancer (28-30,33-34,38,39). Additionally, the importance of DNM3 was investigated in brain tumors glioblastomas (31-32).

RAD51B plays a role in homologous DNA pairing and strand exchange in DNA double-strand break repair (38,40). The importance of RAD51B has previously been investigated in the breast, ovarian, lung cancer and uterine leiomyomas (31,41-43). Also, some studies have been carried out to look for the possible association between the RAD51 gene variants and pancreatic (44-47), prostate cancer (48,49), malignant melanoma (50), colorectal adenocarcinoma (51), endometrial cancer (52), soft tissue sarcoma (53) and glioblastoma (54).

Our study aimed to determine associations between TBX15 rs98422, DNM3 rs1011731, RAD51B rs8017304, rs2588809 single nucleotide polymorphisms and pituitary adenoma invasiveness, development, and recurrence.

These findings support the hypothesised role of TBX15 , DNM3 and RAD 51 as tumour promoters. Based on the TBX15, DNM3 and RAD51B associations with cancerous processes we selected four widely described SNPs located in these genes. According to the dbSNP database (https://www.ncbi.nlm.nih.gov/snp/) the minor allele frequencies of these intronic variants (TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, rs2588809) are more than 0.1 in the Europe population, and none of these variants have been studied with PA development, invasiveness, PA activity and recurrence. The aim of the present study was to determine these associations.

Materials and Methods

Patients and selection. This study was carried out at the Department of Ophthalmology, Hospital of Lithuanian University of Health Sciences and Laboratory of Ophthalmology, Neuroscience Institute, Lithuanian University of Health Sciences. The Ethics Committee for Biomedical Research at Lithuanian University of Health Sciences (LUHS) approved the study (number BE-2-47). All subjects provided written informed consent under the Declaration of Helsinki. Based on our inclusion and exclusion criteria (55), two groups were formed in the study: the PA group (n=113) and the control group (n=283).

Evaluation of PA hormonal activity, invasiveness, recurrence and DNA extraction and genotyping. The analysis of all pituitary adenomas was based on histopathological findings of PA and hormone levels in the blood serum before surgery. All PA subjects were categorized into two groups – active and inactive PA (56).

Since some of the subjects had already had surgery in recent years, we categorized them by recurrence of pituitary adenoma into two groups – PA with and without recurrence.

Pituitary adenoma recurrence was diagnosed when enlargement of a residual tumor or a new growth was documented on follow-up magnetic resonance imaging (MRI) after surgical resection during the period of this study. The residual tumor was considered stable if there no signs of tumor progression on follow-up MRI. Most prolactinomas were surgically treated because of the remaining pressure effects of surrounding structures or ineffective medical treatment.

PA invasiveness has been described previously (55). The suprasellar extension and sphenoid sinus invasion by PA were classified according to the Hardy classification modified by Wilson, and the degree of suprasellar and parasellar extensions was graded as stages A–E. The degree of sellar floor erosion was graded as grades I-IV. Grade III shows localized sellar perforation, and grade IV shows diffuse destruction of sellar floor, which are the signs of invasive PA. The Knosp classification system was used to quantify the invasion of the cavernous sinus. Grade 3 and 4 pituitary tumors were considered to be invasive.

DNA was extracted from 200 μL venous blood (white blood cells) using the silica-based membrane technology utilizing a genomic DNA extraction kit (GeneJET Genomic DNA Purification Kit, Thermo Scientific, MA, USA), according to the manufacturer’s recommendations. The genotyping of TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and rs2588809 was carried out using the real-time PCR. SNPs were genotyped on the Step One Plus real-time PCR system (Applied Biosystems, Foster City, CA, USA). The TaqMan^® SNP genotyping assays (Thermo Scientific) for all SNPs were performed according to the manufacturer’s protocol. The Allelic Discrimination program was used during the real-time PCR. The program determined individual genotypes according to the fluorescence intensity rate of different detectors (VIC and FAM).

Statistical analysis. The age of study participants was presented as the median and interquartile range (IQR). It was compared between both study groups using the nonparametric Mann-Whitney U-test. All categorical variables of TBX15 rs98422, DNM3 rs1011731, RAD51B rs8017304 and rs2588809 genotypes and alleles were expressed as absolute numbers with percentages in brackets and compared using the Pearson’s χ² and Fisher’s exact test (when n<50) in both groups. Binomial logistic regression analysis was performed to evaluate the genotype and allele impact on PA development and reported as odds ratios (ORs) with 95% confidence intervals (CIs). The lowest values of the Akaike information criterion (AIC) showed the best genetic models. Statistically significant differences were reported when p<0.05, but for multiple comparisons, the Bonferroni correction was applied with the p<0.05/4 (since we analyzed four different SNPs).

Results

A total of 396 individuals were included in the study. Two groups of subjects were formed during the study. The first one included patients with pituitary adenoma, the second included healthy subjects (control group). The characteristics of the subjects are presented in Table I. The first group consisted of 113 individuals, of whom 45 (39.8%) were men, and 68 (60.2%) were women. The median age of this group was 54 years. The control group consisted of 100 (35.3%) men and 183 (64.7%) women. In total, the control group consisted of 283 individuals with a median age of 55.5 years.

Table I. Characteristics of study subjects.

TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 genotype frequencies in the pituitary adenoma and healthy population groups. Hardy Weinberg analysis was performed to compare the observed and expected frequencies of TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 using the χ² test in the control group. The genotype distribution of the polymorphisms matched the Hardy-Weinberg equilibrium. (p>0.001) (57). TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 genotypes and allele frequencies did not significantly differ between the PA and control groups. The results are shownin Table II.

Table II. TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and RAD51B rs2588809 genotype and allele frequencies in the PA patient and control groups.

PA: Pituitary adenoma; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4.

Binomial logistic regression analysis was performed to estimate the impact of genotypes and alleles on PA development. Binomial logistic regression analysis of RAD51 rs2588809 revealed that the TT genotype was associated with about 7-fold increased odds of PA development in the co-dominant (OR=6.833; 95% CI=2.557-18.262; p<0.001) and recessive (OR=7.066; 95% CI=2.667-18.722; p<0.001) models. Each copy of the T allele was associated with increased odds of PA development (OR=1.627; 95% CI=1.135-2.334; p=0.008) (Table III). Analysis of TBX15 rs984222, DNM3 rs1011731, and RAD51B rs8017304 did not show any statistically significant results (Supplementary material).

Table III. Binary logistic regression analysis of RAD51B rs2588809.

OR: Odds ratio; AIC: Akaike information criterion; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold.

Comparison of TBX15 rs984222, DNM3 rs101173, RAD51B rs8017304, and rs2588809 polymorphisms in pituitary adenoma patients by gender. Statistical analysis was also performed to compare the TBX15 rs984222, DNM3 rs101731, and RAD51 rs8017304 genotype and allele frequencies between the patients with PA and control group subjects by their gender (Table IV). The analysis of RAD51B rs2588809 showed a statistically significant difference in the CC, CT, and TT genotype distributions between females with PA and control females (69.12%, 16.18%, and 14.7% vs. 67.21%, 27.87%, and 4.92%, respectively, p=0.011). The results are presented in Table IV.

Table IV. TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and RAD51B rs2588809 genotype and allele frequencies in PA patients and controls by gender.

PA: Pituitary adenoma; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4.Significant p-Values are shown in bold.

Binominal logistic regression was performed to evaluate these polymorphisms’ impact on the PA development in men and women, separately. Binominal logistic regression analysis in the women’s group showed that the TT genotype was associated with 6.7-fold higher odds of PA development in the co-dominant model (OR=6.744; 95% CI=2.021-22.583; p=0.002) and with 7.7-fold increased odds of PA development in the recessive model (OR=7.716; 95% CI=2.332-25.533; p=0.001). The results are shown in Table V. The TBX15 rs984222, DNM3 rs1011731, and RAD51B rs8017304 were not associated with female PA development (Supplementary material). Also, no statistically significant variables were found in the men’s group (Supplementary material).

Table V. Binary logistic regression analysis of RAD51B rs2588809 in females.

OR: Odds ratio; AIC: Akaike information criterion; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold.

Association of TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 polymorphisms with clinical and morphological features of PA. One of our study’s objectives was to determine if there is a relationship between TBX15, DNM3, and RAD51B gene polymorphisms with PA’s clinical and morphological features. Comparing the distribution of genotypes and alleles of TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 between the PA groups by recurrence and the control group, we obtained statistically significant differences in the rs2588809 CC, CT, and TT genotype distributions between PA without-recurrence patients and healthy controls (67.03%, 17.58% and 15.39% vs. 69.96%, 24.73%, and 5.31%, respectively; p=0.005). The results are shown in Table VI. Regarding PA recurrence, we performed binominal logistic regression to evaluate the impact of TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 polymorphisms on the development of PAs with and without recurrence. We found that the RAD15B rs2588809 TT genotype was associated with approximately 8-fold increased odds of development of PA without recurrence in the co-dominant (OR=7.842; 95% CI=2.890-21.277; p<0.001) and recessive model (OR=8.394; 95% CI=3.122-22.571; p<0.001). Also, each T allele was associated with 1.7-fold increased odds of development of PA without recurrence in the additive model (OR=1.676; 95% CI=0.114-2.457; p=0.008). The data are presented in Table VII. No associations were found in the recurrent PA group (Supplementary material). The TBX15 rs984222, DNM3 rs1011731, and RAD51B rs8017304 were not associated with PA recurrence (Supplementary material).

Table VI. TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and RAD51B rs2588809 genotype and allele frequencies in patients grouped by PA recurrence and healthy subjects.

PA: Pituitary adenoma; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4.Significant p-Values are shown in bold.

Table VII. RAD51B rs2588809 association with PA without recurrence.

OR: Odds ratio; AIC: Akaike information criterion; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold.

TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 genotypes and allele frequencies were compared between the active and inactive PA and healthy control groups. We found that the RAD51B rs8017304 G allele was detected significantly more frequently in the inactive PA group vs. the control group (48.13% vs. 33.57%; p=0.004) (Table VIII).

Table VIII. TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and RAD51B rs2588809 genotype and allele frequencies in patients grouped by PA hormonal activity and healthy subjects.

p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold.

Binominal logistic regression revealed that the RAD51B rs2588809 TT genotype was associated with increased odds of active PA development in the codominant (OR=6.058; 95% CI=2.146-19.734; p=0.001) and recessive (OR=7.103; 95% CI=2.366-21.320; p<0.001) models (Table IX). Also, the RAD51B rs2588809 TT genotype was associated with increased odds of inactive PA development in the codominant (OR=7.247; 95% CI=2.29-22.906; p=0.001) and recessive (OR=7.260; 95% CI=2.260-21.840; p=0.001) models. Each T allele at rs2588809 was associated with 1.9-fold increased odds of inactive PA development in the additive model (OR=1.865; 95% CI=1.154-3.014; p=0.011). These data are presented in Table IX. The TBX15 rs984222, DNM3 rs1011731 and RAD51 rs8017304 were not associated with PA hormonal activity (Supplementary material).

Table IX. RAD51B rs2588809 associations with PA hormonal activity.

OR: Odds ratio; AIC: Akaike information criterion; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold

We then compared the distribution of TBX15 rs984222, DNM3 rs1011731, RAD51 rs8017304, and rs2588809 genotypes and alleles in patients with invasive and non-invasive PAs vs. healthy controls. The RAD51B rs2588809 genotypes (CC, CT, and TT) were distributed significantly differently in patients with non-invasive PA and healthy subjects (59.09%, 22.72% and 18.19% vs. 69.96%, 24.73%, and 5.31%, respectively, p=0.008) (Table X). Also, the T allele occurred more frequently in patients with non-invasive PA than in control subjects (29.55% vs. 17.67%, p=0.008). The results are presented in Table X.

Table X. TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and RAD51B rs2588809 genotype and allele frequencies in patients grouped by PA invasiveness and healthy subjects.

p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold.

Binominal logistic regression was performed in patients with PA by its invasiveness. It was revealed that the RAD51B rs2588809 TT genotype was associated with about 5-fold increased odds of invasive PA in the codominant (OR=4.881; 95% CI=1.570-15.172; p=0.006) and recessive (OR=5.212; 95% CI=1.693-16.050; p=0.004) models (Table XI). Also, the RAD51B rs2588809 TT genotype was associated with increased odds of non-invasive PA development in the codominant (OR=10.513; 95% CI=3.381-32.688; p<0.001) and recessive (OR=10.259; 95% CI=3.368-31.255; p<0.001) models. Each T allele was associated with 2.2-fold increased odds of non-invasive PA development in the additive model (OR=2.222; 95% CI=1.352-3.652; p=0.002). The results are shown in Table XI. The TBX15 rs984222, DNM3 rs1011731, and RAD51 rs8017304 were not associated with PA invasiveness (Supplementary material).

Table XI. RAD51B gene rs2588809 association with PA invasiveness.

OR: Odds ratio; AIC: Akaike information criterion; p-Value: Bonferroni corrected level of significance, differences are considered statistically significant when p<0.05/4. Significant p-Values are shown in bold.

Discussion

Our study analyzed the TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304 and RAD51B rs2588809 gene polymorphisms in PA patients (n=113) and healthy control subjects (n=283). The results were compared by gender, age, and the clinical course of the disease. Studies of these polymorphisms analyzing PA association with rs984222, rs1011731, rs8017304, and rs2588809 have not been performed yet, to the best of our knowledge.

The role of TBX family genes (TBX2 and TBX3) in oncogenic processes was associated with an increase of their expression level, as they have been found to be overexpressed in different types of cancer, including breast, cervical, ovarian, pancreatic, liver, and bladder cancer (58,59). TBX15 hypermethylation has been evaluated in prostate and ovarian carcinomas (19,20). No studies have been performed in association with any brain tumors, including PA. Our study was the first to find that the C allele of TBX15 rs984222 polymorphism reduced PA’s recurrence (p=0.037).

DNM3 has been shown to be involved in various malignancies (28-37). Marino et al. have reported DNM3 expression in the brain and testicles and less often in the lungs and heart (29). Inokawa et al. and Shen et al. have found that DNM3 is hypermethylated in hepatocellular cancer (HCC) (33-34). Zhang et al. have also studied the mechanism of DNM3 in HCC (60). Teicher et al. have reported liposarcoma 1q24.3 amplifications involving DNM3 (29) while low DNM2 expression has been associated with tumor invasion and metastasis in cervix carcinoma and up-regulation of matrix metalloproteinase 2 (MMP-2) expression (61). The DNM3 gene has also been investigated as a possible molecular marker for diagnosis and gene therapy of malignant diseases (38). Yang et al. have discussed the importance of the DNM3 gene in gliomas. As the DNM3 gene is the target of miR-221, the overexpression of DNM3 could reverse its tumor-promoting effect (31-32). Based on these findings, we sought to examine whether a polymorphism in the DNM3 promoter could impact PA development risk. Unfortunately, in our study, we did not find any statistically significant differences analyzing DNM3 rs1011731 gene polymorphism in relation to PA.

Concerning theother two gene polymorphisms, we found that the RAD51B rs2588809 CC genotype and the rs8017304 AG genotype might increase the probability of PA recurrence and invasiveness. Also, we proved that the RAD51B rs2588809 TT genotype might increase the odds of PA development in women and may be associated with PA development without recurrence. The RAD51B gene has previously been studied in other tumor types (breast, ovarian, and lung cancers (32,41) but not in brain tumors, so we could not compare our results with the results of other authors.

RAD51B has been previously evaluated as a candidate gene for breast cancer predisposition, but no mutation was detected in a study of 188 multiple-case breast cancer families (62). Previous studies have identified chromosomal rearrangements disrupting RAD51B in benign tumors, particularly uterine leiomyomas (42,43). In addition, the findings by Golmard and colleagues must be interpreted in the context of two genome-wide association studies (GWAS), which identified the minor allele of single nucleotide polymorphisms in RAD51B acting as a low-risk factor for breast cancer: the rs999737 (63) and rs1314913 (64), located in RAD51B introns 10 and 7, respectively. Results by Mengyin et al. also suggest that RAD51B could be a candidate prognostic factor for non-small cell lung cancer patients (41).

Overall, the present study of the TBX15 rs984222, DNM3 rs1011731, RAD51B rs8017304, and RAD51B rs2588809 gene polymorphisms requires future replication in studies with higher sample sizes to confirm the association of RAD51B rs2588809 with PA.

Conclusion

The RAD51B rs2588809 TT genotype was more common in women with PA than in healthy women, and the T allele was less frequent in men with PA than in healthy men. The RAD51B rs2588809 T allele increased the potential for PA invasiveness and PA activity. The likelihood of PA recurrence was reduced by the TT genotype and each T allele.

Data Availability

The genotyping data used to support the findings of this study is available from the corresponding author upon request.

Supplementary Material

Available at: https://docs.google.com/document/d/1U5Za-8j3e9mbHEye0LJlEcwVPL1f7-uSt7HlwRpslqM/edit?usp=sharing.

Conflicts of Interest

None of the Authors has any proprietary interests or conflicts of interest related to this submission

Authors’ Contributions

Conceptualization, R.L., and B.G.; Data curation, A.V., G.G., and B.G.; Writing-Original draft preparation, I.L., G.J., and R.L; Methodology, A.V., G.G., B.G., L.K, and R.L.; Investigation, A.V., G.G., G.J., I.L., and L.K.; Validation, L.K., and R.L.; Supervision, R.L., and B.G.; Writing-Reviewing and Editing, R.L.