Evaluation of 06-methylguanine-DNA methyltransferase expression in children and adolescent pituitary adenoma

Xueming Shen Capital Medical University Sanbo Brain Hospital Yakun Yang Capital Medical University Sanbo Brain Hospital Zuocheng Yang Capital Medical University Sanbo Brain Hospital Ning Liu Capital Medical University Sanbo Brain Hospital Xueling Qi Capital Medical University Sanbo Brain Hospital Hongqi Xu Department of Neurosurgery, The People's Hospital of Anyang City Changxiang Yan Capital Medical University Sanbo Brain Hospital Song Han (  13581776844@163.com ) Capital Medical University Sanbo Brain Hospital https://orcid.org/0000-0001-7538-5185


Background
Pituitary adenomas (PA) are uncommon in children and adolescents, accounting for < 3% of childhood supratentorial tumors and between 3-6% of all surgically treated adenomas [1]. PA has a serious impact on the growth, development, vision, and future fertility of the patient, and may be secondary to cardiovascular events [2][3][4][5]. Like adults, pituitary adenomas in children and adolescents are also treated differently depending on the type of tumor. For example, lactotroph adenomas are preferentially treated with dopamine agonists [6,7]; corticotroph adenomas and somatotroph adenomas are treated surgically [3,8,9]. Compared with adults, invasive PAs are more common in children and adolescents, furthermore, the incidence of apoplexy is also higher [4,10]. Although a signi cant progress has been made in the medical management of pituitary adenomas over the past decades, the invasive tumors are di cult to manage and are associated with poor prognosis and fatality, the treatment of these tumors still remains challenging, as conventional chemotherapy exhibits low response rates, whereas radiotherapy fails to control tumor growth and may be associated with considerable side effects [11,12].
Temozolomide (TMZ), a routine chemotherapy for glioblastoma (GBM), is an imidazotetrazine derivative, oral cytotoxic chemotherapeutic agent that inhibits DNA replication. For aggressive PA, pituitary carcinoma and atypical pituitary adenomas (APA) (WHO 2004 classi cation, the WHO 2017 classi cation delete diagnosis [13,14]) treatment in adults could choose temozolomide [15][16][17][18]. This drug has currently become a therapeutic option for aggressive PA, which has been shown to reduce tumor volume and reduce the pituitary hormone levels The cytotoxic activity of temozolomide depends on the expression of O-6 methylguanine DNA transferase (MGMT), a DNA repair enzyme present in tumor cells that could lead to the inactivation of TMZ. The lower expression of MGMT is reportedly predictive of responsiveness to TMZ [19,20]. And several studies report that MGMT expression is associated with PA patients' prognoses [21].
However, the use of temozolomide has not been reported in children and adolescents. The e cacy of temozolomide can be evaluated using MGMT protein expression levels [22,23]. In addition, the expression of MGMT has a positive effect on prognosis [12,24]. Therefore, this study detected the expression of MGMT protein in PA tissue samples of children and adolescents, combined with some biological characteristics and molecular indicators for a comprehensive evaluation.

Methods
This study was a single-center study of 38 patients with surgically-treated PA, admitted to the Sanbo Brain Hospital, Capital Medical University, from June 2012 to September 2017. The age range of reported PA in adolescents is 18-21 years [2,9,[25][26][27]. Patients 20 years or younger with PA were included in this study. The ethics committee of Sanbo Brain Hospital approved the study. Written informed consents were obtained from all the patients involved in the study.

Immunohistochemistry (IHC)
Tumor specimens were xed in 10% formaldehyde and embedded in para n for histological sectioning.
Para n-embedded sections were depara nized in xylene and dehydrated in graded alcohol. Immunohistochemistry studies for the MGMT protein were performed using a mouse monoclonal antibody (1:10) (clone UMAB56; ZSGB-BIO). In addition, we screened for other molecular markers, including Ki-67 (1:200) (clone Mib1; ZSGB-BIO) and p53 (1:300) (clone BP53.12; ZSGB-BIO). Antigen retrieval was performed in citrate buffer at pH 9.0. The sections were incubated overnight at 4℃ with the primary antibody. Following this, the sections were rinsed with 1X phosphate buffered saline (PBS) and incubated with the horseradish peroxidase conjugated secondary antibody, followed by a rinse in PBS, incubation with 3,3'-diaminobenzidine staining and counterstaining with hematoxylin blue. The negative control sections were treated using the same parameters except that PBS without antibody was used for the incubation steps.
Scores 1 and 2 were combined to form the category of low level MGMT expression; scores 3 and 4 represented intermediate and high MGMT expression, respectively [12,28]. The Ki-67 index was de ned as the proportion of MIB-1 positive tumor nuclei out of 500 total cells. The p53 labeling index was determined by the percentage of positively stained nuclei.

Tumor Invasive Capacity and Texture
The invasion of suprasellar and saddle base was determined according to the Hardy classi cation. Invasive PA was determined by III, IV or C, D, E on the Hardy scale [29]. Parasellar invasion was determined using the Knosp grading and grades and were de ned as invasive PA [30]. Based on the Mahmoud classi cation method [31], the tumors were divided into three groups including a. soft texture group: a tumor that is easily suctioned with an aspirator, b. medium texture group: a tumor that is hard to remove with an aspirator, and c. tough texture group: a tumor that cannot be suctioned with an aspirator, and for which bipolar electrocoagulation or sharp segmentation resection is needed.
Statistical Analysis SPSS 21.0 was used to analyze the data. The measurement data were expressed as the mean ± SD (standard deviation). The Student's t-test was used to compare the mean difference between two measurement data. The relationship between MGMT expression and age, sex, tumor diameter, type, texture, p53, Ki-67, apoplexy and recurrence were analyzed using the chi square test or logistic regression.
Bilateral p < 0.05 for the difference was statistically signi cant.

Sample Characteristics
The patient cohort included 38 cases, aged 12-20 years (17.67 ± 1.63 years), of which 23 were males and 15 were females, accounting for 3.5% of the 1073 surgically treated pituitary patients in the same period. The clinical manifestations are: headache, menstrual disorder, visual impairment, accelerated development, galactorrhea, growth retardation, gynaecomastia, etc. (Table 1). Of these, 13 were invasive PA cases and 12 were cases of apoplexy (31.58%). The texture of the tumor was soft in 30 cases (78.95%), moderate in three cases (7.89%), and tough in ve cases (13.16%).

MGMT Expression Instability
Two patients in this cohort were treated with two surgeries each in our center,and MGMT protein expression was detected in the specimens. In these two patients, one of them was a 12-year-old boy with growth hormone pituitary adenoma. The rst transsphenoidal surgery failed to completely remove the tumor, so the tumor recurred 11 months after the surgery, and the tumor was nally completely removed by pterional approach. However, MGMT expression was 95% (high expression), Ki-67 was 4% and p53 was negative in the sample in rst operation (Fig. 1C) and 10% (low expression), Ki-67 was 2% and p53 also was negative in the second sample in the second surgery (Fig. 1F). In the other patient with nonfunctioning pituitary adenoma, after the rst transsphenoidal subtotal resection of pituitary adenoma, the second transsphenoidal resection was performed 10 months later due to tumor recurrence. The rst specimen was negative and the second specimen was 60% positive for MGMT (high expression). Ki-67 and p53 were expressed relatively stably (Fig. 1A, B, D, and E).

MGMT and Tumor Biological and Molecular Characterization
The Chi-square test or logistic regression were used to analyze the relationship between the expression of MGMT and age, gender, tumor biological characteristics such as diameter, invasiveness, texture, apoplexy, Ki-67, and p53 expression. A signi cant correlation was identi ed between the tumor diameter and MGMT expression (B = 2.56, SE = 0.99, P = 0.01, Exp (B) = 12.89). The larger the tumor diameter, the higher the MGMT expression. The remaining indicators did not show a correlation (Table 2).

Discussion
PAs are very rare in children and account for about 3% of intracranial tumors [1]. The incidence of PA increases in adolescence, but is still a relatively rare tumor [32]. In this group of 38 cases, ve were children and 33 were adolescents, suggesting that PA incidence may be related to changes in adolescent hormone levels. Biological characteristics of PAs in children and adolescents include: high rates of functional PAs [25], high rates of invasive PAs [33], and high rates of apoplexy [10]. In adults, temozolomide may be used for aggressive PA, PCs, and APAs [15][16][17][18]. In addition, bromocriptine resistance in patients with temozolomide treatment is equally effective [34,35]. The e cacy of temozolomide can be evaluated using MGMT protein expression levels [22,23]. In addition, the expression of MGMT has a positive effect on prognosis [12,24]. The assessment of MGMT expression in PAs in children and adolescents has not previously been studied.
Immunohistochemical detection of MGMT protein expression is relatively easy in para n-embedded tissue. Low MGMT expression predicts positive response (tumor inhibition) to temozolomide [23]. In this study, the expression of MGMT protein was detected in 38 cases of PA in children and adolescents, and 16/38 (42.11%) of these showed low MGMT expression. A previous study reported low MGMT expression in 26% of functional PA in a cohort of adult-dominated patients [24]. Low MGMT expression was also reported in prolactinoma (78-85.71%) [36,37], corticotroph adenomas (62.5-81.48%) [37,38], and somatotroph adenomas (90.0-91.7%) [37,39]. In the current study, the functional PAs were 32/38 (84.21%), of which 13/32 (34.21%) showed low MGMT expression. Low MGMT expression was seen in 11/22 (50.00%) prolactinomas, 1/9 (11.11%) somatotroph adenomas, and only one case of corticotroph adenoma. Overall, low MGMT expressing somatotroph adenomas are not as common in children and adolescents compared to adults. Therefore, this type of pituitary adenoma may not be suitable for temozolomide treatment.
The proportion of invasive PAs in this cohort was 13/38 (34.21%). Although there was no correlation between MGMT expression and invasiveness (p > 0.05), 6/13 (46.15%) showed low MGMT expression and may be suitable candidates for treatment with temozolomide. In adult-dominated PAs, the correlation between MGMT expression and invasiveness is reported inconsistently. Some studies suggest a correlation between low to moderate MGMT and increased invasiveness [24], while others do not support the same correlation [37,40]. To predict the likelihood of temozolomide response, all invasive pituitary tumors should be evaluated for MGMT expression. Interestingly, the study found that the diameter of the tumor was related to the expression of MGMT (p = 0.01). The larger the tumor, the higher would be the MGMT expression,100% of microadenomas showed low MGMT expression. However, only 25% and 22.2% of macro and giant adenomas showed low MGMT expression, respectively. These results suggest that macroadenoma and giant adenoma are not suitable for temozolomide treatment.
Ki-67, a nuclear antigen, is a proliferation marker in different neoplastic lesions. In this cohort, there was no correlation between MGMT expression and Ki-67 expression (p = 0.09). Inactivation of p53 is known to be involved in aberrant cell proliferation. p53 expression is a prognostic indicator of various primary human tumors. Therefore, immunohistochemistry was performed to detect the expression of p53 in pituitary adenomas. In this cohort, most of the samples were positive for p53 expression; however, there was no correlation between p53 expression and MGMT expression (p = 0.58).
The expression level of MGMT in pituitary tumor tissue may change over time, but Lau Q et al reported stable MGMT expression in most primary and recurrent tumor samples [28]. Radiation therapy has a signi cant impact on the level of MGMT expression in tumors. Among the two patients, one was a somatotroph adenoma, while the other was a non-functioning PA. No radiation therapy was applied. The two cases were treated with prednisone replacement therapy, and one was treated with desmopressin acetate tablets for diabetes insipidus. It remains unclear if the change of MGMT expression is related to surgery, drug treatment, or other unknown factors. Further studies are warranted to understand the differential expression of MGMT, particularly in cases undergoing surgical interventions.
In this article, the MGMT expression of PAs in children and adolescents were evaluated for the rst time. Somatotroph adenomas and large-diameter PAs, both of which not have low MGMT expression, may not be suitable for temozolomide treatment. In the context of PAs in children and adolescents, the measurement of MGMT expression could serve as an indicator for choice of temozolomide as a treatment option.

Declarations
Ethics approval and consent to participate The studies involving human participants were reviewed and approved Ethics Committee of Sanbo Brain Hospital. The patients/participants provided their written informed consent to participate in this study. For participants aged under 16, their parents or legal guardians provided written informed consent for their participation in the study.

Consent for publication
Written, informed consent was obtained from the individual(s) AND/OR minor(s)' legal guardian/next of kin for the publication of any potentially identi able images or data included in this article.

Availability of data and materials
The clinical datasets generated during and/or analyzed during the current study are available from the corresponding authors on reasonable request.