ASSESSMENT AND COMPARISON OF HORMONAL IMMUNOEXPRESSION AND THE CLINICAL PICTURE IN PATIENTS WITH PITUITARY ADENOMAS

Abstract

Introduction

Symptoms related to hypersecretion of hormones in patients with pituitary adenomas do not always correlate with immunohistochemical staining results.

Objective.

To evaluate the relationship between the pituitary adenomas hormone immunoexpressions and endocrine presentations.

Patients and methods.

The clinical status and immunoexpression of 72 patients who underwent transsphenoidal surgery for pituitary adenomas were analyzed.

Results.

Macroadenomas were diagnosed in 51 cases (70.84%), while microadenomas were found in 21 cases (29.16%). The 72 adenoma specimens were divided into 22 monohormonal, 21 plurihormonal, 21 immunonegative and 8 unreliable specimens. The positive immunohistochemical staining results occurred as follows: prolactin and growth hormone 25% each, adrenocorticotropic hormone 13.89%, thyroid-stimulating hormone 5.56%, leuteinizing hormone and follicle-stimulating hormone 12.5%, glycoprotein hormone alpha-subunit 22.22%. Statistically significant relationships between the immunohistochemical presentation and the preoperative diagnosis were found for prolactin and hyperprolactinemia, growth hormone and acromegaly and adrenocorticotropic hormone and Cushing’s syndrome.

Conclusions.

The lack of full concordance between the clinical presentations and immunohistochemical staining was mainly a result of the presence of nonfunctioning adenomas, plurihormonal adenomas and unreliable specimens. The morphometric method introduced in this study, utilizing the immunoexpression index, provided a very precise evaluation of pituitary adenomas pathology.

INTRODUCTION

In patients with pituitary adenomas (PAs) the relationship between the signs or symptoms and immunohistochemical (IHC) staining is controversial (1-6). Positive IHC staining is not always associated with increased pituitary hormone levels in the blood, when measured before surgery, and nor with the clinical presentation (4-8). One possible cause of the aforementioned differences might be specimen reliability errors, such as tissue contamination, nondiagnostic tissue samples or an insufficient quantity thereof taken during surgery, or decreased biological activity of the hormones produced by the PAs. It is also necessary to remember that different hormones have different physiological patterns of secretion. Some of them, such as growth hormone (GH) or thyroid-stimulating hormone (TSH) are secreted in a pulsatile manner and others have a daily biological rhythm, e.g. prolactin (PRL) or adrenocorticotropic hormone (ACTH) (9). Furthermore, the secretion of gonadotropins depends on factors such as gender or age. Gonadotropins in males are secreted in a constant manner, while in females they are released in a pulsatile manner in relation to the menstrual cycle (10). Another relevant issue is the method of assessment of hormonal expression in PAs. Neuropathologists most commonly use a descriptive method that is based on the experience of the neuropathologist and a subjective evaluation. Some authors prefer to use scales to classify the intensity of the IHC staining, but these have not been standardized (11-13). Both qualitative and quantitative methods are unable to classify the hormone expression of PAs with one hundred per cent reliability and one of the crucial problems while diagnosing PAs are specimens with low or diffuse expression (14). In the literature, it is controversial to recognize such IHC stains as positive. Additionally, some authors establish a minimal expression level to classify IHC staining as positive (7, 13, 15-17).

These issues naturally lead to the conclusion that the selection of an appropriate morphometric method for ICH staining assessment is essential. In all cases of clinically non-functioning pituitary adenomas (CNFPA) there is an inconsistency between a lack of signs and symptoms from the hormone hypersecretion of the PA and positive IHC staining for those hormones (8, 18, 19). Hyposecretion, rapid destruction in lysosomes before secretion, and storage or synthesis disorders of the non-functioning hormones can all constitute a reason of the anomaly under discussion (20, 21). It is also worth mentioning that, in most cases, patients with functioning PAs are treated pharmacologically to suppress endocrine hypersecretion before neurosurgical procedures (22, 23). For this reason, comparing IHC results and a patient’s symptoms is not ideal and can be inaccurate.

Despite growing accuracy in the diagnosis of pituitary adenomas, both in IHC and with blood investigations, the clinical picture is still difficult to unravel. The aim of this study was to describe the relation between the IHC status of pituitary adenomas and the clinical presentation of patients.

PATIENTS AND METHODS

The study group consisted of 72 patients (44 females and 28 males) operated on at one university neurosurgical department (the Department of Neurosurgery and Neurotraumatology Jagiellonian University Medical College) during the period of seven years between 2011 and 2018. The mean age of the enrolled patients was 46.9 years, ranging from 10 to 84 years old. Only patients with a complete medical and histological record were admitted into the study. The main factors taken into consideration were as follows: the patient’s history, complaints, physical examination findings, comorbidities, medications, laboratory test results and the report from the neurosurgical procedure (transsphenoidal resection of PAs). During the study design, the patients’ ailments, as well as physical examination findings, were divided into two main categories: neurologic and endocrine.

The laboratory tests consisted of a pituitary hormone panel including glycoprotein hormone alpha-subunit (α-SU). The samples for those tests were taken at 7am for all patients, the day prior to operation. The results were obtained by the immunoradiometric method (IRMA - Immuno-Radio-Metric Assay). The laboratory reference ranges of hormone levels are presented in Table 1.

Table 1.

Reference ranges for hormone levels

Hormone	Sex	Menstrual cycle phase	Min.	Max.	Unit
PRL	F		50	800	μU/mL
	M		50	500	μU/mL
GH			0.2	17	μU/mL
ACTH			6	56	pg/mL
TSH			0,3	4,3	μU/mL
LH	F	Follicular phase	4	12	mU/mL
		Ovulation phase	>20		mU/mL
		Luteal phase	5	20	mU/mL
		Menopause	>10		mU/mL
	M		1	8	mU/mL
FSH	F	Follicular phase	4	12	mU/mL
		Ovulation phase	10	20	mU/mL
		Luteal phase	2	8	mU/mL
		Menopause	>25		mU/mL
	M		4	8	mU/mL

Abbreviations: ACTH- Adrenocorticotropic hormone, F-female, FSH- Follicle-stimulating hormone, GH- Growth hormone, LH- Luteinizing hormone, M-male, PRL- Prolactin, TSH- Thyroid-stimulating hormone.

In every patient, a full IHC panel for pituitary hormones (PRL, GH, ACTH, TSH, LH, FSH and α-SU) was performed. Table 2 summarizes the characteristics of each hormone staining. A control pituitary specimen was stained in order to confirm adequate IHC staining. IHC stained specimens were assessed with an optical microscope, Nikon Optishot-2 at 200x magnification. Non-diagnostic fragments (the presence of a normal pituitary tissue without PA, fibrosis, hemorrhage etc.) were excluded from further analysis. The morphometric evaluation was performed only if at least one field of view (at 200x magnification) included PA material without any damage (thermal or mechanical) or massive necrotic or hemorrhagic lesions. In cases that were not suitable for assessment, staining was repeated after performing another section. The precise quantitative IHC analyses of the PAs were carried out in the following steps. A cell index was used to assess hormonal expression. Under 400x magnification, the percentage of IHC positive stained cells was calculated. Morphometric analysis was performed manually using a morphometric adapter to an optical microscope and a morphometric grid (16 fields of equal area covering the entire field of view of the microscope under 400x magnification). In every field, all cells, including IHC positive cells, were counted. The analyses were performed systematically in 5 clean parts of the preparations, without any artifacts. A small number of dispersed positive cells situated on the perimeter of the adenoma presenting a positive IHC reaction were not counted because this was considered as impurity of “trapped” normal cell adenohypophysis.

Table 2.

Immunohistochemical staining

Antibody	Manufacturer	Dilution	Incubation	Reaction visualization	Comments
anti-PRL	Dako, Denmark	ready to use	15 min., room temp.	EnVISION TM + System Labelled Polymer HRP-ant rabbit	-
anti-GH	Dako, Denmark	ready to use	15 min., room temp.	EnVISION TM+ System Labelled Polymer HRP-ant rabbit	-
anti-ACTH	Dako, Denmark	ready to use	15 min., room temp.	EnVISION TM+ System Labelled Polymer HRP-ant rabbit	-
anti-TSH	Novocastra, UK	01:25	24 hours, 4°C	ABC KIT Vectastain Universal	unmasking antigen by 2×5min. in buffer pH=6.0
anti-LH anti-FSH	Dako, Denmark	01:25	30 min., room temp.	EnVISION TM+ System Labelled Polymer HRP-ant mouse	unmasking antigen by 2×5min. in buffer pH=6.0
anti-α-SU	Novocastra, UK	0,100	24 hours, 4°C	ABC KIT Vectastain Universal	unmasking antigen by 2×5min. in buffer pH=6.0

The study was conducted in accordance with the ethical standards laid down in the Declaration of Helsinki (1964) and its design was approved by the local University Ethical Committee (protocol number KBET/157/B/2012).

The results were shown using descriptive statistical methods, including ranges, means, standard deviations, and percentage distributions. Statistical analyses were performed using Statistica 9.0 PL, (StatSoft, Poland).

RESULTS

Anamnesis

Among neurological ailments, there was a predominance of headaches, reported by 24 patients (33.3%). Both dizziness and nausea were reported in 3 cases (4.17%), whereas vomiting in 2 (2.78%). The majority of the endocrine complaints group was made of female patients with menstrual disorders; a lack of or irregular menstruation was reported in 14 cases (31.81% of females). 23 females were pre- and 21 were post-menopausal. Infertility and pathological alopecia appeared in 2 patients (2.78%). Other complaints were singular cases such as potency disorders, fatigue, hot flashes, muscle aches, facial flushing and hypotonia. Patient comorbidities are presented in Figure 1. All 24 patients with recognized hyperprolactinemia were treated with bromocriptine. In 10 patients (76.92%) with acromegaly the therapy with somatostatin analogues was conducted prior to surgery. One thyrotropinoma patient took thyrostatics.

Figure 1.

Presentation of patients’ comorbidities.

Physical examination

Physical examination revealed that the most frequently occurring neurological signs were disturbances in vision: blurry vision was recognized in 25 (34.72%) cases and bitemporal hemianopsia appreciated in 22 (30.56%) cases. Other symptoms were forgetfulness, balance disturbances, diplopia, psychomotor deterioration, sphincter disturbances, ptosis, anisocoria and abducens nerve palsy. The endocrine symptoms were dominated by acromegaly features in 12 (16.67%) cases. Signs of Cushing’s syndrome were reported in 7 (9.72%) patients and features of hypopituitarism in 2 (2.78%) cases. One patient with precocious puberty was a boy, who revealed the following symptoms: testicular enlargement, phallic growth, the initial appearance of pubic, facial, and axillary hair. The diagnosis has been confirmed by the endocrinologist at the age of 9 years. The transsphenoidal surgery was performed at the age of 10 years.

Radiological investigation

Macroadenomas were diagnosed in 51 cases (70.84%), while microadenomas were found in 21 cases (29.16%). The tumor size and invasiveness evaluation according to the Hardy-Wilson and Knosp scales are demonstrated in Table 3 (24, 25).

Table 3.

The tumor size and invasiveness evaluation according to the Hardy-Wilson and Knosp scales

Hardy -Wilson scale degree	% of evaluated subjects	Knosp scale grade	% of evaluated subjects
I 0	29.17	0	55.56
II 0	8.33	1	8.33
II A	11.11	2	15.28
II B	6.94	3	16.67
II D	1.39	4	4.17
III A	5.56
III B	5.56
III C	5.56
III E	1.39
IV B	4.17
IV C	5.56
IV D	8.33
IV E	6.94

Laboratory tests

The hormone level results have been divided into 3 sections (below, in and above the reference range) and the percentages of presentations in these three groups are shown in Table 4. The endocrine diagnoses, which were made by taking into consideration patients’ complaints, physical examination and hormone laboratory test results, are presented in Figure 2.

Figure 2.

Endocrine preoperative diagnoses.

Table 4.

Hormone levels with the percentage of outcomes divided into 3 sections: below, in and above the reference range

	Below reference		Reference range		Above reference
	n	%	n	%	n	%
PRL	0	0	48	66.67	24	33.33
GH	0	0	61	84.72	11	15.28
ACTH	0	0	66	91.67	6	8.33
TSH	7	9.72	63	87.5	2	2.78
FSH	17	23.61	49	68.06	6	8.33
LH	25	34.72	45	62.5	2	2.78

Immunohistochemical staining

In the presented study group the majority of patients were IHC staining positive (Figure 3). More precisely, monohormonal adenomas were diagnosed in 22 cases (30.56%), plurihormonal in 21 cases (29.17%) and PAs with negative IHC staining in 21 cases (29.17%). In 8 cases (11.11%), the interpretation of the IHC examination was inconclusive. These samples were labeled as unreliable and were excluded from further analysis.

Figure 3.

Positive IHC staining results.

Comparison of hormonal immunoexpression and the clinical picture

We found that 36.11% of patients presented with clinically nonfunctioning adenomas. These patients first presented with mass effect symptoms such as visual field loss. The remaining 63.89% presented with symptoms of hyperprolactinemia (33.33%), acromegaly (18.06%), Cushing syndrome (9.72%), hyperthyroidism (1.39%) and precocious puberty (1.39%). We present the distribution of functioning pituitary adenomas in Figure 4.

Figure 4.

Functioning pituitary adenomas.

The relation of the percentage distribution of immunoexpression on the preoperative diagnosis is demonstrated in Table 5. We managed to determine the relationship between preoperative diagnosis and the percentage intensity of IHC staining for hormones such as PRL, GH, ACTH. The highest mean percentage of immunoexpression in groups of patients diagnosed with hyperprolactinemia, acromegaly and Cushing’s syndrome were measured as PRL, GH, and ACTH respectively. No statistically significant relationship was found in respect to TSH, gonadotropins or α-SU. When comparing the percentage of immunoexpression of PA hormones with the hormone level measured in the serum, statistically significant relationships for PRL and ACTH were found. The results are shown in Table 6.

Table 5.

The relationships between preoperative diagnosis and the percentage of immunoexpression of specific hormones

	Mean percentage of immunoexpression of a specific hormone
Preoperative diagnosis	PRL	GH	ACTH	TSH	LH	FSH	α-SU	p-value
Clinically non- functioning adenomas (n=26), mean± SD	0	0	2.24±8.63	0.46±2.35	3.62±11.01	5.43±16.96	8.28±17.96	<0.001
Acromegaly(n=13), mean± SD	5.82±15.23	58.06±28.64	0.62±2.22	0.33±1.19	6.45±15.86	4.08±14.70	13.32±28.00	<0.001
Cushing Syndrome (n=7), mean± SD	10.57±18.95	11.87±31.41	31.51±33.36	0	0	19.27±34.87	15.09±31.49	<0.001
Hyperprolactinemia (n=24), mean± SD	38.70±36.09	13.07±27.36	4.39±12.28	3.32±15.58	6.58±17.35	6.79±19.25	6.24±19.66	NS
Hyperthyreosis (n=1), mean± SD	0	0	0	7.1	0	0	87.50	NS
Precocious puberty (n=1) mean± SD	0	0	0	0	0	0	0	NS

Table 6.

Pearson’s Correlation between the percentage of IHC positive staining results and the hormone level in the blood serum

Correlated parameters		r	n	P
PRL % immunoexpression	PRL level	0.522	n=18	P=0.026
GH % immunoexpression	GH level	0.171	n=18	P=0.498
ACTH % immunoexpression	ACTH level	0.643	n=10	P=0.045
TSH % immunoexpression	TSH level	-0.278	n=4	P=0.722
LH % immunoexpression	LH level	-0.233	n=9	P=0.546
FSH % immunoexpression	FSH level	-0.476	n=9	P=0.196

DISCUSSION

In this study we determined the correlation between pituitary adenoma IHC status, hormonal blood levels and clinical presentation. We observed the highest correlation between the expression of PRL, ACTH and blood hormone levels. When it comes to TSH, LH, FSH, there was no significant correlation. We also found that the most common presentation was a mass effect from hormonally non-functioning adenomas.

Some studies have demonstrated a correlation between serum hormone levels and IHC results (8). At the same time, there are other authors who have demonstrated a lack of the aforementioned relationship (4, 26, 27). There are some factors which indicate that pituitary hormone levels are not the best parameter reflecting a patient’s clinical condition. Growth hormone level can be in the reference range while examined due to the pulsatile fashion of secretion. ACTH undergoes diurnal variations in secretion. In such cases, hormonal disorders can be proven through functional tests or by detecting abnormal levels of hormonal effectors such as IGF1 or cortisol (28). Furthermore, preoperational pharmacotherapy in most cases modifies circulating hormone levels as it is used to reduce the severity of symptoms or to decrease the size of the PA (29). In our study group, all patients with PRL secreting PAs were treated with bromocriptine whereas majority of GH secreting PAs were treated with somatostatin analogues. For these reasons, we compared the results of IHC staining with both preoperational diagnosis and hormone levels. It was important to compare both of these factors as basal hormone levels do not usually correspond with the patient’s condition and a formal endocrine diagnosis includes all information about the patient including complaints, signs and symptoms.

Figure 5.

Examples of pituitary adenoma stainings. a. Hematoxylin and eosine staining. Magnification 400x b. Prolactinoma. Anti-PRL staining. Magnification 200x c. Somatotropinoma. Anti-GH staining. Magnification 400x d. Corticotropinoma. Anti-ACTH staining. Magnification 400x e. Thyrotropinoma. Anti-TSH staining. Magnification 200x f. Gonadotropinoma. Anti-LH staining. Magnification 400x g. Gonadotropinoma. Anti-FSH staining. Magnification 200x h. Anti- α-SU. Magnification 400x

Without any doubt, using the percentage rate of IHC staining is precise, less subjective and gives a new opportunity to compare similar studies between different authors with an overall lower inaccuracy rate. Our results indicate an association between the IHC staining results and preoperative diagnoses in the cases of PRL, GH, and ACTH and they correlate with hyperprolactinemia, acromegaly and Cushing’s Syndrome respectively. These results correspond well with other studies in the literature and the evaluation of this factor may be helpful in the prognostic process (8). However, the relationship between IHC positive staining for ACTH and Cushing’s Syndrome is low in comparison with other results (8, 30-33). During the study we found one case of hyperthyroidism with positive IHC staining for TSH. This finding is rare and corresponds with the literature, where such cases of functional PAs with TSH hypersecretion are reported in 0.5-2.0% of cases and has a tendency to increase (23, 34). In most cases, TSH is co-expressed with GH in patients with acromegaly and is not accompanied by hyperthyroidism (32). In our study there were no patients with hypersecretion of gonadotropins. This is also consistent with the work of other researchers as immunopositivity for LH or FSH in PAs rarely presents clinically (17, 33). The expression of free α-SU did not correlate with any of the clinical presentations, and this factor alone has not been associated with any kind of endocrinopathy. Nevertheless, assessing α-SU expression is included in the IHC panel while diagnosing PAs and there are reports which highlight greater recurrence of α-SU positive PAs after operations. Desai et al. established that PAs secreting ACTH and α-SU are more aggressive than those without α-SU expression (35). It has been proven that α-SU immunopositive PAs present with a suprasellar extension more frequently and usually have a Ki67 proliferation greater than 3% (36). Preoperative endocrine diagnoses were compared to post operational diagnoses by IHC at a 46% compatibility level by Sung-Ku et al. who included FPA and CNFPAs in their study (4). Similarly, our results showed that preoperative diagnoses and IHC results do not fully correspond with each other. However, we managed to establish a lower than previously reported percentage rank of agreement in all the hormones analyzed. These differences might have been caused by only including patients undergoing neurosurgical treatment. This cohort includes more CNFPA patients with mass effect symptoms rather than patients with endocrine syndromes. CNFPA are a heterogeneous group of tumors including null cell adenomas, oncocytic adenomas, gonadotropinomas, silent corticotropinomas and other silent adenomas (18). Although the silent pituitary adenomas present positive immunohistochemical staining results, they do not secrete hormones at clinically adequate level (19). It has been well established that young patients with silent corticotropic adenomas have a higher frequency of multiple and late recurrences, and more aggressive tumor behavior in comparison to CNFPAs without ACTH immunoreactivity and for this reason these patients require long-term monitoring (30). Hence IHC investigation is obligatory even in group of CNFPA. Lack of full concordance between clinical and IHC presentations is related also to plurihormonality. This group of tumors are often related to more aggressive behavior (7, 35).

Making the comparison between IHC results and hormone levels we revealed a positive correlation for PRL, GH, ACTH and negative correlation for TSH, LH, FSH. IHC positive staining was significantly correlated to hormone levels of PRL and ACTH. This differs from the comparison between the percentage of immunoexpression and preoperative diagnoses of the results of GH, as they were significant. All of this evidence points to the fact that the pituitary hormone level does not necessarily correspond with clinical demonstration.

In the functional PA group, our measurements show that 15.55% of cases were IHC negative. In all of these cases, the patients had symptoms of hyperprolactinemia and all the tumors involved were macroadenomas. In such cases, we assumed a non-primary hyperproduction of hormones, or the so-called stalk effect (37, 38-41).

In conclusion, the discordance between clinical and IHC presentations is caused mostly by CNFPAs and plurihormonal adenomas. Due to the fact that plurihormonality is a frequent occurrence in PAs, IHC staining for all pituitary hormones is obligatory for full categorization. Limiting an IHC staining panel can lead to false or defective diagnoses. A morphometric IHC assessment with an immunoreactive index allows a neuropathologist to meticulously evaluate the pathology of PAs, decreases the subjectivity of the researcher and provides a useful tool for objectifying results. This method should be considered as a useful or possibly standard approach for diagnosing PAs.

Conflict of interest

The authors declare that they have no conflict of interest.