Acute elevation of interleukin 6 and matrix metalloproteinase 9 during the onset of pituitary apoplexy in Cushing’s disease

Abstract

Purpose:

Pituitary apoplexy is a rare endocrine emergency. The purpose of this study is to characterize physiological changes involved in pituitary apoplexy, especially during the acute phase.

Methods:

A Cushing’s disease patient experienced corticotroph releasing hormone (CRH)-induced pituitary apoplexy during inferior petrosal sinus sampling (IPSS). The IPSS blood samples from the Cushing’s disease patient were retrospectively analyzed for cytokine markers. For comparison, we also analyzed cytokine markers in blood samples from two pituitary ACTH-secreting microadenoma patients and one patient with an ectopic ACTH-secreting tumor.

Results:

Acute elevation of interleukin 6 (IL-6) and matrix metalloproteinase 9 (MMP9) was observed in the IPSS blood sample on the apoplectic hemorrhagic site of the tumor. In contrast, such a change was not observed in the blood samples from the contralateral side of the apoplexy patient and in other IPSS samples from two non-apoplexy Cushing’s disease patient and a patient with ectopic Cushing’s syndrome.

Conclusion:

IL-6 and MMP9 may be involved in the acute process of pituitary apoplexy in Cushing’s disease.

Introduction

Pituitary apoplexy is a clinical emergency, and is characterized by the sudden onset of headaches, vomiting, visual impairment, ophthalmoplegia, altered levels of consciousness and hormonal dysfunction [1, 2]. Hemorrhaging or hemorrhagic infarction into an underlying pituitary adenoma is the most common cause of this condition. Pituitary apoplexy occurs in 0.6–9% of all pituitary adenomas [3], and in up to 25% of the pituitary tumors if subclinical spontaneous hemorrhage cases are included [4]. Acute physiological changes involved in pituitary apoplexy are not well characterized, due to the rarity of this condition. Understanding the underlying mechanisms in pituitary apoplexy is an important step toward development of appropriate therapeutic strategies.

Treatments and diagnostic procedures including the administration of endocrine stimulants are possible precipitants of pituitary apoplexy. Endocrine stimulants-induced apoplexy has been described following administering gonadotropin-releasing hormone (GnRH), thyrotropin-releasing hormone (TRH), dexamethasone, and corticotropin releasing hormone (CRH) [4] and dopamine agonist; a medical treatment for prolactinoma [1, 5]. The inferior petrosal sinus sampling (IPSS) procedure is used in the evaluation of Cushing’s disease. This procedure exhibits high sensitivity (80–100%) and specificity (>95 %) [6, 7], and involves serial collection of blood samples from bilaterally placed catheters in the inferior petrosal sinuses. The sensitivity of this procedure can be increased with the intravenous injection of an endocrine stimulant such as CRH or a vasopressin analogue (desmopressin).

The mechanisms of pituitary apoplexy are poorly understood. Current evidence suggests the involvement of several genes including HIF1A (encoding Hypoxia-inducible factor 1 α), TNFA (encoding Tumor necrosis factor-α), IL1B (encoding Interleukin-1β), IL-6 (encoding Interleukin-6), MMP2 (encoding Matrix metallopeptidase 2), MMP9 (encoding Matrix metallopeptidase 9), VEGF (encoding Vascular endothelial growth factor), and PTTG (encoding Pituitary tumor transforming gene) [8–14]. However, previous studies were carried out by retrospective analysis of these genes in resected apoplectic tumors, which only addressed a later event in the apoplectic timeline. Acute changes of molecular markers at the onset of pituitary apoplexy remain uncharacterized.

Matrix metallopeptidases or metalloproteinases (MMPs) are members of a peptidase family that degrade components of the extracellular matrix. MMPs are involved in many pathophysiological processes including inflammation, tissue repair, tumor invasion and metastasis. MMP2 and MMP9 subtypes possess collagenase activity and degrade type IV collagen, which is enriched in the extracellular space of the pituitary gland [15]. Studies have shown that levels of type IV collagen are reduced in pituitary adenomas, suggesting that the activity of MMP2 and/or MMP9 is upregulated in pituitary adenomas [15, 16].

Here we present a rare case of CRH-induced pituitary apoplexy, which was likely precipitated by IPSS. We retrospectively analyzed the expression of the pituitary apoplexy markers HIF1-α, TNF-α, IL-1β, IL-6, MMP2, MMP9 and VEGF-A using the stored blood samples obtained during IPSS. These markers were chosen for this study because previous reports identified elevated expression of these markers by post apoplexy pituitary tumor immunostaining. Additionally, VEGF-A is a key regulator of angiogenesis in pituitary adenomas [13, 14]. These markers were also found to be elevated during other vascular events, such as ischemic stroke [17, 18], ischemic heart disease [19] and other vascular diseases [20, 21]. Moreover, serum HIF1-α levels are elevated in other medical conditions, such as diabetes [22], chronic kidney disease [23], chronic obstructive lung disease [24] and non-small cell lung cancers [25]. Among these markers, we found acute elevation of the levels of IL-6 and MMP9 in the IPSS blood samples after CRH injection, suggesting that early changes in these factors may be associated with onset of a pituitary apoplexy event.

Case

A 42-year-old woman, who was previously healthy, presented with fatigue, weakness and headache. Physical examination showed the subject had a fit BMI (23 kg/m²) with elevated blood pressure 177/103 mmHg and several skin eruptions on extremities. She did not have truncal obesity, hirsutism or striae. Initial lab tests (complete blood count, comprehensive metabolic panel and urinalysis) showed a markedly low potassium level at 2.2 mmol/L (ref: 3.4–5.3 mmol/L). Brain magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) showed left vertebral artery dissections. Carotid arteries were intact and there was no evidence of stroke or brain hemorrhage. A pituitary mass was incidentally detected, measuring 19 × 20 × 23 mm extending into the left cavernous sinus (Fig. 1A, yellow arrowheads). The optic chiasm was not compressed. Further endocrinology workup showed normal level of prolactin 14 μg/L (ref: 3–27 μg/L), IGF1 97 ng/ml (ref: 73–263 ng/ml) and free T4 0.82 mg/dL (ref: 0.76–1.40 mg/dL), low TSH 0.2 mU/L (ref: 0.4–4.0 mU/L), LH <0.2 IU/L and FSH 0.5 IU/L, and high ACTH 270 pg/ml (ref: < 47 pg/ml). Peripheral blood morning cortisol was significantly elevated at 51.2 ug/dL (ref: 4–22 ug/dL). Twenty four-hour urinary free cortisol level was 17,097 ug/dL (ref: 3.5–45 ug/d). Low dose (1 mg) and high dose (8 mg) overnight dexamethasone suppression tests failed to suppress cortisol levels, which were 37.5 ug/dL and 67.3 ug/dL, respectively. Computed tomography chest abdomen and pelvis showed bilateral adrenal hyperplasia. Ketoconazole treatment was started for severe hypercortisolemia management. Her headache resolved and her blood pressure reduced (131/91mmHg), and she was discharged after a 2-day hospital stay.

Figure 1. A case of pituitary apoplexy during IPSS.

(A – B) Brain magnetic resonance imaging (MRI). MRI shows coronal (left) and sagittal (right) images T1-weighted with gadolinium contrast. (A) MRI images before apoplexy show a large solid sellar mass (yellow arrowheads). (B) MRI images after apoplexy show expansion of tumor to left cavernous sinus (red arrowheads). (C) Photos before IPSS (left) and after IPSS with left ophthalmoplegia (right). (D) MRI images after the surgical resection of the adenoma (E) ACTH levels from blood obtained from the left (red) and right (blue) petrosal sinus, and peripheral vessels (gray). (F) H&E-stained image of islands of viable adenoma with large area of adjacent hemorrhage. (G) H&E stained image of viable pituitary tumor cells. (H) Enlarged H&E-stained image of acute necrosis and hemorrhage with cell ghosts. (I) Immunohistochemistry of ACTH in the same area as (F). (J) Immunohistochemistry of KI-67, showing low labeling of viable tumor. (K) Immunohistochemistry of P53, showing low labeling of viable tumor. Scale bar, 30 μm in H, I, J, K, and 20 μm in F and G.

Although there was an evident pituitary macroadenoma, non-secreting pituitary adenoma with ectopic ACTH-secreting tumor was still in our differential diagnosis for the following reasons: the lack of classical physical findings of pituitary ACTH-secreting Cushing’s disease, failure of high dose dexamethasone cortisol suppression test and severe hypokalemia. Additionally, the patient refused a surgical approach without complete confirmation of the origin of Cushing’s disease. Therefore, we performed IPSS to differentiate between a pituitary ACTH-secreting tumor and an ectopic ACTH-secreting tumor concomitant with non-secreting pituitary macroadenoma [26]. It was performed 25 days after the patient hospital discharge. Prior to IPSS, she did not report headaches and her BP was 127/90 mmHg. IPSS results are shown in Fig. 1D. Baseline ACTH level on both sides was significantly elevated compared to peripheral ACTH, and the pituitary-to-peripheral ratio of ACTH was 5.6 on the right, 38.4 on the left (the ratio greater than 2 results in diagnosis of Cushing’s disease of pituitary origin [6, 7]). The ACTH level on the left side of the pituitary was significantly elevated at 4,605 pg/mL. After 100 μg of intravenous ovine CRH injection, ACTH on the right side peaked after 5 minutes at 2,486 pg/mL. In contrast, on the left side, ACTH levels kept rising to 19,290 pg/mL until the end of the study (+20 min after CRH administration) (Fig. 1D). These results confirmed that the pituitary adenoma was the source of elevation of ACTH.

3–4 hours after the IPSS procedure, the patient started to experience headaches, blurry/double vision and developed left eye ptosis with ophthalmoplegia (Fig. 1C), which had worsened by the following day. Due to headache and ophthalmoplegia, cavernous sinus thrombosis, which is rare but known vascular complication of IPSS [27], was suspected. Brain MRI/MRA showed no evidence of cavernous sinus thrombosis but found slight expansion of pituitary adenoma to the left cavernous sinus (Fig. 1B, red arrowhead) and the tumor lesion appeared more hypo enhanced. Given her clinical symptoms and imaging results, pituitary apoplexy was suspected, and she underwent urgent endoscopic trans-sphenoidal resection of the pituitary tumor. Intraoperatively, hemorrhagic foci were found in the tumor. The tumor was successfully removed without complications (Fig. 1D). Pathological analysis revealed fragments of mixture of a necrotic and hemorrhagic pituitary adenoma (Fig. 1F and H) and viable tumor without hemorrhagic finding (Fig. 1G). Immunohistochemistry analysis showed cells strongly positive for ACTH (Fig. 1I), indicating that the tumor was a corticotroph adenoma. In contrast, staining for KI-67 and P53 exhibited low labeling index (Fig. 1J and K), indicating that the tumor did not possess aggressive histological features. Postoperatively the cortisol level significantly reduced. Hydrocortisone treatment was started and slowly tapered to a low daily maintenance dose. Mild secondary hypothyroidism, secondary hypogonadism and mild growth hormone deficiency were managed with supplemental therapies. Repeated MRA showed vertebral artery dissections had resolved. Extraocular dysfunction fully recovered after several months and currently there are no signs of recurrence of Cushing’s disease.

Results

We performed retrospective measurements of pituitary apoplexy marker levels in the stored blood from IPSS. We interpreted that the tumor apoplexy was localized to the left side based on the following reasons; left-side specific acute ophthalmoplegia, histological findings; there are large hemorrhagic areas and areas with tumor cells without hemorrhage, and subtle tumor expansion adjunct to the left cavernous sinus by MRI. From the apoplexy patient (Patient#1), we present data from left (apoplexy) side (Apo-left, red line in Fig. 2) and right (non-apoplexy) side (Apo-right, blue line in Fig. 2). For comparison, we also measured the marker levels from two cases of pituitary ACTH-secreting microadenoma patients and a case with ectopic ACTH-secreting tumor in the same assay. The two ACTH-secreting microadenomas exhibited different tumor sizes; One is a micro lesion undetectable by MRI (Patient#2), and the other showed 8 mm in size but the lesion was not clearly demarcated by MRI (Patient#3). From the Patients #2 and #3, we used blood samples only from the tumor side (Fig. 2, Micro Pt#2, green line and Micro Pt#3, purple line), which were confirmed by IPSS results (Supplemental Fig. 1). In Patient #4, IPSS samples from both sides failed to increase the pituitary-to-peripheral ratio, which ruled out Cushing’s disease of the pituitary origin. Patient #4 was therefore diagnosed with ectopic Cushing’s syndrome. We present left side sample data of this patient because marker levels from both sides were very similar (Fig. 2, Ectopic Pt#4, yellow line) (Supplemental Fig. 1).

Figure 2. Pituitary apoplexy markers in blood samples obtained during the IPSS procedure.

Graphs showing levels of IL-6 (A), MMP9 (B), TNF-α (C), MMP2 (D) and HIF1-α (E). Time point 0 refers to samples collected before CRH stimulation. Samples are color-coded; Apo-left Pt#1 (red), Apo-right Pt#1 (blue), Micro Pt#2 (green), Micro Pt#3 (purple), and Ectopic Pt#4 (orange). Intra and inter assay variations are shown in Supplemental Figure 2. (F-H) Immunofluorescence of apoplexy Cushing’s disease tissue. G and H are enlarged images of the squares in F. Arrowheads in G and an arrow in H point to round-shape cells with lobulated nuclei and mono-nuclear spherical cell, respectively. Scale bar, 20 μm in F and 10 μm in G and H.

Given that blood samples were collected and stored during IPSS, we were able to obtain information from the pituitary apoplexy patient in a real-time monitoring manner. Serum levels of the following markers were measured; HIF1-α, TNF-α, IL-6, MMP2, MMP9 and VEGF. HIF1-α was measured by ELISA and all the other markers were measured by a multiplex immunoassay [28]. For the IPSS procedure, blood samples were collected prior to CRH stimulation (referred to as 0 minutes) and at 5, 10, 15 and 20 minutes after CRH injection according to the clinical protocol in our facility.

IL-6 was detectable in all samples as baseline (0 minute) (Fig. 2A). Its levels in the hemorrhagic side (Apo-left) increased to 1.9 times after 20 minutes post CRH injection, compared to the sample collected before CRH injection (Fig 2A). In contrast, samples from the non-hemorrhagic side (Apo-right) in the same patient did not show any significant changes. In all other samples, there were no significant changes during the time-course after CRH injection. MMP9 was detected in all samples (0 minute) (Fig. 2B). MMP9 levels in the hemorrhagic apoplexy side (Apo-left) samples showed a consistent increase and reached 2.2-fold 20 minutes post CRH injection, as compared to no stimulation (Fig 2B). In contrast, samples from the non-hemorrhagic side (Apo-right) in the same patient did not show any significant changes. There was no significant change of the MMP9 level in other samples. MMP2 and TNF-α were detected in the entire series of samples (Fig. 2C, D). There was no significant change in the levels of either at any time point in both the hemorrhagic and non-hemorrhagic sides in patient #1. Similarly, there was no significant change in both MMP2 and TNF-α levels in any other sample. HIF1-α, whose serum levels have been used as a marker in clinical studies [22–25], was detectable only in two patients (Micro Pt#3 and Ectopic Pt#4) (Fig. 2E). HIF1-α was undetectable in all the rest of the samples. For the detectable samples, there was no significant change in serum levels of HIF1-α from baseline after CRH injection (Fig. 2E). VEGF-A was undetectable in all samples (data not shown).

The tumor tissue was also immunostained with IL-6 and MMP9 (Fig. 2F–H). Both IL-6 and MMP9 were detected in the resected tumor tissue. We detected IL-6 broadly in the tumor tissue, while MMP9 was detected in a focal manner. The morphologies of MMP9-stained cells appeared in 2 types. The majority of the stained cells exhibited round shape with lobulated nuclei (arrowheads, Fig. 2G), suggesting that MMP9 is expressed in macrophages or neutrophils. Some MMP9-positive cells exhibited mono nuclear and spherical shapes, suggesting that these are pituitary cells (arrow, Fig. 2H). In summary, only levels of IL-6 and MMP9, both of which were detected in the tumor tissue, increased on the apoplexy side of the pituitary apoplexy patient after CRH injection during IPSS.

Discussion

Here we report a rare case of Cushing’s disease complicated by pituitary apoplexy presumably induced by CRH injection as a part of IPSS procedures. CRH-induced pituitary apoplexy is extremely rare. Only 2 cases of pituitary apoplexy during a peripheral CRH stimulation test have been previously reported [5, 29]. In these reports, only routine serial measurement of ACTH levels in peripheral blood samples were performed, but levels of the apoplexy cytokine markers were not assessed [5, 29]. To the best of our knowledge, this is the first report of CRH-induced pituitary apoplexy during IPSS and the first report of the associated IPSS cytokine levels in the inferior petrosal sinus blood. The levels of IL-6, MMP9 and ACTH increased only in the pituitary apoplexy side of the Cushing’s disease patient. None of the other markers we measured changed in 2 other Cushing’s disease cases and 1 case of ectopic Cushing’s syndrome. The similar pattern of changes in levels of ACTH, IL-6 and MMP9 in the pituitary apoplexy sample suggests that these changes may be linked to CRH stimulation. CRH stimulates expression of POMC, which encodes the polypeptide precursor for ACTH, via CRH receptor 1 in corticotrophs [30, 31]. CRH also induces IL-6 expression in adrenal glands during acute inflammation and elevates IL-6 levels in the serum [32]. IL-6 induces ACTH secretion independent of CRH in CRH receptor knockout mice [33]. These reports suggest that both CRH and CRH-induced IL-6 lead to increased ACTH levels. By contrast, there is no report about direct associations of CRH and MMP9. However, it has been reported that IL-6 induces rapid elevation of MMP9 levels in macrophages [34], suggesting that CRH-induced IL-6 may have caused increased MMP9 levels by macrophages. In addition, mast cells, which release MMP9 in pathological conditions [35], may have already been resident in the tumor, as suggested by a recent single cell RNA-Seq study of pituitary tumors [36]. Taken together, elevation of levels of ACTH, IL-6 and MMP9 may be a coordinated response to CRH injection in the apoplexy patient.

Mechanisms of pituitary apoplexy and contributions of MMP9 in pituitary apoplexy are totally unknown. Given that MMP9 has a collagenase activity, we speculate that rapid elevation of MMP9 affected vascular integrity in the pituitary tumor, leading to hemorrhage in the pituitary gland. The size of the pituitary tumor itself may be an additive factor for triggering apoplexy, since a large tumor may accelerate the disruption of the blood supply to the tumor. Of note, MMP9 was also detected in cells whose morphology seems to be that of the pituitary cells in our immunofluorescence analysis, which is consistent with a previous report that showed MMP9 expression in pituitary tumor cells [16, 37].

We recognize several limitations in our report. In the case described here, the levels of MMP2, TNF-α and HIF-1 α did not show significant changes after CRH stimulation in the apoplexy patient. The lack of changes suggests that these markers do not exhibit acute changes during the onset of apoplexy; however, it is also possible that levels of these markers may change in later stages of apoplexy since our measurements were limited to 20 minutes after CRH injection. Another limitation is that our analysis was limited to CRH-induced apoplexy in one Cushing’s disease case, and it does not necessarily apply to all apoplexies in Cushing’s disease, nor to apoplexies in any other type of pituitary tumor.

We are aware that our results and a report by Paorette et.al [38] contradict. Specifically, we show that the IL-6 level increased only in the apoplexy side. Additionally, we and Watanobe et al showed that serum IL-6 level did not increase in the non-apoplexy Cushing’s tumor side [39]. In contrast, Paorette et.al reported that IL-6 levels increased in non-apoplexy Cushing’s tumor side [38]. Assay sensitivity is likely not the cause of the discrepancy, because the assay sensitivity in Paorette et. al and our study are similar according to the manufacturer’s data sheet. A difference in these two studies is found in the approach to processing/presenting results. Paorette et. al showed the average value of 11 patients, while we show an individual patient’s data. Due to the difference in data processing, we cannot directly compare our results to results of Paorette et.al. It has been shown that IL-6 levels can be affected by other factors, such as age, circadian rhythm and BMI, which might also have contributed to the difference in the two studies [40].

Currently there are no biomarkers to diagnose pituitary apoplexy and diagnostic modalities are limited for pituitary apoplexy [2]. Since pituitary apoplexy is rare and there is no available in vivo model, information for pituitary apoplexy physiology is also limited. Our observation of acute elevations of IL-6 and MMP9 may be helpful to increase understanding of the pathophysiology of pituitary apoplexy and might be useful information that contributes to the development of diagnostic biomarkers of pituitary apoplexy in the future.

Materials and Methods

ELISA

HIF1-α enzyme-linked immunosorbent assay (ELISA) was performed according to the manufactures’ instruction (RayBiotech, #ELH-HIF1a). The absorbance is measured on the microtiter plate reader (EPOCH, BioTek Instruments, Inc, Gen5 2.9 Soft). Samples were tested in duplicate and values were interpolated from log-log fitted standard curves. Intra and inter assay variations for HIF1-α assays are shown in Supplemental Table 1.

Multiplex immunoassay

All samples were consistently placed on ice immediately after sampling and stored in −80 °C immediately after measurement of ACTH. All samples were thawed once and markers were measured simultaneously. Intra assay variations of each sample are included in the supplemental material. Samples were analyzed using antibodies specific to human IL-6 (R&D Systems, cat#LHSCM206), TNFα (R&D Systems, cat#LHSCM210), VEGF-A(R&D Systems, cat#LHSCM293), IL-1ß (R&D Systems, cat#LHSCM201), MMP2 (R&D Systems, cat#LMPM902) and MMP9 (R&D Systems, cat#LMPM911). We used the Luminex platform and performed high sensitivity multi-plex assays. The magnetic bead set (R&D Systems, cat. # LHSCM000) was purchased from R&D Systems Inc. Samples were assayed according to the manufacturer’s instructions. In brief, fluorescent color-coded beads, coated with specific capture antibodies, were added to each sample. After incubation and washing, biotinylated detection antibodies (R&D Systems, cat# LHSCM000) were added, followed by incubation with phycoerythrin-conjugated streptavidin (R&D Systems, cat# LMPM 000). The fluorescent signals from the beads were detected on a Luminex dual-laser fluidics-based instrument (Bioplex 200). One laser determines the analyte being detected via the color coding; the other measures the magnitude of the PE signal from the detection antibody, which is proportional to the amount of analyte bound to the bead. Samples were run in duplicate and values were interpolated from 5-parameter fitted standard curves. Intra and inter assay variations for these cytokine assays are shown in Supplemental Table 1.

Immunohistochemical Staining

Human corticotroph adenomas were fixed in 10% formalin and embedded in paraffin. After deparaffinization and antigen retrieval, slides were blocked with 5% donkey serum in PBS+0.1% Triton X-100 for 60 min at room temperature, and stained and rocked overnight with goat polyclonal anti-MMP9 (1:20 dilution; AF911, R&D Systems Inc.) and mouse monoclonal anti-IL-6 (1:10 dilution; MAB206, R&D Systems Inc.). After washing with PBS+0.1% Triton X-100, sections were incubated with secondary antibodies at 4°C overnight: Alexa Fluor 488-donkey anti-goat IgG (Invitrogen A32814, 1:500 dilution) and Alexa Fluor 594-donkey anti-mouse IgG (Invitrogen, A32754, 1:500 dilution). The slides were washed, and mounted using DAPI Fluoromount-G (SouthernBiotech, 0100-20). Images were acquired using a Zeiss LSM710 confocal microscope with Zen software.