2018 Volume 65 Issue 10 Pages 1061-1067
A 43-year-old woman with an 8-year history of diabetes, hypertension, and dyslipidemia presented with amenorrhea and convulsion. Her MRI scan revealed a 3.5-cm T2-hyperintense pituitary macroadenoma with suprasellar extension to the frontal lobe and bilateral cavernous sinus invasion. Her serum levels of GH and insulin-like growth factor-I (IGF-I) were elevated to 9.08 ng/mL (normal range: <2.1 ng/mL) and 1,000 ng/mL (normal range: 90–233 ng/mL, SD score +10.6), respectively. Bromocriptine insufficiently suppressed her GH levels, while octreotide paradoxically increased her GH levels. Together with her characteristic features, she was diagnosed with acromegaly caused by an invasive GH-producing pituitary macroadenoma. As performing a one-stage operation would have been extremely difficult, she was first treated with pasireotide long-acting release (40 mg monthly) for 5 months followed by a successful transsphenoidal surgery. One month after the first injection, biochemical control was achieved (IGF-I, 220 ng/mL; GH, 1.26 ng/mL), and tumor shrinkage of approximately 50% was observed. The resected tumor was histologically diagnosed as a sparsely granulated somatotroph adenoma, with higher expression of somatostatin receptor subtype 5 (SSTR5) than that of SSTR2A. The germline aryl hydrocarbon receptor interacting protein (AIP) mutation was negative, and several tumor cells were weakly immunoreactive for AIP. Despite the presence of a residual tumor postoperatively, biochemical control was achieved 6 months after the final injection of pasireotide. In conclusion, this case suggests that pasireotide may be an option for preoperative first-line therapy in invasive and octreotide-resistant sparsely granulated somatotroph adenomas.
ALTHOUGH the gold standard therapy for acromegaly is transsphenoidal surgery, in cases of invasive and aggressive adenomas, preoperative medical therapy, especially using first-generation somatostatin analogues (SSAs) such as octreotide or lanreotide, can be considered as a first-choice treatment [1, 2]. SSAs often achieve tumor shrinkage, decreased levels of GH and insulin-like growth factor-I (IGF-I) by pretreatment, and improved surgical cure rates in GH-secreting pituitary macroadenomas [3].
Among the 5 subtypes of somatostatin receptors (SSTRs), octreotide has a high affinity for SSTR2A, while its affinity is lower for SSTR5. On the other hand, sparsely granulated somatotroph adenoma is characterized by younger age of the patient at diagnosis, high invasiveness, high GH levels, low SSTR2A expression, and treatment resistance to first generation SSAs [4, 5]. To date, the appropriate treatment strategy for sparsely granulated somatotroph adenoma has not been established. We present here a case of inoperable aggressive acromegaly caused by sparsely granulated somatotroph adenoma with resistance to octreotide that achieved tumor shrinkage and biochemical control, and finally was able to undergo successful surgery by preoperative therapy with pasireotide, a new class of second-generation somatostatin analogue.
A 43-year-old woman experienced generalized convulsion and was taken to the hospital in an ambulance. She had a history of amenorrhea for 18 years; diabetes, hypertension, and dyslipidemia for 8 years; and complex partial seizures for 3 months. She had a suprasellar tumor detected on MRI and was referred to our hospital for closer endocrine examination. There was no family history of pituitary tumors or gigantism, however there was a family history of diabetes in her mother. She continued to grow in height after age 20, reaching 166 cm (SD score +1.45), suggesting that onset of her GH-producing pituitary tumor preceded epiphyseal closure. On admission, the patient’s blood pressure was 127/87 mmHg. She had typical acromegalic features such as macroglossia, thick lips, and thickening of her hands and feet. Visual field examination was normal. She had no parathyroid or pancreatic diseases.
Her diabetes was fairly well controlled with sitagliptin (50 mg), dipeptidyl peptidase-4 inhibitor, and luseogliflozin (2.5 mg), sodium-glucose cotransporter 2 inhibitor, each administered once daily (fasting plasma glucose, 104 mg/dL; HbA1c, 6.2%). Her serum levels of GH and IGF-I were elevated to 9.08 ng/mL (normal range: <2.1 ng/mL) and 1,000 ng/mL (normal range: 90–233 ng/mL, SD score +10.6), respectively. GH showed no paradoxical response to the TRH (0.5 mg i.v.) and LHRH (0.1 mg i.v.) tests. Bromocriptine (2.5 mg p.o.) administration did not suppress the GH levels sufficiently, and acute octreotide challenge (50 μg s.c.) paradoxically raised the GH levels from 5.00 to 8.29 ng/mL (Table 1). Her PRL level was slightly increased (31.63 ng/mL), whereas her TSH (0.70 μU/mL) and normal free T4 (1.03 ng/dL) levels were reduced, and her gonadotropin level was undetectable. An MRI scan revealed a 3.5-cm pituitary adenoma with higher intensity signal in the T2-weighted image compared with that of the gray matter (Fig. 1g) and with suprasellar extension to the frontal lobe and bilateral cavernous sinus invasion (Fig. 1).
| TRH test (0.5 mg) | |||||
| Time (min) | 0 | 30 | 60 | 90 | 120 |
| GH (ng/mL) | 5.88 | 5.87 | 5.96 | 5.98 | 6.31 |
| PRL (ng/mL) | 33.39 | 35.45 | 34.66 | 33.42 | 34.95 |
| LHRH test (0.1 mg) | ||||||
| Time (min) | 0 | 15 | 30 | 60 | 90 | 120 |
| GH (ng/mL) | 2.34 | 2.05 | 1.77 | 1.58 | 1.38 | 1.43 |
| Octreotide challenge test (50 μg) | |||||
| Time (h) | 0 | 2 | 4 | 6 | 8 |
| GH (ng/mL) | 5.00 | 4.82 | 5.06 | 6.22 | 8.29 |
| PRL (ng/mL) | 25.59 | 25.59 | 24.93 | 24.65 | 26.30 |
| Bromocriptine challenge test (2.5 mg) | |||||
| Time (h) | 0 | 2 | 4 | 6 | 8 |
| GH (ng/mL) | 6.89 | 6.37 | 6.03 | 5.28 | 4.42 |
| PRL (ng/mL) | 34.62 | 27.60 | 13.97 | 13.06 | 10.21 |
Diagnostic imaging and histopathology of the resected somatotroph adenoma. Coronal contrast-enhanced T1-weighted (a, b), sagittal contrast-enhanced T1-weighted (c, d), sagittal T2-weighted (e, f), coronal T2-weighted (g, h), MRI scans before (a, c, e, g) and 5 months after (b, d, f, h) pasireotide therapy.
Based on her physical, endocrine, and imaging data, she was diagnosed with acromegaly caused by an invasive GH-producing pituitary macroadenoma. Specifically, we speculated her pituitary adenoma to be a sparsely granulated somatotroph adenoma because of its clinical features such as resistance to the first-generation SSA octreotide, invasive nature, and hyperintensity on T2-weighted MRI images, along with her young age.
We considered that the rapidly growing adenoma may increase the risk for impending convulsion. However, it seemed challenging to completely remove this pituitary tumor surgically at once because of the invasive suprasellar extension of the pituitary adenoma. Therefore, we chose to perform preoperative treatment with a monthly dose of 40 mg pasireotide, a new class of second-generation somatostatin analogue (Fig. 2). One month after the first pasireotide injection, her GH and IGF-I levels dramatically decreased to within the normal ranges (1.26 ng/mL and 220 ng/mL, respectively), and were maintained thereafter for 5 months (Fig. 2). Consequently, her acromegalic features recognizably diminished (Fig. 3), and her convulsions disappeared. Follow-up MRI showed dramatic tumor shrinkage of approximately 50%, which plateaued after three months of treatment. The T2-hyperintensity feature of the tumor diminished after 5 months of treatment (Fig. 1). Before pasireotide treatment, the patient’s HbA1c level was 6.3% and her CPR level was 3.93 ng/mL. Her diabetes was being treated with the oral hypoglycemic agents sitagliptin (50 mg), a dipeptidyl peptidase-4 inhibitor, and luseogliflozin (2.5 mg), a sodium-glucose cotransporter 2 inhibitor, each administered once daily. Two weeks before starting pasireotide therapy, her diabetes was treated by mealtime dosing of a rapid-acting insulin analog, insulin aspart. Two-month treatment with pasireotide increased HbA1c to 7.4%, and we added bedtime dosing of the long-acting insulin analog, insulin degludec. Five months after pasireotide treatment, her HbA1c level was decreased to 6.3% with a CPR level 0.95 ng/mL. Pasireotide treatment initially increased her insulin requirement, however this effect decreased over time along with normalization of her IGF-I levels. Finally, 2 months after TSS, and after pasireotide washed out, she no longer required insulin to keep HbA1c levels at 6.3% with diet therapy alone (Fig. 2). Reduction in her C-peptide levels may reflect impaired secretion and ameliorated insulin resistance mediated by pasireotide therapy.
Clinical course
LAR, long-acting release; IGF-I, insulin-like growth factor-I; TSS, transsphenoidal surgery; OGTT, oral glucose tolerance test; OHA, oral hypoglycemic agents.
The fingers and ring of the patient before and after 5 months of pasireotide treatment.
The fingers of the patient demonstrating characteristic acromegalic features (a) before treatment (no space between the ring and finger), significantly diminished (b) after 5 months of pasireotide treatment (a space is present between the ring and finger).
After 5 months of treatment with pasireotide, the patient underwent extended transsphenoidal surgery of the pituitary tumor. Almost the entire tumor was successfully removed except for a vascular invasive lesion. Postoperative MRI scan revealed residual tumor in the upper cavernous sinus and lower frontal lobe. Histological examination revealed that this tumor was composed of pale eosinophilic cells with nuclear pleomorphism (Fig. 4a). The cells were weakly and focally immunopositive for GH (Fig. 4b), and partly positive for PRL (Fig. 4c). CAM5.2 staining showed prominent fibrous bodies (Fig. 4d). Immunohistochemistry for SSTR2A showed negative or only cytoplasmic immunoreactivity (score 0–1 according scoring system proposed by Volante et al. [6]). On the other hand, SSTR5 staining showed circumferential membranous staining in more than 50% of tumor cells (score 3 in the same scoring system) (Fig. 4e, f). Ki-67 index was 0.8% (Fig. 4g), and E-cadherin expression was decreased (Fig. 4h). Several tumor cells were weakly immunoreactive for AIP (Fig. 4i). We analyzed AIP gene using genomic DNA extracted from peripheral blood cells of the patient under approval of the related ethical committees (Kobe University No. 1363, Kanazawa University No. 496) as previously described [7] and there were no mutations.
Immunohistochemical staining of the somatotroph adenoma for HE (a), GH (b), PRL (c), anti-cytokeratin (CAM-5.2) (d), somatostatin receptor 2A (SSTR2A) (e), somatostatin receptor 5 (SSTR5) (f), Ki-67 (g), E-cadherin (h), and aryl hydrocarbon receptor-interacting protein (AIP) (i). (original magnification of all immunohistochemical stained images, ×400)
The following primary antibodies were used: GH (DAKO A0570), PRL (DAKO A0569), CAM-5.2 (Becton Dickinson 349205), SSTR2A (abcam ab134152), SSTR5 (abcam ab109495), E-cadherin (DAKO M3612-01), and AIP (Novus Biologicals NB100-127).
Two months after the operation, a 75-g oral glucose tolerance test showed normal glucose tolerance, with a nadir GH level of 0.28 ng/mL. She needed no medication for her hypertension, dyslipidemia, and diabetes. At 6 months after the final injection of pasireotide, biochemical control was still achieved.
We have described in detail the clinical and morphological courses taken during the preoperative pasireotide therapy in a patient with octreotide-resistant acromegaly caused by a histologically confirmed SSTR5-dominant sparsely granulated somatotroph adenoma.
In a head-to-head superiority study comparing pasireotide and octreotide in tne treatment of acromegaly, pasireotide was effective in cases resistant to octreotide [8]. As a small part of this study, seven patients with naïve active acromegaly were randomized to presurgical treatments with either pasireotide long acting release (LAR) or octreotide LAR [9]. In the 4 patients switched from octreotide to pasireotide therapy after the initial study period, GH and IGF-1 levels were further reduced in 2 patients, and tumor size was reduced in 3 patients. However, histology of tumors and the relationship between pasireotide responsiveness and the clinical features of each patient were not examined. Therefore, the clinical effects of acromegaly-associated features on pasireotide-associated glucose intolerance remain unclear.
The GH-producing pituitary adenoma in the present case was characterized as invasive, large, with negative octreotide acute test, and T2-weighted MRI hyperintensity, all of which predict resistance to first-generation SSAs [10, 11]. Based on these findings, we diagnosed the present case with a suspected GH-producing, sparsely granulated adenoma. To date, satisfactory treatments remain to be established for acromegaly patients resistant to first-generation SSAs, which in turn is associated with the histological expression patterns of the SSTR subtypes [4]. There are 5 subtypes of SSTRs expressed in the tumor that contribute to the response to SSAs [12]. It is reported that sparsely granulated somatotroph adenomas lacking SSTR2 with a low SSTR2/SSTR5 ratio are associated with resistance to first-generation SSAs [10]. Sparsely granulated adenomas are characterized by higher SSTR5 positivity along with low or no SSTR2A positivity, whereas densely granulated somatotroph adenomas dominantly express SSTR2A [13]. Recent investigations have demonstrated that the membranous expression of SSTR2A and SSTR5 is a prerequisite for responsiveness to first generation SSAs and pasireotide treatment, respectively [14]. In addition, these studies have suggested that a higher SSTR5 staining score is the only indicator of the biochemical response to pasireotide. On the other hand, tumors with AIP deficiency express SSTR2A at a lower score (score 0–1) and no difference was observed in SSTR5. These observations might be associated with resistance to first generation SSAs and responsiveness to pasireotide in AIP deficient tumors. As for adenoma subtype, sparsely granulated adenoma had a tendency for resistance to first generation SSAs and showed better response to pasireotide than densely granulated adenoma. In line with a current study [4], sparsely granulated adenomas had a lower SSTR2A expression and higher SSTR5 expression compared to densely granulated adenomas.
Pasireotide, a new class of second-generation somatostatin analogue, binds to SSTR1, 2, 3, and 5, and has a 39x greater binding affinity for SSTR5 but a 0.4x affinity for SSTR2 compared to octreotide [15]. The adenoma in our patient was a sparsely granulated adenoma which dominantly stained with SSTR5 rather than SSTR2A despite being stained after using pasireotide, which binds to the former analog at much higher affinity. These findings may explain the molecular basis underlying octreotide resistance and response to pasireotide in the present case. In this case, although AIP mutation was negative, the decreased expression of E-cadherin and AIP [16] in the tumor accompanied with the feature of sparsely granulated adenoma clearly indicated the predictive characteristics of resistance to first generation SSAs.
Interestingly, pasireotide-mediated shrinkage of the pituitary adenoma was accompanied by an intra-tumoral T2-hyperintense signal in our case, suggestive of intra-tumor hemorrhage or necrosis. Generally, SSAs shrink tumor cells by reducing secretory granules, rather than by direct cytotoxicity [17]. Accumulation of similar cases will clarify the pasireotide-induced alterations in tumor histology.
Hyperglycemia associated with pasireotide is also a concern in 53.7–67.0% of the treated patients [8, 18]. Pasireotide binds to SSTR5 expressed not only in pituitary cells but also in pancreatic beta cells, thereby reducing insulin secretion [19]. Pasireotide-mediated reduction in insulin secretion is accompanied by a decrease in incretins. On the other hand, improving biochemical control of acromegaly rescues insulin sensitivity and glucose tolerance [19]. Recent studies have shown that glucose levels peak after a few months and remain stable or slightly reduced during pasireotide therapy [19, 20]. Similarly, in the present case, the HbA1c levels peaked at 2 months after pasireotide injection, followed by a reduction in the insulin doses and a gradual reduction in HbA1c to 6.3%, along with normalization of IGF-I (Fig. 2). These findings suggest that the benefits of reduced insulin resistance overcome the deleterious suppression of insulin and incretin secretion during the course of pasireotide treatment. Therefore, under careful glucose monitoring and anti-diabetic therapy, we can adequately control pasireotide-induced hyperglycemia.
In conclusion, this case suggests that pasireotide may be an option for preoperative first-line therapy in invasive and octreotide-resistant sparsely granulated somatotroph adenomas.
