Abstract
The expression of steroid receptors by tumours offers a therapeutic advantage if functionally responsive to exogenous hormones. Insulinomas represent a highly symptomatic group of pancreatic tumours and the steroid receptor status of these tumours is poorly understood. The object of the study was to characterise the sex steroid receptor status of human insulinomas and to investigate whether sex steroids alter insulin expression therein. At our tertiary referral University Hospital, archival and prospective tissues from 25 insulinoma patients collected over 14 years were analysed for oestrogen receptor-α (ERα), oestrogen receptor β (ERβ) and progesterone receptor (PR) expression. Tissue explants of insulinoma and control pancreatic tissue from two new insulinoma patients were cultured and treated with oestrogen and progesterone and insulin expression measured by RT-PCR and ELISA. The main outcome measures were established before data collection and included sex steroid receptor status of tumours and insulin expression in fresh tissue in response to exogenous sex steroids. PR was expressed in 24 out of 25, ERα in 10 out of 25 and ERβ in 21 out of 25 insulinomas. In fresh insulinoma cultures, insulin expression was increased by oestrogen or progesterone, whereas no significant effect was observed in adjacent pancreatic tissue. This study demonstrates widespread expression of sex steroid receptors on human insulinoma tissue and provides in vitro evidence of functionality with increased expression of insulin by insulinoma explants in response to exogenous oestrogen or progesterone. Confirmation of these results may provide a therapeutic mechanism for reducing symptomatic insulin secretion by receptor blockade.
Introduction
Hormonal manipulation using sex steroid hormones has an established role as therapy for tumours of the breast and female genital tract. The hormones exert their effect by interacting with specific receptor proteins that must be expressed by cells for them to be responsive. Progesterone receptor (PR) immunoreactivity has been demonstrated in normal pancreas (Targarona et al. 1991) and specifically co-localised to normal insulin-producing β cells in the islets of Langerhans (Doglioni et al. 1990). There have been a limited number of studies characterising the steroid hormone receptor profile of pancreatic tumours. PRs have been detected immunohistochemically in normal pancreas and pancreatic tumours, almost exclusively being confined to endocrine pancreas tissue (Doglioni et al. 1990, Viale et al. 1992). A decade ago, oestrogen receptor (ER) presence was demonstrated in pancreatic tumour tissue using radioligand binding assays and Scatchard analysis (Greenway et al. 1981). However, the more recent studies have reported no detection of ER in either normal or pathological pancreatic tissue (Doglioni et al. 1990, Targarona et al. 1991, Viale et al. 1992, Singh et al. 1997).
In terms of translational significance, trials of anti-oestrogenic therapy in ductal pancreatic cancer have had disappointing results. Some studies have shown beneficial effect of tamoxifen (Wong & Chan 1993, Horimi et al. 1996) but the definitive randomised trials of tamoxifen in pancreatic cancer have been unanimously disappointing (Keating et al. 1989, Bakkevold et al. 1990, Taylor et al. 1993). There have been no clinical trials of anti-progestogenic therapy in pancreatic cancer.
Insulinomas are rare neuroendocrine tumours of the pancreas affecting 1–2 per million per year in the UK. Patients usually present with symptoms of hypoglycaemia secondary to insulin hypersecretion. These symptoms may sometimes be episodic and non-specific, and thus misleadingly prompt a barrage of neurological or psychiatric investigations. The tumours show a female preponderance with a female to male ratio of 4:1 (Service et al. 1991). Most lesions are small, solitary and benign with only 10% showing size ≥2 cm, multiplicity or malignancy. There is a good long-term survival of 88% at 10 years following surgical excision (Service et al. 1991).
We present an analysis of the oestrogen and PR status of 25 insulinoma cases. Finally, we demonstrate the effect of exogenous oestrogen and progesterone on insulin mRNA and protein expression in freshly isolated human insulinoma tissue.
Materials and methods
Tissue source
A total of 25 cases of pancreatic insulinoma were analysed in our series, and 23 cases of pancreatic insulinoma were retrieved from the files of the Department of Pathology, Queen Elizabeth Hospital, Birmingham, UK. Clinical details were reviewed from the files of medical records. In addition, we prospectively collected freshly explanted pancreatic tissue from two new cases of insulinoma. These studies were performed with appropriate ethical approval form the South Birmingham Regional Ethics Committee.
Immunohistochemistry
Light microscopic findings were reviewed and immunohistochemical staining was repeated where necessary. Formalin fixed, paraffin-embedded tissues were cut at 4 μm on to adhesive-coated slides. After dewaxing and rehydration of sections, all slides were treated for 15 min in 3% hydrogen peroxide in water to block endogenous peroxidase activity. Antigens were retrieved by incubation overnight in EDTA buffer on a hot-plate stirrer at 65 °C as previously described (Reynolds et al. 2002). Using a ‘Sequenzer’ for automation, slides were incubated for 1 h with the following primary antibodies diluted in TBS/Tween (pH 7.6):
anti-gastrin antibody (Novo-Castra, GASp)
anti-PP (Novo-Castra, NCL-PPp)
anti-insulin (Novo-Castra, insulin)
anti-ERα mAb (Dako, 1D5)
anti-ERβ mAb (Labvision, RB-10658-P)
anti-PR mAb (Dako, PR636)
anti-somatostatin (Dako, A0566)
anti-glucagon (Dako, A0565)
anti-chromogranin (Dako, A0430)
anti-synaptophysin (Novo-Castra, Synap-299)
anti-Ki-67 (Dako, M7240)
(Novo Castra, Newcastle upon Tyne, UK; Dako, Ely, Cambridgeshire, UK; Labvision, Roncorn, Cheshire, UK)
Following a wash in TBS/Tween, primary antibodies were visualised using Dako ChemMate EnVision detection kit (K5007, Dako), counterstained in haematoxylin, dehydrated, cleared and mounted.
Immunoperoxidase-stained slides were reviewed independently. Each slide for gastrin, pancreatic polypeptide, insulin, somatostatin, glucagons, chromogranin, synaptophysin, Ki67, ER and PR was coded for two parameters of staining: intensity (strong, detectable at ×40; weak, only detectable at greater than ×100) and distribution (diffuse, >50% of tumour cells; focal, <50% of tumour cells). The labelling index of Ki-67 was evaluated by scoring a minimum of 1000 tumour cells in randomly selected fields. For each case, three different counts were performed and the highest score was chosen as the corresponding index value.
In vitro culture insulinoma tissue and insulin secretion
Pancreatic insulinoma tumour explants surplus to requirements were trimmed from two resection specimens. For each culture condition, equal 5 mm3 blocks of insulinoma tissue were finely diced and incubated in 2 ml phenol red-free RPMI 1640 medium containing 5% foetal calf serum. In each insulinoma case, macroscopically uninvolved pancreatic tissue was obtained from the resection margins, distant from the insulinoma tumour, under the supervision of the reporting histopathologist (Phillipe Taniere). Equal 5 mm3 blocks of this tissue were processed in exactly the same way as tumour tissue and served as patient-specific control pancreatic tissue.
Tumour and ‘control’ tissue were incubated for 24 h in four culture conditions namely: phenol red-free RPMI 1640 medium (Invitrogen) alone, medium plus 10−7 M E2 (17β-oestradiol; E2257, Sigma–Aldrich), medium plus 10−7 M progesterone (P6149, Sigma–Aldrich) alone, and or medium plus 10−7 M of E2 and 10−7 M progesterone. Culture was carried out in 24-well plates maintained at 37 °C in 5% CO2. After 24 h in culture, the supernatant was collected for insulin assay and the pancreatic tissue was used for RNA extraction. The supernatant was frozen at −70 °C prior to use in insulin ELISA.
Insulin ELISA
Insulin concentrations in the supernatants were measured using an immunoenzymometric assay which was specific for insulin (BioSource International, Camarillo, CA, USA) and calibrated against IRP 66/304. Fifty microlitres of each standard, control or sample were dispensed into the appropriate number of wells of the insulin antibody-coated 96-well plate. Fifty microlitres of anti-insulin-HRP conjugate were dispensed into all the wells and the plate was incubated for 30 min at room temperature on a horizontal shaker set at 700 r.p.m. The plate was washed thrice with Tween 20 diluted 1:200 in distilled water. Two hundred microlitres of freshly prepared chromogen solution (TMB mixed with H2O2 in acetate/citrate buffer) were added into each well within 15 min following the washing step. The plate was then wrapped in tin foil (to exclude direct sunlight) and incubated for 15 min at room temperature on a horizontal shaker set at 700 r.p.m. The reaction was stopped by dispensing 50 μl of 1.8 M H2SO4 into each well. The microtitre plate absorbances were read at 450 and 490 nm. The detection range of the ELISA kit was <10–3400 pmol/l.
Insulin PCR
RNA was extracted from the pancreatic tissue (obtained from patient 25 only) using the RNeasy mini kit (Qiagen). A total of 30 mg pancreatic tissue was used as starting material for each sample and tissue was disrupted and homogenised using a conventional rotor–stator homogeniser. The optional on-column DNase digestion was used and the rest of the method was as indicated in the standard manufacturer's protocol for total RNA isolation from animal tissue.
RT was carried out in RNase- and DNase-free conditions. Samples were heated at 75 °C for 5 min. A RT mix comprising RNasin RNase Inhibitor (Promega), dNTPs (Amersham), random hexamers (Amersham), 10× PCR buffer (Qiagen) and 25 mM MgCl2 was added to each sample. After 5-min incubation at 42 °C, Superscript II RNase H–reverse transcriptase (Invitrogen) was added to each sample. Incubation was carried out for 1 h at 42 °C and then for 5 min at 95 °C.
Conventional PCR for insulin was performed using a forward primer (5′-GCA-AGC-AGG-TCA-TTG-TTT-CA-3′) and reverse primer (5′-CAC-TTG-TGG-GTC-CTC-CAC-TT-3′). PCR amplification was carried out according to the following cycling programme: 5-min denaturation at 95 °C, followed by 30 cycles of: 95 °C for 30 s; 60 °C for 30 s and 72 °C for 30 s. Samples were held at 72 °C for 7 min and then at 4 °C. After PCR amplification, samples were electrophoresed in 2% agarose gel and stained with ethidium bromide; the photographic negatives were used for assessment. Under these conditions, the insulin PCR amplification product has a band size of 211 bp on the agarose gel.
Results
Clinical details
A total of 25 patients with insulinomas were studied, 19 were female and 6 were male (female to male ratio was 3.2:1; Table 1). The mean age of all patients with insulinomas was 50 years 8 months±17 years 7 months (range=17 years 4 months–83 years 7 months) and their mean follow-up time was 30.5±32.9 months (range=1.7–138.1 months). There were five deaths during the follow-up period and four of these were due to recurrent disease. Of the four patients who died of recurrent disease, two had undergone distal pancreatectomy plus concomitant partial liver resection and one underwent distal pancreatectomy plus concomitant cryotherapy for hepatic metastases. The final patient who died of recurrent disease had undergone complete hepatectomy and liver allograft transplantation. One patient (patient 18) is known to have multiple endocrine neoplasia type 1A (MEN-1A) and is currently having pituitary investigations for reported symptoms. Another patient (patient 3 in Table 1) was diagnosed with acromegaly before presenting with his insulinoma, but gene analysis showed that he had no mutations in the menin gene. Predicted malignant behaviour as defined by the presence of metastases, angioinvasion or local infiltration and/or recurrent disease (Rindi et al. 1998, Schindl et al. 2000) was seen in 6 (24%) of the insulinomas, while the remaining 19 (76%) were benign.
Clinic details of the 25 insulinoma patients in the study
Patient | Sex | Age at surgery (years) | Length of follow-up (months) | Type of surgery | Malignant |
---|---|---|---|---|---|
1 | F | 35.7 | 34.4 | Hepatectomy and orthotopic liver transplantation | Yes |
2 | F | 72.2 | 16.4 | Enucleation of insulinoma, cholecystectomy | No |
3 | M | 57.6 | 66.8 | Excision of lesion of pancreas | No |
4 | F | 32.6 | 34.7 | Excision of lesion of pancreas | No |
5 | F | 38.7 | 46.7 | Cholecystectomy, excision of lesion of pancreas, laparotomy | No |
6 | M | 24.5 | 42.8 | Enucleation of insulinoma | No |
7 | F | 63.2 | 51.5 | Excision of insulinoma from tail of pancreas | No |
8 | F | 83.6 | 6.6 | Distal pancreatectomy, splenectomy | No |
9 | F | 34.4 | 15.9 | Local enucleation of insulinoma in head of pancreas | Yes |
10 | F | 38.7 | 1.7 | Excision of insulinoma from tail of pancreas, cholecystectomy | No |
11 | F | 64.4 | 2.6 | Distal pancreatectomy, cholecystectomy | No |
12 | F | 54.1 | 34.3 | Excision of lesion of pancreas, cholecystectomy | No |
13 | F | 41.4 | 13.4 | Distal pancreatectomy, excision of accessory spleen, cholecystectomy | No |
14 | M | 31.0 | 4.2 | Excision of insulinoma in body of pancreas, excision of splenunculus in tail of pancreas | Yes |
15 | F | 51.8 | 6.9 | Enucleation of insulinoma in head of pancreas, cholecystectomy | No |
16 | M | 51.8 | 18.2 | Right extended hemihepatectomy, distal pancreatectomy, splenectomy | Yes |
17 | M | 72.2 | 13.6 | Distal pancreatectomy, L lateral segmentectomy, excision of liver lesion, splenectomy | Yes |
18 | F | 17.4 | 6.6 | Enucleation of insulinoma from pancreatic head, cholecystectomy | No |
19 | M | 67.9 | 138.1 | Distal pancreatectomy, splenectomy, cholecystectomy | Yes |
20 | F | 33.6 | 94.5 | Excision of lesion of pancreas | No |
21 | F | 40.0 | 46.0 | Excision of insulinoma, cholecystectomy | No |
22 | F | 75.6 | 2.6 | Distal pancreatectomy | No |
23 | F | 59.0 | 2.6 | Enucleation of insulinoma, cholecystectomy | No |
24 | F | 73.6 | n/a | central pancreatic resection | No |
25 | F | 53.5 | n/a | Pylorus-preserving pancreaticoduodenectomy | No |
This table shows the relevant descriptive demographics of the 25 patients in our series. Data were collected form patient records and by the review of pathological and radiological reports. The table gives the patients' sex, age at surgery, length of follow-up, type of surgery and tumour malignancy status.
Immunohistochemistry
The neuroendocrine differentiation of our series of insulinomas is confirmed immunohistochemically by the presence of protein gene product 9.5, neuron-specific enolase, chromogranin, synaptophysin and insulin (Lam et al. 1999).
There was expression of PR in 24 out of 25 tumours. ERα was expressed in 10 out of 25, while ERβ was expressed in 21 out of 25 tumours (Table 2, Fig. 1).
Immunohistochemical analysis of our series of 25 insulinomas
Patient | Gastrin | PP | Ins | Somat | Glucag | Chromo, synapto | Ki-67 (%) | PGP 9.5, NSE | ERα | ERβ | PR |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | Neg | WFP | WFP |
2 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | SFP | SDP | SDP |
3 | Neg | Neg | SDP | SDP | Neg | SDP | <3 | SDP | Neg | SDP | SDP |
4 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | SFP | SDP | SDP |
5 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | Neg | SDP | Neg |
6 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | SDP | SDP | SFP |
7 | Neg | Neg | SFP | Neg | SDP | SDP | <2 | SDP | Neg | SDP | SDP |
8 | WFP | Neg | SDP | SFP | SDP | SDP | <2 | SDP | SDP | SDP | SDP |
9 | Neg | Neg | SDP | SFP | SDP | SDP | <5 | SDP | SFP | SDP | SFP |
10 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | SFP | SDP | SDP |
11 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | WFP | SDP | SDP |
12 | Neg | Neg | SDP | Neg | Neg | SDP | <2 | SDP | SDP | SDP | SDP |
13 | Neg | Neg | SDP | Neg | Neg | SDP | <1 | SDP | Neg | Neg | SDP |
14 | Neg | SDP | SDP | Neg | Neg | SDP | 2–5 | SDP | Neg | SDP | SDP |
15 | Neg | Neg | SDP | Neg | SDP | SDP | <2 | SDP | Neg | SDP | SDP |
16 | Neg | Neg | SDP | Neg | SFP | SDP | <2 | SDP | Neg | Neg | SDP |
17 | SDP | SDP | SDP | Neg | SDP | SDP | <2 | SDP | WFP | Neg | SDP |
18 | ± | ± | SDP | ± | SDP | SDP | 1–2 | SDP | Neg | SDP | SDP |
19 | Neg | Neg | SDP | Neg | Neg | SDP | <2 | SDP | Neg | SDP | SDP |
20 | Neg | ± | SDP | Neg | Neg | SDP | <3 | SDP | Neg | Neg | SDP |
21 | Neg | Neg | SDP | Neg | SDP | SDP | <2 | SDP | SDP | SDP | SDP |
22 | Neg | Neg | SDP | SDP | SDP | SDP | <3 | SDP | Neg | SDP | SDP |
23 | Neg | Neg | SDP | Neg | Neg | SDP | <3 | SDP | Neg | SDP | SDP |
24 | Neg | Neg | SDP | Neg | Neg | SDP | >10 | SDP | Neg | SDP | SDP |
25 | Neg | Neg | SDP | Neg | Neg | SDP | <2 | SDP | Neg | SDP | SDP |
This table shows the immunohistochemical characterisation of the 25 insulinomas in our series. Data on staining were collected by examination of archived slides, re-staining of archived slides and by staining of prospective cases. We have graded staining as negative, weak focal positive, strong focal positive and strong diffuse positive. Staining was scored on two parameters: intensity (strong, detectable at ×40; weak, only detectable at >×100) and distribution (diffuse, >50% of tumour cells; focal, <50% of tumour cells). The labelling index of Ki-67 was evaluated by scoring a minimum of 1000 tumour cells in randomly selected fields. For each case, three different counts were performed and the highest score was chosen as the corresponding index value. Neg, negative; WFP, weak focal positive; SDP, strong diffuse positive; SFP, strong focal positive; ±, equivocal; ERα, oestrogen receptor alpha; ERβ, oestrogen receptor beta; PR, progesterone receptor; Ins, insulin; PP, pancreatic polypeptide; PGP 9.5, protein gene product 9.5; NSE, neuron-specific enolase; chromo, chromogranin; synapto, synaptophysin and somato, somatostatin.
Insulin ELISA
As anticipated insulinoma culture samples secreted several fold more insulin than the control pancreatic tissue. We were able to obtain fresh tissue from two patients in our cohort (patients 23 and 25). In both patients, the highest level of insulin secretion achieved (patient 23: 1621±10 pmol/l, patient 25: 2530±82 pmol/l) was seen in the experiments culturing insulinoma tissue in the presence of E2 alone (Figs 2 and 3). Conversely, in both patients the lowest level of insulin secretion achieved was seen in the experiments culturing control pancreatic tissue in the absence or presence of steroid hormones (patient 23: peak level of 89 pmol/l, patient 25: peak level of 90 pmol/l). The culture of insulinoma tissue with progesterone alone, with oestrogen and progesterone in combination or without steroid hormone(s) led to intermediate levels of insulin secretion. However, these levels were still several folds higher than those from control pancreatic tissue cultures. In patient 23, culture of insulinoma tissue with progesterone alone led to insulin secretion of 889±14 pmol/l, with a combination of oestrogen and progesterone led to insulin secretion of 298±16 pmol/l and without steroid hormone led to insulin secretion of 368±24 pmol/l. In patient 25, culture of insulinoma tissue with progesterone alone led to insulin secretion of 1660±35 pmol/l, with a combination of oestrogen and progesterone led to insulin secretion of 1766±34 pmol/l and without steroid hormone led to insulin secretion of 1420±20 pmol/l.
Insulin PCR
RT-PCR confirmed expression of insulin mRNA in all the samples of control pancreatic tissue and insulinoma tissue. There was relatively higher expression of insulin mRNA in the sample of insulinoma tissue that was cultured in the presence of exogenous E2 (Fig. 4).
Discussion
Insulinoma remains a rare but important medical condition that can be resistant to treatment. The average age of the patients in this series was 50 years, which is similar to that in published series from the Mayo Clinic where the average age was 47 years (Service et al. 1991), the Cleveland Clinic where the average age was 55 years (Dizon et al. 1999) and from the University of Milan where the average age was 49 years (Viale et al. 1992).
Our data confirm that there is ER and PR expression in insulinoma tissue. There was strong expression of ERβ and PR in the two patients from whom we obtained fresh tissue for in vitro culture. Neither ER nor PR was present in normal pancreatic tissue except for some weak expression of ERβ in some of the ‘normal’ margins of pancreatic tissue distant from insulinoma tumour (data not shown). It is evident that fresh insulinoma tissue cultured in the presence or absence of oestrogen and/or progesterone secretes significantly more insulin than control pancreatic tissue in any of these conditions. The highest level of insulinoma-secreted insulin was seen with the addition of oestrogen alone and was evidently consistent in both prospective insulinoma cases. The other culture conditions of insulinoma led to the secretion of lower levels of insulin but the hierarchical order of secretion was not consistent in both patients. In addition, we have demonstrated that there is expression of insulin mRNA in both control pancreatic tissue and insulinoma tissue.
The human ER exists in two forms, ERα and ERβ. Most of the previous studies characterising the ER status of pancreatic tissue have focused on the ‘traditional’ ER, which is now called ERα. The identification and characterisation of ERβ in humans was first reported 10 years ago (Mosselman et al. 1996). Both of the isoforms have been shown to be functional and have different roles in gene regulation as suggested by their contrasting effects on the nuclear transcription factor AP-1 (Paech et al. 1997). Ligand-binding experiments have shown that E2 has the same specific binding affinity for ERβ as it does for ERα (Kuiper et al. 1997). There are a limited number of published studies looking at the expression of ERβ in pancreatic tissue, but there it appears that ERβ has relatively higher expression in pancreatic tissue than ERα. The use of ‘traditional’ ER antibodies with no anti-β activity could explain why ER has previously been scarcely found in pancreatic tumours. ERβ is reported to be relatively more expressed than ERα in papillary cystic neoplasms of the pancreas (Morales et al. 2003), solid pseudopapillary tumours (Geers et al. 2006) and human pancreatic cancer cell lines (Konduri & Schwarz 2007). Our study shows for the first time that there is a considerably high expression of ER in human insulinomas and it is predominantly of the ERβ isoform.
There are two distinct subtypes of the human PR: PR-A and PR-B (Clarke et al. 1987). The anti-PR antibody (clone 636) we used has been demonstrated to react with the PR-A and PR-B forms. We have thus reported the overall level of PR that may comprise both subtypes. There are no data available on the relative expression of the different subtypes of PR in pancreatic tissue, and there are no data available to suggest that any subtype is more predominant. Unlike the case for ER, the single PR antibody employed was immunoreactive for both PR subtypes such that we did not have to interrogate separately with two isoform-specific antibodies. PR subtype tissue expression has been reported in breast (Mote et al. 2002) and endometrial cancers (Saito et al. 2006) where a higher the ratio of PR-A to PR-B affects prognosis. This suggests that both PR isoforms are functional and their expression ratio may have an impact on biological responses with potential important clinical consequences. In our current study, we have shown that there is high expression of PR in human insulinomas but we have not specifically delineated the relative preponderance of the different PR isoforms.
In a series of 156 pancreatic endocrine tumours, Viale et al. characterised 42 insulinomas for expression of oestrogen and PRs. Some differences have emerged between the study by Viale et al. and our study. In their series, they found that the expression of PR was 66% but this was higher in our series with PR being detected in 96% (24/25) of insulinomas. Secondly, in contrast to their study where ER was undetectable in the series of insulinomas, we detected the expression of ER in a significant number of the insulinomas in our study. We found expression of the ERα and ERβ isoforms in 40 and 84% of tumours respectively. Finally, they were able to show that PR expression correlated with the absence of malignant features but our data do not suggest the same. In our series, the expression of PR seen in malignant and benign insulinomas was not significantly different with values of 100% (6/6) and 95% (18/19) respectively. The analysis of our insulinoma dataset did not reveal any distinct patterns of steroid hormone receptor status and tumour behaviour.
There are no reported cases of insulinomas where glycaemic control has been shown to be affected by exogenous oestrogen and progesterone, although in our own series the index case that led to this study showed just such features (data not shown). Our data show that compared with progesterone alone or a combination of oestrogen and progesterone, stimulation with exogenous oestrogen alone causes a more profound increase in insulin secretion from insulinoma tissue in vitro.
The rarity of insulinoma, even in a tertiary referral centre, places unavoidable limitations on studies into this condition. Notwithstanding this limitation, the patterns of receptor expression were consistent and in many cases the high levels of expression allowed easy distinction between insulinoma tissue and adjacent ‘normal’ pancreatic tissue quite apart form the different histological appearance. The data relating to functionality of sex steroid receptor expression and insulin secretion are limited to two patients and in view of this must be treated with some caution although the insulin response to oestrogen in both cases was substantial. Further studies using concomitant receptor blockade will be required to substantiate these findings and are in progress.
Insulinomas express functional ER and PR which increase tumour insulin secretion when stimulated by oestrogen and progesterone in vitro. Hormonal manipulation of insulinomas could thus provide an avenue for the management of the neurogenic and neuroglycopenic symptoms experienced by patients prior to tumour excision or in the event of tumour recurrence.
Acknowledgements
We would like to acknowledge Dr Penelope M Clark (Department of Biochemistry, Queen Elizabeth Hospital, Birmingham, B15 2TH, UK) for her invaluable assistance and guidance with regards to the insulin ELISA, Chris Coldham for maintaining our excellent insulinoma database (Hepatobiliary and Liver Transplant Unit, Queen Elizabeth Hospital, Birmingham, B15 2TH, UK) and Angie Swali for assistance with the insulin PCR (Medical School, University of Birmingham, B15 2TT, UK). The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.
References
Bakkevold KE, Pettersen A, Arnesjo B & Espehaug B 1990 Tamoxifen therapy in unresectable adenocarcinoma of the pancreas and the papilla of Vater. British Journal of Surgery 77 725–730.
Clarke CL, Zaino RJ, Feil PD, Miller JV, Steck ME, Ohlsson-Wilhelm BM & Satyaswaroop PG 1987 Monoclonal antibodies to human progesterone receptor: characterization by biochemical and immunohistochemical techniques. Endocrinology 121 1123–1132.
Dizon AM, Kowalyk S & Hoogwerf BJ 1999 Neuroglycopenic and other symptoms in patients with insulinomas. American Journal of Medicine 106 307–310.
Doglioni C, Gambacorta M, Zamboni G, Coggi G & Viale G 1990 Immunocytochemical localization of progesterone receptors in endocrine cells of the human pancreas. American Journal of Pathology 137 999–1005.
Geers C, Moulin P, Gigot JF, Weynand B, Deprez P, Rahier J & Sempoux C 2006 Solid and pseudopapillary tumor of the pancreas – review and new insights into pathogenesis. American Journal of Surgical Pathology 30 1243–1249.
Greenway B, Iqbal MJ, Johnson PJ & Williams R 1981 Oestrogen receptor proteins in malignant and fetal pancreas. BMJ 283 751–753.
Horimi T, Takasaki M, Toki A, Nishimura W & Morita S 1996 The beneficial effect of tamoxifen therapy in patients with resected adenocarcinoma of the pancreas. Hepatogastroenterology 43 1225–1229.
Keating JJ, Johnson PJ, Cochrane AM, Gazzard BG, Krasner N, Smith PM, Trewby PN, Wheeler P, Wilkinson SP & Williams R 1989 A prospective randomised controlled trial of tamoxifen and cyproterone acetate in pancreatic carcinoma. British Journal of Cancer 60 789–792.
Konduri S & Schwarz RE 2007 Estrogen receptor beta/alpha ratio predicts response of pancreatic cancer cells to estrogens and phytoestrogens. Journal of Surgical Research 140 55–66.
Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S & Gustafsson JA 1997 Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors alpha and beta. Endocrinology 138 863–870.
Lam KY, Lo CY & Fan ST 1999 Pancreatic solid-cystic-papillary tumor: clinicopathologic features in eight patients from Hong Kong and review of the literature. World Journal of Surgery 23 1045–1050.
Morales A, Duarte-Rojo A, Angeles-Angeles A, Mery CM, Ruiz-Molina JM, Diaz-Sanchez V & Robles-Diaz G 2003 The beta form of the estrogen receptor is predominantly expressed in the papillary cystic neoplasm of the pancreas. Pancreas 26 258–263.
Mosselman S, Polman J & Dijkema R 1996 ER beta: identification and characterization of a novel human estrogen receptor. FEBS Letters 392 49–53.
Mote PA, Bartow S, Tran N & Clarke CL 2002 Loss of co-ordinate expression of progesterone receptors A and B is an early event in breast carcinogenesis. Breast Cancer Research and Treatment 72 163–172.
Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J, Kushner PJ & Scanlan TS 1997 Differential ligand activation of estrogen receptors ERα and ERβ at AP1 sites. Science 277 1508–1510.
Reynolds GM, Billingham LJ, Gray LJ, Flavell JR, Najafipour S, Crocker J, Nelson P, Young LS & Murray PG 2002 Interleukin 6 expression by Hodgkin/Reed-Sternberg cells is associated with the presence of ‘B’ symptoms and failure to achieve complete remission in patients with advanced Hodgkin's disease. British Journal of Haematology 118 195–201.
Rindi G, Capella C & Solcia E 1998 Cell biology, clinicopathological profile, and classification of gastro-enteropancreatic endocrine tumors. Journal of Molecular Medicine 76 413–420.
Saito S, Ito K, Nagase S, Suzuki T, Akahira J, Okamura K, Yaegashi N & Sasano H 2006 Progesterone receptor isoforms as a prognostic marker in human endometrial carcinoma. Cancer Science 97 1308–1314.
Schindl M, Kaczirek K, Kaserer K & Niederle B 2000 Is the new classification of neuroendocrine pancreatic tumors of clinical help? World Journal of Surgery 24 1312–1318.
Service FJ, McMahon MM, O'Brien PC & Ballard DJ 1991 Functioning insulinoma – incidence, recurrence, and long-term survival of patients: a 60-year study. Mayo Clinic Proceedings 66 711–719.
Singh S, Baker PR, Poulsom R, Wright NA, Sheppard MC, Langman MJ & Neoptolemos JP 1997 Expression of oestrogen receptor and oestrogen-inducible genes in pancreatic cancer. British Journal of Surgery 84 1085–1089.
Targarona EM, Pons MD, Gonzalez G, Boix L, Marco V & Marco C 1991 Is exocrine pancreatic cancer a hormone-dependent tumor? A study of the existence of sex hormone receptors in normal and neoplastic pancreas. Hepatogastroenterology 38 165–169.
Taylor OM, Benson EA & McMahon MJ 1993 Clinical trial of tamoxifen in patients with irresectable pancreatic adenocarcinoma. The Yorkshire Gastrointestinal Tumour Group. British Journal of Surgery 80 384–386.
Viale G, Doglioni C, Gambacorta M, Zamboni G, Coggi G & Bordi C 1992 Progesterone receptor immunoreactivity in pancreatic endocrine tumors. An immunocytochemical study of 156 neuroendocrine tumors of the pancreas, gastrointestinal and respiratory tracts, and skin. Cancer 70 2268–2277.
Wong A & Chan A 1993 Survival benefit of tamoxifen therapy in adenocarcinoma of pancreas. A case–control study. Cancer 71 2200–2203.