R. V. Jeffreys

Cytokine expression in human anterior pituitary adenomas

OBJECTIVE There is increasing evidence for the role of cytokines in pituitary differentiated function and tumorigenesis, but the spectrum of cytokines found in the pituitary is unknown. Therefore profiles of cytokine expression were determined in different human anterior pituitary adenoma sub‐types.

Effects of insulin-like growth factor-I on growth hormone and prolactin secretion and cell proliferation of human somatotrophinomas and prolactinomas in vitro

OBJECTIVE IGF‐I inhibits GH secretion from normal and some tumorous pituitary tissue, and has been shown to be mitogenic for gonadotrophinoma cells in vitro. It is not known whether IGF‐l affects somatotrophinoma cellular proliferation or the secretion of other hormones, such as PRL and α‐subunit, which are often co‐secreted by these tumours. We have therefore examined the effects of IGF‐l on proliferation and hormonal secretion of human somatotrophinomas and prolactinomas in vitro.

Effect of Cell Density on Hormonal Secretion from Human Pituitary Adenomas in vitro

Cell density effects were investigated on tumorous hormonal secretion from 10 pituitary adenomas: 3 somatotrophinomas secreting GH and PRL; 7 gonadotrophinomas, 3 co-secreted both FSH and LH, all 7 secreted LH. Enzymatically dispersed tissue was plated out in 24-well plates at 5 ×105, 105, 5 ×104 and 104 cells/well in serum-free media. Media were collected weekly for 2 weeks. Results: In 3 of 3 somatotrophinomas, GH and PRL secretion was higher (p < 0.05) at both week 1 and 2 from 104 cells/well, but similar at other cell densities. In all 3 gonadotrophinomas, the FSH secretory rate was highest at 5 ×105 cells/well which fell as cell density decreased. Conversely, in 7 of 7 gonadotrophinomas the LH secretory rate was highest at 104 cells/well (p < 0.01) which fell as cell density increased. Conclusion: These data suggest that paracrine factors may modulate tumorous GH, PRL, FSH and LH secretion, and show that FSH and LH secretion vary inversely as cell density increases.

Human anterior pituitary adenoma cell attachment in vitro

Dear Editor: The purpose of this letter is to inform fellow scientists of a method for the attachment of enzymatically dispersed human anterior pituitary adenoma cells, which often adhere poorly to the surface of cell culture dishes preventing dynamic in vitro studies: this is especially recognized for prolactin secreting and nonfunctioning pituitary adenomas (3,10-12). Attempts to enhance pituitary adenoma attachment have included the use of an extracellular matrix derived from bovine corneal endothelium, which has been reported to be successful for prolactinomas (3) and Cushings adenomas (15). It has also been reported that rat pituitary endocrine cells express laminin and collagen IV, and that these cells grow as small clusters on fibroblast monolayers, which in turn are composed of collagen IV, heparin sulphate, and glycoprotein (13). Others have shown that the extracellular matrix components present among epithelial ceils forming Rathkes pouch consist of laminin, fibronectin, and collagen IV (8). Therefore, we set out to determine the optimal conditions for the attachment of different human anterior pituitary adenoma subtypes in dispersed culture using a panel of attachment factors: an extracellular matrix derived from bovine corneal endothelial cells, collagen type I and IV, fibronectin, laminin, human extracellular matrix (HEM), poly-l-lysine, plastic, and glass. Sixteen human pituitary adenomas--three macroprolactinomas, eight acromegalic adenomas, four nonfunctional tumors, and one corticotrophinoma removed by transsphenoidal surgery--were dispersed to single cells using a combination of a 200 mOsm hypoosmolar medium, clostripain inhibited crude collagenase, and dispase type 2 (1). Adenoma cells were plated at a density of 10 ~ cells per well in plastic 24-well plates with and without attachment factors in Iscoves modified Dulbeccos medium (IMDM) supplemented with 10% fetal calf serum (FCS) and gentamicin (50 rag/l). Cultures were maintained at 37 ° C in a humidity controlled atmosphere of 95% air and 5% CO2. After the cells had been allowed to adhere for 4 d, the total attachment was assessed using a calibrated graticule and viability was determined using fluorescein and propidium iodide (2), following which the wells were gently washed three times in IMDM to remove nonattached cells, and attachment and viability were reassessed. Four replicates were used for each group. Bovine corneal endothelial matrix (BCEM) were prepared from cells 7 d postconfluence: 1-6 d postconfluence gave a BCEM that was too thin, contained multiple holes, and peeled off from the culture surface; 9 d postconfluence the membrane was thicker, but peeled off after preparation. Three methods were compared for BCEM preparation: firstly, BCE cells were washed with phosphate-buffered saline (PBS) and 5 ml of 0.5% Triton X100 (Sigma Chemical Co., St. Louis, MO) (6); secondly, 0.1% Triton X100 ]adapted from (6)]; or thirdly, 20 mmol freshly prepared ammonium hydroxide (5) was added and incubated for 15-20 min and monitored by phase contrast microscopy. When the nuclei and extracellular matrix became visihie, the flasks were washed three times with PBS and once with IMDM. Cells from each tumor were seeded onto plastic plates treated with BCEM, collagen type I and type I¥, fibronectin, laminin, human extracellular matrix, poly-l-lysine, 24 plates alone and new glass coverslips (detailed in Table 1). The resultant attachment (percent) to each surface/attachment factor and cell viability (percent) is shown in Table 1. These studies suggest that the BCEM may be a universal attachment method for human pituitary adenoma cells in vitro, although the technique is limited by its technical complexity and the fragility of the BCEM. All pituitary adenoma subtypes attached and were viable on the BCEM (range 20-96% attachment; viability range 5097%) and the results are in accord with BCEM attachment of prolactinomas and Cushings adenomas reported previously (2,15). The BCEM was superior to collagen type I and IV, fibronectin, laminin, HEM, poly-l-lysine, plastic, and glass. There was no difference between the BCEM preparation methods for the Triton X100 or the sodium hydroxide method (data not shown). The BCEM is composed mainly of type III and type IV collagen (14), but it has been suggested that other factors may also be present (6). The BCEM may also play an active role in pituitary function; prolactin secretion from rat GH3 ceils has been found to be higher on BCEM compared to plastic alone, and cells cultured on plastic secreted more prolactin in response to TRH stimulation than those cultured on BCEM (4,9). Although poly-l-lysine showed good adenoma cell attachment generally, cell viability was markedly diminished compared to the BCEM, suggesting that poly-l-lysine is not suitable for dispersed human anterior pituitary adenoma cell culture. Collagen IV appeared to be superior to collagen I for the attachment of adenoma cells, which may reflect the reports of rat endocrine epithelial cells and Rathkes pouch expressing collagen IV rather than collagen I (8,13). However, neither attached the 3 prolactinomas studied and collagen I attached only 6 of 13 tumors, while collagen IV attached only 12 of 13 tumors; there was no difference in viability between collagen I and collagen IV for attached tumors. The poor attachment of adenoma cel~s to fibronectin (attaching 6 of 13 tumors) and laminin (attaching 5 of 13 tumors) was unexpected because both of these components have been found in the basement membrane in rat endocrine cells and Rathkes pouch (8,13), suggesting that a combination of extracellular components may be required for attachment of pituitary adenoma cells in vitro. No adhesion to the commercial HEM preparation occurred; however, this preparation is composed of a high percentage of laminin to which pituitary cells were seen to attach poorly. Plastic alone attached all adenomas apart from the three prolactinomas and one nonfunctional adenoma; there was no difference in attachment between the three plastic surfaces examined.

How many beds do we really need—for example, in neurosurgery?

I committed two errors. Firstly, I did not consider the possibility of a second lesion seriously enough when I found the ulcer. Secondly, in retrospect, during colonoscopy when I thought I was in the caecum I was in fact in the proximal transverse colon, which commonly hangs down to the pelvis. This mistake was due to lack of experience (I had just started colonoscopy) compounded by not having x-ray screening facilities.